diff options
Diffstat (limited to 'clang')
514 files changed, 19697 insertions, 6759 deletions
diff --git a/clang/bindings/python/clang/cindex.py b/clang/bindings/python/clang/cindex.py index 80140d2..da97bd3 100644 --- a/clang/bindings/python/clang/cindex.py +++ b/clang/bindings/python/clang/cindex.py @@ -3939,6 +3939,8 @@ class Token(Structure): cursor._tu = self._tu conf.lib.clang_annotateTokens(self._tu, byref(self), 1, byref(cursor)) + if cursor.is_null(): + return None return cursor diff --git a/clang/bindings/python/tests/cindex/test_tokens.py b/clang/bindings/python/tests/cindex/test_tokens.py index b6c1fc8..6658579 100644 --- a/clang/bindings/python/tests/cindex/test_tokens.py +++ b/clang/bindings/python/tests/cindex/test_tokens.py @@ -53,3 +53,9 @@ class TestTokens(unittest.TestCase): self.assertEqual(extent.start.offset, 4) self.assertEqual(extent.end.offset, 7) + + def test_null_cursor(self): + """Ensure that the cursor property converts null cursors to None""" + tu = get_tu("int i = 5;") + tokens = list(tu.get_tokens(extent=tu.cursor.extent)) + self.assertEqual(tokens[-1].cursor, None) diff --git a/clang/cmake/modules/CMakeLists.txt b/clang/cmake/modules/CMakeLists.txt index d2d6812..9ad2f98 100644 --- a/clang/cmake/modules/CMakeLists.txt +++ b/clang/cmake/modules/CMakeLists.txt @@ -8,15 +8,19 @@ include(FindPrefixFromConfig) # the usual CMake convention seems to be ${Project}Targets.cmake. set(CLANG_INSTALL_PACKAGE_DIR "${CMAKE_INSTALL_PACKAGEDIR}/clang" CACHE STRING "Path for CMake subdirectory for Clang (defaults to '${CMAKE_INSTALL_PACKAGEDIR}/clang')") -# CMAKE_INSTALL_PACKAGEDIR might be absolute, so don't reuse below. -set(clang_cmake_builddir "${CMAKE_BINARY_DIR}/lib${LLVM_LIBDIR_SUFFIX}/cmake/clang") # Keep this in sync with llvm/cmake/CMakeLists.txt! set(LLVM_INSTALL_PACKAGE_DIR "${CMAKE_INSTALL_PACKAGEDIR}/llvm" CACHE STRING "Path for CMake subdirectory for LLVM (defaults to '${CMAKE_INSTALL_PACKAGEDIR}/llvm')") # CMAKE_INSTALL_PACKAGEDIR might be absolute, so don't reuse below. -string(REPLACE "${CMAKE_CFG_INTDIR}" "." llvm_cmake_builddir "${LLVM_LIBRARY_DIR}") -set(llvm_cmake_builddir "${llvm_cmake_builddir}/cmake/llvm") +string(REPLACE "${CMAKE_CFG_INTDIR}" "." llvm_builddir "${LLVM_LIBRARY_DIR}") +set(llvm_cmake_builddir "${llvm_builddir}/cmake/llvm") +if(CLANG_BUILT_STANDALONE) + # CMAKE_INSTALL_PACKAGEDIR might be absolute, so don't reuse below. + set(clang_cmake_builddir "${CMAKE_BINARY_DIR}/lib${LLVM_LIBDIR_SUFFIX}/cmake/clang") +else() + set(clang_cmake_builddir "${llvm_builddir}/cmake/clang") +endif() get_property(CLANG_EXPORTS GLOBAL PROPERTY CLANG_EXPORTS) export(TARGETS ${CLANG_EXPORTS} FILE ${clang_cmake_builddir}/ClangTargets.cmake) diff --git a/clang/docs/OpenMPSupport.rst b/clang/docs/OpenMPSupport.rst index cdb3b33..5c73e24 100644 --- a/clang/docs/OpenMPSupport.rst +++ b/clang/docs/OpenMPSupport.rst @@ -505,7 +505,7 @@ implementation. +-------------------------------------------------------------+---------------------------+---------------------------+--------------------------------------------------------------------------+
| pure directives in DO CONCURRENT | | :none:`unclaimed` | |
+-------------------------------------------------------------+---------------------------+---------------------------+--------------------------------------------------------------------------+
-| Optional argument for all clauses | :none:`unclaimed` | :none:`unclaimed` | |
+| Optional argument for all clauses | :none:`partial` | :none:`In Progress` | Parse/Sema (nowait): https://github.com/llvm/llvm-project/pull/159628 |
+-------------------------------------------------------------+---------------------------+---------------------------+--------------------------------------------------------------------------+
| Function references for locator list items | :none:`unclaimed` | :none:`unclaimed` | |
+-------------------------------------------------------------+---------------------------+---------------------------+--------------------------------------------------------------------------+
diff --git a/clang/docs/PointerAuthentication.rst b/clang/docs/PointerAuthentication.rst index 96eb498..7e65f4b 100644 --- a/clang/docs/PointerAuthentication.rst +++ b/clang/docs/PointerAuthentication.rst @@ -592,6 +592,36 @@ The result value is never zero and always within range for both the This can be used in constant expressions. +``ptrauth_type_discriminator`` +^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + +.. code-block:: c + + ptrauth_type_discriminator(type) + +Compute the constant discriminator derived from the given type, as is computed +for automatically type diversified schemas. + +``type`` must be a type. The result has the type ``ptrauth_extra_data_t``. + +This can be used in constant expressions. + +``ptrauth_function_pointer_type_discriminator`` +^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + +.. code-block:: c + + ptrauth_function_pointer_type_discriminator(function_type) + +Compute the constant discriminator derived from the provided function type, for +use in contexts where the default function authentication schema. If function +pointer type diversity is enabled, this is equivalent to +`ptrauth_type_discriminator(function_type)`, if it is not enabled this is `0`. + +``function_type`` must be a function type. The result has the type ``ptrauth_extra_data_t``. + +This can be used in constant expressions. + ``ptrauth_strip`` ^^^^^^^^^^^^^^^^^ diff --git a/clang/docs/ReleaseNotes.rst b/clang/docs/ReleaseNotes.rst index 65b086c..8b42468 100644 --- a/clang/docs/ReleaseNotes.rst +++ b/clang/docs/ReleaseNotes.rst @@ -136,6 +136,7 @@ Clang Frontend Potentially Breaking Changes Clang Python Bindings Potentially Breaking Changes -------------------------------------------------- +- Return ``None`` instead of null cursors from ``Token.cursor`` - TypeKind ``ELABORATED`` is not used anymore, per clang AST changes removing ElaboratedTypes. The value becomes unused, and all the existing users should expect the former underlying type to be reported instead. @@ -187,6 +188,11 @@ C2y Feature Support C23 Feature Support ^^^^^^^^^^^^^^^^^^^ +- Added ``FLT_SNAN``, ``DBL_SNAN``, and ``LDBL_SNAN`` to Clang's ``<float.h>`` + header in C23 and later modes. This implements + `WG14 N2710 <https://www.open-std.org/jtc1/sc22/wg14/www/docs/n2710.htm>`_. +- Fixed accepting as compatible unnamed tag types with the same fields within + the same translation unit but from different types. Non-comprehensive list of changes in this release ------------------------------------------------- @@ -265,12 +271,14 @@ Non-comprehensive list of changes in this release allocation functions with a token ID can be enabled via the ``-fsanitize=alloc-token`` flag. +- Clang now rejects the invalid use of ``constexpr`` with ``auto`` and an explicit type in C. (#GH163090) + New Compiler Flags ------------------ - New option ``-fno-sanitize-debug-trap-reasons`` added to disable emitting trap reasons into the debug info when compiling with trapping UBSan (e.g. ``-fsanitize-trap=undefined``). - New option ``-fsanitize-debug-trap-reasons=`` added to control emitting trap reasons into the debug info when compiling with trapping UBSan (e.g. ``-fsanitize-trap=undefined``). - New options for enabling allocation token instrumentation: ``-fsanitize=alloc-token``, ``-falloc-token-max=``, ``-fsanitize-alloc-token-fast-abi``, ``-fsanitize-alloc-token-extended``. - +- The ``-resource-dir`` option is now displayed in the list of options shown by ``--help``. Lanai Support ^^^^^^^^^^^^^^ @@ -282,6 +290,7 @@ Deprecated Compiler Flags Modified Compiler Flags ----------------------- - The `-gkey-instructions` compiler flag is now enabled by default when DWARF is emitted for plain C/C++ and optimizations are enabled. (#GH149509) +- The `-fconstexpr-steps` compiler flag now accepts value `0` to opt out of this limit. (#GH160440) Removed Compiler Flags ------------------------- @@ -297,6 +306,8 @@ Attribute Changes in Clang Improvements to Clang's diagnostics ----------------------------------- +- Diagnostics messages now refer to ``structured binding`` instead of ``decomposition``, + to align with `P0615R0 <https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2017/p0615r0.html>`_ changing the term. (#GH157880) - Added a separate diagnostic group ``-Wfunction-effect-redeclarations``, for the more pedantic diagnostics for function effects (``[[clang::nonblocking]]`` and ``[[clang::nonallocating]]``). Moved the warning for a missing (though implied) attribute on a redeclaration into this group. @@ -405,6 +416,8 @@ Bug Fixes in This Version a function without arguments caused us to try to access a non-existent argument. (#GH159080) - Fixed a failed assertion with empty filename arguments in ``__has_embed``. (#GH159898) +- Fixed a failed assertion with empty filename in ``#embed`` directive. (#GH162951) +- Fixed a crash triggered by unterminated ``__has_embed``. (#GH162953) Bug Fixes to Compiler Builtins ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ @@ -418,7 +431,8 @@ Bug Fixes to Attribute Support (#GH141504) and on types returned from indirect calls (#GH142453). - Fixes some late parsed attributes, when applied to function definitions, not being parsed in function try blocks, and some situations where parsing of the function body - is skipped, such as error recovery and code completion. (#GH153551) + is skipped, such as error recovery, code completion, and msvc-compatible delayed + template parsing. (#GH153551) - Using ``[[gnu::cleanup(some_func)]]`` where some_func is annotated with ``[[gnu::error("some error")]]`` now correctly triggers an error. (#GH146520) - Fix a crash when the function name is empty in the `swift_name` attribute. (#GH157075) @@ -429,6 +443,7 @@ Bug Fixes to C++ Support - Suppress ``-Wdeprecated-declarations`` in implicitly generated functions. (#GH147293) - Fix a crash when deleting a pointer to an incomplete array (#GH150359). - Fixed a mismatched lambda scope bug when propagating up ``consteval`` within nested lambdas. (#GH145776) +- Disallow immediate escalation in destructors. (#GH109096) - Fix an assertion failure when expression in assumption attribute (``[[assume(expr)]]``) creates temporary objects. - Fix the dynamic_cast to final class optimization to correctly handle @@ -506,6 +521,8 @@ X86 Support driver. - Remove `[no-]evex512` feature request from intrinsics and builtins. - Change features `avx10.x-[256,512]` to `avx10.x`. +- `-march=wildcatlake` is now supported. +- `-march=novalake` is now supported. Arm and AArch64 Support ^^^^^^^^^^^^^^^^^^^^^^^ @@ -635,6 +652,8 @@ OpenMP Support - Added support for ``defaultmap`` directive implicit-behavior ``private``. - Added parsing and semantic analysis support for ``groupprivate`` directive. - Added support for 'omp fuse' directive. +- Updated parsing and semantic analysis support for ``nowait`` clause to accept + optional argument in OpenMP >= 60. Improvements ^^^^^^^^^^^^ diff --git a/clang/docs/UsersManual.rst b/clang/docs/UsersManual.rst index 5745e4b..e82b16f 100644 --- a/clang/docs/UsersManual.rst +++ b/clang/docs/UsersManual.rst @@ -2325,7 +2325,7 @@ are listed below. devirtualization and virtual constant propagation, for classes with :doc:`hidden LTO visibility <LTOVisibility>`. Requires ``-flto``. -.. option:: -f[no]split-lto-unit +.. option:: -f[no-]split-lto-unit Controls splitting the :doc:`LTO unit <LTOVisibility>` into regular LTO and :doc:`ThinLTO` portions, when compiling with -flto=thin. Defaults to false @@ -2518,7 +2518,7 @@ are listed below. .. _funique_internal_linkage_names: -.. option:: -f[no]-unique-internal-linkage-names +.. option:: -f[no-]unique-internal-linkage-names Controls whether Clang emits a unique (best-effort) symbol name for internal linkage symbols. When this option is set, compiler hashes the main source @@ -2539,7 +2539,7 @@ are listed below. $ cd $P/bar && clang -c -funique-internal-linkage-names name_conflict.c $ cd $P && clang foo/name_conflict.o && bar/name_conflict.o -.. option:: -f[no]-basic-block-address-map: +.. option:: -f[no-]basic-block-address-map: Emits a ``SHT_LLVM_BB_ADDR_MAP`` section which includes address offsets for each basic block in the program, relative to the parent function address. @@ -2780,6 +2780,25 @@ usual build cycle when using sample profilers for optimization: /clang:-fdebug-info-for-profiling /clang:-funique-internal-linkage-names ^ code.cc /Fe:code -fuse-ld=lld /link /debug:dwarf + [OPTIONAL] Pseudo instrumentation can be used as the anchor for accurate + profile mapping with the ``-fpseudo-probe-for-profiling`` option. + + On Linux: + + .. code-block:: console + + $ clang++ -O2 -gline-tables-only \ + -fpseudo-probe-for-profiling -funique-internal-linkage-names \ + code.cc -o code + + On Windows: + + .. code-block:: winbatch + + > clang-cl /O2 -gdwarf -gline-tables-only ^ + -fpseudo-probe-for-profiling /clang:-funique-internal-linkage-names ^ + code.cc /Fe:code -fuse-ld=lld /link /debug:dwarf + .. note:: :ref:`-funique-internal-linkage-names <funique_internal_linkage_names>` @@ -2881,6 +2900,25 @@ usual build cycle when using sample profilers for optimization: /clang:-fdebug-info-for-profiling /clang:-funique-internal-linkage-names ^ -fprofile-sample-use=code.prof code.cc /Fe:code + [OPTIONAL] Pseudo instrumentation can be used as the anchor for accurate + profile mapping with the ``-fpseudo-probe-for-profiling`` option. + + On Linux: + + .. code-block:: console + + $ clang++ -O2 \ + -fpseudo-probe-for-profiling -funique-internal-linkage-names \ + -fprofile-sample-use=code.prof code.cc -o code + + On Windows: + + .. code-block:: winbatch + + > clang-cl /O2 ^ + -fpseudo-probe-for-profiling /clang:-funique-internal-linkage-names ^ + -fprofile-sample-use=code.prof code.cc /Fe:code + [OPTIONAL] Sampling-based profiles can have inaccuracies or missing block/ edge counters. The profile inference algorithm (profi) can be used to infer missing blocks and edge counts, and improve the quality of profile data. @@ -4036,7 +4074,7 @@ Controlling implementation limits Sets the limit for the number of full-expressions evaluated in a single constant expression evaluation. This also controls the maximum size of array and dynamic array allocation that can be constant evaluated. - The default is 1048576. + The default is 1048576, and the limit can be disabled with `-fconstexpr-steps=0`. .. option:: -ftemplate-depth=N diff --git a/clang/docs/analyzer/checkers.rst b/clang/docs/analyzer/checkers.rst index d942578..fd0b304 100644 --- a/clang/docs/analyzer/checkers.rst +++ b/clang/docs/analyzer/checkers.rst @@ -205,6 +205,50 @@ pointers with a specified address space. If the option is set to false, then reports from the specific x86 address spaces 256, 257 and 258 are still suppressed, but null dereferences from other address spaces are reported. +.. _core-NullPointerArithm: + +core.NullPointerArithm (C, C++) +""""""""""""""""""""""""""""""" +Check for undefined arithmetic operations with null pointers. + +The checker can detect the following cases: + + - ``p + x`` and ``x + p`` where ``p`` is a null pointer and ``x`` is a nonzero + integer value. + - ``p - x`` where ``p`` is a null pointer and ``x`` is a nonzero integer + value. + - ``p1 - p2`` where one of ``p1`` and ``p2`` is null and the other a + non-null pointer. + +Result of these operations is undefined according to the standard. +In the above listed cases, the checker will warn even if the expression +described to be "nonzero" or "non-null" has unknown value, because it is likely +that it can have non-zero value during the program execution. + +.. code-block:: c + + void test1(int *p, int offset) { + if (p) + return; + + int *p1 = p + offset; // warn: 'p' is null, 'offset' is unknown but likely non-zero + } + + void test2(int *p, int offset) { + if (p) { } // this indicates that it is possible for 'p' to be null + if (offset == 0) + return; + + int *p1 = p - offset; // warn: 'p' is null, 'offset' is known to be non-zero + } + + void test3(char *p1, char *p2) { + if (p1) + return; + + int a = p1 - p2; // warn: 'p1' is null, 'p2' can be likely non-null + } + .. _core-StackAddressEscape: core.StackAddressEscape (C) @@ -3421,12 +3465,6 @@ Check for an out-of-bound pointer being returned to callers. return x; // warn: undefined or garbage returned } - -alpha.security.cert -^^^^^^^^^^^^^^^^^^^ - -SEI CERT checkers which tries to find errors based on their `C coding rules <https://wiki.sei.cmu.edu/confluence/display/c/2+Rules>`_. - alpha.unix ^^^^^^^^^^ diff --git a/clang/docs/analyzer/developer-docs/Statistics.rst b/clang/docs/analyzer/developer-docs/Statistics.rst index 4f2484a..355759d 100644 --- a/clang/docs/analyzer/developer-docs/Statistics.rst +++ b/clang/docs/analyzer/developer-docs/Statistics.rst @@ -22,7 +22,7 @@ However, note that with ``LLVM_ENABLE_STATS`` disabled, only storage of the valu If you want to define a statistic only for entry point, EntryPointStats.h has four classes at your disposal: -- ``UnsignedEPStat`` - an unsigned value assigned at most once per entry point. For example: "the number of source characters in an entry-point body". +- ``UnsignedEPStat`` - an unsigned value assigned at most once per entry point. For example: "the number of source characters in an entry-point body". If no value is assigned during analysis of an entry point, the corresponding CSV cell will be empty. - ``CounterEPStat`` - an additive statistic. It starts with 0 and you can add to it as many times as needed. For example: "the number of bugs discovered". - ``UnsignedMaxEPStat`` - a maximizing statistic. It starts with 0 and when you join it with a value, it picks the maximum of the previous value and the new one. For example, "the longest execution path of a bug". diff --git a/clang/include/clang/AST/ASTContext.h b/clang/include/clang/AST/ASTContext.h index 78220d4..33aa2d3 100644 --- a/clang/include/clang/AST/ASTContext.h +++ b/clang/include/clang/AST/ASTContext.h @@ -2874,11 +2874,11 @@ public: /// returned type is guaranteed to be free of any of these, allowing two /// canonical types to be compared for exact equality with a simple pointer /// comparison. - CanQualType getCanonicalType(QualType T) const { + static CanQualType getCanonicalType(QualType T) { return CanQualType::CreateUnsafe(T.getCanonicalType()); } - const Type *getCanonicalType(const Type *T) const { + static const Type *getCanonicalType(const Type *T) { return T->getCanonicalTypeInternal().getTypePtr(); } @@ -2890,10 +2890,10 @@ public: CanQualType getCanonicalParamType(QualType T) const; /// Determine whether the given types \p T1 and \p T2 are equivalent. - bool hasSameType(QualType T1, QualType T2) const { + static bool hasSameType(QualType T1, QualType T2) { return getCanonicalType(T1) == getCanonicalType(T2); } - bool hasSameType(const Type *T1, const Type *T2) const { + static bool hasSameType(const Type *T1, const Type *T2) { return getCanonicalType(T1) == getCanonicalType(T2); } @@ -2921,7 +2921,7 @@ public: /// Determine whether the given types are equivalent after /// cvr-qualifiers have been removed. - bool hasSameUnqualifiedType(QualType T1, QualType T2) const { + static bool hasSameUnqualifiedType(QualType T1, QualType T2) { return getCanonicalType(T1).getTypePtr() == getCanonicalType(T2).getTypePtr(); } diff --git a/clang/include/clang/AST/ASTNodeTraverser.h b/clang/include/clang/AST/ASTNodeTraverser.h index 0921604..e74bb72 100644 --- a/clang/include/clang/AST/ASTNodeTraverser.h +++ b/clang/include/clang/AST/ASTNodeTraverser.h @@ -770,7 +770,7 @@ public: // it will not be in the parent context: if (auto *TT = D->getFriendType()->getType()->getAs<TagType>()) if (TT->isTagOwned()) - Visit(TT->getOriginalDecl()); + Visit(TT->getDecl()); } else { Visit(D->getFriendDecl()); } diff --git a/clang/include/clang/AST/CanonicalType.h b/clang/include/clang/AST/CanonicalType.h index b5a4e94e13..87bbd7b 100644 --- a/clang/include/clang/AST/CanonicalType.h +++ b/clang/include/clang/AST/CanonicalType.h @@ -551,18 +551,18 @@ struct CanProxyAdaptor<UnaryTransformType> template<> struct CanProxyAdaptor<TagType> : public CanProxyBase<TagType> { - LLVM_CLANG_CANPROXY_SIMPLE_ACCESSOR(TagDecl *, getOriginalDecl) + LLVM_CLANG_CANPROXY_SIMPLE_ACCESSOR(TagDecl *, getDecl) }; template<> struct CanProxyAdaptor<RecordType> : public CanProxyBase<RecordType> { - LLVM_CLANG_CANPROXY_SIMPLE_ACCESSOR(RecordDecl *, getOriginalDecl) + LLVM_CLANG_CANPROXY_SIMPLE_ACCESSOR(RecordDecl *, getDecl) LLVM_CLANG_CANPROXY_SIMPLE_ACCESSOR(bool, hasConstFields) }; template<> struct CanProxyAdaptor<EnumType> : public CanProxyBase<EnumType> { - LLVM_CLANG_CANPROXY_SIMPLE_ACCESSOR(EnumDecl *, getOriginalDecl) + LLVM_CLANG_CANPROXY_SIMPLE_ACCESSOR(EnumDecl *, getDecl) }; template<> diff --git a/clang/include/clang/AST/DeclCXX.h b/clang/include/clang/AST/DeclCXX.h index 898487b..dfa3bef 100644 --- a/clang/include/clang/AST/DeclCXX.h +++ b/clang/include/clang/AST/DeclCXX.h @@ -3832,7 +3832,7 @@ public: public: EnumDecl *getEnumDecl() const { - return EnumType->getType()->castAs<clang::EnumType>()->getOriginalDecl(); + return EnumType->getType()->castAs<clang::EnumType>()->getDecl(); } static UsingEnumDecl *Create(ASTContext &C, DeclContext *DC, diff --git a/clang/include/clang/AST/NestedNameSpecifierBase.h b/clang/include/clang/AST/NestedNameSpecifierBase.h index 73c60ba..8f4bbe9 100644 --- a/clang/include/clang/AST/NestedNameSpecifierBase.h +++ b/clang/include/clang/AST/NestedNameSpecifierBase.h @@ -361,6 +361,9 @@ public: /// Retrieve the source range covering just the last part of /// this nested-name-specifier, not including the prefix. /// + /// Note that this is the source range of this NestedNameSpecifier chunk, + /// and for a type this includes the prefix of that type. + /// /// For example, if this instance refers to a nested-name-specifier /// \c \::std::vector<int>::, the returned source range would cover /// from "vector" to the last '::'. diff --git a/clang/include/clang/AST/OpenACCClause.h b/clang/include/clang/AST/OpenACCClause.h index 79cffeb..83f2b18 100644 --- a/clang/include/clang/AST/OpenACCClause.h +++ b/clang/include/clang/AST/OpenACCClause.h @@ -1277,7 +1277,7 @@ public: }; // A structure to stand in for the recipe on a reduction. RecipeDecl is the -// 'main' declaration used for initializaiton, which is fixed. +// 'main' declaration used for initializaiton, which is fixed. struct OpenACCReductionRecipe { VarDecl *AllocaDecl; @@ -1297,36 +1297,72 @@ struct OpenACCReductionRecipe { // -For a struct without the operator, this will be 1 element per field, which // should be the combiner for that element. // -For an array of any of the above, it will be the above for the element. - llvm::SmallVector<CombinerRecipe, 1> CombinerRecipes; + // Note: These are necessarily stored in either Trailing Storage (when in the + // AST), or in a separate collection when being semantically analyzed. + llvm::ArrayRef<CombinerRecipe> CombinerRecipes; + bool isSet() const { return AllocaDecl; } + +private: + friend class OpenACCReductionClause; OpenACCReductionRecipe(VarDecl *A, llvm::ArrayRef<CombinerRecipe> Combiners) : AllocaDecl(A), CombinerRecipes(Combiners) {} +}; - bool isSet() const { return AllocaDecl; } - static OpenACCReductionRecipe Empty() { - return OpenACCReductionRecipe(/*AllocaDecl=*/nullptr, {}); +// A version of the above that is used for semantic analysis, at a time before +// the OpenACCReductionClause node has been created. This one has storage for +// the CombinerRecipe, since Trailing storage for it doesn't exist yet. +struct OpenACCReductionRecipeWithStorage { + VarDecl *AllocaDecl; + llvm::SmallVector<OpenACCReductionRecipe::CombinerRecipe, 1> CombinerRecipes; + + OpenACCReductionRecipeWithStorage( + VarDecl *A, + llvm::ArrayRef<OpenACCReductionRecipe::CombinerRecipe> Combiners) + : AllocaDecl(A), CombinerRecipes(Combiners) {} + + static OpenACCReductionRecipeWithStorage Empty() { + return OpenACCReductionRecipeWithStorage(/*AllocaDecl=*/nullptr, {}); } }; class OpenACCReductionClause final : public OpenACCClauseWithVarList, private llvm::TrailingObjects<OpenACCReductionClause, Expr *, - OpenACCReductionRecipe> { + OpenACCReductionRecipe, + OpenACCReductionRecipe::CombinerRecipe> { friend TrailingObjects; OpenACCReductionOperator Op; OpenACCReductionClause(SourceLocation BeginLoc, SourceLocation LParenLoc, OpenACCReductionOperator Operator, ArrayRef<Expr *> VarList, - ArrayRef<OpenACCReductionRecipe> Recipes, + ArrayRef<OpenACCReductionRecipeWithStorage> Recipes, SourceLocation EndLoc) : OpenACCClauseWithVarList(OpenACCClauseKind::Reduction, BeginLoc, LParenLoc, EndLoc), Op(Operator) { - assert(VarList.size() == Recipes.size()); + assert(VarList.size() == Recipes.size()); setExprs(getTrailingObjects<Expr *>(VarList.size()), VarList); - llvm::uninitialized_copy(Recipes, getTrailingObjects< - OpenACCReductionRecipe > ()); + + // Since we're using trailing storage on this node to store the 'combiner' + // recipes of the Reduction Recipes (which have a 1:M relationship), we need + // to ensure we get the ArrayRef of each of our combiner 'correct'. + OpenACCReductionRecipe::CombinerRecipe *CurCombinerLoc = + getTrailingObjects<OpenACCReductionRecipe::CombinerRecipe>(); + for (const auto &[Idx, R] : llvm::enumerate(Recipes)) { + + // ArrayRef to the 'correct' data location in trailing storage. + llvm::MutableArrayRef<OpenACCReductionRecipe::CombinerRecipe> + NewCombiners{CurCombinerLoc, R.CombinerRecipes.size()}; + CurCombinerLoc += R.CombinerRecipes.size(); + + llvm::uninitialized_copy(R.CombinerRecipes, NewCombiners.begin()); + + // Placement new into the correct location in trailng storage. + new (&getTrailingObjects<OpenACCReductionRecipe>()[Idx]) + OpenACCReductionRecipe(R.AllocaDecl, NewCombiners); + } } public: @@ -1347,13 +1383,17 @@ public: static OpenACCReductionClause * Create(const ASTContext &C, SourceLocation BeginLoc, SourceLocation LParenLoc, OpenACCReductionOperator Operator, ArrayRef<Expr *> VarList, - ArrayRef<OpenACCReductionRecipe> Recipes, SourceLocation EndLoc); + ArrayRef<OpenACCReductionRecipeWithStorage> Recipes, + SourceLocation EndLoc); OpenACCReductionOperator getReductionOp() const { return Op; } size_t numTrailingObjects(OverloadToken<Expr *>) const { return getExprs().size(); } + size_t numTrailingObjects(OverloadToken<OpenACCReductionRecipe>) const { + return getExprs().size(); + } }; class OpenACCLinkClause final diff --git a/clang/include/clang/AST/OpenMPClause.h b/clang/include/clang/AST/OpenMPClause.h index 68d220a..bc791e4 100644 --- a/clang/include/clang/AST/OpenMPClause.h +++ b/clang/include/clang/AST/OpenMPClause.h @@ -2291,18 +2291,68 @@ public: /// This represents 'nowait' clause in the '#pragma omp ...' directive. /// /// \code -/// #pragma omp for nowait +/// #pragma omp for nowait (cond) /// \endcode -/// In this example directive '#pragma omp for' has 'nowait' clause. -class OMPNowaitClause final : public OMPNoChildClause<llvm::omp::OMPC_nowait> { +/// In this example directive '#pragma omp for' has simple 'nowait' clause with +/// condition 'cond'. +class OMPNowaitClause final : public OMPClause { + friend class OMPClauseReader; + + /// Location of '('. + SourceLocation LParenLoc; + + /// Condition of the 'nowait' clause. + Stmt *Condition = nullptr; + + /// Set condition. + void setCondition(Expr *Cond) { Condition = Cond; } + public: - /// Build 'nowait' clause. + /// Build 'nowait' clause with condition \a Cond. /// + /// \param Cond Condition of the clause. /// \param StartLoc Starting location of the clause. + /// \param LParenLoc Location of '('. /// \param EndLoc Ending location of the clause. - OMPNowaitClause(SourceLocation StartLoc = SourceLocation(), - SourceLocation EndLoc = SourceLocation()) - : OMPNoChildClause(StartLoc, EndLoc) {} + OMPNowaitClause(Expr *Cond, SourceLocation StartLoc, SourceLocation LParenLoc, + SourceLocation EndLoc) + : OMPClause(llvm::omp::OMPC_nowait, StartLoc, EndLoc), + LParenLoc(LParenLoc), Condition(Cond) {} + + /// Build an empty clause. + OMPNowaitClause() + : OMPClause(llvm::omp::OMPC_nowait, SourceLocation(), SourceLocation()) {} + + /// Sets the location of '('. + void setLParenLoc(SourceLocation Loc) { LParenLoc = Loc; } + + /// Returns the location of '('. + SourceLocation getLParenLoc() const { return LParenLoc; } + + /// Returns condition. + Expr *getCondition() const { return cast_or_null<Expr>(Condition); } + + child_range children() { + if (Condition) + return child_range(&Condition, &Condition + 1); + return child_range(child_iterator(), child_iterator()); + } + + const_child_range children() const { + if (Condition) + return const_child_range(&Condition, &Condition + 1); + return const_child_range(const_child_iterator(), const_child_iterator()); + } + + child_range used_children(); + const_child_range used_children() const { + auto Children = const_cast<OMPNowaitClause *>(this)->used_children(); + return const_child_range(Children.begin(), Children.end()); + } + + static bool classof(const OMPClause *T) { + return T->getClauseKind() == llvm::omp::OMPC_nowait; + } }; /// This represents 'untied' clause in the '#pragma omp ...' directive. diff --git a/clang/include/clang/AST/RecursiveASTVisitor.h b/clang/include/clang/AST/RecursiveASTVisitor.h index 7a2881f..32b2b6b 100644 --- a/clang/include/clang/AST/RecursiveASTVisitor.h +++ b/clang/include/clang/AST/RecursiveASTVisitor.h @@ -1710,7 +1710,7 @@ DEF_TRAVERSE_DECL(FriendDecl, { // it will not be in the parent context: if (auto *TT = D->getFriendType()->getType()->getAs<TagType>(); TT && TT->isTagOwned()) - TRY_TO(TraverseDecl(TT->getOriginalDecl())); + TRY_TO(TraverseDecl(TT->getDecl())); } else { TRY_TO(TraverseDecl(D->getFriendDecl())); } @@ -3594,7 +3594,8 @@ bool RecursiveASTVisitor<Derived>::VisitOMPOrderedClause(OMPOrderedClause *C) { } template <typename Derived> -bool RecursiveASTVisitor<Derived>::VisitOMPNowaitClause(OMPNowaitClause *) { +bool RecursiveASTVisitor<Derived>::VisitOMPNowaitClause(OMPNowaitClause *C) { + TRY_TO(TraverseStmt(C->getCondition())); return true; } diff --git a/clang/include/clang/AST/Type.h b/clang/include/clang/AST/Type.h index df106d5..7bd2441 100644 --- a/clang/include/clang/AST/Type.h +++ b/clang/include/clang/AST/Type.h @@ -27,7 +27,7 @@ inline CXXRecordDecl *Type::getAsCXXRecordDecl() const { const auto *TT = dyn_cast<TagType>(CanonicalType); if (!isa_and_present<RecordType, InjectedClassNameType>(TT)) return nullptr; - auto *TD = TT->getOriginalDecl(); + auto *TD = TT->getDecl(); if (isa<RecordType>(TT) && !isa<CXXRecordDecl>(TD)) return nullptr; return cast<CXXRecordDecl>(TD)->getDefinitionOrSelf(); @@ -35,41 +35,39 @@ inline CXXRecordDecl *Type::getAsCXXRecordDecl() const { inline CXXRecordDecl *Type::castAsCXXRecordDecl() const { const auto *TT = cast<TagType>(CanonicalType); - return cast<CXXRecordDecl>(TT->getOriginalDecl())->getDefinitionOrSelf(); + return cast<CXXRecordDecl>(TT->getDecl())->getDefinitionOrSelf(); } inline RecordDecl *Type::getAsRecordDecl() const { const auto *TT = dyn_cast<TagType>(CanonicalType); if (!isa_and_present<RecordType, InjectedClassNameType>(TT)) return nullptr; - return cast<RecordDecl>(TT->getOriginalDecl())->getDefinitionOrSelf(); + return cast<RecordDecl>(TT->getDecl())->getDefinitionOrSelf(); } inline RecordDecl *Type::castAsRecordDecl() const { const auto *TT = cast<TagType>(CanonicalType); - return cast<RecordDecl>(TT->getOriginalDecl())->getDefinitionOrSelf(); + return cast<RecordDecl>(TT->getDecl())->getDefinitionOrSelf(); } inline EnumDecl *Type::getAsEnumDecl() const { if (const auto *TT = dyn_cast<EnumType>(CanonicalType)) - return TT->getOriginalDecl()->getDefinitionOrSelf(); + return TT->getDecl()->getDefinitionOrSelf(); return nullptr; } inline EnumDecl *Type::castAsEnumDecl() const { - return cast<EnumType>(CanonicalType) - ->getOriginalDecl() - ->getDefinitionOrSelf(); + return cast<EnumType>(CanonicalType)->getDecl()->getDefinitionOrSelf(); } inline TagDecl *Type::getAsTagDecl() const { if (const auto *TT = dyn_cast<TagType>(CanonicalType)) - return TT->getOriginalDecl()->getDefinitionOrSelf(); + return TT->getDecl()->getDefinitionOrSelf(); return nullptr; } inline TagDecl *Type::castAsTagDecl() const { - return cast<TagType>(CanonicalType)->getOriginalDecl()->getDefinitionOrSelf(); + return cast<TagType>(CanonicalType)->getDecl()->getDefinitionOrSelf(); } inline bool QualType::hasNonTrivialToPrimitiveDefaultInitializeCUnion() const { diff --git a/clang/include/clang/AST/TypeBase.h b/clang/include/clang/AST/TypeBase.h index 625cc77..f07861f 100644 --- a/clang/include/clang/AST/TypeBase.h +++ b/clang/include/clang/AST/TypeBase.h @@ -4378,8 +4378,6 @@ protected: unsigned NumRows; unsigned NumColumns; - static constexpr unsigned MaxElementsPerDimension = (1 << 20) - 1; - ConstantMatrixType(QualType MatrixElementType, unsigned NRows, unsigned NColumns, QualType CanonElementType); @@ -4398,16 +4396,6 @@ public: return getNumRows() * getNumColumns(); } - /// Returns true if \p NumElements is a valid matrix dimension. - static constexpr bool isDimensionValid(size_t NumElements) { - return NumElements > 0 && NumElements <= MaxElementsPerDimension; - } - - /// Returns the maximum number of elements per dimension. - static constexpr unsigned getMaxElementsPerDimension() { - return MaxElementsPerDimension; - } - void Profile(llvm::FoldingSetNodeID &ID) { Profile(ID, getElementType(), getNumRows(), getNumColumns(), getTypeClass()); @@ -6419,10 +6407,10 @@ protected: bool IsInjected, const Type *CanonicalType); public: - // FIXME: Temporarily renamed from `getDecl` in order to facilitate - // rebasing, due to change in behaviour. This should be renamed back - // to `getDecl` once the change is settled. - TagDecl *getOriginalDecl() const { return decl; } + TagDecl *getDecl() const { return decl; } + [[deprecated("Use getDecl instead")]] TagDecl *getOriginalDecl() const { + return decl; + } NestedNameSpecifier getQualifier() const; @@ -6463,7 +6451,7 @@ struct TagTypeFoldingSetPlaceholder : public llvm::FoldingSetNode { void Profile(llvm::FoldingSetNodeID &ID) const { const TagType *T = getTagType(); - Profile(ID, T->getKeyword(), T->getQualifier(), T->getOriginalDecl(), + Profile(ID, T->getKeyword(), T->getQualifier(), T->getDecl(), T->isTagOwned(), T->isInjected()); } @@ -6487,11 +6475,11 @@ class RecordType final : public TagType { using TagType::TagType; public: - // FIXME: Temporarily renamed from `getDecl` in order to facilitate - // rebasing, due to change in behaviour. This should be renamed back - // to `getDecl` once the change is settled. - RecordDecl *getOriginalDecl() const { - return reinterpret_cast<RecordDecl *>(TagType::getOriginalDecl()); + RecordDecl *getDecl() const { + return reinterpret_cast<RecordDecl *>(TagType::getDecl()); + } + [[deprecated("Use getDecl instead")]] RecordDecl *getOriginalDecl() const { + return getDecl(); } /// Recursively check all fields in the record for const-ness. If any field @@ -6507,11 +6495,11 @@ class EnumType final : public TagType { using TagType::TagType; public: - // FIXME: Temporarily renamed from `getDecl` in order to facilitate - // rebasing, due to change in behaviour. This should be renamed back - // to `getDecl` once the change is settled. - EnumDecl *getOriginalDecl() const { - return reinterpret_cast<EnumDecl *>(TagType::getOriginalDecl()); + EnumDecl *getDecl() const { + return reinterpret_cast<EnumDecl *>(TagType::getDecl()); + } + [[deprecated("Use getDecl instead")]] EnumDecl *getOriginalDecl() const { + return getDecl(); } static bool classof(const Type *T) { return T->getTypeClass() == Enum; } @@ -6542,11 +6530,11 @@ class InjectedClassNameType final : public TagType { bool IsInjected, const Type *CanonicalType); public: - // FIXME: Temporarily renamed from `getDecl` in order to facilitate - // rebasing, due to change in behaviour. This should be renamed back - // to `getDecl` once the change is settled. - CXXRecordDecl *getOriginalDecl() const { - return reinterpret_cast<CXXRecordDecl *>(TagType::getOriginalDecl()); + CXXRecordDecl *getDecl() const { + return reinterpret_cast<CXXRecordDecl *>(TagType::getDecl()); + } + [[deprecated("Use getDecl instead")]] CXXRecordDecl *getOriginalDecl() const { + return getDecl(); } static bool classof(const Type *T) { @@ -8930,8 +8918,8 @@ inline bool Type::isIntegerType() const { if (const EnumType *ET = dyn_cast<EnumType>(CanonicalType)) { // Incomplete enum types are not treated as integer types. // FIXME: In C++, enum types are never integer types. - return IsEnumDeclComplete(ET->getOriginalDecl()) && - !IsEnumDeclScoped(ET->getOriginalDecl()); + return IsEnumDeclComplete(ET->getDecl()) && + !IsEnumDeclScoped(ET->getDecl()); } return isBitIntType(); } @@ -8989,7 +8977,7 @@ inline bool Type::isScalarType() const { if (const EnumType *ET = dyn_cast<EnumType>(CanonicalType)) // Enums are scalar types, but only if they are defined. Incomplete enums // are not treated as scalar types. - return IsEnumDeclComplete(ET->getOriginalDecl()); + return IsEnumDeclComplete(ET->getDecl()); return isa<PointerType>(CanonicalType) || isa<BlockPointerType>(CanonicalType) || isa<MemberPointerType>(CanonicalType) || @@ -9005,7 +8993,7 @@ inline bool Type::isIntegralOrEnumerationType() const { // Check for a complete enum type; incomplete enum types are not properly an // enumeration type in the sense required here. if (const auto *ET = dyn_cast<EnumType>(CanonicalType)) - return IsEnumDeclComplete(ET->getOriginalDecl()); + return IsEnumDeclComplete(ET->getDecl()); return isBitIntType(); } diff --git a/clang/include/clang/AST/TypeLoc.h b/clang/include/clang/AST/TypeLoc.h index 38e8fba..2cefaa9 100644 --- a/clang/include/clang/AST/TypeLoc.h +++ b/clang/include/clang/AST/TypeLoc.h @@ -199,11 +199,6 @@ public: NestedNameSpecifierLoc getPrefix() const; /// This returns the position of the type after any elaboration, such as the - /// 'struct' keyword, and name qualifiers. This will the 'template' keyword if - /// present, or the name location otherwise. - SourceLocation getNonPrefixBeginLoc() const; - - /// This returns the position of the type after any elaboration, such as the /// 'struct' keyword. This may be the position of the name qualifiers, /// 'template' keyword, or the name location otherwise. SourceLocation getNonElaboratedBeginLoc() const; @@ -798,7 +793,7 @@ struct TagTypeLocInfo { class TagTypeLoc : public ConcreteTypeLoc<UnqualTypeLoc, TagTypeLoc, TagType, TagTypeLocInfo> { public: - TagDecl *getOriginalDecl() const { return getTypePtr()->getOriginalDecl(); } + TagDecl *getDecl() const { return getTypePtr()->getDecl(); } /// True if the tag was defined in this type specifier. bool isDefinition() const; @@ -859,9 +854,7 @@ class RecordTypeLoc : public InheritingConcreteTypeLoc<TagTypeLoc, RecordTypeLoc, RecordType> { public: - RecordDecl *getOriginalDecl() const { - return getTypePtr()->getOriginalDecl(); - } + RecordDecl *getDecl() const { return getTypePtr()->getDecl(); } }; /// Wrapper for source info for enum types. @@ -869,7 +862,7 @@ class EnumTypeLoc : public InheritingConcreteTypeLoc<TagTypeLoc, EnumTypeLoc, EnumType> { public: - EnumDecl *getOriginalDecl() const { return getTypePtr()->getOriginalDecl(); } + EnumDecl *getDecl() const { return getTypePtr()->getDecl(); } }; /// Wrapper for source info for injected class names of class @@ -878,9 +871,7 @@ class InjectedClassNameTypeLoc : public InheritingConcreteTypeLoc<TagTypeLoc, InjectedClassNameTypeLoc, InjectedClassNameType> { public: - CXXRecordDecl *getOriginalDecl() const { - return getTypePtr()->getOriginalDecl(); - } + CXXRecordDecl *getDecl() const { return getTypePtr()->getDecl(); } }; /// Wrapper for template type parameters. diff --git a/clang/include/clang/AST/TypeProperties.td b/clang/include/clang/AST/TypeProperties.td index 9dc85fb..03613d5 100644 --- a/clang/include/clang/AST/TypeProperties.td +++ b/clang/include/clang/AST/TypeProperties.td @@ -575,7 +575,7 @@ let Class = TagType in { let Conditional = [{ !IsCanonical }]; let Read = [{ node->getQualifier() }]; } - def : Property<"TD", TagDeclRef> { let Read = [{ node->getOriginalDecl() }]; } + def : Property<"TD", TagDeclRef> { let Read = [{ node->getDecl() }]; } } let Class = EnumType in { diff --git a/clang/include/clang/ASTMatchers/ASTMatchersInternal.h b/clang/include/clang/ASTMatchers/ASTMatchersInternal.h index 1ab6f11..c050fb7 100644 --- a/clang/include/clang/ASTMatchers/ASTMatchersInternal.h +++ b/clang/include/clang/ASTMatchers/ASTMatchersInternal.h @@ -1017,7 +1017,7 @@ private: // First, for any types that have a declaration, extract the declaration and // match on it. if (const auto *S = dyn_cast<TagType>(&Node)) { - return matchesDecl(S->getOriginalDecl(), Finder, Builder); + return matchesDecl(S->getDecl(), Finder, Builder); } if (const auto *S = dyn_cast<TemplateTypeParmType>(&Node)) { return matchesDecl(S->getDecl(), Finder, Builder); diff --git a/clang/include/clang/Basic/Attr.td b/clang/include/clang/Basic/Attr.td index 3cde249..eb48a0c 100644 --- a/clang/include/clang/Basic/Attr.td +++ b/clang/include/clang/Basic/Attr.td @@ -1572,6 +1572,23 @@ def HIPManaged : InheritableAttr { let Documentation = [HIPManagedAttrDocs]; } +def CUDAClusterDims : InheritableAttr { + let Spellings = [GNU<"cluster_dims">]; + let Args = [ExprArgument<"X">, ExprArgument<"Y", /*opt=*/1>, ExprArgument<"Z", /*opt=*/1>]; + let Subjects = SubjectList<[ObjCMethod, FunctionLike]>; + let LangOpts = [CUDA]; + let Documentation = [CUDAClusterDimsAttrDoc]; +} + +def CUDANoCluster : InheritableAttr { + let Spellings = [GNU<"no_cluster">]; + let Subjects = SubjectList<[ObjCMethod, FunctionLike]>; + let LangOpts = [CUDA]; + let Documentation = [CUDANoClusterAttrDoc]; +} + +def : MutualExclusions<[CUDAClusterDims, CUDANoCluster]>; + def CUDAInvalidTarget : InheritableAttr { let Spellings = []; let Subjects = SubjectList<[Function]>; @@ -1599,8 +1616,15 @@ def CUDAShared : InheritableAttr { } def : MutualExclusions<[CUDAConstant, CUDAShared, HIPManaged]>; +def SYCLKernel : InheritableAttr { + let Spellings = [Clang<"sycl_kernel">]; + let Subjects = SubjectList<[FunctionTmpl]>; + let LangOpts = [SYCLDevice]; + let Documentation = [SYCLKernelDocs]; +} + def DeviceKernel : DeclOrTypeAttr { - let Spellings = [Clang<"device_kernel">, Clang<"sycl_kernel">, + let Spellings = [Clang<"device_kernel">, Clang<"nvptx_kernel">, Clang<"amdgpu_kernel">, CustomKeyword<"__kernel">, CustomKeyword<"kernel">]; let Documentation = [DeviceKernelDocs]; @@ -1624,15 +1648,6 @@ def DeviceKernel : DeclOrTypeAttr { if(!A) return false; return isNVPTXSpelling(*A); } - static inline bool isSYCLSpelling(const AttributeCommonInfo& A) { - return A.getAttributeSpellingListIndex() == GNU_sycl_kernel || - A.getAttributeSpellingListIndex() == CXX11_clang_sycl_kernel || - A.getAttributeSpellingListIndex() == C23_clang_sycl_kernel; - } - static inline bool isSYCLSpelling(const AttributeCommonInfo* A) { - if(!A) return false; - return isSYCLSpelling(*A); - } static inline bool isOpenCLSpelling(const AttributeCommonInfo& A) { // Tablegen trips underscores from spellings to build the spelling // list, but here we have the same spelling with unscores and without, diff --git a/clang/include/clang/Basic/AttrDocs.td b/clang/include/clang/Basic/AttrDocs.td index 20a52b4..2fdd041 100644 --- a/clang/include/clang/Basic/AttrDocs.td +++ b/clang/include/clang/Basic/AttrDocs.td @@ -396,13 +396,10 @@ any option of a multiversioned function is undefined. }]; } -def DeviceKernelDocs : Documentation { +def SYCLKernelDocs : Documentation { let Category = DocCatFunction; - let Heading = "device_kernel, sycl_kernel, nvptx_kernel, amdgpu_kernel, " - "kernel, __kernel"; + let Heading = "sycl_kernel"; let Content = [{ -These attributes specify that the function represents a kernel for device offloading. -The specific semantics depend on the offloading language, target, and attribute spelling. The ``sycl_kernel`` attribute specifies that a function template will be used to outline device code and to generate an OpenCL kernel. Here is a code example of the SYCL program, which demonstrates the compiler's @@ -476,6 +473,21 @@ The SYCL kernel in the previous code sample meets these expectations. }]; } +def DeviceKernelDocs : Documentation { + let Category = DocCatFunction; + let Heading = "device_kernel, nvptx_kernel, amdgpu_kernel, " + "kernel, __kernel"; + let Content = [{ +These attributes specify that the function represents a kernel for device offloading. +The specific semantics depend on the offloading language, target, and attribute spelling. +Here is a code example using the attribute to mark a function as a kernel: + +.. code-block:: c++ + + [[clang::device_kernel]] int foo(int x) { return ++x; } + }]; +} + def SYCLExternalDocs : Documentation { let Category = DocCatFunction; let Heading = "sycl_external"; @@ -7533,6 +7545,45 @@ A managed variable can be accessed in both device and host code. }]; } +def CUDAClusterDimsAttrDoc : Documentation { + let Category = DocCatDecl; + let Content = [{ +In CUDA/HIP programming, the ``cluster_dims`` attribute, conventionally exposed as the +``__cluster_dims__`` macro, can be applied to a kernel function to set the dimensions of a +thread block cluster, which is an optional level of hierarchy and made up of thread blocks. +``__cluster_dims__`` defines the cluster size as ``(X, Y, Z)``, where each value is the number +of thread blocks in that dimension. The ``cluster_dims`` and `no_cluster`` attributes are +mutually exclusive. + +.. code:: + + __global__ __cluster_dims__(2, 1, 1) void kernel(...) { + ... + } + + }]; +} + +def CUDANoClusterAttrDoc : Documentation { + let Category = DocCatDecl; + let Content = [{ +In CUDA/HIP programming, a kernel function can still be launched with the cluster feature enabled +at runtime, even without being annotated with ``__cluster_dims__``. The LLVM/Clang-exclusive +``no_cluster`` attribute, conventionally exposed as the ``__no_cluster__`` macro, can be applied to +a kernel function to explicitly indicate that the cluster feature will not be enabled either at +compile time or at kernel launch time. This allows the compiler to apply certain optimizations +without assuming that clustering could be enabled at runtime. It is undefined behavior to launch a +kernel annotated with ``__no_cluster__`` if the cluster feature is enabled at runtime. +The ``cluster_dims`` and ``no_cluster`` attributes are mutually exclusive. + +.. code:: + + __global__ __no_cluster__ void kernel(...) { + ... + } + }]; +} + def LifetimeOwnerDocs : Documentation { let Category = DocCatDecl; let Content = [{ diff --git a/clang/include/clang/Basic/BuiltinsX86.td b/clang/include/clang/Basic/BuiltinsX86.td index 217589d..62c70fba 100644 --- a/clang/include/clang/Basic/BuiltinsX86.td +++ b/clang/include/clang/Basic/BuiltinsX86.td @@ -110,21 +110,21 @@ let Attributes = [Const, NoThrow, RequiredVectorWidth<128>] in { } let Features = "sse3" in { - foreach Op = ["addsub", "hadd", "hsub"] in { + foreach Op = ["addsub"] in { def Op#ps : X86Builtin<"_Vector<4, float>(_Vector<4, float>, _Vector<4, float>)">; def Op#pd : X86Builtin<"_Vector<2, double>(_Vector<2, double>, _Vector<2, double>)">; } } - let Features = "ssse3" in { - foreach Op = ["phadd", "phsub"] in { - def Op#w128 : X86Builtin<"_Vector<8, short>(_Vector<8, short>, _Vector<8, short>)">; - def Op#sw128 : X86Builtin<"_Vector<8, short>(_Vector<8, short>, _Vector<8, short>)">; - def Op#d128 : X86Builtin<"_Vector<4, int>(_Vector<4, int>, _Vector<4, int>)">; + let Features = "sse3", Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<128>] in { + foreach Op = ["hadd", "hsub"] in { + def Op#ps : X86Builtin<"_Vector<4, float>(_Vector<4, float>, _Vector<4, float>)">; + def Op#pd : X86Builtin<"_Vector<2, double>(_Vector<2, double>, _Vector<2, double>)">; } + } + let Features = "ssse3" in { def pmulhrsw128 : X86Builtin<"_Vector<8, short>(_Vector<8, short>, _Vector<8, short>)">; - def pshufb128 : X86Builtin<"_Vector<16, char>(_Vector<16, char>, _Vector<16, char>)">; def psignb128 : X86Builtin<"_Vector<16, char>(_Vector<16, char>, _Vector<16, char>)">; def psignw128 : X86Builtin<"_Vector<8, short>(_Vector<8, short>, _Vector<8, short>)">; def psignd128 : X86Builtin<"_Vector<4, int>(_Vector<4, int>, _Vector<4, int>)">; @@ -132,12 +132,13 @@ let Attributes = [Const, NoThrow, RequiredVectorWidth<128>] in { let Features = "ssse3", Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<128>] in { def pmaddubsw128 : X86Builtin<"_Vector<8, short>(_Vector<16, char>, _Vector<16, char>)">; + def pshufb128 : X86Builtin<"_Vector<16, char>(_Vector<16, char>, _Vector<16, char>)">; } } // AVX let Attributes = [Const, NoThrow, RequiredVectorWidth<256>], Features = "avx" in { - foreach Op = ["addsub", "hadd", "hsub", "max", "min"] in { + foreach Op = ["addsub", "max", "min"] in { def Op#pd256 : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>)">; def Op#ps256 : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>)">; } @@ -316,6 +317,14 @@ let Features = "ssse3", Attributes = [NoThrow, Const, RequiredVectorWidth<128>] def palignr128 : X86Builtin<"_Vector<16, char>(_Vector<16, char>, _Vector<16, char>, _Constant int)">; } +let Features = "ssse3", Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<128>] in { + foreach Op = ["phadd", "phsub"] in { + def Op#w128 : X86Builtin<"_Vector<8, short>(_Vector<8, short>, _Vector<8, short>)">; + def Op#sw128 : X86Builtin<"_Vector<8, short>(_Vector<8, short>, _Vector<8, short>)">; + def Op#d128 : X86Builtin<"_Vector<4, int>(_Vector<4, int>, _Vector<4, int>)">; + } +} + let Features = "sse4.1", Attributes = [NoThrow, Const, RequiredVectorWidth<128>] in { def insertps128 : X86Builtin<"_Vector<4, float>(_Vector<4, float>, _Vector<4, float>, _Constant char)">; def roundps : X86Builtin<"_Vector<4, float>(_Vector<4, float>, _Constant int)">; @@ -323,14 +332,22 @@ let Features = "sse4.1", Attributes = [NoThrow, Const, RequiredVectorWidth<128>] def roundsd : X86Builtin<"_Vector<2, double>(_Vector<2, double>, _Vector<2, double>, _Constant int)">; def roundpd : X86Builtin<"_Vector<2, double>(_Vector<2, double>, _Constant int)">; def dpps : X86Builtin<"_Vector<4, float>(_Vector<4, float>, _Vector<4, float>, _Constant char)">; - def dppd : X86Builtin<"_Vector<2, double>(_Vector<2, double>, _Vector<2, double>, _Constant char)">; - def ptestz128 : X86Builtin<"int(_Vector<2, long long int>, _Vector<2, long long int>)">; - def ptestc128 : X86Builtin<"int(_Vector<2, long long int>, _Vector<2, long long int>)">; - def ptestnzc128 : X86Builtin<"int(_Vector<2, long long int>, _Vector<2, long long int>)">; + def dppd : X86Builtin<"_Vector<2, double>(_Vector<2, double>, " + "_Vector<2,double>, _Constant char)">; def mpsadbw128 : X86Builtin<"_Vector<16, char>(_Vector<16, char>, _Vector<16, char>, _Constant char)">; def phminposuw128 : X86Builtin<"_Vector<8, short>(_Vector<8, short>)">; } +let Features = "sse4.1", + Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<128>] in { + def ptestz128 + : X86Builtin<"int(_Vector<2, long long int>, _Vector<2, long long int>)">; + def ptestc128 + : X86Builtin<"int(_Vector<2, long long int>, _Vector<2, long long int>)">; + def ptestnzc128 + : X86Builtin<"int(_Vector<2, long long int>, _Vector<2, long long int>)">; +} + let Features = "sse4.1", Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<128>] in { def pblendw128 : X86Builtin<"_Vector<8, short>(_Vector<8, short>, _Vector<8, short>, _Constant int)">; def blendpd : X86Builtin<"_Vector<2, double>(_Vector<2, double>, _Vector<2, double>, _Constant int)">; @@ -507,6 +524,11 @@ let Features = "avx", Attributes = [NoThrow, Const, Constexpr, RequiredVectorWid def vinsertf128_pd256 : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<2, double>, _Constant int)">; def vinsertf128_ps256 : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<4, float>, _Constant int)">; def vinsertf128_si256 : X86Builtin<"_Vector<8, int>(_Vector<8, int>, _Vector<4, int>, _Constant int)">; + + foreach Op = ["hadd", "hsub"] in { + def Op#pd256 : X86Builtin<"_Vector<4, double>(_Vector<4, double>, _Vector<4, double>)">; + def Op#ps256 : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Vector<8, float>)">; + } } let Features = "avx", Attributes = [NoThrow, Const, RequiredVectorWidth<256>] in { @@ -520,8 +542,8 @@ let Features = "avx", Attributes = [NoThrow, Const, RequiredVectorWidth<256>] in def roundps256 : X86Builtin<"_Vector<8, float>(_Vector<8, float>, _Constant int)">; } - -let Features = "avx", Attributes = [NoThrow, Const, RequiredVectorWidth<128>] in { +let Features = "avx", + Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<128>] in { def vtestzpd : X86Builtin<"int(_Vector<2, double>, _Vector<2, double>)">; def vtestcpd : X86Builtin<"int(_Vector<2, double>, _Vector<2, double>)">; def vtestnzcpd : X86Builtin<"int(_Vector<2, double>, _Vector<2, double>)">; @@ -530,7 +552,8 @@ let Features = "avx", Attributes = [NoThrow, Const, RequiredVectorWidth<128>] in def vtestnzcps : X86Builtin<"int(_Vector<4, float>, _Vector<4, float>)">; } -let Features = "avx", Attributes = [NoThrow, Const, RequiredVectorWidth<256>] in { +let Features = "avx", + Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<256>] in { def vtestzpd256 : X86Builtin<"int(_Vector<4, double>, _Vector<4, double>)">; def vtestcpd256 : X86Builtin<"int(_Vector<4, double>, _Vector<4, double>)">; def vtestnzcpd256 : X86Builtin<"int(_Vector<4, double>, _Vector<4, double>)">; @@ -540,6 +563,10 @@ let Features = "avx", Attributes = [NoThrow, Const, RequiredVectorWidth<256>] in def ptestz256 : X86Builtin<"int(_Vector<4, long long int>, _Vector<4, long long int>)">; def ptestc256 : X86Builtin<"int(_Vector<4, long long int>, _Vector<4, long long int>)">; def ptestnzc256 : X86Builtin<"int(_Vector<4, long long int>, _Vector<4, long long int>)">; +} + +let Features = "avx", + Attributes = [NoThrow, Const, RequiredVectorWidth<256>] in { def movmskpd256 : X86Builtin<"int(_Vector<4, double>)">; def movmskps256 : X86Builtin<"int(_Vector<8, float>)">; } @@ -579,16 +606,10 @@ let Features = "avx", Attributes = [NoThrow, Const, Constexpr, RequiredVectorWid let Features = "avx2", Attributes = [NoThrow, Const, RequiredVectorWidth<256>] in { def mpsadbw256 : X86Builtin<"_Vector<32, char>(_Vector<32, char>, _Vector<32, char>, _Constant char)">; def palignr256 : X86Builtin<"_Vector<32, char>(_Vector<32, char>, _Vector<32, char>, _Constant int)">; - def phaddw256 : X86Builtin<"_Vector<16, short>(_Vector<16, short>, _Vector<16, short>)">; - def phaddd256 : X86Builtin<"_Vector<8, int>(_Vector<8, int>, _Vector<8, int>)">; - def phaddsw256 : X86Builtin<"_Vector<16, short>(_Vector<16, short>, _Vector<16, short>)">; - def phsubw256 : X86Builtin<"_Vector<16, short>(_Vector<16, short>, _Vector<16, short>)">; - def phsubd256 : X86Builtin<"_Vector<8, int>(_Vector<8, int>, _Vector<8, int>)">; - def phsubsw256 : X86Builtin<"_Vector<16, short>(_Vector<16, short>, _Vector<16, short>)">; + def pmovmskb256 : X86Builtin<"int(_Vector<32, char>)">; def pmulhrsw256 : X86Builtin<"_Vector<16, short>(_Vector<16, short>, _Vector<16, short>)">; def psadbw256 : X86Builtin<"_Vector<4, long long int>(_Vector<32, char>, _Vector<32, char>)">; - def pshufb256 : X86Builtin<"_Vector<32, char>(_Vector<32, char>, _Vector<32, char>)">; def psignb256 : X86Builtin<"_Vector<32, char>(_Vector<32, char>, _Vector<32, char>)">; def psignw256 : X86Builtin<"_Vector<16, short>(_Vector<16, short>, _Vector<16, short>)">; def psignd256 : X86Builtin<"_Vector<8, int>(_Vector<8, int>, _Vector<8, int>)">; @@ -627,6 +648,8 @@ let Features = "avx2", Attributes = [NoThrow, Const, Constexpr, RequiredVectorWi def pmuldq256 : X86Builtin<"_Vector<4, long long int>(_Vector<8, int>, _Vector<8, int>)">; def pmuludq256 : X86Builtin<"_Vector<4, long long int>(_Vector<8, int>, _Vector<8, int>)">; + def pshufb256 : X86Builtin<"_Vector<32, char>(_Vector<32, char>, _Vector<32, char>)">; + def psllwi256 : X86Builtin<"_Vector<16, short>(_Vector<16, short>, int)">; def pslldi256 : X86Builtin<"_Vector<8, int>(_Vector<8, int>, int)">; def psllqi256 : X86Builtin<"_Vector<4, long long int>(_Vector<4, long long int>, int)">; @@ -653,6 +676,13 @@ let Features = "avx2", Attributes = [NoThrow, Const, Constexpr, RequiredVectorWi def packssdw256 : X86Builtin<"_Vector<16, short>(_Vector<8, int>, _Vector<8, int>)">; def packuswb256 : X86Builtin<"_Vector<32, char>(_Vector<16, short>, _Vector<16, short>)">; + def phaddw256 : X86Builtin<"_Vector<16, short>(_Vector<16, short>, _Vector<16, short>)">; + def phaddd256 : X86Builtin<"_Vector<8, int>(_Vector<8, int>, _Vector<8, int>)">; + def phaddsw256 : X86Builtin<"_Vector<16, short>(_Vector<16, short>, _Vector<16, short>)">; + def phsubw256 : X86Builtin<"_Vector<16, short>(_Vector<16, short>, _Vector<16, short>)">; + def phsubd256 : X86Builtin<"_Vector<8, int>(_Vector<8, int>, _Vector<8, int>)">; + def phsubsw256 : X86Builtin<"_Vector<16, short>(_Vector<16, short>, _Vector<16, short>)">; + def pshuflw256 : X86Builtin<"_Vector<16, short>(_Vector<16, short>, _Constant int)">; def pshufhw256 : X86Builtin<"_Vector<16, short>(_Vector<16, short>, _Constant int)">; def pshufd256 : X86Builtin<"_Vector<8, int>(_Vector<8, int>, _Constant int)">; @@ -1318,7 +1348,6 @@ let Features = "avx512f", Attributes = [NoThrow, Const, RequiredVectorWidth<512> let Features = "avx512bw", Attributes = [NoThrow, Const, RequiredVectorWidth<512>] in { def ucmpw512_mask : X86Builtin<"unsigned int(_Vector<32, short>, _Vector<32, short>, _Constant int, unsigned int)">; - def pshufb512 : X86Builtin<"_Vector<64, char>(_Vector<64, char>, _Vector<64, char>)">; } let Features = "avx512bw", Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<512>] in { @@ -1326,25 +1355,21 @@ let Features = "avx512bw", Attributes = [NoThrow, Const, Constexpr, RequiredVect def packssdw512 : X86Builtin<"_Vector<32, short>(_Vector<16, int>, _Vector<16, int>)">; def packuswb512 : X86Builtin<"_Vector<64, char>(_Vector<32, short>, _Vector<32, short>)">; def packusdw512 : X86Builtin<"_Vector<32, short>(_Vector<16, int>, _Vector<16, int>)">; -} - -let Features = "avx512cd,avx512vl", Attributes = [NoThrow, Const, RequiredVectorWidth<128>] in { - def vpconflictdi_128 : X86Builtin<"_Vector<2, long long int>(_Vector<2, long long int>)">; -} -let Features = "avx512cd,avx512vl", Attributes = [NoThrow, Const, RequiredVectorWidth<256>] in { - def vpconflictdi_256 : X86Builtin<"_Vector<4, long long int>(_Vector<4, long long int>)">; + def pshufb512 : X86Builtin<"_Vector<64, char>(_Vector<64, char>, _Vector<64, char>)">; } -let Features = "avx512cd,avx512vl", Attributes = [NoThrow, Const, RequiredVectorWidth<128>] in { +let Features = "avx512cd,avx512vl", Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<128>] in { + def vpconflictdi_128 : X86Builtin<"_Vector<2, long long int>(_Vector<2, long long int>)">; def vpconflictsi_128 : X86Builtin<"_Vector<4, int>(_Vector<4, int>)">; } -let Features = "avx512cd,avx512vl", Attributes = [NoThrow, Const, RequiredVectorWidth<256>] in { +let Features = "avx512cd,avx512vl", Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<256>] in { + def vpconflictdi_256 : X86Builtin<"_Vector<4, long long int>(_Vector<4, long long int>)">; def vpconflictsi_256 : X86Builtin<"_Vector<8, int>(_Vector<8, int>)">; } -let Features = "avx512cd", Attributes = [NoThrow, Const, RequiredVectorWidth<512>] in { +let Features = "avx512cd", Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<512>] in { def vpconflictdi_512 : X86Builtin<"_Vector<8, long long int>(_Vector<8, long long int>)">; def vpconflictsi_512 : X86Builtin<"_Vector<16, int>(_Vector<16, int>)">; } @@ -2108,24 +2133,18 @@ let Features = "avx512vl", Attributes = [NoThrow, RequiredVectorWidth<256>] in { def movdqa64store256_mask : X86Builtin<"void(_Vector<4, long long int *>, _Vector<4, long long int>, unsigned char)">; } -let Features = "avx512ifma", Attributes = [NoThrow, Const, RequiredVectorWidth<512>] in { +let Features = "avx512ifma", Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<512>] in { def vpmadd52huq512 : X86Builtin<"_Vector<8, long long int>(_Vector<8, long long int>, _Vector<8, long long int>, _Vector<8, long long int>)">; def vpmadd52luq512 : X86Builtin<"_Vector<8, long long int>(_Vector<8, long long int>, _Vector<8, long long int>, _Vector<8, long long int>)">; } -let Features = "avx512ifma,avx512vl|avxifma", Attributes = [NoThrow, Const, RequiredVectorWidth<128>] in { +let Features = "avx512ifma,avx512vl|avxifma", Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<128>] in { def vpmadd52huq128 : X86Builtin<"_Vector<2, long long int>(_Vector<2, long long int>, _Vector<2, long long int>, _Vector<2, long long int>)">; -} - -let Features = "avx512ifma,avx512vl|avxifma", Attributes = [NoThrow, Const, RequiredVectorWidth<256>] in { - def vpmadd52huq256 : X86Builtin<"_Vector<4, long long int>(_Vector<4, long long int>, _Vector<4, long long int>, _Vector<4, long long int>)">; -} - -let Features = "avx512ifma,avx512vl|avxifma", Attributes = [NoThrow, Const, RequiredVectorWidth<128>] in { def vpmadd52luq128 : X86Builtin<"_Vector<2, long long int>(_Vector<2, long long int>, _Vector<2, long long int>, _Vector<2, long long int>)">; } -let Features = "avx512ifma,avx512vl|avxifma", Attributes = [NoThrow, Const, RequiredVectorWidth<256>] in { +let Features = "avx512ifma,avx512vl|avxifma", Attributes = [NoThrow, Const, Constexpr, RequiredVectorWidth<256>] in { + def vpmadd52huq256 : X86Builtin<"_Vector<4, long long int>(_Vector<4, long long int>, _Vector<4, long long int>, _Vector<4, long long int>)">; def vpmadd52luq256 : X86Builtin<"_Vector<4, long long int>(_Vector<4, long long int>, _Vector<4, long long int>, _Vector<4, long long int>)">; } diff --git a/clang/include/clang/Basic/DiagnosticGroups.td b/clang/include/clang/Basic/DiagnosticGroups.td index ef3f59f..8aa3489 100644 --- a/clang/include/clang/Basic/DiagnosticGroups.td +++ b/clang/include/clang/Basic/DiagnosticGroups.td @@ -631,6 +631,7 @@ def MissingFieldInitializers : DiagGroup<"missing-field-initializers", def ModuleLock : DiagGroup<"module-lock">; def ModuleBuild : DiagGroup<"module-build">; def ModuleImport : DiagGroup<"module-import">; +def ModuleValidation : DiagGroup<"module-validation">; def ModuleConflict : DiagGroup<"module-conflict">; def ModuleFileExtension : DiagGroup<"module-file-extension">; def ModuleIncludeDirectiveTranslation : DiagGroup<"module-include-translation">; diff --git a/clang/include/clang/Basic/DiagnosticParseKinds.td b/clang/include/clang/Basic/DiagnosticParseKinds.td index c724136..e5e071f 100644 --- a/clang/include/clang/Basic/DiagnosticParseKinds.td +++ b/clang/include/clang/Basic/DiagnosticParseKinds.td @@ -505,7 +505,7 @@ def err_expected_coloncolon_after_super : Error< "expected '::' after '__super'">; def ext_decomp_decl_empty : ExtWarn< - "ISO C++17 does not allow a decomposition group to be empty">, + "ISO C++17 does not allow a structured binding group to be empty">, InGroup<DiagGroup<"empty-decomposition">>; def err_function_parameter_limit_exceeded : Error< diff --git a/clang/include/clang/Basic/DiagnosticSemaKinds.td b/clang/include/clang/Basic/DiagnosticSemaKinds.td index 3df28f2..22de85d 100644 --- a/clang/include/clang/Basic/DiagnosticSemaKinds.td +++ b/clang/include/clang/Basic/DiagnosticSemaKinds.td @@ -31,12 +31,12 @@ defm constexpr_body_multiple_return : CXX14Compat< defm variable_template : CXX14Compat<"variable templates are">; // C++17 compatibility with C++14 and earlier. -defm decomp_decl : CXX17Compat<"decomposition declarations are">; +defm decomp_decl : CXX17Compat<"structured binding declarations are">; defm inline_variable : CXX17Compat<"inline variables are">; // C++20 compatibility with C++17 and earlier. defm decomp_decl_spec - : CXX20Compat<"decomposition declaration declared '%0' is">; + : CXX20Compat<"structured binding declaration declared '%0' is">; defm constexpr_local_var_no_init : CXX20Compat< "uninitialized variable in a constexpr %select{function|constructor}0 is">; defm constexpr_function_try_block : CXX20Compat< @@ -591,58 +591,58 @@ def warn_modifying_shadowing_decl : // C++ decomposition declarations def err_decomp_decl_context : Error< - "decomposition declaration not permitted in this context">; + "structured binding declaration not permitted in this context">; def err_decomp_decl_spec - : Error<"decomposition declaration cannot be declared '%0'">; + : Error<"structured binding declaration cannot be declared '%0'">; def err_decomp_decl_type : Error< - "decomposition declaration cannot be declared with type %0; " + "structured binding declaration cannot be declared with type %0; " "declared type must be 'auto' or reference to 'auto'">; def err_decomp_decl_constraint : Error< - "decomposition declaration cannot be declared with constrained 'auto'">; + "structured binding declaration cannot be declared with constrained 'auto'">; def err_decomp_decl_parens : Error< - "decomposition declaration cannot be declared with parentheses">; + "structured binding declaration cannot be declared with parentheses">; def err_decomp_decl_template : Error< - "decomposition declaration cannot be a template">; + "structured binding declaration cannot be a template">; def err_decomp_decl_not_alone : Error< - "decomposition declaration must be the only declaration in its group">; + "structured binding declaration must be the only declaration in its group">; def err_decomp_decl_requires_init : Error< - "decomposition declaration %0 requires an initializer">; + "structured binding declaration %0 requires an initializer">; def err_decomp_decl_wrong_number_bindings : Error< - "type %0 decomposes into %3 %plural{1:element|:elements}2, but " + "type %0 binds to %3 %plural{1:element|:elements}2, but " "%select{%plural{0:no|:only %1}1|%1}4 " "%plural{1:name was|:names were}1 provided">; def err_decomp_decl_unbindable_type : Error< - "cannot decompose %select{union|non-class, non-array}1 type %2">; + "cannot bind %select{union|non-class, non-array}1 type %2">; def err_decomp_decl_multiple_bases_with_members : Error< - "cannot decompose class type %1: " + "cannot bind class type %1: " "%select{its base classes %2 and|both it and its base class}0 %3 " "have non-static data members">; def err_decomp_decl_ambiguous_base : Error< - "cannot decompose members of ambiguous base class %1 of %0:%2">; + "cannot bind members of ambiguous base class %1 of %0:%2">; def err_decomp_decl_inaccessible_base : Error< - "cannot decompose members of inaccessible base class %1 of %0">, + "cannot bind members of inaccessible base class %1 of %0">, AccessControl; def err_decomp_decl_inaccessible_field : Error< - "cannot decompose %select{private|protected}0 member %1 of %3">, + "cannot bind %select{private|protected}0 member %1 of %3">, AccessControl; def err_decomp_decl_lambda : Error< - "cannot decompose lambda closure type">; + "cannot bind lambda closure type">; def err_decomp_decl_anon_union_member : Error< - "cannot decompose class type %0 because it has an anonymous " + "cannot bind class type %0 because it has an anonymous " "%select{struct|union}1 member">; def err_decomp_decl_std_tuple_element_not_specialized : Error< - "cannot decompose this type; 'std::tuple_element<%0>::type' " + "cannot bind this type; 'std::tuple_element<%0>::type' " "does not name a type">; def err_decomp_decl_std_tuple_size_not_constant : Error< - "cannot decompose this type; 'std::tuple_size<%0>::value' " + "cannot bind this type; 'std::tuple_size<%0>::value' " "is not a valid integral constant expression">; def err_decomp_decl_std_tuple_size_invalid - : Error<"cannot decompose this type; 'std::tuple_size<%0>::value' " + : Error<"cannot bind this type; 'std::tuple_size<%0>::value' " "is not a valid size: %1">; def note_in_binding_decl_init : Note< "in implicit initialization of binding declaration %0">; def err_arg_is_not_destructurable : Error< - "type %0 cannot be decomposed">; + "type %0 cannot be bound">; def err_std_type_trait_not_class_template : Error< "unsupported standard library implementation: " @@ -2620,7 +2620,7 @@ def err_auto_variable_cannot_appear_in_own_initializer "declared with deduced type %2 cannot appear in its own initializer">; def err_binding_cannot_appear_in_own_initializer : Error< "binding %0 cannot appear in the initializer of its own " - "decomposition declaration">; + "structured binding declaration">; def err_new_array_of_auto : Error< "cannot allocate array of 'auto'">; def err_auto_not_allowed : Error< @@ -13070,6 +13070,12 @@ def warn_cuda_maxclusterrank_sm_90 : Warning< "maxclusterrank requires sm_90 or higher, CUDA arch provided: %0, ignoring " "%1 attribute">, InGroup<IgnoredAttributes>; +def err_cluster_attr_not_supported : Error< + "%0 is not supported for this GPU architecture">; + +def err_cluster_dims_too_large : Error< + "cluster does not support more than %0 thread blocks; %1 provided">; + // VTable pointer authentication errors def err_non_polymorphic_vtable_pointer_auth : Error< "cannot set vtable pointer authentication on monomorphic type %0">; @@ -13665,6 +13671,11 @@ def warn_acc_var_referenced_lacks_op "reference has no effect">, InGroup<DiagGroup<"openacc-var-lacks-operation">>, DefaultError; +def err_acc_reduction_recipe_no_op + : Error<"variable of type %0 referenced in OpenACC 'reduction' clause does " + "not have a valid operation available">; +def note_acc_reduction_recipe_noop_field + : Note<"while forming combiner for compound type %0">; // AMDGCN builtins diagnostics def err_amdgcn_load_lds_size_invalid_value : Error<"invalid size value">; diff --git a/clang/include/clang/Basic/DiagnosticSerializationKinds.td b/clang/include/clang/Basic/DiagnosticSerializationKinds.td index fc3585f..b80aff3 100644 --- a/clang/include/clang/Basic/DiagnosticSerializationKinds.td +++ b/clang/include/clang/Basic/DiagnosticSerializationKinds.td @@ -86,6 +86,10 @@ def remark_module_import : Remark< "importing module '%0'%select{| into '%3'}2 from '%1'">, ShowInSystemHeader, InGroup<ModuleImport>; +def remark_module_validation : Remark< + "validating %0 input files in module '%1' from '%2'">, + ShowInSystemHeader, + InGroup<ModuleValidation>; def err_imported_module_not_found : Error< "module '%0' in precompiled file '%1' %select{(imported by precompiled file '%2') |}4" diff --git a/clang/include/clang/Basic/LangOptions.def b/clang/include/clang/Basic/LangOptions.def index 9e85008..8d6b8a1 100644 --- a/clang/include/clang/Basic/LangOptions.def +++ b/clang/include/clang/Basic/LangOptions.def @@ -243,6 +243,7 @@ ENUM_LANGOPT(HLSLVersion, HLSLLangStd, 16, HLSL_Unset, NotCompatible, "HLSL Vers LANGOPT(HLSLStrictAvailability, 1, 0, NotCompatible, "Strict availability diagnostic mode for HLSL built-in functions.") LANGOPT(HLSLSpvUseUnknownImageFormat, 1, 0, NotCompatible, "For storage images and texel buffers, sets the default format to 'Unknown' when not specified via the `vk::image_format` attribute. If this option is not used, the format is inferred from the resource's data type.") +LANGOPT(HLSLSpvEnableMaximalReconvergence, 1, 0, NotCompatible, "Enables the MaximallyReconvergesKHR execution mode for this module. This ensures that control flow reconverges at well-defined merge points as defined by the Vulkan spec.") LANGOPT(CUDAIsDevice , 1, 0, NotCompatible, "compiling for CUDA device") LANGOPT(CUDAHostDeviceConstexpr, 1, 1, NotCompatible, "treating unattributed constexpr functions as __host__ __device__") @@ -432,6 +433,7 @@ ENUM_LANGOPT(RegisterStaticDestructors, RegisterStaticDestructorsKind, 2, LANGOPT(RegCall4, 1, 0, NotCompatible, "Set __regcall4 as a default calling convention to respect __regcall ABI v.4") LANGOPT(MatrixTypes, 1, 0, NotCompatible, "Enable or disable the builtin matrix type") +VALUE_LANGOPT(MaxMatrixDimension, 32, (1 << 20) - 1, NotCompatible, "maximum allowed matrix dimension") LANGOPT(CXXAssumptions, 1, 1, NotCompatible, "Enable or disable codegen and compile-time checks for C++23's [[assume]] attribute") diff --git a/clang/include/clang/Basic/LangOptions.h b/clang/include/clang/Basic/LangOptions.h index 41595ec..260a753 100644 --- a/clang/include/clang/Basic/LangOptions.h +++ b/clang/include/clang/Basic/LangOptions.h @@ -756,6 +756,15 @@ public: bool isTargetDevice() const { return OpenMPIsTargetDevice || CUDAIsDevice || SYCLIsDevice; } + + /// Returns the most applicable C standard-compliant language version code. + /// If none could be determined, returns \ref std::nullopt. + std::optional<uint32_t> getCLangStd() const; + + /// Returns the most applicable C++ standard-compliant language + /// version code. + /// If none could be determined, returns \ref std::nullopt. + std::optional<uint32_t> getCPlusPlusLangStd() const; }; /// Floating point control options diff --git a/clang/include/clang/Basic/LangStandard.h b/clang/include/clang/Basic/LangStandard.h index 4941223..64645fe 100644 --- a/clang/include/clang/Basic/LangStandard.h +++ b/clang/include/clang/Basic/LangStandard.h @@ -70,8 +70,7 @@ enum LangFeatures { /// standard. struct LangStandard { enum Kind { -#define LANGSTANDARD(id, name, lang, desc, features) \ - lang_##id, +#define LANGSTANDARD(id, name, lang, desc, features, version) lang_##id, #include "clang/Basic/LangStandards.def" lang_unspecified }; @@ -80,6 +79,7 @@ struct LangStandard { const char *Description; unsigned Flags; clang::Language Language; + std::optional<uint32_t> Version; public: /// getName - Get the name of this standard. @@ -91,6 +91,9 @@ public: /// Get the language that this standard describes. clang::Language getLanguage() const { return Language; } + /// Get the version code for this language standard. + std::optional<uint32_t> getVersion() const { return Version; } + /// Language supports '//' comments. bool hasLineComments() const { return Flags & LineComment; } diff --git a/clang/include/clang/Basic/LangStandards.def b/clang/include/clang/Basic/LangStandards.def index 244692a..4edc935 100644 --- a/clang/include/clang/Basic/LangStandards.def +++ b/clang/include/clang/Basic/LangStandards.def @@ -10,7 +10,7 @@ #error "LANGSTANDARD must be defined before including this file" #endif -/// LANGSTANDARD(IDENT, NAME, LANG, DESC, FEATURES) +/// LANGSTANDARD(IDENT, NAME, LANG, DESC, FEATURES, VERSION) /// /// \param IDENT - The name of the standard as a C++ identifier. /// \param NAME - The name of the standard. @@ -18,6 +18,8 @@ /// \param DESC - A short description of the standard. /// \param FEATURES - The standard features as flags, these are enums from the /// clang::frontend namespace, which is assumed to be available. +/// \param VERSION - The official version code for this standard. +/// Has value 'std::nullopt' if no official version exists. /// LANGSTANDARD_ALIAS(IDENT, ALIAS) /// \param IDENT - The name of the standard as a C++ identifier. @@ -36,186 +38,188 @@ // C89-ish modes. LANGSTANDARD(c89, "c89", - C, "ISO C 1990", 0) + C, "ISO C 1990", 0, std::nullopt) LANGSTANDARD_ALIAS(c89, "c90") LANGSTANDARD_ALIAS(c89, "iso9899:1990") LANGSTANDARD(c94, "iso9899:199409", C, "ISO C 1990 with amendment 1", - Digraphs) + Digraphs, 199409) LANGSTANDARD(gnu89, "gnu89", C, "ISO C 1990 with GNU extensions", - LineComment | Digraphs | GNUMode) + LineComment | Digraphs | GNUMode, std::nullopt) LANGSTANDARD_ALIAS(gnu89, "gnu90") // C99-ish modes LANGSTANDARD(c99, "c99", C, "ISO C 1999", - LineComment | C99 | Digraphs | HexFloat) + LineComment | C99 | Digraphs | HexFloat, 199901) LANGSTANDARD_ALIAS(c99, "iso9899:1999") LANGSTANDARD_ALIAS_DEPR(c99, "c9x") LANGSTANDARD_ALIAS_DEPR(c99, "iso9899:199x") LANGSTANDARD(gnu99, "gnu99", C, "ISO C 1999 with GNU extensions", - LineComment | C99 | Digraphs | GNUMode | HexFloat) + LineComment | C99 | Digraphs | GNUMode | HexFloat, 199901) LANGSTANDARD_ALIAS_DEPR(gnu99, "gnu9x") // C11 modes LANGSTANDARD(c11, "c11", C, "ISO C 2011", - LineComment | C99 | C11 | Digraphs | HexFloat) + LineComment | C99 | C11 | Digraphs | HexFloat, 201112) LANGSTANDARD_ALIAS(c11, "iso9899:2011") LANGSTANDARD_ALIAS_DEPR(c11, "c1x") LANGSTANDARD_ALIAS_DEPR(c11, "iso9899:201x") LANGSTANDARD(gnu11, "gnu11", C, "ISO C 2011 with GNU extensions", - LineComment | C99 | C11 | Digraphs | GNUMode | HexFloat) + LineComment | C99 | C11 | Digraphs | GNUMode | HexFloat, 201112) LANGSTANDARD_ALIAS_DEPR(gnu11, "gnu1x") // C17 modes LANGSTANDARD(c17, "c17", C, "ISO C 2017", - LineComment | C99 | C11 | C17 | Digraphs | HexFloat) + LineComment | C99 | C11 | C17 | Digraphs | HexFloat, 201710) LANGSTANDARD_ALIAS(c17, "iso9899:2017") LANGSTANDARD_ALIAS(c17, "c18") LANGSTANDARD_ALIAS(c17, "iso9899:2018") LANGSTANDARD(gnu17, "gnu17", C, "ISO C 2017 with GNU extensions", - LineComment | C99 | C11 | C17 | Digraphs | GNUMode | HexFloat) + LineComment | C99 | C11 | C17 | Digraphs | GNUMode | HexFloat, 201710) LANGSTANDARD_ALIAS(gnu17, "gnu18") // C23 modes LANGSTANDARD(c23, "c23", C, "ISO C 2023", - LineComment | C99 | C11 | C17 | C23 | Digraphs | HexFloat) + LineComment | C99 | C11 | C17 | C23 | Digraphs | HexFloat, 202311) LANGSTANDARD_ALIAS(c23, "iso9899:2024") LANGSTANDARD_ALIAS_DEPR(c23, "c2x") LANGSTANDARD(gnu23, "gnu23", C, "ISO C 2023 with GNU extensions", - LineComment | C99 | C11 | C17 | C23 | Digraphs | GNUMode | HexFloat) + LineComment | C99 | C11 | C17 | C23 | Digraphs | GNUMode | HexFloat, 202311) LANGSTANDARD_ALIAS_DEPR(gnu23, "gnu2x") // C2y modes +// FIXME: Use correct version code for C2y once published. LANGSTANDARD(c2y, "c2y", C, "Working Draft for ISO C2y", - LineComment | C99 | C11 | C17 | C23 | C2y | Digraphs | HexFloat) + LineComment | C99 | C11 | C17 | C23 | C2y | Digraphs | HexFloat, 202400) LANGSTANDARD(gnu2y, "gnu2y", C, "Working Draft for ISO C2y with GNU extensions", - LineComment | C99 | C11 | C17 | C23 | C2y | Digraphs | GNUMode | HexFloat) + LineComment | C99 | C11 | C17 | C23 | C2y | Digraphs | GNUMode | HexFloat, 202400) // TODO: Add the iso9899:202y alias once ISO publishes the standard. // C++ modes LANGSTANDARD(cxx98, "c++98", CXX, "ISO C++ 1998 with amendments", - LineComment | CPlusPlus | Digraphs) + LineComment | CPlusPlus | Digraphs, 199711) LANGSTANDARD_ALIAS(cxx98, "c++03") LANGSTANDARD(gnucxx98, "gnu++98", CXX, "ISO C++ 1998 with amendments and GNU extensions", - LineComment | CPlusPlus | Digraphs | GNUMode) + LineComment | CPlusPlus | Digraphs | GNUMode, 199711) LANGSTANDARD_ALIAS(gnucxx98, "gnu++03") LANGSTANDARD(cxx11, "c++11", CXX, "ISO C++ 2011 with amendments", - LineComment | CPlusPlus | CPlusPlus11 | Digraphs) + LineComment | CPlusPlus | CPlusPlus11 | Digraphs, 201103) LANGSTANDARD_ALIAS_DEPR(cxx11, "c++0x") LANGSTANDARD(gnucxx11, "gnu++11", CXX, "ISO C++ 2011 with amendments and GNU extensions", - LineComment | CPlusPlus | CPlusPlus11 | Digraphs | GNUMode) + LineComment | CPlusPlus | CPlusPlus11 | Digraphs | GNUMode, 201103) LANGSTANDARD_ALIAS_DEPR(gnucxx11, "gnu++0x") LANGSTANDARD(cxx14, "c++14", CXX, "ISO C++ 2014 with amendments", - LineComment | CPlusPlus | CPlusPlus11 | CPlusPlus14 | Digraphs) + LineComment | CPlusPlus | CPlusPlus11 | CPlusPlus14 | Digraphs, 201402) LANGSTANDARD_ALIAS_DEPR(cxx14, "c++1y") LANGSTANDARD(gnucxx14, "gnu++14", CXX, "ISO C++ 2014 with amendments and GNU extensions", LineComment | CPlusPlus | CPlusPlus11 | CPlusPlus14 | Digraphs | - GNUMode) + GNUMode, 201402) LANGSTANDARD_ALIAS_DEPR(gnucxx14, "gnu++1y") LANGSTANDARD(cxx17, "c++17", CXX, "ISO C++ 2017 with amendments", LineComment | CPlusPlus | CPlusPlus11 | CPlusPlus14 | CPlusPlus17 | - Digraphs | HexFloat) + Digraphs | HexFloat, 201703) LANGSTANDARD_ALIAS_DEPR(cxx17, "c++1z") LANGSTANDARD(gnucxx17, "gnu++17", CXX, "ISO C++ 2017 with amendments and GNU extensions", LineComment | CPlusPlus | CPlusPlus11 | CPlusPlus14 | CPlusPlus17 | - Digraphs | HexFloat | GNUMode) + Digraphs | HexFloat | GNUMode, 201703) LANGSTANDARD_ALIAS_DEPR(gnucxx17, "gnu++1z") LANGSTANDARD(cxx20, "c++20", CXX, "ISO C++ 2020 DIS", LineComment | CPlusPlus | CPlusPlus11 | CPlusPlus14 | CPlusPlus17 | - CPlusPlus20 | Digraphs | HexFloat) + CPlusPlus20 | Digraphs | HexFloat, 202002) LANGSTANDARD_ALIAS_DEPR(cxx20, "c++2a") LANGSTANDARD(gnucxx20, "gnu++20", CXX, "ISO C++ 2020 DIS with GNU extensions", LineComment | CPlusPlus | CPlusPlus11 | CPlusPlus14 | CPlusPlus17 | - CPlusPlus20 | Digraphs | HexFloat | GNUMode) + CPlusPlus20 | Digraphs | HexFloat | GNUMode, 202002) LANGSTANDARD_ALIAS_DEPR(gnucxx20, "gnu++2a") LANGSTANDARD(cxx23, "c++23", CXX, "ISO C++ 2023 DIS", LineComment | CPlusPlus | CPlusPlus11 | CPlusPlus14 | CPlusPlus17 | - CPlusPlus20 | CPlusPlus23 | Digraphs | HexFloat) + CPlusPlus20 | CPlusPlus23 | Digraphs | HexFloat, 202302) LANGSTANDARD_ALIAS_DEPR(cxx23, "c++2b") LANGSTANDARD(gnucxx23, "gnu++23", CXX, "ISO C++ 2023 DIS with GNU extensions", LineComment | CPlusPlus | CPlusPlus11 | CPlusPlus14 | CPlusPlus17 | - CPlusPlus20 | CPlusPlus23 | Digraphs | HexFloat | GNUMode) + CPlusPlus20 | CPlusPlus23 | Digraphs | HexFloat | GNUMode, 202302) LANGSTANDARD_ALIAS_DEPR(gnucxx23, "gnu++2b") +// FIXME: Use correct version code for C++26 once published. LANGSTANDARD(cxx26, "c++2c", CXX, "Working draft for C++2c", LineComment | CPlusPlus | CPlusPlus11 | CPlusPlus14 | CPlusPlus17 | - CPlusPlus20 | CPlusPlus23 | CPlusPlus26 | Digraphs | HexFloat) + CPlusPlus20 | CPlusPlus23 | CPlusPlus26 | Digraphs | HexFloat, 202400) LANGSTANDARD_ALIAS(cxx26, "c++26") LANGSTANDARD(gnucxx26, "gnu++2c", CXX, "Working draft for C++2c with GNU extensions", LineComment | CPlusPlus | CPlusPlus11 | CPlusPlus14 | CPlusPlus17 | - CPlusPlus20 | CPlusPlus23 | CPlusPlus26 | Digraphs | HexFloat | GNUMode) + CPlusPlus20 | CPlusPlus23 | CPlusPlus26 | Digraphs | HexFloat | GNUMode, 202400) LANGSTANDARD_ALIAS(gnucxx26, "gnu++26") // OpenCL LANGSTANDARD(opencl10, "cl1.0", OpenCL, "OpenCL 1.0", - LineComment | C99 | Digraphs | HexFloat | OpenCL) + LineComment | C99 | Digraphs | HexFloat | OpenCL, std::nullopt) LANGSTANDARD_ALIAS_DEPR(opencl10, "cl") LANGSTANDARD(opencl11, "cl1.1", OpenCL, "OpenCL 1.1", - LineComment | C99 | Digraphs | HexFloat | OpenCL) + LineComment | C99 | Digraphs | HexFloat | OpenCL, std::nullopt) LANGSTANDARD(opencl12, "cl1.2", OpenCL, "OpenCL 1.2", - LineComment | C99 | Digraphs | HexFloat | OpenCL) + LineComment | C99 | Digraphs | HexFloat | OpenCL, std::nullopt) LANGSTANDARD(opencl20, "cl2.0", OpenCL, "OpenCL 2.0", - LineComment | C99 | Digraphs | HexFloat | OpenCL) + LineComment | C99 | Digraphs | HexFloat | OpenCL, std::nullopt) LANGSTANDARD(opencl30, "cl3.0", OpenCL, "OpenCL 3.0", - LineComment | C99 | Digraphs | HexFloat | OpenCL) + LineComment | C99 | Digraphs | HexFloat | OpenCL, std::nullopt) LANGSTANDARD(openclcpp10, "clc++1.0", OpenCL, "C++ for OpenCL 1.0", LineComment | CPlusPlus | CPlusPlus11 | CPlusPlus14 | CPlusPlus17 | - Digraphs | HexFloat | OpenCL) + Digraphs | HexFloat | OpenCL, std::nullopt) LANGSTANDARD_ALIAS(openclcpp10, "clc++") LANGSTANDARD(openclcpp2021, "clc++2021", OpenCL, "C++ for OpenCL 2021", LineComment | CPlusPlus | CPlusPlus11 | CPlusPlus14 | CPlusPlus17 | - Digraphs | HexFloat | OpenCL) + Digraphs | HexFloat | OpenCL, std::nullopt) LANGSTANDARD_ALIAS_DEPR(opencl10, "CL") LANGSTANDARD_ALIAS_DEPR(opencl11, "CL1.1") @@ -229,35 +233,35 @@ LANGSTANDARD_ALIAS_DEPR(openclcpp2021, "CLC++2021") // HLSL LANGSTANDARD(hlsl, "hlsl", HLSL, "High Level Shader Language", - LineComment | HLSL | CPlusPlus | CPlusPlus11) + LineComment | HLSL | CPlusPlus | CPlusPlus11, std::nullopt) LANGSTANDARD(hlsl2015, "hlsl2015", HLSL, "High Level Shader Language 2015", - LineComment | HLSL | CPlusPlus | CPlusPlus11) + LineComment | HLSL | CPlusPlus | CPlusPlus11, std::nullopt) LANGSTANDARD(hlsl2016, "hlsl2016", HLSL, "High Level Shader Language 2016", - LineComment | HLSL | CPlusPlus | CPlusPlus11) + LineComment | HLSL | CPlusPlus | CPlusPlus11, std::nullopt) LANGSTANDARD(hlsl2017, "hlsl2017", HLSL, "High Level Shader Language 2017", - LineComment | HLSL | CPlusPlus | CPlusPlus11) + LineComment | HLSL | CPlusPlus | CPlusPlus11, std::nullopt) LANGSTANDARD(hlsl2018, "hlsl2018", HLSL, "High Level Shader Language 2018", - LineComment | HLSL | CPlusPlus | CPlusPlus11) + LineComment | HLSL | CPlusPlus | CPlusPlus11, std::nullopt) LANGSTANDARD(hlsl2021, "hlsl2021", HLSL, "High Level Shader Language 2021", - LineComment | HLSL | CPlusPlus | CPlusPlus11) + LineComment | HLSL | CPlusPlus | CPlusPlus11, std::nullopt) LANGSTANDARD(hlsl202x, "hlsl202x", HLSL, "High Level Shader Language 202x", - LineComment | HLSL | CPlusPlus | CPlusPlus11) + LineComment | HLSL | CPlusPlus | CPlusPlus11, std::nullopt) LANGSTANDARD(hlsl202y, "hlsl202y", HLSL, "High Level Shader Language 202y", - LineComment | HLSL | CPlusPlus | CPlusPlus11) + LineComment | HLSL | CPlusPlus | CPlusPlus11, std::nullopt) #undef LANGSTANDARD diff --git a/clang/include/clang/Basic/OpenACCKinds.h b/clang/include/clang/Basic/OpenACCKinds.h index 9383856..8d52e13 100644 --- a/clang/include/clang/Basic/OpenACCKinds.h +++ b/clang/include/clang/Basic/OpenACCKinds.h @@ -494,7 +494,7 @@ inline StreamTy &printOpenACCClauseKind(StreamTy &Out, OpenACCClauseKind K) { case OpenACCClauseKind::Shortloop: llvm_unreachable("Shortloop shouldn't be generated in clang"); - LLVM_FALLTHROUGH; + [[fallthrough]]; case OpenACCClauseKind::Invalid: return Out << "<invalid>"; } diff --git a/clang/include/clang/Basic/TargetInfo.h b/clang/include/clang/Basic/TargetInfo.h index ceb16174..ea73ed9 100644 --- a/clang/include/clang/Basic/TargetInfo.h +++ b/clang/include/clang/Basic/TargetInfo.h @@ -1211,6 +1211,25 @@ public: TiedOperand = N; // Don't copy Name or constraint string. } + + // For output operand constraints, the target can set bounds to indicate + // that the result value is guaranteed to fall within a certain range. + // This will cause corresponding assertions to be emitted that will allow + // for potential optimization based of that guarantee. + // + // NOTE: This re-uses the `ImmRange` fields to store the range, which are + // otherwise unused for constraint types used for output operands. + void setOutputOperandBounds(unsigned Min, unsigned Max) { + ImmRange.Min = Min; + ImmRange.Max = Max; + ImmRange.isConstrained = true; + } + std::optional<std::pair<unsigned, unsigned>> + getOutputOperandBounds() const { + return ImmRange.isConstrained + ? std::make_pair(ImmRange.Min, ImmRange.Max) + : std::optional<std::pair<unsigned, unsigned>>(); + } }; /// Validate register name used for global register variables. diff --git a/clang/include/clang/Basic/arm_mve.td b/clang/include/clang/Basic/arm_mve.td index 412ef9a..2e5e1d9 100644 --- a/clang/include/clang/Basic/arm_mve.td +++ b/clang/include/clang/Basic/arm_mve.td @@ -831,9 +831,8 @@ multiclass contiguous_load<string mnemonic, PrimitiveType memtype, NameOverride<mnemonic>; def: Intrinsic<Vector, (args CPtr<CopyKind<same_size[0], Scalar>>:$addr, Predicate:$pred), - (IRIntBase<"masked_load", [Vector, CPtr<Vector>]> - (CPtr<Vector> $addr), !srl(memtype.size,3), - $pred, (zeroinit Vector))>, + (masked_load Vector, (CPtr<Vector> $addr), + !srl(memtype.size,3), $pred, (zeroinit Vector))>, NameOverride<mnemonic # "_z">; } @@ -846,9 +845,8 @@ multiclass contiguous_load<string mnemonic, PrimitiveType memtype, NameOverride<"vld1q">; def: Intrinsic<Vector, (args CPtr<CopyKind<same_size[0], Scalar>>:$addr, Predicate:$pred), - (IRIntBase<"masked_load", [Vector, CPtr<Vector>]> - (CPtr<Vector> $addr), !srl(memtype.size,3), - $pred, (zeroinit Vector))>, + (masked_load Vector, (CPtr<Vector> $addr), + !srl(memtype.size,3), $pred, (zeroinit Vector))>, NameOverride<"vld1q_z">; } @@ -863,9 +861,7 @@ multiclass contiguous_load<string mnemonic, PrimitiveType memtype, NameOverride<mnemonic>; def: Intrinsic<Vector, (args CPtr<CopyKind<same_size[0], Scalar>>:$addr, Predicate:$pred), - (extend (IRIntBase<"masked_load", - [NarrowedVecOf<memtype,Vector>, - CPtr<NarrowedVecOf<memtype,Vector>>]> + (extend (masked_load NarrowedVecOf<memtype,Vector>, (CPtr<NarrowedVecOf<memtype,Vector>> $addr), !srl(memtype.size,3), $pred, (zeroinit NarrowedVecOf<memtype,Vector>)), @@ -890,8 +886,7 @@ multiclass contiguous_store<string mnemonic, PrimitiveType memtype, NameOverride<mnemonic>; def: Intrinsic<Void, (args Ptr<CopyKind<same_size[0], Scalar>>:$addr, Vector:$value, Predicate:$pred), - (IRIntBase<"masked_store", [Vector, Ptr<Vector>]> - $value, (Ptr<Vector> $addr), + (masked_store $value, (Ptr<Vector> $addr), !srl(memtype.size,3), $pred)>, NameOverride<mnemonic # "_p">; } @@ -907,8 +902,7 @@ multiclass contiguous_store<string mnemonic, PrimitiveType memtype, NameOverride<"vst1q">; def: Intrinsic<Void, (args Ptr<CopyKind<same_size[0], Scalar>>:$addr, Vector:$value, Predicate:$pred), - (IRIntBase<"masked_store", [Vector, Ptr<Vector>]> - $value, (Ptr<Vector> $addr), + (masked_store $value, (Ptr<Vector> $addr), !srl(memtype.size,3), $pred)>, NameOverride<"vst1q_p">; } @@ -925,9 +919,7 @@ multiclass contiguous_store<string mnemonic, PrimitiveType memtype, NameOverride<mnemonic>; def: Intrinsic<Void, (args Ptr<CopyKind<same_size[0], Scalar>>:$addr, Vector:$value, Predicate:$pred), - (IRIntBase<"masked_store", - [NarrowedVecOf<memtype,Vector>, - Ptr<NarrowedVecOf<memtype,Vector>>]> + (masked_store (trunc $value, NarrowedVecOf<memtype,Vector>), (Ptr<NarrowedVecOf<memtype,Vector>> $addr), !srl(memtype.size,3), $pred)>, diff --git a/clang/include/clang/Basic/arm_mve_defs.td b/clang/include/clang/Basic/arm_mve_defs.td index 083d03a..c1562a0 100644 --- a/clang/include/clang/Basic/arm_mve_defs.td +++ b/clang/include/clang/Basic/arm_mve_defs.td @@ -134,6 +134,13 @@ def unzip: CGHelperFn<"VectorUnzip"> { } def zip: CGHelperFn<"VectorZip">; +def masked_load: IRBuilder<"CreateMaskedLoad"> { + let special_params = [IRBuilderIntParam<2, "Align">]; +} +def masked_store: IRBuilder<"CreateMaskedStore"> { + let special_params = [IRBuilderIntParam<2, "Align">]; +} + // Trivial 'codegen' function that just returns its argument. Useful // for wrapping up a variable name like $foo into a thing you can pass // around as type 'dag'. diff --git a/clang/include/clang/CIR/Dialect/Builder/CIRBaseBuilder.h b/clang/include/clang/CIR/Dialect/Builder/CIRBaseBuilder.h index b4c24d7..3ac8987 100644 --- a/clang/include/clang/CIR/Dialect/Builder/CIRBaseBuilder.h +++ b/clang/include/clang/CIR/Dialect/Builder/CIRBaseBuilder.h @@ -127,6 +127,14 @@ public: cir::BoolType getBoolTy() { return cir::BoolType::get(getContext()); } cir::VoidType getVoidTy() { return cir::VoidType::get(getContext()); } + cir::IntType getUIntNTy(int n) { + return cir::IntType::get(getContext(), n, false); + } + + cir::IntType getSIntNTy(int n) { + return cir::IntType::get(getContext(), n, true); + } + cir::PointerType getPointerTo(mlir::Type ty) { return cir::PointerType::get(ty); } diff --git a/clang/include/clang/CIR/Dialect/IR/CIRAttrConstraints.td b/clang/include/clang/CIR/Dialect/IR/CIRAttrConstraints.td index 8f72ff4..2548d46 100644 --- a/clang/include/clang/CIR/Dialect/IR/CIRAttrConstraints.td +++ b/clang/include/clang/CIR/Dialect/IR/CIRAttrConstraints.td @@ -39,13 +39,31 @@ def CIR_AnyIntOrFloatAttr : AnyAttrOf<[CIR_AnyIntAttr, CIR_AnyFPAttr], } //===----------------------------------------------------------------------===// +// Exceptions constraints +//===----------------------------------------------------------------------===// + +def CIR_AnyCatchAllAttr + : CIR_AttrConstraint<"::cir::CatchAllAttr", "catch all attribute">; + +def CIR_AnyUnwindAttr + : CIR_AttrConstraint<"::cir::UnwindAttr", "unwind attribute">; + +//===----------------------------------------------------------------------===// // GlobalViewAttr constraints //===----------------------------------------------------------------------===// -def CIR_AnyGlobalViewAttr : CIR_AttrConstraint<"::cir::GlobalViewAttr", "GlobalView attribute">; +def CIR_AnyGlobalViewAttr + : CIR_AttrConstraint<"::cir::GlobalViewAttr", "GlobalView attribute">; -def CIR_AnyIntOrGlobalViewAttr : AnyAttrOf<[CIR_AnyIntAttr, CIR_AnyGlobalViewAttr], - "integer or global view attribute"> { +def CIR_AnyIntOrGlobalViewAttr + : AnyAttrOf<[CIR_AnyIntAttr, CIR_AnyGlobalViewAttr], + "integer or global view attribute"> { + string cppType = "::mlir::TypedAttr"; +} + +def CIR_TryHandlerAttr + : AnyAttrOf<[CIR_AnyGlobalViewAttr, CIR_AnyCatchAllAttr, CIR_AnyUnwindAttr], + "catch all or unwind or global view attribute"> { string cppType = "::mlir::TypedAttr"; } @@ -61,4 +79,7 @@ def CIR_IntOrGlobalViewArrayAttr : TypedArrayAttrBase<CIR_AnyIntOrGlobalViewAttr string cppType = "::mlir::ArrayAttr"; } -#endif // CLANG_CIR_DIALECT_IR_CIRATTRCONSTRAINTS_TD
\ No newline at end of file +def CIR_TryHandlerArrayAttr : TypedArrayAttrBase<CIR_TryHandlerAttr, + "catch all or unwind or global view array attribute">; + +#endif // CLANG_CIR_DIALECT_IR_CIRATTRCONSTRAINTS_TD diff --git a/clang/include/clang/CIR/Dialect/IR/CIRAttrs.td b/clang/include/clang/CIR/Dialect/IR/CIRAttrs.td index bb62223..1e0fb03 100644 --- a/clang/include/clang/CIR/Dialect/IR/CIRAttrs.td +++ b/clang/include/clang/CIR/Dialect/IR/CIRAttrs.td @@ -814,6 +814,14 @@ def CIR_GlobalCtorAttr : CIR_GlobalCtorDtor<"Ctor", "ctor"> { }]; } +def CIR_GlobalDtorAttr : CIR_GlobalCtorDtor<"Dtor", "dtor"> { + let summary = "Marks a function as a global destructor"; + let description = [{ + Marks a function as a global destructor in the module dtors list. + The function will be executed before the module unloading. + }]; +} + //===----------------------------------------------------------------------===// // BitfieldInfoAttr //===----------------------------------------------------------------------===// @@ -959,5 +967,63 @@ def CIR_TypeInfoAttr : CIR_Attr<"TypeInfo", "typeinfo", [TypedAttrInterface]> { `<` custom<RecordMembers>($data) `>` }]; } +//===----------------------------------------------------------------------===// +// InlineAttr +//===----------------------------------------------------------------------===// + +def CIR_InlineKind : CIR_I32EnumAttr<"InlineKind", "inlineKind", [ + I32EnumAttrCase<"NoInline", 1, "never">, + I32EnumAttrCase<"AlwaysInline", 2, "always">, + I32EnumAttrCase<"InlineHint", 3, "hint"> +]> { + let genSpecializedAttr = 0; +} + +def CIR_InlineAttr : CIR_EnumAttr<CIR_InlineKind, "inline"> { + let summary = "Inline attribute"; + let description = [{ + Inline attribute represents user directives for inlining behavior. + This attribute is only used by `cir.func` operations. + + Values: + - `never`: Prevents the function from being inlined (__attribute__((noinline))) + - `always`: Forces the function to be inlined (__attribute__((always_inline))) + - `hint`: Suggests the function should be inlined (inline keyword) + + Example: + ``` + cir.func @noinline_func(%arg0: !s32i) -> !s32i inline(never) { + cir.return %arg0 : !s32i + } + cir.func @always_inline_func() -> !s32i inline(always) { + %0 = cir.const #cir.int<42> : !s32i + cir.return %0 : !s32i + } + ``` + }]; + + let cppClassName = "InlineAttr"; + + let extraClassDeclaration = [{ + bool isNoInline() const { return getValue() == InlineKind::NoInline; }; + bool isAlwaysInline() const { return getValue() == InlineKind::AlwaysInline; }; + bool isInlineHint() const { return getValue() == InlineKind::InlineHint; }; + }]; +} + +//===----------------------------------------------------------------------===// +// CatchAllAttr & UnwindAttr +//===----------------------------------------------------------------------===// + +// Represents the catch_all region. +def CIR_CatchAllAttr : CIR_UnitAttr<"CatchAll", "all"> { + let storageType = [{ CatchAllAttr }]; +} + +// Represents the unwind region where unwind continues or +// the program std::terminate's. +def CIR_UnwindAttr : CIR_UnitAttr<"Unwind", "unwind"> { + let storageType = [{ CatchUnwind }]; +} #endif // CLANG_CIR_DIALECT_IR_CIRATTRS_TD diff --git a/clang/include/clang/CIR/Dialect/IR/CIRDialect.td b/clang/include/clang/CIR/Dialect/IR/CIRDialect.td index feb08d60..e915371 100644 --- a/clang/include/clang/CIR/Dialect/IR/CIRDialect.td +++ b/clang/include/clang/CIR/Dialect/IR/CIRDialect.td @@ -43,6 +43,7 @@ def CIR_Dialect : Dialect { static llvm::StringRef getSideEffectAttrName() { return "side_effect"; } static llvm::StringRef getModuleLevelAsmAttrName() { return "cir.module_asm"; } static llvm::StringRef getGlobalCtorsAttrName() { return "cir.global_ctors"; } + static llvm::StringRef getGlobalDtorsAttrName() { return "cir.global_dtors"; } void registerAttributes(); void registerTypes(); diff --git a/clang/include/clang/CIR/Dialect/IR/CIROps.td b/clang/include/clang/CIR/Dialect/IR/CIROps.td index 27fe0cc..e0163a4 100644 --- a/clang/include/clang/CIR/Dialect/IR/CIROps.td +++ b/clang/include/clang/CIR/Dialect/IR/CIROps.td @@ -89,6 +89,19 @@ class CIR_Op<string mnemonic, list<Trait> traits = []> : } //===----------------------------------------------------------------------===// +// CIR Operation Traits +//===----------------------------------------------------------------------===// + +class HasAtMostOneOfAttrsPred<list<string> names> : + CPred<!foldl("0", names, acc, name, acc # " + (" # name # " ? 1 : 0)") + # " <= 1">; + +class HasAtMostOneOfAttrs<list<string> names> : PredOpTrait< + "has only one of the optional attributes: " # !interleave(names, ", "), + HasAtMostOneOfAttrsPred<!foreach(name, names, "$" # name)> +>; + +//===----------------------------------------------------------------------===// // CastOp //===----------------------------------------------------------------------===// @@ -631,7 +644,7 @@ def CIR_StoreOp : CIR_Op<"store", [ defvar CIR_ReturnableScopes = [ "FuncOp", "ScopeOp", "IfOp", "SwitchOp", "CaseOp", - "DoWhileOp", "WhileOp", "ForOp" + "DoWhileOp", "WhileOp", "ForOp", "TryOp" ]; def CIR_ReturnOp : CIR_Op<"return", [ @@ -778,7 +791,7 @@ def CIR_ConditionOp : CIR_Op<"condition", [ defvar CIR_YieldableScopes = [ "ArrayCtor", "ArrayDtor", "CaseOp", "DoWhileOp", "ForOp", "GlobalOp", "IfOp", - "ScopeOp", "SwitchOp", "TernaryOp", "WhileOp" + "ScopeOp", "SwitchOp", "TernaryOp", "WhileOp", "TryOp" ]; def CIR_YieldOp : CIR_Op<"yield", [ @@ -2422,9 +2435,17 @@ def CIR_GetMemberOp : CIR_Op<"get_member"> { // TODO(CIR): FuncOp is still a tiny shell of what it will become. Many more // properties and attributes will be added as upstreaming continues. +def CIR_OptionalPriorityAttr : OptionalAttr< + DefaultValuedAttr< + ConfinedAttr<I32Attr, [IntMinValue<101>, IntMaxValue<65535>]>, + "65535" + > +>; + def CIR_FuncOp : CIR_Op<"func", [ AutomaticAllocationScope, CallableOpInterface, FunctionOpInterface, DeclareOpInterfaceMethods<CIRGlobalValueInterface>, + HasAtMostOneOfAttrs<["global_ctor_priority", "global_dtor_priority"]>, IsolatedFromAbove ]> { let summary = "Declare or define a function"; @@ -2449,6 +2470,16 @@ def CIR_FuncOp : CIR_Op<"func", [ without a prototype and, consequently, may contain calls with invalid arguments and undefined behavior. + The `global_ctor` keyword indicates whether a function should execute before + `main()` function, as specified by `__attribute__((constructor))`. An + execution priority can also be specified `global_ctor(<priority>)`. + Similarly, for global destructors both `global_dtor` and + `global_dtor(<priority>)` are available. + + The `inline(never)` keyword marks a function that should not be inlined. + The `inline(always)` keyword marks a function that should always be inlined. + The `inline(hint)` keyword suggests that the function should be inlined. + Example: ```mlir @@ -2483,11 +2514,14 @@ def CIR_FuncOp : CIR_Op<"func", [ UnitAttr:$dso_local, DefaultValuedAttr<CIR_GlobalLinkageKind, "cir::GlobalLinkageKind::ExternalLinkage">:$linkage, + OptionalAttr<CIR_InlineAttr>:$inline_kind, OptionalAttr<StrAttr>:$sym_visibility, UnitAttr:$comdat, OptionalAttr<DictArrayAttr>:$arg_attrs, OptionalAttr<DictArrayAttr>:$res_attrs, - OptionalAttr<FlatSymbolRefAttr>:$aliasee); + OptionalAttr<FlatSymbolRefAttr>:$aliasee, + CIR_OptionalPriorityAttr:$global_ctor_priority, + CIR_OptionalPriorityAttr:$global_dtor_priority); let regions = (region AnyRegion:$body); @@ -4016,6 +4050,43 @@ def CIR_ExpectOp : CIR_Op<"expect", [ } //===----------------------------------------------------------------------===// +// PtrDiffOp +//===----------------------------------------------------------------------===// + +def CIR_PtrDiffOp : CIR_Op<"ptr_diff", [Pure, SameTypeOperands]> { + let summary = "Pointer subtraction arithmetic"; + let description = [{ + The cir.ptr_diff operation computes the difference between two pointers that + have the same element type. + + The result reflects the ABI-defined size of the pointed-to type. For example, + subtracting two !cir.ptr<!u64i> values may yield 1, representing an 8-byte + difference. In contrast, for pointers to void or function types, a result of + 8 corresponds to an 8-byte difference. + + For pointers to types whose size are not aligned with the target data + layout, the size is generally rounded to the next power of 2 bits. For + example, subtracting two !cir.ptr<!s24i> values for the _BitInt(24) type may + yield 1, representing a 4-byte difference (as opposed to a 3-byte + difference). + + Example: + + ```mlir + %7 = cir.ptr_diff %0, %1 : !cir.ptr<!u64i> -> !u64i + ``` + }]; + + let arguments = (ins CIR_PointerType:$lhs, CIR_PointerType:$rhs); + let results = (outs CIR_AnyFundamentalIntType:$result); + + let assemblyFormat = [{ + $lhs `,` $rhs `:` qualified(type($lhs)) `->` qualified(type($result)) + attr-dict + }]; +} + +//===----------------------------------------------------------------------===// // Floating Point Ops //===----------------------------------------------------------------------===// @@ -4297,10 +4368,93 @@ def CIR_AllocExceptionOp : CIR_Op<"alloc.exception"> { } //===----------------------------------------------------------------------===// +// TryOp +//===----------------------------------------------------------------------===// + +def CIR_TryOp : CIR_Op<"try",[ + DeclareOpInterfaceMethods<RegionBranchOpInterface>, + RecursivelySpeculatable, AutomaticAllocationScope, NoRegionArguments +]> { + let summary = "C++ try block"; + let description = [{ + Holds the lexical scope of `try {}`. Note that resources used on catch + clauses are usually allocated in the same parent as `cir.try`. + + `synthetic`: use `cir.try` to represent try/catches not originally + present in the source code. For example, a synthetic `cir.try` region + is created around the constructor call when `operator new` is used + so that the memory allocated will be freed if the constructor throws + an exception. + + `cleanup`: indicates that there are cleanups that must be performed + when exiting the try region via exception, even if the exception is not + caught. + + Example: + + ```mlir + cir.try { + cir.call exception @function() : () -> () + cir.yield + } catch [type #cir.global_view<@_ZTIPf> : !cir.ptr<!u8i>] { + ... + cir.yield + } unwind { + cir.resume + } + ``` + }]; + + let arguments = (ins + UnitAttr:$synthetic, + UnitAttr:$cleanup, + DefaultValuedAttr<CIR_TryHandlerArrayAttr, "{}">:$handler_types + ); + + let regions = (region + AnyRegion:$try_region, + VariadicRegion<AnyRegion>:$handler_regions + ); + + let assemblyFormat = [{ + (`synthetic` $synthetic^)? + (`cleanup` $cleanup^)? + $try_region + custom<TryHandlerRegions>($handler_regions, $handler_types) + attr-dict + }]; + + let builders = [ + OpBuilder<(ins + "llvm::function_ref<void(mlir::OpBuilder &, " + "mlir::Location)>":$tryBuilder, + "llvm::function_ref<void(mlir::OpBuilder &, mlir::Location, " + "mlir::OperationState &)>":$handlersBuilder), + [{ + assert(tryBuilder && "expected builder callback for 'cir.try' body"); + assert(handlersBuilder + && "expected builder callback for 'handlers' body"); + + OpBuilder::InsertionGuard guard($_builder); + + // Try body region + mlir::Region *tryBodyRegion = $_state.addRegion(); + + // Create try body region and set insertion point + $_builder.createBlock(tryBodyRegion); + tryBuilder($_builder, $_state.location); + handlersBuilder($_builder, $_state.location, $_state); + }]> + ]; + + let hasLLVMLowering = false; +} + +//===----------------------------------------------------------------------===// // Atomic operations //===----------------------------------------------------------------------===// -def CIR_AtomicXchg : CIR_Op<"atomic.xchg", [ +def CIR_AtomicXchgOp : CIR_Op<"atomic.xchg", [ AllTypesMatch<["result", "val"]>, TypesMatchWith<"type of 'val' must match the pointee type of 'ptr'", "ptr", "val", "mlir::cast<cir::PointerType>($_self).getPointee()"> @@ -4318,9 +4472,7 @@ def CIR_AtomicXchg : CIR_Op<"atomic.xchg", [ Example: ```mlir - %res = cir.atomic.xchg(%ptr : !cir.ptr<!u64i>, - %val : !u64i, - seq_cst) : !u64i + %res = cir.atomic.xchg seq_cst %ptr, %val : !cir.ptr<!u64i> -> !u64i ``` }]; @@ -4335,12 +4487,16 @@ def CIR_AtomicXchg : CIR_Op<"atomic.xchg", [ let assemblyFormat = [{ $mem_order (`volatile` $is_volatile^)? $ptr `,` $val - `:` qualified(type($ptr)) `->` type($result) attr-dict + `:` functional-type(operands, results) attr-dict }]; } -def CIR_AtomicCmpXchg : CIR_Op<"atomic.cmpxchg", [ - AllTypesMatch<["old", "expected", "desired"]> +def CIR_AtomicCmpXchgOp : CIR_Op<"atomic.cmpxchg", [ + AllTypesMatch<["old", "expected", "desired"]>, + TypesMatchWith<"type of 'expected' must match the pointee type of 'ptr'", + "ptr", "expected", "mlir::cast<cir::PointerType>($_self).getPointee()">, + TypesMatchWith<"type of 'desired' must match the pointee type of 'ptr'", + "ptr", "desired", "mlir::cast<cir::PointerType>($_self).getPointee()"> ]> { let summary = "Atomic compare and exchange"; let description = [{ @@ -4373,12 +4529,9 @@ def CIR_AtomicCmpXchg : CIR_Op<"atomic.cmpxchg", [ Example: ```mlir - %old, %success = cir.atomic.cmpxchg(%ptr : !cir.ptr<!u64i>, - %expected : !u64i, - %desired : !u64i, - success = seq_cst, - failure = seq_cst) weak - : (!u64i, !cir.bool) + %old, %success = cir.atomic.cmpxchg weak success(seq_cst) failure(acquire) + %ptr, %expected, %desired + : (!cir.ptr<!u64i>, !u64i, !u64i) -> (!u64i, !cir.bool) ``` }]; let results = (outs CIR_AnyType:$old, CIR_BoolType:$success); @@ -4392,20 +4545,13 @@ def CIR_AtomicCmpXchg : CIR_Op<"atomic.cmpxchg", [ UnitAttr:$is_volatile); let assemblyFormat = [{ - `(` - $ptr `:` qualified(type($ptr)) `,` - $expected `:` type($expected) `,` - $desired `:` type($desired) `,` - `success` `=` $succ_order `,` - `failure` `=` $fail_order - `)` - (`align` `(` $alignment^ `)`)? (`weak` $weak^)? + `success` `(` $succ_order `)` `failure` `(` $fail_order `)` + $ptr `,` $expected `,` $desired + (`align` `(` $alignment^ `)`)? (`volatile` $is_volatile^)? - `:` `(` type($old) `,` type($success) `)` attr-dict + `:` functional-type(operands, results) attr-dict }]; - - let hasVerifier = 1; } #endif // CLANG_CIR_DIALECT_IR_CIROPS_TD diff --git a/clang/include/clang/CIR/Dialect/IR/CIROpsEnums.h b/clang/include/clang/CIR/Dialect/IR/CIROpsEnums.h index 17fddae..dbd0304 100644 --- a/clang/include/clang/CIR/Dialect/IR/CIROpsEnums.h +++ b/clang/include/clang/CIR/Dialect/IR/CIROpsEnums.h @@ -54,10 +54,10 @@ static bool isLocalLinkage(GlobalLinkageKind linkage) { static bool isExternalWeakLinkage(GlobalLinkageKind linkage) { return linkage == GlobalLinkageKind::ExternalWeakLinkage; } -LLVM_ATTRIBUTE_UNUSED static bool isCommonLinkage(GlobalLinkageKind linkage) { +[[maybe_unused]] static bool isCommonLinkage(GlobalLinkageKind linkage) { return linkage == GlobalLinkageKind::CommonLinkage; } -LLVM_ATTRIBUTE_UNUSED static bool +[[maybe_unused]] static bool isValidDeclarationLinkage(GlobalLinkageKind linkage) { return isExternalWeakLinkage(linkage) || isExternalLinkage(linkage); } @@ -65,8 +65,7 @@ isValidDeclarationLinkage(GlobalLinkageKind linkage) { /// Whether the definition of this global may be replaced by something /// non-equivalent at link time. For example, if a function has weak linkage /// then the code defining it may be replaced by different code. -LLVM_ATTRIBUTE_UNUSED static bool -isInterposableLinkage(GlobalLinkageKind linkage) { +[[maybe_unused]] static bool isInterposableLinkage(GlobalLinkageKind linkage) { switch (linkage) { case GlobalLinkageKind::WeakAnyLinkage: case GlobalLinkageKind::LinkOnceAnyLinkage: @@ -89,8 +88,7 @@ isInterposableLinkage(GlobalLinkageKind linkage) { /// Whether the definition of this global may be discarded if it is not used /// in its compilation unit. -LLVM_ATTRIBUTE_UNUSED static bool -isDiscardableIfUnused(GlobalLinkageKind linkage) { +[[maybe_unused]] static bool isDiscardableIfUnused(GlobalLinkageKind linkage) { return isLinkOnceLinkage(linkage) || isLocalLinkage(linkage) || isAvailableExternallyLinkage(linkage); } @@ -99,7 +97,7 @@ isDiscardableIfUnused(GlobalLinkageKind linkage) { /// Using this method outside of the code generators is almost always a /// mistake: when working at the IR level use isInterposable instead as it /// knows about ODR semantics. -LLVM_ATTRIBUTE_UNUSED static bool isWeakForLinker(GlobalLinkageKind linkage) { +[[maybe_unused]] static bool isWeakForLinker(GlobalLinkageKind linkage) { return linkage == GlobalLinkageKind::WeakAnyLinkage || linkage == GlobalLinkageKind::WeakODRLinkage || linkage == GlobalLinkageKind::LinkOnceAnyLinkage || @@ -108,7 +106,7 @@ LLVM_ATTRIBUTE_UNUSED static bool isWeakForLinker(GlobalLinkageKind linkage) { linkage == GlobalLinkageKind::ExternalWeakLinkage; } -LLVM_ATTRIBUTE_UNUSED static bool isValidLinkage(GlobalLinkageKind gl) { +[[maybe_unused]] static bool isValidLinkage(GlobalLinkageKind gl) { return isExternalLinkage(gl) || isLocalLinkage(gl) || isWeakLinkage(gl) || isLinkOnceLinkage(gl); } diff --git a/clang/include/clang/CIR/MissingFeatures.h b/clang/include/clang/CIR/MissingFeatures.h index df82ca1..090cf35 100644 --- a/clang/include/clang/CIR/MissingFeatures.h +++ b/clang/include/clang/CIR/MissingFeatures.h @@ -39,7 +39,6 @@ struct MissingFeatures { static bool opGlobalUsedOrCompilerUsed() { return false; } static bool opGlobalAnnotations() { return false; } static bool opGlobalCtorPriority() { return false; } - static bool opGlobalDtorList() { return false; } static bool setDSOLocal() { return false; } static bool setComdat() { return false; } @@ -70,24 +69,31 @@ struct MissingFeatures { static bool opAllocaCaptureByInit() { return false; } // FuncOp handling - static bool opFuncOpenCLKernelMetadata() { return false; } + static bool opFuncArmNewAttr() { return false; } + static bool opFuncArmStreamingAttr() { return false; } static bool opFuncAstDeclAttr() { return false; } - static bool opFuncAttributesForDefinition() { return false; } static bool opFuncCallingConv() { return false; } + static bool opFuncColdHotAttr() { return false; } static bool opFuncCPUAndFeaturesAttributes() { return false; } static bool opFuncExceptions() { return false; } static bool opFuncExtraAttrs() { return false; } static bool opFuncMaybeHandleStaticInExternC() { return false; } + static bool opFuncMinSizeAttr() { return false; } static bool opFuncMultipleReturnVals() { return false; } + static bool opFuncNakedAttr() { return false; } + static bool opFuncNoDuplicateAttr() { return false; } static bool opFuncNoUnwind() { return false; } + static bool opFuncOpenCLKernelMetadata() { return false; } static bool opFuncOperandBundles() { return false; } + static bool opFuncOptNoneAttr() { return false; } static bool opFuncParameterAttributes() { return false; } static bool opFuncReadOnly() { return false; } static bool opFuncSection() { return false; } + static bool opFuncUnwindTablesAttr() { return false; } static bool opFuncWillReturn() { return false; } static bool opFuncNoReturn() { return false; } - static bool setLLVMFunctionFEnvAttributes() { return false; } static bool setFunctionAttributes() { return false; } + static bool setLLVMFunctionFEnvAttributes() { return false; } // CallOp handling static bool opCallAggregateArgs() { return false; } @@ -175,6 +181,10 @@ struct MissingFeatures { static bool atomicScope() { return false; } static bool atomicSyncScopeID() { return false; } + // Global ctor handling + static bool globalCtorLexOrder() { return false; } + static bool globalCtorAssociatedData() { return false; } + // Misc static bool abiArgInfo() { return false; } static bool addHeapAllocSiteMetadata() { return false; } @@ -268,6 +278,7 @@ struct MissingFeatures { static bool objCBlocks() { return false; } static bool objCGC() { return false; } static bool objCLifetime() { return false; } + static bool hlsl() { return false; } static bool openCL() { return false; } static bool openMP() { return false; } static bool opTBAA() { return false; } @@ -285,6 +296,7 @@ struct MissingFeatures { static bool sourceLanguageCases() { return false; } static bool stackBase() { return false; } static bool stackSaveOp() { return false; } + static bool stackProtector() { return false; } static bool targetCIRGenInfoArch() { return false; } static bool targetCIRGenInfoOS() { return false; } static bool targetCodeGenInfoGetNullPointer() { return false; } @@ -319,6 +331,7 @@ struct MissingFeatures { static bool invokeOp() { return false; } static bool labelOp() { return false; } static bool ptrDiffOp() { return false; } + static bool llvmLoweringPtrDiffConsidersPointee() { return false; } static bool ptrStrideOp() { return false; } static bool switchOp() { return false; } static bool throwOp() { return false; } diff --git a/clang/include/clang/Driver/Distro.h b/clang/include/clang/Driver/Distro.h index 008de0d..a515cbf 100644 --- a/clang/include/clang/Driver/Distro.h +++ b/clang/include/clang/Driver/Distro.h @@ -79,6 +79,7 @@ public: UbuntuOracular, UbuntuPlucky, UbuntuQuesting, + UbuntuResolute, UnknownDistro }; @@ -130,7 +131,7 @@ public: } bool IsUbuntu() const { - return DistroVal >= UbuntuMaverick && DistroVal <= UbuntuQuesting; + return DistroVal >= UbuntuMaverick && DistroVal <= UbuntuResolute; } bool IsAlpineLinux() const { return DistroVal == AlpineLinux; } diff --git a/clang/include/clang/Driver/Options.td b/clang/include/clang/Driver/Options.td index a55a523..7ae153d 100644 --- a/clang/include/clang/Driver/Options.td +++ b/clang/include/clang/Driver/Options.td @@ -2047,7 +2047,7 @@ def fconstexpr_depth_EQ : Joined<["-"], "fconstexpr-depth=">, Group<f_Group>, MarshallingInfoInt<LangOpts<"ConstexprCallDepth">, "512">; def fconstexpr_steps_EQ : Joined<["-"], "fconstexpr-steps=">, Group<f_Group>, Visibility<[ClangOption, CC1Option]>, - HelpText<"Set the maximum number of steps in constexpr function evaluation">, + HelpText<"Set the maximum number of steps in constexpr function evaluation (0 = no limit)">, MarshallingInfoInt<LangOpts<"ConstexprStepLimit">, "1048576">; def fexperimental_new_constant_interpreter : Flag<["-"], "fexperimental-new-constant-interpreter">, Group<f_Group>, HelpText<"Enable the experimental new constant interpreter">, @@ -6122,7 +6122,7 @@ def rewrite_legacy_objc : Flag<["-"], "rewrite-legacy-objc">, def rdynamic : Flag<["-"], "rdynamic">, Group<Link_Group>, Visibility<[ClangOption, FlangOption]>; def resource_dir : Separate<["-"], "resource-dir">, - Flags<[NoXarchOption, HelpHidden]>, + Flags<[NoXarchOption]>, Visibility<[ClangOption, CC1Option, CLOption, DXCOption, FlangOption, FC1Option]>, HelpText<"The directory which holds the compiler resource files">, MarshallingInfoString<HeaderSearchOpts<"ResourceDir">>; @@ -6211,11 +6211,12 @@ def static : Flag<["-", "--"], "static">, Group<Link_Group>, Flags<[NoArgumentUnused]>; def std_default_EQ : Joined<["-"], "std-default=">; def std_EQ : Joined<["-", "--"], "std=">, - Visibility<[ClangOption, CC1Option, FlangOption, FC1Option]>, - Group<CompileOnly_Group>, HelpText<"Language standard to compile for">, - ValuesCode<[{ + Visibility<[ClangOption, CC1Option, FlangOption, FC1Option]>, + Group<CompileOnly_Group>, + HelpText<"Language standard to compile for">, + ValuesCode<[{ static constexpr const char VALUES_CODE [] = - #define LANGSTANDARD(id, name, lang, desc, features) name "," + #define LANGSTANDARD(id, name, lang, desc, features, version) name "," #define LANGSTANDARD_ALIAS(id, alias) alias "," #include "clang/Basic/LangStandards.def" ; @@ -9600,6 +9601,15 @@ def fhlsl_spv_use_unknown_image_format "from the resource's data type.">, MarshallingInfoFlag<LangOpts<"HLSLSpvUseUnknownImageFormat">>; +def fhlsl_spv_enable_maximal_reconvergence + : Flag<["-"], "fspv-enable-maximal-reconvergence">, + Group<dxc_Group>, + Visibility<[CC1Option, DXCOption]>, + HelpText<"Enables the MaximallyReconvergesKHR execution mode for this " + "module. This ensures that control flow reconverges at " + "well-defined merge points as defined by the Vulkan spec.">, + MarshallingInfoFlag<LangOpts<"HLSLSpvEnableMaximalReconvergence">>; + def no_wasm_opt : Flag<["--"], "no-wasm-opt">, Group<m_Group>, HelpText<"Disable the wasm-opt optimizer">, diff --git a/clang/include/clang/ExtractAPI/API.h b/clang/include/clang/ExtractAPI/API.h index 1ace535..bea5416 100644 --- a/clang/include/clang/ExtractAPI/API.h +++ b/clang/include/clang/ExtractAPI/API.h @@ -618,17 +618,17 @@ struct TagRecord : APIRecord, RecordContext { static bool classofKind(RecordKind K) { switch (K) { case RK_Enum: - LLVM_FALLTHROUGH; + [[fallthrough]]; case RK_Struct: - LLVM_FALLTHROUGH; + [[fallthrough]]; case RK_Union: - LLVM_FALLTHROUGH; + [[fallthrough]]; case RK_CXXClass: - LLVM_FALLTHROUGH; + [[fallthrough]]; case RK_ClassTemplate: - LLVM_FALLTHROUGH; + [[fallthrough]]; case RK_ClassTemplateSpecialization: - LLVM_FALLTHROUGH; + [[fallthrough]]; case RK_ClassTemplatePartialSpecialization: return true; default: @@ -704,15 +704,15 @@ struct RecordRecord : TagRecord { static bool classofKind(RecordKind K) { switch (K) { case RK_Struct: - LLVM_FALLTHROUGH; + [[fallthrough]]; case RK_Union: - LLVM_FALLTHROUGH; + [[fallthrough]]; case RK_CXXClass: - LLVM_FALLTHROUGH; + [[fallthrough]]; case RK_ClassTemplate: - LLVM_FALLTHROUGH; + [[fallthrough]]; case RK_ClassTemplateSpecialization: - LLVM_FALLTHROUGH; + [[fallthrough]]; case RK_ClassTemplatePartialSpecialization: return true; default: diff --git a/clang/include/clang/Sema/Sema.h b/clang/include/clang/Sema/Sema.h index 37598f8..e3faaac 100644 --- a/clang/include/clang/Sema/Sema.h +++ b/clang/include/clang/Sema/Sema.h @@ -5010,6 +5010,14 @@ public: void AddLaunchBoundsAttr(Decl *D, const AttributeCommonInfo &CI, Expr *MaxThreads, Expr *MinBlocks, Expr *MaxBlocks); + /// Add a cluster_dims attribute to a particular declaration. + CUDAClusterDimsAttr *createClusterDimsAttr(const AttributeCommonInfo &CI, + Expr *X, Expr *Y, Expr *Z); + void addClusterDimsAttr(Decl *D, const AttributeCommonInfo &CI, Expr *X, + Expr *Y, Expr *Z); + /// Add a no_cluster attribute to a particular declaration. + void addNoClusterAttr(Decl *D, const AttributeCommonInfo &CI); + enum class RetainOwnershipKind { NS, CF, OS }; UuidAttr *mergeUuidAttr(Decl *D, const AttributeCommonInfo &CI, @@ -13385,6 +13393,13 @@ public: const MultiLevelTemplateArgumentList &TemplateArgs, TemplateArgumentListInfo &Outputs); + /// Substitute concept template arguments in the constraint expression + /// of a concept-id. This is used to implement [temp.constr.normal]. + ExprResult + SubstConceptTemplateArguments(const ConceptSpecializationExpr *CSE, + const Expr *ConstraintExpr, + const MultiLevelTemplateArgumentList &MLTAL); + bool SubstTemplateArgumentsInParameterMapping( ArrayRef<TemplateArgumentLoc> Args, SourceLocation BaseLoc, const MultiLevelTemplateArgumentList &TemplateArgs, diff --git a/clang/include/clang/Sema/SemaHLSL.h b/clang/include/clang/Sema/SemaHLSL.h index 46b088c..f9d3a4ea 100644 --- a/clang/include/clang/Sema/SemaHLSL.h +++ b/clang/include/clang/Sema/SemaHLSL.h @@ -215,7 +215,6 @@ public: bool diagnosePositionType(QualType T, const ParsedAttr &AL); bool CanPerformScalarCast(QualType SrcTy, QualType DestTy); - bool ContainsBitField(QualType BaseTy); bool CanPerformElementwiseCast(Expr *Src, QualType DestType); bool CanPerformAggregateSplatCast(Expr *Src, QualType DestType); ExprResult ActOnOutParamExpr(ParmVarDecl *Param, Expr *Arg); diff --git a/clang/include/clang/Sema/SemaOpenACC.h b/clang/include/clang/Sema/SemaOpenACC.h index 6cadc34..f751e98 100644 --- a/clang/include/clang/Sema/SemaOpenACC.h +++ b/clang/include/clang/Sema/SemaOpenACC.h @@ -228,6 +228,11 @@ private: bool DiagnoseAllowedClauses(OpenACCDirectiveKind DK, OpenACCClauseKind CK, SourceLocation ClauseLoc); + bool CreateReductionCombinerRecipe( + SourceLocation loc, OpenACCReductionOperator ReductionOperator, + QualType VarTy, + llvm::SmallVectorImpl<OpenACCReductionRecipe::CombinerRecipe> + &CombinerRecipes); public: // Needed from the visitor, so should be public. @@ -240,7 +245,7 @@ public: OpenACCPrivateRecipe CreatePrivateInitRecipe(const Expr *VarExpr); OpenACCFirstPrivateRecipe CreateFirstPrivateInitRecipe(const Expr *VarExpr); - OpenACCReductionRecipe + OpenACCReductionRecipeWithStorage CreateReductionInitRecipe(OpenACCReductionOperator ReductionOperator, const Expr *VarExpr); @@ -946,12 +951,14 @@ public: ArrayRef<Expr *> IntExprs, SourceLocation EndLoc); // Does the checking for a 'reduction ' clause that needs to be done in // dependent and not dependent cases. - OpenACCClause *CheckReductionClause( - ArrayRef<const OpenACCClause *> ExistingClauses, - OpenACCDirectiveKind DirectiveKind, SourceLocation BeginLoc, - SourceLocation LParenLoc, OpenACCReductionOperator ReductionOp, - ArrayRef<Expr *> Vars, ArrayRef<OpenACCReductionRecipe> Recipes, - SourceLocation EndLoc); + OpenACCClause * + CheckReductionClause(ArrayRef<const OpenACCClause *> ExistingClauses, + OpenACCDirectiveKind DirectiveKind, + SourceLocation BeginLoc, SourceLocation LParenLoc, + OpenACCReductionOperator ReductionOp, + ArrayRef<Expr *> Vars, + ArrayRef<OpenACCReductionRecipeWithStorage> Recipes, + SourceLocation EndLoc); ExprResult BuildOpenACCAsteriskSizeExpr(SourceLocation AsteriskLoc); ExprResult ActOnOpenACCAsteriskSizeExpr(SourceLocation AsteriskLoc); diff --git a/clang/include/clang/Sema/SemaOpenMP.h b/clang/include/clang/Sema/SemaOpenMP.h index daf58b1..f9baeed 100644 --- a/clang/include/clang/Sema/SemaOpenMP.h +++ b/clang/include/clang/Sema/SemaOpenMP.h @@ -1022,7 +1022,8 @@ public: SourceLocation EndLoc); /// Called on well-formed 'nowait' clause. OMPClause *ActOnOpenMPNowaitClause(SourceLocation StartLoc, - SourceLocation EndLoc); + SourceLocation EndLoc, + SourceLocation LParenLoc, Expr *Condition); /// Called on well-formed 'untied' clause. OMPClause *ActOnOpenMPUntiedClause(SourceLocation StartLoc, SourceLocation EndLoc); diff --git a/clang/include/clang/Sema/Template.h b/clang/include/clang/Sema/Template.h index 60c7d27..e963439b 100644 --- a/clang/include/clang/Sema/Template.h +++ b/clang/include/clang/Sema/Template.h @@ -205,8 +205,8 @@ enum class TemplateSubstitutionKind : char { /// Add a new outmost level to the multi-level template argument /// list. - /// A 'Final' substitution means that Subst* nodes won't be built - /// for the replacements. + /// A 'Final' substitution means that these Args don't need to be + /// resugared later. void addOuterTemplateArguments(Decl *AssociatedDecl, ArgList Args, bool Final) { assert(!NumRetainedOuterLevels && diff --git a/clang/include/clang/StaticAnalyzer/Checkers/Checkers.td b/clang/include/clang/StaticAnalyzer/Checkers/Checkers.td index 4473c54..ffae3b9 100644 --- a/clang/include/clang/StaticAnalyzer/Checkers/Checkers.td +++ b/clang/include/clang/StaticAnalyzer/Checkers/Checkers.td @@ -76,9 +76,6 @@ def SecurityAlpha : Package<"security">, ParentPackage<Alpha>; def CERT : Package<"cert">, ParentPackage<Security>; def ENV : Package<"env">, ParentPackage<CERT>; -def CERTAlpha : Package<"cert">, ParentPackage<SecurityAlpha>; -def POSAlpha : Package<"pos">, ParentPackage<CERTAlpha>; - def Unix : Package<"unix">; def UnixAlpha : Package<"unix">, ParentPackage<Alpha>; def CString : Package<"cstring">, ParentPackage<Unix>; @@ -195,6 +192,11 @@ def NullDereferenceChecker HelpText<"Check for dereferences of null pointers">, Documentation<HasDocumentation>; +def NullPointerArithmChecker + : Checker<"NullPointerArithm">, + HelpText<"Check for undefined arithmetic operations on null pointers">, + Documentation<HasDocumentation>; + def NonNullParamChecker : Checker<"NonNullParamChecker">, HelpText<"Check for null pointers passed as arguments to a function whose " "arguments are references or marked with the 'nonnull' attribute">, diff --git a/clang/include/clang/StaticAnalyzer/Core/PathSensitive/AnalysisManager.h b/clang/include/clang/StaticAnalyzer/Core/PathSensitive/AnalysisManager.h index e3cf1ba..1e87b47 100644 --- a/clang/include/clang/StaticAnalyzer/Core/PathSensitive/AnalysisManager.h +++ b/clang/include/clang/StaticAnalyzer/Core/PathSensitive/AnalysisManager.h @@ -140,8 +140,8 @@ public: // It might be great to reuse FrontendOptions::getInputKindForExtension() // but for now it doesn't discriminate between code and header files. return llvm::StringSwitch<bool>(SM.getFilename(SL).rsplit('.').second) - .Cases("c", "m", "mm", "C", "cc", "cp", true) - .Cases("cpp", "CPP", "c++", "cxx", "cppm", true) + .Cases({"c", "m", "mm", "C", "cc", "cp"}, true) + .Cases({"cpp", "CPP", "c++", "cxx", "cppm"}, true) .Default(false); } diff --git a/clang/include/clang/StaticAnalyzer/Core/PathSensitive/EntryPointStats.h b/clang/include/clang/StaticAnalyzer/Core/PathSensitive/EntryPointStats.h index 389f17d..f3c6a62 100644 --- a/clang/include/clang/StaticAnalyzer/Core/PathSensitive/EntryPointStats.h +++ b/clang/include/clang/StaticAnalyzer/Core/PathSensitive/EntryPointStats.h @@ -9,6 +9,7 @@ #ifndef CLANG_INCLUDE_CLANG_STATICANALYZER_CORE_PATHSENSITIVE_ENTRYPOINTSTATS_H #define CLANG_INCLUDE_CLANG_STATICANALYZER_CORE_PATHSENSITIVE_ENTRYPOINTSTATS_H +#include "clang/AST/ASTContext.h" #include "llvm/ADT/Statistic.h" #include "llvm/ADT/StringRef.h" @@ -25,7 +26,7 @@ class EntryPointStat { public: llvm::StringLiteral name() const { return Name; } - static void lockRegistry(llvm::StringRef CPPFileName); + static void lockRegistry(llvm::StringRef CPPFileName, ASTContext &Ctx); static void takeSnapshot(const Decl *EntryPoint); static void dumpStatsAsCSV(llvm::raw_ostream &OS); @@ -85,7 +86,7 @@ class UnsignedEPStat : public EntryPointStat { public: explicit UnsignedEPStat(llvm::StringLiteral Name); - unsigned value() const { return Value.value_or(0); } + std::optional<unsigned> value() const { return Value; } void reset() { Value.reset(); } void set(unsigned V) { assert(!Value.has_value()); diff --git a/clang/lib/AST/ASTContext.cpp b/clang/lib/AST/ASTContext.cpp index a8b41ba..32c8f62 100644 --- a/clang/lib/AST/ASTContext.cpp +++ b/clang/lib/AST/ASTContext.cpp @@ -563,8 +563,7 @@ comments::FullComment *ASTContext::getCommentForDecl( // does not have one of its own. QualType QT = TD->getUnderlyingType(); if (const auto *TT = QT->getAs<TagType>()) - if (comments::FullComment *FC = - getCommentForDecl(TT->getOriginalDecl(), PP)) + if (comments::FullComment *FC = getCommentForDecl(TT->getDecl(), PP)) return cloneFullComment(FC, D); } else if (const auto *IC = dyn_cast<ObjCInterfaceDecl>(D)) { @@ -2387,7 +2386,7 @@ TypeInfo ASTContext::getTypeInfoImpl(const Type *T) const { case Type::Record: case Type::Enum: { const auto *TT = cast<TagType>(T); - const TagDecl *TD = TT->getOriginalDecl()->getDefinitionOrSelf(); + const TagDecl *TD = TT->getDecl()->getDefinitionOrSelf(); if (TD->isInvalidDecl()) { Width = 8; @@ -2531,7 +2530,7 @@ unsigned ASTContext::getTypeUnadjustedAlign(const Type *T) const { unsigned UnadjustedAlign; if (const auto *RT = T->getAsCanonical<RecordType>()) { - const ASTRecordLayout &Layout = getASTRecordLayout(RT->getOriginalDecl()); + const ASTRecordLayout &Layout = getASTRecordLayout(RT->getDecl()); UnadjustedAlign = toBits(Layout.getUnadjustedAlignment()); } else if (const auto *ObjCI = T->getAsCanonical<ObjCInterfaceType>()) { const ASTRecordLayout &Layout = getASTObjCInterfaceLayout(ObjCI->getDecl()); @@ -3469,7 +3468,7 @@ static void encodeTypeForFunctionPointerAuth(const ASTContext &Ctx, llvm_unreachable("should never get here"); } case Type::Record: { - const RecordDecl *RD = T->castAsCanonical<RecordType>()->getOriginalDecl(); + const RecordDecl *RD = T->castAsCanonical<RecordType>()->getDecl(); const IdentifierInfo *II = RD->getIdentifier(); // In C++, an immediate typedef of an anonymous struct or union @@ -4713,8 +4712,8 @@ QualType ASTContext::getConstantMatrixType(QualType ElementTy, unsigned NumRows, assert(MatrixType::isValidElementType(ElementTy) && "need a valid element type"); - assert(ConstantMatrixType::isDimensionValid(NumRows) && - ConstantMatrixType::isDimensionValid(NumColumns) && + assert(NumRows > 0 && NumRows <= LangOpts.MaxMatrixDimension && + NumColumns > 0 && NumColumns <= LangOpts.MaxMatrixDimension && "need valid matrix dimensions"); void *InsertPos = nullptr; if (ConstantMatrixType *MTP = MatrixTypes.FindNodeOrInsertPos(ID, InsertPos)) @@ -5377,7 +5376,7 @@ TagType *ASTContext::getTagTypeInternal(ElaboratedTypeKeyword Keyword, }(); assert(T->getKeyword() == Keyword); assert(T->getQualifier() == Qualifier); - assert(T->getOriginalDecl() == TD); + assert(T->getDecl() == TD); assert(T->isInjected() == IsInjected); assert(T->isTagOwned() == OwnsTag); assert((T->isCanonicalUnqualified() @@ -8271,7 +8270,7 @@ Qualifiers::ObjCLifetime ASTContext::getInnerObjCOwnership(QualType T) const { static const Type *getIntegerTypeForEnum(const EnumType *ET) { // Incomplete enum types are not treated as integer types. // FIXME: In C++, enum types are never integer types. - const EnumDecl *ED = ET->getOriginalDecl()->getDefinitionOrSelf(); + const EnumDecl *ED = ET->getDecl()->getDefinitionOrSelf(); if (ED->isComplete() && !ED->isScoped()) return ED->getIntegerType().getTypePtr(); return nullptr; @@ -9189,7 +9188,7 @@ static void EncodeBitField(const ASTContext *Ctx, std::string& S, S += llvm::utostr(Offset); if (const auto *ET = T->getAsCanonical<EnumType>()) - S += ObjCEncodingForEnumDecl(Ctx, ET->getOriginalDecl()); + S += ObjCEncodingForEnumDecl(Ctx, ET->getDecl()); else { const auto *BT = T->castAs<BuiltinType>(); S += getObjCEncodingForPrimitiveType(Ctx, BT); @@ -9246,7 +9245,7 @@ void ASTContext::getObjCEncodingForTypeImpl(QualType T, std::string &S, if (const auto *BT = dyn_cast<BuiltinType>(CT)) S += getObjCEncodingForPrimitiveType(this, BT); else - S += ObjCEncodingForEnumDecl(this, cast<EnumType>(CT)->getOriginalDecl()); + S += ObjCEncodingForEnumDecl(this, cast<EnumType>(CT)->getDecl()); return; case Type::Complex: @@ -9314,7 +9313,7 @@ void ASTContext::getObjCEncodingForTypeImpl(QualType T, std::string &S, return; } } else if (const auto *RTy = PointeeTy->getAsCanonical<RecordType>()) { - const IdentifierInfo *II = RTy->getOriginalDecl()->getIdentifier(); + const IdentifierInfo *II = RTy->getDecl()->getIdentifier(); // GCC binary compat: Need to convert "struct objc_class *" to "#". if (II == &Idents.get("objc_class")) { S += '#'; @@ -9386,7 +9385,7 @@ void ASTContext::getObjCEncodingForTypeImpl(QualType T, std::string &S, return; case Type::Record: { - RecordDecl *RDecl = cast<RecordType>(CT)->getOriginalDecl(); + RecordDecl *RDecl = cast<RecordType>(CT)->getDecl(); S += RDecl->isUnion() ? '(' : '{'; // Anonymous structures print as '?' if (const IdentifierInfo *II = RDecl->getIdentifier()) { @@ -11290,7 +11289,7 @@ QualType ASTContext::mergeTransparentUnionType(QualType T, QualType SubType, bool OfBlockPointer, bool Unqualified) { if (const RecordType *UT = T->getAsUnionType()) { - RecordDecl *UD = UT->getOriginalDecl()->getMostRecentDecl(); + RecordDecl *UD = UT->getDecl()->getMostRecentDecl(); if (UD->hasAttr<TransparentUnionAttr>()) { for (const auto *I : UD->fields()) { QualType ET = I->getType().getUnqualifiedType(); @@ -11560,7 +11559,7 @@ static QualType mergeEnumWithInteger(ASTContext &Context, const EnumType *ET, // Compatibility is based on the underlying type, not the promotion // type. QualType underlyingType = - ET->getOriginalDecl()->getDefinitionOrSelf()->getIntegerType(); + ET->getDecl()->getDefinitionOrSelf()->getIntegerType(); if (underlyingType.isNull()) return {}; if (Context.hasSameType(underlyingType, other)) @@ -11581,6 +11580,12 @@ QualType ASTContext::mergeTagDefinitions(QualType LHS, QualType RHS) { if (LangOpts.CPlusPlus || !LangOpts.C23) return {}; + // Nameless tags are comparable only within outer definitions. At the top + // level they are not comparable. + const TagDecl *LTagD = LHS->castAsTagDecl(), *RTagD = RHS->castAsTagDecl(); + if (!LTagD->getIdentifier() || !RTagD->getIdentifier()) + return {}; + // C23, on the other hand, requires the members to be "the same enough", so // we use a structural equivalence check. StructuralEquivalenceContext::NonEquivalentDeclSet NonEquivalentDecls; @@ -14154,11 +14159,10 @@ static QualType getCommonNonSugarTypeNode(const ASTContext &Ctx, const Type *X, case Type::Record: case Type::InjectedClassName: { const auto *TX = cast<TagType>(X), *TY = cast<TagType>(Y); - return Ctx.getTagType( - ::getCommonTypeKeyword(TX, TY, /*IsSame=*/false), - ::getCommonQualifier(Ctx, TX, TY, /*IsSame=*/false), - ::getCommonDeclChecked(TX->getOriginalDecl(), TY->getOriginalDecl()), - /*OwnedTag=*/false); + return Ctx.getTagType(::getCommonTypeKeyword(TX, TY, /*IsSame=*/false), + ::getCommonQualifier(Ctx, TX, TY, /*IsSame=*/false), + ::getCommonDeclChecked(TX->getDecl(), TY->getDecl()), + /*OwnedTag=*/false); } case Type::TemplateSpecialization: { const auto *TX = cast<TemplateSpecializationType>(X), diff --git a/clang/lib/AST/ASTDiagnostic.cpp b/clang/lib/AST/ASTDiagnostic.cpp index d7fd411..b8023cb 100644 --- a/clang/lib/AST/ASTDiagnostic.cpp +++ b/clang/lib/AST/ASTDiagnostic.cpp @@ -196,8 +196,7 @@ break; \ // Don't desugar through the primary typedef of an anonymous type. if (const TagType *UTT = Underlying->getAs<TagType>()) if (const TypedefType *QTT = dyn_cast<TypedefType>(QT)) - if (UTT->getOriginalDecl()->getTypedefNameForAnonDecl() == - QTT->getDecl()) + if (UTT->getDecl()->getTypedefNameForAnonDecl() == QTT->getDecl()) break; // Record that we actually looked through an opaque type here. @@ -1147,14 +1146,11 @@ class TemplateDiff { if (const auto* SubstType = Ty->getAs<SubstTemplateTypeParmType>()) Ty = SubstType->getReplacementType(); - const RecordType *RT = Ty->getAs<RecordType>(); - + const auto *RT = Ty->getAs<RecordType>(); if (!RT) return nullptr; - const ClassTemplateSpecializationDecl *CTSD = - dyn_cast<ClassTemplateSpecializationDecl>(RT->getOriginalDecl()); - + const auto *CTSD = dyn_cast<ClassTemplateSpecializationDecl>(RT->getDecl()); if (!CTSD) return nullptr; diff --git a/clang/lib/AST/ASTImporter.cpp b/clang/lib/AST/ASTImporter.cpp index f43fa8c..bf51c3e 100644 --- a/clang/lib/AST/ASTImporter.cpp +++ b/clang/lib/AST/ASTImporter.cpp @@ -1740,7 +1740,7 @@ ExpectedType ASTNodeImporter::VisitDeducedTemplateSpecializationType( } ExpectedType ASTNodeImporter::VisitTagType(const TagType *T) { - TagDecl *DeclForType = T->getOriginalDecl(); + TagDecl *DeclForType = T->getDecl(); Expected<TagDecl *> ToDeclOrErr = import(DeclForType); if (!ToDeclOrErr) return ToDeclOrErr.takeError(); @@ -2155,7 +2155,7 @@ Error ASTNodeImporter::ImportDeclParts( const Type *LeafT = getLeafPointeeType(P->getType().getCanonicalType().getTypePtr()); auto *RT = dyn_cast<RecordType>(LeafT); - if (RT && RT->getOriginalDecl() == D) { + if (RT && RT->getDecl() == D) { Importer.FromDiag(D->getLocation(), diag::err_unsupported_ast_node) << D->getDeclKindName(); return make_error<ASTImportError>(ASTImportError::UnsupportedConstruct); @@ -2408,8 +2408,8 @@ Error ASTNodeImporter::ImportFieldDeclDefinition(const FieldDecl *From, const RecordType *RecordTo = ToType->getAs<RecordType>(); if (RecordFrom && RecordTo) { - FromRecordDecl = RecordFrom->getOriginalDecl(); - ToRecordDecl = RecordTo->getOriginalDecl(); + FromRecordDecl = RecordFrom->getDecl(); + ToRecordDecl = RecordTo->getDecl(); } } @@ -3205,7 +3205,7 @@ ExpectedDecl ASTNodeImporter::VisitEnumDecl(EnumDecl *D) { if (auto *Typedef = dyn_cast<TypedefNameDecl>(FoundDecl)) { if (const auto *Tag = Typedef->getUnderlyingType()->getAs<TagType>()) - FoundDecl = Tag->getOriginalDecl(); + FoundDecl = Tag->getDecl(); } if (auto *FoundEnum = dyn_cast<EnumDecl>(FoundDecl)) { @@ -3336,7 +3336,7 @@ ExpectedDecl ASTNodeImporter::VisitRecordDecl(RecordDecl *D) { Decl *Found = FoundDecl; if (auto *Typedef = dyn_cast<TypedefNameDecl>(Found)) { if (const auto *Tag = Typedef->getUnderlyingType()->getAs<TagType>()) - Found = Tag->getOriginalDecl(); + Found = Tag->getDecl(); } if (auto *FoundRecord = dyn_cast<RecordDecl>(Found)) { @@ -3757,12 +3757,11 @@ public: } std::optional<bool> VisitTagType(const TagType *T) { - if (auto *Spec = - dyn_cast<ClassTemplateSpecializationDecl>(T->getOriginalDecl())) + if (auto *Spec = dyn_cast<ClassTemplateSpecializationDecl>(T->getDecl())) for (const auto &Arg : Spec->getTemplateArgs().asArray()) if (checkTemplateArgument(Arg)) return true; - return isAncestorDeclContextOf(ParentDC, T->getOriginalDecl()); + return isAncestorDeclContextOf(ParentDC, T->getDecl()); } std::optional<bool> VisitPointerType(const PointerType *T) { diff --git a/clang/lib/AST/ASTStructuralEquivalence.cpp b/clang/lib/AST/ASTStructuralEquivalence.cpp index 1557346..da64c92 100644 --- a/clang/lib/AST/ASTStructuralEquivalence.cpp +++ b/clang/lib/AST/ASTStructuralEquivalence.cpp @@ -878,10 +878,10 @@ static bool IsStructurallyEquivalent(StructuralEquivalenceContext &Context, // Treat the enumeration as its underlying type and use the builtin type // class comparison. if (T1->getTypeClass() == Type::Enum) { - T1 = cast<EnumType>(T1)->getOriginalDecl()->getIntegerType(); + T1 = cast<EnumType>(T1)->getDecl()->getIntegerType(); assert(T2->isBuiltinType() && !T1.isNull()); // Sanity check } else if (T2->getTypeClass() == Type::Enum) { - T2 = cast<EnumType>(T2)->getOriginalDecl()->getIntegerType(); + T2 = cast<EnumType>(T2)->getDecl()->getIntegerType(); assert(T1->isBuiltinType() && !T2.isNull()); // Sanity check } TC = Type::Builtin; @@ -1300,8 +1300,7 @@ static bool IsStructurallyEquivalent(StructuralEquivalenceContext &Context, if (!IsStructurallyEquivalent(Context, TT1->getQualifier(), TT2->getQualifier())) return false; - if (!IsStructurallyEquivalent(Context, TT1->getOriginalDecl(), - TT2->getOriginalDecl())) + if (!IsStructurallyEquivalent(Context, TT1->getDecl(), TT2->getDecl())) return false; break; } @@ -1531,8 +1530,8 @@ static bool IsStructurallyEquivalent(StructuralEquivalenceContext &Context, // types if (Field1->isAnonymousStructOrUnion() && Field2->isAnonymousStructOrUnion()) { - RecordDecl *D1 = Field1->getType()->castAs<RecordType>()->getOriginalDecl(); - RecordDecl *D2 = Field2->getType()->castAs<RecordType>()->getOriginalDecl(); + RecordDecl *D1 = Field1->getType()->castAs<RecordType>()->getDecl(); + RecordDecl *D2 = Field2->getType()->castAs<RecordType>()->getDecl(); return IsStructurallyEquivalent(Context, D1, D2); } @@ -1763,19 +1762,6 @@ static bool IsStructurallyEquivalent(StructuralEquivalenceContext &Context, // another anonymous structure or union, respectively, if their members // fulfill the preceding requirements. ... Otherwise, the structure, union, // or enumerated types are incompatible. - - // Note: "the same tag" refers to the identifier for the structure; two - // structures without names are not compatible within a TU. In C23, if either - // declaration has no name, they're not equivalent. However, the paragraph - // after the bulleted list goes on to talk about compatibility of anonymous - // structure and union members, so this prohibition only applies to top-level - // declarations; if either declaration is not a member, they cannot be - // compatible. - if (Context.LangOpts.C23 && (!D1->getIdentifier() || !D2->getIdentifier()) && - (!D1->getDeclContext()->isRecord() || !D2->getDeclContext()->isRecord())) - return false; - - // Otherwise, check the names for equivalence. if (!NameIsStructurallyEquivalent(*D1, *D2)) return false; @@ -2600,7 +2586,7 @@ StructuralEquivalenceContext::findUntaggedStructOrUnionIndex(RecordDecl *Anon) { // struct { ... } A; QualType FieldType = F->getType(); if (const auto *RecType = dyn_cast<RecordType>(FieldType)) { - const RecordDecl *RecDecl = RecType->getOriginalDecl(); + const RecordDecl *RecDecl = RecType->getDecl(); if (RecDecl->getDeclContext() == Owner && !RecDecl->getIdentifier()) { if (Context.hasSameType(FieldType, AnonTy)) break; diff --git a/clang/lib/AST/ByteCode/Compiler.cpp b/clang/lib/AST/ByteCode/Compiler.cpp index c71fd22..74cae03 100644 --- a/clang/lib/AST/ByteCode/Compiler.cpp +++ b/clang/lib/AST/ByteCode/Compiler.cpp @@ -4660,7 +4660,7 @@ const RecordType *Compiler<Emitter>::getRecordTy(QualType Ty) { template <class Emitter> Record *Compiler<Emitter>::getRecord(QualType Ty) { if (const auto *RecordTy = getRecordTy(Ty)) - return getRecord(RecordTy->getOriginalDecl()->getDefinitionOrSelf()); + return getRecord(RecordTy->getDecl()->getDefinitionOrSelf()); return nullptr; } diff --git a/clang/lib/AST/ByteCode/Interp.cpp b/clang/lib/AST/ByteCode/Interp.cpp index 8904396..a72282c 100644 --- a/clang/lib/AST/ByteCode/Interp.cpp +++ b/clang/lib/AST/ByteCode/Interp.cpp @@ -1358,9 +1358,6 @@ bool Free(InterpState &S, CodePtr OpPC, bool DeleteIsArrayForm, void diagnoseEnumValue(InterpState &S, CodePtr OpPC, const EnumDecl *ED, const APSInt &Value) { - if (S.EvaluatingDecl && !S.EvaluatingDecl->isConstexpr()) - return; - llvm::APInt Min; llvm::APInt Max; ED->getValueRange(Max, Min); diff --git a/clang/lib/AST/ByteCode/Interp.h b/clang/lib/AST/ByteCode/Interp.h index bb0c458..2f7e2d9 100644 --- a/clang/lib/AST/ByteCode/Interp.h +++ b/clang/lib/AST/ByteCode/Interp.h @@ -2258,6 +2258,8 @@ std::optional<Pointer> OffsetHelper(InterpState &S, CodePtr OpPC, S.CCEDiag(S.Current->getSource(OpPC), diag::note_constexpr_array_index) << N << /*non-array*/ true << 0; return Pointer(Ptr.asFunctionPointer().getFunction(), N); + } else if (!Ptr.isBlockPointer()) { + return std::nullopt; } assert(Ptr.isBlockPointer()); @@ -3096,7 +3098,8 @@ inline bool ArrayElemPtr(InterpState &S, CodePtr OpPC) { } if (Offset.isZero()) { - if (Ptr.getFieldDesc()->isArray() && Ptr.getIndex() == 0) { + if (const Descriptor *Desc = Ptr.getFieldDesc(); + Desc && Desc->isArray() && Ptr.getIndex() == 0) { S.Stk.push<Pointer>(Ptr.atIndex(0).narrow()); return true; } @@ -3126,7 +3129,8 @@ inline bool ArrayElemPtrPop(InterpState &S, CodePtr OpPC) { } if (Offset.isZero()) { - if (Ptr.getFieldDesc()->isArray() && Ptr.getIndex() == 0) { + if (const Descriptor *Desc = Ptr.getFieldDesc(); + Desc && Desc->isArray() && Ptr.getIndex() == 0) { S.Stk.push<Pointer>(Ptr.atIndex(0).narrow()); return true; } @@ -3183,9 +3187,8 @@ inline bool CopyArray(InterpState &S, CodePtr OpPC, uint32_t SrcIndex, if (!CheckLoad(S, OpPC, SP)) return false; - const Pointer &DP = DestPtr.atIndex(DestIndex + I); - DP.deref<T>() = SP.deref<T>(); - DP.initialize(); + DestPtr.elem<T>(DestIndex + I) = SrcPtr.elem<T>(SrcIndex + I); + DestPtr.initializeElement(DestIndex + I); } return true; } @@ -3700,7 +3703,7 @@ inline bool CheckDestruction(InterpState &S, CodePtr OpPC) { inline bool CheckArraySize(InterpState &S, CodePtr OpPC, uint64_t NumElems) { uint64_t Limit = S.getLangOpts().ConstexprStepLimit; - if (NumElems > Limit) { + if (Limit != 0 && NumElems > Limit) { S.FFDiag(S.Current->getSource(OpPC), diag::note_constexpr_new_exceeds_limits) << NumElems << Limit; diff --git a/clang/lib/AST/ByteCode/InterpBlock.cpp b/clang/lib/AST/ByteCode/InterpBlock.cpp index ac6f01f..24825ad 100644 --- a/clang/lib/AST/ByteCode/InterpBlock.cpp +++ b/clang/lib/AST/ByteCode/InterpBlock.cpp @@ -100,6 +100,19 @@ bool Block::hasPointer(const Pointer *P) const { } #endif +void Block::movePointersTo(Block *B) { + assert(B != this); + + while (Pointers) { + Pointer *P = Pointers; + + this->removePointer(P); + P->BS.Pointee = B; + B->addPointer(P); + } + assert(!this->hasPointers()); +} + DeadBlock::DeadBlock(DeadBlock *&Root, Block *Blk) : Root(Root), B(~0u, Blk->Desc, Blk->isExtern(), Blk->IsStatic, Blk->isWeak(), Blk->isDummy(), /*IsDead=*/true) { diff --git a/clang/lib/AST/ByteCode/InterpBlock.h b/clang/lib/AST/ByteCode/InterpBlock.h index 9b3dadc..73fdc8d 100644 --- a/clang/lib/AST/ByteCode/InterpBlock.h +++ b/clang/lib/AST/ByteCode/InterpBlock.h @@ -92,6 +92,8 @@ public: bool isInitialized() const { return IsInitialized; } /// The Evaluation ID this block was created in. unsigned getEvalID() const { return EvalID; } + /// Move all pointers from this block to \param B. + void movePointersTo(Block *B); /// Returns a pointer to the stored data. /// You are allowed to read Desc->getSize() bytes from this address. diff --git a/clang/lib/AST/ByteCode/InterpBuiltin.cpp b/clang/lib/AST/ByteCode/InterpBuiltin.cpp index 922d679..5838cf8 100644 --- a/clang/lib/AST/ByteCode/InterpBuiltin.cpp +++ b/clang/lib/AST/ByteCode/InterpBuiltin.cpp @@ -23,7 +23,7 @@ namespace clang { namespace interp { -LLVM_ATTRIBUTE_UNUSED static bool isNoopBuiltin(unsigned ID) { +[[maybe_unused]] static bool isNoopBuiltin(unsigned ID) { switch (ID) { case Builtin::BIas_const: case Builtin::BIforward: @@ -1633,8 +1633,8 @@ static bool interp__builtin_elementwise_countzeroes(InterpState &S, const InterpFrame *Frame, const CallExpr *Call, unsigned BuiltinID) { - const bool HasZeroArg = Call->getNumArgs() == 2; - const bool IsCTTZ = BuiltinID == Builtin::BI__builtin_elementwise_ctzg; + bool HasZeroArg = Call->getNumArgs() == 2; + bool IsCTTZ = BuiltinID == Builtin::BI__builtin_elementwise_ctzg; assert(Call->getNumArgs() == 1 || HasZeroArg); if (Call->getArg(0)->getType()->isIntegerType()) { PrimType ArgT = *S.getContext().classify(Call->getArg(0)->getType()); @@ -2447,18 +2447,18 @@ interp__builtin_x86_pack(InterpState &S, CodePtr, const CallExpr *E, const Pointer &Dst = S.Stk.peek<Pointer>(); const ASTContext &ASTCtx = S.getASTContext(); - const unsigned SrcBits = ASTCtx.getIntWidth(VT0->getElementType()); - const unsigned LHSVecLen = VT0->getNumElements(); - const unsigned SrcPerLane = 128 / SrcBits; - const unsigned Lanes = LHSVecLen * SrcBits / 128; + unsigned SrcBits = ASTCtx.getIntWidth(VT0->getElementType()); + unsigned LHSVecLen = VT0->getNumElements(); + unsigned SrcPerLane = 128 / SrcBits; + unsigned Lanes = LHSVecLen * SrcBits / 128; PrimType SrcT = *S.getContext().classify(VT0->getElementType()); PrimType DstT = *S.getContext().classify(getElemType(Dst)); - const bool IsUnsigend = getElemType(Dst)->isUnsignedIntegerType(); + bool IsUnsigend = getElemType(Dst)->isUnsignedIntegerType(); for (unsigned Lane = 0; Lane != Lanes; ++Lane) { - const unsigned BaseSrc = Lane * SrcPerLane; - const unsigned BaseDst = Lane * (2 * SrcPerLane); + unsigned BaseSrc = Lane * SrcPerLane; + unsigned BaseDst = Lane * (2 * SrcPerLane); for (unsigned I = 0; I != SrcPerLane; ++I) { INT_TYPE_SWITCH_NO_BOOL(SrcT, { @@ -2587,6 +2587,82 @@ static bool interp__builtin_ia32_pmul( return true; } +static bool interp_builtin_horizontal_int_binop( + InterpState &S, CodePtr OpPC, const CallExpr *Call, + llvm::function_ref<APInt(const APSInt &, const APSInt &)> Fn) { + const auto *VT = Call->getArg(0)->getType()->castAs<VectorType>(); + PrimType ElemT = *S.getContext().classify(VT->getElementType()); + bool DestUnsigned = Call->getType()->isUnsignedIntegerOrEnumerationType(); + + const Pointer &RHS = S.Stk.pop<Pointer>(); + const Pointer &LHS = S.Stk.pop<Pointer>(); + const Pointer &Dst = S.Stk.peek<Pointer>(); + unsigned NumElts = VT->getNumElements(); + unsigned EltBits = S.getASTContext().getIntWidth(VT->getElementType()); + unsigned EltsPerLane = 128 / EltBits; + unsigned Lanes = NumElts * EltBits / 128; + unsigned DestIndex = 0; + + for (unsigned Lane = 0; Lane < Lanes; ++Lane) { + unsigned LaneStart = Lane * EltsPerLane; + for (unsigned I = 0; I < EltsPerLane; I += 2) { + INT_TYPE_SWITCH_NO_BOOL(ElemT, { + APSInt Elem1 = LHS.elem<T>(LaneStart + I).toAPSInt(); + APSInt Elem2 = LHS.elem<T>(LaneStart + I + 1).toAPSInt(); + APSInt ResL = APSInt(Fn(Elem1, Elem2), DestUnsigned); + Dst.elem<T>(DestIndex++) = static_cast<T>(ResL); + }); + } + + for (unsigned I = 0; I < EltsPerLane; I += 2) { + INT_TYPE_SWITCH_NO_BOOL(ElemT, { + APSInt Elem1 = RHS.elem<T>(LaneStart + I).toAPSInt(); + APSInt Elem2 = RHS.elem<T>(LaneStart + I + 1).toAPSInt(); + APSInt ResR = APSInt(Fn(Elem1, Elem2), DestUnsigned); + Dst.elem<T>(DestIndex++) = static_cast<T>(ResR); + }); + } + } + Dst.initializeAllElements(); + return true; +} + +static bool interp_builtin_horizontal_fp_binop( + InterpState &S, CodePtr OpPC, const CallExpr *Call, + llvm::function_ref<APFloat(const APFloat &, const APFloat &, + llvm::RoundingMode)> + Fn) { + const Pointer &RHS = S.Stk.pop<Pointer>(); + const Pointer &LHS = S.Stk.pop<Pointer>(); + const Pointer &Dst = S.Stk.peek<Pointer>(); + FPOptions FPO = Call->getFPFeaturesInEffect(S.Ctx.getLangOpts()); + llvm::RoundingMode RM = getRoundingMode(FPO); + const auto *VT = Call->getArg(0)->getType()->castAs<VectorType>(); + + unsigned NumElts = VT->getNumElements(); + unsigned EltBits = S.getASTContext().getTypeSize(VT->getElementType()); + unsigned NumLanes = NumElts * EltBits / 128; + unsigned NumElemsPerLane = NumElts / NumLanes; + unsigned HalfElemsPerLane = NumElemsPerLane / 2; + + for (unsigned L = 0; L != NumElts; L += NumElemsPerLane) { + using T = PrimConv<PT_Float>::T; + for (unsigned E = 0; E != HalfElemsPerLane; ++E) { + APFloat Elem1 = LHS.elem<T>(L + (2 * E) + 0).getAPFloat(); + APFloat Elem2 = LHS.elem<T>(L + (2 * E) + 1).getAPFloat(); + Dst.elem<T>(L + E) = static_cast<T>(Fn(Elem1, Elem2, RM)); + } + for (unsigned E = 0; E != HalfElemsPerLane; ++E) { + APFloat Elem1 = RHS.elem<T>(L + (2 * E) + 0).getAPFloat(); + APFloat Elem2 = RHS.elem<T>(L + (2 * E) + 1).getAPFloat(); + Dst.elem<T>(L + E + HalfElemsPerLane) = + static_cast<T>(Fn(Elem1, Elem2, RM)); + } + } + Dst.initializeAllElements(); + return true; +} + static bool interp__builtin_elementwise_triop_fp( InterpState &S, CodePtr OpPC, const CallExpr *Call, llvm::function_ref<APFloat(const APFloat &, const APFloat &, @@ -2596,9 +2672,9 @@ static bool interp__builtin_elementwise_triop_fp( FPOptions FPO = Call->getFPFeaturesInEffect(S.Ctx.getLangOpts()); llvm::RoundingMode RM = getRoundingMode(FPO); - const QualType Arg1Type = Call->getArg(0)->getType(); - const QualType Arg2Type = Call->getArg(1)->getType(); - const QualType Arg3Type = Call->getArg(2)->getType(); + QualType Arg1Type = Call->getArg(0)->getType(); + QualType Arg2Type = Call->getArg(1)->getType(); + QualType Arg3Type = Call->getArg(2)->getType(); // Non-vector floating point types. if (!Arg1Type->isVectorType()) { @@ -2621,16 +2697,16 @@ static bool interp__builtin_elementwise_triop_fp( assert(Arg1Type->isVectorType() && Arg2Type->isVectorType() && Arg3Type->isVectorType()); - const VectorType *VecT = Arg1Type->castAs<VectorType>(); - const QualType ElemT = VecT->getElementType(); - unsigned NumElems = VecT->getNumElements(); + const VectorType *VecTy = Arg1Type->castAs<VectorType>(); + QualType ElemQT = VecTy->getElementType(); + unsigned NumElems = VecTy->getNumElements(); - assert(ElemT == Arg2Type->castAs<VectorType>()->getElementType() && - ElemT == Arg3Type->castAs<VectorType>()->getElementType()); + assert(ElemQT == Arg2Type->castAs<VectorType>()->getElementType() && + ElemQT == Arg3Type->castAs<VectorType>()->getElementType()); assert(NumElems == Arg2Type->castAs<VectorType>()->getNumElements() && NumElems == Arg3Type->castAs<VectorType>()->getNumElements()); - assert(ElemT->isRealFloatingType()); - (void)ElemT; + assert(ElemQT->isRealFloatingType()); + (void)ElemQT; const Pointer &VZ = S.Stk.pop<Pointer>(); const Pointer &VY = S.Stk.pop<Pointer>(); @@ -2714,6 +2790,34 @@ static bool interp__builtin_blend(InterpState &S, CodePtr OpPC, return true; } +static bool interp__builtin_ia32_pshufb(InterpState &S, CodePtr OpPC, + const CallExpr *Call) { + assert(Call->getNumArgs() == 2 && "masked forms handled via select*"); + const Pointer &Control = S.Stk.pop<Pointer>(); + const Pointer &Src = S.Stk.pop<Pointer>(); + const Pointer &Dst = S.Stk.peek<Pointer>(); + + unsigned NumElems = Dst.getNumElems(); + assert(NumElems == Control.getNumElems()); + assert(NumElems == Dst.getNumElems()); + + for (unsigned Idx = 0; Idx != NumElems; ++Idx) { + uint8_t Ctlb = static_cast<uint8_t>(Control.elem<int8_t>(Idx)); + + if (Ctlb & 0x80) { + Dst.elem<int8_t>(Idx) = 0; + } else { + unsigned LaneBase = (Idx / 16) * 16; + unsigned SrcOffset = Ctlb & 0x0F; + unsigned SrcIdx = LaneBase + SrcOffset; + + Dst.elem<int8_t>(Idx) = Src.elem<int8_t>(SrcIdx); + } + } + Dst.initializeAllElements(); + return true; +} + static bool interp__builtin_ia32_pshuf(InterpState &S, CodePtr OpPC, const CallExpr *Call, bool IsShufHW) { assert(Call->getNumArgs() == 2 && "masked forms handled via select*"); @@ -2756,6 +2860,45 @@ static bool interp__builtin_ia32_pshuf(InterpState &S, CodePtr OpPC, return true; } +static bool interp__builtin_ia32_test_op( + InterpState &S, CodePtr OpPC, const CallExpr *Call, + llvm::function_ref<bool(const APInt &A, const APInt &B)> Fn) { + const Pointer &RHS = S.Stk.pop<Pointer>(); + const Pointer &LHS = S.Stk.pop<Pointer>(); + + assert(LHS.getNumElems() == RHS.getNumElems()); + + unsigned SourceLen = LHS.getNumElems(); + QualType ElemQT = getElemType(LHS); + OptPrimType ElemPT = S.getContext().classify(ElemQT); + unsigned LaneWidth = S.getASTContext().getTypeSize(ElemQT); + + APInt AWide(LaneWidth * SourceLen, 0); + APInt BWide(LaneWidth * SourceLen, 0); + + for (unsigned I = 0; I != SourceLen; ++I) { + APInt ALane; + APInt BLane; + + if (ElemQT->isIntegerType()) { // Get value. + INT_TYPE_SWITCH_NO_BOOL(*ElemPT, { + ALane = LHS.elem<T>(I).toAPSInt(); + BLane = RHS.elem<T>(I).toAPSInt(); + }); + } else if (ElemQT->isFloatingType()) { // Get only sign bit. + using T = PrimConv<PT_Float>::T; + ALane = LHS.elem<T>(I).getAPFloat().bitcastToAPInt().isNegative(); + BLane = RHS.elem<T>(I).getAPFloat().bitcastToAPInt().isNegative(); + } else { // Must be integer or floating type. + return false; + } + AWide.insertBits(ALane, I * LaneWidth); + BWide.insertBits(BLane, I * LaneWidth); + } + pushInteger(S, Fn(AWide, BWide), Call->getType()); + return true; +} + static bool interp__builtin_elementwise_triop( InterpState &S, CodePtr OpPC, const CallExpr *Call, llvm::function_ref<APInt(const APSInt &, const APSInt &, const APSInt &)> @@ -2775,7 +2918,7 @@ static bool interp__builtin_elementwise_triop( } const auto *VecT = Arg0Type->castAs<VectorType>(); - const PrimType &ElemT = *S.getContext().classify(VecT->getElementType()); + PrimType ElemT = *S.getContext().classify(VecT->getElementType()); unsigned NumElems = VecT->getNumElements(); bool DestUnsigned = Call->getType()->isUnsignedIntegerOrEnumerationType(); @@ -2847,9 +2990,9 @@ static bool interp__builtin_x86_insert_subvector(InterpState &S, CodePtr OpPC, unsigned Lane = static_cast<unsigned>(Index % NumLanes); unsigned InsertPos = Lane * SubElements; - PrimType ElemPT = BaseVec.getFieldDesc()->getPrimType(); + PrimType ElemT = BaseVec.getFieldDesc()->getPrimType(); - TYPE_SWITCH(ElemPT, { + TYPE_SWITCH(ElemT, { for (unsigned I = 0; I != BaseElements; ++I) Dst.elem<T>(I) = BaseVec.elem<T>(I); for (unsigned I = 0; I != SubElements; ++I) @@ -2872,12 +3015,12 @@ static bool interp__builtin_ia32_pternlog(InterpState &S, CodePtr OpPC, const Pointer &Dst = S.Stk.peek<Pointer>(); unsigned DstLen = A.getNumElems(); - const QualType ElemQT = getElemType(A); - const OptPrimType ElemPT = S.getContext().classify(ElemQT); + QualType ElemQT = getElemType(A); + OptPrimType ElemT = S.getContext().classify(ElemQT); unsigned LaneWidth = S.getASTContext().getTypeSize(ElemQT); bool DstUnsigned = ElemQT->isUnsignedIntegerOrEnumerationType(); - INT_TYPE_SWITCH_NO_BOOL(*ElemPT, { + INT_TYPE_SWITCH_NO_BOOL(*ElemT, { for (unsigned I = 0; I != DstLen; ++I) { APInt ALane = A.elem<T>(I).toAPSInt(); APInt BLane = B.elem<T>(I).toAPSInt(); @@ -2916,13 +3059,13 @@ static bool interp__builtin_vec_ext(InterpState &S, CodePtr OpPC, unsigned Index = static_cast<unsigned>(ImmAPS.getZExtValue() & (NumElems - 1)); - PrimType ElemPT = Vec.getFieldDesc()->getPrimType(); + PrimType ElemT = Vec.getFieldDesc()->getPrimType(); // FIXME(#161685): Replace float+int split with a numeric-only type switch - if (ElemPT == PT_Float) { + if (ElemT == PT_Float) { S.Stk.push<Floating>(Vec.elem<Floating>(Index)); return true; } - INT_TYPE_SWITCH_NO_BOOL(ElemPT, { + INT_TYPE_SWITCH_NO_BOOL(ElemT, { APSInt V = Vec.elem<T>(Index).toAPSInt(); pushInteger(S, V, Call->getType()); }); @@ -2947,8 +3090,8 @@ static bool interp__builtin_vec_set(InterpState &S, CodePtr OpPC, unsigned Index = static_cast<unsigned>(ImmAPS.getZExtValue() & (NumElems - 1)); - PrimType ElemPT = Base.getFieldDesc()->getPrimType(); - INT_TYPE_SWITCH_NO_BOOL(ElemPT, { + PrimType ElemT = Base.getFieldDesc()->getPrimType(); + INT_TYPE_SWITCH_NO_BOOL(ElemT, { for (unsigned I = 0; I != NumElems; ++I) Dst.elem<T>(I) = Base.elem<T>(I); Dst.elem<T>(Index) = static_cast<T>(ValAPS); @@ -2958,6 +3101,33 @@ static bool interp__builtin_vec_set(InterpState &S, CodePtr OpPC, return true; } +static bool interp__builtin_ia32_vpconflict(InterpState &S, CodePtr OpPC, + const CallExpr *Call) { + assert(Call->getNumArgs() == 1); + + QualType Arg0Type = Call->getArg(0)->getType(); + const auto *VecT = Arg0Type->castAs<VectorType>(); + PrimType ElemT = *S.getContext().classify(VecT->getElementType()); + unsigned NumElems = VecT->getNumElements(); + bool DestUnsigned = Call->getType()->isUnsignedIntegerOrEnumerationType(); + const Pointer &Src = S.Stk.pop<Pointer>(); + const Pointer &Dst = S.Stk.peek<Pointer>(); + + for (unsigned I = 0; I != NumElems; ++I) { + INT_TYPE_SWITCH_NO_BOOL(ElemT, { + APSInt ElemI = Src.elem<T>(I).toAPSInt(); + APInt ConflictMask(ElemI.getBitWidth(), 0); + for (unsigned J = 0; J != I; ++J) { + APSInt ElemJ = Src.elem<T>(J).toAPSInt(); + ConflictMask.setBitVal(J, ElemI == ElemJ); + } + Dst.elem<T>(I) = static_cast<T>(APSInt(ConflictMask, DestUnsigned)); + }); + } + Dst.initializeAllElements(); + return true; +} + bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const CallExpr *Call, uint32_t BuiltinID) { if (!S.getASTContext().BuiltinInfo.isConstantEvaluated(BuiltinID)) @@ -3115,14 +3285,14 @@ bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const CallExpr *Call, case Builtin::BI__builtin_parityl: case Builtin::BI__builtin_parityll: return interp__builtin_elementwise_int_unaryop( - S, OpPC, Call, [](const APSInt &Val) -> APInt { + S, OpPC, Call, [](const APSInt &Val) { return APInt(Val.getBitWidth(), Val.popcount() % 2); }); case Builtin::BI__builtin_clrsb: case Builtin::BI__builtin_clrsbl: case Builtin::BI__builtin_clrsbll: return interp__builtin_elementwise_int_unaryop( - S, OpPC, Call, [](const APSInt &Val) -> APInt { + S, OpPC, Call, [](const APSInt &Val) { return APInt(Val.getBitWidth(), Val.getBitWidth() - Val.getSignificantBits()); }); @@ -3131,8 +3301,7 @@ bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const CallExpr *Call, case Builtin::BI__builtin_bitreverse32: case Builtin::BI__builtin_bitreverse64: return interp__builtin_elementwise_int_unaryop( - S, OpPC, Call, - [](const APSInt &Val) -> APInt { return Val.reverseBits(); }); + S, OpPC, Call, [](const APSInt &Val) { return Val.reverseBits(); }); case Builtin::BI__builtin_classify_type: return interp__builtin_classify_type(S, OpPC, Frame, Call); @@ -3626,6 +3795,53 @@ bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const CallExpr *Call, case Builtin::BI__builtin_elementwise_min: return interp__builtin_elementwise_maxmin(S, OpPC, Call, BuiltinID); + case clang::X86::BI__builtin_ia32_phaddw128: + case clang::X86::BI__builtin_ia32_phaddw256: + case clang::X86::BI__builtin_ia32_phaddd128: + case clang::X86::BI__builtin_ia32_phaddd256: + return interp_builtin_horizontal_int_binop( + S, OpPC, Call, + [](const APSInt &LHS, const APSInt &RHS) { return LHS + RHS; }); + case clang::X86::BI__builtin_ia32_phaddsw128: + case clang::X86::BI__builtin_ia32_phaddsw256: + return interp_builtin_horizontal_int_binop( + S, OpPC, Call, + [](const APSInt &LHS, const APSInt &RHS) { return LHS.sadd_sat(RHS); }); + case clang::X86::BI__builtin_ia32_phsubw128: + case clang::X86::BI__builtin_ia32_phsubw256: + case clang::X86::BI__builtin_ia32_phsubd128: + case clang::X86::BI__builtin_ia32_phsubd256: + return interp_builtin_horizontal_int_binop( + S, OpPC, Call, + [](const APSInt &LHS, const APSInt &RHS) { return LHS - RHS; }); + case clang::X86::BI__builtin_ia32_phsubsw128: + case clang::X86::BI__builtin_ia32_phsubsw256: + return interp_builtin_horizontal_int_binop( + S, OpPC, Call, + [](const APSInt &LHS, const APSInt &RHS) { return LHS.ssub_sat(RHS); }); + case clang::X86::BI__builtin_ia32_haddpd: + case clang::X86::BI__builtin_ia32_haddps: + case clang::X86::BI__builtin_ia32_haddpd256: + case clang::X86::BI__builtin_ia32_haddps256: + return interp_builtin_horizontal_fp_binop( + S, OpPC, Call, + [](const APFloat &LHS, const APFloat &RHS, llvm::RoundingMode RM) { + APFloat F = LHS; + F.add(RHS, RM); + return F; + }); + case clang::X86::BI__builtin_ia32_hsubpd: + case clang::X86::BI__builtin_ia32_hsubps: + case clang::X86::BI__builtin_ia32_hsubpd256: + case clang::X86::BI__builtin_ia32_hsubps256: + return interp_builtin_horizontal_fp_binop( + S, OpPC, Call, + [](const APFloat &LHS, const APFloat &RHS, llvm::RoundingMode RM) { + APFloat F = LHS; + F.subtract(RHS, RM); + return F; + }); + case clang::X86::BI__builtin_ia32_pmuldq128: case clang::X86::BI__builtin_ia32_pmuldq256: case clang::X86::BI__builtin_ia32_pmuldq512: @@ -3656,6 +3872,21 @@ bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const CallExpr *Call, return F; }); + case X86::BI__builtin_ia32_vpmadd52luq128: + case X86::BI__builtin_ia32_vpmadd52luq256: + case X86::BI__builtin_ia32_vpmadd52luq512: + return interp__builtin_elementwise_triop( + S, OpPC, Call, [](const APSInt &A, const APSInt &B, const APSInt &C) { + return A + (B.trunc(52) * C.trunc(52)).zext(64); + }); + case X86::BI__builtin_ia32_vpmadd52huq128: + case X86::BI__builtin_ia32_vpmadd52huq256: + case X86::BI__builtin_ia32_vpmadd52huq512: + return interp__builtin_elementwise_triop( + S, OpPC, Call, [](const APSInt &A, const APSInt &B, const APSInt &C) { + return A + llvm::APIntOps::mulhu(B.trunc(52), C.trunc(52)).zext(64); + }); + case X86::BI__builtin_ia32_vpshldd128: case X86::BI__builtin_ia32_vpshldd256: case X86::BI__builtin_ia32_vpshldd512: @@ -3686,7 +3917,13 @@ bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const CallExpr *Call, [](const APSInt &Lo, const APSInt &Hi, const APSInt &Amt) { return llvm::APIntOps::fshr(Hi, Lo, Amt); }); - + case X86::BI__builtin_ia32_vpconflictsi_128: + case X86::BI__builtin_ia32_vpconflictsi_256: + case X86::BI__builtin_ia32_vpconflictsi_512: + case X86::BI__builtin_ia32_vpconflictdi_128: + case X86::BI__builtin_ia32_vpconflictdi_256: + case X86::BI__builtin_ia32_vpconflictdi_512: + return interp__builtin_ia32_vpconflict(S, OpPC, Call); case clang::X86::BI__builtin_ia32_blendpd: case clang::X86::BI__builtin_ia32_blendpd256: case clang::X86::BI__builtin_ia32_blendps: @@ -3712,7 +3949,34 @@ bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const CallExpr *Call, S, OpPC, Call, [](const APSInt &F, const APSInt &T, const APSInt &C) { return ((APInt)C).isNegative() ? T : F; }); - + case X86::BI__builtin_ia32_ptestz128: + case X86::BI__builtin_ia32_ptestz256: + case X86::BI__builtin_ia32_vtestzps: + case X86::BI__builtin_ia32_vtestzps256: + case X86::BI__builtin_ia32_vtestzpd: + case X86::BI__builtin_ia32_vtestzpd256: + return interp__builtin_ia32_test_op( + S, OpPC, Call, + [](const APInt &A, const APInt &B) { return (A & B) == 0; }); + case X86::BI__builtin_ia32_ptestc128: + case X86::BI__builtin_ia32_ptestc256: + case X86::BI__builtin_ia32_vtestcps: + case X86::BI__builtin_ia32_vtestcps256: + case X86::BI__builtin_ia32_vtestcpd: + case X86::BI__builtin_ia32_vtestcpd256: + return interp__builtin_ia32_test_op( + S, OpPC, Call, + [](const APInt &A, const APInt &B) { return (~A & B) == 0; }); + case X86::BI__builtin_ia32_ptestnzc128: + case X86::BI__builtin_ia32_ptestnzc256: + case X86::BI__builtin_ia32_vtestnzcps: + case X86::BI__builtin_ia32_vtestnzcps256: + case X86::BI__builtin_ia32_vtestnzcpd: + case X86::BI__builtin_ia32_vtestnzcpd256: + return interp__builtin_ia32_test_op( + S, OpPC, Call, [](const APInt &A, const APInt &B) { + return ((A & B) != 0) && ((~A & B) != 0); + }); case X86::BI__builtin_ia32_selectb_128: case X86::BI__builtin_ia32_selectb_256: case X86::BI__builtin_ia32_selectb_512: @@ -3739,6 +4003,11 @@ bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const CallExpr *Call, case X86::BI__builtin_ia32_selectpd_512: return interp__builtin_select(S, OpPC, Call); + case X86::BI__builtin_ia32_pshufb128: + case X86::BI__builtin_ia32_pshufb256: + case X86::BI__builtin_ia32_pshufb512: + return interp__builtin_ia32_pshufb(S, OpPC, Call); + case X86::BI__builtin_ia32_pshuflw: case X86::BI__builtin_ia32_pshuflw256: case X86::BI__builtin_ia32_pshuflw512: diff --git a/clang/lib/AST/ByteCode/InterpState.h b/clang/lib/AST/ByteCode/InterpState.h index a13244b..e2e4d5c 100644 --- a/clang/lib/AST/ByteCode/InterpState.h +++ b/clang/lib/AST/ByteCode/InterpState.h @@ -114,7 +114,7 @@ public: Alloc = std::make_unique<DynamicAllocator>(); } - return *Alloc.get(); + return *Alloc; } /// Diagnose any dynamic allocations that haven't been freed yet. diff --git a/clang/lib/AST/ByteCode/Pointer.cpp b/clang/lib/AST/ByteCode/Pointer.cpp index 663134c..e417bdf 100644 --- a/clang/lib/AST/ByteCode/Pointer.cpp +++ b/clang/lib/AST/ByteCode/Pointer.cpp @@ -751,7 +751,7 @@ std::optional<APValue> Pointer::toRValue(const Context &Ctx, assert(Record && "Missing record descriptor"); bool Ok = true; - if (RT->getOriginalDecl()->isUnion()) { + if (RT->getDecl()->isUnion()) { const FieldDecl *ActiveField = nullptr; APValue Value; for (const auto &F : Record->fields()) { diff --git a/clang/lib/AST/ByteCode/Program.cpp b/clang/lib/AST/ByteCode/Program.cpp index 75bfd9f..e653782 100644 --- a/clang/lib/AST/ByteCode/Program.cpp +++ b/clang/lib/AST/ByteCode/Program.cpp @@ -226,11 +226,7 @@ UnsignedOrNone Program::createGlobal(const ValueDecl *VD, const Expr *Init) { Globals[PIdx] = NewGlobal; // All pointers pointing to the previous extern decl now point to the // new decl. - for (Pointer *Ptr = RedeclBlock->Pointers; Ptr; Ptr = Ptr->BS.Next) { - RedeclBlock->removePointer(Ptr); - Ptr->BS.Pointee = NewGlobal->block(); - NewGlobal->block()->addPointer(Ptr); - } + RedeclBlock->movePointersTo(NewGlobal->block()); } } PIdx = *Idx; diff --git a/clang/lib/AST/Comment.cpp b/clang/lib/AST/Comment.cpp index 37e21c3..361a8a7e 100644 --- a/clang/lib/AST/Comment.cpp +++ b/clang/lib/AST/Comment.cpp @@ -56,16 +56,16 @@ good implements_child_begin_end(Comment::child_iterator (T::*)() const) { return good(); } -LLVM_ATTRIBUTE_UNUSED -static inline bad implements_child_begin_end( - Comment::child_iterator (Comment::*)() const) { +[[maybe_unused]] +static inline bad +implements_child_begin_end(Comment::child_iterator (Comment::*)() const) { return bad(); } #define ASSERT_IMPLEMENTS_child_begin(function) \ (void) good(implements_child_begin_end(function)) -LLVM_ATTRIBUTE_UNUSED +[[maybe_unused]] static inline void CheckCommentASTNodes() { #define ABSTRACT_COMMENT(COMMENT) #define COMMENT(CLASS, PARENT) \ diff --git a/clang/lib/AST/Decl.cpp b/clang/lib/AST/Decl.cpp index c734155..f048076 100644 --- a/clang/lib/AST/Decl.cpp +++ b/clang/lib/AST/Decl.cpp @@ -2989,10 +2989,7 @@ bool ParmVarDecl::isDestroyedInCallee() const { // FIXME: isParamDestroyedInCallee() should probably imply // isDestructedType() const auto *RT = getType()->getAsCanonical<RecordType>(); - if (RT && - RT->getOriginalDecl() - ->getDefinitionOrSelf() - ->isParamDestroyedInCallee() && + if (RT && RT->getDecl()->getDefinitionOrSelf()->isParamDestroyedInCallee() && getType().isDestructedType()) return true; @@ -3316,6 +3313,10 @@ bool FunctionDecl::isImmediateEscalating() const { CD && CD->isInheritingConstructor()) return CD->getInheritedConstructor().getConstructor(); + // Destructors are not immediate escalating. + if (isa<CXXDestructorDecl>(this)) + return false; + // - a function that results from the instantiation of a templated entity // defined with the constexpr specifier. TemplatedKind TK = getTemplatedKind(); @@ -3503,7 +3504,7 @@ bool FunctionDecl::isUsableAsGlobalAllocationFunctionInConstantEvaluation( while (const auto *TD = T->getAs<TypedefType>()) T = TD->getDecl()->getUnderlyingType(); const IdentifierInfo *II = - T->castAsCanonical<EnumType>()->getOriginalDecl()->getIdentifier(); + T->castAsCanonical<EnumType>()->getDecl()->getIdentifier(); if (II && II->isStr("__hot_cold_t")) Consume(); } @@ -4705,7 +4706,7 @@ bool FieldDecl::isAnonymousStructOrUnion() const { return false; if (const auto *Record = getType()->getAsCanonical<RecordType>()) - return Record->getOriginalDecl()->isAnonymousStructOrUnion(); + return Record->getDecl()->isAnonymousStructOrUnion(); return false; } @@ -4765,7 +4766,7 @@ bool FieldDecl::isZeroSize(const ASTContext &Ctx) const { const auto *RT = getType()->getAsCanonical<RecordType>(); if (!RT) return false; - const RecordDecl *RD = RT->getOriginalDecl()->getDefinition(); + const RecordDecl *RD = RT->getDecl()->getDefinition(); if (!RD) { assert(isInvalidDecl() && "valid field has incomplete type"); return false; @@ -5190,7 +5191,7 @@ bool RecordDecl::isOrContainsUnion() const { if (const RecordDecl *Def = getDefinition()) { for (const FieldDecl *FD : Def->fields()) { const RecordType *RT = FD->getType()->getAsCanonical<RecordType>(); - if (RT && RT->getOriginalDecl()->isOrContainsUnion()) + if (RT && RT->getDecl()->isOrContainsUnion()) return true; } } @@ -5688,14 +5689,14 @@ void TypedefNameDecl::anchor() {} TagDecl *TypedefNameDecl::getAnonDeclWithTypedefName(bool AnyRedecl) const { if (auto *TT = getTypeSourceInfo()->getType()->getAs<TagType>()) { - auto *OwningTypedef = TT->getOriginalDecl()->getTypedefNameForAnonDecl(); + auto *OwningTypedef = TT->getDecl()->getTypedefNameForAnonDecl(); auto *ThisTypedef = this; if (AnyRedecl && OwningTypedef) { OwningTypedef = OwningTypedef->getCanonicalDecl(); ThisTypedef = ThisTypedef->getCanonicalDecl(); } if (OwningTypedef == ThisTypedef) - return TT->getOriginalDecl()->getDefinitionOrSelf(); + return TT->getDecl()->getDefinitionOrSelf(); } return nullptr; @@ -5704,7 +5705,7 @@ TagDecl *TypedefNameDecl::getAnonDeclWithTypedefName(bool AnyRedecl) const { bool TypedefNameDecl::isTransparentTagSlow() const { auto determineIsTransparent = [&]() { if (auto *TT = getUnderlyingType()->getAs<TagType>()) { - if (auto *TD = TT->getOriginalDecl()) { + if (auto *TD = TT->getDecl()) { if (TD->getName() != getName()) return false; SourceLocation TTLoc = getLocation(); diff --git a/clang/lib/AST/DeclBase.cpp b/clang/lib/AST/DeclBase.cpp index b244f0a..30c6d3e 100644 --- a/clang/lib/AST/DeclBase.cpp +++ b/clang/lib/AST/DeclBase.cpp @@ -77,8 +77,11 @@ void *Decl::operator new(std::size_t Size, const ASTContext &Context, *PrefixPtr = ID.getRawValue(); // We leave the upper 16 bits to store the module IDs. 48 bits should be - // sufficient to store a declaration ID. - assert(*PrefixPtr < llvm::maskTrailingOnes<uint64_t>(48)); + // sufficient to store a declaration ID. See the comments in setOwningModuleID + // for details. + assert((*PrefixPtr < llvm::maskTrailingOnes<uint64_t>(48)) && + "Current Implementation limits the number of module files to not " + "exceed 2^16. Contact Clang Developers to remove the limitation."); return Result; } @@ -122,6 +125,25 @@ unsigned Decl::getOwningModuleID() const { void Decl::setOwningModuleID(unsigned ID) { assert(isFromASTFile() && "Only works on a deserialized declaration"); + // Currently, we use 64 bits to store the GlobalDeclID and the module ID + // to save the space. See `Decl::operator new` for details. To make it, + // we split the higher 32 bits to 2 16bits for the module file index of + // GlobalDeclID and the module ID. This introduces a limitation that the + // number of modules can't exceed 2^16. (The number of module files should be + // less than the number of modules). + // + // It is counter-intuitive to store both the module file index and the + // module ID as it seems redundant. However, this is not true. + // The module ID may be different from the module file where it is serialized + // from for implicit template instantiations. See + // https://github.com/llvm/llvm-project/issues/101939 + // + // If we reach the limitation, we have to remove the limitation by asking + // every deserialized declaration to pay for yet another 32 bits, or we have + // to review the above issue to decide what we should do for it. + assert((ID < llvm::maskTrailingOnes<unsigned>(16)) && + "Current Implementation limits the number of modules to not exceed " + "2^16. Contact Clang Developers to remove the limitation."); uint64_t *IDAddress = (uint64_t *)this - 1; *IDAddress &= llvm::maskTrailingOnes<uint64_t>(48); *IDAddress |= (uint64_t)ID << 48; diff --git a/clang/lib/AST/DeclCXX.cpp b/clang/lib/AST/DeclCXX.cpp index 43264f8..24e4f18 100644 --- a/clang/lib/AST/DeclCXX.cpp +++ b/clang/lib/AST/DeclCXX.cpp @@ -2314,7 +2314,7 @@ bool CXXRecordDecl::mayBeAbstract() const { for (const auto &B : bases()) { const auto *BaseDecl = cast<CXXRecordDecl>( - B.getType()->castAsCanonical<RecordType>()->getOriginalDecl()); + B.getType()->castAsCanonical<RecordType>()->getDecl()); if (BaseDecl->isAbstract()) return true; } diff --git a/clang/lib/AST/DeclPrinter.cpp b/clang/lib/AST/DeclPrinter.cpp index 7f3dcca..47ae613 100644 --- a/clang/lib/AST/DeclPrinter.cpp +++ b/clang/lib/AST/DeclPrinter.cpp @@ -485,7 +485,7 @@ void DeclPrinter::VisitDeclContext(DeclContext *DC, bool Indent) { QualType BaseType = GetBaseType(CurDeclType); if (const auto *TT = dyn_cast_or_null<TagType>(BaseType); TT && TT->isTagOwned()) { - if (TT->getOriginalDecl() == Decls[0]) { + if (TT->getDecl() == Decls[0]) { Decls.push_back(*D); continue; } diff --git a/clang/lib/AST/DeclarationName.cpp b/clang/lib/AST/DeclarationName.cpp index 55f5a99..9a89a66 100644 --- a/clang/lib/AST/DeclarationName.cpp +++ b/clang/lib/AST/DeclarationName.cpp @@ -116,12 +116,12 @@ static void printCXXConstructorDestructorName(QualType ClassType, Policy.SuppressScope = true; if (const RecordType *ClassRec = ClassType->getAsCanonical<RecordType>()) { - ClassRec->getOriginalDecl()->printName(OS, Policy); + ClassRec->getDecl()->printName(OS, Policy); return; } if (Policy.SuppressTemplateArgsInCXXConstructors) { if (auto *InjTy = ClassType->getAsCanonical<InjectedClassNameType>()) { - InjTy->getOriginalDecl()->printName(OS, Policy); + InjTy->getDecl()->printName(OS, Policy); return; } } @@ -185,7 +185,7 @@ void DeclarationName::print(raw_ostream &OS, OS << "operator "; QualType Type = getCXXNameType(); if (const RecordType *Rec = Type->getAs<RecordType>()) { - OS << *Rec->getOriginalDecl(); + OS << *Rec->getDecl(); return; } // We know we're printing C++ here, ensure we print 'bool' properly. diff --git a/clang/lib/AST/Expr.cpp b/clang/lib/AST/Expr.cpp index 597cbd8..340bb4b 100644 --- a/clang/lib/AST/Expr.cpp +++ b/clang/lib/AST/Expr.cpp @@ -4126,9 +4126,7 @@ Expr::isNullPointerConstant(ASTContext &Ctx, if (const RecordType *UT = getType()->getAsUnionType()) if (!Ctx.getLangOpts().CPlusPlus11 && UT && - UT->getOriginalDecl() - ->getMostRecentDecl() - ->hasAttr<TransparentUnionAttr>()) + UT->getDecl()->getMostRecentDecl()->hasAttr<TransparentUnionAttr>()) if (const CompoundLiteralExpr *CLE = dyn_cast<CompoundLiteralExpr>(this)){ const Expr *InitExpr = CLE->getInitializer(); if (const InitListExpr *ILE = dyn_cast<InitListExpr>(InitExpr)) diff --git a/clang/lib/AST/ExprConstant.cpp b/clang/lib/AST/ExprConstant.cpp index 35a866e..16141b2 100644 --- a/clang/lib/AST/ExprConstant.cpp +++ b/clang/lib/AST/ExprConstant.cpp @@ -990,7 +990,7 @@ namespace { // of arrays to avoid exhausting the system resources, as initialization // of each element is likely to take some number of steps anyway. uint64_t Limit = Ctx.getLangOpts().ConstexprStepLimit; - if (ElemCount > Limit) { + if (Limit != 0 && ElemCount > Limit) { if (Diag) FFDiag(Loc, diag::note_constexpr_new_exceeds_limits) << ElemCount << Limit; @@ -1016,6 +1016,9 @@ namespace { } bool nextStep(const Stmt *S) { + if (Ctx.getLangOpts().ConstexprStepLimit == 0) + return true; + if (!StepsLeft) { FFDiag(S->getBeginLoc(), diag::note_constexpr_step_limit_exceeded); return false; @@ -1186,7 +1189,8 @@ namespace { /// Should we continue evaluation as much as possible after encountering a /// construct which can't be reduced to a value? bool keepEvaluatingAfterFailure() const override { - if (!StepsLeft) + uint64_t Limit = Ctx.getLangOpts().ConstexprStepLimit; + if (Limit != 0 && !StepsLeft) return false; switch (EvalMode) { @@ -4070,8 +4074,7 @@ findSubobject(EvalInfo &Info, const Expr *E, const CompleteObject &Obj, } // Next subobject is a class, struct or union field. - RecordDecl *RD = - ObjType->castAsCanonical<RecordType>()->getOriginalDecl(); + RecordDecl *RD = ObjType->castAsCanonical<RecordType>()->getDecl(); if (RD->isUnion()) { const FieldDecl *UnionField = O->getUnionField(); if (!UnionField || @@ -7806,7 +7809,7 @@ class BufferToAPValueConverter { std::optional<APValue> visit(const EnumType *Ty, CharUnits Offset) { QualType RepresentationType = - Ty->getOriginalDecl()->getDefinitionOrSelf()->getIntegerType(); + Ty->getDecl()->getDefinitionOrSelf()->getIntegerType(); assert(!RepresentationType.isNull() && "enum forward decl should be caught by Sema"); const auto *AsBuiltin = @@ -8603,7 +8606,7 @@ public: const FieldDecl *FD = dyn_cast<FieldDecl>(E->getMemberDecl()); if (!FD) return Error(E); assert(!FD->getType()->isReferenceType() && "prvalue reference?"); - assert(BaseTy->castAsCanonical<RecordType>()->getOriginalDecl() == + assert(BaseTy->castAsCanonical<RecordType>()->getDecl() == FD->getParent()->getCanonicalDecl() && "record / field mismatch"); @@ -8832,7 +8835,7 @@ public: const ValueDecl *MD = E->getMemberDecl(); if (const FieldDecl *FD = dyn_cast<FieldDecl>(E->getMemberDecl())) { - assert(BaseTy->castAsCanonical<RecordType>()->getOriginalDecl() == + assert(BaseTy->castAsCanonical<RecordType>()->getDecl() == FD->getParent()->getCanonicalDecl() && "record / field mismatch"); (void)BaseTy; @@ -11615,6 +11618,44 @@ static bool evalPackBuiltin(const CallExpr *E, EvalInfo &Info, APValue &Result, return true; } +static bool evalPshufbBuiltin(EvalInfo &Info, const CallExpr *Call, + APValue &Out) { + APValue SrcVec, ControlVec; + if (!EvaluateAsRValue(Info, Call->getArg(0), SrcVec)) + return false; + if (!EvaluateAsRValue(Info, Call->getArg(1), ControlVec)) + return false; + + const auto *VT = Call->getType()->getAs<VectorType>(); + if (!VT) + return false; + + QualType ElemT = VT->getElementType(); + unsigned NumElts = VT->getNumElements(); + + SmallVector<APValue, 64> ResultElements; + ResultElements.reserve(NumElts); + + for (unsigned Idx = 0; Idx != NumElts; ++Idx) { + APValue CtlVal = ControlVec.getVectorElt(Idx); + APSInt CtlByte = CtlVal.getInt(); + uint8_t Ctl = static_cast<uint8_t>(CtlByte.getZExtValue()); + + if (Ctl & 0x80) { + APValue Zero(Info.Ctx.MakeIntValue(0, ElemT)); + ResultElements.push_back(Zero); + } else { + unsigned LaneBase = (Idx / 16) * 16; + unsigned SrcOffset = Ctl & 0x0F; + unsigned SrcIdx = LaneBase + SrcOffset; + + ResultElements.push_back(SrcVec.getVectorElt(SrcIdx)); + } + } + Out = APValue(ResultElements.data(), ResultElements.size()); + return true; +} + static bool evalPshufBuiltin(EvalInfo &Info, const CallExpr *Call, bool IsShufHW, APValue &Out) { APValue Vec; @@ -11970,6 +12011,54 @@ bool VectorExprEvaluator::VisitCallExpr(const CallExpr *E) { return Success(APValue(ResultElements.data(), ResultElements.size()), E); } + + case X86::BI__builtin_ia32_vpmadd52luq128: + case X86::BI__builtin_ia32_vpmadd52luq256: + case X86::BI__builtin_ia32_vpmadd52luq512: { + APValue A, B, C; + if (!EvaluateAsRValue(Info, E->getArg(0), A) || + !EvaluateAsRValue(Info, E->getArg(1), B) || + !EvaluateAsRValue(Info, E->getArg(2), C)) + return false; + + unsigned ALen = A.getVectorLength(); + SmallVector<APValue, 4> ResultElements; + ResultElements.reserve(ALen); + + for (unsigned EltNum = 0; EltNum < ALen; EltNum += 1) { + APInt AElt = A.getVectorElt(EltNum).getInt(); + APInt BElt = B.getVectorElt(EltNum).getInt().trunc(52); + APInt CElt = C.getVectorElt(EltNum).getInt().trunc(52); + APSInt ResElt(AElt + (BElt * CElt).zext(64), false); + ResultElements.push_back(APValue(ResElt)); + } + + return Success(APValue(ResultElements.data(), ResultElements.size()), E); + } + case X86::BI__builtin_ia32_vpmadd52huq128: + case X86::BI__builtin_ia32_vpmadd52huq256: + case X86::BI__builtin_ia32_vpmadd52huq512: { + APValue A, B, C; + if (!EvaluateAsRValue(Info, E->getArg(0), A) || + !EvaluateAsRValue(Info, E->getArg(1), B) || + !EvaluateAsRValue(Info, E->getArg(2), C)) + return false; + + unsigned ALen = A.getVectorLength(); + SmallVector<APValue, 4> ResultElements; + ResultElements.reserve(ALen); + + for (unsigned EltNum = 0; EltNum < ALen; EltNum += 1) { + APInt AElt = A.getVectorElt(EltNum).getInt(); + APInt BElt = B.getVectorElt(EltNum).getInt().trunc(52); + APInt CElt = C.getVectorElt(EltNum).getInt().trunc(52); + APSInt ResElt(AElt + llvm::APIntOps::mulhu(BElt, CElt).zext(64), false); + ResultElements.push_back(APValue(ResElt)); + } + + return Success(APValue(ResultElements.data(), ResultElements.size()), E); + } + case clang::X86::BI__builtin_ia32_vprotbi: case clang::X86::BI__builtin_ia32_vprotdi: case clang::X86::BI__builtin_ia32_vprotqi: @@ -12090,6 +12179,37 @@ bool VectorExprEvaluator::VisitCallExpr(const CallExpr *E) { return Success(APValue(ResultElements.data(), ResultElements.size()), E); } + case X86::BI__builtin_ia32_vpconflictsi_128: + case X86::BI__builtin_ia32_vpconflictsi_256: + case X86::BI__builtin_ia32_vpconflictsi_512: + case X86::BI__builtin_ia32_vpconflictdi_128: + case X86::BI__builtin_ia32_vpconflictdi_256: + case X86::BI__builtin_ia32_vpconflictdi_512: { + APValue Source; + + if (!EvaluateAsRValue(Info, E->getArg(0), Source)) + return false; + + unsigned SourceLen = Source.getVectorLength(); + SmallVector<APValue, 32> ResultElements; + ResultElements.reserve(SourceLen); + + const auto *VecT = E->getType()->castAs<VectorType>(); + bool DestUnsigned = + VecT->getElementType()->isUnsignedIntegerOrEnumerationType(); + + for (unsigned I = 0; I != SourceLen; ++I) { + const APValue &EltI = Source.getVectorElt(I); + + APInt ConflictMask(EltI.getInt().getBitWidth(), 0); + for (unsigned J = 0; J != I; ++J) { + const APValue &EltJ = Source.getVectorElt(J); + ConflictMask.setBitVal(J, EltI.getInt() == EltJ.getInt()); + } + ResultElements.push_back(APValue(APSInt(ConflictMask, DestUnsigned))); + } + return Success(APValue(ResultElements.data(), ResultElements.size()), E); + } case X86::BI__builtin_ia32_blendpd: case X86::BI__builtin_ia32_blendpd256: case X86::BI__builtin_ia32_blendps: @@ -12189,6 +12309,15 @@ bool VectorExprEvaluator::VisitCallExpr(const CallExpr *E) { return Success(APValue(ResultElements.data(), ResultElements.size()), E); } + case X86::BI__builtin_ia32_pshufb128: + case X86::BI__builtin_ia32_pshufb256: + case X86::BI__builtin_ia32_pshufb512: { + APValue R; + if (!evalPshufbBuiltin(Info, E, R)) + return false; + return Success(R, E); + } + case X86::BI__builtin_ia32_pshuflw: case X86::BI__builtin_ia32_pshuflw256: case X86::BI__builtin_ia32_pshuflw512: { @@ -12377,6 +12506,169 @@ bool VectorExprEvaluator::VisitCallExpr(const CallExpr *E) { return Success(APValue(ResultElements.data(), ResultElements.size()), E); } + case clang::X86::BI__builtin_ia32_phaddw128: + case clang::X86::BI__builtin_ia32_phaddw256: + case clang::X86::BI__builtin_ia32_phaddd128: + case clang::X86::BI__builtin_ia32_phaddd256: + case clang::X86::BI__builtin_ia32_phaddsw128: + case clang::X86::BI__builtin_ia32_phaddsw256: + + case clang::X86::BI__builtin_ia32_phsubw128: + case clang::X86::BI__builtin_ia32_phsubw256: + case clang::X86::BI__builtin_ia32_phsubd128: + case clang::X86::BI__builtin_ia32_phsubd256: + case clang::X86::BI__builtin_ia32_phsubsw128: + case clang::X86::BI__builtin_ia32_phsubsw256: { + APValue SourceLHS, SourceRHS; + if (!EvaluateAsRValue(Info, E->getArg(0), SourceLHS) || + !EvaluateAsRValue(Info, E->getArg(1), SourceRHS)) + return false; + QualType DestEltTy = E->getType()->castAs<VectorType>()->getElementType(); + bool DestUnsigned = DestEltTy->isUnsignedIntegerOrEnumerationType(); + + unsigned NumElts = SourceLHS.getVectorLength(); + unsigned EltBits = Info.Ctx.getIntWidth(DestEltTy); + unsigned EltsPerLane = 128 / EltBits; + SmallVector<APValue, 4> ResultElements; + ResultElements.reserve(NumElts); + + for (unsigned LaneStart = 0; LaneStart != NumElts; + LaneStart += EltsPerLane) { + for (unsigned I = 0; I != EltsPerLane; I += 2) { + APSInt LHSA = SourceLHS.getVectorElt(LaneStart + I).getInt(); + APSInt LHSB = SourceLHS.getVectorElt(LaneStart + I + 1).getInt(); + switch (E->getBuiltinCallee()) { + case clang::X86::BI__builtin_ia32_phaddw128: + case clang::X86::BI__builtin_ia32_phaddw256: + case clang::X86::BI__builtin_ia32_phaddd128: + case clang::X86::BI__builtin_ia32_phaddd256: { + APSInt Res(LHSA + LHSB, DestUnsigned); + ResultElements.push_back(APValue(Res)); + break; + } + case clang::X86::BI__builtin_ia32_phaddsw128: + case clang::X86::BI__builtin_ia32_phaddsw256: { + APSInt Res(LHSA.sadd_sat(LHSB)); + ResultElements.push_back(APValue(Res)); + break; + } + case clang::X86::BI__builtin_ia32_phsubw128: + case clang::X86::BI__builtin_ia32_phsubw256: + case clang::X86::BI__builtin_ia32_phsubd128: + case clang::X86::BI__builtin_ia32_phsubd256: { + APSInt Res(LHSA - LHSB, DestUnsigned); + ResultElements.push_back(APValue(Res)); + break; + } + case clang::X86::BI__builtin_ia32_phsubsw128: + case clang::X86::BI__builtin_ia32_phsubsw256: { + APSInt Res(LHSA.ssub_sat(LHSB)); + ResultElements.push_back(APValue(Res)); + break; + } + } + } + for (unsigned I = 0; I != EltsPerLane; I += 2) { + APSInt RHSA = SourceRHS.getVectorElt(LaneStart + I).getInt(); + APSInt RHSB = SourceRHS.getVectorElt(LaneStart + I + 1).getInt(); + switch (E->getBuiltinCallee()) { + case clang::X86::BI__builtin_ia32_phaddw128: + case clang::X86::BI__builtin_ia32_phaddw256: + case clang::X86::BI__builtin_ia32_phaddd128: + case clang::X86::BI__builtin_ia32_phaddd256: { + APSInt Res(RHSA + RHSB, DestUnsigned); + ResultElements.push_back(APValue(Res)); + break; + } + case clang::X86::BI__builtin_ia32_phaddsw128: + case clang::X86::BI__builtin_ia32_phaddsw256: { + APSInt Res(RHSA.sadd_sat(RHSB)); + ResultElements.push_back(APValue(Res)); + break; + } + case clang::X86::BI__builtin_ia32_phsubw128: + case clang::X86::BI__builtin_ia32_phsubw256: + case clang::X86::BI__builtin_ia32_phsubd128: + case clang::X86::BI__builtin_ia32_phsubd256: { + APSInt Res(RHSA - RHSB, DestUnsigned); + ResultElements.push_back(APValue(Res)); + break; + } + case clang::X86::BI__builtin_ia32_phsubsw128: + case clang::X86::BI__builtin_ia32_phsubsw256: { + APSInt Res(RHSA.ssub_sat(RHSB)); + ResultElements.push_back(APValue(Res)); + break; + } + } + } + } + return Success(APValue(ResultElements.data(), ResultElements.size()), E); + } + case clang::X86::BI__builtin_ia32_haddpd: + case clang::X86::BI__builtin_ia32_haddps: + case clang::X86::BI__builtin_ia32_haddps256: + case clang::X86::BI__builtin_ia32_haddpd256: + case clang::X86::BI__builtin_ia32_hsubpd: + case clang::X86::BI__builtin_ia32_hsubps: + case clang::X86::BI__builtin_ia32_hsubps256: + case clang::X86::BI__builtin_ia32_hsubpd256: { + APValue SourceLHS, SourceRHS; + if (!EvaluateAsRValue(Info, E->getArg(0), SourceLHS) || + !EvaluateAsRValue(Info, E->getArg(1), SourceRHS)) + return false; + unsigned NumElts = SourceLHS.getVectorLength(); + SmallVector<APValue, 4> ResultElements; + ResultElements.reserve(NumElts); + llvm::RoundingMode RM = getActiveRoundingMode(getEvalInfo(), E); + QualType DestEltTy = E->getType()->castAs<VectorType>()->getElementType(); + unsigned EltBits = Info.Ctx.getTypeSize(DestEltTy); + unsigned NumLanes = NumElts * EltBits / 128; + unsigned NumElemsPerLane = NumElts / NumLanes; + unsigned HalfElemsPerLane = NumElemsPerLane / 2; + + for (unsigned L = 0; L != NumElts; L += NumElemsPerLane) { + for (unsigned I = 0; I != HalfElemsPerLane; ++I) { + APFloat LHSA = SourceLHS.getVectorElt(L + (2 * I) + 0).getFloat(); + APFloat LHSB = SourceLHS.getVectorElt(L + (2 * I) + 1).getFloat(); + switch (E->getBuiltinCallee()) { + case clang::X86::BI__builtin_ia32_haddpd: + case clang::X86::BI__builtin_ia32_haddps: + case clang::X86::BI__builtin_ia32_haddps256: + case clang::X86::BI__builtin_ia32_haddpd256: + LHSA.add(LHSB, RM); + break; + case clang::X86::BI__builtin_ia32_hsubpd: + case clang::X86::BI__builtin_ia32_hsubps: + case clang::X86::BI__builtin_ia32_hsubps256: + case clang::X86::BI__builtin_ia32_hsubpd256: + LHSA.subtract(LHSB, RM); + break; + } + ResultElements.push_back(APValue(LHSA)); + } + for (unsigned I = 0; I != HalfElemsPerLane; ++I) { + APFloat RHSA = SourceRHS.getVectorElt(L + (2 * I) + 0).getFloat(); + APFloat RHSB = SourceRHS.getVectorElt(L + (2 * I) + 1).getFloat(); + switch (E->getBuiltinCallee()) { + case clang::X86::BI__builtin_ia32_haddpd: + case clang::X86::BI__builtin_ia32_haddps: + case clang::X86::BI__builtin_ia32_haddps256: + case clang::X86::BI__builtin_ia32_haddpd256: + RHSA.add(RHSB, RM); + break; + case clang::X86::BI__builtin_ia32_hsubpd: + case clang::X86::BI__builtin_ia32_hsubps: + case clang::X86::BI__builtin_ia32_hsubps256: + case clang::X86::BI__builtin_ia32_hsubpd256: + RHSA.subtract(RHSB, RM); + break; + } + ResultElements.push_back(APValue(RHSA)); + } + } + return Success(APValue(ResultElements.data(), ResultElements.size()), E); + } case Builtin::BI__builtin_elementwise_fshl: case Builtin::BI__builtin_elementwise_fshr: { APValue SourceHi, SourceLo, SourceShift; @@ -13905,6 +14197,40 @@ static bool getBuiltinAlignArguments(const CallExpr *E, EvalInfo &Info, bool IntExprEvaluator::VisitBuiltinCallExpr(const CallExpr *E, unsigned BuiltinOp) { + auto EvalTestOp = [&](llvm::function_ref<bool(const APInt &, const APInt &)> + Fn) { + APValue SourceLHS, SourceRHS; + if (!EvaluateAsRValue(Info, E->getArg(0), SourceLHS) || + !EvaluateAsRValue(Info, E->getArg(1), SourceRHS)) + return false; + + unsigned SourceLen = SourceLHS.getVectorLength(); + const VectorType *VT = E->getArg(0)->getType()->castAs<VectorType>(); + QualType ElemQT = VT->getElementType(); + unsigned LaneWidth = Info.Ctx.getTypeSize(ElemQT); + + APInt AWide(LaneWidth * SourceLen, 0); + APInt BWide(LaneWidth * SourceLen, 0); + + for (unsigned I = 0; I != SourceLen; ++I) { + APInt ALane; + APInt BLane; + if (ElemQT->isIntegerType()) { // Get value. + ALane = SourceLHS.getVectorElt(I).getInt(); + BLane = SourceRHS.getVectorElt(I).getInt(); + } else if (ElemQT->isFloatingType()) { // Get only sign bit. + ALane = + SourceLHS.getVectorElt(I).getFloat().bitcastToAPInt().isNegative(); + BLane = + SourceRHS.getVectorElt(I).getFloat().bitcastToAPInt().isNegative(); + } else { // Must be integer or floating type. + return false; + } + AWide.insertBits(ALane, I * LaneWidth); + BWide.insertBits(BLane, I * LaneWidth); + } + return Success(Fn(AWide, BWide), E); + }; auto HandleMaskBinOp = [&](llvm::function_ref<APSInt(const APSInt &, const APSInt &)> Fn) @@ -15018,7 +15344,34 @@ bool IntExprEvaluator::VisitBuiltinCallExpr(const CallExpr *E, Result.setBitVal(P++, Val[I]); return Success(Result, E); } - + case X86::BI__builtin_ia32_ptestz128: + case X86::BI__builtin_ia32_ptestz256: + case X86::BI__builtin_ia32_vtestzps: + case X86::BI__builtin_ia32_vtestzps256: + case X86::BI__builtin_ia32_vtestzpd: + case X86::BI__builtin_ia32_vtestzpd256: { + return EvalTestOp( + [](const APInt &A, const APInt &B) { return (A & B) == 0; }); + } + case X86::BI__builtin_ia32_ptestc128: + case X86::BI__builtin_ia32_ptestc256: + case X86::BI__builtin_ia32_vtestcps: + case X86::BI__builtin_ia32_vtestcps256: + case X86::BI__builtin_ia32_vtestcpd: + case X86::BI__builtin_ia32_vtestcpd256: { + return EvalTestOp( + [](const APInt &A, const APInt &B) { return (~A & B) == 0; }); + } + case X86::BI__builtin_ia32_ptestnzc128: + case X86::BI__builtin_ia32_ptestnzc256: + case X86::BI__builtin_ia32_vtestnzcps: + case X86::BI__builtin_ia32_vtestnzcps256: + case X86::BI__builtin_ia32_vtestnzcpd: + case X86::BI__builtin_ia32_vtestnzcpd256: { + return EvalTestOp([](const APInt &A, const APInt &B) { + return ((A & B) != 0) && ((~A & B) != 0); + }); + } case X86::BI__builtin_ia32_kandqi: case X86::BI__builtin_ia32_kandhi: case X86::BI__builtin_ia32_kandsi: diff --git a/clang/lib/AST/InheritViz.cpp b/clang/lib/AST/InheritViz.cpp index 3c4a5a8..c5f4c2b 100644 --- a/clang/lib/AST/InheritViz.cpp +++ b/clang/lib/AST/InheritViz.cpp @@ -89,8 +89,8 @@ void InheritanceHierarchyWriter::WriteNode(QualType Type, bool FromVirtual) { Out << " \"];\n"; // Display the base classes. - const auto *Decl = cast<CXXRecordDecl>( - Type->castAsCanonical<RecordType>()->getOriginalDecl()); + const auto *Decl = + cast<CXXRecordDecl>(Type->castAsCanonical<RecordType>()->getDecl()); for (const auto &Base : Decl->bases()) { QualType CanonBaseType = Context.getCanonicalType(Base.getType()); diff --git a/clang/lib/AST/ItaniumMangle.cpp b/clang/lib/AST/ItaniumMangle.cpp index 844db79..5572e0a 100644 --- a/clang/lib/AST/ItaniumMangle.cpp +++ b/clang/lib/AST/ItaniumMangle.cpp @@ -2491,7 +2491,7 @@ bool CXXNameMangler::mangleUnresolvedTypeOrSimpleId(QualType Ty, case Type::Enum: case Type::Record: mangleSourceNameWithAbiTags( - cast<TagType>(Ty)->getOriginalDecl()->getDefinitionOrSelf()); + cast<TagType>(Ty)->getDecl()->getDefinitionOrSelf()); break; case Type::TemplateSpecialization: { @@ -2556,9 +2556,8 @@ bool CXXNameMangler::mangleUnresolvedTypeOrSimpleId(QualType Ty, } case Type::InjectedClassName: - mangleSourceNameWithAbiTags(cast<InjectedClassNameType>(Ty) - ->getOriginalDecl() - ->getDefinitionOrSelf()); + mangleSourceNameWithAbiTags( + cast<InjectedClassNameType>(Ty)->getDecl()->getDefinitionOrSelf()); break; case Type::DependentName: @@ -3795,7 +3794,7 @@ void CXXNameMangler::mangleType(const RecordType *T) { mangleType(static_cast<const TagType*>(T)); } void CXXNameMangler::mangleType(const TagType *T) { - mangleName(T->getOriginalDecl()->getDefinitionOrSelf()); + mangleName(T->getDecl()->getDefinitionOrSelf()); } // <type> ::= <array-type> @@ -4430,8 +4429,8 @@ void CXXNameMangler::mangleType(const InjectedClassNameType *T) { // Mangle injected class name types as if the user had written the // specialization out fully. It may not actually be possible to see // this mangling, though. - mangleType(T->getOriginalDecl()->getCanonicalTemplateSpecializationType( - getASTContext())); + mangleType( + T->getDecl()->getCanonicalTemplateSpecializationType(getASTContext())); } void CXXNameMangler::mangleType(const TemplateSpecializationType *T) { @@ -4692,7 +4691,7 @@ void CXXNameMangler::mangleIntegerLiteral(QualType T, void CXXNameMangler::mangleMemberExprBase(const Expr *Base, bool IsArrow) { // Ignore member expressions involving anonymous unions. while (const auto *RT = Base->getType()->getAsCanonical<RecordType>()) { - if (!RT->getOriginalDecl()->isAnonymousStructOrUnion()) + if (!RT->getDecl()->isAnonymousStructOrUnion()) break; const auto *ME = dyn_cast<MemberExpr>(Base); if (!ME) @@ -7010,8 +7009,7 @@ bool CXXNameMangler::isSpecializedAs(QualType S, llvm::StringRef Name, if (!RT) return false; - const ClassTemplateSpecializationDecl *SD = - dyn_cast<ClassTemplateSpecializationDecl>(RT->getOriginalDecl()); + const auto *SD = dyn_cast<ClassTemplateSpecializationDecl>(RT->getDecl()); if (!SD || !SD->getIdentifier()->isStr(Name)) return false; diff --git a/clang/lib/AST/JSONNodeDumper.cpp b/clang/lib/AST/JSONNodeDumper.cpp index 0ef6328..9f4dba9 100644 --- a/clang/lib/AST/JSONNodeDumper.cpp +++ b/clang/lib/AST/JSONNodeDumper.cpp @@ -396,7 +396,7 @@ llvm::json::Array JSONNodeDumper::createCastPath(const CastExpr *C) { for (auto I = C->path_begin(), E = C->path_end(); I != E; ++I) { const CXXBaseSpecifier *Base = *I; const auto *RD = cast<CXXRecordDecl>( - Base->getType()->castAsCanonical<RecordType>()->getOriginalDecl()); + Base->getType()->castAsCanonical<RecordType>()->getDecl()); llvm::json::Object Val{{"name", RD->getName()}}; if (Base->isVirtual()) @@ -764,7 +764,7 @@ void JSONNodeDumper::VisitTagType(const TagType *TT) { Qualifier.print(OS, PrintPolicy, /*ResolveTemplateArguments=*/true); JOS.attribute("qualifier", Str); } - JOS.attribute("decl", createBareDeclRef(TT->getOriginalDecl())); + JOS.attribute("decl", createBareDeclRef(TT->getDecl())); if (TT->isTagOwned()) JOS.attribute("isTagOwned", true); } @@ -816,7 +816,7 @@ void JSONNodeDumper::VisitTemplateSpecializationType( void JSONNodeDumper::VisitInjectedClassNameType( const InjectedClassNameType *ICNT) { - JOS.attribute("decl", createBareDeclRef(ICNT->getOriginalDecl())); + JOS.attribute("decl", createBareDeclRef(ICNT->getDecl())); } void JSONNodeDumper::VisitObjCInterfaceType(const ObjCInterfaceType *OIT) { diff --git a/clang/lib/AST/MicrosoftMangle.cpp b/clang/lib/AST/MicrosoftMangle.cpp index 8cbc72b..f1baf9f 100644 --- a/clang/lib/AST/MicrosoftMangle.cpp +++ b/clang/lib/AST/MicrosoftMangle.cpp @@ -3248,11 +3248,11 @@ void MicrosoftCXXNameMangler::mangleTagTypeKind(TagTypeKind TTK) { } void MicrosoftCXXNameMangler::mangleType(const EnumType *T, Qualifiers, SourceRange) { - mangleType(cast<TagType>(T)->getOriginalDecl()); + mangleType(cast<TagType>(T)->getDecl()); } void MicrosoftCXXNameMangler::mangleType(const RecordType *T, Qualifiers, SourceRange) { - mangleType(cast<TagType>(T)->getOriginalDecl()); + mangleType(cast<TagType>(T)->getDecl()); } void MicrosoftCXXNameMangler::mangleType(const TagDecl *TD) { // MSVC chooses the tag kind of the definition if it exists, otherwise it diff --git a/clang/lib/AST/ODRHash.cpp b/clang/lib/AST/ODRHash.cpp index 6842038..46a4e25 100644 --- a/clang/lib/AST/ODRHash.cpp +++ b/clang/lib/AST/ODRHash.cpp @@ -913,7 +913,7 @@ public: return false; if (TypedefT->getDecl()->getIdentifier() != - TagT->getOriginalDecl()->getIdentifier()) + TagT->getDecl()->getIdentifier()) return false; ID.AddInteger(TagT->getTypeClass()); @@ -1059,7 +1059,7 @@ public: } void VisitInjectedClassNameType(const InjectedClassNameType *T) { - AddDecl(T->getOriginalDecl()->getDefinitionOrSelf()); + AddDecl(T->getDecl()->getDefinitionOrSelf()); VisitType(T); } @@ -1164,7 +1164,7 @@ public: AddNestedNameSpecifier(ElaboratedOverride ? ElaboratedOverride->getQualifier() : T->getQualifier()); - AddDecl(T->getOriginalDecl()->getDefinitionOrSelf()); + AddDecl(T->getDecl()->getDefinitionOrSelf()); VisitType(T); } diff --git a/clang/lib/AST/OpenACCClause.cpp b/clang/lib/AST/OpenACCClause.cpp index 6c4bc7c..142c932 100644 --- a/clang/lib/AST/OpenACCClause.cpp +++ b/clang/lib/AST/OpenACCClause.cpp @@ -506,11 +506,17 @@ OpenACCDeviceTypeClause *OpenACCDeviceTypeClause::Create( OpenACCReductionClause *OpenACCReductionClause::Create( const ASTContext &C, SourceLocation BeginLoc, SourceLocation LParenLoc, OpenACCReductionOperator Operator, ArrayRef<Expr *> VarList, - ArrayRef<OpenACCReductionRecipe> Recipes, + ArrayRef<OpenACCReductionRecipeWithStorage> Recipes, SourceLocation EndLoc) { - void *Mem = C.Allocate( - OpenACCReductionClause::totalSizeToAlloc<Expr *, OpenACCReductionRecipe>( - VarList.size(), Recipes.size())); + size_t NumCombiners = llvm::accumulate( + Recipes, 0, [](size_t Num, const OpenACCReductionRecipeWithStorage &R) { + return Num + R.CombinerRecipes.size(); + }); + + void *Mem = C.Allocate(OpenACCReductionClause::totalSizeToAlloc< + Expr *, OpenACCReductionRecipe, + OpenACCReductionRecipe::CombinerRecipe>( + VarList.size(), Recipes.size(), NumCombiners)); return new (Mem) OpenACCReductionClause(BeginLoc, LParenLoc, Operator, VarList, Recipes, EndLoc); } diff --git a/clang/lib/AST/OpenMPClause.cpp b/clang/lib/AST/OpenMPClause.cpp index 2ce4419..791df7e 100644 --- a/clang/lib/AST/OpenMPClause.cpp +++ b/clang/lib/AST/OpenMPClause.cpp @@ -309,6 +309,12 @@ OMPClause::child_range OMPIfClause::used_children() { return child_range(&Condition, &Condition + 1); } +OMPClause::child_range OMPNowaitClause::used_children() { + if (Condition) + return child_range(&Condition, &Condition + 1); + return children(); +} + OMPClause::child_range OMPGrainsizeClause::used_children() { if (Stmt **C = getAddrOfExprAsWritten(getPreInitStmt())) return child_range(C, C + 1); @@ -2113,8 +2119,13 @@ void OMPClausePrinter::VisitOMPOrderedClause(OMPOrderedClause *Node) { } } -void OMPClausePrinter::VisitOMPNowaitClause(OMPNowaitClause *) { +void OMPClausePrinter::VisitOMPNowaitClause(OMPNowaitClause *Node) { OS << "nowait"; + if (auto *Cond = Node->getCondition()) { + OS << "("; + Cond->printPretty(OS, nullptr, Policy, 0); + OS << ")"; + } } void OMPClausePrinter::VisitOMPUntiedClause(OMPUntiedClause *) { diff --git a/clang/lib/AST/ParentMapContext.cpp b/clang/lib/AST/ParentMapContext.cpp index acc011c..7138dff 100644 --- a/clang/lib/AST/ParentMapContext.cpp +++ b/clang/lib/AST/ParentMapContext.cpp @@ -20,36 +20,6 @@ using namespace clang; -ParentMapContext::ParentMapContext(ASTContext &Ctx) : ASTCtx(Ctx) {} - -ParentMapContext::~ParentMapContext() = default; - -void ParentMapContext::clear() { Parents.reset(); } - -const Expr *ParentMapContext::traverseIgnored(const Expr *E) const { - return traverseIgnored(const_cast<Expr *>(E)); -} - -Expr *ParentMapContext::traverseIgnored(Expr *E) const { - if (!E) - return nullptr; - - switch (Traversal) { - case TK_AsIs: - return E; - case TK_IgnoreUnlessSpelledInSource: - return E->IgnoreUnlessSpelledInSource(); - } - llvm_unreachable("Invalid Traversal type!"); -} - -DynTypedNode ParentMapContext::traverseIgnored(const DynTypedNode &N) const { - if (const auto *E = N.get<Expr>()) { - return DynTypedNode::create(*traverseIgnored(E)); - } - return N; -} - template <typename T, typename... U> static std::tuple<bool, DynTypedNodeList, const T *, const U *...> matchParents(const DynTypedNodeList &NodeList, @@ -334,6 +304,36 @@ matchParents(const DynTypedNodeList &NodeList, return MatchParents<T, U...>::match(NodeList, ParentMap); } +ParentMapContext::ParentMapContext(ASTContext &Ctx) : ASTCtx(Ctx) {} + +ParentMapContext::~ParentMapContext() = default; + +void ParentMapContext::clear() { Parents.reset(); } + +const Expr *ParentMapContext::traverseIgnored(const Expr *E) const { + return traverseIgnored(const_cast<Expr *>(E)); +} + +Expr *ParentMapContext::traverseIgnored(Expr *E) const { + if (!E) + return nullptr; + + switch (Traversal) { + case TK_AsIs: + return E; + case TK_IgnoreUnlessSpelledInSource: + return E->IgnoreUnlessSpelledInSource(); + } + llvm_unreachable("Invalid Traversal type!"); +} + +DynTypedNode ParentMapContext::traverseIgnored(const DynTypedNode &N) const { + if (const auto *E = N.get<Expr>()) { + return DynTypedNode::create(*traverseIgnored(E)); + } + return N; +} + /// Template specializations to abstract away from pointers and TypeLocs. /// @{ template <typename T> static DynTypedNode createDynTypedNode(const T &Node) { diff --git a/clang/lib/AST/QualTypeNames.cpp b/clang/lib/AST/QualTypeNames.cpp index a2f9309..1918416 100644 --- a/clang/lib/AST/QualTypeNames.cpp +++ b/clang/lib/AST/QualTypeNames.cpp @@ -125,7 +125,7 @@ static const Type *getFullyQualifiedTemplateType(const ASTContext &Ctx, // which can point to a template instantiation with no sugar in any of // its template argument, however we still need to fully qualify them. - const auto *TD = TSTRecord->getOriginalDecl(); + const auto *TD = TSTRecord->getDecl(); const auto *TSTDecl = dyn_cast<ClassTemplateSpecializationDecl>(TD); if (!TSTDecl) return Ctx.getTagType(Keyword, Qualifier, TD, /*OwnsTag=*/false) @@ -232,7 +232,7 @@ static NestedNameSpecifier getFullyQualifiedNestedNameSpecifier( // Find decl context. const TypeDecl *TD; if (const TagType *TagDeclType = Type->getAs<TagType>()) - TD = TagDeclType->getOriginalDecl(); + TD = TagDeclType->getDecl(); else if (const auto *D = dyn_cast<TypedefType>(Type)) TD = D->getDecl(); else @@ -316,7 +316,7 @@ createNestedNameSpecifierForScopeOf(const ASTContext &Ctx, const Type *TypePtr, if (const auto *TDT = dyn_cast<TypedefType>(TypePtr)) { Decl = TDT->getDecl(); } else if (const auto *TagDeclType = dyn_cast<TagType>(TypePtr)) { - Decl = TagDeclType->getOriginalDecl(); + Decl = TagDeclType->getDecl(); } else if (const auto *TST = dyn_cast<TemplateSpecializationType>(TypePtr)) { Decl = TST->getTemplateName().getAsTemplateDecl(); } else { diff --git a/clang/lib/AST/RecordLayoutBuilder.cpp b/clang/lib/AST/RecordLayoutBuilder.cpp index 00b938b..ac18d4d 100644 --- a/clang/lib/AST/RecordLayoutBuilder.cpp +++ b/clang/lib/AST/RecordLayoutBuilder.cpp @@ -2009,7 +2009,7 @@ void ItaniumRecordLayoutBuilder::LayoutField(const FieldDecl *D, } else if (const BuiltinType *BTy = BaseTy->getAs<BuiltinType>()) { performBuiltinTypeAlignmentUpgrade(BTy); } else if (const RecordType *RT = BaseTy->getAsCanonical<RecordType>()) { - const RecordDecl *RD = RT->getOriginalDecl(); + const RecordDecl *RD = RT->getDecl(); const ASTRecordLayout &FieldRecord = Context.getASTRecordLayout(RD); PreferredAlign = FieldRecord.getPreferredAlignment(); } @@ -2710,7 +2710,7 @@ MicrosoftRecordLayoutBuilder::getAdjustedElementInfo( if (const auto *RT = FD->getType() ->getBaseElementTypeUnsafe() ->getAsCanonical<RecordType>()) { - auto const &Layout = Context.getASTRecordLayout(RT->getOriginalDecl()); + auto const &Layout = Context.getASTRecordLayout(RT->getDecl()); EndsWithZeroSizedObject = Layout.endsWithZeroSizedObject(); FieldRequiredAlignment = std::max(FieldRequiredAlignment, Layout.getRequiredAlignment()); diff --git a/clang/lib/AST/Stmt.cpp b/clang/lib/AST/Stmt.cpp index 9ae8aea..11ece49 100644 --- a/clang/lib/AST/Stmt.cpp +++ b/clang/lib/AST/Stmt.cpp @@ -252,7 +252,7 @@ namespace { template <class T> good implements_children(children_t T::*) { return good(); } - LLVM_ATTRIBUTE_UNUSED + [[maybe_unused]] static bad implements_children(children_t Stmt::*) { return bad(); } @@ -261,15 +261,19 @@ namespace { template <class T> good implements_getBeginLoc(getBeginLoc_t T::*) { return good(); } - LLVM_ATTRIBUTE_UNUSED - static bad implements_getBeginLoc(getBeginLoc_t Stmt::*) { return bad(); } + [[maybe_unused]] + static bad implements_getBeginLoc(getBeginLoc_t Stmt::*) { + return bad(); + } typedef SourceLocation getLocEnd_t() const; template <class T> good implements_getEndLoc(getLocEnd_t T::*) { return good(); } - LLVM_ATTRIBUTE_UNUSED - static bad implements_getEndLoc(getLocEnd_t Stmt::*) { return bad(); } + [[maybe_unused]] + static bad implements_getEndLoc(getLocEnd_t Stmt::*) { + return bad(); + } #define ASSERT_IMPLEMENTS_children(type) \ (void) is_good(implements_children(&type::children)) @@ -282,7 +286,7 @@ namespace { /// Check whether the various Stmt classes implement their member /// functions. -LLVM_ATTRIBUTE_UNUSED +[[maybe_unused]] static inline void check_implementations() { #define ABSTRACT_STMT(type) #define STMT(type, base) \ diff --git a/clang/lib/AST/StmtPrinter.cpp b/clang/lib/AST/StmtPrinter.cpp index 586c300..ff8ca01 100644 --- a/clang/lib/AST/StmtPrinter.cpp +++ b/clang/lib/AST/StmtPrinter.cpp @@ -151,11 +151,11 @@ namespace { else StmtVisitor<StmtPrinter>::Visit(S); } - void VisitStmt(Stmt *Node) LLVM_ATTRIBUTE_UNUSED { + [[maybe_unused]] void VisitStmt(Stmt *Node) { Indent() << "<<unknown stmt type>>" << NL; } - void VisitExpr(Expr *Node) LLVM_ATTRIBUTE_UNUSED { + [[maybe_unused]] void VisitExpr(Expr *Node) { OS << "<<unknown expr type>>"; } diff --git a/clang/lib/AST/StmtProfile.cpp b/clang/lib/AST/StmtProfile.cpp index 3cd033e..05b64cc 100644 --- a/clang/lib/AST/StmtProfile.cpp +++ b/clang/lib/AST/StmtProfile.cpp @@ -585,7 +585,10 @@ void OMPClauseProfiler::VisitOMPOrderedClause(const OMPOrderedClause *C) { Profiler->VisitStmt(Num); } -void OMPClauseProfiler::VisitOMPNowaitClause(const OMPNowaitClause *) {} +void OMPClauseProfiler::VisitOMPNowaitClause(const OMPNowaitClause *C) { + if (C->getCondition()) + Profiler->VisitStmt(C->getCondition()); +} void OMPClauseProfiler::VisitOMPUntiedClause(const OMPUntiedClause *) {} diff --git a/clang/lib/AST/TemplateBase.cpp b/clang/lib/AST/TemplateBase.cpp index 76f96fb..131ae6e 100644 --- a/clang/lib/AST/TemplateBase.cpp +++ b/clang/lib/AST/TemplateBase.cpp @@ -340,13 +340,14 @@ bool TemplateArgument::isPackExpansion() const { } bool TemplateArgument::isConceptOrConceptTemplateParameter() const { - if (getKind() == TemplateArgument::Template) { - if (isa<ConceptDecl>(getAsTemplate().getAsTemplateDecl())) - return true; - else if (auto *TTP = dyn_cast_if_present<TemplateTemplateParmDecl>( - getAsTemplate().getAsTemplateDecl())) - return TTP->templateParameterKind() == TNK_Concept_template; - } + if (getKind() != TemplateArgument::Template) + return false; + + if (isa_and_nonnull<ConceptDecl>(getAsTemplate().getAsTemplateDecl())) + return true; + if (auto *TTP = llvm::dyn_cast_or_null<TemplateTemplateParmDecl>( + getAsTemplate().getAsTemplateDecl())) + return TTP->templateParameterKind() == TNK_Concept_template; return false; } diff --git a/clang/lib/AST/TextNodeDumper.cpp b/clang/lib/AST/TextNodeDumper.cpp index cf5e914..41aebdb 100644 --- a/clang/lib/AST/TextNodeDumper.cpp +++ b/clang/lib/AST/TextNodeDumper.cpp @@ -1401,7 +1401,7 @@ static void dumpBasePath(raw_ostream &OS, const CastExpr *Node) { OS << " -> "; const auto *RD = cast<CXXRecordDecl>( - Base->getType()->castAsCanonical<RecordType>()->getOriginalDecl()); + Base->getType()->castAsCanonical<RecordType>()->getDecl()); if (Base->isVirtual()) OS << "virtual "; @@ -2180,7 +2180,7 @@ void TextNodeDumper::VisitTagType(const TagType *T) { K != ElaboratedTypeKeyword::None) OS << ' ' << TypeWithKeyword::getKeywordName(K); dumpNestedNameSpecifier(T->getQualifier()); - dumpDeclRef(T->getOriginalDecl()); + dumpDeclRef(T->getDecl()); } void TextNodeDumper::VisitTemplateTypeParmType(const TemplateTypeParmType *T) { @@ -2232,7 +2232,7 @@ void TextNodeDumper::VisitTemplateSpecializationType( void TextNodeDumper::VisitInjectedClassNameType( const InjectedClassNameType *T) { - dumpDeclRef(T->getOriginalDecl()); + dumpDeclRef(T->getDecl()); } void TextNodeDumper::VisitObjCInterfaceType(const ObjCInterfaceType *T) { diff --git a/clang/lib/AST/Type.cpp b/clang/lib/AST/Type.cpp index ee7a68e..4548af1 100644 --- a/clang/lib/AST/Type.cpp +++ b/clang/lib/AST/Type.cpp @@ -114,7 +114,7 @@ const IdentifierInfo *QualType::getBaseTypeIdentifier() const { if (ty->isPointerOrReferenceType()) return ty->getPointeeType().getBaseTypeIdentifier(); if (const auto *TT = ty->getAs<TagType>()) - ND = TT->getOriginalDecl(); + ND = TT->getDecl(); else if (ty->getTypeClass() == Type::Typedef) ND = ty->castAs<TypedefType>()->getDecl(); else if (ty->isArrayType()) @@ -671,13 +671,13 @@ const Type *Type::getUnqualifiedDesugaredType() const { bool Type::isClassType() const { if (const auto *RT = getAsCanonical<RecordType>()) - return RT->getOriginalDecl()->isClass(); + return RT->getDecl()->isClass(); return false; } bool Type::isStructureType() const { if (const auto *RT = getAsCanonical<RecordType>()) - return RT->getOriginalDecl()->isStruct(); + return RT->getDecl()->isStruct(); return false; } @@ -685,7 +685,7 @@ bool Type::isStructureTypeWithFlexibleArrayMember() const { const auto *RT = getAsCanonical<RecordType>(); if (!RT) return false; - const auto *Decl = RT->getOriginalDecl(); + const auto *Decl = RT->getDecl(); if (!Decl->isStruct()) return false; return Decl->getDefinitionOrSelf()->hasFlexibleArrayMember(); @@ -699,13 +699,13 @@ bool Type::isObjCBoxableRecordType() const { bool Type::isInterfaceType() const { if (const auto *RT = getAsCanonical<RecordType>()) - return RT->getOriginalDecl()->isInterface(); + return RT->getDecl()->isInterface(); return false; } bool Type::isStructureOrClassType() const { if (const auto *RT = getAsCanonical<RecordType>()) - return RT->getOriginalDecl()->isStructureOrClass(); + return RT->getDecl()->isStructureOrClass(); return false; } @@ -717,7 +717,7 @@ bool Type::isVoidPointerType() const { bool Type::isUnionType() const { if (const auto *RT = getAsCanonical<RecordType>()) - return RT->getOriginalDecl()->isUnion(); + return RT->getDecl()->isUnion(); return false; } @@ -734,7 +734,7 @@ bool Type::isComplexIntegerType() const { bool Type::isScopedEnumeralType() const { if (const auto *ET = getAsCanonical<EnumType>()) - return ET->getOriginalDecl()->isScoped(); + return ET->getDecl()->isScoped(); return false; } @@ -768,13 +768,13 @@ QualType Type::getPointeeType() const { const RecordType *Type::getAsStructureType() const { // If this is directly a structure type, return it. if (const auto *RT = dyn_cast<RecordType>(this)) { - if (RT->getOriginalDecl()->isStruct()) + if (RT->getDecl()->isStruct()) return RT; } // If the canonical form of this type isn't the right kind, reject it. if (const auto *RT = dyn_cast<RecordType>(CanonicalType)) { - if (!RT->getOriginalDecl()->isStruct()) + if (!RT->getDecl()->isStruct()) return nullptr; // If this is a typedef for a structure type, strip the typedef off without @@ -787,13 +787,13 @@ const RecordType *Type::getAsStructureType() const { const RecordType *Type::getAsUnionType() const { // If this is directly a union type, return it. if (const auto *RT = dyn_cast<RecordType>(this)) { - if (RT->getOriginalDecl()->isUnion()) + if (RT->getDecl()->isUnion()) return RT; } // If the canonical form of this type isn't the right kind, reject it. if (const auto *RT = dyn_cast<RecordType>(CanonicalType)) { - if (!RT->getOriginalDecl()->isUnion()) + if (!RT->getDecl()->isUnion()) return nullptr; // If this is a typedef for a union type, strip the typedef off without @@ -2107,7 +2107,7 @@ bool Type::isIntegralType(const ASTContext &Ctx) const { // Complete enum types are integral in C. if (!Ctx.getLangOpts().CPlusPlus) if (const auto *ET = dyn_cast<EnumType>(CanonicalType)) - return IsEnumDeclComplete(ET->getOriginalDecl()); + return IsEnumDeclComplete(ET->getDecl()); return isBitIntType(); } @@ -2124,7 +2124,7 @@ bool Type::isIntegralOrUnscopedEnumerationType() const { bool Type::isUnscopedEnumerationType() const { if (const auto *ET = dyn_cast<EnumType>(CanonicalType)) - return !ET->getOriginalDecl()->isScoped(); + return !ET->getDecl()->isScoped(); return false; } @@ -2328,7 +2328,7 @@ bool Type::isRealType() const { return BT->getKind() >= BuiltinType::Bool && BT->getKind() <= BuiltinType::Ibm128; if (const auto *ET = dyn_cast<EnumType>(CanonicalType)) { - const auto *ED = ET->getOriginalDecl(); + const auto *ED = ET->getDecl(); return !ED->isScoped() && ED->getDefinitionOrSelf()->isComplete(); } return isBitIntType(); @@ -2345,7 +2345,7 @@ bool Type::isArithmeticType() const { // C++0x: Enumerations are not arithmetic types. For now, just return // false for scoped enumerations since that will disable any // unwanted implicit conversions. - const auto *ED = ET->getOriginalDecl(); + const auto *ED = ET->getDecl(); return !ED->isScoped() && ED->getDefinitionOrSelf()->isComplete(); } return isa<ComplexType>(CanonicalType) || isBitIntType(); @@ -2410,8 +2410,7 @@ Type::ScalarTypeKind Type::getScalarTypeKind() const { /// includes union types. bool Type::isAggregateType() const { if (const auto *Record = dyn_cast<RecordType>(CanonicalType)) { - if (const auto *ClassDecl = - dyn_cast<CXXRecordDecl>(Record->getOriginalDecl())) + if (const auto *ClassDecl = dyn_cast<CXXRecordDecl>(Record->getDecl())) return ClassDecl->isAggregate(); return true; @@ -2746,8 +2745,8 @@ bool QualType::isCXX98PODType(const ASTContext &Context) const { return true; case Type::Record: - if (const auto *ClassDecl = dyn_cast<CXXRecordDecl>( - cast<RecordType>(CanonicalType)->getOriginalDecl())) + if (const auto *ClassDecl = + dyn_cast<CXXRecordDecl>(cast<RecordType>(CanonicalType)->getDecl())) return ClassDecl->isPOD(); // C struct/union is POD. @@ -3179,7 +3178,7 @@ bool Type::isNothrowT() const { bool Type::isAlignValT() const { if (const auto *ET = getAsCanonical<EnumType>()) { - const auto *ED = ET->getOriginalDecl(); + const auto *ED = ET->getDecl(); IdentifierInfo *II = ED->getIdentifier(); if (II && II->isStr("align_val_t") && ED->isInStdNamespace()) return true; @@ -3189,7 +3188,7 @@ bool Type::isAlignValT() const { bool Type::isStdByteType() const { if (const auto *ET = getAsCanonical<EnumType>()) { - const auto *ED = ET->getOriginalDecl(); + const auto *ED = ET->getDecl(); IdentifierInfo *II = ED->getIdentifier(); if (II && II->isStr("byte") && ED->isInStdNamespace()) return true; @@ -4321,7 +4320,7 @@ bool RecordType::hasConstFields() const { while (RecordTypeList.size() > NextToCheckIndex) { for (FieldDecl *FD : RecordTypeList[NextToCheckIndex] - ->getOriginalDecl() + ->getDecl() ->getDefinitionOrSelf() ->fields()) { QualType FieldTy = FD->getType(); @@ -4815,8 +4814,7 @@ static CachedProperties computeCachedProperties(const Type *T) { case Type::Record: case Type::Enum: { - const TagDecl *Tag = - cast<TagType>(T)->getOriginalDecl()->getDefinitionOrSelf(); + const auto *Tag = cast<TagType>(T)->getDecl()->getDefinitionOrSelf(); // C++ [basic.link]p8: // - it is a class or enumeration type that is named (or has a name @@ -4926,7 +4924,7 @@ LinkageInfo LinkageComputer::computeTypeLinkageInfo(const Type *T) { case Type::Record: case Type::Enum: return getDeclLinkageAndVisibility( - cast<TagType>(T)->getOriginalDecl()->getDefinitionOrSelf()); + cast<TagType>(T)->getDecl()->getDefinitionOrSelf()); case Type::Complex: return computeTypeLinkageInfo(cast<ComplexType>(T)->getElementType()); @@ -5129,7 +5127,7 @@ bool Type::canHaveNullability(bool ResultIfUnknown) const { llvm_unreachable("unknown builtin type"); case Type::Record: { - const RecordDecl *RD = cast<RecordType>(type)->getOriginalDecl(); + const auto *RD = cast<RecordType>(type)->getDecl(); // For template specializations, look only at primary template attributes. // This is a consistent regardless of whether the instantiation is known. if (const auto *CTSD = dyn_cast<ClassTemplateSpecializationDecl>(RD)) @@ -5327,7 +5325,7 @@ bool Type::isCARCBridgableType() const { /// Check if the specified type is the CUDA device builtin surface type. bool Type::isCUDADeviceBuiltinSurfaceType() const { if (const auto *RT = getAsCanonical<RecordType>()) - return RT->getOriginalDecl() + return RT->getDecl() ->getMostRecentDecl() ->hasAttr<CUDADeviceBuiltinSurfaceTypeAttr>(); return false; @@ -5336,7 +5334,7 @@ bool Type::isCUDADeviceBuiltinSurfaceType() const { /// Check if the specified type is the CUDA device builtin texture type. bool Type::isCUDADeviceBuiltinTextureType() const { if (const auto *RT = getAsCanonical<RecordType>()) - return RT->getOriginalDecl() + return RT->getDecl() ->getMostRecentDecl() ->hasAttr<CUDADeviceBuiltinTextureTypeAttr>(); return false; diff --git a/clang/lib/AST/TypeLoc.cpp b/clang/lib/AST/TypeLoc.cpp index 55476e2..f54ccf0 100644 --- a/clang/lib/AST/TypeLoc.cpp +++ b/clang/lib/AST/TypeLoc.cpp @@ -303,8 +303,7 @@ bool TypeSpecTypeLoc::isKind(const TypeLoc &TL) { } bool TagTypeLoc::isDefinition() const { - return getTypePtr()->isTagOwned() && - getOriginalDecl()->isCompleteDefinition(); + return getTypePtr()->isTagOwned() && getDecl()->isCompleteDefinition(); } // Reimplemented to account for GNU/C++ extension @@ -494,39 +493,6 @@ NestedNameSpecifierLoc TypeLoc::getPrefix() const { } } -SourceLocation TypeLoc::getNonPrefixBeginLoc() const { - switch (getTypeLocClass()) { - case TypeLoc::TemplateSpecialization: { - auto TL = castAs<TemplateSpecializationTypeLoc>(); - SourceLocation Loc = TL.getTemplateKeywordLoc(); - if (!Loc.isValid()) - Loc = TL.getTemplateNameLoc(); - return Loc; - } - case TypeLoc::DeducedTemplateSpecialization: { - auto TL = castAs<DeducedTemplateSpecializationTypeLoc>(); - SourceLocation Loc = TL.getTemplateKeywordLoc(); - if (!Loc.isValid()) - Loc = TL.getTemplateNameLoc(); - return Loc; - } - case TypeLoc::DependentName: - return castAs<DependentNameTypeLoc>().getNameLoc(); - case TypeLoc::Enum: - case TypeLoc::Record: - case TypeLoc::InjectedClassName: - return castAs<TagTypeLoc>().getNameLoc(); - case TypeLoc::Typedef: - return castAs<TypedefTypeLoc>().getNameLoc(); - case TypeLoc::UnresolvedUsing: - return castAs<UnresolvedUsingTypeLoc>().getNameLoc(); - case TypeLoc::Using: - return castAs<UsingTypeLoc>().getNameLoc(); - default: - return getBeginLoc(); - } -} - SourceLocation TypeLoc::getNonElaboratedBeginLoc() const { // For elaborated types (e.g. `struct a::A`) we want the portion after the // `struct` but including the namespace qualifier, `a::`. diff --git a/clang/lib/AST/TypePrinter.cpp b/clang/lib/AST/TypePrinter.cpp index 66a1b68..2da7789 100644 --- a/clang/lib/AST/TypePrinter.cpp +++ b/clang/lib/AST/TypePrinter.cpp @@ -1421,7 +1421,7 @@ void TypePrinter::printDeducedTemplateSpecializationBefore( } else { // Should only get here for canonical types. const auto *CD = cast<ClassTemplateSpecializationDecl>( - cast<RecordType>(T->getDeducedType())->getOriginalDecl()); + cast<RecordType>(T->getDeducedType())->getDecl()); DeducedTD = CD->getSpecializedTemplate(); Args = CD->getTemplateArgs().asArray(); } @@ -1565,7 +1565,7 @@ void TypePrinter::AppendScope(DeclContext *DC, raw_ostream &OS, } void TypePrinter::printTagType(const TagType *T, raw_ostream &OS) { - TagDecl *D = T->getOriginalDecl(); + TagDecl *D = T->getDecl(); if (Policy.IncludeTagDefinition && T->isTagOwned()) { D->print(OS, Policy, Indentation); @@ -1669,11 +1669,11 @@ void TypePrinter::printTagType(const TagType *T, raw_ostream &OS) { void TypePrinter::printRecordBefore(const RecordType *T, raw_ostream &OS) { // Print the preferred name if we have one for this type. if (Policy.UsePreferredNames) { - for (const auto *PNA : T->getOriginalDecl() + for (const auto *PNA : T->getDecl() ->getMostRecentDecl() ->specific_attrs<PreferredNameAttr>()) { if (!declaresSameEntity(PNA->getTypedefType()->getAsCXXRecordDecl(), - T->getOriginalDecl())) + T->getDecl())) continue; // Find the outermost typedef or alias template. QualType T = PNA->getTypedefType(); @@ -1700,11 +1700,11 @@ void TypePrinter::printEnumAfter(const EnumType *T, raw_ostream &OS) {} void TypePrinter::printInjectedClassNameBefore(const InjectedClassNameType *T, raw_ostream &OS) { - const ASTContext &Ctx = T->getOriginalDecl()->getASTContext(); + const ASTContext &Ctx = T->getDecl()->getASTContext(); IncludeStrongLifetimeRAII Strong(Policy); T->getTemplateName(Ctx).print(OS, Policy); if (Policy.PrintInjectedClassNameWithArguments) { - auto *Decl = T->getOriginalDecl(); + auto *Decl = T->getDecl(); // FIXME: Use T->getTemplateArgs(Ctx) when that supports as-written // arguments. if (auto *RD = dyn_cast<ClassTemplateSpecializationDecl>(Decl)) { diff --git a/clang/lib/AST/VTableBuilder.cpp b/clang/lib/AST/VTableBuilder.cpp index 6cec526..3ded3a5 100644 --- a/clang/lib/AST/VTableBuilder.cpp +++ b/clang/lib/AST/VTableBuilder.cpp @@ -312,13 +312,12 @@ ComputeReturnAdjustmentBaseOffset(ASTContext &Context, return BaseOffset(); } - const CXXRecordDecl *DerivedRD = - cast<CXXRecordDecl>( - cast<RecordType>(CanDerivedReturnType)->getOriginalDecl()) + const auto *DerivedRD = + cast<CXXRecordDecl>(cast<RecordType>(CanDerivedReturnType)->getDecl()) ->getDefinitionOrSelf(); - const CXXRecordDecl *BaseRD = cast<CXXRecordDecl>( - cast<RecordType>(CanBaseReturnType)->getOriginalDecl()); + const auto *BaseRD = + cast<CXXRecordDecl>(cast<RecordType>(CanBaseReturnType)->getDecl()); return ComputeBaseOffset(Context, BaseRD, DerivedRD); } diff --git a/clang/lib/Analysis/LifetimeSafety/FactsGenerator.cpp b/clang/lib/Analysis/LifetimeSafety/FactsGenerator.cpp index 485308f..9b68de1 100644 --- a/clang/lib/Analysis/LifetimeSafety/FactsGenerator.cpp +++ b/clang/lib/Analysis/LifetimeSafety/FactsGenerator.cpp @@ -9,9 +9,11 @@ #include "clang/Analysis/Analyses/LifetimeSafety/FactsGenerator.h" #include "clang/Analysis/Analyses/LifetimeSafety/LifetimeAnnotations.h" #include "clang/Analysis/Analyses/PostOrderCFGView.h" +#include "llvm/Support/Casting.h" #include "llvm/Support/TimeProfiler.h" namespace clang::lifetimes::internal { +using llvm::isa_and_present; static bool isGslPointerType(QualType QT) { if (const auto *RD = QT->getAsCXXRecordDecl()) { @@ -108,7 +110,7 @@ void FactsGenerator::VisitCXXMemberCallExpr(const CXXMemberCallExpr *MCE) { // Specifically for conversion operators, // like `std::string_view p = std::string{};` if (isGslPointerType(MCE->getType()) && - isa<CXXConversionDecl>(MCE->getCalleeDecl())) { + isa_and_present<CXXConversionDecl>(MCE->getCalleeDecl())) { // The argument is the implicit object itself. handleFunctionCall(MCE, MCE->getMethodDecl(), {MCE->getImplicitObjectArgument()}, diff --git a/clang/lib/Analysis/ThreadSafety.cpp b/clang/lib/Analysis/ThreadSafety.cpp index a56fdb1..77750cf 100644 --- a/clang/lib/Analysis/ThreadSafety.cpp +++ b/clang/lib/Analysis/ThreadSafety.cpp @@ -2032,9 +2032,7 @@ void BuildLockset::handleCall(const Expr *Exp, const NamedDecl *D, assert(inserted.second && "Are we visiting the same expression again?"); if (isa<CXXConstructExpr>(Exp)) Self = Placeholder; - if (TagT->getOriginalDecl() - ->getMostRecentDecl() - ->hasAttr<ScopedLockableAttr>()) + if (TagT->getDecl()->getMostRecentDecl()->hasAttr<ScopedLockableAttr>()) Scp = CapabilityExpr(Placeholder, Exp->getType(), /*Neg=*/false); } diff --git a/clang/lib/Basic/Diagnostic.cpp b/clang/lib/Basic/Diagnostic.cpp index 8ecbd3c..2dec26e 100644 --- a/clang/lib/Basic/Diagnostic.cpp +++ b/clang/lib/Basic/Diagnostic.cpp @@ -525,7 +525,8 @@ std::unique_ptr<WarningsSpecialCaseList> WarningsSpecialCaseList::create(const llvm::MemoryBuffer &Input, std::string &Err) { auto WarningSuppressionList = std::make_unique<WarningsSpecialCaseList>(); - if (!WarningSuppressionList->createInternal(&Input, Err)) + if (!WarningSuppressionList->createInternal(&Input, Err, + /*OrderBySize=*/true)) return nullptr; return WarningSuppressionList; } diff --git a/clang/lib/Basic/LangOptions.cpp b/clang/lib/Basic/LangOptions.cpp index f034514..19b5576 100644 --- a/clang/lib/Basic/LangOptions.cpp +++ b/clang/lib/Basic/LangOptions.cpp @@ -11,6 +11,7 @@ //===----------------------------------------------------------------------===// #include "clang/Basic/LangOptions.h" +#include "clang/Basic/LangStandard.h" #include "llvm/Support/Path.h" using namespace clang; @@ -131,8 +132,12 @@ void LangOptions::setLangDefaults(LangOptions &Opts, Language Lang, Opts.NamedLoops = Std.isC2y(); Opts.HLSL = Lang == Language::HLSL; - if (Opts.HLSL && Opts.IncludeDefaultHeader) - Includes.push_back("hlsl.h"); + if (Opts.HLSL) { + if (Opts.IncludeDefaultHeader) + Includes.push_back("hlsl.h"); + // Set maximum matrix dimension to 4 for HLSL + Opts.MaxMatrixDimension = 4; + } // Set OpenCL Version. Opts.OpenCL = Std.isOpenCL(); @@ -243,3 +248,46 @@ LLVM_DUMP_METHOD void FPOptionsOverride::dump() { #include "clang/Basic/FPOptions.def" llvm::errs() << "\n"; } + +std::optional<uint32_t> LangOptions::getCPlusPlusLangStd() const { + if (!CPlusPlus) + return std::nullopt; + + LangStandard::Kind Std; + if (CPlusPlus26) + Std = LangStandard::lang_cxx26; + else if (CPlusPlus23) + Std = LangStandard::lang_cxx23; + else if (CPlusPlus20) + Std = LangStandard::lang_cxx20; + else if (CPlusPlus17) + Std = LangStandard::lang_cxx17; + else if (CPlusPlus14) + Std = LangStandard::lang_cxx14; + else if (CPlusPlus11) + Std = LangStandard::lang_cxx11; + else + Std = LangStandard::lang_cxx98; + + return LangStandard::getLangStandardForKind(Std).getVersion(); +} + +std::optional<uint32_t> LangOptions::getCLangStd() const { + LangStandard::Kind Std; + if (C2y) + Std = LangStandard::lang_c2y; + else if (C23) + Std = LangStandard::lang_c23; + else if (C17) + Std = LangStandard::lang_c17; + else if (C11) + Std = LangStandard::lang_c11; + else if (C99) + Std = LangStandard::lang_c99; + else if (!GNUMode && Digraphs) + Std = LangStandard::lang_c94; + else + return std::nullopt; + + return LangStandard::getLangStandardForKind(Std).getVersion(); +} diff --git a/clang/lib/Basic/LangStandards.cpp b/clang/lib/Basic/LangStandards.cpp index c49d095..01c524b 100644 --- a/clang/lib/Basic/LangStandards.cpp +++ b/clang/lib/Basic/LangStandards.cpp @@ -46,16 +46,18 @@ StringRef clang::languageToString(Language L) { llvm_unreachable("unhandled language kind"); } -#define LANGSTANDARD(id, name, lang, desc, features) \ - static const LangStandard Lang_##id = {name, desc, features, Language::lang}; +#define LANGSTANDARD(id, name, lang, desc, features, version) \ + static const LangStandard Lang_##id = {name, desc, features, Language::lang, \ + version}; #include "clang/Basic/LangStandards.def" const LangStandard &LangStandard::getLangStandardForKind(Kind K) { switch (K) { case lang_unspecified: llvm::report_fatal_error("getLangStandardForKind() on unspecified kind"); -#define LANGSTANDARD(id, name, lang, desc, features) \ - case lang_##id: return Lang_##id; +#define LANGSTANDARD(id, name, lang, desc, features, version) \ + case lang_##id: \ + return Lang_##id; #include "clang/Basic/LangStandards.def" } llvm_unreachable("Invalid language kind!"); @@ -63,7 +65,8 @@ const LangStandard &LangStandard::getLangStandardForKind(Kind K) { LangStandard::Kind LangStandard::getLangKind(StringRef Name) { return llvm::StringSwitch<Kind>(Name) -#define LANGSTANDARD(id, name, lang, desc, features) .Case(name, lang_##id) +#define LANGSTANDARD(id, name, lang, desc, features, version) \ + .Case(name, lang_##id) #define LANGSTANDARD_ALIAS(id, alias) .Case(alias, lang_##id) #include "clang/Basic/LangStandards.def" .Default(lang_unspecified); diff --git a/clang/lib/Basic/Targets/AArch64.cpp b/clang/lib/Basic/Targets/AArch64.cpp index 9e03a08..18641a9 100644 --- a/clang/lib/Basic/Targets/AArch64.cpp +++ b/clang/lib/Basic/Targets/AArch64.cpp @@ -1568,6 +1568,7 @@ bool AArch64TargetInfo::validateAsmConstraint( if (const unsigned Len = matchAsmCCConstraint(Name)) { Name += Len - 1; Info.setAllowsRegister(); + Info.setOutputOperandBounds(0, 2); return true; } } diff --git a/clang/lib/Basic/Targets/AMDGPU.cpp b/clang/lib/Basic/Targets/AMDGPU.cpp index 87de9e6..d4de704 100644 --- a/clang/lib/Basic/Targets/AMDGPU.cpp +++ b/clang/lib/Basic/Targets/AMDGPU.cpp @@ -27,11 +27,11 @@ namespace targets { // getPointerWidthV(). static const char *const DataLayoutStringR600 = - "e-p:32:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128" + "e-m:e-p:32:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128" "-v192:256-v256:256-v512:512-v1024:1024-v2048:2048-n32:64-S32-A5-G1"; static const char *const DataLayoutStringAMDGCN = - "e-p:64:64-p1:64:64-p2:32:32-p3:32:32-p4:64:64-p5:32:32-p6:32:32" + "e-m:e-p:64:64-p1:64:64-p2:32:32-p3:32:32-p4:64:64-p5:32:32-p6:32:32" "-p7:160:256:256:32-p8:128:128:128:48-p9:192:256:256:32-i64:64-" "v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-" "v2048:2048-n32:64-S32-A5-G1-ni:7:8:9"; diff --git a/clang/lib/Basic/Targets/AVR.cpp b/clang/lib/Basic/Targets/AVR.cpp index bbe7b01..2673669 100644 --- a/clang/lib/Basic/Targets/AVR.cpp +++ b/clang/lib/Basic/Targets/AVR.cpp @@ -420,23 +420,23 @@ static MCUInfo AVRMcus[] = { static bool ArchHasELPM(StringRef Arch) { return llvm::StringSwitch<bool>(Arch) - .Cases("31", "51", "6", true) - .Cases("102", "104", "105", "106", "107", true) - .Default(false); + .Cases({"31", "51", "6"}, true) + .Cases({"102", "104", "105", "106", "107"}, true) + .Default(false); } static bool ArchHasELPMX(StringRef Arch) { return llvm::StringSwitch<bool>(Arch) - .Cases("51", "6", true) - .Cases("102", "104", "105", "106", "107", true) - .Default(false); + .Cases({"51", "6"}, true) + .Cases({"102", "104", "105", "106", "107"}, true) + .Default(false); } static bool ArchHasMOVW(StringRef Arch) { return llvm::StringSwitch<bool>(Arch) - .Cases("25", "35", "4", "5", "51", "6", true) - .Cases("102", "103", "104", "105", "106", "107", true) - .Default(false); + .Cases({"25", "35", "4", "5", "51", "6"}, true) + .Cases({"102", "103", "104", "105", "106", "107"}, true) + .Default(false); } static bool ArchHasLPMX(StringRef Arch) { @@ -445,16 +445,16 @@ static bool ArchHasLPMX(StringRef Arch) { static bool ArchHasMUL(StringRef Arch) { return llvm::StringSwitch<bool>(Arch) - .Cases("4", "5", "51", "6", true) - .Cases("102", "103", "104", "105", "106", "107", true) - .Default(false); + .Cases({"4", "5", "51", "6"}, true) + .Cases({"102", "103", "104", "105", "106", "107"}, true) + .Default(false); } static bool ArchHasJMPCALL(StringRef Arch) { return llvm::StringSwitch<bool>(Arch) - .Cases("3", "31", "35", "5", "51", "6", true) - .Cases("102", "103", "104", "105", "106", "107", true) - .Default(false); + .Cases({"3", "31", "35", "5", "51", "6"}, true) + .Cases({"102", "103", "104", "105", "106", "107"}, true) + .Default(false); } static bool ArchHas3BytePC(StringRef Arch) { diff --git a/clang/lib/Basic/Targets/DirectX.h b/clang/lib/Basic/Targets/DirectX.h index bd13c9e..a21a593 100644 --- a/clang/lib/Basic/Targets/DirectX.h +++ b/clang/lib/Basic/Targets/DirectX.h @@ -64,8 +64,11 @@ public: NoAsmVariants = true; PlatformMinVersion = Triple.getOSVersion(); PlatformName = llvm::Triple::getOSTypeName(Triple.getOS()); + // TODO: We need to align vectors on the element size generally, but for now + // we hard code this for 3-element 32- and 64-bit vectors as a workaround. + // See https://github.com/llvm/llvm-project/issues/123968 resetDataLayout("e-m:e-p:32:32-i1:32-i8:8-i16:16-i32:32-i64:64-f16:16-f32:" - "32-f64:64-n8:16:32:64"); + "32-f64:64-n8:16:32:64-v48:16:16-v96:32:32-v192:64:64"); TheCXXABI.set(TargetCXXABI::GenericItanium); } bool useFP16ConversionIntrinsics() const override { return false; } diff --git a/clang/lib/Basic/Targets/Mips.cpp b/clang/lib/Basic/Targets/Mips.cpp index 34837cc..de6ccff 100644 --- a/clang/lib/Basic/Targets/Mips.cpp +++ b/clang/lib/Basic/Targets/Mips.cpp @@ -72,7 +72,7 @@ unsigned MipsTargetInfo::getISARev() const { .Cases("mips32r2", "mips64r2", "octeon", "octeon+", 2) .Cases("mips32r3", "mips64r3", 3) .Cases("mips32r5", "mips64r5", "p5600", 5) - .Cases("mips32r6", "mips64r6", 6) + .Cases("mips32r6", "mips64r6", "i6400", "i6500", 6) .Default(0); } @@ -270,8 +270,9 @@ bool MipsTargetInfo::validateTarget(DiagnosticsEngine &Diags) const { return false; } // Mips revision 6 and -mfp32 are incompatible - if (FPMode != FP64 && FPMode != FPXX && (CPU == "mips32r6" || - CPU == "mips64r6")) { + if (FPMode != FP64 && FPMode != FPXX && + (CPU == "mips32r6" || CPU == "mips64r6" || CPU == "i6400" || + CPU == "i6500")) { Diags.Report(diag::err_opt_not_valid_with_opt) << "-mfp32" << CPU; return false; } diff --git a/clang/lib/Basic/Targets/Mips.h b/clang/lib/Basic/Targets/Mips.h index e199df3..930271c 100644 --- a/clang/lib/Basic/Targets/Mips.h +++ b/clang/lib/Basic/Targets/Mips.h @@ -83,7 +83,8 @@ public: } bool isIEEE754_2008Default() const { - return CPU == "mips32r6" || CPU == "mips64r6"; + return CPU == "mips32r6" || CPU == "mips64r6" || CPU == "i6400" || + CPU == "i6500"; } enum FPModeEnum getDefaultFPMode() const { diff --git a/clang/lib/Basic/Targets/SystemZ.cpp b/clang/lib/Basic/Targets/SystemZ.cpp index 13b8623..30f846c 100644 --- a/clang/lib/Basic/Targets/SystemZ.cpp +++ b/clang/lib/Basic/Targets/SystemZ.cpp @@ -99,6 +99,16 @@ bool SystemZTargetInfo::validateAsmConstraint( case 'T': // Likewise, plus an index Info.setAllowsMemory(); return true; + case '@': + // CC condition changes. + if (StringRef(Name) == "@cc") { + Name += 2; + Info.setAllowsRegister(); + // SystemZ has 2-bits CC, and hence Interval [0, 4). + Info.setOutputOperandBounds(0, 4); + return true; + } + return false; } } @@ -161,6 +171,9 @@ unsigned SystemZTargetInfo::getMinGlobalAlign(uint64_t Size, void SystemZTargetInfo::getTargetDefines(const LangOptions &Opts, MacroBuilder &Builder) const { + // Inline assembly supports SystemZ flag outputs. + Builder.defineMacro("__GCC_ASM_FLAG_OUTPUTS__"); + Builder.defineMacro("__s390__"); Builder.defineMacro("__s390x__"); Builder.defineMacro("__zarch__"); diff --git a/clang/lib/Basic/Targets/SystemZ.h b/clang/lib/Basic/Targets/SystemZ.h index dc2185e..4e15d5a 100644 --- a/clang/lib/Basic/Targets/SystemZ.h +++ b/clang/lib/Basic/Targets/SystemZ.h @@ -136,6 +136,12 @@ public: std::string convertConstraint(const char *&Constraint) const override { switch (Constraint[0]) { + case '@': // Flag output operand. + if (llvm::StringRef(Constraint) == "@cc") { + Constraint += 2; + return std::string("{@cc}"); + } + break; case 'p': // Keep 'p' constraint. return std::string("p"); case 'Z': diff --git a/clang/lib/Basic/Targets/X86.cpp b/clang/lib/Basic/Targets/X86.cpp index 6eb4db5..e71f10c 100644 --- a/clang/lib/Basic/Targets/X86.cpp +++ b/clang/lib/Basic/Targets/X86.cpp @@ -625,6 +625,8 @@ void X86TargetInfo::getTargetDefines(const LangOptions &Opts, case CK_ArrowlakeS: case CK_Lunarlake: case CK_Pantherlake: + case CK_Wildcatlake: + case CK_Novalake: case CK_Sierraforest: case CK_Grandridge: case CK_Graniterapids: @@ -1516,6 +1518,7 @@ bool X86TargetInfo::validateAsmConstraint( if (auto Len = matchAsmCCConstraint(Name)) { Name += Len - 1; Info.setAllowsRegister(); + Info.setOutputOperandBounds(0, 2); return true; } return false; @@ -1612,6 +1615,8 @@ std::optional<unsigned> X86TargetInfo::getCPUCacheLineSize() const { case CK_ArrowlakeS: case CK_Lunarlake: case CK_Pantherlake: + case CK_Wildcatlake: + case CK_Novalake: case CK_Sierraforest: case CK_Grandridge: case CK_Graniterapids: diff --git a/clang/lib/CIR/CodeGen/Address.h b/clang/lib/CIR/CodeGen/Address.h index fb74aa0..a67cbad 100644 --- a/clang/lib/CIR/CodeGen/Address.h +++ b/clang/lib/CIR/CodeGen/Address.h @@ -17,6 +17,7 @@ #include "mlir/IR/Value.h" #include "clang/AST/CharUnits.h" #include "clang/CIR/Dialect/IR/CIRTypes.h" +#include "clang/CIR/MissingFeatures.h" #include "llvm/ADT/PointerIntPair.h" namespace clang::CIRGen { @@ -90,6 +91,13 @@ public: return getPointer(); } + /// Return the pointer contained in this class after authenticating it and + /// adding offset to it if necessary. + mlir::Value emitRawPointer() const { + assert(!cir::MissingFeatures::addressPointerAuthInfo()); + return getBasePointer(); + } + mlir::Type getType() const { assert(mlir::cast<cir::PointerType>( pointerAndKnownNonNull.getPointer().getType()) diff --git a/clang/lib/CIR/CodeGen/CIRGenAtomic.cpp b/clang/lib/CIR/CodeGen/CIRGenAtomic.cpp index 0f4d6d2..a9983f8 100644 --- a/clang/lib/CIR/CodeGen/CIRGenAtomic.cpp +++ b/clang/lib/CIR/CodeGen/CIRGenAtomic.cpp @@ -255,7 +255,7 @@ static void emitAtomicCmpXchg(CIRGenFunction &cgf, AtomicExpr *e, bool isWeak, mlir::Value expected = builder.createLoad(loc, val1); mlir::Value desired = builder.createLoad(loc, val2); - auto cmpxchg = cir::AtomicCmpXchg::create( + auto cmpxchg = cir::AtomicCmpXchgOp::create( builder, loc, expected.getType(), builder.getBoolTy(), ptr.getPointer(), expected, desired, cir::MemOrderAttr::get(&cgf.getMLIRContext(), successOrder), @@ -404,7 +404,7 @@ static void emitAtomicOp(CIRGenFunction &cgf, AtomicExpr *expr, Address dest, case AtomicExpr::AO__c11_atomic_exchange: case AtomicExpr::AO__atomic_exchange_n: case AtomicExpr::AO__atomic_exchange: - opName = cir::AtomicXchg::getOperationName(); + opName = cir::AtomicXchgOp::getOperationName(); break; case AtomicExpr::AO__opencl_atomic_init: diff --git a/clang/lib/CIR/CodeGen/CIRGenBuilder.h b/clang/lib/CIR/CodeGen/CIRGenBuilder.h index a6f10e6..50d585d 100644 --- a/clang/lib/CIR/CodeGen/CIRGenBuilder.h +++ b/clang/lib/CIR/CodeGen/CIRGenBuilder.h @@ -380,6 +380,16 @@ public: /*relative_layout=*/false); } + mlir::Value createDynCastToVoid(mlir::Location loc, mlir::Value src, + bool vtableUseRelativeLayout) { + // TODO(cir): consider address space here. + assert(!cir::MissingFeatures::addressSpace()); + cir::PointerType destTy = getVoidPtrTy(); + return cir::DynamicCastOp::create( + *this, loc, destTy, cir::DynamicCastKind::Ptr, src, + cir::DynamicCastInfoAttr{}, vtableUseRelativeLayout); + } + Address createBaseClassAddr(mlir::Location loc, Address addr, mlir::Type destType, unsigned offset, bool assumeNotNull) { @@ -519,6 +529,14 @@ public: return createGlobal(module, loc, uniqueName, type, isConstant, linkage); } + cir::StackSaveOp createStackSave(mlir::Location loc, mlir::Type ty) { + return cir::StackSaveOp::create(*this, loc, ty); + } + + cir::StackRestoreOp createStackRestore(mlir::Location loc, mlir::Value v) { + return cir::StackRestoreOp::create(*this, loc, v); + } + mlir::Value createSetBitfield(mlir::Location loc, mlir::Type resultType, Address dstAddr, mlir::Type storageType, mlir::Value src, const CIRGenBitFieldInfo &info, diff --git a/clang/lib/CIR/CodeGen/CIRGenBuiltin.cpp b/clang/lib/CIR/CodeGen/CIRGenBuiltin.cpp index 4cfa91e..ea31871 100644 --- a/clang/lib/CIR/CodeGen/CIRGenBuiltin.cpp +++ b/clang/lib/CIR/CodeGen/CIRGenBuiltin.cpp @@ -463,7 +463,9 @@ RValue CIRGenFunction::emitBuiltinExpr(const GlobalDecl &gd, unsigned builtinID, return emitLibraryCall(*this, fd, e, cgm.getBuiltinLibFunction(fd, builtinID)); - cgm.errorNYI(e->getSourceRange(), "unimplemented builtin call"); + cgm.errorNYI(e->getSourceRange(), + std::string("unimplemented builtin call: ") + + getContext().BuiltinInfo.getName(builtinID)); return getUndefRValue(e->getType()); } diff --git a/clang/lib/CIR/CodeGen/CIRGenCXX.cpp b/clang/lib/CIR/CodeGen/CIRGenCXX.cpp index 274d11b..171ce1c 100644 --- a/clang/lib/CIR/CodeGen/CIRGenCXX.cpp +++ b/clang/lib/CIR/CodeGen/CIRGenCXX.cpp @@ -171,7 +171,8 @@ cir::FuncOp CIRGenModule::codegenCXXStructor(GlobalDecl gd) { curCGF = nullptr; setNonAliasAttributes(gd, fn); - assert(!cir::MissingFeatures::opFuncAttributesForDefinition()); + setCIRFunctionAttributesForDefinition(mlir::cast<FunctionDecl>(gd.getDecl()), + fn); return fn; } diff --git a/clang/lib/CIR/CodeGen/CIRGenCXXABI.h b/clang/lib/CIR/CodeGen/CIRGenCXXABI.h index 06f41cd..6d3741c4 100644 --- a/clang/lib/CIR/CodeGen/CIRGenCXXABI.h +++ b/clang/lib/CIR/CodeGen/CIRGenCXXABI.h @@ -191,6 +191,15 @@ public: virtual void emitVTableDefinitions(CIRGenVTables &cgvt, const CXXRecordDecl *rd) = 0; + using DeleteOrMemberCallExpr = + llvm::PointerUnion<const CXXDeleteExpr *, const CXXMemberCallExpr *>; + + virtual mlir::Value emitVirtualDestructorCall(CIRGenFunction &cgf, + const CXXDestructorDecl *dtor, + CXXDtorType dtorType, + Address thisAddr, + DeleteOrMemberCallExpr e) = 0; + /// Emit any tables needed to implement virtual inheritance. For Itanium, /// this emits virtual table tables. virtual void emitVirtualInheritanceTables(const CXXRecordDecl *rd) = 0; diff --git a/clang/lib/CIR/CodeGen/CIRGenClass.cpp b/clang/lib/CIR/CodeGen/CIRGenClass.cpp index 485b2c8..89f4926 100644 --- a/clang/lib/CIR/CodeGen/CIRGenClass.cpp +++ b/clang/lib/CIR/CodeGen/CIRGenClass.cpp @@ -478,8 +478,7 @@ void CIRGenFunction::getVTablePointers(BaseSubobject base, for (const auto &nextBase : rd->bases()) { const auto *baseDecl = - cast<CXXRecordDecl>( - nextBase.getType()->castAs<RecordType>()->getOriginalDecl()) + cast<CXXRecordDecl>(nextBase.getType()->castAs<RecordType>()->getDecl()) ->getDefinitionOrSelf(); // Ignore classes without a vtable. @@ -895,6 +894,26 @@ void CIRGenFunction::destroyCXXObject(CIRGenFunction &cgf, Address addr, } namespace { +mlir::Value loadThisForDtorDelete(CIRGenFunction &cgf, + const CXXDestructorDecl *dd) { + if (Expr *thisArg = dd->getOperatorDeleteThisArg()) + return cgf.emitScalarExpr(thisArg); + return cgf.loadCXXThis(); +} + +/// Call the operator delete associated with the current destructor. +struct CallDtorDelete final : EHScopeStack::Cleanup { + CallDtorDelete() {} + + void emit(CIRGenFunction &cgf) override { + const CXXDestructorDecl *dtor = cast<CXXDestructorDecl>(cgf.curFuncDecl); + const CXXRecordDecl *classDecl = dtor->getParent(); + cgf.emitDeleteCall(dtor->getOperatorDelete(), + loadThisForDtorDelete(cgf, dtor), + cgf.getContext().getCanonicalTagType(classDecl)); + } +}; + class DestroyField final : public EHScopeStack::Cleanup { const FieldDecl *field; CIRGenFunction::Destroyer *destroyer; @@ -932,7 +951,18 @@ void CIRGenFunction::enterDtorCleanups(const CXXDestructorDecl *dd, // The deleting-destructor phase just needs to call the appropriate // operator delete that Sema picked up. if (dtorType == Dtor_Deleting) { - cgm.errorNYI(dd->getSourceRange(), "deleting destructor cleanups"); + assert(dd->getOperatorDelete() && + "operator delete missing - EnterDtorCleanups"); + if (cxxStructorImplicitParamValue) { + cgm.errorNYI(dd->getSourceRange(), "deleting destructor with vtt"); + } else { + if (dd->getOperatorDelete()->isDestroyingOperatorDelete()) { + cgm.errorNYI(dd->getSourceRange(), + "deleting destructor with destroying operator delete"); + } else { + ehStack.pushCleanup<CallDtorDelete>(NormalAndEHCleanup); + } + } return; } @@ -994,7 +1024,7 @@ void CIRGenFunction::enterDtorCleanups(const CXXDestructorDecl *dd, // Anonymous union members do not have their destructors called. const RecordType *rt = type->getAsUnionType(); - if (rt && rt->getOriginalDecl()->isAnonymousStructOrUnion()) + if (rt && rt->getDecl()->isAnonymousStructOrUnion()) continue; CleanupKind cleanupKind = getCleanupKind(dtorKind); diff --git a/clang/lib/CIR/CodeGen/CIRGenDecl.cpp b/clang/lib/CIR/CodeGen/CIRGenDecl.cpp index 039d290..4a19d91 100644 --- a/clang/lib/CIR/CodeGen/CIRGenDecl.cpp +++ b/clang/lib/CIR/CodeGen/CIRGenDecl.cpp @@ -44,38 +44,70 @@ CIRGenFunction::emitAutoVarAlloca(const VarDecl &d, // If the type is variably-modified, emit all the VLA sizes for it. if (ty->isVariablyModifiedType()) - cgm.errorNYI(d.getSourceRange(), "emitAutoVarDecl: variably modified type"); + emitVariablyModifiedType(ty); assert(!cir::MissingFeatures::openMP()); Address address = Address::invalid(); - if (!ty->isConstantSizeType()) - cgm.errorNYI(d.getSourceRange(), "emitAutoVarDecl: non-constant size type"); - - // A normal fixed sized variable becomes an alloca in the entry block, - // unless: - // - it's an NRVO variable. - // - we are compiling OpenMP and it's an OpenMP local variable. - if (nrvo) { - // The named return value optimization: allocate this variable in the - // return slot, so that we can elide the copy when returning this - // variable (C++0x [class.copy]p34). - address = returnValue; - - if (const RecordDecl *rd = ty->getAsRecordDecl()) { - if (const auto *cxxrd = dyn_cast<CXXRecordDecl>(rd); - (cxxrd && !cxxrd->hasTrivialDestructor()) || - rd->isNonTrivialToPrimitiveDestroy()) - cgm.errorNYI(d.getSourceRange(), "emitAutoVarAlloca: set NRVO flag"); + if (ty->isConstantSizeType()) { + // A normal fixed sized variable becomes an alloca in the entry block, + // unless: + // - it's an NRVO variable. + // - we are compiling OpenMP and it's an OpenMP local variable. + if (nrvo) { + // The named return value optimization: allocate this variable in the + // return slot, so that we can elide the copy when returning this + // variable (C++0x [class.copy]p34). + address = returnValue; + + if (const RecordDecl *rd = ty->getAsRecordDecl()) { + if (const auto *cxxrd = dyn_cast<CXXRecordDecl>(rd); + (cxxrd && !cxxrd->hasTrivialDestructor()) || + rd->isNonTrivialToPrimitiveDestroy()) + cgm.errorNYI(d.getSourceRange(), "emitAutoVarAlloca: set NRVO flag"); + } + } else { + // A normal fixed sized variable becomes an alloca in the entry block, + mlir::Type allocaTy = convertTypeForMem(ty); + // Create the temp alloca and declare variable using it. + address = createTempAlloca(allocaTy, alignment, loc, d.getName(), + /*arraySize=*/nullptr, /*alloca=*/nullptr, ip); + declare(address.getPointer(), &d, ty, getLoc(d.getSourceRange()), + alignment); } } else { - // A normal fixed sized variable becomes an alloca in the entry block, - mlir::Type allocaTy = convertTypeForMem(ty); - // Create the temp alloca and declare variable using it. - address = createTempAlloca(allocaTy, alignment, loc, d.getName(), - /*arraySize=*/nullptr, /*alloca=*/nullptr, ip); - declare(address.getPointer(), &d, ty, getLoc(d.getSourceRange()), - alignment); + // Non-constant size type + assert(!cir::MissingFeatures::openMP()); + if (!didCallStackSave) { + // Save the stack. + cir::PointerType defaultTy = AllocaInt8PtrTy; + CharUnits align = CharUnits::fromQuantity( + cgm.getDataLayout().getAlignment(defaultTy, false)); + Address stack = createTempAlloca(defaultTy, align, loc, "saved_stack"); + + mlir::Value v = builder.createStackSave(loc, defaultTy); + assert(v.getType() == AllocaInt8PtrTy); + builder.createStore(loc, v, stack); + + didCallStackSave = true; + + // Push a cleanup block and restore the stack there. + // FIXME: in general circumstances, this should be an EH cleanup. + pushStackRestore(NormalCleanup, stack); + } + + VlaSizePair vlaSize = getVLASize(ty); + mlir::Type memTy = convertTypeForMem(vlaSize.type); + + // Allocate memory for the array. + address = + createTempAlloca(memTy, alignment, loc, d.getName(), vlaSize.numElts, + /*alloca=*/nullptr, builder.saveInsertionPoint()); + + // If we have debug info enabled, properly describe the VLA dimensions for + // this type by registering the vla size expression for each of the + // dimensions. + assert(!cir::MissingFeatures::generateDebugInfo()); } emission.addr = address; @@ -696,6 +728,16 @@ struct DestroyObject final : EHScopeStack::Cleanup { cgf.emitDestroy(addr, type, destroyer); } }; + +struct CallStackRestore final : EHScopeStack::Cleanup { + Address stack; + CallStackRestore(Address stack) : stack(stack) {} + void emit(CIRGenFunction &cgf) override { + mlir::Location loc = stack.getPointer().getLoc(); + mlir::Value v = cgf.getBuilder().createLoad(loc, stack); + cgf.getBuilder().createStackRestore(loc, v); + } +}; } // namespace void CIRGenFunction::pushDestroy(CleanupKind cleanupKind, Address addr, @@ -805,6 +847,10 @@ CIRGenFunction::getDestroyer(QualType::DestructionKind kind) { llvm_unreachable("Unknown DestructionKind"); } +void CIRGenFunction::pushStackRestore(CleanupKind kind, Address spMem) { + ehStack.pushCleanup<CallStackRestore>(kind, spMem); +} + /// Enter a destroy cleanup for the given local variable. void CIRGenFunction::emitAutoVarTypeCleanup( const CIRGenFunction::AutoVarEmission &emission, diff --git a/clang/lib/CIR/CodeGen/CIRGenExpr.cpp b/clang/lib/CIR/CodeGen/CIRGenExpr.cpp index f416571..4897c29 100644 --- a/clang/lib/CIR/CodeGen/CIRGenExpr.cpp +++ b/clang/lib/CIR/CodeGen/CIRGenExpr.cpp @@ -2068,7 +2068,7 @@ mlir::Value CIRGenFunction::emitAlloca(StringRef name, mlir::Type ty, mlir::OpBuilder::InsertionGuard guard(builder); builder.restoreInsertionPoint(ip); addr = builder.createAlloca(loc, /*addr type*/ localVarPtrTy, - /*var type*/ ty, name, alignIntAttr); + /*var type*/ ty, name, alignIntAttr, arraySize); assert(!cir::MissingFeatures::astVarDeclInterface()); } return addr; diff --git a/clang/lib/CIR/CodeGen/CIRGenExprCXX.cpp b/clang/lib/CIR/CodeGen/CIRGenExprCXX.cpp index 97c0944..b1e9e76 100644 --- a/clang/lib/CIR/CodeGen/CIRGenExprCXX.cpp +++ b/clang/lib/CIR/CodeGen/CIRGenExprCXX.cpp @@ -130,13 +130,11 @@ RValue CIRGenFunction::emitCXXMemberOrOperatorMemberCallExpr( const CXXMethodDecl *calleeDecl = devirtualizedMethod ? devirtualizedMethod : md; const CIRGenFunctionInfo *fInfo = nullptr; - if (isa<CXXDestructorDecl>(calleeDecl)) { - cgm.errorNYI(ce->getSourceRange(), - "emitCXXMemberOrOperatorMemberCallExpr: destructor call"); - return RValue::get(nullptr); - } - - fInfo = &cgm.getTypes().arrangeCXXMethodDeclaration(calleeDecl); + if (const auto *dtor = dyn_cast<CXXDestructorDecl>(calleeDecl)) + fInfo = &cgm.getTypes().arrangeCXXStructorDeclaration( + GlobalDecl(dtor, Dtor_Complete)); + else + fInfo = &cgm.getTypes().arrangeCXXMethodDeclaration(calleeDecl); cir::FuncType ty = cgm.getTypes().getFunctionType(*fInfo); @@ -151,9 +149,34 @@ RValue CIRGenFunction::emitCXXMemberOrOperatorMemberCallExpr( // because then we know what the type is. bool useVirtualCall = canUseVirtualCall && !devirtualizedMethod; - if (isa<CXXDestructorDecl>(calleeDecl)) { - cgm.errorNYI(ce->getSourceRange(), - "emitCXXMemberOrOperatorMemberCallExpr: destructor call"); + if (const auto *dtor = dyn_cast<CXXDestructorDecl>(calleeDecl)) { + assert(ce->arg_begin() == ce->arg_end() && + "Destructor shouldn't have explicit parameters"); + assert(returnValue.isNull() && "Destructor shouldn't have return value"); + if (useVirtualCall) { + cgm.getCXXABI().emitVirtualDestructorCall(*this, dtor, Dtor_Complete, + thisPtr.getAddress(), + cast<CXXMemberCallExpr>(ce)); + } else { + GlobalDecl globalDecl(dtor, Dtor_Complete); + CIRGenCallee callee; + assert(!cir::MissingFeatures::appleKext()); + if (!devirtualizedMethod) { + callee = CIRGenCallee::forDirect( + cgm.getAddrOfCXXStructor(globalDecl, fInfo, ty), globalDecl); + } else { + cgm.errorNYI(ce->getSourceRange(), "devirtualized destructor call"); + return RValue::get(nullptr); + } + + QualType thisTy = + isArrow ? base->getType()->getPointeeType() : base->getType(); + // CIRGen does not pass CallOrInvoke here (different from OG LLVM codegen) + // because in practice it always null even in OG. + emitCXXDestructorCall(globalDecl, callee, thisPtr.getPointer(), thisTy, + /*implicitParam=*/nullptr, + /*implicitParamTy=*/QualType(), ce); + } return RValue::get(nullptr); } diff --git a/clang/lib/CIR/CodeGen/CIRGenExprConstant.cpp b/clang/lib/CIR/CodeGen/CIRGenExprConstant.cpp index 89e9ec4..19ed656 100644 --- a/clang/lib/CIR/CodeGen/CIRGenExprConstant.cpp +++ b/clang/lib/CIR/CodeGen/CIRGenExprConstant.cpp @@ -614,7 +614,7 @@ bool ConstRecordBuilder::applyZeroInitPadding(const ASTRecordLayout &layout, bool ConstRecordBuilder::build(InitListExpr *ile, bool allowOverwrite) { RecordDecl *rd = ile->getType() ->castAs<clang::RecordType>() - ->getOriginalDecl() + ->getDecl() ->getDefinitionOrSelf(); const ASTRecordLayout &layout = cgm.getASTContext().getASTRecordLayout(rd); @@ -817,9 +817,8 @@ bool ConstRecordBuilder::build(const APValue &val, const RecordDecl *rd, mlir::Attribute ConstRecordBuilder::finalize(QualType type) { type = type.getNonReferenceType(); - RecordDecl *rd = type->castAs<clang::RecordType>() - ->getOriginalDecl() - ->getDefinitionOrSelf(); + RecordDecl *rd = + type->castAs<clang::RecordType>()->getDecl()->getDefinitionOrSelf(); mlir::Type valTy = cgm.convertType(type); return builder.build(valTy, rd->hasFlexibleArrayMember()); } @@ -842,9 +841,8 @@ mlir::Attribute ConstRecordBuilder::buildRecord(ConstantEmitter &emitter, ConstantAggregateBuilder constant(emitter.cgm); ConstRecordBuilder builder(emitter, constant, CharUnits::Zero()); - const RecordDecl *rd = valTy->castAs<clang::RecordType>() - ->getOriginalDecl() - ->getDefinitionOrSelf(); + const RecordDecl *rd = + valTy->castAs<clang::RecordType>()->getDecl()->getDefinitionOrSelf(); const CXXRecordDecl *cd = dyn_cast<CXXRecordDecl>(rd); if (!builder.build(val, rd, false, cd, CharUnits::Zero())) return nullptr; @@ -873,7 +871,7 @@ bool ConstRecordBuilder::updateRecord(ConstantEmitter &emitter, class ConstExprEmitter : public StmtVisitor<ConstExprEmitter, mlir::Attribute, QualType> { CIRGenModule &cgm; - LLVM_ATTRIBUTE_UNUSED ConstantEmitter &emitter; + [[maybe_unused]] ConstantEmitter &emitter; public: ConstExprEmitter(ConstantEmitter &emitter) diff --git a/clang/lib/CIR/CodeGen/CIRGenExprScalar.cpp b/clang/lib/CIR/CodeGen/CIRGenExprScalar.cpp index 637f9ef..138082b 100644 --- a/clang/lib/CIR/CodeGen/CIRGenExprScalar.cpp +++ b/clang/lib/CIR/CodeGen/CIRGenExprScalar.cpp @@ -1734,9 +1734,9 @@ mlir::Value ScalarExprEmitter::emitSub(const BinOpInfo &ops) { // LLVM we shall take VLA's, division by element size, etc. // // See more in `EmitSub` in CGExprScalar.cpp. - assert(!cir::MissingFeatures::ptrDiffOp()); - cgf.cgm.errorNYI("ptrdiff"); - return {}; + assert(!cir::MissingFeatures::llvmLoweringPtrDiffConsidersPointee()); + return cir::PtrDiffOp::create(builder, cgf.getLoc(ops.loc), cgf.PtrDiffTy, + ops.lhs, ops.rhs); } mlir::Value ScalarExprEmitter::emitShl(const BinOpInfo &ops) { diff --git a/clang/lib/CIR/CodeGen/CIRGenFunction.cpp b/clang/lib/CIR/CodeGen/CIRGenFunction.cpp index 7a774e0..ba36cbe 100644 --- a/clang/lib/CIR/CodeGen/CIRGenFunction.cpp +++ b/clang/lib/CIR/CodeGen/CIRGenFunction.cpp @@ -410,6 +410,8 @@ void CIRGenFunction::startFunction(GlobalDecl gd, QualType returnType, curFn = fn; const Decl *d = gd.getDecl(); + + didCallStackSave = false; curCodeDecl = d; const auto *fd = dyn_cast_or_null<FunctionDecl>(d); curFuncDecl = d->getNonClosureContext(); @@ -678,7 +680,13 @@ void CIRGenFunction::emitDestructorBody(FunctionArgList &args) { // possible to delegate the destructor body to the complete // destructor. Do so. if (dtorType == Dtor_Deleting) { - cgm.errorNYI(dtor->getSourceRange(), "deleting destructor"); + RunCleanupsScope dtorEpilogue(*this); + enterDtorCleanups(dtor, Dtor_Deleting); + if (haveInsertPoint()) { + QualType thisTy = dtor->getFunctionObjectParameterType(); + emitCXXDestructorCall(dtor, Dtor_Complete, /*forVirtualBase=*/false, + /*delegating=*/false, loadCXXThisAddress(), thisTy); + } return; } @@ -1000,6 +1008,41 @@ mlir::Value CIRGenFunction::emitAlignmentAssumption( offsetValue); } +CIRGenFunction::VlaSizePair CIRGenFunction::getVLASize(QualType type) { + const VariableArrayType *vla = + cgm.getASTContext().getAsVariableArrayType(type); + assert(vla && "type was not a variable array type!"); + return getVLASize(vla); +} + +CIRGenFunction::VlaSizePair +CIRGenFunction::getVLASize(const VariableArrayType *type) { + // The number of elements so far; always size_t. + mlir::Value numElements; + + QualType elementType; + do { + elementType = type->getElementType(); + mlir::Value vlaSize = vlaSizeMap[type->getSizeExpr()]; + assert(vlaSize && "no size for VLA!"); + assert(vlaSize.getType() == SizeTy); + + if (!numElements) { + numElements = vlaSize; + } else { + // It's undefined behavior if this wraps around, so mark it that way. + // FIXME: Teach -fsanitize=undefined to trap this. + + numElements = + builder.createMul(numElements.getLoc(), numElements, vlaSize, + cir::OverflowBehavior::NoUnsignedWrap); + } + } while ((type = getContext().getAsVariableArrayType(elementType))); + + assert(numElements && "Undefined elements number"); + return {numElements, elementType}; +} + // TODO(cir): Most of this function can be shared between CIRGen // and traditional LLVM codegen void CIRGenFunction::emitVariablyModifiedType(QualType type) { @@ -1080,7 +1123,26 @@ void CIRGenFunction::emitVariablyModifiedType(QualType type) { break; case Type::VariableArray: { - cgm.errorNYI("CIRGenFunction::emitVariablyModifiedType VLA"); + // Losing element qualification here is fine. + const VariableArrayType *vat = cast<clang::VariableArrayType>(ty); + + // Unknown size indication requires no size computation. + // Otherwise, evaluate and record it. + if (const Expr *sizeExpr = vat->getSizeExpr()) { + // It's possible that we might have emitted this already, + // e.g. with a typedef and a pointer to it. + mlir::Value &entry = vlaSizeMap[sizeExpr]; + if (!entry) { + mlir::Value size = emitScalarExpr(sizeExpr); + assert(!cir::MissingFeatures::sanitizers()); + + // Always zexting here would be wrong if it weren't + // undefined behavior to have a negative bound. + // FIXME: What about when size's type is larger than size_t? + entry = builder.createIntCast(size, SizeTy); + } + } + type = vat->getElementType(); break; } diff --git a/clang/lib/CIR/CodeGen/CIRGenFunction.h b/clang/lib/CIR/CodeGen/CIRGenFunction.h index d71de2f..3c36f5c 100644 --- a/clang/lib/CIR/CodeGen/CIRGenFunction.h +++ b/clang/lib/CIR/CodeGen/CIRGenFunction.h @@ -149,6 +149,10 @@ public: using SymTableTy = llvm::ScopedHashTable<const clang::Decl *, mlir::Value>; SymTableTy symbolTable; + /// Whether a cir.stacksave operation has been added. Used to avoid + /// inserting cir.stacksave for multiple VLAs in the same scope. + bool didCallStackSave = false; + /// Whether or not a Microsoft-style asm block has been processed within /// this fuction. These can potentially set the return value. bool sawAsmBlock = false; @@ -188,6 +192,14 @@ public: llvm::DenseMap<const OpaqueValueExpr *, LValue> opaqueLValues; llvm::DenseMap<const OpaqueValueExpr *, RValue> opaqueRValues; + // This keeps track of the associated size for each VLA type. + // We track this by the size expression rather than the type itself because + // in certain situations, like a const qualifier applied to an VLA typedef, + // multiple VLA types can share the same size expression. + // FIXME: Maybe this could be a stack of maps that is pushed/popped as we + // enter/leave scopes. + llvm::DenseMap<const Expr *, mlir::Value> vlaSizeMap; + public: /// A non-RAII class containing all the information about a bound /// opaque value. OpaqueValueMapping, below, is a RAII wrapper for @@ -436,6 +448,20 @@ public: } }; + struct VlaSizePair { + mlir::Value numElts; + QualType type; + + VlaSizePair(mlir::Value num, QualType ty) : numElts(num), type(ty) {} + }; + + /// Returns an MLIR::Value+QualType pair that corresponds to the size, + /// in non-variably-sized elements, of a variable length array type, + /// plus that largest non-variably-sized element type. Assumes that + /// the type has already been emitted with emitVariablyModifiedType. + VlaSizePair getVLASize(const VariableArrayType *type); + VlaSizePair getVLASize(QualType type); + void finishFunction(SourceLocation endLoc); /// Determine whether the given initializer is trivial in the sense @@ -583,6 +609,8 @@ public: return needsEHCleanup(kind) ? NormalAndEHCleanup : NormalCleanup; } + void pushStackRestore(CleanupKind kind, Address spMem); + /// Set the address of a local variable. void setAddrOfLocalVar(const clang::VarDecl *vd, Address addr) { assert(!localDeclMap.count(vd) && "Decl already exists in LocalDeclMap!"); @@ -854,6 +882,7 @@ public: protected: bool performCleanup; + bool oldDidCallStackSave; private: RunCleanupsScope(const RunCleanupsScope &) = delete; @@ -867,6 +896,8 @@ public: explicit RunCleanupsScope(CIRGenFunction &cgf) : performCleanup(true), cgf(cgf) { cleanupStackDepth = cgf.ehStack.stable_begin(); + oldDidCallStackSave = cgf.didCallStackSave; + cgf.didCallStackSave = false; oldCleanupStackDepth = cgf.currentCleanupStackDepth; cgf.currentCleanupStackDepth = cleanupStackDepth; } @@ -883,6 +914,7 @@ public: assert(performCleanup && "Already forced cleanup"); { mlir::OpBuilder::InsertionGuard guard(cgf.getBuilder()); + cgf.didCallStackSave = oldDidCallStackSave; cgf.popCleanupBlocks(cleanupStackDepth); performCleanup = false; cgf.currentCleanupStackDepth = oldCleanupStackDepth; @@ -1281,10 +1313,10 @@ public: mlir::Value emitCXXNewExpr(const CXXNewExpr *e); - void emitNewArrayInitializer(const CXXNewExpr *E, QualType ElementType, - mlir::Type ElementTy, Address BeginPtr, - mlir::Value NumElements, - mlir::Value AllocSizeWithoutCookie); + void emitNewArrayInitializer(const CXXNewExpr *e, QualType elementType, + mlir::Type elementTy, Address beginPtr, + mlir::Value numElements, + mlir::Value allocSizeWithoutCookie); RValue emitCXXOperatorMemberCallExpr(const CXXOperatorCallExpr *e, const CXXMethodDecl *md, diff --git a/clang/lib/CIR/CodeGen/CIRGenItaniumCXXABI.cpp b/clang/lib/CIR/CodeGen/CIRGenItaniumCXXABI.cpp index 9e490c6d..c184d4a 100644 --- a/clang/lib/CIR/CodeGen/CIRGenItaniumCXXABI.cpp +++ b/clang/lib/CIR/CodeGen/CIRGenItaniumCXXABI.cpp @@ -95,7 +95,10 @@ public: clang::GlobalDecl gd, Address thisAddr, mlir::Type ty, SourceLocation loc) override; - + mlir::Value emitVirtualDestructorCall(CIRGenFunction &cgf, + const CXXDestructorDecl *dtor, + CXXDtorType dtorType, Address thisAddr, + DeleteOrMemberCallExpr e) override; mlir::Value getVTableAddressPoint(BaseSubobject base, const CXXRecordDecl *vtableClass) override; mlir::Value getVTableAddressPointInStructorWithVTT( @@ -465,6 +468,29 @@ void CIRGenItaniumCXXABI::emitVTableDefinitions(CIRGenVTables &cgvt, } } +mlir::Value CIRGenItaniumCXXABI::emitVirtualDestructorCall( + CIRGenFunction &cgf, const CXXDestructorDecl *dtor, CXXDtorType dtorType, + Address thisAddr, DeleteOrMemberCallExpr expr) { + auto *callExpr = dyn_cast<const CXXMemberCallExpr *>(expr); + auto *delExpr = dyn_cast<const CXXDeleteExpr *>(expr); + assert((callExpr != nullptr) ^ (delExpr != nullptr)); + assert(callExpr == nullptr || callExpr->arg_begin() == callExpr->arg_end()); + assert(dtorType == Dtor_Deleting || dtorType == Dtor_Complete); + + GlobalDecl globalDecl(dtor, dtorType); + const CIRGenFunctionInfo *fnInfo = + &cgm.getTypes().arrangeCXXStructorDeclaration(globalDecl); + const cir::FuncType &fnTy = cgm.getTypes().getFunctionType(*fnInfo); + auto callee = CIRGenCallee::forVirtual(callExpr, globalDecl, thisAddr, fnTy); + + QualType thisTy = + callExpr ? callExpr->getObjectType() : delExpr->getDestroyedType(); + + cgf.emitCXXDestructorCall(globalDecl, callee, thisAddr.emitRawPointer(), + thisTy, nullptr, QualType(), nullptr); + return nullptr; +} + void CIRGenItaniumCXXABI::emitVirtualInheritanceTables( const CXXRecordDecl *rd) { CIRGenVTables &vtables = cgm.getVTables(); @@ -718,8 +744,8 @@ static bool shouldUseExternalRttiDescriptor(CIRGenModule &cgm, QualType ty) { return false; if (const auto *recordTy = dyn_cast<RecordType>(ty)) { - const CXXRecordDecl *rd = - cast<CXXRecordDecl>(recordTy->getOriginalDecl())->getDefinitionOrSelf(); + const auto *rd = + cast<CXXRecordDecl>(recordTy->getDecl())->getDefinitionOrSelf(); if (!rd->hasDefinition()) return false; @@ -833,9 +859,7 @@ static bool canUseSingleInheritance(const CXXRecordDecl *rd) { /// IsIncompleteClassType - Returns whether the given record type is incomplete. static bool isIncompleteClassType(const RecordType *recordTy) { - return !recordTy->getOriginalDecl() - ->getDefinitionOrSelf() - ->isCompleteDefinition(); + return !recordTy->getDecl()->getDefinitionOrSelf()->isCompleteDefinition(); } /// Returns whether the given type contains an @@ -913,8 +937,7 @@ const char *vTableClassNameForType(const CIRGenModule &cgm, const Type *ty) { case Type::Atomic: // FIXME: GCC treats block pointers as fundamental types?! case Type::BlockPointer: - cgm.errorNYI("VTableClassNameForType: __fundamental_type_info"); - break; + return "_ZTVN10__cxxabiv123__fundamental_type_infoE"; case Type::ConstantArray: case Type::IncompleteArray: case Type::VariableArray: @@ -927,13 +950,11 @@ const char *vTableClassNameForType(const CIRGenModule &cgm, const Type *ty) { break; case Type::Enum: - cgm.errorNYI("VTableClassNameForType: Enum"); - break; + return "_ZTVN10__cxxabiv116__enum_type_infoE"; case Type::Record: { - const CXXRecordDecl *rd = - cast<CXXRecordDecl>(cast<RecordType>(ty)->getOriginalDecl()) - ->getDefinitionOrSelf(); + const auto *rd = cast<CXXRecordDecl>(cast<RecordType>(ty)->getDecl()) + ->getDefinitionOrSelf(); if (!rd->hasDefinition() || !rd->getNumBases()) { return classTypeInfo; @@ -1005,8 +1026,8 @@ static cir::GlobalLinkageKind getTypeInfoLinkage(CIRGenModule &cgm, return cir::GlobalLinkageKind::LinkOnceODRLinkage; if (const RecordType *record = dyn_cast<RecordType>(ty)) { - const CXXRecordDecl *rd = - cast<CXXRecordDecl>(record->getOriginalDecl())->getDefinitionOrSelf(); + const auto *rd = + cast<CXXRecordDecl>(record->getDecl())->getDefinitionOrSelf(); if (rd->hasAttr<WeakAttr>()) return cir::GlobalLinkageKind::WeakODRLinkage; @@ -1356,9 +1377,8 @@ mlir::Attribute CIRGenItaniumRTTIBuilder::buildTypeInfo( break; case Type::Record: { - const auto *rd = - cast<CXXRecordDecl>(cast<RecordType>(ty)->getOriginalDecl()) - ->getDefinitionOrSelf(); + const auto *rd = cast<CXXRecordDecl>(cast<RecordType>(ty)->getDecl()) + ->getDefinitionOrSelf(); if (!rd->hasDefinition() || !rd->getNumBases()) { // We don't need to emit any fields. break; @@ -1625,8 +1645,7 @@ void CIRGenItaniumCXXABI::emitThrow(CIRGenFunction &cgf, // Lowering pass to skip passing the trivial function. // if (const RecordType *recordTy = clangThrowType->getAs<RecordType>()) { - CXXRecordDecl *rec = - cast<CXXRecordDecl>(recordTy->getOriginalDecl()->getDefinition()); + auto *rec = cast<CXXRecordDecl>(recordTy->getDecl()->getDefinition()); assert(!cir::MissingFeatures::isTrivialCtorOrDtor()); if (!rec->hasTrivialDestructor()) { cgm.errorNYI("emitThrow: non-trivial destructor"); @@ -1925,6 +1944,15 @@ static cir::FuncOp getItaniumDynamicCastFn(CIRGenFunction &cgf) { return cgf.cgm.createRuntimeFunction(FTy, "__dynamic_cast"); } +static Address emitDynamicCastToVoid(CIRGenFunction &cgf, mlir::Location loc, + QualType srcRecordTy, Address src) { + bool vtableUsesRelativeLayout = + cgf.cgm.getItaniumVTableContext().isRelativeLayout(); + mlir::Value ptr = cgf.getBuilder().createDynCastToVoid( + loc, src.getPointer(), vtableUsesRelativeLayout); + return Address{ptr, src.getAlignment()}; +} + static cir::DynamicCastInfoAttr emitDynamicCastInfo(CIRGenFunction &cgf, mlir::Location loc, QualType srcRecordTy, @@ -1959,10 +1987,8 @@ mlir::Value CIRGenItaniumCXXABI::emitDynamicCast(CIRGenFunction &cgf, bool isCastToVoid = destRecordTy.isNull(); assert((!isCastToVoid || !isRefCast) && "cannot cast to void reference"); - if (isCastToVoid) { - cgm.errorNYI(loc, "emitDynamicCastToVoid"); - return {}; - } + if (isCastToVoid) + return emitDynamicCastToVoid(cgf, loc, srcRecordTy, src).getPointer(); // If the destination is effectively final, the cast succeeds if and only // if the dynamic type of the pointer is exactly the destination type. diff --git a/clang/lib/CIR/CodeGen/CIRGenModule.cpp b/clang/lib/CIR/CodeGen/CIRGenModule.cpp index fe1ea56..127f763 100644 --- a/clang/lib/CIR/CodeGen/CIRGenModule.cpp +++ b/clang/lib/CIR/CodeGen/CIRGenModule.cpp @@ -88,6 +88,8 @@ CIRGenModule::CIRGenModule(mlir::MLIRContext &mlirContext, FP80Ty = cir::FP80Type::get(&getMLIRContext()); FP128Ty = cir::FP128Type::get(&getMLIRContext()); + AllocaInt8PtrTy = cir::PointerType::get(UInt8Ty, cirAllocaAddressSpace); + PointerAlignInBytes = astContext .toCharUnitsFromBits( @@ -449,17 +451,49 @@ void CIRGenModule::emitGlobalFunctionDefinition(clang::GlobalDecl gd, curCGF = nullptr; setNonAliasAttributes(gd, funcOp); - assert(!cir::MissingFeatures::opFuncAttributesForDefinition()); + setCIRFunctionAttributesForDefinition(funcDecl, funcOp); + + auto getPriority = [this](const auto *attr) -> int { + Expr *e = attr->getPriority(); + if (e) + return e->EvaluateKnownConstInt(this->getASTContext()).getExtValue(); + return attr->DefaultPriority; + }; - if (funcDecl->getAttr<ConstructorAttr>()) - errorNYI(funcDecl->getSourceRange(), "constructor attribute"); - if (funcDecl->getAttr<DestructorAttr>()) - errorNYI(funcDecl->getSourceRange(), "destructor attribute"); + if (const ConstructorAttr *ca = funcDecl->getAttr<ConstructorAttr>()) + addGlobalCtor(funcOp, getPriority(ca)); + if (const DestructorAttr *da = funcDecl->getAttr<DestructorAttr>()) + addGlobalDtor(funcOp, getPriority(da)); if (funcDecl->getAttr<AnnotateAttr>()) errorNYI(funcDecl->getSourceRange(), "deferredAnnotations"); } +/// Track functions to be called before main() runs. +void CIRGenModule::addGlobalCtor(cir::FuncOp ctor, + std::optional<int> priority) { + assert(!cir::MissingFeatures::globalCtorLexOrder()); + assert(!cir::MissingFeatures::globalCtorAssociatedData()); + + // Traditional LLVM codegen directly adds the function to the list of global + // ctors. In CIR we just add a global_ctor attribute to the function. The + // global list is created in LoweringPrepare. + // + // FIXME(from traditional LLVM): Type coercion of void()* types. + ctor.setGlobalCtorPriority(priority); +} + +/// Add a function to the list that will be called when the module is unloaded. +void CIRGenModule::addGlobalDtor(cir::FuncOp dtor, + std::optional<int> priority) { + if (codeGenOpts.RegisterGlobalDtorsWithAtExit && + (!getASTContext().getTargetInfo().getTriple().isOSAIX())) + errorNYI(dtor.getLoc(), "registerGlobalDtorsWithAtExit"); + + // FIXME(from traditional LLVM): Type coercion of void()* types. + dtor.setGlobalDtorPriority(priority); +} + void CIRGenModule::handleCXXStaticMemberVarInstantiation(VarDecl *vd) { VarDecl::DefinitionKind dk = vd->isThisDeclarationADefinition(); if (dk == VarDecl::Definition && vd->hasAttr<DLLImportAttr>()) @@ -1885,6 +1919,91 @@ void CIRGenModule::setFunctionAttributes(GlobalDecl globalDecl, } } +void CIRGenModule::setCIRFunctionAttributesForDefinition( + const clang::FunctionDecl *decl, cir::FuncOp f) { + assert(!cir::MissingFeatures::opFuncUnwindTablesAttr()); + assert(!cir::MissingFeatures::stackProtector()); + + std::optional<cir::InlineKind> existingInlineKind = f.getInlineKind(); + bool isNoInline = + existingInlineKind && *existingInlineKind == cir::InlineKind::NoInline; + bool isAlwaysInline = existingInlineKind && + *existingInlineKind == cir::InlineKind::AlwaysInline; + + if (!decl) { + assert(!cir::MissingFeatures::hlsl()); + + if (!isAlwaysInline && + codeGenOpts.getInlining() == CodeGenOptions::OnlyAlwaysInlining) { + // If inlining is disabled and we don't have a declaration to control + // inlining, mark the function as 'noinline' unless it is explicitly + // marked as 'alwaysinline'. + f.setInlineKindAttr( + cir::InlineAttr::get(&getMLIRContext(), cir::InlineKind::NoInline)); + } + + return; + } + + assert(!cir::MissingFeatures::opFuncArmStreamingAttr()); + assert(!cir::MissingFeatures::opFuncArmNewAttr()); + assert(!cir::MissingFeatures::opFuncOptNoneAttr()); + assert(!cir::MissingFeatures::opFuncMinSizeAttr()); + assert(!cir::MissingFeatures::opFuncNakedAttr()); + assert(!cir::MissingFeatures::opFuncNoDuplicateAttr()); + assert(!cir::MissingFeatures::hlsl()); + + // Handle inline attributes + if (decl->hasAttr<NoInlineAttr>() && !isAlwaysInline) { + // Add noinline if the function isn't always_inline. + f.setInlineKindAttr( + cir::InlineAttr::get(&getMLIRContext(), cir::InlineKind::NoInline)); + } else if (decl->hasAttr<AlwaysInlineAttr>() && !isNoInline) { + // Don't override AlwaysInline with NoInline, or vice versa, since we can't + // specify both in IR. + f.setInlineKindAttr( + cir::InlineAttr::get(&getMLIRContext(), cir::InlineKind::AlwaysInline)); + } else if (codeGenOpts.getInlining() == CodeGenOptions::OnlyAlwaysInlining) { + // If inlining is disabled, force everything that isn't always_inline + // to carry an explicit noinline attribute. + if (!isAlwaysInline) { + f.setInlineKindAttr( + cir::InlineAttr::get(&getMLIRContext(), cir::InlineKind::NoInline)); + } + } else { + // Otherwise, propagate the inline hint attribute and potentially use its + // absence to mark things as noinline. + // Search function and template pattern redeclarations for inline. + if (auto *fd = dyn_cast<FunctionDecl>(decl)) { + // TODO: Share this checkForInline implementation with classic codegen. + // This logic is likely to change over time, so sharing would help ensure + // consistency. + auto checkForInline = [](const FunctionDecl *decl) { + auto checkRedeclForInline = [](const FunctionDecl *redecl) { + return redecl->isInlineSpecified(); + }; + if (any_of(decl->redecls(), checkRedeclForInline)) + return true; + const FunctionDecl *pattern = decl->getTemplateInstantiationPattern(); + if (!pattern) + return false; + return any_of(pattern->redecls(), checkRedeclForInline); + }; + if (checkForInline(fd)) { + f.setInlineKindAttr(cir::InlineAttr::get(&getMLIRContext(), + cir::InlineKind::InlineHint)); + } else if (codeGenOpts.getInlining() == + CodeGenOptions::OnlyHintInlining && + !fd->isInlined() && !isAlwaysInline) { + f.setInlineKindAttr( + cir::InlineAttr::get(&getMLIRContext(), cir::InlineKind::NoInline)); + } + } + } + + assert(!cir::MissingFeatures::opFuncColdHotAttr()); +} + cir::FuncOp CIRGenModule::getOrCreateCIRFunction( StringRef mangledName, mlir::Type funcType, GlobalDecl gd, bool forVTable, bool dontDefer, bool isThunk, ForDefinition_t isForDefinition, diff --git a/clang/lib/CIR/CodeGen/CIRGenModule.h b/clang/lib/CIR/CodeGen/CIRGenModule.h index f627bae..1fc116d 100644 --- a/clang/lib/CIR/CodeGen/CIRGenModule.h +++ b/clang/lib/CIR/CodeGen/CIRGenModule.h @@ -159,6 +159,13 @@ public: bool isConstant = false, mlir::Operation *insertPoint = nullptr); + /// Add a global constructor or destructor to the module. + /// The priority is optional, if not specified, the default priority is used. + void addGlobalCtor(cir::FuncOp ctor, + std::optional<int> priority = std::nullopt); + void addGlobalDtor(cir::FuncOp dtor, + std::optional<int> priority = std::nullopt); + bool shouldZeroInitPadding() const { // In C23 (N3096) $6.7.10: // """ @@ -422,6 +429,10 @@ public: void setFunctionAttributes(GlobalDecl gd, cir::FuncOp f, bool isIncompleteFunction, bool isThunk); + /// Set extra attributes (inline, etc.) for a function. + void setCIRFunctionAttributesForDefinition(const clang::FunctionDecl *fd, + cir::FuncOp f); + void emitGlobalDefinition(clang::GlobalDecl gd, mlir::Operation *op = nullptr); void emitGlobalFunctionDefinition(clang::GlobalDecl gd, mlir::Operation *op); diff --git a/clang/lib/CIR/CodeGen/CIRGenOpenACCClause.cpp b/clang/lib/CIR/CodeGen/CIRGenOpenACCClause.cpp index 3d86f71..ce4ae7e 100644 --- a/clang/lib/CIR/CodeGen/CIRGenOpenACCClause.cpp +++ b/clang/lib/CIR/CodeGen/CIRGenOpenACCClause.cpp @@ -1005,7 +1005,7 @@ public: /*temporary=*/nullptr, OpenACCReductionOperator::Invalid, Decl::castToDeclContext(cgf.curFuncDecl), opInfo.origType, opInfo.bounds.size(), opInfo.boundTypes, opInfo.baseType, - privateOp); + privateOp, /*reductionCombinerRecipes=*/{}); // TODO: OpenACC: The dialect is going to change in the near future to // have these be on a different operation, so when that changes, we // probably need to change these here. @@ -1046,7 +1046,7 @@ public: OpenACCReductionOperator::Invalid, Decl::castToDeclContext(cgf.curFuncDecl), opInfo.origType, opInfo.bounds.size(), opInfo.boundTypes, opInfo.baseType, - firstPrivateOp); + firstPrivateOp, /*reductionCombinerRecipe=*/{}); // TODO: OpenACC: The dialect is going to change in the near future to // have these be on a different operation, so when that changes, we @@ -1088,7 +1088,7 @@ public: /*temporary=*/nullptr, clause.getReductionOp(), Decl::castToDeclContext(cgf.curFuncDecl), opInfo.origType, opInfo.bounds.size(), opInfo.boundTypes, opInfo.baseType, - reductionOp); + reductionOp, varRecipe.CombinerRecipes); operation.addReduction(builder.getContext(), reductionOp, recipe); } diff --git a/clang/lib/CIR/CodeGen/CIRGenOpenACCRecipe.cpp b/clang/lib/CIR/CodeGen/CIRGenOpenACCRecipe.cpp index 24a5fc2..f638d39 100644 --- a/clang/lib/CIR/CodeGen/CIRGenOpenACCRecipe.cpp +++ b/clang/lib/CIR/CodeGen/CIRGenOpenACCRecipe.cpp @@ -398,6 +398,7 @@ void OpenACCRecipeBuilderBase::createRecipeDestroySection( emitDestroy(block->getArgument(1), elementTy); } + ls.forceCleanup(); mlir::acc::YieldOp::create(builder, locEnd); } void OpenACCRecipeBuilderBase::makeBoundsInit( @@ -480,6 +481,7 @@ void OpenACCRecipeBuilderBase::createInitRecipe( /*isInitSection=*/true); } + ls.forceCleanup(); mlir::acc::YieldOp::create(builder, locEnd); } @@ -518,6 +520,7 @@ void OpenACCRecipeBuilderBase::createFirstprivateRecipeCopy( cgf.emitAutoVarInit(tempDeclEmission); builder.setInsertionPointToEnd(©Region.back()); + ls.forceCleanup(); mlir::acc::YieldOp::create(builder, locEnd); } @@ -527,16 +530,143 @@ void OpenACCRecipeBuilderBase::createFirstprivateRecipeCopy( // doesn't restore it aftewards. void OpenACCRecipeBuilderBase::createReductionRecipeCombiner( mlir::Location loc, mlir::Location locEnd, mlir::Value mainOp, - mlir::acc::ReductionRecipeOp recipe, size_t numBounds) { + mlir::acc::ReductionRecipeOp recipe, size_t numBounds, QualType origType, + llvm::ArrayRef<OpenACCReductionRecipe::CombinerRecipe> combinerRecipes) { mlir::Block *block = createRecipeBlock(recipe.getCombinerRegion(), mainOp.getType(), loc, numBounds, /*isInit=*/false); builder.setInsertionPointToEnd(&recipe.getCombinerRegion().back()); CIRGenFunction::LexicalScope ls(cgf, loc, block); - mlir::BlockArgument lhsArg = block->getArgument(0); + mlir::Value lhsArg = block->getArgument(0); + mlir::Value rhsArg = block->getArgument(1); + llvm::MutableArrayRef<mlir::BlockArgument> boundsRange = + block->getArguments().drop_front(2); + + if (llvm::any_of(combinerRecipes, [](auto &r) { return r.Op == nullptr; })) { + cgf.cgm.errorNYI(loc, "OpenACC Reduction combiner not generated"); + mlir::acc::YieldOp::create(builder, locEnd, block->getArgument(0)); + return; + } + + // apply the bounds so that we can get our bounds emitted correctly. + for (mlir::BlockArgument boundArg : llvm::reverse(boundsRange)) + std::tie(lhsArg, rhsArg) = + createBoundsLoop(lhsArg, rhsArg, boundArg, loc, /*inverse=*/false); + + // Emitter for when we know this isn't a struct or array we have to loop + // through. This should work for the 'field' once the get-element call has + // been made. + auto emitSingleCombiner = + [&](mlir::Value lhsArg, mlir::Value rhsArg, + const OpenACCReductionRecipe::CombinerRecipe &combiner) { + mlir::Type elementTy = + mlir::cast<cir::PointerType>(lhsArg.getType()).getPointee(); + CIRGenFunction::DeclMapRevertingRAII declMapRAIILhs{cgf, combiner.LHS}; + cgf.setAddrOfLocalVar( + combiner.LHS, Address{lhsArg, elementTy, + cgf.getContext().getDeclAlign(combiner.LHS)}); + CIRGenFunction::DeclMapRevertingRAII declMapRAIIRhs{cgf, combiner.RHS}; + cgf.setAddrOfLocalVar( + combiner.RHS, Address{rhsArg, elementTy, + cgf.getContext().getDeclAlign(combiner.RHS)}); + + [[maybe_unused]] mlir::LogicalResult stmtRes = + cgf.emitStmt(combiner.Op, /*useCurrentScope=*/true); + }; + + // Emitter for when we know this is either a non-array or element of an array + // (which also shouldn't be an array type?). This function should generate the + // initialization code for an entire 'array-element'/non-array, including + // diving into each element of a struct (if necessary). + auto emitCombiner = [&](mlir::Value lhsArg, mlir::Value rhsArg, QualType ty) { + assert(!ty->isArrayType() && "Array type shouldn't get here"); + if (const auto *rd = ty->getAsRecordDecl()) { + if (combinerRecipes.size() == 1 && + cgf.getContext().hasSameType(ty, combinerRecipes[0].LHS->getType())) { + // If this is a 'top level' operator on the type we can just emit this + // as a simple one. + emitSingleCombiner(lhsArg, rhsArg, combinerRecipes[0]); + } else { + // else we have to handle each individual field after after a + // get-element. + for (const auto &[field, combiner] : + llvm::zip_equal(rd->fields(), combinerRecipes)) { + mlir::Type fieldType = cgf.convertType(field->getType()); + auto fieldPtr = cir::PointerType::get(fieldType); + + mlir::Value lhsField = builder.createGetMember( + loc, fieldPtr, lhsArg, field->getName(), field->getFieldIndex()); + mlir::Value rhsField = builder.createGetMember( + loc, fieldPtr, rhsArg, field->getName(), field->getFieldIndex()); + + emitSingleCombiner(lhsField, rhsField, combiner); + } + } + + } else { + // if this is a single-thing (because we should know this isn't an array, + // as Sema wouldn't let us get here), we can just do a normal emit call. + emitSingleCombiner(lhsArg, rhsArg, combinerRecipes[0]); + } + }; + + if (const auto *cat = cgf.getContext().getAsConstantArrayType(origType)) { + // If we're in an array, we have to emit the combiner for each element of + // the array. + auto itrTy = mlir::cast<cir::IntType>(cgf.PtrDiffTy); + auto itrPtrTy = cir::PointerType::get(itrTy); + + mlir::Value zero = + builder.getConstInt(loc, mlir::cast<cir::IntType>(cgf.PtrDiffTy), 0); + mlir::Value itr = + cir::AllocaOp::create(builder, loc, itrPtrTy, itrTy, "itr", + cgf.cgm.getSize(cgf.getPointerAlign())); + builder.CIRBaseBuilderTy::createStore(loc, zero, itr); + + builder.setInsertionPointAfter(builder.createFor( + loc, + /*condBuilder=*/ + [&](mlir::OpBuilder &b, mlir::Location loc) { + auto loadItr = cir::LoadOp::create(builder, loc, {itr}); + mlir::Value arraySize = builder.getConstInt( + loc, mlir::cast<cir::IntType>(cgf.PtrDiffTy), cat->getZExtSize()); + auto cmp = builder.createCompare(loc, cir::CmpOpKind::lt, loadItr, + arraySize); + builder.createCondition(cmp); + }, + /*bodyBuilder=*/ + [&](mlir::OpBuilder &b, mlir::Location loc) { + auto loadItr = cir::LoadOp::create(builder, loc, {itr}); + auto lhsElt = builder.getArrayElement( + loc, loc, lhsArg, cgf.convertType(cat->getElementType()), loadItr, + /*shouldDecay=*/true); + auto rhsElt = builder.getArrayElement( + loc, loc, rhsArg, cgf.convertType(cat->getElementType()), loadItr, + /*shouldDecay=*/true); + + emitCombiner(lhsElt, rhsElt, cat->getElementType()); + builder.createYield(loc); + }, + /*stepBuilder=*/ + [&](mlir::OpBuilder &b, mlir::Location loc) { + auto loadItr = cir::LoadOp::create(builder, loc, {itr}); + auto inc = cir::UnaryOp::create(builder, loc, loadItr.getType(), + cir::UnaryOpKind::Inc, loadItr); + builder.CIRBaseBuilderTy::createStore(loc, inc, itr); + builder.createYield(loc); + })); - mlir::acc::YieldOp::create(builder, locEnd, lhsArg); + } else if (origType->isArrayType()) { + cgf.cgm.errorNYI(loc, + "OpenACC Reduction combiner non-constant array recipe"); + } else { + emitCombiner(lhsArg, rhsArg, origType); + } + + builder.setInsertionPointToEnd(&recipe.getCombinerRegion().back()); + ls.forceCleanup(); + mlir::acc::YieldOp::create(builder, locEnd, block->getArgument(0)); } } // namespace clang::CIRGen diff --git a/clang/lib/CIR/CodeGen/CIRGenOpenACCRecipe.h b/clang/lib/CIR/CodeGen/CIRGenOpenACCRecipe.h index a5da744..745d424 100644 --- a/clang/lib/CIR/CodeGen/CIRGenOpenACCRecipe.h +++ b/clang/lib/CIR/CodeGen/CIRGenOpenACCRecipe.h @@ -64,10 +64,10 @@ protected: // that this function is not 'insertion point' clean, in that it alters the // insertion point to be inside of the 'combiner' section of the recipe, but // doesn't restore it aftewards. - void createReductionRecipeCombiner(mlir::Location loc, mlir::Location locEnd, - mlir::Value mainOp, - mlir::acc::ReductionRecipeOp recipe, - size_t numBounds); + void createReductionRecipeCombiner( + mlir::Location loc, mlir::Location locEnd, mlir::Value mainOp, + mlir::acc::ReductionRecipeOp recipe, size_t numBounds, QualType origType, + llvm::ArrayRef<OpenACCReductionRecipe::CombinerRecipe> combinerRecipes); void createInitRecipe(mlir::Location loc, mlir::Location locEnd, SourceRange exprRange, mlir::Value mainOp, @@ -169,7 +169,9 @@ public: const Expr *varRef, const VarDecl *varRecipe, const VarDecl *temporary, OpenACCReductionOperator reductionOp, DeclContext *dc, QualType origType, size_t numBounds, llvm::ArrayRef<QualType> boundTypes, QualType baseType, - mlir::Value mainOp) { + mlir::Value mainOp, + llvm::ArrayRef<OpenACCReductionRecipe::CombinerRecipe> + reductionCombinerRecipes) { assert(!varRecipe->getType()->isSpecificBuiltinType( BuiltinType::ArraySection) && "array section shouldn't make it to recipe creation"); @@ -208,7 +210,8 @@ public: createInitRecipe(loc, locEnd, varRef->getSourceRange(), mainOp, recipe.getInitRegion(), numBounds, boundTypes, varRecipe, origType, /*emitInitExpr=*/true); - createReductionRecipeCombiner(loc, locEnd, mainOp, recipe, numBounds); + createReductionRecipeCombiner(loc, locEnd, mainOp, recipe, numBounds, + origType, reductionCombinerRecipes); } else { static_assert(std::is_same_v<RecipeTy, mlir::acc::FirstprivateRecipeOp>); createInitRecipe(loc, locEnd, varRef->getSourceRange(), mainOp, diff --git a/clang/lib/CIR/CodeGen/CIRGenStmt.cpp b/clang/lib/CIR/CodeGen/CIRGenStmt.cpp index cfd48a2..5ba64dd 100644 --- a/clang/lib/CIR/CodeGen/CIRGenStmt.cpp +++ b/clang/lib/CIR/CodeGen/CIRGenStmt.cpp @@ -536,7 +536,7 @@ mlir::LogicalResult CIRGenFunction::emitLabel(const clang::LabelDecl &d) { mlir::Block *currBlock = builder.getBlock(); mlir::Block *labelBlock = currBlock; - if (!currBlock->empty()) { + if (!currBlock->empty() || currBlock->isEntryBlock()) { { mlir::OpBuilder::InsertionGuard guard(builder); labelBlock = builder.createBlock(builder.getBlock()->getParent()); diff --git a/clang/lib/CIR/CodeGen/CIRGenTypeCache.h b/clang/lib/CIR/CodeGen/CIRGenTypeCache.h index 273ec7f..b5612d9 100644 --- a/clang/lib/CIR/CodeGen/CIRGenTypeCache.h +++ b/clang/lib/CIR/CodeGen/CIRGenTypeCache.h @@ -65,6 +65,9 @@ struct CIRGenTypeCache { cir::PointerType VoidPtrTy; cir::PointerType UInt8PtrTy; + /// void* in alloca address space + cir::PointerType AllocaInt8PtrTy; + /// The size and alignment of a pointer into the generic address space. union { unsigned char PointerAlignInBytes; diff --git a/clang/lib/CIR/CodeGen/CIRGenTypes.cpp b/clang/lib/CIR/CodeGen/CIRGenTypes.cpp index e65896a..d1b91d0 100644 --- a/clang/lib/CIR/CodeGen/CIRGenTypes.cpp +++ b/clang/lib/CIR/CodeGen/CIRGenTypes.cpp @@ -159,7 +159,7 @@ isSafeToConvert(const RecordDecl *rd, CIRGenTypes &cgt, for (const clang::CXXBaseSpecifier &i : crd->bases()) if (!isSafeToConvert(i.getType() ->castAs<RecordType>() - ->getOriginalDecl() + ->getDecl() ->getDefinitionOrSelf(), cgt, alreadyChecked)) return false; @@ -279,8 +279,7 @@ mlir::Type CIRGenTypes::convertType(QualType type) { // Process record types before the type cache lookup. if (const auto *recordType = dyn_cast<RecordType>(type)) - return convertRecordDeclType( - recordType->getOriginalDecl()->getDefinitionOrSelf()); + return convertRecordDeclType(recordType->getDecl()->getDefinitionOrSelf()); // Has the type already been processed? TypeCacheTy::iterator tci = typeCache.find(ty); @@ -421,6 +420,16 @@ mlir::Type CIRGenTypes::convertType(QualType type) { break; } + case Type::VariableArray: { + const VariableArrayType *a = cast<VariableArrayType>(ty); + if (a->getIndexTypeCVRQualifiers() != 0) + cgm.errorNYI(SourceLocation(), "non trivial array types", type); + // VLAs resolve to the innermost element type; this matches + // the return of alloca, and there isn't any obviously better choice. + resultType = convertTypeForMem(a->getElementType()); + break; + } + case Type::IncompleteArray: { const IncompleteArrayType *arrTy = cast<IncompleteArrayType>(ty); if (arrTy->getIndexTypeCVRQualifiers() != 0) @@ -619,10 +628,8 @@ const CIRGenFunctionInfo &CIRGenTypes::arrangeGlobalDeclaration(GlobalDecl gd) { const auto *fd = cast<FunctionDecl>(gd.getDecl()); if (isa<CXXConstructorDecl>(gd.getDecl()) || - isa<CXXDestructorDecl>(gd.getDecl())) { - cgm.errorNYI(SourceLocation(), - "arrangeGlobalDeclaration for C++ constructor or destructor"); - } + isa<CXXDestructorDecl>(gd.getDecl())) + return arrangeCXXStructorDeclaration(gd); return arrangeFunctionDeclaration(fd); } diff --git a/clang/lib/CIR/CodeGen/CIRGenVTables.cpp b/clang/lib/CIR/CodeGen/CIRGenVTables.cpp index 84f5977..36bab62 100644 --- a/clang/lib/CIR/CodeGen/CIRGenVTables.cpp +++ b/clang/lib/CIR/CodeGen/CIRGenVTables.cpp @@ -120,12 +120,6 @@ mlir::Attribute CIRGenVTables::getVTableComponent( assert(!cir::MissingFeatures::vtableRelativeLayout()); switch (component.getKind()) { - case VTableComponent::CK_CompleteDtorPointer: - cgm.errorNYI("getVTableComponent: CompleteDtorPointer"); - return mlir::Attribute(); - case VTableComponent::CK_DeletingDtorPointer: - cgm.errorNYI("getVTableComponent: DeletingDtorPointer"); - return mlir::Attribute(); case VTableComponent::CK_UnusedFunctionPointer: cgm.errorNYI("getVTableComponent: UnusedFunctionPointer"); return mlir::Attribute(); @@ -148,7 +142,9 @@ mlir::Attribute CIRGenVTables::getVTableComponent( "expected GlobalViewAttr or ConstPtrAttr"); return rtti; - case VTableComponent::CK_FunctionPointer: { + case VTableComponent::CK_FunctionPointer: + case VTableComponent::CK_CompleteDtorPointer: + case VTableComponent::CK_DeletingDtorPointer: { GlobalDecl gd = component.getGlobalDecl(); assert(!cir::MissingFeatures::cudaSupport()); diff --git a/clang/lib/CIR/CodeGen/CIRGenValue.h b/clang/lib/CIR/CodeGen/CIRGenValue.h index 25b6ecb..08d9913 100644 --- a/clang/lib/CIR/CodeGen/CIRGenValue.h +++ b/clang/lib/CIR/CodeGen/CIRGenValue.h @@ -308,7 +308,7 @@ class AggValueSlot { /// destructor for the slot. Otherwise the code which constructs it should /// push the appropriate cleanup. LLVM_PREFERRED_TYPE(bool) - LLVM_ATTRIBUTE_UNUSED unsigned destructedFlag : 1; + [[maybe_unused]] unsigned destructedFlag : 1; /// This is set to true if the memory in the slot is known to be zero before /// the assignment into it. This means that zero fields don't need to be set. @@ -327,7 +327,7 @@ class AggValueSlot { /// object, it's important that this flag never be set when /// evaluating an expression which constructs such an object. LLVM_PREFERRED_TYPE(bool) - LLVM_ATTRIBUTE_UNUSED unsigned aliasedFlag : 1; + [[maybe_unused]] unsigned aliasedFlag : 1; /// This is set to true if the tail padding of this slot might overlap /// another object that may have already been initialized (and whose @@ -335,7 +335,7 @@ class AggValueSlot { /// store up to the dsize of the type. Otherwise we can widen stores to /// the size of the type. LLVM_PREFERRED_TYPE(bool) - LLVM_ATTRIBUTE_UNUSED unsigned overlapFlag : 1; + [[maybe_unused]] unsigned overlapFlag : 1; public: enum IsDestructed_t { IsNotDestructed, IsDestructed }; diff --git a/clang/lib/CIR/Dialect/IR/CIRDialect.cpp b/clang/lib/CIR/Dialect/IR/CIRDialect.cpp index 5f88590..b4c3704 100644 --- a/clang/lib/CIR/Dialect/IR/CIRDialect.cpp +++ b/clang/lib/CIR/Dialect/IR/CIRDialect.cpp @@ -15,6 +15,7 @@ #include "clang/CIR/Dialect/IR/CIROpsEnums.h" #include "clang/CIR/Dialect/IR/CIRTypes.h" +#include "mlir/IR/DialectImplementation.h" #include "mlir/Interfaces/ControlFlowInterfaces.h" #include "mlir/Interfaces/FunctionImplementation.h" #include "mlir/Support/LLVM.h" @@ -1720,6 +1721,73 @@ ParseResult cir::FuncOp::parse(OpAsmParser &parser, OperationState &state) { hasAlias = true; } + auto parseGlobalDtorCtor = + [&](StringRef keyword, + llvm::function_ref<void(std::optional<int> prio)> createAttr) + -> mlir::LogicalResult { + if (mlir::succeeded(parser.parseOptionalKeyword(keyword))) { + std::optional<int> priority; + if (mlir::succeeded(parser.parseOptionalLParen())) { + auto parsedPriority = mlir::FieldParser<int>::parse(parser); + if (mlir::failed(parsedPriority)) + return parser.emitError(parser.getCurrentLocation(), + "failed to parse 'priority', of type 'int'"); + priority = parsedPriority.value_or(int()); + // Parse literal ')' + if (parser.parseRParen()) + return failure(); + } + createAttr(priority); + } + return success(); + }; + + if (parseGlobalDtorCtor("global_ctor", [&](std::optional<int> priority) { + mlir::IntegerAttr globalCtorPriorityAttr = + builder.getI32IntegerAttr(priority.value_or(65535)); + state.addAttribute(getGlobalCtorPriorityAttrName(state.name), + globalCtorPriorityAttr); + }).failed()) + return failure(); + + if (parseGlobalDtorCtor("global_dtor", [&](std::optional<int> priority) { + mlir::IntegerAttr globalDtorPriorityAttr = + builder.getI32IntegerAttr(priority.value_or(65535)); + state.addAttribute(getGlobalDtorPriorityAttrName(state.name), + globalDtorPriorityAttr); + }).failed()) + return failure(); + + // Parse optional inline kind: inline(never|always|hint) + if (parser.parseOptionalKeyword("inline").succeeded()) { + if (parser.parseLParen().failed()) + return failure(); + + llvm::StringRef inlineKindStr; + const std::array<llvm::StringRef, cir::getMaxEnumValForInlineKind()> + allowedInlineKindStrs{ + cir::stringifyInlineKind(cir::InlineKind::NoInline), + cir::stringifyInlineKind(cir::InlineKind::AlwaysInline), + cir::stringifyInlineKind(cir::InlineKind::InlineHint), + }; + if (parser.parseOptionalKeyword(&inlineKindStr, allowedInlineKindStrs) + .failed()) + return parser.emitError(parser.getCurrentLocation(), + "expected 'never', 'always', or 'hint'"); + + std::optional<InlineKind> inlineKind = + cir::symbolizeInlineKind(inlineKindStr); + if (!inlineKind) + return parser.emitError(parser.getCurrentLocation(), + "invalid inline kind"); + + state.addAttribute(getInlineKindAttrName(state.name), + cir::InlineAttr::get(builder.getContext(), *inlineKind)); + + if (parser.parseRParen().failed()) + return failure(); + } + // Parse the optional function body. auto *body = state.addRegion(); OptionalParseResult parseResult = parser.parseOptionalRegion( @@ -1801,6 +1869,22 @@ void cir::FuncOp::print(OpAsmPrinter &p) { p << ")"; } + if (auto globalCtorPriority = getGlobalCtorPriority()) { + p << " global_ctor"; + if (globalCtorPriority.value() != 65535) + p << "(" << globalCtorPriority.value() << ")"; + } + + if (auto globalDtorPriority = getGlobalDtorPriority()) { + p << " global_dtor"; + if (globalDtorPriority.value() != 65535) + p << "(" << globalDtorPriority.value() << ")"; + } + + if (cir::InlineAttr inlineAttr = getInlineKindAttr()) { + p << " inline(" << cir::stringifyInlineKind(inlineAttr.getValue()) << ")"; + } + // Print the body if this is not an external function. Region &body = getOperation()->getRegion(0); if (!body.empty()) { @@ -2851,31 +2935,144 @@ mlir::LogicalResult cir::ThrowOp::verify() { } //===----------------------------------------------------------------------===// -// AtomicCmpXchg +// TypeInfoAttr //===----------------------------------------------------------------------===// -LogicalResult cir::AtomicCmpXchg::verify() { - mlir::Type pointeeType = getPtr().getType().getPointee(); +LogicalResult cir::TypeInfoAttr::verify( + ::llvm::function_ref<::mlir::InFlightDiagnostic()> emitError, + ::mlir::Type type, ::mlir::ArrayAttr typeInfoData) { - if (pointeeType != getExpected().getType() || - pointeeType != getDesired().getType()) - return emitOpError("ptr, expected and desired types must match"); + if (cir::ConstRecordAttr::verify(emitError, type, typeInfoData).failed()) + return failure(); return success(); } //===----------------------------------------------------------------------===// -// TypeInfoAttr +// TryOp //===----------------------------------------------------------------------===// -LogicalResult cir::TypeInfoAttr::verify( - ::llvm::function_ref<::mlir::InFlightDiagnostic()> emitError, - ::mlir::Type type, ::mlir::ArrayAttr typeInfoData) { +void cir::TryOp::getSuccessorRegions( + mlir::RegionBranchPoint point, + llvm::SmallVectorImpl<mlir::RegionSuccessor> ®ions) { + // The `try` and the `catchers` region branch back to the parent operation. + if (!point.isParent()) { + regions.push_back(mlir::RegionSuccessor()); + return; + } - if (cir::ConstRecordAttr::verify(emitError, type, typeInfoData).failed()) - return failure(); + regions.push_back(mlir::RegionSuccessor(&getTryRegion())); - return success(); + // TODO(CIR): If we know a target function never throws a specific type, we + // can remove the catch handler. + for (mlir::Region &handlerRegion : this->getHandlerRegions()) + regions.push_back(mlir::RegionSuccessor(&handlerRegion)); +} + +static void +printTryHandlerRegions(mlir::OpAsmPrinter &printer, cir::TryOp op, + mlir::MutableArrayRef<mlir::Region> handlerRegions, + mlir::ArrayAttr handlerTypes) { + if (!handlerTypes) + return; + + for (const auto [typeIdx, typeAttr] : llvm::enumerate(handlerTypes)) { + if (typeIdx) + printer << " "; + + if (mlir::isa<cir::CatchAllAttr>(typeAttr)) { + printer << "catch all "; + } else if (mlir::isa<cir::UnwindAttr>(typeAttr)) { + printer << "unwind "; + } else { + printer << "catch [type "; + printer.printAttribute(typeAttr); + printer << "] "; + } + + printer.printRegion(handlerRegions[typeIdx], + /*printEntryBLockArgs=*/false, + /*printBlockTerminators=*/true); + } +} + +static mlir::ParseResult parseTryHandlerRegions( + mlir::OpAsmParser &parser, + llvm::SmallVectorImpl<std::unique_ptr<mlir::Region>> &handlerRegions, + mlir::ArrayAttr &handlerTypes) { + + auto parseCheckedCatcherRegion = [&]() -> mlir::ParseResult { + handlerRegions.emplace_back(new mlir::Region); + + mlir::Region &currRegion = *handlerRegions.back(); + mlir::SMLoc regionLoc = parser.getCurrentLocation(); + if (parser.parseRegion(currRegion)) { + handlerRegions.clear(); + return failure(); + } + + if (currRegion.empty()) + return parser.emitError(regionLoc, "handler region shall not be empty"); + + if (!(currRegion.back().mightHaveTerminator() && + currRegion.back().getTerminator())) + return parser.emitError( + regionLoc, "blocks are expected to be explicitly terminated"); + + return success(); + }; + + bool hasCatchAll = false; + llvm::SmallVector<mlir::Attribute, 4> catcherAttrs; + while (parser.parseOptionalKeyword("catch").succeeded()) { + bool hasLSquare = parser.parseOptionalLSquare().succeeded(); + + llvm::StringRef attrStr; + if (parser.parseOptionalKeyword(&attrStr, {"all", "type"}).failed()) + return parser.emitError(parser.getCurrentLocation(), + "expected 'all' or 'type' keyword"); + + bool isCatchAll = attrStr == "all"; + if (isCatchAll) { + if (hasCatchAll) + return parser.emitError(parser.getCurrentLocation(), + "can't have more than one catch all"); + hasCatchAll = true; + } + + mlir::Attribute exceptionRTTIAttr; + if (!isCatchAll && parser.parseAttribute(exceptionRTTIAttr).failed()) + return parser.emitError(parser.getCurrentLocation(), + "expected valid RTTI info attribute"); + + catcherAttrs.push_back(isCatchAll + ? cir::CatchAllAttr::get(parser.getContext()) + : exceptionRTTIAttr); + + if (hasLSquare && isCatchAll) + return parser.emitError(parser.getCurrentLocation(), + "catch all dosen't need RTTI info attribute"); + + if (hasLSquare && parser.parseRSquare().failed()) + return parser.emitError(parser.getCurrentLocation(), + "expected `]` after RTTI info attribute"); + + if (parseCheckedCatcherRegion().failed()) + return mlir::failure(); + } + + if (parser.parseOptionalKeyword("unwind").succeeded()) { + if (hasCatchAll) + return parser.emitError(parser.getCurrentLocation(), + "unwind can't be used with catch all"); + + catcherAttrs.push_back(cir::UnwindAttr::get(parser.getContext())); + if (parseCheckedCatcherRegion().failed()) + return mlir::failure(); + } + + handlerTypes = parser.getBuilder().getArrayAttr(catcherAttrs); + return mlir::success(); } //===----------------------------------------------------------------------===// diff --git a/clang/lib/CIR/Dialect/Transforms/LoweringPrepare.cpp b/clang/lib/CIR/Dialect/Transforms/LoweringPrepare.cpp index dbff0b9..d99c362 100644 --- a/clang/lib/CIR/Dialect/Transforms/LoweringPrepare.cpp +++ b/clang/lib/CIR/Dialect/Transforms/LoweringPrepare.cpp @@ -105,6 +105,8 @@ struct LoweringPreparePass : public LoweringPrepareBase<LoweringPreparePass> { /// List of ctors and their priorities to be called before main() llvm::SmallVector<std::pair<std::string, uint32_t>, 4> globalCtorList; + /// List of dtors and their priorities to be called when unloading module. + llvm::SmallVector<std::pair<std::string, uint32_t>, 4> globalDtorList; void setASTContext(clang::ASTContext *c) { astCtx = c; @@ -823,10 +825,13 @@ void LoweringPreparePass::buildGlobalCtorDtorList() { mlir::ArrayAttr::get(&getContext(), globalCtors)); } - // We will eventual need to populate a global_dtor list, but that's not - // needed for globals with destructors. It will only be needed for functions - // that are marked as global destructors with an attribute. - assert(!cir::MissingFeatures::opGlobalDtorList()); + if (!globalDtorList.empty()) { + llvm::SmallVector<mlir::Attribute> globalDtors = + prepareCtorDtorAttrList<cir::GlobalDtorAttr>(&getContext(), + globalDtorList); + mlirModule->setAttr(cir::CIRDialect::getGlobalDtorsAttrName(), + mlir::ArrayAttr::get(&getContext(), globalDtors)); + } } void LoweringPreparePass::buildCXXGlobalInitFunc() { @@ -975,22 +980,28 @@ void LoweringPreparePass::lowerArrayCtor(cir::ArrayCtor op) { } void LoweringPreparePass::runOnOp(mlir::Operation *op) { - if (auto arrayCtor = dyn_cast<ArrayCtor>(op)) + if (auto arrayCtor = dyn_cast<cir::ArrayCtor>(op)) { lowerArrayCtor(arrayCtor); - else if (auto arrayDtor = dyn_cast<cir::ArrayDtor>(op)) + } else if (auto arrayDtor = dyn_cast<cir::ArrayDtor>(op)) { lowerArrayDtor(arrayDtor); - else if (auto cast = mlir::dyn_cast<cir::CastOp>(op)) + } else if (auto cast = mlir::dyn_cast<cir::CastOp>(op)) { lowerCastOp(cast); - else if (auto complexDiv = mlir::dyn_cast<cir::ComplexDivOp>(op)) + } else if (auto complexDiv = mlir::dyn_cast<cir::ComplexDivOp>(op)) { lowerComplexDivOp(complexDiv); - else if (auto complexMul = mlir::dyn_cast<cir::ComplexMulOp>(op)) + } else if (auto complexMul = mlir::dyn_cast<cir::ComplexMulOp>(op)) { lowerComplexMulOp(complexMul); - else if (auto glob = mlir::dyn_cast<cir::GlobalOp>(op)) + } else if (auto glob = mlir::dyn_cast<cir::GlobalOp>(op)) { lowerGlobalOp(glob); - else if (auto dynamicCast = mlir::dyn_cast<cir::DynamicCastOp>(op)) + } else if (auto dynamicCast = mlir::dyn_cast<cir::DynamicCastOp>(op)) { lowerDynamicCastOp(dynamicCast); - else if (auto unary = mlir::dyn_cast<cir::UnaryOp>(op)) + } else if (auto unary = mlir::dyn_cast<cir::UnaryOp>(op)) { lowerUnaryOp(unary); + } else if (auto fnOp = dyn_cast<cir::FuncOp>(op)) { + if (auto globalCtor = fnOp.getGlobalCtorPriority()) + globalCtorList.emplace_back(fnOp.getName(), globalCtor.value()); + else if (auto globalDtor = fnOp.getGlobalDtorPriority()) + globalDtorList.emplace_back(fnOp.getName(), globalDtor.value()); + } } void LoweringPreparePass::runOnOperation() { @@ -1003,7 +1014,7 @@ void LoweringPreparePass::runOnOperation() { op->walk([&](mlir::Operation *op) { if (mlir::isa<cir::ArrayCtor, cir::ArrayDtor, cir::CastOp, cir::ComplexMulOp, cir::ComplexDivOp, cir::DynamicCastOp, - cir::GlobalOp, cir::UnaryOp>(op)) + cir::FuncOp, cir::GlobalOp, cir::UnaryOp>(op)) opsToTransform.push_back(op); }); diff --git a/clang/lib/CIR/Dialect/Transforms/LoweringPrepareItaniumCXXABI.cpp b/clang/lib/CIR/Dialect/Transforms/LoweringPrepareItaniumCXXABI.cpp index 7d3c711..11ce2a8 100644 --- a/clang/lib/CIR/Dialect/Transforms/LoweringPrepareItaniumCXXABI.cpp +++ b/clang/lib/CIR/Dialect/Transforms/LoweringPrepareItaniumCXXABI.cpp @@ -92,7 +92,53 @@ static mlir::Value buildDynamicCastToVoidAfterNullCheck(cir::CIRBaseBuilderTy &builder, clang::ASTContext &astCtx, cir::DynamicCastOp op) { - llvm_unreachable("dynamic cast to void is NYI"); + mlir::Location loc = op.getLoc(); + bool vtableUsesRelativeLayout = op.getRelativeLayout(); + + // TODO(cir): consider address space in this function. + assert(!cir::MissingFeatures::addressSpace()); + + mlir::Type vtableElemTy; + uint64_t vtableElemAlign; + if (vtableUsesRelativeLayout) { + vtableElemTy = builder.getSIntNTy(32); + vtableElemAlign = 4; + } else { + const auto &targetInfo = astCtx.getTargetInfo(); + auto ptrdiffTy = targetInfo.getPtrDiffType(clang::LangAS::Default); + bool ptrdiffTyIsSigned = clang::TargetInfo::isTypeSigned(ptrdiffTy); + uint64_t ptrdiffTyWidth = targetInfo.getTypeWidth(ptrdiffTy); + + vtableElemTy = cir::IntType::get(builder.getContext(), ptrdiffTyWidth, + ptrdiffTyIsSigned); + vtableElemAlign = + llvm::divideCeil(targetInfo.getPointerAlign(clang::LangAS::Default), 8); + } + + // Access vtable to get the offset from the given object to its containing + // complete object. + // TODO: Add a specialized operation to get the object offset? + auto vptrTy = cir::VPtrType::get(builder.getContext()); + cir::PointerType vptrPtrTy = builder.getPointerTo(vptrTy); + auto vptrPtr = + cir::VTableGetVPtrOp::create(builder, loc, vptrPtrTy, op.getSrc()); + mlir::Value vptr = builder.createLoad(loc, vptrPtr); + mlir::Value elementPtr = + builder.createBitcast(vptr, builder.getPointerTo(vtableElemTy)); + mlir::Value minusTwo = builder.getSignedInt(loc, -2, 64); + auto offsetToTopSlotPtr = cir::PtrStrideOp::create( + builder, loc, builder.getPointerTo(vtableElemTy), elementPtr, minusTwo); + mlir::Value offsetToTop = + builder.createAlignedLoad(loc, offsetToTopSlotPtr, vtableElemAlign); + + // Add the offset to the given pointer to get the cast result. + // Cast the input pointer to a uint8_t* to allow pointer arithmetic. + cir::PointerType u8PtrTy = builder.getPointerTo(builder.getUIntNTy(8)); + mlir::Value srcBytePtr = builder.createBitcast(op.getSrc(), u8PtrTy); + auto dstBytePtr = + cir::PtrStrideOp::create(builder, loc, u8PtrTy, srcBytePtr, offsetToTop); + // Cast the result to a void*. + return builder.createBitcast(dstBytePtr, builder.getVoidPtrTy()); } mlir::Value diff --git a/clang/lib/CIR/Lowering/DirectToLLVM/LowerToLLVM.cpp b/clang/lib/CIR/Lowering/DirectToLLVM/LowerToLLVM.cpp index 26e0ba9..0243bf1 100644 --- a/clang/lib/CIR/Lowering/DirectToLLVM/LowerToLLVM.cpp +++ b/clang/lib/CIR/Lowering/DirectToLLVM/LowerToLLVM.cpp @@ -694,8 +694,8 @@ getLLVMMemOrder(std::optional<cir::MemOrder> memorder) { llvm_unreachable("unknown memory order"); } -mlir::LogicalResult CIRToLLVMAtomicCmpXchgLowering::matchAndRewrite( - cir::AtomicCmpXchg op, OpAdaptor adaptor, +mlir::LogicalResult CIRToLLVMAtomicCmpXchgOpLowering::matchAndRewrite( + cir::AtomicCmpXchgOp op, OpAdaptor adaptor, mlir::ConversionPatternRewriter &rewriter) const { mlir::Value expected = adaptor.getExpected(); mlir::Value desired = adaptor.getDesired(); @@ -719,8 +719,8 @@ mlir::LogicalResult CIRToLLVMAtomicCmpXchgLowering::matchAndRewrite( return mlir::success(); } -mlir::LogicalResult CIRToLLVMAtomicXchgLowering::matchAndRewrite( - cir::AtomicXchg op, OpAdaptor adaptor, +mlir::LogicalResult CIRToLLVMAtomicXchgOpLowering::matchAndRewrite( + cir::AtomicXchgOp op, OpAdaptor adaptor, mlir::ConversionPatternRewriter &rewriter) const { assert(!cir::MissingFeatures::atomicSyncScopeID()); mlir::LLVM::AtomicOrdering llvmOrder = getLLVMMemOrder(adaptor.getMemOrder()); @@ -1499,6 +1499,54 @@ mlir::LogicalResult CIRToLLVMConstantOpLowering::matchAndRewrite( return mlir::success(); } +static uint64_t getTypeSize(mlir::Type type, mlir::Operation &op) { + mlir::DataLayout layout(op.getParentOfType<mlir::ModuleOp>()); + // For LLVM purposes we treat void as u8. + if (isa<cir::VoidType>(type)) + type = cir::IntType::get(type.getContext(), 8, /*isSigned=*/false); + return llvm::divideCeil(layout.getTypeSizeInBits(type), 8); +} + +mlir::LogicalResult CIRToLLVMPtrDiffOpLowering::matchAndRewrite( + cir::PtrDiffOp op, OpAdaptor adaptor, + mlir::ConversionPatternRewriter &rewriter) const { + auto dstTy = mlir::cast<cir::IntType>(op.getType()); + mlir::Type llvmDstTy = getTypeConverter()->convertType(dstTy); + + auto lhs = rewriter.create<mlir::LLVM::PtrToIntOp>(op.getLoc(), llvmDstTy, + adaptor.getLhs()); + auto rhs = rewriter.create<mlir::LLVM::PtrToIntOp>(op.getLoc(), llvmDstTy, + adaptor.getRhs()); + + auto diff = + rewriter.create<mlir::LLVM::SubOp>(op.getLoc(), llvmDstTy, lhs, rhs); + + cir::PointerType ptrTy = op.getLhs().getType(); + assert(!cir::MissingFeatures::llvmLoweringPtrDiffConsidersPointee()); + uint64_t typeSize = getTypeSize(ptrTy.getPointee(), *op); + + // Avoid silly division by 1. + mlir::Value resultVal = diff.getResult(); + if (typeSize != 1) { + auto typeSizeVal = rewriter.create<mlir::LLVM::ConstantOp>( + op.getLoc(), llvmDstTy, typeSize); + + if (dstTy.isUnsigned()) { + auto uDiv = + rewriter.create<mlir::LLVM::UDivOp>(op.getLoc(), diff, typeSizeVal); + uDiv.setIsExact(true); + resultVal = uDiv.getResult(); + } else { + auto sDiv = + rewriter.create<mlir::LLVM::SDivOp>(op.getLoc(), diff, typeSizeVal); + sDiv.setIsExact(true); + resultVal = sDiv.getResult(); + } + } + rewriter.replaceOp(op, resultVal); + return mlir::success(); +} + mlir::LogicalResult CIRToLLVMExpectOpLowering::matchAndRewrite( cir::ExpectOp op, OpAdaptor adaptor, mlir::ConversionPatternRewriter &rewriter) const { @@ -1539,6 +1587,7 @@ void CIRToLLVMFuncOpLowering::lowerFuncAttributes( attr.getName() == getLinkageAttrNameString() || attr.getName() == func.getGlobalVisibilityAttrName() || attr.getName() == func.getDsoLocalAttrName() || + attr.getName() == func.getInlineKindAttrName() || (filterArgAndResAttrs && (attr.getName() == func.getArgAttrsAttrName() || attr.getName() == func.getResAttrsAttrName()))) @@ -1623,6 +1672,12 @@ mlir::LogicalResult CIRToLLVMFuncOpLowering::matchAndRewrite( assert(!cir::MissingFeatures::opFuncMultipleReturnVals()); + if (auto inlineKind = op.getInlineKind()) { + fn.setNoInline(inlineKind == cir::InlineKind::NoInline); + fn.setInlineHint(inlineKind == cir::InlineKind::InlineHint); + fn.setAlwaysInline(inlineKind == cir::InlineKind::AlwaysInline); + } + fn.setVisibility_Attr(mlir::LLVM::VisibilityAttr::get( getContext(), lowerCIRVisibilityToLLVMVisibility( op.getGlobalVisibilityAttr().getValue()))); @@ -1793,12 +1848,20 @@ CIRToLLVMGlobalOpLowering::getComdatAttr(cir::GlobalOp &op, if (!comdatOp) { builder.setInsertionPointToStart(module.getBody()); comdatOp = - builder.create<mlir::LLVM::ComdatOp>(module.getLoc(), comdatName); + mlir::LLVM::ComdatOp::create(builder, module.getLoc(), comdatName); + } + + if (auto comdatSelector = comdatOp.lookupSymbol<mlir::LLVM::ComdatSelectorOp>( + op.getSymName())) { + return mlir::SymbolRefAttr::get( + builder.getContext(), comdatName, + mlir::FlatSymbolRefAttr::get(comdatSelector.getSymNameAttr())); } builder.setInsertionPointToStart(&comdatOp.getBody().back()); - auto selectorOp = builder.create<mlir::LLVM::ComdatSelectorOp>( - comdatOp.getLoc(), op.getSymName(), mlir::LLVM::comdat::Comdat::Any); + auto selectorOp = mlir::LLVM::ComdatSelectorOp::create( + builder, comdatOp.getLoc(), op.getSymName(), + mlir::LLVM::comdat::Comdat::Any); return mlir::SymbolRefAttr::get( builder.getContext(), comdatName, mlir::FlatSymbolRefAttr::get(selectorOp.getSymNameAttr())); @@ -2598,7 +2661,13 @@ void ConvertCIRToLLVMPass::runOnOperation() { return std::make_pair(ctorAttr.getName(), ctorAttr.getPriority()); }); - assert(!cir::MissingFeatures::opGlobalDtorList()); + // Emit the llvm.global_dtors array. + buildCtorDtorList(module, cir::CIRDialect::getGlobalDtorsAttrName(), + "llvm.global_dtors", [](mlir::Attribute attr) { + auto dtorAttr = mlir::cast<cir::GlobalDtorAttr>(attr); + return std::make_pair(dtorAttr.getName(), + dtorAttr.getPriority()); + }); } mlir::LogicalResult CIRToLLVMBrOpLowering::matchAndRewrite( diff --git a/clang/lib/CodeGen/ABIInfoImpl.cpp b/clang/lib/CodeGen/ABIInfoImpl.cpp index 13c837a..1e3ac2e 100644 --- a/clang/lib/CodeGen/ABIInfoImpl.cpp +++ b/clang/lib/CodeGen/ABIInfoImpl.cpp @@ -105,7 +105,7 @@ llvm::Type *CodeGen::getVAListElementType(CodeGenFunction &CGF) { CGCXXABI::RecordArgABI CodeGen::getRecordArgABI(const RecordType *RT, CGCXXABI &CXXABI) { - const RecordDecl *RD = RT->getOriginalDecl()->getDefinitionOrSelf(); + const RecordDecl *RD = RT->getDecl()->getDefinitionOrSelf(); if (const auto *CXXRD = dyn_cast<CXXRecordDecl>(RD)) return CXXABI.getRecordArgABI(CXXRD); if (!RD->canPassInRegisters()) @@ -136,7 +136,7 @@ bool CodeGen::classifyReturnType(const CGCXXABI &CXXABI, CGFunctionInfo &FI, QualType CodeGen::useFirstFieldIfTransparentUnion(QualType Ty) { if (const RecordType *UT = Ty->getAsUnionType()) { - const RecordDecl *UD = UT->getOriginalDecl()->getDefinitionOrSelf(); + const RecordDecl *UD = UT->getDecl()->getDefinitionOrSelf(); if (UD->hasAttr<TransparentUnionAttr>()) { assert(!UD->field_empty() && "sema created an empty transparent union"); return UD->field_begin()->getType(); @@ -274,7 +274,7 @@ bool CodeGen::isEmptyField(ASTContext &Context, const FieldDecl *FD, // according to the Itanium ABI. The exception applies only to records, // not arrays of records, so we must also check whether we stripped off an // array type above. - if (isa<CXXRecordDecl>(RT->getOriginalDecl()) && + if (isa<CXXRecordDecl>(RT->getDecl()) && (WasArray || (!AsIfNoUniqueAddr && !FD->hasAttr<NoUniqueAddressAttr>()))) return false; diff --git a/clang/lib/CodeGen/BackendUtil.cpp b/clang/lib/CodeGen/BackendUtil.cpp index f8e8086..6020684 100644 --- a/clang/lib/CodeGen/BackendUtil.cpp +++ b/clang/lib/CodeGen/BackendUtil.cpp @@ -1200,7 +1200,8 @@ void EmitAssemblyHelper::RunOptimizationPipeline( } } - if (shouldEmitUnifiedLTOModueFlag()) + if (shouldEmitUnifiedLTOModueFlag() && + !TheModule->getModuleFlag("UnifiedLTO")) TheModule->addModuleFlag(llvm::Module::Error, "UnifiedLTO", uint32_t(1)); } diff --git a/clang/lib/CodeGen/CGAtomic.cpp b/clang/lib/CodeGen/CGAtomic.cpp index 4a3446a..d95dab3 100644 --- a/clang/lib/CodeGen/CGAtomic.cpp +++ b/clang/lib/CodeGen/CGAtomic.cpp @@ -507,6 +507,18 @@ static llvm::Value *EmitPostAtomicMinMax(CGBuilderTy &Builder, bool IsSigned, llvm::Value *OldVal, llvm::Value *RHS) { + const bool IsFP = OldVal->getType()->isFloatingPointTy(); + + if (IsFP) { + llvm::Intrinsic::ID IID = (Op == AtomicExpr::AO__atomic_max_fetch || + Op == AtomicExpr::AO__scoped_atomic_max_fetch) + ? llvm::Intrinsic::maxnum + : llvm::Intrinsic::minnum; + + return Builder.CreateBinaryIntrinsic(IID, OldVal, RHS, llvm::FMFSource(), + "newval"); + } + llvm::CmpInst::Predicate Pred; switch (Op) { default: diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp index 9ee810c..92dba32 100644 --- a/clang/lib/CodeGen/CGBuiltin.cpp +++ b/clang/lib/CodeGen/CGBuiltin.cpp @@ -4283,15 +4283,11 @@ RValue CodeGenFunction::EmitBuiltinExpr(const GlobalDecl GD, unsigned BuiltinID, CharUnits Align = CGM.getNaturalTypeAlignment( E->getType()->getAs<VectorType>()->getElementType(), nullptr); - llvm::Value *AlignVal = - llvm::ConstantInt::get(Int32Ty, Align.getQuantity()); llvm::Value *Result; if (BuiltinID == Builtin::BI__builtin_masked_load) { - Function *F = - CGM.getIntrinsic(Intrinsic::masked_load, {RetTy, Ptr->getType()}); - Result = - Builder.CreateCall(F, {Ptr, AlignVal, Mask, PassThru}, "masked_load"); + Result = Builder.CreateMaskedLoad(RetTy, Ptr, Align.getAsAlign(), Mask, + PassThru, "masked_load"); } else { Function *F = CGM.getIntrinsic(Intrinsic::masked_expandload, {RetTy}); Result = @@ -4307,8 +4303,6 @@ RValue CodeGenFunction::EmitBuiltinExpr(const GlobalDecl GD, unsigned BuiltinID, llvm::Type *RetTy = CGM.getTypes().ConvertType(E->getType()); CharUnits Align = CGM.getNaturalTypeAlignment( E->getType()->getAs<VectorType>()->getElementType(), nullptr); - llvm::Value *AlignVal = - llvm::ConstantInt::get(Int32Ty, Align.getQuantity()); llvm::Value *PassThru = llvm::PoisonValue::get(RetTy); if (E->getNumArgs() > 3) @@ -4318,12 +4312,8 @@ RValue CodeGenFunction::EmitBuiltinExpr(const GlobalDecl GD, unsigned BuiltinID, E->getType()->getAs<VectorType>()->getElementType()); llvm::Value *PtrVec = Builder.CreateGEP(ElemTy, Ptr, Idx); - llvm::Value *Result; - Function *F = - CGM.getIntrinsic(Intrinsic::masked_gather, {RetTy, PtrVec->getType()}); - - Result = Builder.CreateCall(F, {PtrVec, AlignVal, Mask, PassThru}, - "masked_gather"); + llvm::Value *Result = Builder.CreateMaskedGather( + RetTy, PtrVec, Align.getAsAlign(), Mask, PassThru, "masked_gather"); return RValue::get(Result); } case Builtin::BI__builtin_masked_store: @@ -4338,13 +4328,9 @@ RValue CodeGenFunction::EmitBuiltinExpr(const GlobalDecl GD, unsigned BuiltinID, CharUnits Align = CGM.getNaturalTypeAlignment( E->getArg(1)->getType()->getAs<VectorType>()->getElementType(), nullptr); - llvm::Value *AlignVal = - llvm::ConstantInt::get(Int32Ty, Align.getQuantity()); if (BuiltinID == Builtin::BI__builtin_masked_store) { - llvm::Function *F = CGM.getIntrinsic(llvm::Intrinsic::masked_store, - {ValLLTy, Ptr->getType()}); - Builder.CreateCall(F, {Val, Ptr, AlignVal, Mask}); + Builder.CreateMaskedStore(Val, Ptr, Align.getAsAlign(), Mask); } else { llvm::Function *F = CGM.getIntrinsic(llvm::Intrinsic::masked_compressstore, {ValLLTy}); @@ -4361,17 +4347,12 @@ RValue CodeGenFunction::EmitBuiltinExpr(const GlobalDecl GD, unsigned BuiltinID, CharUnits Align = CGM.getNaturalTypeAlignment( E->getArg(2)->getType()->getAs<VectorType>()->getElementType(), nullptr); - llvm::Value *AlignVal = - llvm::ConstantInt::get(Int32Ty, Align.getQuantity()); llvm::Type *ElemTy = CGM.getTypes().ConvertType( E->getArg(1)->getType()->getAs<VectorType>()->getElementType()); llvm::Value *PtrVec = Builder.CreateGEP(ElemTy, Ptr, Idx); - Function *F = CGM.getIntrinsic(Intrinsic::masked_scatter, - {Val->getType(), PtrVec->getType()}); - - Builder.CreateCall(F, {Val, PtrVec, AlignVal, Mask}); + Builder.CreateMaskedScatter(Val, PtrVec, Align.getAsAlign(), Mask); return RValue(); } case Builtin::BI__builtin_isinf_sign: { diff --git a/clang/lib/CodeGen/CGCall.cpp b/clang/lib/CodeGen/CGCall.cpp index df28641..741fa44 100644 --- a/clang/lib/CodeGen/CGCall.cpp +++ b/clang/lib/CodeGen/CGCall.cpp @@ -1896,7 +1896,7 @@ bool CodeGenModule::MayDropFunctionReturn(const ASTContext &Context, // complex destructor or a non-trivially copyable type. if (const RecordType *RT = ReturnType.getCanonicalType()->getAsCanonical<RecordType>()) { - if (const auto *ClassDecl = dyn_cast<CXXRecordDecl>(RT->getOriginalDecl())) + if (const auto *ClassDecl = dyn_cast<CXXRecordDecl>(RT->getDecl())) return ClassDecl->hasTrivialDestructor(); } return ReturnType.isTriviallyCopyableType(Context); @@ -3853,7 +3853,7 @@ static void setUsedBits(CodeGenModule &CGM, const RecordType *RTy, int Offset, SmallVectorImpl<uint64_t> &Bits) { ASTContext &Context = CGM.getContext(); int CharWidth = Context.getCharWidth(); - const RecordDecl *RD = RTy->getOriginalDecl()->getDefinition(); + const RecordDecl *RD = RTy->getDecl()->getDefinition(); const ASTRecordLayout &ASTLayout = Context.getASTRecordLayout(RD); const CGRecordLayout &Layout = CGM.getTypes().getCGRecordLayout(RD); diff --git a/clang/lib/CodeGen/CGClass.cpp b/clang/lib/CodeGen/CGClass.cpp index f31f0a2..f782b0c 100644 --- a/clang/lib/CodeGen/CGClass.cpp +++ b/clang/lib/CodeGen/CGClass.cpp @@ -2026,7 +2026,7 @@ void CodeGenFunction::EnterDtorCleanups(const CXXDestructorDecl *DD, // Anonymous union members do not have their destructors called. const RecordType *RT = type->getAsUnionType(); - if (RT && RT->getOriginalDecl()->isAnonymousStructOrUnion()) + if (RT && RT->getDecl()->isAnonymousStructOrUnion()) continue; CleanupKind cleanupKind = getCleanupKind(dtorKind); diff --git a/clang/lib/CodeGen/CGDebugInfo.cpp b/clang/lib/CodeGen/CGDebugInfo.cpp index 9fe9a13..12e2813ef 100644 --- a/clang/lib/CodeGen/CGDebugInfo.cpp +++ b/clang/lib/CodeGen/CGDebugInfo.cpp @@ -647,6 +647,68 @@ StringRef CGDebugInfo::getCurrentDirname() { return CGM.getCodeGenOpts().DebugCompilationDir; } +static llvm::dwarf::SourceLanguage GetSourceLanguage(const CodeGenModule &CGM) { + const CodeGenOptions &CGO = CGM.getCodeGenOpts(); + const LangOptions &LO = CGM.getLangOpts(); + + assert(CGO.DwarfVersion <= 5); + + llvm::dwarf::SourceLanguage LangTag; + if (LO.CPlusPlus) { + if (LO.ObjC) + LangTag = llvm::dwarf::DW_LANG_ObjC_plus_plus; + else if (CGO.DebugStrictDwarf && CGO.DwarfVersion < 5) + LangTag = llvm::dwarf::DW_LANG_C_plus_plus; + else if (LO.CPlusPlus14) + LangTag = llvm::dwarf::DW_LANG_C_plus_plus_14; + else if (LO.CPlusPlus11) + LangTag = llvm::dwarf::DW_LANG_C_plus_plus_11; + else + LangTag = llvm::dwarf::DW_LANG_C_plus_plus; + } else if (LO.ObjC) { + LangTag = llvm::dwarf::DW_LANG_ObjC; + } else if (LO.OpenCL && (!CGO.DebugStrictDwarf || CGO.DwarfVersion >= 5)) { + LangTag = llvm::dwarf::DW_LANG_OpenCL; + } else if (LO.C11 && !(CGO.DebugStrictDwarf && CGO.DwarfVersion < 5)) { + LangTag = llvm::dwarf::DW_LANG_C11; + } else if (LO.C99) { + LangTag = llvm::dwarf::DW_LANG_C99; + } else { + LangTag = llvm::dwarf::DW_LANG_C89; + } + + return LangTag; +} + +static llvm::DISourceLanguageName +GetDISourceLanguageName(const CodeGenModule &CGM) { + // Emit pre-DWARFv6 language codes. + if (CGM.getCodeGenOpts().DwarfVersion < 6) + return llvm::DISourceLanguageName(GetSourceLanguage(CGM)); + + const LangOptions &LO = CGM.getLangOpts(); + + uint32_t LangVersion = 0; + llvm::dwarf::SourceLanguageName LangTag; + if (LO.CPlusPlus) { + if (LO.ObjC) { + LangTag = llvm::dwarf::DW_LNAME_ObjC_plus_plus; + } else { + LangTag = llvm::dwarf::DW_LNAME_C_plus_plus; + LangVersion = LO.getCPlusPlusLangStd().value_or(0); + } + } else if (LO.ObjC) { + LangTag = llvm::dwarf::DW_LNAME_ObjC; + } else if (LO.OpenCL) { + LangTag = llvm::dwarf::DW_LNAME_OpenCL_C; + } else { + LangTag = llvm::dwarf::DW_LNAME_C; + LangVersion = LO.getCLangStd().value_or(0); + } + + return llvm::DISourceLanguageName(LangTag, LangVersion); +} + void CGDebugInfo::CreateCompileUnit() { SmallString<64> Checksum; std::optional<llvm::DIFile::ChecksumKind> CSKind; @@ -702,31 +764,6 @@ void CGDebugInfo::CreateCompileUnit() { } } - llvm::dwarf::SourceLanguage LangTag; - if (LO.CPlusPlus) { - if (LO.ObjC) - LangTag = llvm::dwarf::DW_LANG_ObjC_plus_plus; - else if (CGO.DebugStrictDwarf && CGO.DwarfVersion < 5) - LangTag = llvm::dwarf::DW_LANG_C_plus_plus; - else if (LO.CPlusPlus14) - LangTag = llvm::dwarf::DW_LANG_C_plus_plus_14; - else if (LO.CPlusPlus11) - LangTag = llvm::dwarf::DW_LANG_C_plus_plus_11; - else - LangTag = llvm::dwarf::DW_LANG_C_plus_plus; - } else if (LO.ObjC) { - LangTag = llvm::dwarf::DW_LANG_ObjC; - } else if (LO.OpenCL && (!CGM.getCodeGenOpts().DebugStrictDwarf || - CGM.getCodeGenOpts().DwarfVersion >= 5)) { - LangTag = llvm::dwarf::DW_LANG_OpenCL; - } else if (LO.C11 && !(CGO.DebugStrictDwarf && CGO.DwarfVersion < 5)) { - LangTag = llvm::dwarf::DW_LANG_C11; - } else if (LO.C99) { - LangTag = llvm::dwarf::DW_LANG_C99; - } else { - LangTag = llvm::dwarf::DW_LANG_C89; - } - std::string Producer = getClangFullVersion(); // Figure out which version of the ObjC runtime we have. @@ -787,7 +824,7 @@ void CGDebugInfo::CreateCompileUnit() { // Create new compile unit. TheCU = DBuilder.createCompileUnit( - llvm::DISourceLanguageName(LangTag), CUFile, + GetDISourceLanguageName(CGM), CUFile, CGOpts.EmitVersionIdentMetadata ? Producer : "", CGOpts.OptimizationLevel != 0 || CGOpts.PrepareForLTO || CGOpts.PrepareForThinLTO, @@ -1234,20 +1271,46 @@ llvm::DIType *CGDebugInfo::CreateType(const PointerType *Ty, Ty->getPointeeType(), Unit); } -/// \return whether a C++ mangling exists for the type defined by TD. -static bool hasCXXMangling(const TagDecl *TD, llvm::DICompileUnit *TheCU) { - switch (TheCU->getSourceLanguage().getUnversionedName()) { +static bool hasCXXMangling(llvm::dwarf::SourceLanguage Lang, bool IsTagDecl) { + switch (Lang) { case llvm::dwarf::DW_LANG_C_plus_plus: case llvm::dwarf::DW_LANG_C_plus_plus_11: case llvm::dwarf::DW_LANG_C_plus_plus_14: return true; case llvm::dwarf::DW_LANG_ObjC_plus_plus: - return isa<CXXRecordDecl>(TD) || isa<EnumDecl>(TD); + return IsTagDecl; + default: + return false; + } +} + +static bool hasCXXMangling(llvm::dwarf::SourceLanguageName Lang, + bool IsTagDecl) { + switch (Lang) { + case llvm::dwarf::DW_LNAME_C_plus_plus: + return true; + case llvm::dwarf::DW_LNAME_ObjC_plus_plus: + return IsTagDecl; default: return false; } } +/// \return whether a C++ mangling exists for the type defined by TD. +static bool hasCXXMangling(const TagDecl *TD, llvm::DICompileUnit *TheCU) { + const bool IsTagDecl = isa<CXXRecordDecl>(TD) || isa<EnumDecl>(TD); + + if (llvm::DISourceLanguageName SourceLang = TheCU->getSourceLanguage(); + SourceLang.hasVersionedName()) + return hasCXXMangling( + static_cast<llvm::dwarf::SourceLanguageName>(SourceLang.getName()), + IsTagDecl); + else + return hasCXXMangling( + static_cast<llvm::dwarf::SourceLanguage>(SourceLang.getName()), + IsTagDecl); +} + // Determines if the debug info for this tag declaration needs a type // identifier. The purpose of the unique identifier is to deduplicate type // information for identical types across TUs. Because of the C++ one definition @@ -1290,7 +1353,7 @@ static bool needsTypeIdentifier(const TagDecl *TD, CodeGenModule &CGM, static SmallString<256> getTypeIdentifier(const TagType *Ty, CodeGenModule &CGM, llvm::DICompileUnit *TheCU) { SmallString<256> Identifier; - const TagDecl *TD = Ty->getOriginalDecl()->getDefinitionOrSelf(); + const TagDecl *TD = Ty->getDecl()->getDefinitionOrSelf(); if (!needsTypeIdentifier(TD, CGM, TheCU)) return Identifier; @@ -1326,7 +1389,7 @@ static llvm::dwarf::Tag getTagForRecord(const RecordDecl *RD) { llvm::DICompositeType * CGDebugInfo::getOrCreateRecordFwdDecl(const RecordType *Ty, llvm::DIScope *Ctx) { - const RecordDecl *RD = Ty->getOriginalDecl()->getDefinitionOrSelf(); + const RecordDecl *RD = Ty->getDecl()->getDefinitionOrSelf(); if (llvm::DIType *T = getTypeOrNull(QualType(Ty, 0))) return cast<llvm::DICompositeType>(T); llvm::DIFile *DefUnit = getOrCreateFile(RD->getLocation()); @@ -2402,7 +2465,7 @@ void CGDebugInfo::CollectCXXBasesAux( for (const auto &BI : Bases) { const auto *Base = cast<CXXRecordDecl>( - BI.getType()->castAsCanonical<RecordType>()->getOriginalDecl()) + BI.getType()->castAsCanonical<RecordType>()->getDecl()) ->getDefinition(); if (!SeenTypes.insert(Base).second) continue; @@ -3077,7 +3140,7 @@ void CGDebugInfo::completeRequiredType(const RecordDecl *RD) { } llvm::DIType *CGDebugInfo::CreateType(const RecordType *Ty) { - RecordDecl *RD = Ty->getOriginalDecl()->getDefinitionOrSelf(); + RecordDecl *RD = Ty->getDecl()->getDefinitionOrSelf(); llvm::DIType *T = cast_or_null<llvm::DIType>(getTypeOrNull(QualType(Ty, 0))); if (T || shouldOmitDefinition(DebugKind, DebugTypeExtRefs, RD, CGM.getLangOpts())) { @@ -3105,7 +3168,7 @@ llvm::DIType *CGDebugInfo::GetPreferredNameType(const CXXRecordDecl *RD, std::pair<llvm::DIType *, llvm::DIType *> CGDebugInfo::CreateTypeDefinition(const RecordType *Ty) { - RecordDecl *RD = Ty->getOriginalDecl()->getDefinitionOrSelf(); + RecordDecl *RD = Ty->getDecl()->getDefinitionOrSelf(); // Get overall information about the record type for the debug info. llvm::DIFile *DefUnit = getOrCreateFile(RD->getLocation()); @@ -3748,7 +3811,7 @@ llvm::DIType *CGDebugInfo::CreateType(const HLSLInlineSpirvType *Ty, static auto getEnumInfo(CodeGenModule &CGM, llvm::DICompileUnit *TheCU, const EnumType *Ty) { - const EnumDecl *ED = Ty->getOriginalDecl()->getDefinitionOrSelf(); + const EnumDecl *ED = Ty->getDecl()->getDefinitionOrSelf(); uint64_t Size = 0; uint32_t Align = 0; @@ -4151,7 +4214,7 @@ CGDebugInfo::getOrCreateLimitedType(const RecordType *Ty) { // TODO: Currently used for context chains when limiting debug info. llvm::DICompositeType *CGDebugInfo::CreateLimitedType(const RecordType *Ty) { - RecordDecl *RD = Ty->getOriginalDecl()->getDefinitionOrSelf(); + RecordDecl *RD = Ty->getDecl()->getDefinitionOrSelf(); // Get overall information about the record type for the debug info. StringRef RDName = getClassName(RD); @@ -4238,8 +4301,7 @@ llvm::DICompositeType *CGDebugInfo::CreateLimitedType(const RecordType *Ty) { break; } - if (auto *CTSD = - dyn_cast<ClassTemplateSpecializationDecl>(Ty->getOriginalDecl())) { + if (auto *CTSD = dyn_cast<ClassTemplateSpecializationDecl>(Ty->getDecl())) { CXXRecordDecl *TemplateDecl = CTSD->getSpecializedTemplate()->getTemplatedDecl(); RegionMap[TemplateDecl].reset(RealDecl); @@ -5141,7 +5203,7 @@ llvm::DILocalVariable *CGDebugInfo::EmitDeclare(const VarDecl *VD, } else if (const auto *RT = dyn_cast<RecordType>(VD->getType())) { // If VD is an anonymous union then Storage represents value for // all union fields. - const RecordDecl *RD = RT->getOriginalDecl()->getDefinitionOrSelf(); + const RecordDecl *RD = RT->getDecl()->getDefinitionOrSelf(); if (RD->isUnion() && RD->isAnonymousStructOrUnion()) { // GDB has trouble finding local variables in anonymous unions, so we emit // artificial local variables for each of the members. @@ -5691,9 +5753,8 @@ llvm::DIGlobalVariableExpression *CGDebugInfo::CollectAnonRecordDecls( // Ignore unnamed fields, but recurse into anonymous records. if (FieldName.empty()) { if (const auto *RT = dyn_cast<RecordType>(Field->getType())) - GVE = - CollectAnonRecordDecls(RT->getOriginalDecl()->getDefinitionOrSelf(), - Unit, LineNo, LinkageName, Var, DContext); + GVE = CollectAnonRecordDecls(RT->getDecl()->getDefinitionOrSelf(), Unit, + LineNo, LinkageName, Var, DContext); continue; } // Use VarDecl's Tag, Scope and Line number. @@ -5712,7 +5773,7 @@ static bool ReferencesAnonymousEntity(RecordType *RT) { // But so long as it's not one of those, it doesn't matter if some sub-type // of the record (a template parameter) can't be reconstituted - because the // un-reconstitutable type itself will carry its own name. - const auto *RD = dyn_cast<CXXRecordDecl>(RT->getOriginalDecl()); + const auto *RD = dyn_cast<CXXRecordDecl>(RT->getDecl()); if (!RD) return false; if (!RD->getIdentifier()) @@ -5774,7 +5835,7 @@ struct ReconstitutableType : public RecursiveASTVisitor<ReconstitutableType> { bool TraverseEnumType(EnumType *ET, bool = false) { // Unnamed enums can't be reconstituted due to a lack of column info we // produce in the DWARF, so we can't get Clang's full name back. - if (const auto *ED = dyn_cast<EnumDecl>(ET->getOriginalDecl())) { + if (const auto *ED = dyn_cast<EnumDecl>(ET->getDecl())) { if (!ED->getIdentifier()) { Reconstitutable = false; return false; diff --git a/clang/lib/CodeGen/CGExpr.cpp b/clang/lib/CodeGen/CGExpr.cpp index e8255b0..8439ec7 100644 --- a/clang/lib/CodeGen/CGExpr.cpp +++ b/clang/lib/CodeGen/CGExpr.cpp @@ -2011,7 +2011,7 @@ static bool isConstantEmittableObjectType(QualType type) { // Otherwise, all object types satisfy this except C++ classes with // mutable subobjects or non-trivial copy/destroy behavior. if (const auto *RT = dyn_cast<RecordType>(type)) - if (const auto *RD = dyn_cast<CXXRecordDecl>(RT->getOriginalDecl())) { + if (const auto *RD = dyn_cast<CXXRecordDecl>(RT->getDecl())) { RD = RD->getDefinitionOrSelf(); if (RD->hasMutableFields() || !RD->isTrivial()) return false; @@ -4564,7 +4564,7 @@ static bool IsPreserveAIArrayBase(CodeGenFunction &CGF, const Expr *ArrayBase) { const auto *PointeeT = PtrT->getPointeeType() ->getUnqualifiedDesugaredType(); if (const auto *RecT = dyn_cast<RecordType>(PointeeT)) - return RecT->getOriginalDecl() + return RecT->getDecl() ->getMostRecentDecl() ->hasAttr<BPFPreserveAccessIndexAttr>(); return false; @@ -7008,7 +7008,7 @@ void CodeGenFunction::FlattenAccessAndTypeLValue( WorkList.emplace_back(LVal, CAT->getElementType(), IdxListCopy); } } else if (const auto *RT = dyn_cast<RecordType>(T)) { - const RecordDecl *Record = RT->getOriginalDecl()->getDefinitionOrSelf(); + const RecordDecl *Record = RT->getDecl()->getDefinitionOrSelf(); assert(!Record->isUnion() && "Union types not supported in flat cast."); const CXXRecordDecl *CXXD = dyn_cast<CXXRecordDecl>(Record); diff --git a/clang/lib/CodeGen/CGExprAgg.cpp b/clang/lib/CodeGen/CGExprAgg.cpp index 07b9aeb..eee397f 100644 --- a/clang/lib/CodeGen/CGExprAgg.cpp +++ b/clang/lib/CodeGen/CGExprAgg.cpp @@ -2080,7 +2080,7 @@ static CharUnits GetNumNonZeroBytesInInit(const Expr *E, CodeGenFunction &CGF) { // referencee. InitListExprs for unions and arrays can't have references. if (const RecordType *RT = E->getType()->getAsCanonical<RecordType>()) { if (!RT->isUnionType()) { - RecordDecl *SD = RT->getOriginalDecl()->getDefinitionOrSelf(); + RecordDecl *SD = RT->getDecl()->getDefinitionOrSelf(); CharUnits NumNonZeroBytes = CharUnits::Zero(); unsigned ILEElement = 0; @@ -2133,7 +2133,7 @@ static void CheckAggExprForMemSetUse(AggValueSlot &Slot, const Expr *E, if (const RecordType *RT = CGF.getContext() .getBaseElementType(E->getType()) ->getAsCanonical<RecordType>()) { - const CXXRecordDecl *RD = cast<CXXRecordDecl>(RT->getOriginalDecl()); + const auto *RD = cast<CXXRecordDecl>(RT->getDecl()); if (RD->hasUserDeclaredConstructor()) return; } diff --git a/clang/lib/CodeGen/CGExprCXX.cpp b/clang/lib/CodeGen/CGExprCXX.cpp index 31ac266..14d8db3 100644 --- a/clang/lib/CodeGen/CGExprCXX.cpp +++ b/clang/lib/CodeGen/CGExprCXX.cpp @@ -1236,8 +1236,8 @@ void CodeGenFunction::EmitNewArrayInitializer( if (auto *ILE = dyn_cast<InitListExpr>(Init)) { if (const RecordType *RType = ILE->getType()->getAsCanonical<RecordType>()) { - if (RType->getOriginalDecl()->isStruct()) { - const RecordDecl *RD = RType->getOriginalDecl()->getDefinitionOrSelf(); + if (RType->getDecl()->isStruct()) { + const RecordDecl *RD = RType->getDecl()->getDefinitionOrSelf(); unsigned NumElements = 0; if (auto *CXXRD = dyn_cast<CXXRecordDecl>(RD)) NumElements = CXXRD->getNumBases(); diff --git a/clang/lib/CodeGen/CGExprScalar.cpp b/clang/lib/CodeGen/CGExprScalar.cpp index 715160d..714192d 100644 --- a/clang/lib/CodeGen/CGExprScalar.cpp +++ b/clang/lib/CodeGen/CGExprScalar.cpp @@ -3583,7 +3583,7 @@ Value *ScalarExprEmitter::VisitOffsetOfExpr(OffsetOfExpr *E) { } const ASTRecordLayout &RL = CGF.getContext().getASTRecordLayout( - CurrentType->castAsCanonical<RecordType>()->getOriginalDecl()); + CurrentType->castAsCanonical<RecordType>()->getDecl()); // Save the element type. CurrentType = ON.getBase()->getType(); diff --git a/clang/lib/CodeGen/CGHLSLRuntime.cpp b/clang/lib/CodeGen/CGHLSLRuntime.cpp index 603cef9..ecab933 100644 --- a/clang/lib/CodeGen/CGHLSLRuntime.cpp +++ b/clang/lib/CodeGen/CGHLSLRuntime.cpp @@ -519,6 +519,10 @@ void clang::CodeGen::CGHLSLRuntime::setHLSLEntryAttributes( if (CGM.getCodeGenOpts().OptimizationLevel == 0) Fn->addFnAttr(llvm::Attribute::OptimizeNone); Fn->addFnAttr(llvm::Attribute::NoInline); + + if (CGM.getLangOpts().HLSLSpvEnableMaximalReconvergence) { + Fn->addFnAttr("enable-maximal-reconvergence", "true"); + } } static Value *buildVectorInput(IRBuilder<> &B, Function *F, llvm::Type *Ty) { diff --git a/clang/lib/CodeGen/CGNonTrivialStruct.cpp b/clang/lib/CodeGen/CGNonTrivialStruct.cpp index 2d70e4c..0a383c8 100644 --- a/clang/lib/CodeGen/CGNonTrivialStruct.cpp +++ b/clang/lib/CodeGen/CGNonTrivialStruct.cpp @@ -464,8 +464,7 @@ template <class Derived> struct GenFuncBase { if (WrongType) { std::string FuncName = std::string(F->getName()); - SourceLocation Loc = - QT->castAs<RecordType>()->getOriginalDecl()->getLocation(); + SourceLocation Loc = QT->castAs<RecordType>()->getDecl()->getLocation(); CGM.Error(Loc, "special function " + FuncName + " for non-trivial C struct has incorrect type"); return nullptr; diff --git a/clang/lib/CodeGen/CGObjCMac.cpp b/clang/lib/CodeGen/CGObjCMac.cpp index dbcce9b..c571821a 100644 --- a/clang/lib/CodeGen/CGObjCMac.cpp +++ b/clang/lib/CodeGen/CGObjCMac.cpp @@ -2495,7 +2495,7 @@ void CGObjCCommonMac::BuildRCBlockVarRecordLayout(const RecordType *RT, CharUnits BytePos, bool &HasUnion, bool ByrefLayout) { - const RecordDecl *RD = RT->getOriginalDecl()->getDefinitionOrSelf(); + const RecordDecl *RD = RT->getDecl()->getDefinitionOrSelf(); SmallVector<const FieldDecl *, 16> Fields(RD->fields()); llvm::Type *Ty = CGM.getTypes().ConvertType(QualType(RT, 0)); const llvm::StructLayout *RecLayout = @@ -5184,7 +5184,7 @@ CGObjCCommonMac::GetIvarLayoutName(IdentifierInfo *Ident, } void IvarLayoutBuilder::visitRecord(const RecordType *RT, CharUnits offset) { - const RecordDecl *RD = RT->getOriginalDecl()->getDefinitionOrSelf(); + const RecordDecl *RD = RT->getDecl()->getDefinitionOrSelf(); // If this is a union, remember that we had one, because it might mess // up the ordering of layout entries. diff --git a/clang/lib/CodeGen/CGOpenMPRuntime.cpp b/clang/lib/CodeGen/CGOpenMPRuntime.cpp index fa94692..1ff2be7 100644 --- a/clang/lib/CodeGen/CGOpenMPRuntime.cpp +++ b/clang/lib/CodeGen/CGOpenMPRuntime.cpp @@ -1762,8 +1762,11 @@ void CGOpenMPRuntime::emitDeclareTargetFunction(const FunctionDecl *FD, // access its value. llvm::GlobalValue *Addr = GV; if (CGM.getLangOpts().OpenMPIsTargetDevice) { + llvm::PointerType *FnPtrTy = llvm::PointerType::get( + CGM.getLLVMContext(), + CGM.getModule().getDataLayout().getProgramAddressSpace()); Addr = new llvm::GlobalVariable( - CGM.getModule(), CGM.VoidPtrTy, + CGM.getModule(), FnPtrTy, /*isConstant=*/true, llvm::GlobalValue::ExternalLinkage, GV, Name, nullptr, llvm::GlobalValue::NotThreadLocal, CGM.getModule().getDataLayout().getDefaultGlobalsAddressSpace()); diff --git a/clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp b/clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp index 3613b6a..fddeba9 100644 --- a/clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp +++ b/clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp @@ -7,7 +7,7 @@ //===----------------------------------------------------------------------===// // // This provides a generalized class for OpenMP runtime code generation -// specialized by GPU targets NVPTX and AMDGCN. +// specialized by GPU targets NVPTX, AMDGCN and SPIR-V. // //===----------------------------------------------------------------------===// @@ -1242,12 +1242,13 @@ void CGOpenMPRuntimeGPU::emitParallelCall( CGBuilderTy &Bld = CGF.Builder; llvm::Value *NumThreadsVal = NumThreads; llvm::Function *WFn = WrapperFunctionsMap[OutlinedFn]; - llvm::Value *ID = llvm::ConstantPointerNull::get(CGM.Int8PtrTy); - if (WFn) - ID = Bld.CreateBitOrPointerCast(WFn, CGM.Int8PtrTy); - llvm::Type *FnPtrTy = llvm::PointerType::get( + llvm::PointerType *FnPtrTy = llvm::PointerType::get( CGF.getLLVMContext(), CGM.getDataLayout().getProgramAddressSpace()); + llvm::Value *ID = llvm::ConstantPointerNull::get(FnPtrTy); + if (WFn) + ID = Bld.CreateBitOrPointerCast(WFn, FnPtrTy); + llvm::Value *FnPtr = Bld.CreateBitOrPointerCast(OutlinedFn, FnPtrTy); // Create a private scope that will globalize the arguments diff --git a/clang/lib/CodeGen/CGStmt.cpp b/clang/lib/CodeGen/CGStmt.cpp index 92636f2..fdc1a11 100644 --- a/clang/lib/CodeGen/CGStmt.cpp +++ b/clang/lib/CodeGen/CGStmt.cpp @@ -2674,7 +2674,8 @@ EmitAsmStores(CodeGenFunction &CGF, const AsmStmt &S, const llvm::ArrayRef<LValue> ResultRegDests, const llvm::ArrayRef<QualType> ResultRegQualTys, const llvm::BitVector &ResultTypeRequiresCast, - const llvm::BitVector &ResultRegIsFlagReg) { + const std::vector<std::optional<std::pair<unsigned, unsigned>>> + &ResultBounds) { CGBuilderTy &Builder = CGF.Builder; CodeGenModule &CGM = CGF.CGM; llvm::LLVMContext &CTX = CGF.getLLVMContext(); @@ -2685,18 +2686,20 @@ EmitAsmStores(CodeGenFunction &CGF, const AsmStmt &S, // ResultRegDests can be also populated by addReturnRegisterOutputs() above, // in which case its size may grow. assert(ResultTypeRequiresCast.size() <= ResultRegDests.size()); - assert(ResultRegIsFlagReg.size() <= ResultRegDests.size()); + assert(ResultBounds.size() <= ResultRegDests.size()); for (unsigned i = 0, e = RegResults.size(); i != e; ++i) { llvm::Value *Tmp = RegResults[i]; llvm::Type *TruncTy = ResultTruncRegTypes[i]; - if ((i < ResultRegIsFlagReg.size()) && ResultRegIsFlagReg[i]) { - // Target must guarantee the Value `Tmp` here is lowered to a boolean - // value. - llvm::Constant *Two = llvm::ConstantInt::get(Tmp->getType(), 2); + if ((i < ResultBounds.size()) && ResultBounds[i].has_value()) { + const auto [LowerBound, UpperBound] = ResultBounds[i].value(); + // FIXME: Support for nonzero lower bounds not yet implemented. + assert(LowerBound == 0 && "Output operand lower bound is not zero."); + llvm::Constant *UpperBoundConst = + llvm::ConstantInt::get(Tmp->getType(), UpperBound); llvm::Value *IsBooleanValue = - Builder.CreateCmp(llvm::CmpInst::ICMP_ULT, Tmp, Two); + Builder.CreateCmp(llvm::CmpInst::ICMP_ULT, Tmp, UpperBoundConst); llvm::Function *FnAssume = CGM.getIntrinsic(llvm::Intrinsic::assume); Builder.CreateCall(FnAssume, IsBooleanValue); } @@ -2825,7 +2828,7 @@ void CodeGenFunction::EmitAsmStmt(const AsmStmt &S) { std::vector<llvm::Type *> ArgElemTypes; std::vector<llvm::Value*> Args; llvm::BitVector ResultTypeRequiresCast; - llvm::BitVector ResultRegIsFlagReg; + std::vector<std::optional<std::pair<unsigned, unsigned>>> ResultBounds; // Keep track of inout constraints. std::string InOutConstraints; @@ -2883,8 +2886,7 @@ void CodeGenFunction::EmitAsmStmt(const AsmStmt &S) { ResultRegQualTys.push_back(QTy); ResultRegDests.push_back(Dest); - bool IsFlagReg = llvm::StringRef(OutputConstraint).starts_with("{@cc"); - ResultRegIsFlagReg.push_back(IsFlagReg); + ResultBounds.emplace_back(Info.getOutputOperandBounds()); llvm::Type *Ty = ConvertTypeForMem(QTy); const bool RequiresCast = Info.allowsRegister() && @@ -3231,7 +3233,7 @@ void CodeGenFunction::EmitAsmStmt(const AsmStmt &S) { EmitAsmStores(*this, S, RegResults, ResultRegTypes, ResultTruncRegTypes, ResultRegDests, ResultRegQualTys, ResultTypeRequiresCast, - ResultRegIsFlagReg); + ResultBounds); // If this is an asm goto with outputs, repeat EmitAsmStores, but with a // different insertion point; one for each indirect destination and with @@ -3242,7 +3244,7 @@ void CodeGenFunction::EmitAsmStmt(const AsmStmt &S) { Builder.SetInsertPoint(Succ, --(Succ->end())); EmitAsmStores(*this, S, CBRRegResults[Succ], ResultRegTypes, ResultTruncRegTypes, ResultRegDests, ResultRegQualTys, - ResultTypeRequiresCast, ResultRegIsFlagReg); + ResultTypeRequiresCast, ResultBounds); } } } diff --git a/clang/lib/CodeGen/CodeGenModule.cpp b/clang/lib/CodeGen/CodeGenModule.cpp index 8d019d4..c5eb14e 100644 --- a/clang/lib/CodeGen/CodeGenModule.cpp +++ b/clang/lib/CodeGen/CodeGenModule.cpp @@ -2348,7 +2348,7 @@ static QualType GeneralizeTransparentUnion(QualType Ty) { const RecordType *UT = Ty->getAsUnionType(); if (!UT) return Ty; - const RecordDecl *UD = UT->getOriginalDecl()->getDefinitionOrSelf(); + const RecordDecl *UD = UT->getDecl()->getDefinitionOrSelf(); if (!UD->hasAttr<TransparentUnionAttr>()) return Ty; for (const auto *it : UD->fields()) { @@ -4230,7 +4230,7 @@ void CodeGenModule::EmitGlobal(GlobalDecl GD) { static bool HasNonDllImportDtor(QualType T) { if (const auto *RT = T->getBaseElementTypeUnsafe()->getAsCanonical<RecordType>()) - if (auto *RD = dyn_cast<CXXRecordDecl>(RT->getOriginalDecl())) { + if (auto *RD = dyn_cast<CXXRecordDecl>(RT->getDecl())) { RD = RD->getDefinitionOrSelf(); if (RD->getDestructor() && !RD->getDestructor()->hasAttr<DLLImportAttr>()) return true; diff --git a/clang/lib/CodeGen/CodeGenTBAA.cpp b/clang/lib/CodeGen/CodeGenTBAA.cpp index f8c7d64..cd08f3e 100644 --- a/clang/lib/CodeGen/CodeGenTBAA.cpp +++ b/clang/lib/CodeGen/CodeGenTBAA.cpp @@ -310,7 +310,7 @@ llvm::MDNode *CodeGenTBAA::getTypeInfoHelper(const Type *Ty) { // This also covers anonymous structs and unions, which have a different // compatibility rule, but it doesn't matter because you can never have a // pointer to an anonymous struct or union. - if (!RT->getOriginalDecl()->getDeclName()) + if (!RT->getDecl()->getDeclName()) return getAnyPtr(PtrDepth); // For non-builtin types use the mangled name of the canonical type. @@ -332,7 +332,7 @@ llvm::MDNode *CodeGenTBAA::getTypeInfoHelper(const Type *Ty) { // Enum types are distinct types. In C++ they have "underlying types", // however they aren't related for TBAA. if (const EnumType *ETy = dyn_cast<EnumType>(Ty)) { - const EnumDecl *ED = ETy->getOriginalDecl()->getDefinitionOrSelf(); + const EnumDecl *ED = ETy->getDecl()->getDefinitionOrSelf(); if (!Features.CPlusPlus) return getTypeInfo(ED->getIntegerType()); @@ -433,7 +433,7 @@ CodeGenTBAA::CollectFields(uint64_t BaseOffset, llvm::MDBuilder::TBAAStructField(BaseOffset, Size, TBAATag)); return true; } - const RecordDecl *RD = TTy->getOriginalDecl()->getDefinition(); + const RecordDecl *RD = TTy->getDecl()->getDefinition(); if (RD->hasFlexibleArrayMember()) return false; @@ -514,7 +514,7 @@ CodeGenTBAA::getTBAAStructInfo(QualType QTy) { llvm::MDNode *CodeGenTBAA::getBaseTypeInfoHelper(const Type *Ty) { if (auto *TTy = dyn_cast<RecordType>(Ty)) { - const RecordDecl *RD = TTy->getOriginalDecl()->getDefinition(); + const RecordDecl *RD = TTy->getDecl()->getDefinition(); const ASTRecordLayout &Layout = Context.getASTRecordLayout(RD); using TBAAStructField = llvm::MDBuilder::TBAAStructField; SmallVector<TBAAStructField, 4> Fields; @@ -609,8 +609,7 @@ llvm::MDNode *CodeGenTBAA::getValidBaseTypeInfo(QualType QTy) { // First calculate the metadata, before recomputing the insertion point, as // the helper can recursively call us. llvm::MDNode *TypeNode = getBaseTypeInfoHelper(Ty); - LLVM_ATTRIBUTE_UNUSED auto inserted = - BaseTypeMetadataCache.insert({Ty, TypeNode}); + [[maybe_unused]] auto inserted = BaseTypeMetadataCache.insert({Ty, TypeNode}); assert(inserted.second && "BaseType metadata was already inserted"); return TypeNode; diff --git a/clang/lib/CodeGen/CodeGenTypes.cpp b/clang/lib/CodeGen/CodeGenTypes.cpp index 3ffe999..ea31195 100644 --- a/clang/lib/CodeGen/CodeGenTypes.cpp +++ b/clang/lib/CodeGen/CodeGenTypes.cpp @@ -373,8 +373,8 @@ llvm::Type *CodeGenTypes::ConvertType(QualType T) { } // RecordTypes are cached and processed specially. - if (const RecordType *RT = dyn_cast<RecordType>(Ty)) - return ConvertRecordDeclType(RT->getOriginalDecl()->getDefinitionOrSelf()); + if (const auto *RT = dyn_cast<RecordType>(Ty)) + return ConvertRecordDeclType(RT->getDecl()->getDefinitionOrSelf()); llvm::Type *CachedType = nullptr; auto TCI = TypeCache.find(Ty); diff --git a/clang/lib/CodeGen/ItaniumCXXABI.cpp b/clang/lib/CodeGen/ItaniumCXXABI.cpp index 7dc2eaf..9e195a9 100644 --- a/clang/lib/CodeGen/ItaniumCXXABI.cpp +++ b/clang/lib/CodeGen/ItaniumCXXABI.cpp @@ -3816,7 +3816,7 @@ static bool ShouldUseExternalRTTIDescriptor(CodeGenModule &CGM, if (const RecordType *RecordTy = dyn_cast<RecordType>(Ty)) { const CXXRecordDecl *RD = - cast<CXXRecordDecl>(RecordTy->getOriginalDecl())->getDefinitionOrSelf(); + cast<CXXRecordDecl>(RecordTy->getDecl())->getDefinitionOrSelf(); if (!RD->hasDefinition()) return false; @@ -3850,9 +3850,7 @@ static bool ShouldUseExternalRTTIDescriptor(CodeGenModule &CGM, /// IsIncompleteClassType - Returns whether the given record type is incomplete. static bool IsIncompleteClassType(const RecordType *RecordTy) { - return !RecordTy->getOriginalDecl() - ->getDefinitionOrSelf() - ->isCompleteDefinition(); + return !RecordTy->getDecl()->getDefinitionOrSelf()->isCompleteDefinition(); } /// ContainsIncompleteClassType - Returns whether the given type contains an @@ -3985,9 +3983,8 @@ void ItaniumRTTIBuilder::BuildVTablePointer(const Type *Ty, break; case Type::Record: { - const CXXRecordDecl *RD = - cast<CXXRecordDecl>(cast<RecordType>(Ty)->getOriginalDecl()) - ->getDefinitionOrSelf(); + const auto *RD = cast<CXXRecordDecl>(cast<RecordType>(Ty)->getDecl()) + ->getDefinitionOrSelf(); if (!RD->hasDefinition() || !RD->getNumBases()) { VTableName = ClassTypeInfo; @@ -4109,8 +4106,8 @@ static llvm::GlobalVariable::LinkageTypes getTypeInfoLinkage(CodeGenModule &CGM, return llvm::GlobalValue::LinkOnceODRLinkage; if (const RecordType *Record = dyn_cast<RecordType>(Ty)) { - const CXXRecordDecl *RD = - cast<CXXRecordDecl>(Record->getOriginalDecl())->getDefinitionOrSelf(); + const auto *RD = + cast<CXXRecordDecl>(Record->getDecl())->getDefinitionOrSelf(); if (RD->hasAttr<WeakAttr>()) return llvm::GlobalValue::WeakODRLinkage; if (CGM.getTriple().isWindowsItaniumEnvironment()) @@ -4273,9 +4270,8 @@ llvm::Constant *ItaniumRTTIBuilder::BuildTypeInfo( break; case Type::Record: { - const CXXRecordDecl *RD = - cast<CXXRecordDecl>(cast<RecordType>(Ty)->getOriginalDecl()) - ->getDefinitionOrSelf(); + const auto *RD = cast<CXXRecordDecl>(cast<RecordType>(Ty)->getDecl()) + ->getDefinitionOrSelf(); if (!RD->hasDefinition() || !RD->getNumBases()) { // We don't need to emit any fields. break; @@ -4322,8 +4318,8 @@ llvm::Constant *ItaniumRTTIBuilder::BuildTypeInfo( if (CGM.getTarget().hasPS4DLLImportExport() && GVDLLStorageClass != llvm::GlobalVariable::DLLExportStorageClass) { if (const RecordType *RecordTy = dyn_cast<RecordType>(Ty)) { - const CXXRecordDecl *RD = cast<CXXRecordDecl>(RecordTy->getOriginalDecl()) - ->getDefinitionOrSelf(); + const auto *RD = + cast<CXXRecordDecl>(RecordTy->getDecl())->getDefinitionOrSelf(); if (RD->hasAttr<DLLExportAttr>() || CXXRecordNonInlineHasAttr<DLLExportAttr>(RD)) GVDLLStorageClass = llvm::GlobalVariable::DLLExportStorageClass; diff --git a/clang/lib/CodeGen/SwiftCallingConv.cpp b/clang/lib/CodeGen/SwiftCallingConv.cpp index 4d894fd..20965430 100644 --- a/clang/lib/CodeGen/SwiftCallingConv.cpp +++ b/clang/lib/CodeGen/SwiftCallingConv.cpp @@ -66,7 +66,7 @@ void SwiftAggLowering::addTypedData(QualType type, CharUnits begin) { // Record types. if (auto recType = type->getAsCanonical<RecordType>()) { - addTypedData(recType->getOriginalDecl(), begin); + addTypedData(recType->getDecl(), begin); // Array types. } else if (type->isArrayType()) { @@ -814,7 +814,7 @@ static ABIArgInfo classifyType(CodeGenModule &CGM, CanQualType type, bool forReturn) { unsigned IndirectAS = CGM.getDataLayout().getAllocaAddrSpace(); if (auto recordType = dyn_cast<RecordType>(type)) { - auto record = recordType->getOriginalDecl(); + auto record = recordType->getDecl(); auto &layout = CGM.getContext().getASTRecordLayout(record); if (mustPassRecordIndirectly(CGM, record)) @@ -822,8 +822,7 @@ static ABIArgInfo classifyType(CodeGenModule &CGM, CanQualType type, /*AddrSpace=*/IndirectAS, /*byval=*/false); SwiftAggLowering lowering(CGM); - lowering.addTypedData(recordType->getOriginalDecl(), CharUnits::Zero(), - layout); + lowering.addTypedData(recordType->getDecl(), CharUnits::Zero(), layout); lowering.finish(); return classifyExpandedType(lowering, forReturn, layout.getAlignment(), diff --git a/clang/lib/CodeGen/TargetBuiltins/ARM.cpp b/clang/lib/CodeGen/TargetBuiltins/ARM.cpp index 82b71e3..2429a43 100644 --- a/clang/lib/CodeGen/TargetBuiltins/ARM.cpp +++ b/clang/lib/CodeGen/TargetBuiltins/ARM.cpp @@ -7795,7 +7795,7 @@ Value *CodeGenFunction::EmitAArch64BuiltinExpr(unsigned BuiltinID, } case NEON::BI__builtin_neon_vcvt1_low_bf16_mf8_fpm: ExtractLow = true; - LLVM_FALLTHROUGH; + [[fallthrough]]; case NEON::BI__builtin_neon_vcvt1_bf16_mf8_fpm: case NEON::BI__builtin_neon_vcvt1_high_bf16_mf8_fpm: return EmitFP8NeonCvtCall(Intrinsic::aarch64_neon_fp8_cvtl1, @@ -7803,7 +7803,7 @@ Value *CodeGenFunction::EmitAArch64BuiltinExpr(unsigned BuiltinID, Ops[0]->getType(), ExtractLow, Ops, E, "vbfcvt1"); case NEON::BI__builtin_neon_vcvt2_low_bf16_mf8_fpm: ExtractLow = true; - LLVM_FALLTHROUGH; + [[fallthrough]]; case NEON::BI__builtin_neon_vcvt2_bf16_mf8_fpm: case NEON::BI__builtin_neon_vcvt2_high_bf16_mf8_fpm: return EmitFP8NeonCvtCall(Intrinsic::aarch64_neon_fp8_cvtl2, @@ -7811,7 +7811,7 @@ Value *CodeGenFunction::EmitAArch64BuiltinExpr(unsigned BuiltinID, Ops[0]->getType(), ExtractLow, Ops, E, "vbfcvt2"); case NEON::BI__builtin_neon_vcvt1_low_f16_mf8_fpm: ExtractLow = true; - LLVM_FALLTHROUGH; + [[fallthrough]]; case NEON::BI__builtin_neon_vcvt1_f16_mf8_fpm: case NEON::BI__builtin_neon_vcvt1_high_f16_mf8_fpm: return EmitFP8NeonCvtCall(Intrinsic::aarch64_neon_fp8_cvtl1, @@ -7819,7 +7819,7 @@ Value *CodeGenFunction::EmitAArch64BuiltinExpr(unsigned BuiltinID, Ops[0]->getType(), ExtractLow, Ops, E, "vbfcvt1"); case NEON::BI__builtin_neon_vcvt2_low_f16_mf8_fpm: ExtractLow = true; - LLVM_FALLTHROUGH; + [[fallthrough]]; case NEON::BI__builtin_neon_vcvt2_f16_mf8_fpm: case NEON::BI__builtin_neon_vcvt2_high_f16_mf8_fpm: return EmitFP8NeonCvtCall(Intrinsic::aarch64_neon_fp8_cvtl2, @@ -7854,7 +7854,7 @@ Value *CodeGenFunction::EmitAArch64BuiltinExpr(unsigned BuiltinID, case NEON::BI__builtin_neon_vdot_lane_f16_mf8_fpm: case NEON::BI__builtin_neon_vdotq_lane_f16_mf8_fpm: ExtendLaneArg = true; - LLVM_FALLTHROUGH; + [[fallthrough]]; case NEON::BI__builtin_neon_vdot_laneq_f16_mf8_fpm: case NEON::BI__builtin_neon_vdotq_laneq_f16_mf8_fpm: return EmitFP8NeonFDOTCall(Intrinsic::aarch64_neon_fp8_fdot2_lane, @@ -7866,7 +7866,7 @@ Value *CodeGenFunction::EmitAArch64BuiltinExpr(unsigned BuiltinID, case NEON::BI__builtin_neon_vdot_lane_f32_mf8_fpm: case NEON::BI__builtin_neon_vdotq_lane_f32_mf8_fpm: ExtendLaneArg = true; - LLVM_FALLTHROUGH; + [[fallthrough]]; case NEON::BI__builtin_neon_vdot_laneq_f32_mf8_fpm: case NEON::BI__builtin_neon_vdotq_laneq_f32_mf8_fpm: return EmitFP8NeonFDOTCall(Intrinsic::aarch64_neon_fp8_fdot4_lane, @@ -7898,37 +7898,37 @@ Value *CodeGenFunction::EmitAArch64BuiltinExpr(unsigned BuiltinID, "vmlall"); case NEON::BI__builtin_neon_vmlalbq_lane_f16_mf8_fpm: ExtendLaneArg = true; - LLVM_FALLTHROUGH; + [[fallthrough]]; case NEON::BI__builtin_neon_vmlalbq_laneq_f16_mf8_fpm: return EmitFP8NeonFMLACall(Intrinsic::aarch64_neon_fp8_fmlalb_lane, ExtendLaneArg, HalfTy, Ops, E, "vmlal_lane"); case NEON::BI__builtin_neon_vmlaltq_lane_f16_mf8_fpm: ExtendLaneArg = true; - LLVM_FALLTHROUGH; + [[fallthrough]]; case NEON::BI__builtin_neon_vmlaltq_laneq_f16_mf8_fpm: return EmitFP8NeonFMLACall(Intrinsic::aarch64_neon_fp8_fmlalt_lane, ExtendLaneArg, HalfTy, Ops, E, "vmlal_lane"); case NEON::BI__builtin_neon_vmlallbbq_lane_f32_mf8_fpm: ExtendLaneArg = true; - LLVM_FALLTHROUGH; + [[fallthrough]]; case NEON::BI__builtin_neon_vmlallbbq_laneq_f32_mf8_fpm: return EmitFP8NeonFMLACall(Intrinsic::aarch64_neon_fp8_fmlallbb_lane, ExtendLaneArg, FloatTy, Ops, E, "vmlall_lane"); case NEON::BI__builtin_neon_vmlallbtq_lane_f32_mf8_fpm: ExtendLaneArg = true; - LLVM_FALLTHROUGH; + [[fallthrough]]; case NEON::BI__builtin_neon_vmlallbtq_laneq_f32_mf8_fpm: return EmitFP8NeonFMLACall(Intrinsic::aarch64_neon_fp8_fmlallbt_lane, ExtendLaneArg, FloatTy, Ops, E, "vmlall_lane"); case NEON::BI__builtin_neon_vmlalltbq_lane_f32_mf8_fpm: ExtendLaneArg = true; - LLVM_FALLTHROUGH; + [[fallthrough]]; case NEON::BI__builtin_neon_vmlalltbq_laneq_f32_mf8_fpm: return EmitFP8NeonFMLACall(Intrinsic::aarch64_neon_fp8_fmlalltb_lane, ExtendLaneArg, FloatTy, Ops, E, "vmlall_lane"); case NEON::BI__builtin_neon_vmlallttq_lane_f32_mf8_fpm: ExtendLaneArg = true; - LLVM_FALLTHROUGH; + [[fallthrough]]; case NEON::BI__builtin_neon_vmlallttq_laneq_f32_mf8_fpm: return EmitFP8NeonFMLACall(Intrinsic::aarch64_neon_fp8_fmlalltt_lane, ExtendLaneArg, FloatTy, Ops, E, "vmlall_lane"); diff --git a/clang/lib/CodeGen/TargetInfo.cpp b/clang/lib/CodeGen/TargetInfo.cpp index 1e58c3f..342a3af 100644 --- a/clang/lib/CodeGen/TargetInfo.cpp +++ b/clang/lib/CodeGen/TargetInfo.cpp @@ -82,6 +82,8 @@ TargetCodeGenInfo::~TargetCodeGenInfo() = default; // If someone can figure out a general rule for this, that would be great. // It's probably just doomed to be platform-dependent, though. unsigned TargetCodeGenInfo::getSizeOfUnwindException() const { + if (getABIInfo().getCodeGenOpts().hasSEHExceptions()) + return getABIInfo().getDataLayout().getPointerSizeInBits() > 32 ? 64 : 48; // Verified for: // x86-64 FreeBSD, Linux, Darwin // x86-32 FreeBSD, Linux, Darwin diff --git a/clang/lib/CodeGen/Targets/AArch64.cpp b/clang/lib/CodeGen/Targets/AArch64.cpp index d7deece..bb41a14 100644 --- a/clang/lib/CodeGen/Targets/AArch64.cpp +++ b/clang/lib/CodeGen/Targets/AArch64.cpp @@ -743,7 +743,7 @@ bool AArch64ABIInfo::passAsPureScalableType( return false; // Pure scalable types are never unions and never contain unions. - const RecordDecl *RD = RT->getOriginalDecl()->getDefinitionOrSelf(); + const RecordDecl *RD = RT->getDecl()->getDefinitionOrSelf(); if (RD->isUnion()) return false; diff --git a/clang/lib/CodeGen/Targets/AMDGPU.cpp b/clang/lib/CodeGen/Targets/AMDGPU.cpp index 0fcbf7e..16d5919 100644 --- a/clang/lib/CodeGen/Targets/AMDGPU.cpp +++ b/clang/lib/CodeGen/Targets/AMDGPU.cpp @@ -402,6 +402,26 @@ void AMDGPUTargetCodeGenInfo::setFunctionDeclAttributes( F->addFnAttr("amdgpu-max-num-workgroups", AttrVal.str()); } + + if (auto *Attr = FD->getAttr<CUDAClusterDimsAttr>()) { + auto GetExprVal = [&](const auto &E) { + return E ? E->EvaluateKnownConstInt(M.getContext()).getExtValue() : 1; + }; + unsigned X = GetExprVal(Attr->getX()); + unsigned Y = GetExprVal(Attr->getY()); + unsigned Z = GetExprVal(Attr->getZ()); + llvm::SmallString<32> AttrVal; + llvm::raw_svector_ostream OS(AttrVal); + OS << X << ',' << Y << ',' << Z; + F->addFnAttr("amdgpu-cluster-dims", AttrVal.str()); + } + + // OpenCL doesn't support cluster feature. + const TargetInfo &TTI = M.getContext().getTargetInfo(); + if ((IsOpenCLKernel && + TTI.hasFeatureEnabled(TTI.getTargetOpts().FeatureMap, "clusters")) || + FD->hasAttr<CUDANoClusterAttr>()) + F->addFnAttr("amdgpu-cluster-dims", "0,0,0"); } void AMDGPUTargetCodeGenInfo::setTargetAttributes( diff --git a/clang/lib/CodeGen/Targets/ARC.cpp b/clang/lib/CodeGen/Targets/ARC.cpp index 6727587..6c9444d 100644 --- a/clang/lib/CodeGen/Targets/ARC.cpp +++ b/clang/lib/CodeGen/Targets/ARC.cpp @@ -112,8 +112,7 @@ ABIArgInfo ARCABIInfo::classifyArgumentType(QualType Ty, if (isAggregateTypeForABI(Ty)) { // Structures with flexible arrays are always indirect. - if (RT && - RT->getOriginalDecl()->getDefinitionOrSelf()->hasFlexibleArrayMember()) + if (RT && RT->getDecl()->getDefinitionOrSelf()->hasFlexibleArrayMember()) return getIndirectByValue(Ty); // Ignore empty structs/unions. diff --git a/clang/lib/CodeGen/Targets/ARM.cpp b/clang/lib/CodeGen/Targets/ARM.cpp index c84c9f2..4d05217 100644 --- a/clang/lib/CodeGen/Targets/ARM.cpp +++ b/clang/lib/CodeGen/Targets/ARM.cpp @@ -516,7 +516,7 @@ static bool isIntegerLikeType(QualType Ty, ASTContext &Context, if (!RT) return false; // Ignore records with flexible arrays. - const RecordDecl *RD = RT->getOriginalDecl()->getDefinitionOrSelf(); + const RecordDecl *RD = RT->getDecl()->getDefinitionOrSelf(); if (RD->hasFlexibleArrayMember()) return false; diff --git a/clang/lib/CodeGen/Targets/Lanai.cpp b/clang/lib/CodeGen/Targets/Lanai.cpp index e76431a..871a135 100644 --- a/clang/lib/CodeGen/Targets/Lanai.cpp +++ b/clang/lib/CodeGen/Targets/Lanai.cpp @@ -102,8 +102,7 @@ ABIArgInfo LanaiABIInfo::classifyArgumentType(QualType Ty, if (isAggregateTypeForABI(Ty)) { // Structures with flexible arrays are always indirect. - if (RT && - RT->getOriginalDecl()->getDefinitionOrSelf()->hasFlexibleArrayMember()) + if (RT && RT->getDecl()->getDefinitionOrSelf()->hasFlexibleArrayMember()) return getIndirectResult(Ty, /*ByVal=*/true, State); // Ignore empty structs/unions. diff --git a/clang/lib/CodeGen/Targets/LoongArch.cpp b/clang/lib/CodeGen/Targets/LoongArch.cpp index 1f344d6..878723d 100644 --- a/clang/lib/CodeGen/Targets/LoongArch.cpp +++ b/clang/lib/CodeGen/Targets/LoongArch.cpp @@ -150,7 +150,7 @@ bool LoongArchABIInfo::detectFARsEligibleStructHelper( // Non-zero-length arrays of empty records make the struct ineligible to be // passed via FARs in C++. if (const auto *RTy = EltTy->getAsCanonical<RecordType>()) { - if (ArraySize != 0 && isa<CXXRecordDecl>(RTy->getOriginalDecl()) && + if (ArraySize != 0 && isa<CXXRecordDecl>(RTy->getDecl()) && isEmptyRecord(getContext(), EltTy, true, true)) return false; } @@ -169,7 +169,7 @@ bool LoongArchABIInfo::detectFARsEligibleStructHelper( // copy constructor are not eligible for the FP calling convention. if (getRecordArgABI(Ty, CGT.getCXXABI())) return false; - const RecordDecl *RD = RTy->getOriginalDecl()->getDefinitionOrSelf(); + const RecordDecl *RD = RTy->getDecl()->getDefinitionOrSelf(); if (isEmptyRecord(getContext(), Ty, true, true) && (!RD->isUnion() || !isa<CXXRecordDecl>(RD))) return true; diff --git a/clang/lib/CodeGen/Targets/Mips.cpp b/clang/lib/CodeGen/Targets/Mips.cpp index f26ab97..22fdcd9 100644 --- a/clang/lib/CodeGen/Targets/Mips.cpp +++ b/clang/lib/CodeGen/Targets/Mips.cpp @@ -161,7 +161,7 @@ llvm::Type* MipsABIInfo::HandleAggregates(QualType Ty, uint64_t TySize) const { return llvm::StructType::get(getVMContext(), ArgList); } - const RecordDecl *RD = RT->getOriginalDecl()->getDefinitionOrSelf(); + const RecordDecl *RD = RT->getDecl()->getDefinitionOrSelf(); const ASTRecordLayout &Layout = getContext().getASTRecordLayout(RD); assert(!(TySize % 8) && "Size of structure must be multiple of 8."); @@ -265,7 +265,7 @@ MipsABIInfo::returnAggregateInRegs(QualType RetTy, uint64_t Size) const { SmallVector<llvm::Type*, 8> RTList; if (RT && RT->isStructureOrClassType()) { - const RecordDecl *RD = RT->getOriginalDecl()->getDefinitionOrSelf(); + const RecordDecl *RD = RT->getDecl()->getDefinitionOrSelf(); const ASTRecordLayout &Layout = getContext().getASTRecordLayout(RD); unsigned FieldCnt = Layout.getFieldCount(); diff --git a/clang/lib/CodeGen/Targets/RISCV.cpp b/clang/lib/CodeGen/Targets/RISCV.cpp index 0ef39b6..d134589 100644 --- a/clang/lib/CodeGen/Targets/RISCV.cpp +++ b/clang/lib/CodeGen/Targets/RISCV.cpp @@ -234,7 +234,7 @@ bool RISCVABIInfo::detectFPCCEligibleStructHelper(QualType Ty, CharUnits CurOff, // Non-zero-length arrays of empty records make the struct ineligible for // the FP calling convention in C++. if (const auto *RTy = EltTy->getAsCanonical<RecordType>()) { - if (ArraySize != 0 && isa<CXXRecordDecl>(RTy->getOriginalDecl()) && + if (ArraySize != 0 && isa<CXXRecordDecl>(RTy->getDecl()) && isEmptyRecord(getContext(), EltTy, true, true)) return false; } @@ -256,7 +256,7 @@ bool RISCVABIInfo::detectFPCCEligibleStructHelper(QualType Ty, CharUnits CurOff, return false; if (isEmptyRecord(getContext(), Ty, true, true)) return true; - const RecordDecl *RD = RTy->getOriginalDecl()->getDefinitionOrSelf(); + const RecordDecl *RD = RTy->getDecl()->getDefinitionOrSelf(); // Unions aren't eligible unless they're empty (which is caught above). if (RD->isUnion()) return false; @@ -680,22 +680,22 @@ ABIArgInfo RISCVABIInfo::classifyArgumentType(QualType Ty, bool IsFixed, if (const auto *ED = Ty->getAsEnumDecl()) Ty = ED->getIntegerType(); - // All integral types are promoted to XLen width - if (Size < XLen && Ty->isIntegralOrEnumerationType()) { - return extendType(Ty, CGT.ConvertType(Ty)); - } - if (const auto *EIT = Ty->getAs<BitIntType>()) { - if (EIT->getNumBits() < XLen) + + if (XLen == 64 && EIT->getNumBits() == 32) return extendType(Ty, CGT.ConvertType(Ty)); - if (EIT->getNumBits() > 128 || - (!getContext().getTargetInfo().hasInt128Type() && - EIT->getNumBits() > 64)) - return getNaturalAlignIndirect( - Ty, /*AddrSpace=*/getDataLayout().getAllocaAddrSpace(), - /*ByVal=*/false); + + if (EIT->getNumBits() <= 2 * XLen) + return ABIArgInfo::getExtend(Ty, CGT.ConvertType(Ty)); + return getNaturalAlignIndirect( + Ty, /*AddrSpace=*/getDataLayout().getAllocaAddrSpace(), + /*ByVal=*/false); } + // All integral types are promoted to XLen width + if (Size < XLen && Ty->isIntegralOrEnumerationType()) + return extendType(Ty, CGT.ConvertType(Ty)); + return ABIArgInfo::getDirect(); } diff --git a/clang/lib/CodeGen/Targets/SystemZ.cpp b/clang/lib/CodeGen/Targets/SystemZ.cpp index 9b6b72b1..e50f06c 100644 --- a/clang/lib/CodeGen/Targets/SystemZ.cpp +++ b/clang/lib/CodeGen/Targets/SystemZ.cpp @@ -193,11 +193,11 @@ llvm::Type *SystemZABIInfo::getFPArgumentType(QualType Ty, case BuiltinType::Float16: if (Size == 16) return llvm::Type::getHalfTy(getVMContext()); - LLVM_FALLTHROUGH; + [[fallthrough]]; case BuiltinType::Float: if (Size == 32) return llvm::Type::getFloatTy(getVMContext()); - LLVM_FALLTHROUGH; + [[fallthrough]]; case BuiltinType::Double: return llvm::Type::getDoubleTy(getVMContext()); default: diff --git a/clang/lib/CodeGen/Targets/X86.cpp b/clang/lib/CodeGen/Targets/X86.cpp index fb78948..8daf8eb 100644 --- a/clang/lib/CodeGen/Targets/X86.cpp +++ b/clang/lib/CodeGen/Targets/X86.cpp @@ -795,8 +795,7 @@ ABIArgInfo X86_32ABIInfo::classifyArgumentType(QualType Ty, CCState &State, if (isAggregateTypeForABI(Ty)) { // Structures with flexible arrays are always indirect. // FIXME: This should not be byval! - if (RT && - RT->getOriginalDecl()->getDefinitionOrSelf()->hasFlexibleArrayMember()) + if (RT && RT->getDecl()->getDefinitionOrSelf()->hasFlexibleArrayMember()) return getIndirectResult(Ty, true, State); // Ignore empty structs/unions on non-Windows. @@ -831,7 +830,7 @@ ABIArgInfo X86_32ABIInfo::classifyArgumentType(QualType Ty, CCState &State, unsigned AlignInBits = 0; if (RT) { const ASTRecordLayout &Layout = - getContext().getASTRecordLayout(RT->getOriginalDecl()); + getContext().getASTRecordLayout(RT->getDecl()); AlignInBits = getContext().toBits(Layout.getRequiredAlignment()); } else if (TI.isAlignRequired()) { AlignInBits = TI.Align; @@ -2042,7 +2041,7 @@ void X86_64ABIInfo::classify(QualType Ty, uint64_t OffsetBase, Class &Lo, if (getRecordArgABI(RT, getCXXABI())) return; - const RecordDecl *RD = RT->getOriginalDecl()->getDefinitionOrSelf(); + const RecordDecl *RD = RT->getDecl()->getDefinitionOrSelf(); // Assume variable sized types are passed in memory. if (RD->hasFlexibleArrayMember()) @@ -2851,9 +2850,8 @@ ABIArgInfo X86_64ABIInfo::classifyRegCallStructTypeImpl(QualType Ty, unsigned &NeededInt, unsigned &NeededSSE, unsigned &MaxVectorWidth) const { - auto *RD = cast<RecordType>(Ty.getCanonicalType()) - ->getOriginalDecl() - ->getDefinitionOrSelf(); + auto *RD = + cast<RecordType>(Ty.getCanonicalType())->getDecl()->getDefinitionOrSelf(); if (RD->hasFlexibleArrayMember()) return getIndirectReturnResult(Ty); @@ -3313,7 +3311,7 @@ ABIArgInfo WinX86_64ABIInfo::classify(QualType Ty, unsigned &FreeSSERegs, RAA == CGCXXABI::RAA_DirectInMemory); } - if (RT->getOriginalDecl()->getDefinitionOrSelf()->hasFlexibleArrayMember()) + if (RT->getDecl()->getDefinitionOrSelf()->hasFlexibleArrayMember()) return getNaturalAlignIndirect(Ty, getDataLayout().getAllocaAddrSpace(), /*ByVal=*/false); } diff --git a/clang/lib/CodeGen/Targets/XCore.cpp b/clang/lib/CodeGen/Targets/XCore.cpp index ab01154..f9726ec 100644 --- a/clang/lib/CodeGen/Targets/XCore.cpp +++ b/clang/lib/CodeGen/Targets/XCore.cpp @@ -380,7 +380,7 @@ static bool appendRecordType(SmallStringEnc &Enc, const RecordType *RT, // We collect all encoded fields and order as necessary. bool IsRecursive = false; - const RecordDecl *RD = RT->getOriginalDecl()->getDefinition(); + const RecordDecl *RD = RT->getDecl()->getDefinition(); if (RD && !RD->field_empty()) { // An incomplete TypeString stub is placed in the cache for this RecordType // so that recursive calls to this RecordType will use it whilst building a @@ -429,7 +429,7 @@ static bool appendEnumType(SmallStringEnc &Enc, const EnumType *ET, Enc += "){"; // We collect all encoded enumerations and order them alphanumerically. - if (const EnumDecl *ED = ET->getOriginalDecl()->getDefinition()) { + if (const EnumDecl *ED = ET->getDecl()->getDefinition()) { SmallVector<FieldEncoding, 16> FE; for (auto I = ED->enumerator_begin(), E = ED->enumerator_end(); I != E; ++I) { diff --git a/clang/lib/Driver/Distro.cpp b/clang/lib/Driver/Distro.cpp index 8a5a9fc..838e087 100644 --- a/clang/lib/Driver/Distro.cpp +++ b/clang/lib/Driver/Distro.cpp @@ -92,6 +92,7 @@ static Distro::DistroType DetectLsbRelease(llvm::vfs::FileSystem &VFS) { .Case("oracular", Distro::UbuntuOracular) .Case("plucky", Distro::UbuntuPlucky) .Case("questing", Distro::UbuntuQuesting) + .Case("resolute", Distro::UbuntuResolute) .Default(Distro::UnknownDistro); return Version; } diff --git a/clang/lib/Driver/Job.cpp b/clang/lib/Driver/Job.cpp index 880e9e3..715429b 100644 --- a/clang/lib/Driver/Job.cpp +++ b/clang/lib/Driver/Job.cpp @@ -57,24 +57,25 @@ static bool skipArgs(const char *Flag, bool HaveCrashVFS, int &SkipNum, SkipNum = 2; // These flags are all of the form -Flag <Arg> and are treated as two // arguments. Therefore, we need to skip the flag and the next argument. - bool ShouldSkip = llvm::StringSwitch<bool>(Flag) - .Cases("-MF", "-MT", "-MQ", "-serialize-diagnostic-file", true) - .Cases("-o", "-dependency-file", true) - .Cases("-fdebug-compilation-dir", "-diagnostic-log-file", true) - .Cases("-dwarf-debug-flags", "-ivfsoverlay", true) - .Default(false); + bool ShouldSkip = + llvm::StringSwitch<bool>(Flag) + .Cases({"-MF", "-MT", "-MQ", "-serialize-diagnostic-file"}, true) + .Cases({"-o", "-dependency-file"}, true) + .Cases({"-fdebug-compilation-dir", "-diagnostic-log-file"}, true) + .Cases({"-dwarf-debug-flags", "-ivfsoverlay"}, true) + .Default(false); if (ShouldSkip) return true; // Some include flags shouldn't be skipped if we have a crash VFS IsInclude = llvm::StringSwitch<bool>(Flag) - .Cases("-include", "-header-include-file", true) - .Cases("-idirafter", "-internal-isystem", "-iwithprefix", true) - .Cases("-internal-externc-isystem", "-iprefix", true) - .Cases("-iwithprefixbefore", "-isystem", "-iquote", true) - .Cases("-isysroot", "-I", "-F", "-resource-dir", true) - .Cases("-internal-iframework", "-iframework", "-include-pch", true) + .Cases({"-include", "-header-include-file"}, true) + .Cases({"-idirafter", "-internal-isystem", "-iwithprefix"}, true) + .Cases({"-internal-externc-isystem", "-iprefix"}, true) + .Cases({"-iwithprefixbefore", "-isystem", "-iquote"}, true) + .Cases({"-isysroot", "-I", "-F", "-resource-dir"}, true) + .Cases({"-internal-iframework", "-iframework", "-include-pch"}, true) .Default(false); if (IsInclude) return !HaveCrashVFS; @@ -83,9 +84,9 @@ static bool skipArgs(const char *Flag, bool HaveCrashVFS, int &SkipNum, // These flags are all of the form -Flag and have no second argument. ShouldSkip = llvm::StringSwitch<bool>(Flag) - .Cases("-M", "-MM", "-MG", "-MP", "-MD", true) - .Case("-MMD", true) - .Default(false); + .Cases({"-M", "-MM", "-MG", "-MP", "-MD"}, true) + .Case("-MMD", true) + .Default(false); // Match found. SkipNum = 1; diff --git a/clang/lib/Driver/ToolChains/Arch/Mips.cpp b/clang/lib/Driver/ToolChains/Arch/Mips.cpp index 8787c82..bac8681 100644 --- a/clang/lib/Driver/ToolChains/Arch/Mips.cpp +++ b/clang/lib/Driver/ToolChains/Arch/Mips.cpp @@ -442,6 +442,8 @@ bool mips::hasCompactBranches(StringRef &CPU) { return llvm::StringSwitch<bool>(CPU) .Case("mips32r6", true) .Case("mips64r6", true) + .Case("i6400", true) + .Case("i6500", true) .Default(false); } diff --git a/clang/lib/Driver/ToolChains/Clang.cpp b/clang/lib/Driver/ToolChains/Clang.cpp index bf75573..a7310ba 100644 --- a/clang/lib/Driver/ToolChains/Clang.cpp +++ b/clang/lib/Driver/ToolChains/Clang.cpp @@ -3755,7 +3755,8 @@ static void RenderHLSLOptions(const ArgList &Args, ArgStringList &CmdArgs, options::OPT_hlsl_entrypoint, options::OPT_fdx_rootsignature_define, options::OPT_fdx_rootsignature_version, - options::OPT_fhlsl_spv_use_unknown_image_format}; + options::OPT_fhlsl_spv_use_unknown_image_format, + options::OPT_fhlsl_spv_enable_maximal_reconvergence}; if (!types::isHLSL(InputType)) return; for (const auto &Arg : ForwardedArguments) @@ -9214,8 +9215,9 @@ void LinkerWrapper::ConstructJob(Compilation &C, const JobAction &JA, options::OPT_nogpulibc)) { forAllAssociatedToolChains(C, JA, getToolChain(), [&](const ToolChain &TC) { // The device C library is only available for NVPTX and AMDGPU targets - // currently. - if (!TC.getTriple().isNVPTX() && !TC.getTriple().isAMDGPU()) + // and we only link it by default for OpenMP currently. + if ((!TC.getTriple().isNVPTX() && !TC.getTriple().isAMDGPU()) || + !JA.isHostOffloading(Action::OFK_OpenMP)) return; bool HasLibC = TC.getStdlibIncludePath().has_value(); if (HasLibC) { diff --git a/clang/lib/Driver/ToolChains/HLSL.cpp b/clang/lib/Driver/ToolChains/HLSL.cpp index 2869549..20a320e 100644 --- a/clang/lib/Driver/ToolChains/HLSL.cpp +++ b/clang/lib/Driver/ToolChains/HLSL.cpp @@ -191,23 +191,35 @@ void getSpirvExtOperand(StringRef SpvExtensionArg, raw_ostream &out) { // The extensions that are commented out are supported in DXC, but the SPIR-V // backend does not know about them yet. static const std::vector<StringRef> DxcSupportedExtensions = { - "SPV_KHR_16bit_storage", "SPV_KHR_device_group", - "SPV_KHR_fragment_shading_rate", "SPV_KHR_multiview", - "SPV_KHR_post_depth_coverage", "SPV_KHR_non_semantic_info", - "SPV_KHR_shader_draw_parameters", "SPV_KHR_ray_tracing", - "SPV_KHR_shader_clock", "SPV_EXT_demote_to_helper_invocation", - "SPV_EXT_descriptor_indexing", "SPV_EXT_fragment_fully_covered", + "SPV_KHR_16bit_storage", + "SPV_KHR_device_group", + "SPV_KHR_fragment_shading_rate", + "SPV_KHR_multiview", + "SPV_KHR_post_depth_coverage", + "SPV_KHR_non_semantic_info", + "SPV_KHR_shader_draw_parameters", + "SPV_KHR_ray_tracing", + "SPV_KHR_shader_clock", + "SPV_EXT_demote_to_helper_invocation", + "SPV_EXT_descriptor_indexing", + "SPV_EXT_fragment_fully_covered", "SPV_EXT_fragment_invocation_density", - "SPV_EXT_fragment_shader_interlock", "SPV_EXT_mesh_shader", - "SPV_EXT_shader_stencil_export", "SPV_EXT_shader_viewport_index_layer", + "SPV_EXT_fragment_shader_interlock", + "SPV_EXT_mesh_shader", + "SPV_EXT_shader_stencil_export", + "SPV_EXT_shader_viewport_index_layer", // "SPV_AMD_shader_early_and_late_fragment_tests", - "SPV_GOOGLE_hlsl_functionality1", "SPV_GOOGLE_user_type", - "SPV_KHR_ray_query", "SPV_EXT_shader_image_int64", - "SPV_KHR_fragment_shader_barycentric", "SPV_KHR_physical_storage_buffer", + "SPV_GOOGLE_hlsl_functionality1", + "SPV_GOOGLE_user_type", + "SPV_KHR_ray_query", + "SPV_EXT_shader_image_int64", + "SPV_KHR_fragment_shader_barycentric", + "SPV_KHR_physical_storage_buffer", "SPV_KHR_vulkan_memory_model", // "SPV_KHR_compute_shader_derivatives", - // "SPV_KHR_maximal_reconvergence", - "SPV_KHR_float_controls", "SPV_NV_shader_subgroup_partitioned", + "SPV_KHR_maximal_reconvergence", + "SPV_KHR_float_controls", + "SPV_NV_shader_subgroup_partitioned", // "SPV_KHR_quad_control" }; @@ -218,7 +230,6 @@ void getSpirvExtOperand(StringRef SpvExtensionArg, raw_ostream &out) { if (SpvExtensionArg.compare_insensitive("DXC") == 0) { bool first = true; - std::string Operand; for (StringRef E : DxcSupportedExtensions) { if (!first) out << ","; diff --git a/clang/lib/Driver/XRayArgs.cpp b/clang/lib/Driver/XRayArgs.cpp index ceed7cb..0325296 100644 --- a/clang/lib/Driver/XRayArgs.cpp +++ b/clang/lib/Driver/XRayArgs.cpp @@ -105,8 +105,9 @@ XRayArgs::XRayArgs(const ToolChain &TC, const ArgList &Args) { for (const auto &P : BundleParts) { // TODO: Automate the generation of the string case table. auto Valid = llvm::StringSwitch<bool>(P) - .Cases("none", "all", "function", "function-entry", - "function-exit", "custom", true) + .Cases({"none", "all", "function", "function-entry", + "function-exit", "custom"}, + true) .Default(false); if (!Valid) { diff --git a/clang/lib/ExtractAPI/DeclarationFragments.cpp b/clang/lib/ExtractAPI/DeclarationFragments.cpp index 541af6d..e5eda46 100644 --- a/clang/lib/ExtractAPI/DeclarationFragments.cpp +++ b/clang/lib/ExtractAPI/DeclarationFragments.cpp @@ -422,7 +422,7 @@ DeclarationFragments DeclarationFragmentsBuilder::getFragmentsForType( Fragments.append(getFragmentsForNNS(TagTy->getQualifier(), Context, After)); - const TagDecl *Decl = TagTy->getOriginalDecl(); + const TagDecl *Decl = TagTy->getDecl(); // Anonymous decl, skip this fragment. if (Decl->getName().empty()) return Fragments.append("{ ... }", diff --git a/clang/lib/ExtractAPI/TypedefUnderlyingTypeResolver.cpp b/clang/lib/ExtractAPI/TypedefUnderlyingTypeResolver.cpp index 5adbbc6..41e4e0c 100644 --- a/clang/lib/ExtractAPI/TypedefUnderlyingTypeResolver.cpp +++ b/clang/lib/ExtractAPI/TypedefUnderlyingTypeResolver.cpp @@ -26,7 +26,7 @@ TypedefUnderlyingTypeResolver::getUnderlyingTypeDecl(QualType Type) const { if (TypedefTy) TypeDecl = TypedefTy->getDecl(); if (const TagType *TagTy = Type->getAs<TagType>()) { - TypeDecl = TagTy->getOriginalDecl(); + TypeDecl = TagTy->getDecl(); } else if (const ObjCInterfaceType *ObjCITy = Type->getAs<ObjCInterfaceType>()) { TypeDecl = ObjCITy->getDecl(); diff --git a/clang/lib/Format/Format.cpp b/clang/lib/Format/Format.cpp index 686e541..093e88f 100644 --- a/clang/lib/Format/Format.cpp +++ b/clang/lib/Format/Format.cpp @@ -2184,8 +2184,9 @@ std::error_code parseConfiguration(llvm::MemoryBufferRef Config, Input >> Styles; if (Input.error()) return Input.error(); + if (Styles.empty()) + return make_error_code(ParseError::Success); - assert(!Styles.empty()); const auto StyleCount = Styles.size(); // Start from the second style as (only) the first one may be the default. diff --git a/clang/lib/Format/FormatToken.cpp b/clang/lib/Format/FormatToken.cpp index c2956a1..cb3fc1c 100644 --- a/clang/lib/Format/FormatToken.cpp +++ b/clang/lib/Format/FormatToken.cpp @@ -41,8 +41,7 @@ static constexpr std::array<StringRef, 14> QtPropertyKeywords = { bool FormatToken::isQtProperty() const { assert(llvm::is_sorted(QtPropertyKeywords)); - return std::binary_search(QtPropertyKeywords.begin(), - QtPropertyKeywords.end(), TokenText); + return llvm::binary_search(QtPropertyKeywords, TokenText); } // Sorted common C++ non-keyword types. diff --git a/clang/lib/Format/UnwrappedLineParser.cpp b/clang/lib/Format/UnwrappedLineParser.cpp index 2879743..5e2584e 100644 --- a/clang/lib/Format/UnwrappedLineParser.cpp +++ b/clang/lib/Format/UnwrappedLineParser.cpp @@ -60,7 +60,7 @@ void printLine(llvm::raw_ostream &OS, const UnwrappedLine &Line, OS << "\n"; } -LLVM_ATTRIBUTE_UNUSED static void printDebugInfo(const UnwrappedLine &Line) { +[[maybe_unused]] static void printDebugInfo(const UnwrappedLine &Line) { printLine(llvm::dbgs(), Line); } @@ -2569,6 +2569,12 @@ bool UnwrappedLineParser::parseBracedList(bool IsAngleBracket, bool IsEnum) { if (IsEnum && !Style.AllowShortEnumsOnASingleLine) addUnwrappedLine(); break; + case tok::kw_requires: { + auto *RequiresToken = FormatTok; + nextToken(); + parseRequiresExpression(RequiresToken); + break; + } default: nextToken(); break; diff --git a/clang/lib/Format/WhitespaceManager.cpp b/clang/lib/Format/WhitespaceManager.cpp index 54f366f..7348a3a 100644 --- a/clang/lib/Format/WhitespaceManager.cpp +++ b/clang/lib/Format/WhitespaceManager.cpp @@ -289,17 +289,20 @@ AlignTokenSequence(const FormatStyle &Style, unsigned Start, unsigned End, SmallVector<WhitespaceManager::Change, 16> &Changes) { int Shift = 0; - // ScopeStack keeps track of the current scope depth. It contains indices of - // the first token on each scope. + // ScopeStack keeps track of the current scope depth. It contains the levels + // of at most 2 scopes. The first one is the one that the matched token is + // in. The second one is the one that should not be moved by this procedure. // The "Matches" indices should only have tokens from the outer-most scope. // However, we do need to pay special attention to one class of tokens - // that are not in the outer-most scope, and that is function parameters - // which are split across multiple lines, as illustrated by this example: + // that are not in the outer-most scope, and that is the continuations of an + // unwrapped line whose positions are derived from a token to the right of the + // aligned token, as illustrated by this example: // double a(int x); // int b(int y, // double z); // In the above example, we need to take special care to ensure that - // 'double z' is indented along with it's owning function 'b'. + // 'double z' is indented along with its owning function 'b', because its + // position is derived from the '(' token to the right of the 'b' token. // The same holds for calling a function: // double a = foo(x); // int b = bar(foo(y), @@ -309,32 +312,28 @@ AlignTokenSequence(const FormatStyle &Style, unsigned Start, unsigned End, // auto s = "Hello" // "World"; // Special handling is required for 'nested' ternary operators. - SmallVector<unsigned, 16> ScopeStack; + SmallVector<std::tuple<unsigned, unsigned, unsigned>, 2> ScopeStack; for (unsigned i = Start; i != End; ++i) { auto &CurrentChange = Changes[i]; if (!Matches.empty() && Matches[0] < i) Matches.consume_front(); assert(Matches.empty() || Matches[0] >= i); - if (!ScopeStack.empty() && - CurrentChange.indentAndNestingLevel() < - Changes[ScopeStack.back()].indentAndNestingLevel()) { + while (!ScopeStack.empty() && + CurrentChange.indentAndNestingLevel() < ScopeStack.back()) { ScopeStack.pop_back(); } - // Compare current token to previous non-comment token to ensure whether - // it is in a deeper scope or not. - unsigned PreviousNonComment = i - 1; - while (PreviousNonComment > Start && - Changes[PreviousNonComment].Tok->is(tok::comment)) { - --PreviousNonComment; - } - if (i != Start && CurrentChange.indentAndNestingLevel() > - Changes[PreviousNonComment].indentAndNestingLevel()) { - ScopeStack.push_back(i); + // Keep track of the level that should not move with the aligned token. + if (ScopeStack.size() == 1u && CurrentChange.NewlinesBefore != 0u && + CurrentChange.indentAndNestingLevel() > ScopeStack[0] && + !CurrentChange.IsAligned) { + ScopeStack.push_back(CurrentChange.indentAndNestingLevel()); } - bool InsideNestedScope = !ScopeStack.empty(); + bool InsideNestedScope = + !ScopeStack.empty() && + CurrentChange.indentAndNestingLevel() > ScopeStack[0]; bool ContinuedStringLiteral = i > Start && CurrentChange.Tok->is(tok::string_literal) && Changes[i - 1].Tok->is(tok::string_literal); @@ -349,103 +348,20 @@ AlignTokenSequence(const FormatStyle &Style, unsigned Start, unsigned End, if (!Matches.empty() && Matches[0] == i) { Shift = Column - (RightJustify ? CurrentChange.TokenLength : 0) - CurrentChange.StartOfTokenColumn; + ScopeStack = {CurrentChange.indentAndNestingLevel()}; CurrentChange.Spaces += Shift; } if (Shift == 0) continue; - // This is for function parameters that are split across multiple lines, - // as mentioned in the ScopeStack comment. - if (InsideNestedScope && CurrentChange.NewlinesBefore > 0) { - unsigned ScopeStart = ScopeStack.back(); - auto ShouldShiftBeAdded = [&] { - // Function declaration - if (Changes[ScopeStart - 1].Tok->is(TT_FunctionDeclarationName)) - return true; - - // Lambda. - if (Changes[ScopeStart - 1].Tok->is(TT_LambdaLBrace)) - return false; - - // Continued function declaration - if (ScopeStart > Start + 1 && - Changes[ScopeStart - 2].Tok->is(TT_FunctionDeclarationName)) { - return true; - } - - // Continued (template) function call. - if (ScopeStart > Start + 1 && - Changes[ScopeStart - 2].Tok->isOneOf(tok::identifier, - TT_TemplateCloser) && - Changes[ScopeStart - 1].Tok->is(tok::l_paren) && - Changes[ScopeStart].Tok->isNot(TT_LambdaLSquare)) { - if (CurrentChange.Tok->MatchingParen && - CurrentChange.Tok->MatchingParen->is(TT_LambdaLBrace)) { - return false; - } - if (Changes[ScopeStart].NewlinesBefore > 0) - return false; - if (CurrentChange.Tok->is(tok::l_brace) && - CurrentChange.Tok->is(BK_BracedInit)) { - return true; - } - return Style.BinPackArguments; - } - - // Ternary operator - if (CurrentChange.Tok->is(TT_ConditionalExpr)) - return true; - - // Period Initializer .XXX = 1. - if (CurrentChange.Tok->is(TT_DesignatedInitializerPeriod)) - return true; - - // Continued ternary operator - if (CurrentChange.Tok->Previous && - CurrentChange.Tok->Previous->is(TT_ConditionalExpr)) { - return true; - } - - // Continued direct-list-initialization using braced list. - if (ScopeStart > Start + 1 && - Changes[ScopeStart - 2].Tok->is(tok::identifier) && - Changes[ScopeStart - 1].Tok->is(tok::l_brace) && - CurrentChange.Tok->is(tok::l_brace) && - CurrentChange.Tok->is(BK_BracedInit)) { - return true; - } - - // Continued braced list. - if (ScopeStart > Start + 1 && - Changes[ScopeStart - 2].Tok->isNot(tok::identifier) && - Changes[ScopeStart - 1].Tok->is(tok::l_brace) && - CurrentChange.Tok->isNot(tok::r_brace)) { - for (unsigned OuterScopeStart : llvm::reverse(ScopeStack)) { - // Lambda. - if (OuterScopeStart > Start && - Changes[OuterScopeStart - 1].Tok->is(TT_LambdaLBrace)) { - return false; - } - } - if (Changes[ScopeStart].NewlinesBefore > 0) - return false; - return true; - } - - // Continued template parameter. - if (Changes[ScopeStart - 1].Tok->is(TT_TemplateOpener)) - return true; - - return false; - }; - - if (ShouldShiftBeAdded()) - CurrentChange.Spaces += Shift; - } - - if (ContinuedStringLiteral) + // This is for lines that are split across multiple lines, as mentioned in + // the ScopeStack comment. The stack size being 1 means that the token is + // not in a scope that should not move. + if (ScopeStack.size() == 1u && CurrentChange.NewlinesBefore > 0 && + (ContinuedStringLiteral || InsideNestedScope)) { CurrentChange.Spaces += Shift; + } // We should not remove required spaces unless we break the line before. assert(Shift > 0 || Changes[i].NewlinesBefore > 0 || diff --git a/clang/lib/Frontend/CompilerInvocation.cpp b/clang/lib/Frontend/CompilerInvocation.cpp index 292adce..5bd15f5 100644 --- a/clang/lib/Frontend/CompilerInvocation.cpp +++ b/clang/lib/Frontend/CompilerInvocation.cpp @@ -4012,13 +4012,13 @@ bool CompilerInvocation::ParseLangArgs(LangOptions &Opts, ArgList &Args, auto Diag = Diags.Report(diag::note_drv_use_standard); Diag << Std.getName() << Std.getDescription(); unsigned NumAliases = 0; -#define LANGSTANDARD(id, name, lang, desc, features) +#define LANGSTANDARD(id, name, lang, desc, features, version) #define LANGSTANDARD_ALIAS(id, alias) \ if (KindValue == LangStandard::lang_##id) ++NumAliases; #define LANGSTANDARD_ALIAS_DEPR(id, alias) #include "clang/Basic/LangStandards.def" Diag << NumAliases; -#define LANGSTANDARD(id, name, lang, desc, features) +#define LANGSTANDARD(id, name, lang, desc, features, version) #define LANGSTANDARD_ALIAS(id, alias) \ if (KindValue == LangStandard::lang_##id) Diag << alias; #define LANGSTANDARD_ALIAS_DEPR(id, alias) diff --git a/clang/lib/Frontend/FrontendOptions.cpp b/clang/lib/Frontend/FrontendOptions.cpp index 32ed995..fb178b6 100644 --- a/clang/lib/Frontend/FrontendOptions.cpp +++ b/clang/lib/Frontend/FrontendOptions.cpp @@ -14,25 +14,26 @@ using namespace clang; InputKind FrontendOptions::getInputKindForExtension(StringRef Extension) { return llvm::StringSwitch<InputKind>(Extension) - .Cases("ast", "pcm", InputKind(Language::Unknown, InputKind::Precompiled)) + .Cases({"ast", "pcm"}, + InputKind(Language::Unknown, InputKind::Precompiled)) .Case("c", Language::C) - .Cases("S", "s", Language::Asm) + .Cases({"S", "s"}, Language::Asm) .Case("i", InputKind(Language::C).getPreprocessed()) .Case("ii", InputKind(Language::CXX).getPreprocessed()) .Case("cui", InputKind(Language::CUDA).getPreprocessed()) .Case("m", Language::ObjC) .Case("mi", InputKind(Language::ObjC).getPreprocessed()) - .Cases("mm", "M", Language::ObjCXX) + .Cases({"mm", "M"}, Language::ObjCXX) .Case("mii", InputKind(Language::ObjCXX).getPreprocessed()) - .Cases("C", "cc", "cp", Language::CXX) - .Cases("cpp", "CPP", "c++", "cxx", "hpp", "hxx", Language::CXX) + .Cases({"C", "cc", "cp"}, Language::CXX) + .Cases({"cpp", "CPP", "c++", "cxx", "hpp", "hxx"}, Language::CXX) .Case("cppm", Language::CXX) - .Cases("iim", "iih", InputKind(Language::CXX).getPreprocessed()) + .Cases({"iim", "iih"}, InputKind(Language::CXX).getPreprocessed()) .Case("cl", Language::OpenCL) .Case("clcpp", Language::OpenCLCXX) - .Cases("cu", "cuh", Language::CUDA) + .Cases({"cu", "cuh"}, Language::CUDA) .Case("hip", Language::HIP) - .Cases("ll", "bc", Language::LLVM_IR) + .Cases({"ll", "bc"}, Language::LLVM_IR) .Case("hlsl", Language::HLSL) .Case("cir", Language::CIR) .Default(Language::Unknown); diff --git a/clang/lib/Frontend/InitPreprocessor.cpp b/clang/lib/Frontend/InitPreprocessor.cpp index b899fb9..baad6317 100644 --- a/clang/lib/Frontend/InitPreprocessor.cpp +++ b/clang/lib/Frontend/InitPreprocessor.cpp @@ -459,43 +459,12 @@ static void InitializeStandardPredefinedMacros(const TargetInfo &TI, // value is, are implementation-defined. // (Removed in C++20.) if (!LangOpts.CPlusPlus) { - if (LangOpts.C2y) - Builder.defineMacro("__STDC_VERSION__", "202400L"); - else if (LangOpts.C23) - Builder.defineMacro("__STDC_VERSION__", "202311L"); - else if (LangOpts.C17) - Builder.defineMacro("__STDC_VERSION__", "201710L"); - else if (LangOpts.C11) - Builder.defineMacro("__STDC_VERSION__", "201112L"); - else if (LangOpts.C99) - Builder.defineMacro("__STDC_VERSION__", "199901L"); - else if (!LangOpts.GNUMode && LangOpts.Digraphs) - Builder.defineMacro("__STDC_VERSION__", "199409L"); + if (std::optional<uint32_t> Lang = LangOpts.getCLangStd()) + Builder.defineMacro("__STDC_VERSION__", Twine(*Lang) + "L"); } else { // -- __cplusplus - if (LangOpts.CPlusPlus26) - // FIXME: Use correct value for C++26. - Builder.defineMacro("__cplusplus", "202400L"); - else if (LangOpts.CPlusPlus23) - Builder.defineMacro("__cplusplus", "202302L"); - // [C++20] The integer literal 202002L. - else if (LangOpts.CPlusPlus20) - Builder.defineMacro("__cplusplus", "202002L"); - // [C++17] The integer literal 201703L. - else if (LangOpts.CPlusPlus17) - Builder.defineMacro("__cplusplus", "201703L"); - // [C++14] The name __cplusplus is defined to the value 201402L when - // compiling a C++ translation unit. - else if (LangOpts.CPlusPlus14) - Builder.defineMacro("__cplusplus", "201402L"); - // [C++11] The name __cplusplus is defined to the value 201103L when - // compiling a C++ translation unit. - else if (LangOpts.CPlusPlus11) - Builder.defineMacro("__cplusplus", "201103L"); - // [C++03] The name __cplusplus is defined to the value 199711L when - // compiling a C++ translation unit. - else - Builder.defineMacro("__cplusplus", "199711L"); + Builder.defineMacro("__cplusplus", + Twine(*LangOpts.getCPlusPlusLangStd()) + "L"); // -- __STDCPP_DEFAULT_NEW_ALIGNMENT__ // [C++17] An integer literal of type std::size_t whose value is the diff --git a/clang/lib/Frontend/Rewrite/RewriteModernObjC.cpp b/clang/lib/Frontend/Rewrite/RewriteModernObjC.cpp index 42f2d65..dee29fe 100644 --- a/clang/lib/Frontend/Rewrite/RewriteModernObjC.cpp +++ b/clang/lib/Frontend/Rewrite/RewriteModernObjC.cpp @@ -852,7 +852,7 @@ RewriteModernObjC::getIvarAccessString(ObjCIvarDecl *D) { IvarT = GetGroupRecordTypeForObjCIvarBitfield(D); if (!IvarT->getAs<TypedefType>() && IvarT->isRecordType()) { - RecordDecl *RD = IvarT->castAsCanonical<RecordType>()->getOriginalDecl(); + RecordDecl *RD = IvarT->castAsCanonical<RecordType>()->getDecl(); RD = RD->getDefinition(); if (RD && !RD->getDeclName().getAsIdentifierInfo()) { // decltype(((Foo_IMPL*)0)->bar) * @@ -7453,8 +7453,7 @@ Stmt *RewriteModernObjC::RewriteObjCIvarRefExpr(ObjCIvarRefExpr *IV) { IvarT = GetGroupRecordTypeForObjCIvarBitfield(D); if (!IvarT->getAs<TypedefType>() && IvarT->isRecordType()) { - RecordDecl *RD = - IvarT->castAsCanonical<RecordType>()->getOriginalDecl(); + RecordDecl *RD = IvarT->castAsCanonical<RecordType>()->getDecl(); RD = RD->getDefinition(); if (RD && !RD->getDeclName().getAsIdentifierInfo()) { // decltype(((Foo_IMPL*)0)->bar) * diff --git a/clang/lib/Headers/__clang_hip_runtime_wrapper.h b/clang/lib/Headers/__clang_hip_runtime_wrapper.h index da1e39a..fb0ece9 100644 --- a/clang/lib/Headers/__clang_hip_runtime_wrapper.h +++ b/clang/lib/Headers/__clang_hip_runtime_wrapper.h @@ -25,6 +25,8 @@ #define __constant__ __attribute__((constant)) #define __managed__ __attribute__((managed)) +#define __cluster_dims__(...) __attribute__((cluster_dims(__VA_ARGS__))) + #if !defined(__cplusplus) || __cplusplus < 201103L #define nullptr NULL; #endif diff --git a/clang/lib/Headers/avx2intrin.h b/clang/lib/Headers/avx2intrin.h index 4aaca2d..fa7f4c2 100644 --- a/clang/lib/Headers/avx2intrin.h +++ b/clang/lib/Headers/avx2intrin.h @@ -834,10 +834,9 @@ _mm256_cmpgt_epi64(__m256i __a, __m256i __b) /// \param __b /// A 256-bit vector of [16 x i16] containing one of the source operands. /// \returns A 256-bit vector of [16 x i16] containing the sums. -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_hadd_epi16(__m256i __a, __m256i __b) -{ - return (__m256i)__builtin_ia32_phaddw256((__v16hi)__a, (__v16hi)__b); +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_hadd_epi16(__m256i __a, __m256i __b) { + return (__m256i)__builtin_ia32_phaddw256((__v16hi)__a, (__v16hi)__b); } /// Horizontally adds the adjacent pairs of 32-bit integers from two 256-bit @@ -866,10 +865,9 @@ _mm256_hadd_epi16(__m256i __a, __m256i __b) /// \param __b /// A 256-bit vector of [8 x i32] containing one of the source operands. /// \returns A 256-bit vector of [8 x i32] containing the sums. -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_hadd_epi32(__m256i __a, __m256i __b) -{ - return (__m256i)__builtin_ia32_phaddd256((__v8si)__a, (__v8si)__b); +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_hadd_epi32(__m256i __a, __m256i __b) { + return (__m256i)__builtin_ia32_phaddd256((__v8si)__a, (__v8si)__b); } /// Horizontally adds the adjacent pairs of 16-bit integers from two 256-bit @@ -901,10 +899,9 @@ _mm256_hadd_epi32(__m256i __a, __m256i __b) /// \param __b /// A 256-bit vector of [16 x i16] containing one of the source operands. /// \returns A 256-bit vector of [16 x i16] containing the sums. -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_hadds_epi16(__m256i __a, __m256i __b) -{ - return (__m256i)__builtin_ia32_phaddsw256((__v16hi)__a, (__v16hi)__b); +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_hadds_epi16(__m256i __a, __m256i __b) { + return (__m256i)__builtin_ia32_phaddsw256((__v16hi)__a, (__v16hi)__b); } /// Horizontally subtracts adjacent pairs of 16-bit integers from two 256-bit @@ -937,10 +934,9 @@ _mm256_hadds_epi16(__m256i __a, __m256i __b) /// \param __b /// A 256-bit vector of [16 x i16] containing one of the source operands. /// \returns A 256-bit vector of [16 x i16] containing the differences. -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_hsub_epi16(__m256i __a, __m256i __b) -{ - return (__m256i)__builtin_ia32_phsubw256((__v16hi)__a, (__v16hi)__b); +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_hsub_epi16(__m256i __a, __m256i __b) { + return (__m256i)__builtin_ia32_phsubw256((__v16hi)__a, (__v16hi)__b); } /// Horizontally subtracts adjacent pairs of 32-bit integers from two 256-bit @@ -969,10 +965,9 @@ _mm256_hsub_epi16(__m256i __a, __m256i __b) /// \param __b /// A 256-bit vector of [8 x i32] containing one of the source operands. /// \returns A 256-bit vector of [8 x i32] containing the differences. -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_hsub_epi32(__m256i __a, __m256i __b) -{ - return (__m256i)__builtin_ia32_phsubd256((__v8si)__a, (__v8si)__b); +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_hsub_epi32(__m256i __a, __m256i __b) { + return (__m256i)__builtin_ia32_phsubd256((__v8si)__a, (__v8si)__b); } /// Horizontally subtracts adjacent pairs of 16-bit integers from two 256-bit @@ -1005,10 +1000,9 @@ _mm256_hsub_epi32(__m256i __a, __m256i __b) /// \param __b /// A 256-bit vector of [16 x i16] containing one of the source operands. /// \returns A 256-bit vector of [16 x i16] containing the differences. -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_hsubs_epi16(__m256i __a, __m256i __b) -{ - return (__m256i)__builtin_ia32_phsubsw256((__v16hi)__a, (__v16hi)__b); +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_hsubs_epi16(__m256i __a, __m256i __b) { + return (__m256i)__builtin_ia32_phsubsw256((__v16hi)__a, (__v16hi)__b); } /// Multiplies each unsigned byte from the 256-bit integer vector in \a __a @@ -1858,9 +1852,8 @@ _mm256_sad_epu8(__m256i __a, __m256i __b) /// control byte specify the index (within the same 128-bit half) of \a __a /// to copy to the result byte. /// \returns A 256-bit integer vector containing the result. -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_shuffle_epi8(__m256i __a, __m256i __b) -{ +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_shuffle_epi8(__m256i __a, __m256i __b) { return (__m256i)__builtin_ia32_pshufb256((__v32qi)__a, (__v32qi)__b); } diff --git a/clang/lib/Headers/avx512bwintrin.h b/clang/lib/Headers/avx512bwintrin.h index 473fe94..23b2d29 100644 --- a/clang/lib/Headers/avx512bwintrin.h +++ b/clang/lib/Headers/avx512bwintrin.h @@ -866,23 +866,20 @@ _mm512_mask_min_epu16(__m512i __W, __mmask32 __M, __m512i __A, __m512i __B) { (__v32hi)__W); } -static __inline__ __m512i __DEFAULT_FN_ATTRS512 -_mm512_shuffle_epi8(__m512i __A, __m512i __B) -{ +static __inline__ __m512i __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_shuffle_epi8(__m512i __A, __m512i __B) { return (__m512i)__builtin_ia32_pshufb512((__v64qi)__A,(__v64qi)__B); } -static __inline__ __m512i __DEFAULT_FN_ATTRS512 -_mm512_mask_shuffle_epi8(__m512i __W, __mmask64 __U, __m512i __A, __m512i __B) -{ +static __inline__ __m512i __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_mask_shuffle_epi8(__m512i __W, __mmask64 __U, __m512i __A, __m512i __B) { return (__m512i)__builtin_ia32_selectb_512((__mmask64)__U, (__v64qi)_mm512_shuffle_epi8(__A, __B), (__v64qi)__W); } -static __inline__ __m512i __DEFAULT_FN_ATTRS512 -_mm512_maskz_shuffle_epi8(__mmask64 __U, __m512i __A, __m512i __B) -{ +static __inline__ __m512i __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_maskz_shuffle_epi8(__mmask64 __U, __m512i __A, __m512i __B) { return (__m512i)__builtin_ia32_selectb_512((__mmask64)__U, (__v64qi)_mm512_shuffle_epi8(__A, __B), (__v64qi)_mm512_setzero_si512()); diff --git a/clang/lib/Headers/avx512cdintrin.h b/clang/lib/Headers/avx512cdintrin.h index 8899298..fb6dcb6 100644 --- a/clang/lib/Headers/avx512cdintrin.h +++ b/clang/lib/Headers/avx512cdintrin.h @@ -15,94 +15,82 @@ #define __AVX512CDINTRIN_H /* Define the default attributes for the functions in this file. */ +#if defined(__cplusplus) && (__cplusplus >= 201103L) +#define __DEFAULT_FN_ATTRS \ + constexpr __attribute__((__always_inline__, __nodebug__, \ + __target__("avx512cd"), __min_vector_width__(512))) +#else #define __DEFAULT_FN_ATTRS \ __attribute__((__always_inline__, __nodebug__, __target__("avx512cd"), \ __min_vector_width__(512))) - -#if defined(__cplusplus) && (__cplusplus >= 201103L) -#define __DEFAULT_FN_ATTRS_CONSTEXPR __DEFAULT_FN_ATTRS constexpr -#else -#define __DEFAULT_FN_ATTRS_CONSTEXPR __DEFAULT_FN_ATTRS #endif static __inline__ __m512i __DEFAULT_FN_ATTRS -_mm512_conflict_epi64 (__m512i __A) -{ - return (__m512i) __builtin_ia32_vpconflictdi_512 ((__v8di) __A); +_mm512_conflict_epi64(__m512i __A) { + return (__m512i)__builtin_ia32_vpconflictdi_512((__v8di)__A); } static __inline__ __m512i __DEFAULT_FN_ATTRS -_mm512_mask_conflict_epi64 (__m512i __W, __mmask8 __U, __m512i __A) -{ - return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U, - (__v8di)_mm512_conflict_epi64(__A), - (__v8di)__W); +_mm512_mask_conflict_epi64(__m512i __W, __mmask8 __U, __m512i __A) { + return (__m512i)__builtin_ia32_selectq_512( + (__mmask8)__U, (__v8di)_mm512_conflict_epi64(__A), (__v8di)__W); } static __inline__ __m512i __DEFAULT_FN_ATTRS -_mm512_maskz_conflict_epi64 (__mmask8 __U, __m512i __A) -{ +_mm512_maskz_conflict_epi64(__mmask8 __U, __m512i __A) { return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U, (__v8di)_mm512_conflict_epi64(__A), - (__v8di)_mm512_setzero_si512 ()); + (__v8di)_mm512_setzero_si512()); } static __inline__ __m512i __DEFAULT_FN_ATTRS -_mm512_conflict_epi32 (__m512i __A) -{ - return (__m512i) __builtin_ia32_vpconflictsi_512 ((__v16si) __A); +_mm512_conflict_epi32(__m512i __A) { + return (__m512i)__builtin_ia32_vpconflictsi_512((__v16si)__A); } static __inline__ __m512i __DEFAULT_FN_ATTRS -_mm512_mask_conflict_epi32 (__m512i __W, __mmask16 __U, __m512i __A) -{ - return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U, - (__v16si)_mm512_conflict_epi32(__A), - (__v16si)__W); +_mm512_mask_conflict_epi32(__m512i __W, __mmask16 __U, __m512i __A) { + return (__m512i)__builtin_ia32_selectd_512( + (__mmask16)__U, (__v16si)_mm512_conflict_epi32(__A), (__v16si)__W); } static __inline__ __m512i __DEFAULT_FN_ATTRS -_mm512_maskz_conflict_epi32 (__mmask16 __U, __m512i __A) -{ - return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U, - (__v16si)_mm512_conflict_epi32(__A), - (__v16si)_mm512_setzero_si512()); +_mm512_maskz_conflict_epi32(__mmask16 __U, __m512i __A) { + return (__m512i)__builtin_ia32_selectd_512( + (__mmask16)__U, (__v16si)_mm512_conflict_epi32(__A), + (__v16si)_mm512_setzero_si512()); } -static __inline__ __m512i __DEFAULT_FN_ATTRS_CONSTEXPR -_mm512_lzcnt_epi32(__m512i __A) { +static __inline__ __m512i __DEFAULT_FN_ATTRS _mm512_lzcnt_epi32(__m512i __A) { return (__m512i)__builtin_elementwise_clzg((__v16si)__A, (__v16si)_mm512_set1_epi32(32)); } -static __inline__ __m512i __DEFAULT_FN_ATTRS_CONSTEXPR +static __inline__ __m512i __DEFAULT_FN_ATTRS _mm512_mask_lzcnt_epi32(__m512i __W, __mmask16 __U, __m512i __A) { - return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U, - (__v16si)_mm512_lzcnt_epi32(__A), - (__v16si)__W); + return (__m512i)__builtin_ia32_selectd_512( + (__mmask16)__U, (__v16si)_mm512_lzcnt_epi32(__A), (__v16si)__W); } -static __inline__ __m512i __DEFAULT_FN_ATTRS_CONSTEXPR +static __inline__ __m512i __DEFAULT_FN_ATTRS _mm512_maskz_lzcnt_epi32(__mmask16 __U, __m512i __A) { return (__m512i)__builtin_ia32_selectd_512((__mmask16)__U, (__v16si)_mm512_lzcnt_epi32(__A), (__v16si)_mm512_setzero_si512()); } -static __inline__ __m512i __DEFAULT_FN_ATTRS_CONSTEXPR -_mm512_lzcnt_epi64(__m512i __A) { +static __inline__ __m512i __DEFAULT_FN_ATTRS _mm512_lzcnt_epi64(__m512i __A) { return (__m512i)__builtin_elementwise_clzg( (__v8di)__A, (__v8di)_mm512_set1_epi64((long long)64)); } -static __inline__ __m512i __DEFAULT_FN_ATTRS_CONSTEXPR +static __inline__ __m512i __DEFAULT_FN_ATTRS _mm512_mask_lzcnt_epi64(__m512i __W, __mmask8 __U, __m512i __A) { - return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U, - (__v8di)_mm512_lzcnt_epi64(__A), - (__v8di)__W); + return (__m512i)__builtin_ia32_selectq_512( + (__mmask8)__U, (__v8di)_mm512_lzcnt_epi64(__A), (__v8di)__W); } -static __inline__ __m512i __DEFAULT_FN_ATTRS_CONSTEXPR +static __inline__ __m512i __DEFAULT_FN_ATTRS _mm512_maskz_lzcnt_epi64(__mmask8 __U, __m512i __A) { return (__m512i)__builtin_ia32_selectq_512((__mmask8)__U, (__v8di)_mm512_lzcnt_epi64(__A), @@ -110,19 +98,15 @@ _mm512_maskz_lzcnt_epi64(__mmask8 __U, __m512i __A) { } static __inline__ __m512i __DEFAULT_FN_ATTRS -_mm512_broadcastmb_epi64 (__mmask8 __A) -{ - return (__m512i) _mm512_set1_epi64((long long) __A); +_mm512_broadcastmb_epi64(__mmask8 __A) { + return (__m512i)_mm512_set1_epi64((long long)__A); } static __inline__ __m512i __DEFAULT_FN_ATTRS -_mm512_broadcastmw_epi32 (__mmask16 __A) -{ - return (__m512i) _mm512_set1_epi32((int) __A); - +_mm512_broadcastmw_epi32(__mmask16 __A) { + return (__m512i)_mm512_set1_epi32((int)__A); } #undef __DEFAULT_FN_ATTRS -#undef __DEFAULT_FN_ATTRS_CONSTEXPR #endif diff --git a/clang/lib/Headers/avx512dqintrin.h b/clang/lib/Headers/avx512dqintrin.h index fb65bf9..3681cca 100644 --- a/clang/lib/Headers/avx512dqintrin.h +++ b/clang/lib/Headers/avx512dqintrin.h @@ -1083,17 +1083,15 @@ _mm512_broadcast_f32x2(__m128 __A) { 0, 1, 0, 1, 0, 1, 0, 1); } -static __inline__ __m512 __DEFAULT_FN_ATTRS512 -_mm512_mask_broadcast_f32x2 (__m512 __O, __mmask16 __M, __m128 __A) -{ +static __inline__ __m512 __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_mask_broadcast_f32x2(__m512 __O, __mmask16 __M, __m128 __A) { return (__m512)__builtin_ia32_selectps_512((__mmask16)__M, (__v16sf)_mm512_broadcast_f32x2(__A), (__v16sf)__O); } -static __inline__ __m512 __DEFAULT_FN_ATTRS512 -_mm512_maskz_broadcast_f32x2 (__mmask16 __M, __m128 __A) -{ +static __inline__ __m512 __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_maskz_broadcast_f32x2(__mmask16 __M, __m128 __A) { return (__m512)__builtin_ia32_selectps_512((__mmask16)__M, (__v16sf)_mm512_broadcast_f32x2(__A), (__v16sf)_mm512_setzero_ps()); @@ -1106,17 +1104,15 @@ _mm512_broadcast_f32x8(__m256 __A) { 0, 1, 2, 3, 4, 5, 6, 7); } -static __inline__ __m512 __DEFAULT_FN_ATTRS512 -_mm512_mask_broadcast_f32x8(__m512 __O, __mmask16 __M, __m256 __A) -{ +static __inline__ __m512 __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_mask_broadcast_f32x8(__m512 __O, __mmask16 __M, __m256 __A) { return (__m512)__builtin_ia32_selectps_512((__mmask16)__M, (__v16sf)_mm512_broadcast_f32x8(__A), (__v16sf)__O); } -static __inline__ __m512 __DEFAULT_FN_ATTRS512 -_mm512_maskz_broadcast_f32x8(__mmask16 __M, __m256 __A) -{ +static __inline__ __m512 __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_maskz_broadcast_f32x8(__mmask16 __M, __m256 __A) { return (__m512)__builtin_ia32_selectps_512((__mmask16)__M, (__v16sf)_mm512_broadcast_f32x8(__A), (__v16sf)_mm512_setzero_ps()); @@ -1128,17 +1124,15 @@ _mm512_broadcast_f64x2(__m128d __A) { 0, 1, 0, 1, 0, 1, 0, 1); } -static __inline__ __m512d __DEFAULT_FN_ATTRS512 -_mm512_mask_broadcast_f64x2(__m512d __O, __mmask8 __M, __m128d __A) -{ +static __inline__ __m512d __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_mask_broadcast_f64x2(__m512d __O, __mmask8 __M, __m128d __A) { return (__m512d)__builtin_ia32_selectpd_512((__mmask8)__M, (__v8df)_mm512_broadcast_f64x2(__A), (__v8df)__O); } -static __inline__ __m512d __DEFAULT_FN_ATTRS512 -_mm512_maskz_broadcast_f64x2(__mmask8 __M, __m128d __A) -{ +static __inline__ __m512d __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_maskz_broadcast_f64x2(__mmask8 __M, __m128d __A) { return (__m512d)__builtin_ia32_selectpd_512((__mmask8)__M, (__v8df)_mm512_broadcast_f64x2(__A), (__v8df)_mm512_setzero_pd()); @@ -1151,17 +1145,15 @@ _mm512_broadcast_i32x2(__m128i __A) { 0, 1, 0, 1, 0, 1, 0, 1); } -static __inline__ __m512i __DEFAULT_FN_ATTRS512 -_mm512_mask_broadcast_i32x2 (__m512i __O, __mmask16 __M, __m128i __A) -{ +static __inline__ __m512i __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_mask_broadcast_i32x2(__m512i __O, __mmask16 __M, __m128i __A) { return (__m512i)__builtin_ia32_selectd_512((__mmask16)__M, (__v16si)_mm512_broadcast_i32x2(__A), (__v16si)__O); } -static __inline__ __m512i __DEFAULT_FN_ATTRS512 -_mm512_maskz_broadcast_i32x2 (__mmask16 __M, __m128i __A) -{ +static __inline__ __m512i __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_maskz_broadcast_i32x2(__mmask16 __M, __m128i __A) { return (__m512i)__builtin_ia32_selectd_512((__mmask16)__M, (__v16si)_mm512_broadcast_i32x2(__A), (__v16si)_mm512_setzero_si512()); @@ -1174,17 +1166,15 @@ _mm512_broadcast_i32x8(__m256i __A) { 0, 1, 2, 3, 4, 5, 6, 7); } -static __inline__ __m512i __DEFAULT_FN_ATTRS512 -_mm512_mask_broadcast_i32x8(__m512i __O, __mmask16 __M, __m256i __A) -{ +static __inline__ __m512i __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_mask_broadcast_i32x8(__m512i __O, __mmask16 __M, __m256i __A) { return (__m512i)__builtin_ia32_selectd_512((__mmask16)__M, (__v16si)_mm512_broadcast_i32x8(__A), (__v16si)__O); } -static __inline__ __m512i __DEFAULT_FN_ATTRS512 -_mm512_maskz_broadcast_i32x8(__mmask16 __M, __m256i __A) -{ +static __inline__ __m512i __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_maskz_broadcast_i32x8(__mmask16 __M, __m256i __A) { return (__m512i)__builtin_ia32_selectd_512((__mmask16)__M, (__v16si)_mm512_broadcast_i32x8(__A), (__v16si)_mm512_setzero_si512()); @@ -1196,17 +1186,15 @@ _mm512_broadcast_i64x2(__m128i __A) { 0, 1, 0, 1, 0, 1, 0, 1); } -static __inline__ __m512i __DEFAULT_FN_ATTRS512 -_mm512_mask_broadcast_i64x2(__m512i __O, __mmask8 __M, __m128i __A) -{ +static __inline__ __m512i __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_mask_broadcast_i64x2(__m512i __O, __mmask8 __M, __m128i __A) { return (__m512i)__builtin_ia32_selectq_512((__mmask8)__M, (__v8di)_mm512_broadcast_i64x2(__A), (__v8di)__O); } -static __inline__ __m512i __DEFAULT_FN_ATTRS512 -_mm512_maskz_broadcast_i64x2(__mmask8 __M, __m128i __A) -{ +static __inline__ __m512i __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_maskz_broadcast_i64x2(__mmask8 __M, __m128i __A) { return (__m512i)__builtin_ia32_selectq_512((__mmask8)__M, (__v8di)_mm512_broadcast_i64x2(__A), (__v8di)_mm512_setzero_si512()); diff --git a/clang/lib/Headers/avx512fintrin.h b/clang/lib/Headers/avx512fintrin.h index 80e5842..07de036 100644 --- a/clang/lib/Headers/avx512fintrin.h +++ b/clang/lib/Headers/avx512fintrin.h @@ -225,17 +225,15 @@ _mm512_broadcastd_epi32(__m128i __A) { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); } -static __inline__ __m512i __DEFAULT_FN_ATTRS512 -_mm512_mask_broadcastd_epi32 (__m512i __O, __mmask16 __M, __m128i __A) -{ +static __inline__ __m512i __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_mask_broadcastd_epi32(__m512i __O, __mmask16 __M, __m128i __A) { return (__m512i)__builtin_ia32_selectd_512(__M, (__v16si) _mm512_broadcastd_epi32(__A), (__v16si) __O); } -static __inline__ __m512i __DEFAULT_FN_ATTRS512 -_mm512_maskz_broadcastd_epi32 (__mmask16 __M, __m128i __A) -{ +static __inline__ __m512i __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_maskz_broadcastd_epi32(__mmask16 __M, __m128i __A) { return (__m512i)__builtin_ia32_selectd_512(__M, (__v16si) _mm512_broadcastd_epi32(__A), (__v16si) _mm512_setzero_si512()); @@ -247,18 +245,14 @@ _mm512_broadcastq_epi64(__m128i __A) { 0, 0, 0, 0, 0, 0, 0, 0); } -static __inline__ __m512i __DEFAULT_FN_ATTRS512 -_mm512_mask_broadcastq_epi64 (__m512i __O, __mmask8 __M, __m128i __A) -{ - return (__m512i)__builtin_ia32_selectq_512(__M, - (__v8di) _mm512_broadcastq_epi64(__A), - (__v8di) __O); - +static __inline__ __m512i __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_mask_broadcastq_epi64(__m512i __O, __mmask8 __M, __m128i __A) { + return (__m512i)__builtin_ia32_selectq_512( + __M, (__v8di)_mm512_broadcastq_epi64(__A), (__v8di)__O); } -static __inline__ __m512i __DEFAULT_FN_ATTRS512 -_mm512_maskz_broadcastq_epi64 (__mmask8 __M, __m128i __A) -{ +static __inline__ __m512i __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_maskz_broadcastq_epi64(__mmask8 __M, __m128i __A) { return (__m512i)__builtin_ia32_selectq_512(__M, (__v8di) _mm512_broadcastq_epi64(__A), (__v8di) _mm512_setzero_si512()); @@ -321,9 +315,8 @@ _mm512_set1_epi32(int __s) __s, __s, __s, __s, __s, __s, __s, __s }; } -static __inline __m512i __DEFAULT_FN_ATTRS512 -_mm512_maskz_set1_epi32(__mmask16 __M, int __A) -{ +static __inline __m512i __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_maskz_set1_epi32(__mmask16 __M, int __A) { return (__m512i)__builtin_ia32_selectd_512(__M, (__v16si)_mm512_set1_epi32(__A), (__v16si)_mm512_setzero_si512()); @@ -335,9 +328,8 @@ _mm512_set1_epi64(long long __d) return __extension__(__m512i)(__v8di){ __d, __d, __d, __d, __d, __d, __d, __d }; } -static __inline __m512i __DEFAULT_FN_ATTRS512 -_mm512_maskz_set1_epi64(__mmask8 __M, long long __A) -{ +static __inline __m512i __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_maskz_set1_epi64(__mmask8 __M, long long __A) { return (__m512i)__builtin_ia32_selectq_512(__M, (__v8di)_mm512_set1_epi64(__A), (__v8di)_mm512_setzero_si512()); @@ -6552,17 +6544,15 @@ _mm512_broadcast_f32x4(__m128 __A) { 0, 1, 2, 3, 0, 1, 2, 3); } -static __inline__ __m512 __DEFAULT_FN_ATTRS512 -_mm512_mask_broadcast_f32x4(__m512 __O, __mmask16 __M, __m128 __A) -{ +static __inline__ __m512 __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_mask_broadcast_f32x4(__m512 __O, __mmask16 __M, __m128 __A) { return (__m512)__builtin_ia32_selectps_512((__mmask16)__M, (__v16sf)_mm512_broadcast_f32x4(__A), (__v16sf)__O); } -static __inline__ __m512 __DEFAULT_FN_ATTRS512 -_mm512_maskz_broadcast_f32x4(__mmask16 __M, __m128 __A) -{ +static __inline__ __m512 __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_maskz_broadcast_f32x4(__mmask16 __M, __m128 __A) { return (__m512)__builtin_ia32_selectps_512((__mmask16)__M, (__v16sf)_mm512_broadcast_f32x4(__A), (__v16sf)_mm512_setzero_ps()); @@ -6597,17 +6587,15 @@ _mm512_broadcast_i32x4(__m128i __A) { 0, 1, 2, 3, 0, 1, 2, 3); } -static __inline__ __m512i __DEFAULT_FN_ATTRS512 -_mm512_mask_broadcast_i32x4(__m512i __O, __mmask16 __M, __m128i __A) -{ +static __inline__ __m512i __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_mask_broadcast_i32x4(__m512i __O, __mmask16 __M, __m128i __A) { return (__m512i)__builtin_ia32_selectd_512((__mmask16)__M, (__v16si)_mm512_broadcast_i32x4(__A), (__v16si)__O); } -static __inline__ __m512i __DEFAULT_FN_ATTRS512 -_mm512_maskz_broadcast_i32x4(__mmask16 __M, __m128i __A) -{ +static __inline__ __m512i __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_maskz_broadcast_i32x4(__mmask16 __M, __m128i __A) { return (__m512i)__builtin_ia32_selectd_512((__mmask16)__M, (__v16si)_mm512_broadcast_i32x4(__A), (__v16si)_mm512_setzero_si512()); @@ -6635,33 +6623,29 @@ _mm512_maskz_broadcast_i64x4(__mmask8 __M, __m256i __A) (__v8di)_mm512_setzero_si512()); } -static __inline__ __m512d __DEFAULT_FN_ATTRS512 -_mm512_mask_broadcastsd_pd (__m512d __O, __mmask8 __M, __m128d __A) -{ +static __inline__ __m512d __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_mask_broadcastsd_pd(__m512d __O, __mmask8 __M, __m128d __A) { return (__m512d)__builtin_ia32_selectpd_512(__M, (__v8df) _mm512_broadcastsd_pd(__A), (__v8df) __O); } -static __inline__ __m512d __DEFAULT_FN_ATTRS512 -_mm512_maskz_broadcastsd_pd (__mmask8 __M, __m128d __A) -{ +static __inline__ __m512d __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_maskz_broadcastsd_pd(__mmask8 __M, __m128d __A) { return (__m512d)__builtin_ia32_selectpd_512(__M, (__v8df) _mm512_broadcastsd_pd(__A), (__v8df) _mm512_setzero_pd()); } -static __inline__ __m512 __DEFAULT_FN_ATTRS512 -_mm512_mask_broadcastss_ps (__m512 __O, __mmask16 __M, __m128 __A) -{ +static __inline__ __m512 __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_mask_broadcastss_ps(__m512 __O, __mmask16 __M, __m128 __A) { return (__m512)__builtin_ia32_selectps_512(__M, (__v16sf) _mm512_broadcastss_ps(__A), (__v16sf) __O); } -static __inline__ __m512 __DEFAULT_FN_ATTRS512 -_mm512_maskz_broadcastss_ps (__mmask16 __M, __m128 __A) -{ +static __inline__ __m512 __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_maskz_broadcastss_ps(__mmask16 __M, __m128 __A) { return (__m512)__builtin_ia32_selectps_512(__M, (__v16sf) _mm512_broadcastss_ps(__A), (__v16sf) _mm512_setzero_ps()); @@ -8381,17 +8365,15 @@ _mm512_movehdup_ps (__m512 __A) 1, 1, 3, 3, 5, 5, 7, 7, 9, 9, 11, 11, 13, 13, 15, 15); } -static __inline__ __m512 __DEFAULT_FN_ATTRS512 -_mm512_mask_movehdup_ps (__m512 __W, __mmask16 __U, __m512 __A) -{ +static __inline__ __m512 __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_mask_movehdup_ps(__m512 __W, __mmask16 __U, __m512 __A) { return (__m512)__builtin_ia32_selectps_512((__mmask16)__U, (__v16sf)_mm512_movehdup_ps(__A), (__v16sf)__W); } -static __inline__ __m512 __DEFAULT_FN_ATTRS512 -_mm512_maskz_movehdup_ps (__mmask16 __U, __m512 __A) -{ +static __inline__ __m512 __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_maskz_movehdup_ps(__mmask16 __U, __m512 __A) { return (__m512)__builtin_ia32_selectps_512((__mmask16)__U, (__v16sf)_mm512_movehdup_ps(__A), (__v16sf)_mm512_setzero_ps()); @@ -8404,44 +8386,38 @@ _mm512_moveldup_ps (__m512 __A) 0, 0, 2, 2, 4, 4, 6, 6, 8, 8, 10, 10, 12, 12, 14, 14); } -static __inline__ __m512 __DEFAULT_FN_ATTRS512 -_mm512_mask_moveldup_ps (__m512 __W, __mmask16 __U, __m512 __A) -{ +static __inline__ __m512 __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_mask_moveldup_ps(__m512 __W, __mmask16 __U, __m512 __A) { return (__m512)__builtin_ia32_selectps_512((__mmask16)__U, (__v16sf)_mm512_moveldup_ps(__A), (__v16sf)__W); } -static __inline__ __m512 __DEFAULT_FN_ATTRS512 -_mm512_maskz_moveldup_ps (__mmask16 __U, __m512 __A) -{ +static __inline__ __m512 __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_maskz_moveldup_ps(__mmask16 __U, __m512 __A) { return (__m512)__builtin_ia32_selectps_512((__mmask16)__U, (__v16sf)_mm512_moveldup_ps(__A), (__v16sf)_mm512_setzero_ps()); } -static __inline__ __m128 __DEFAULT_FN_ATTRS128 -_mm_mask_move_ss (__m128 __W, __mmask8 __U, __m128 __A, __m128 __B) -{ +static __inline__ __m128 __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask_move_ss(__m128 __W, __mmask8 __U, __m128 __A, __m128 __B) { return __builtin_ia32_selectss_128(__U, _mm_move_ss(__A, __B), __W); } -static __inline__ __m128 __DEFAULT_FN_ATTRS128 -_mm_maskz_move_ss (__mmask8 __U, __m128 __A, __m128 __B) -{ +static __inline__ __m128 __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_maskz_move_ss(__mmask8 __U, __m128 __A, __m128 __B) { return __builtin_ia32_selectss_128(__U, _mm_move_ss(__A, __B), _mm_setzero_ps()); } -static __inline__ __m128d __DEFAULT_FN_ATTRS128 -_mm_mask_move_sd (__m128d __W, __mmask8 __U, __m128d __A, __m128d __B) -{ +static __inline__ __m128d __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask_move_sd(__m128d __W, __mmask8 __U, __m128d __A, __m128d __B) { return __builtin_ia32_selectsd_128(__U, _mm_move_sd(__A, __B), __W); } -static __inline__ __m128d __DEFAULT_FN_ATTRS128 -_mm_maskz_move_sd (__mmask8 __U, __m128d __A, __m128d __B) -{ +static __inline__ __m128d __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_maskz_move_sd(__mmask8 __U, __m128d __A, __m128d __B) { return __builtin_ia32_selectsd_128(__U, _mm_move_sd(__A, __B), _mm_setzero_pd()); } @@ -8884,17 +8860,15 @@ _mm_cvtu64_ss (__m128 __A, unsigned long long __B) } #endif -static __inline__ __m512i __DEFAULT_FN_ATTRS512 -_mm512_mask_set1_epi32 (__m512i __O, __mmask16 __M, int __A) -{ +static __inline__ __m512i __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_mask_set1_epi32(__m512i __O, __mmask16 __M, int __A) { return (__m512i) __builtin_ia32_selectd_512(__M, (__v16si) _mm512_set1_epi32(__A), (__v16si) __O); } -static __inline__ __m512i __DEFAULT_FN_ATTRS512 -_mm512_mask_set1_epi64 (__m512i __O, __mmask8 __M, long long __A) -{ +static __inline__ __m512i __DEFAULT_FN_ATTRS512_CONSTEXPR +_mm512_mask_set1_epi64(__m512i __O, __mmask8 __M, long long __A) { return (__m512i) __builtin_ia32_selectq_512(__M, (__v8di) _mm512_set1_epi64(__A), (__v8di) __O); diff --git a/clang/lib/Headers/avx512ifmaintrin.h b/clang/lib/Headers/avx512ifmaintrin.h index f01b322..625a8ff 100644 --- a/clang/lib/Headers/avx512ifmaintrin.h +++ b/clang/lib/Headers/avx512ifmaintrin.h @@ -15,54 +15,53 @@ #define __IFMAINTRIN_H /* Define the default attributes for the functions in this file. */ +#if defined(__cplusplus) && (__cplusplus >= 201103L) +#define __DEFAULT_FN_ATTRS \ + constexpr \ + __attribute__((__always_inline__, __nodebug__, __target__("avx512ifma"), \ + __min_vector_width__(512))) +#else #define __DEFAULT_FN_ATTRS \ __attribute__((__always_inline__, __nodebug__, __target__("avx512ifma"), \ __min_vector_width__(512))) +#endif static __inline__ __m512i __DEFAULT_FN_ATTRS -_mm512_madd52hi_epu64 (__m512i __X, __m512i __Y, __m512i __Z) -{ - return (__m512i)__builtin_ia32_vpmadd52huq512((__v8di) __X, (__v8di) __Y, - (__v8di) __Z); +_mm512_madd52hi_epu64(__m512i __X, __m512i __Y, __m512i __Z) { + return (__m512i)__builtin_ia32_vpmadd52huq512((__v8di)__X, (__v8di)__Y, + (__v8di)__Z); } -static __inline__ __m512i __DEFAULT_FN_ATTRS -_mm512_mask_madd52hi_epu64 (__m512i __W, __mmask8 __M, __m512i __X, __m512i __Y) -{ - return (__m512i)__builtin_ia32_selectq_512(__M, - (__v8di)_mm512_madd52hi_epu64(__W, __X, __Y), - (__v8di)__W); +static __inline__ __m512i __DEFAULT_FN_ATTRS _mm512_mask_madd52hi_epu64( + __m512i __W, __mmask8 __M, __m512i __X, __m512i __Y) { + return (__m512i)__builtin_ia32_selectq_512( + __M, (__v8di)_mm512_madd52hi_epu64(__W, __X, __Y), (__v8di)__W); } -static __inline__ __m512i __DEFAULT_FN_ATTRS -_mm512_maskz_madd52hi_epu64 (__mmask8 __M, __m512i __X, __m512i __Y, __m512i __Z) -{ - return (__m512i)__builtin_ia32_selectq_512(__M, - (__v8di)_mm512_madd52hi_epu64(__X, __Y, __Z), - (__v8di)_mm512_setzero_si512()); +static __inline__ __m512i __DEFAULT_FN_ATTRS _mm512_maskz_madd52hi_epu64( + __mmask8 __M, __m512i __X, __m512i __Y, __m512i __Z) { + return (__m512i)__builtin_ia32_selectq_512( + __M, (__v8di)_mm512_madd52hi_epu64(__X, __Y, __Z), + (__v8di)_mm512_setzero_si512()); } static __inline__ __m512i __DEFAULT_FN_ATTRS -_mm512_madd52lo_epu64 (__m512i __X, __m512i __Y, __m512i __Z) -{ - return (__m512i)__builtin_ia32_vpmadd52luq512((__v8di) __X, (__v8di) __Y, - (__v8di) __Z); +_mm512_madd52lo_epu64(__m512i __X, __m512i __Y, __m512i __Z) { + return (__m512i)__builtin_ia32_vpmadd52luq512((__v8di)__X, (__v8di)__Y, + (__v8di)__Z); } -static __inline__ __m512i __DEFAULT_FN_ATTRS -_mm512_mask_madd52lo_epu64 (__m512i __W, __mmask8 __M, __m512i __X, __m512i __Y) -{ - return (__m512i)__builtin_ia32_selectq_512(__M, - (__v8di)_mm512_madd52lo_epu64(__W, __X, __Y), - (__v8di)__W); +static __inline__ __m512i __DEFAULT_FN_ATTRS _mm512_mask_madd52lo_epu64( + __m512i __W, __mmask8 __M, __m512i __X, __m512i __Y) { + return (__m512i)__builtin_ia32_selectq_512( + __M, (__v8di)_mm512_madd52lo_epu64(__W, __X, __Y), (__v8di)__W); } -static __inline__ __m512i __DEFAULT_FN_ATTRS -_mm512_maskz_madd52lo_epu64 (__mmask8 __M, __m512i __X, __m512i __Y, __m512i __Z) -{ - return (__m512i)__builtin_ia32_selectq_512(__M, - (__v8di)_mm512_madd52lo_epu64(__X, __Y, __Z), - (__v8di)_mm512_setzero_si512()); +static __inline__ __m512i __DEFAULT_FN_ATTRS _mm512_maskz_madd52lo_epu64( + __mmask8 __M, __m512i __X, __m512i __Y, __m512i __Z) { + return (__m512i)__builtin_ia32_selectq_512( + __M, (__v8di)_mm512_madd52lo_epu64(__X, __Y, __Z), + (__v8di)_mm512_setzero_si512()); } #undef __DEFAULT_FN_ATTRS diff --git a/clang/lib/Headers/avx512ifmavlintrin.h b/clang/lib/Headers/avx512ifmavlintrin.h index a72b561..c4449c7 100644 --- a/clang/lib/Headers/avx512ifmavlintrin.h +++ b/clang/lib/Headers/avx512ifmavlintrin.h @@ -8,13 +8,24 @@ *===-----------------------------------------------------------------------=== */ #ifndef __IMMINTRIN_H -#error "Never use <avx512ifmavlintrin.h> directly; include <immintrin.h> instead." +#error \ + "Never use <avx512ifmavlintrin.h> directly; include <immintrin.h> instead." #endif #ifndef __IFMAVLINTRIN_H #define __IFMAVLINTRIN_H /* Define the default attributes for the functions in this file. */ +#if defined(__cplusplus) && (__cplusplus >= 201103L) +#define __DEFAULT_FN_ATTRS128 \ + constexpr __attribute__((__always_inline__, __nodebug__, \ + __target__("avx512ifma,avx512vl"), \ + __min_vector_width__(128))) +#define __DEFAULT_FN_ATTRS256 \ + constexpr __attribute__((__always_inline__, __nodebug__, \ + __target__("avx512ifma,avx512vl"), \ + __min_vector_width__(256))) +#else #define __DEFAULT_FN_ATTRS128 \ __attribute__((__always_inline__, __nodebug__, \ __target__("avx512ifma,avx512vl"), \ @@ -24,6 +35,8 @@ __target__("avx512ifma,avx512vl"), \ __min_vector_width__(256))) +#endif + #define _mm_madd52hi_epu64(X, Y, Z) \ ((__m128i)__builtin_ia32_vpmadd52huq128((__v2di)(X), (__v2di)(Y), \ (__v2di)(Z))) @@ -41,70 +54,57 @@ (__v4di)(Z))) static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_mask_madd52hi_epu64 (__m128i __W, __mmask8 __M, __m128i __X, __m128i __Y) -{ - return (__m128i)__builtin_ia32_selectq_128(__M, - (__v2di)_mm_madd52hi_epu64(__W, __X, __Y), - (__v2di)__W); +_mm_mask_madd52hi_epu64(__m128i __W, __mmask8 __M, __m128i __X, __m128i __Y) { + return (__m128i)__builtin_ia32_selectq_128( + __M, (__v2di)_mm_madd52hi_epu64(__W, __X, __Y), (__v2di)__W); } static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_maskz_madd52hi_epu64 (__mmask8 __M, __m128i __X, __m128i __Y, __m128i __Z) -{ - return (__m128i)__builtin_ia32_selectq_128(__M, - (__v2di)_mm_madd52hi_epu64(__X, __Y, __Z), - (__v2di)_mm_setzero_si128()); +_mm_maskz_madd52hi_epu64(__mmask8 __M, __m128i __X, __m128i __Y, __m128i __Z) { + return (__m128i)__builtin_ia32_selectq_128( + __M, (__v2di)_mm_madd52hi_epu64(__X, __Y, __Z), + (__v2di)_mm_setzero_si128()); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_mask_madd52hi_epu64 (__m256i __W, __mmask8 __M, __m256i __X, __m256i __Y) -{ - return (__m256i)__builtin_ia32_selectq_256(__M, - (__v4di)_mm256_madd52hi_epu64(__W, __X, __Y), - (__v4di)__W); +static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_mask_madd52hi_epu64( + __m256i __W, __mmask8 __M, __m256i __X, __m256i __Y) { + return (__m256i)__builtin_ia32_selectq_256( + __M, (__v4di)_mm256_madd52hi_epu64(__W, __X, __Y), (__v4di)__W); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_maskz_madd52hi_epu64 (__mmask8 __M, __m256i __X, __m256i __Y, __m256i __Z) -{ - return (__m256i)__builtin_ia32_selectq_256(__M, - (__v4di)_mm256_madd52hi_epu64(__X, __Y, __Z), - (__v4di)_mm256_setzero_si256()); +static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_maskz_madd52hi_epu64( + __mmask8 __M, __m256i __X, __m256i __Y, __m256i __Z) { + return (__m256i)__builtin_ia32_selectq_256( + __M, (__v4di)_mm256_madd52hi_epu64(__X, __Y, __Z), + (__v4di)_mm256_setzero_si256()); } static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_mask_madd52lo_epu64 (__m128i __W, __mmask8 __M, __m128i __X, __m128i __Y) -{ - return (__m128i)__builtin_ia32_selectq_128(__M, - (__v2di)_mm_madd52lo_epu64(__W, __X, __Y), - (__v2di)__W); +_mm_mask_madd52lo_epu64(__m128i __W, __mmask8 __M, __m128i __X, __m128i __Y) { + return (__m128i)__builtin_ia32_selectq_128( + __M, (__v2di)_mm_madd52lo_epu64(__W, __X, __Y), (__v2di)__W); } static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_maskz_madd52lo_epu64 (__mmask8 __M, __m128i __X, __m128i __Y, __m128i __Z) -{ - return (__m128i)__builtin_ia32_selectq_128(__M, - (__v2di)_mm_madd52lo_epu64(__X, __Y, __Z), - (__v2di)_mm_setzero_si128()); +_mm_maskz_madd52lo_epu64(__mmask8 __M, __m128i __X, __m128i __Y, __m128i __Z) { + return (__m128i)__builtin_ia32_selectq_128( + __M, (__v2di)_mm_madd52lo_epu64(__X, __Y, __Z), + (__v2di)_mm_setzero_si128()); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_mask_madd52lo_epu64 (__m256i __W, __mmask8 __M, __m256i __X, __m256i __Y) -{ - return (__m256i)__builtin_ia32_selectq_256(__M, - (__v4di)_mm256_madd52lo_epu64(__W, __X, __Y), - (__v4di)__W); +static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_mask_madd52lo_epu64( + __m256i __W, __mmask8 __M, __m256i __X, __m256i __Y) { + return (__m256i)__builtin_ia32_selectq_256( + __M, (__v4di)_mm256_madd52lo_epu64(__W, __X, __Y), (__v4di)__W); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_maskz_madd52lo_epu64 (__mmask8 __M, __m256i __X, __m256i __Y, __m256i __Z) -{ - return (__m256i)__builtin_ia32_selectq_256(__M, - (__v4di)_mm256_madd52lo_epu64(__X, __Y, __Z), - (__v4di)_mm256_setzero_si256()); +static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_maskz_madd52lo_epu64( + __mmask8 __M, __m256i __X, __m256i __Y, __m256i __Z) { + return (__m256i)__builtin_ia32_selectq_256( + __M, (__v4di)_mm256_madd52lo_epu64(__X, __Y, __Z), + (__v4di)_mm256_setzero_si256()); } - #undef __DEFAULT_FN_ATTRS128 #undef __DEFAULT_FN_ATTRS256 diff --git a/clang/lib/Headers/avx512vlbwintrin.h b/clang/lib/Headers/avx512vlbwintrin.h index 81e4cbb9..639fb60 100644 --- a/clang/lib/Headers/avx512vlbwintrin.h +++ b/clang/lib/Headers/avx512vlbwintrin.h @@ -1067,33 +1067,29 @@ _mm256_mask_min_epu16(__m256i __W, __mmask16 __M, __m256i __A, __m256i __B) { (__v16hi)__W); } -static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_mask_shuffle_epi8(__m128i __W, __mmask16 __U, __m128i __A, __m128i __B) -{ +static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask_shuffle_epi8(__m128i __W, __mmask16 __U, __m128i __A, __m128i __B) { return (__m128i)__builtin_ia32_selectb_128((__mmask16)__U, (__v16qi)_mm_shuffle_epi8(__A, __B), (__v16qi)__W); } -static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_maskz_shuffle_epi8(__mmask16 __U, __m128i __A, __m128i __B) -{ +static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_maskz_shuffle_epi8(__mmask16 __U, __m128i __A, __m128i __B) { return (__m128i)__builtin_ia32_selectb_128((__mmask16)__U, (__v16qi)_mm_shuffle_epi8(__A, __B), (__v16qi)_mm_setzero_si128()); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_mask_shuffle_epi8(__m256i __W, __mmask32 __U, __m256i __A, __m256i __B) -{ +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_shuffle_epi8(__m256i __W, __mmask32 __U, __m256i __A, __m256i __B) { return (__m256i)__builtin_ia32_selectb_256((__mmask32)__U, (__v32qi)_mm256_shuffle_epi8(__A, __B), (__v32qi)__W); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_maskz_shuffle_epi8(__mmask32 __U, __m256i __A, __m256i __B) -{ +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_maskz_shuffle_epi8(__mmask32 __U, __m256i __A, __m256i __B) { return (__m256i)__builtin_ia32_selectb_256((__mmask32)__U, (__v32qi)_mm256_shuffle_epi8(__A, __B), (__v32qi)_mm256_setzero_si256()); diff --git a/clang/lib/Headers/avx512vlcdintrin.h b/clang/lib/Headers/avx512vlcdintrin.h index 30c9f90..7719680f 100644 --- a/clang/lib/Headers/avx512vlcdintrin.h +++ b/clang/lib/Headers/avx512vlcdintrin.h @@ -14,208 +14,182 @@ #define __AVX512VLCDINTRIN_H /* Define the default attributes for the functions in this file. */ +#if defined(__cplusplus) && (__cplusplus >= 201103L) +#define __DEFAULT_FN_ATTRS128 \ + constexpr __attribute__((__always_inline__, __nodebug__, \ + __target__("avx512vl,avx512cd"), \ + __min_vector_width__(128))) +#define __DEFAULT_FN_ATTRS256 \ + constexpr __attribute__((__always_inline__, __nodebug__, \ + __target__("avx512vl,avx512cd"), \ + __min_vector_width__(256))) +#else #define __DEFAULT_FN_ATTRS128 \ __attribute__((__always_inline__, __nodebug__, \ __target__("avx512vl,avx512cd"), __min_vector_width__(128))) #define __DEFAULT_FN_ATTRS256 \ __attribute__((__always_inline__, __nodebug__, \ __target__("avx512vl,avx512cd"), __min_vector_width__(256))) - -#if defined(__cplusplus) && (__cplusplus >= 201103L) -#define __DEFAULT_FN_ATTRS256_CONSTEXPR __DEFAULT_FN_ATTRS256 constexpr -#define __DEFAULT_FN_ATTRS128_CONSTEXPR __DEFAULT_FN_ATTRS128 constexpr -#else -#define __DEFAULT_FN_ATTRS256_CONSTEXPR __DEFAULT_FN_ATTRS256 -#define __DEFAULT_FN_ATTRS128_CONSTEXPR __DEFAULT_FN_ATTRS128 #endif static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_broadcastmb_epi64 (__mmask8 __A) -{ +_mm_broadcastmb_epi64(__mmask8 __A) { return (__m128i) _mm_set1_epi64x((long long) __A); } static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_broadcastmb_epi64 (__mmask8 __A) -{ - return (__m256i) _mm256_set1_epi64x((long long)__A); +_mm256_broadcastmb_epi64(__mmask8 __A) { + return (__m256i)_mm256_set1_epi64x((long long)__A); } static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_broadcastmw_epi32 (__mmask16 __A) -{ +_mm_broadcastmw_epi32(__mmask16 __A) { return (__m128i) _mm_set1_epi32((int)__A); } static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_broadcastmw_epi32 (__mmask16 __A) -{ +_mm256_broadcastmw_epi32(__mmask16 __A) { return (__m256i) _mm256_set1_epi32((int)__A); } - static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_conflict_epi64 (__m128i __A) -{ - return (__m128i) __builtin_ia32_vpconflictdi_128 ((__v2di) __A); +_mm_conflict_epi64(__m128i __A) { + return (__m128i)__builtin_ia32_vpconflictdi_128((__v2di)__A); } static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_mask_conflict_epi64 (__m128i __W, __mmask8 __U, __m128i __A) -{ - return (__m128i)__builtin_ia32_selectq_128((__mmask8)__U, - (__v2di)_mm_conflict_epi64(__A), - (__v2di)__W); +_mm_mask_conflict_epi64(__m128i __W, __mmask8 __U, __m128i __A) { + return (__m128i)__builtin_ia32_selectq_128( + (__mmask8)__U, (__v2di)_mm_conflict_epi64(__A), (__v2di)__W); } static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_maskz_conflict_epi64 (__mmask8 __U, __m128i __A) -{ +_mm_maskz_conflict_epi64(__mmask8 __U, __m128i __A) { return (__m128i)__builtin_ia32_selectq_128((__mmask8)__U, (__v2di)_mm_conflict_epi64(__A), (__v2di)_mm_setzero_si128()); } static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_conflict_epi64 (__m256i __A) -{ - return (__m256i) __builtin_ia32_vpconflictdi_256 ((__v4di) __A); +_mm256_conflict_epi64(__m256i __A) { + return (__m256i)__builtin_ia32_vpconflictdi_256((__v4di)__A); } static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_mask_conflict_epi64 (__m256i __W, __mmask8 __U, __m256i __A) -{ - return (__m256i)__builtin_ia32_selectq_256((__mmask8)__U, - (__v4di)_mm256_conflict_epi64(__A), - (__v4di)__W); +_mm256_mask_conflict_epi64(__m256i __W, __mmask8 __U, __m256i __A) { + return (__m256i)__builtin_ia32_selectq_256( + (__mmask8)__U, (__v4di)_mm256_conflict_epi64(__A), (__v4di)__W); } static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_maskz_conflict_epi64 (__mmask8 __U, __m256i __A) -{ +_mm256_maskz_conflict_epi64(__mmask8 __U, __m256i __A) { return (__m256i)__builtin_ia32_selectq_256((__mmask8)__U, (__v4di)_mm256_conflict_epi64(__A), (__v4di)_mm256_setzero_si256()); } static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_conflict_epi32 (__m128i __A) -{ - return (__m128i) __builtin_ia32_vpconflictsi_128 ((__v4si) __A); +_mm_conflict_epi32(__m128i __A) { + return (__m128i)__builtin_ia32_vpconflictsi_128((__v4si)__A); } static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_mask_conflict_epi32 (__m128i __W, __mmask8 __U, __m128i __A) -{ - return (__m128i)__builtin_ia32_selectd_128((__mmask8)__U, - (__v4si)_mm_conflict_epi32(__A), - (__v4si)__W); +_mm_mask_conflict_epi32(__m128i __W, __mmask8 __U, __m128i __A) { + return (__m128i)__builtin_ia32_selectd_128( + (__mmask8)__U, (__v4si)_mm_conflict_epi32(__A), (__v4si)__W); } static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_maskz_conflict_epi32 (__mmask8 __U, __m128i __A) -{ +_mm_maskz_conflict_epi32(__mmask8 __U, __m128i __A) { return (__m128i)__builtin_ia32_selectd_128((__mmask8)__U, (__v4si)_mm_conflict_epi32(__A), (__v4si)_mm_setzero_si128()); } static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_conflict_epi32 (__m256i __A) -{ - return (__m256i) __builtin_ia32_vpconflictsi_256 ((__v8si) __A); +_mm256_conflict_epi32(__m256i __A) { + return (__m256i)__builtin_ia32_vpconflictsi_256((__v8si)__A); } static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_mask_conflict_epi32 (__m256i __W, __mmask8 __U, __m256i __A) -{ - return (__m256i)__builtin_ia32_selectd_256((__mmask8)__U, - (__v8si)_mm256_conflict_epi32(__A), - (__v8si)__W); +_mm256_mask_conflict_epi32(__m256i __W, __mmask8 __U, __m256i __A) { + return (__m256i)__builtin_ia32_selectd_256( + (__mmask8)__U, (__v8si)_mm256_conflict_epi32(__A), (__v8si)__W); } static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_maskz_conflict_epi32 (__mmask8 __U, __m256i __A) -{ +_mm256_maskz_conflict_epi32(__mmask8 __U, __m256i __A) { return (__m256i)__builtin_ia32_selectd_256((__mmask8)__U, (__v8si)_mm256_conflict_epi32(__A), (__v8si)_mm256_setzero_si256()); } -static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR -_mm_lzcnt_epi32(__m128i __A) { +static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_lzcnt_epi32(__m128i __A) { return (__m128i)__builtin_elementwise_clzg((__v4si)__A, (__v4si)_mm_set1_epi32(32)); } -static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR +static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_mask_lzcnt_epi32(__m128i __W, __mmask8 __U, __m128i __A) { - return (__m128i)__builtin_ia32_selectd_128((__mmask8)__U, - (__v4si)_mm_lzcnt_epi32(__A), - (__v4si)__W); + return (__m128i)__builtin_ia32_selectd_128( + (__mmask8)__U, (__v4si)_mm_lzcnt_epi32(__A), (__v4si)__W); } -static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR +static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_maskz_lzcnt_epi32(__mmask8 __U, __m128i __A) { - return (__m128i)__builtin_ia32_selectd_128((__mmask8)__U, - (__v4si)_mm_lzcnt_epi32(__A), - (__v4si)_mm_setzero_si128()); + return (__m128i)__builtin_ia32_selectd_128( + (__mmask8)__U, (__v4si)_mm_lzcnt_epi32(__A), (__v4si)_mm_setzero_si128()); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_lzcnt_epi32(__m256i __A) { return (__m256i)__builtin_elementwise_clzg((__v8si)__A, (__v8si)_mm256_set1_epi32(32)); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_mask_lzcnt_epi32(__m256i __W, __mmask8 __U, __m256i __A) { - return (__m256i)__builtin_ia32_selectd_256((__mmask8)__U, - (__v8si)_mm256_lzcnt_epi32(__A), - (__v8si)__W); + return (__m256i)__builtin_ia32_selectd_256( + (__mmask8)__U, (__v8si)_mm256_lzcnt_epi32(__A), (__v8si)__W); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_maskz_lzcnt_epi32(__mmask8 __U, __m256i __A) { return (__m256i)__builtin_ia32_selectd_256((__mmask8)__U, (__v8si)_mm256_lzcnt_epi32(__A), (__v8si)_mm256_setzero_si256()); } -static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR -_mm_lzcnt_epi64(__m128i __A) { +static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_lzcnt_epi64(__m128i __A) { return (__m128i)__builtin_elementwise_clzg( (__v2di)__A, (__v2di)_mm_set1_epi64x((long long)64)); } -static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR +static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_mask_lzcnt_epi64(__m128i __W, __mmask8 __U, __m128i __A) { - return (__m128i)__builtin_ia32_selectq_128((__mmask8)__U, - (__v2di)_mm_lzcnt_epi64(__A), - (__v2di)__W); + return (__m128i)__builtin_ia32_selectq_128( + (__mmask8)__U, (__v2di)_mm_lzcnt_epi64(__A), (__v2di)__W); } -static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR +static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_maskz_lzcnt_epi64(__mmask8 __U, __m128i __A) { - return (__m128i)__builtin_ia32_selectq_128((__mmask8)__U, - (__v2di)_mm_lzcnt_epi64(__A), - (__v2di)_mm_setzero_si128()); + return (__m128i)__builtin_ia32_selectq_128( + (__mmask8)__U, (__v2di)_mm_lzcnt_epi64(__A), (__v2di)_mm_setzero_si128()); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_lzcnt_epi64(__m256i __A) { return (__m256i)__builtin_elementwise_clzg( (__v4di)__A, (__v4di)_mm256_set1_epi64x((long long)64)); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_mask_lzcnt_epi64(__m256i __W, __mmask8 __U, __m256i __A) { - return (__m256i)__builtin_ia32_selectq_256((__mmask8)__U, - (__v4di)_mm256_lzcnt_epi64(__A), - (__v4di)__W); + return (__m256i)__builtin_ia32_selectq_256( + (__mmask8)__U, (__v4di)_mm256_lzcnt_epi64(__A), (__v4di)__W); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_maskz_lzcnt_epi64(__mmask8 __U, __m256i __A) { return (__m256i)__builtin_ia32_selectq_256((__mmask8)__U, (__v4di)_mm256_lzcnt_epi64(__A), @@ -224,7 +198,5 @@ _mm256_maskz_lzcnt_epi64(__mmask8 __U, __m256i __A) { #undef __DEFAULT_FN_ATTRS128 #undef __DEFAULT_FN_ATTRS256 -#undef __DEFAULT_FN_ATTRS128_CONSTEXPR -#undef __DEFAULT_FN_ATTRS256_CONSTEXPR #endif /* __AVX512VLCDINTRIN_H */ diff --git a/clang/lib/Headers/avx512vldqintrin.h b/clang/lib/Headers/avx512vldqintrin.h index 68bd52e..ee7974e 100644 --- a/clang/lib/Headers/avx512vldqintrin.h +++ b/clang/lib/Headers/avx512vldqintrin.h @@ -968,17 +968,15 @@ _mm256_broadcast_f32x2(__m128 __A) { 0, 1, 0, 1, 0, 1, 0, 1); } -static __inline__ __m256 __DEFAULT_FN_ATTRS256 -_mm256_mask_broadcast_f32x2 (__m256 __O, __mmask8 __M, __m128 __A) -{ +static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_broadcast_f32x2(__m256 __O, __mmask8 __M, __m128 __A) { return (__m256)__builtin_ia32_selectps_256((__mmask8)__M, (__v8sf)_mm256_broadcast_f32x2(__A), (__v8sf)__O); } -static __inline__ __m256 __DEFAULT_FN_ATTRS256 -_mm256_maskz_broadcast_f32x2 (__mmask8 __M, __m128 __A) -{ +static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_maskz_broadcast_f32x2(__mmask8 __M, __m128 __A) { return (__m256)__builtin_ia32_selectps_256((__mmask8)__M, (__v8sf)_mm256_broadcast_f32x2(__A), (__v8sf)_mm256_setzero_ps()); @@ -990,17 +988,15 @@ _mm256_broadcast_f64x2(__m128d __A) { 0, 1, 0, 1); } -static __inline__ __m256d __DEFAULT_FN_ATTRS256 -_mm256_mask_broadcast_f64x2(__m256d __O, __mmask8 __M, __m128d __A) -{ +static __inline__ __m256d __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_broadcast_f64x2(__m256d __O, __mmask8 __M, __m128d __A) { return (__m256d)__builtin_ia32_selectpd_256((__mmask8)__M, (__v4df)_mm256_broadcast_f64x2(__A), (__v4df)__O); } -static __inline__ __m256d __DEFAULT_FN_ATTRS256 -_mm256_maskz_broadcast_f64x2 (__mmask8 __M, __m128d __A) -{ +static __inline__ __m256d __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_maskz_broadcast_f64x2(__mmask8 __M, __m128d __A) { return (__m256d)__builtin_ia32_selectpd_256((__mmask8)__M, (__v4df)_mm256_broadcast_f64x2(__A), (__v4df)_mm256_setzero_pd()); @@ -1012,17 +1008,15 @@ _mm_broadcast_i32x2(__m128i __A) { 0, 1, 0, 1); } -static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_mask_broadcast_i32x2 (__m128i __O, __mmask8 __M, __m128i __A) -{ +static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask_broadcast_i32x2(__m128i __O, __mmask8 __M, __m128i __A) { return (__m128i)__builtin_ia32_selectd_128((__mmask8)__M, (__v4si)_mm_broadcast_i32x2(__A), (__v4si)__O); } -static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_maskz_broadcast_i32x2 (__mmask8 __M, __m128i __A) -{ +static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_maskz_broadcast_i32x2(__mmask8 __M, __m128i __A) { return (__m128i)__builtin_ia32_selectd_128((__mmask8)__M, (__v4si)_mm_broadcast_i32x2(__A), (__v4si)_mm_setzero_si128()); @@ -1034,17 +1028,15 @@ _mm256_broadcast_i32x2(__m128i __A) { 0, 1, 0, 1, 0, 1, 0, 1); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_mask_broadcast_i32x2 (__m256i __O, __mmask8 __M, __m128i __A) -{ +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_broadcast_i32x2(__m256i __O, __mmask8 __M, __m128i __A) { return (__m256i)__builtin_ia32_selectd_256((__mmask8)__M, (__v8si)_mm256_broadcast_i32x2(__A), (__v8si)__O); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_maskz_broadcast_i32x2 (__mmask8 __M, __m128i __A) -{ +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_maskz_broadcast_i32x2(__mmask8 __M, __m128i __A) { return (__m256i)__builtin_ia32_selectd_256((__mmask8)__M, (__v8si)_mm256_broadcast_i32x2(__A), (__v8si)_mm256_setzero_si256()); @@ -1056,17 +1048,15 @@ _mm256_broadcast_i64x2(__m128i __A) { 0, 1, 0, 1); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_mask_broadcast_i64x2(__m256i __O, __mmask8 __M, __m128i __A) -{ +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_broadcast_i64x2(__m256i __O, __mmask8 __M, __m128i __A) { return (__m256i)__builtin_ia32_selectq_256((__mmask8)__M, (__v4di)_mm256_broadcast_i64x2(__A), (__v4di)__O); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_maskz_broadcast_i64x2 (__mmask8 __M, __m128i __A) -{ +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_maskz_broadcast_i64x2(__mmask8 __M, __m128i __A) { return (__m256i)__builtin_ia32_selectq_256((__mmask8)__M, (__v4di)_mm256_broadcast_i64x2(__A), (__v4di)_mm256_setzero_si256()); diff --git a/clang/lib/Headers/avx512vlintrin.h b/clang/lib/Headers/avx512vlintrin.h index 965741f..676b5a0 100644 --- a/clang/lib/Headers/avx512vlintrin.h +++ b/clang/lib/Headers/avx512vlintrin.h @@ -5101,69 +5101,55 @@ _mm256_maskz_movedup_pd (__mmask8 __U, __m256d __A) (__v4df)_mm256_setzero_pd()); } -static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_mask_set1_epi32(__m128i __O, __mmask8 __M, int __A) -{ - return (__m128i)__builtin_ia32_selectd_128(__M, - (__v4si) _mm_set1_epi32(__A), - (__v4si)__O); +static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask_set1_epi32(__m128i __O, __mmask8 __M, int __A) { + return (__m128i)__builtin_ia32_selectd_128(__M, (__v4si)_mm_set1_epi32(__A), + (__v4si)__O); } -static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_maskz_set1_epi32( __mmask8 __M, int __A) -{ - return (__m128i)__builtin_ia32_selectd_128(__M, - (__v4si) _mm_set1_epi32(__A), - (__v4si)_mm_setzero_si128()); +static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_maskz_set1_epi32(__mmask8 __M, int __A) { + return (__m128i)__builtin_ia32_selectd_128(__M, (__v4si)_mm_set1_epi32(__A), + (__v4si)_mm_setzero_si128()); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_mask_set1_epi32(__m256i __O, __mmask8 __M, int __A) -{ - return (__m256i)__builtin_ia32_selectd_256(__M, - (__v8si) _mm256_set1_epi32(__A), - (__v8si)__O); +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_set1_epi32(__m256i __O, __mmask8 __M, int __A) { + return (__m256i)__builtin_ia32_selectd_256( + __M, (__v8si)_mm256_set1_epi32(__A), (__v8si)__O); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_maskz_set1_epi32( __mmask8 __M, int __A) -{ - return (__m256i)__builtin_ia32_selectd_256(__M, - (__v8si) _mm256_set1_epi32(__A), - (__v8si)_mm256_setzero_si256()); +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_maskz_set1_epi32(__mmask8 __M, int __A) { + return (__m256i)__builtin_ia32_selectd_256( + __M, (__v8si)_mm256_set1_epi32(__A), (__v8si)_mm256_setzero_si256()); } - -static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_mask_set1_epi64 (__m128i __O, __mmask8 __M, long long __A) -{ +static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask_set1_epi64(__m128i __O, __mmask8 __M, long long __A) { return (__m128i) __builtin_ia32_selectq_128(__M, (__v2di) _mm_set1_epi64x(__A), (__v2di) __O); } -static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_maskz_set1_epi64 (__mmask8 __M, long long __A) -{ +static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_maskz_set1_epi64(__mmask8 __M, long long __A) { return (__m128i) __builtin_ia32_selectq_128(__M, (__v2di) _mm_set1_epi64x(__A), (__v2di) _mm_setzero_si128()); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_mask_set1_epi64 (__m256i __O, __mmask8 __M, long long __A) -{ +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_set1_epi64(__m256i __O, __mmask8 __M, long long __A) { return (__m256i) __builtin_ia32_selectq_256(__M, (__v4di) _mm256_set1_epi64x(__A), (__v4di) __O) ; } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_maskz_set1_epi64 (__mmask8 __M, long long __A) -{ - return (__m256i) __builtin_ia32_selectq_256(__M, - (__v4di) _mm256_set1_epi64x(__A), - (__v4di) _mm256_setzero_si256()); +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_maskz_set1_epi64(__mmask8 __M, long long __A) { + return (__m256i)__builtin_ia32_selectq_256( + __M, (__v4di)_mm256_set1_epi64x(__A), (__v4di)_mm256_setzero_si256()); } #define _mm_fixupimm_pd(A, B, C, imm) \ @@ -5610,130 +5596,113 @@ _mm256_mask_storeu_ps (void *__P, __mmask8 __U, __m256 __A) (__mmask8) __U); } - -static __inline__ __m128d __DEFAULT_FN_ATTRS128 -_mm_mask_unpackhi_pd(__m128d __W, __mmask8 __U, __m128d __A, __m128d __B) -{ +static __inline__ __m128d __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask_unpackhi_pd(__m128d __W, __mmask8 __U, __m128d __A, __m128d __B) { return (__m128d)__builtin_ia32_selectpd_128((__mmask8)__U, (__v2df)_mm_unpackhi_pd(__A, __B), (__v2df)__W); } -static __inline__ __m128d __DEFAULT_FN_ATTRS128 -_mm_maskz_unpackhi_pd(__mmask8 __U, __m128d __A, __m128d __B) -{ +static __inline__ __m128d __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_maskz_unpackhi_pd(__mmask8 __U, __m128d __A, __m128d __B) { return (__m128d)__builtin_ia32_selectpd_128((__mmask8)__U, (__v2df)_mm_unpackhi_pd(__A, __B), (__v2df)_mm_setzero_pd()); } -static __inline__ __m256d __DEFAULT_FN_ATTRS256 -_mm256_mask_unpackhi_pd(__m256d __W, __mmask8 __U, __m256d __A, __m256d __B) -{ +static __inline__ __m256d __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_unpackhi_pd(__m256d __W, __mmask8 __U, __m256d __A, __m256d __B) { return (__m256d)__builtin_ia32_selectpd_256((__mmask8)__U, (__v4df)_mm256_unpackhi_pd(__A, __B), (__v4df)__W); } -static __inline__ __m256d __DEFAULT_FN_ATTRS256 -_mm256_maskz_unpackhi_pd(__mmask8 __U, __m256d __A, __m256d __B) -{ +static __inline__ __m256d __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_maskz_unpackhi_pd(__mmask8 __U, __m256d __A, __m256d __B) { return (__m256d)__builtin_ia32_selectpd_256((__mmask8)__U, (__v4df)_mm256_unpackhi_pd(__A, __B), (__v4df)_mm256_setzero_pd()); } -static __inline__ __m128 __DEFAULT_FN_ATTRS128 -_mm_mask_unpackhi_ps(__m128 __W, __mmask8 __U, __m128 __A, __m128 __B) -{ +static __inline__ __m128 __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask_unpackhi_ps(__m128 __W, __mmask8 __U, __m128 __A, __m128 __B) { return (__m128)__builtin_ia32_selectps_128((__mmask8)__U, (__v4sf)_mm_unpackhi_ps(__A, __B), (__v4sf)__W); } -static __inline__ __m128 __DEFAULT_FN_ATTRS128 -_mm_maskz_unpackhi_ps(__mmask8 __U, __m128 __A, __m128 __B) -{ +static __inline__ __m128 __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_maskz_unpackhi_ps(__mmask8 __U, __m128 __A, __m128 __B) { return (__m128)__builtin_ia32_selectps_128((__mmask8)__U, (__v4sf)_mm_unpackhi_ps(__A, __B), (__v4sf)_mm_setzero_ps()); } -static __inline__ __m256 __DEFAULT_FN_ATTRS256 -_mm256_mask_unpackhi_ps(__m256 __W, __mmask8 __U, __m256 __A, __m256 __B) -{ +static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_unpackhi_ps(__m256 __W, __mmask8 __U, __m256 __A, __m256 __B) { return (__m256)__builtin_ia32_selectps_256((__mmask8)__U, (__v8sf)_mm256_unpackhi_ps(__A, __B), (__v8sf)__W); } -static __inline__ __m256 __DEFAULT_FN_ATTRS256 -_mm256_maskz_unpackhi_ps(__mmask8 __U, __m256 __A, __m256 __B) -{ +static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_maskz_unpackhi_ps(__mmask8 __U, __m256 __A, __m256 __B) { return (__m256)__builtin_ia32_selectps_256((__mmask8)__U, (__v8sf)_mm256_unpackhi_ps(__A, __B), (__v8sf)_mm256_setzero_ps()); } -static __inline__ __m128d __DEFAULT_FN_ATTRS128 -_mm_mask_unpacklo_pd(__m128d __W, __mmask8 __U, __m128d __A, __m128d __B) -{ +static __inline__ __m128d __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask_unpacklo_pd(__m128d __W, __mmask8 __U, __m128d __A, __m128d __B) { return (__m128d)__builtin_ia32_selectpd_128((__mmask8)__U, (__v2df)_mm_unpacklo_pd(__A, __B), (__v2df)__W); } -static __inline__ __m128d __DEFAULT_FN_ATTRS128 -_mm_maskz_unpacklo_pd(__mmask8 __U, __m128d __A, __m128d __B) -{ +static __inline__ __m128d __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_maskz_unpacklo_pd(__mmask8 __U, __m128d __A, __m128d __B) { return (__m128d)__builtin_ia32_selectpd_128((__mmask8)__U, (__v2df)_mm_unpacklo_pd(__A, __B), (__v2df)_mm_setzero_pd()); } -static __inline__ __m256d __DEFAULT_FN_ATTRS256 -_mm256_mask_unpacklo_pd(__m256d __W, __mmask8 __U, __m256d __A, __m256d __B) -{ +static __inline__ __m256d __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_unpacklo_pd(__m256d __W, __mmask8 __U, __m256d __A, __m256d __B) { return (__m256d)__builtin_ia32_selectpd_256((__mmask8)__U, (__v4df)_mm256_unpacklo_pd(__A, __B), (__v4df)__W); } -static __inline__ __m256d __DEFAULT_FN_ATTRS256 -_mm256_maskz_unpacklo_pd(__mmask8 __U, __m256d __A, __m256d __B) -{ +static __inline__ __m256d __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_maskz_unpacklo_pd(__mmask8 __U, __m256d __A, __m256d __B) { return (__m256d)__builtin_ia32_selectpd_256((__mmask8)__U, (__v4df)_mm256_unpacklo_pd(__A, __B), (__v4df)_mm256_setzero_pd()); } -static __inline__ __m128 __DEFAULT_FN_ATTRS128 -_mm_mask_unpacklo_ps(__m128 __W, __mmask8 __U, __m128 __A, __m128 __B) -{ +static __inline__ __m128 __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask_unpacklo_ps(__m128 __W, __mmask8 __U, __m128 __A, __m128 __B) { return (__m128)__builtin_ia32_selectps_128((__mmask8)__U, (__v4sf)_mm_unpacklo_ps(__A, __B), (__v4sf)__W); } -static __inline__ __m128 __DEFAULT_FN_ATTRS128 -_mm_maskz_unpacklo_ps(__mmask8 __U, __m128 __A, __m128 __B) -{ +static __inline__ __m128 __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_maskz_unpacklo_ps(__mmask8 __U, __m128 __A, __m128 __B) { return (__m128)__builtin_ia32_selectps_128((__mmask8)__U, (__v4sf)_mm_unpacklo_ps(__A, __B), (__v4sf)_mm_setzero_ps()); } -static __inline__ __m256 __DEFAULT_FN_ATTRS256 -_mm256_mask_unpacklo_ps(__m256 __W, __mmask8 __U, __m256 __A, __m256 __B) -{ +static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_unpacklo_ps(__m256 __W, __mmask8 __U, __m256 __A, __m256 __B) { return (__m256)__builtin_ia32_selectps_256((__mmask8)__U, (__v8sf)_mm256_unpacklo_ps(__A, __B), (__v8sf)__W); } -static __inline__ __m256 __DEFAULT_FN_ATTRS256 -_mm256_maskz_unpacklo_ps(__mmask8 __U, __m256 __A, __m256 __B) -{ +static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_maskz_unpacklo_ps(__mmask8 __U, __m256 __A, __m256 __B) { return (__m256)__builtin_ia32_selectps_256((__mmask8)__U, (__v8sf)_mm256_unpacklo_ps(__A, __B), (__v8sf)_mm256_setzero_ps()); @@ -6055,129 +6024,117 @@ _mm256_mask_testn_epi64_mask (__mmask8 __U, __m256i __A, __m256i __B) _mm256_setzero_si256()); } -static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_mask_unpackhi_epi32(__m128i __W, __mmask8 __U, __m128i __A, __m128i __B) -{ +static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask_unpackhi_epi32(__m128i __W, __mmask8 __U, __m128i __A, __m128i __B) { return (__m128i)__builtin_ia32_selectd_128((__mmask8)__U, (__v4si)_mm_unpackhi_epi32(__A, __B), (__v4si)__W); } -static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_maskz_unpackhi_epi32(__mmask8 __U, __m128i __A, __m128i __B) -{ +static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_maskz_unpackhi_epi32(__mmask8 __U, __m128i __A, __m128i __B) { return (__m128i)__builtin_ia32_selectd_128((__mmask8)__U, (__v4si)_mm_unpackhi_epi32(__A, __B), (__v4si)_mm_setzero_si128()); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_mask_unpackhi_epi32(__m256i __W, __mmask8 __U, __m256i __A, __m256i __B) -{ +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_unpackhi_epi32(__m256i __W, __mmask8 __U, __m256i __A, + __m256i __B) { return (__m256i)__builtin_ia32_selectd_256((__mmask8)__U, (__v8si)_mm256_unpackhi_epi32(__A, __B), (__v8si)__W); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_maskz_unpackhi_epi32(__mmask8 __U, __m256i __A, __m256i __B) -{ +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_maskz_unpackhi_epi32(__mmask8 __U, __m256i __A, __m256i __B) { return (__m256i)__builtin_ia32_selectd_256((__mmask8)__U, (__v8si)_mm256_unpackhi_epi32(__A, __B), (__v8si)_mm256_setzero_si256()); } -static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_mask_unpackhi_epi64(__m128i __W, __mmask8 __U, __m128i __A, __m128i __B) -{ +static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask_unpackhi_epi64(__m128i __W, __mmask8 __U, __m128i __A, __m128i __B) { return (__m128i)__builtin_ia32_selectq_128((__mmask8)__U, (__v2di)_mm_unpackhi_epi64(__A, __B), (__v2di)__W); } -static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_maskz_unpackhi_epi64(__mmask8 __U, __m128i __A, __m128i __B) -{ +static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_maskz_unpackhi_epi64(__mmask8 __U, __m128i __A, __m128i __B) { return (__m128i)__builtin_ia32_selectq_128((__mmask8)__U, (__v2di)_mm_unpackhi_epi64(__A, __B), (__v2di)_mm_setzero_si128()); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_mask_unpackhi_epi64(__m256i __W, __mmask8 __U, __m256i __A, __m256i __B) -{ +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_unpackhi_epi64(__m256i __W, __mmask8 __U, __m256i __A, + __m256i __B) { return (__m256i)__builtin_ia32_selectq_256((__mmask8)__U, (__v4di)_mm256_unpackhi_epi64(__A, __B), (__v4di)__W); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_maskz_unpackhi_epi64(__mmask8 __U, __m256i __A, __m256i __B) -{ +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_maskz_unpackhi_epi64(__mmask8 __U, __m256i __A, __m256i __B) { return (__m256i)__builtin_ia32_selectq_256((__mmask8)__U, (__v4di)_mm256_unpackhi_epi64(__A, __B), (__v4di)_mm256_setzero_si256()); } -static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_mask_unpacklo_epi32(__m128i __W, __mmask8 __U, __m128i __A, __m128i __B) -{ +static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask_unpacklo_epi32(__m128i __W, __mmask8 __U, __m128i __A, __m128i __B) { return (__m128i)__builtin_ia32_selectd_128((__mmask8)__U, (__v4si)_mm_unpacklo_epi32(__A, __B), (__v4si)__W); } -static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_maskz_unpacklo_epi32(__mmask8 __U, __m128i __A, __m128i __B) -{ +static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_maskz_unpacklo_epi32(__mmask8 __U, __m128i __A, __m128i __B) { return (__m128i)__builtin_ia32_selectd_128((__mmask8)__U, (__v4si)_mm_unpacklo_epi32(__A, __B), (__v4si)_mm_setzero_si128()); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_mask_unpacklo_epi32(__m256i __W, __mmask8 __U, __m256i __A, __m256i __B) -{ +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_unpacklo_epi32(__m256i __W, __mmask8 __U, __m256i __A, + __m256i __B) { return (__m256i)__builtin_ia32_selectd_256((__mmask8)__U, (__v8si)_mm256_unpacklo_epi32(__A, __B), (__v8si)__W); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_maskz_unpacklo_epi32(__mmask8 __U, __m256i __A, __m256i __B) -{ +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_maskz_unpacklo_epi32(__mmask8 __U, __m256i __A, __m256i __B) { return (__m256i)__builtin_ia32_selectd_256((__mmask8)__U, (__v8si)_mm256_unpacklo_epi32(__A, __B), (__v8si)_mm256_setzero_si256()); } -static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_mask_unpacklo_epi64(__m128i __W, __mmask8 __U, __m128i __A, __m128i __B) -{ +static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask_unpacklo_epi64(__m128i __W, __mmask8 __U, __m128i __A, __m128i __B) { return (__m128i)__builtin_ia32_selectq_128((__mmask8)__U, (__v2di)_mm_unpacklo_epi64(__A, __B), (__v2di)__W); } -static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_maskz_unpacklo_epi64(__mmask8 __U, __m128i __A, __m128i __B) -{ +static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_maskz_unpacklo_epi64(__mmask8 __U, __m128i __A, __m128i __B) { return (__m128i)__builtin_ia32_selectq_128((__mmask8)__U, (__v2di)_mm_unpacklo_epi64(__A, __B), (__v2di)_mm_setzero_si128()); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_mask_unpacklo_epi64(__m256i __W, __mmask8 __U, __m256i __A, __m256i __B) -{ +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_unpacklo_epi64(__m256i __W, __mmask8 __U, __m256i __A, + __m256i __B) { return (__m256i)__builtin_ia32_selectq_256((__mmask8)__U, (__v4di)_mm256_unpacklo_epi64(__A, __B), (__v4di)__W); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_maskz_unpacklo_epi64(__mmask8 __U, __m256i __A, __m256i __B) -{ +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_maskz_unpacklo_epi64(__mmask8 __U, __m256i __A, __m256i __B) { return (__m256i)__builtin_ia32_selectq_256((__mmask8)__U, (__v4di)_mm256_unpacklo_epi64(__A, __B), (__v4di)_mm256_setzero_si256()); @@ -6594,17 +6551,15 @@ _mm256_broadcast_f32x4(__m128 __A) { 0, 1, 2, 3, 0, 1, 2, 3); } -static __inline__ __m256 __DEFAULT_FN_ATTRS256 -_mm256_mask_broadcast_f32x4(__m256 __O, __mmask8 __M, __m128 __A) -{ +static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_broadcast_f32x4(__m256 __O, __mmask8 __M, __m128 __A) { return (__m256)__builtin_ia32_selectps_256((__mmask8)__M, (__v8sf)_mm256_broadcast_f32x4(__A), (__v8sf)__O); } -static __inline__ __m256 __DEFAULT_FN_ATTRS256 -_mm256_maskz_broadcast_f32x4 (__mmask8 __M, __m128 __A) -{ +static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_maskz_broadcast_f32x4(__mmask8 __M, __m128 __A) { return (__m256)__builtin_ia32_selectps_256((__mmask8)__M, (__v8sf)_mm256_broadcast_f32x4(__A), (__v8sf)_mm256_setzero_ps()); @@ -6616,129 +6571,113 @@ _mm256_broadcast_i32x4(__m128i __A) { 0, 1, 2, 3, 0, 1, 2, 3); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_mask_broadcast_i32x4(__m256i __O, __mmask8 __M, __m128i __A) -{ +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_broadcast_i32x4(__m256i __O, __mmask8 __M, __m128i __A) { return (__m256i)__builtin_ia32_selectd_256((__mmask8)__M, (__v8si)_mm256_broadcast_i32x4(__A), (__v8si)__O); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_maskz_broadcast_i32x4(__mmask8 __M, __m128i __A) -{ +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_maskz_broadcast_i32x4(__mmask8 __M, __m128i __A) { return (__m256i)__builtin_ia32_selectd_256((__mmask8)__M, (__v8si)_mm256_broadcast_i32x4(__A), (__v8si)_mm256_setzero_si256()); } -static __inline__ __m256d __DEFAULT_FN_ATTRS256 -_mm256_mask_broadcastsd_pd (__m256d __O, __mmask8 __M, __m128d __A) -{ +static __inline__ __m256d __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_broadcastsd_pd(__m256d __O, __mmask8 __M, __m128d __A) { return (__m256d)__builtin_ia32_selectpd_256(__M, (__v4df) _mm256_broadcastsd_pd(__A), (__v4df) __O); } -static __inline__ __m256d __DEFAULT_FN_ATTRS256 -_mm256_maskz_broadcastsd_pd (__mmask8 __M, __m128d __A) -{ +static __inline__ __m256d __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_maskz_broadcastsd_pd(__mmask8 __M, __m128d __A) { return (__m256d)__builtin_ia32_selectpd_256(__M, (__v4df) _mm256_broadcastsd_pd(__A), (__v4df) _mm256_setzero_pd()); } -static __inline__ __m128 __DEFAULT_FN_ATTRS128 -_mm_mask_broadcastss_ps (__m128 __O, __mmask8 __M, __m128 __A) -{ +static __inline__ __m128 __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask_broadcastss_ps(__m128 __O, __mmask8 __M, __m128 __A) { return (__m128)__builtin_ia32_selectps_128(__M, (__v4sf) _mm_broadcastss_ps(__A), (__v4sf) __O); } -static __inline__ __m128 __DEFAULT_FN_ATTRS128 -_mm_maskz_broadcastss_ps (__mmask8 __M, __m128 __A) -{ +static __inline__ __m128 __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_maskz_broadcastss_ps(__mmask8 __M, __m128 __A) { return (__m128)__builtin_ia32_selectps_128(__M, (__v4sf) _mm_broadcastss_ps(__A), (__v4sf) _mm_setzero_ps()); } -static __inline__ __m256 __DEFAULT_FN_ATTRS256 -_mm256_mask_broadcastss_ps (__m256 __O, __mmask8 __M, __m128 __A) -{ +static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_broadcastss_ps(__m256 __O, __mmask8 __M, __m128 __A) { return (__m256)__builtin_ia32_selectps_256(__M, (__v8sf) _mm256_broadcastss_ps(__A), (__v8sf) __O); } -static __inline__ __m256 __DEFAULT_FN_ATTRS256 -_mm256_maskz_broadcastss_ps (__mmask8 __M, __m128 __A) -{ +static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_maskz_broadcastss_ps(__mmask8 __M, __m128 __A) { return (__m256)__builtin_ia32_selectps_256(__M, (__v8sf) _mm256_broadcastss_ps(__A), (__v8sf) _mm256_setzero_ps()); } -static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_mask_broadcastd_epi32 (__m128i __O, __mmask8 __M, __m128i __A) -{ +static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask_broadcastd_epi32(__m128i __O, __mmask8 __M, __m128i __A) { return (__m128i)__builtin_ia32_selectd_128(__M, (__v4si) _mm_broadcastd_epi32(__A), (__v4si) __O); } -static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_maskz_broadcastd_epi32 (__mmask8 __M, __m128i __A) -{ +static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_maskz_broadcastd_epi32(__mmask8 __M, __m128i __A) { return (__m128i)__builtin_ia32_selectd_128(__M, (__v4si) _mm_broadcastd_epi32(__A), (__v4si) _mm_setzero_si128()); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_mask_broadcastd_epi32 (__m256i __O, __mmask8 __M, __m128i __A) -{ +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_broadcastd_epi32(__m256i __O, __mmask8 __M, __m128i __A) { return (__m256i)__builtin_ia32_selectd_256(__M, (__v8si) _mm256_broadcastd_epi32(__A), (__v8si) __O); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_maskz_broadcastd_epi32 (__mmask8 __M, __m128i __A) -{ +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_maskz_broadcastd_epi32(__mmask8 __M, __m128i __A) { return (__m256i)__builtin_ia32_selectd_256(__M, (__v8si) _mm256_broadcastd_epi32(__A), (__v8si) _mm256_setzero_si256()); } -static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_mask_broadcastq_epi64 (__m128i __O, __mmask8 __M, __m128i __A) -{ +static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask_broadcastq_epi64(__m128i __O, __mmask8 __M, __m128i __A) { return (__m128i)__builtin_ia32_selectq_128(__M, (__v2di) _mm_broadcastq_epi64(__A), (__v2di) __O); } -static __inline__ __m128i __DEFAULT_FN_ATTRS128 -_mm_maskz_broadcastq_epi64 (__mmask8 __M, __m128i __A) -{ +static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_maskz_broadcastq_epi64(__mmask8 __M, __m128i __A) { return (__m128i)__builtin_ia32_selectq_128(__M, (__v2di) _mm_broadcastq_epi64(__A), (__v2di) _mm_setzero_si128()); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_mask_broadcastq_epi64 (__m256i __O, __mmask8 __M, __m128i __A) -{ +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_broadcastq_epi64(__m256i __O, __mmask8 __M, __m128i __A) { return (__m256i)__builtin_ia32_selectq_256(__M, (__v4di) _mm256_broadcastq_epi64(__A), (__v4di) __O); } -static __inline__ __m256i __DEFAULT_FN_ATTRS256 -_mm256_maskz_broadcastq_epi64 (__mmask8 __M, __m128i __A) -{ +static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_maskz_broadcastq_epi64(__mmask8 __M, __m128i __A) { return (__m256i)__builtin_ia32_selectq_256(__M, (__v4di) _mm256_broadcastq_epi64(__A), (__v4di) _mm256_setzero_si256()); @@ -8003,65 +7942,57 @@ _mm256_maskz_permutexvar_epi32(__mmask8 __M, __m256i __X, __m256i __Y) (__v4di)_mm256_alignr_epi64((A), (B), (imm)), \ (__v4di)_mm256_setzero_si256())) -static __inline__ __m128 __DEFAULT_FN_ATTRS128 -_mm_mask_movehdup_ps (__m128 __W, __mmask8 __U, __m128 __A) -{ +static __inline__ __m128 __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask_movehdup_ps(__m128 __W, __mmask8 __U, __m128 __A) { return (__m128)__builtin_ia32_selectps_128((__mmask8)__U, (__v4sf)_mm_movehdup_ps(__A), (__v4sf)__W); } -static __inline__ __m128 __DEFAULT_FN_ATTRS128 -_mm_maskz_movehdup_ps (__mmask8 __U, __m128 __A) -{ +static __inline__ __m128 __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_maskz_movehdup_ps(__mmask8 __U, __m128 __A) { return (__m128)__builtin_ia32_selectps_128((__mmask8)__U, (__v4sf)_mm_movehdup_ps(__A), (__v4sf)_mm_setzero_ps()); } -static __inline__ __m256 __DEFAULT_FN_ATTRS256 -_mm256_mask_movehdup_ps (__m256 __W, __mmask8 __U, __m256 __A) -{ +static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_movehdup_ps(__m256 __W, __mmask8 __U, __m256 __A) { return (__m256)__builtin_ia32_selectps_256((__mmask8)__U, (__v8sf)_mm256_movehdup_ps(__A), (__v8sf)__W); } -static __inline__ __m256 __DEFAULT_FN_ATTRS256 -_mm256_maskz_movehdup_ps (__mmask8 __U, __m256 __A) -{ +static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_maskz_movehdup_ps(__mmask8 __U, __m256 __A) { return (__m256)__builtin_ia32_selectps_256((__mmask8)__U, (__v8sf)_mm256_movehdup_ps(__A), (__v8sf)_mm256_setzero_ps()); } -static __inline__ __m128 __DEFAULT_FN_ATTRS128 -_mm_mask_moveldup_ps (__m128 __W, __mmask8 __U, __m128 __A) -{ +static __inline__ __m128 __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_mask_moveldup_ps(__m128 __W, __mmask8 __U, __m128 __A) { return (__m128)__builtin_ia32_selectps_128((__mmask8)__U, (__v4sf)_mm_moveldup_ps(__A), (__v4sf)__W); } -static __inline__ __m128 __DEFAULT_FN_ATTRS128 -_mm_maskz_moveldup_ps (__mmask8 __U, __m128 __A) -{ +static __inline__ __m128 __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_maskz_moveldup_ps(__mmask8 __U, __m128 __A) { return (__m128)__builtin_ia32_selectps_128((__mmask8)__U, (__v4sf)_mm_moveldup_ps(__A), (__v4sf)_mm_setzero_ps()); } -static __inline__ __m256 __DEFAULT_FN_ATTRS256 -_mm256_mask_moveldup_ps (__m256 __W, __mmask8 __U, __m256 __A) -{ +static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_mask_moveldup_ps(__m256 __W, __mmask8 __U, __m256 __A) { return (__m256)__builtin_ia32_selectps_256((__mmask8)__U, (__v8sf)_mm256_moveldup_ps(__A), (__v8sf)__W); } -static __inline__ __m256 __DEFAULT_FN_ATTRS256 -_mm256_maskz_moveldup_ps (__mmask8 __U, __m256 __A) -{ +static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR +_mm256_maskz_moveldup_ps(__mmask8 __U, __m256 __A) { return (__m256)__builtin_ia32_selectps_256((__mmask8)__U, (__v8sf)_mm256_moveldup_ps(__A), (__v8sf)_mm256_setzero_ps()); diff --git a/clang/lib/Headers/avxifmaintrin.h b/clang/lib/Headers/avxifmaintrin.h index 5c782d2a..a2ef601 100644 --- a/clang/lib/Headers/avxifmaintrin.h +++ b/clang/lib/Headers/avxifmaintrin.h @@ -15,12 +15,21 @@ #define __AVXIFMAINTRIN_H /* Define the default attributes for the functions in this file. */ +#if defined(__cplusplus) && (__cplusplus >= 201103L) +#define __DEFAULT_FN_ATTRS128 \ + constexpr __attribute__((__always_inline__, __nodebug__, \ + __target__("avxifma"), __min_vector_width__(128))) +#define __DEFAULT_FN_ATTRS256 \ + constexpr __attribute__((__always_inline__, __nodebug__, \ + __target__("avxifma"), __min_vector_width__(256))) +#else #define __DEFAULT_FN_ATTRS128 \ __attribute__((__always_inline__, __nodebug__, __target__("avxifma"), \ __min_vector_width__(128))) #define __DEFAULT_FN_ATTRS256 \ __attribute__((__always_inline__, __nodebug__, __target__("avxifma"), \ __min_vector_width__(256))) +#endif // must vex-encoding diff --git a/clang/lib/Headers/avxintrin.h b/clang/lib/Headers/avxintrin.h index d6ba19a..696ec31 100644 --- a/clang/lib/Headers/avxintrin.h +++ b/clang/lib/Headers/avxintrin.h @@ -694,9 +694,8 @@ _mm256_xor_ps(__m256 __a, __m256 __b) /// elements of a vector of [4 x double]. /// \returns A 256-bit vector of [4 x double] containing the horizontal sums of /// both operands. -static __inline __m256d __DEFAULT_FN_ATTRS -_mm256_hadd_pd(__m256d __a, __m256d __b) -{ +static __inline __m256d __DEFAULT_FN_ATTRS_CONSTEXPR +_mm256_hadd_pd(__m256d __a, __m256d __b) { return (__m256d)__builtin_ia32_haddpd256((__v4df)__a, (__v4df)__b); } @@ -717,9 +716,8 @@ _mm256_hadd_pd(__m256d __a, __m256d __b) /// index 2, 3, 6, 7 of a vector of [8 x float]. /// \returns A 256-bit vector of [8 x float] containing the horizontal sums of /// both operands. -static __inline __m256 __DEFAULT_FN_ATTRS -_mm256_hadd_ps(__m256 __a, __m256 __b) -{ +static __inline __m256 __DEFAULT_FN_ATTRS_CONSTEXPR _mm256_hadd_ps(__m256 __a, + __m256 __b) { return (__m256)__builtin_ia32_haddps256((__v8sf)__a, (__v8sf)__b); } @@ -740,9 +738,8 @@ _mm256_hadd_ps(__m256 __a, __m256 __b) /// odd-indexed elements of a vector of [4 x double]. /// \returns A 256-bit vector of [4 x double] containing the horizontal /// differences of both operands. -static __inline __m256d __DEFAULT_FN_ATTRS -_mm256_hsub_pd(__m256d __a, __m256d __b) -{ +static __inline __m256d __DEFAULT_FN_ATTRS_CONSTEXPR +_mm256_hsub_pd(__m256d __a, __m256d __b) { return (__m256d)__builtin_ia32_hsubpd256((__v4df)__a, (__v4df)__b); } @@ -763,9 +760,8 @@ _mm256_hsub_pd(__m256d __a, __m256d __b) /// elements with index 2, 3, 6, 7 of a vector of [8 x float]. /// \returns A 256-bit vector of [8 x float] containing the horizontal /// differences of both operands. -static __inline __m256 __DEFAULT_FN_ATTRS -_mm256_hsub_ps(__m256 __a, __m256 __b) -{ +static __inline __m256 __DEFAULT_FN_ATTRS_CONSTEXPR _mm256_hsub_ps(__m256 __a, + __m256 __b) { return (__m256)__builtin_ia32_hsubps256((__v8sf)__a, (__v8sf)__b); } @@ -2539,9 +2535,8 @@ _mm256_unpacklo_ps(__m256 __a, __m256 __b) { /// \param __b /// A 128-bit vector of [2 x double]. /// \returns the ZF flag in the EFLAGS register. -static __inline int __DEFAULT_FN_ATTRS128 -_mm_testz_pd(__m128d __a, __m128d __b) -{ +static __inline int __DEFAULT_FN_ATTRS128_CONSTEXPR _mm_testz_pd(__m128d __a, + __m128d __b) { return __builtin_ia32_vtestzpd((__v2df)__a, (__v2df)__b); } @@ -2568,9 +2563,8 @@ _mm_testz_pd(__m128d __a, __m128d __b) /// \param __b /// A 128-bit vector of [2 x double]. /// \returns the CF flag in the EFLAGS register. -static __inline int __DEFAULT_FN_ATTRS128 -_mm_testc_pd(__m128d __a, __m128d __b) -{ +static __inline int __DEFAULT_FN_ATTRS128_CONSTEXPR _mm_testc_pd(__m128d __a, + __m128d __b) { return __builtin_ia32_vtestcpd((__v2df)__a, (__v2df)__b); } @@ -2598,9 +2592,8 @@ _mm_testc_pd(__m128d __a, __m128d __b) /// \param __b /// A 128-bit vector of [2 x double]. /// \returns 1 if both the ZF and CF flags are set to 0, otherwise returns 0. -static __inline int __DEFAULT_FN_ATTRS128 -_mm_testnzc_pd(__m128d __a, __m128d __b) -{ +static __inline int __DEFAULT_FN_ATTRS128_CONSTEXPR +_mm_testnzc_pd(__m128d __a, __m128d __b) { return __builtin_ia32_vtestnzcpd((__v2df)__a, (__v2df)__b); } @@ -2627,9 +2620,8 @@ _mm_testnzc_pd(__m128d __a, __m128d __b) /// \param __b /// A 128-bit vector of [4 x float]. /// \returns the ZF flag. -static __inline int __DEFAULT_FN_ATTRS128 -_mm_testz_ps(__m128 __a, __m128 __b) -{ +static __inline int __DEFAULT_FN_ATTRS128_CONSTEXPR _mm_testz_ps(__m128 __a, + __m128 __b) { return __builtin_ia32_vtestzps((__v4sf)__a, (__v4sf)__b); } @@ -2656,9 +2648,8 @@ _mm_testz_ps(__m128 __a, __m128 __b) /// \param __b /// A 128-bit vector of [4 x float]. /// \returns the CF flag. -static __inline int __DEFAULT_FN_ATTRS128 -_mm_testc_ps(__m128 __a, __m128 __b) -{ +static __inline int __DEFAULT_FN_ATTRS128_CONSTEXPR _mm_testc_ps(__m128 __a, + __m128 __b) { return __builtin_ia32_vtestcps((__v4sf)__a, (__v4sf)__b); } @@ -2686,9 +2677,8 @@ _mm_testc_ps(__m128 __a, __m128 __b) /// \param __b /// A 128-bit vector of [4 x float]. /// \returns 1 if both the ZF and CF flags are set to 0, otherwise returns 0. -static __inline int __DEFAULT_FN_ATTRS128 -_mm_testnzc_ps(__m128 __a, __m128 __b) -{ +static __inline int __DEFAULT_FN_ATTRS128_CONSTEXPR _mm_testnzc_ps(__m128 __a, + __m128 __b) { return __builtin_ia32_vtestnzcps((__v4sf)__a, (__v4sf)__b); } @@ -2715,9 +2705,8 @@ _mm_testnzc_ps(__m128 __a, __m128 __b) /// \param __b /// A 256-bit vector of [4 x double]. /// \returns the ZF flag. -static __inline int __DEFAULT_FN_ATTRS -_mm256_testz_pd(__m256d __a, __m256d __b) -{ +static __inline int __DEFAULT_FN_ATTRS_CONSTEXPR _mm256_testz_pd(__m256d __a, + __m256d __b) { return __builtin_ia32_vtestzpd256((__v4df)__a, (__v4df)__b); } @@ -2744,9 +2733,8 @@ _mm256_testz_pd(__m256d __a, __m256d __b) /// \param __b /// A 256-bit vector of [4 x double]. /// \returns the CF flag. -static __inline int __DEFAULT_FN_ATTRS -_mm256_testc_pd(__m256d __a, __m256d __b) -{ +static __inline int __DEFAULT_FN_ATTRS_CONSTEXPR _mm256_testc_pd(__m256d __a, + __m256d __b) { return __builtin_ia32_vtestcpd256((__v4df)__a, (__v4df)__b); } @@ -2774,9 +2762,8 @@ _mm256_testc_pd(__m256d __a, __m256d __b) /// \param __b /// A 256-bit vector of [4 x double]. /// \returns 1 if both the ZF and CF flags are set to 0, otherwise returns 0. -static __inline int __DEFAULT_FN_ATTRS -_mm256_testnzc_pd(__m256d __a, __m256d __b) -{ +static __inline int __DEFAULT_FN_ATTRS_CONSTEXPR +_mm256_testnzc_pd(__m256d __a, __m256d __b) { return __builtin_ia32_vtestnzcpd256((__v4df)__a, (__v4df)__b); } @@ -2803,9 +2790,8 @@ _mm256_testnzc_pd(__m256d __a, __m256d __b) /// \param __b /// A 256-bit vector of [8 x float]. /// \returns the ZF flag. -static __inline int __DEFAULT_FN_ATTRS -_mm256_testz_ps(__m256 __a, __m256 __b) -{ +static __inline int __DEFAULT_FN_ATTRS_CONSTEXPR _mm256_testz_ps(__m256 __a, + __m256 __b) { return __builtin_ia32_vtestzps256((__v8sf)__a, (__v8sf)__b); } @@ -2832,9 +2818,8 @@ _mm256_testz_ps(__m256 __a, __m256 __b) /// \param __b /// A 256-bit vector of [8 x float]. /// \returns the CF flag. -static __inline int __DEFAULT_FN_ATTRS -_mm256_testc_ps(__m256 __a, __m256 __b) -{ +static __inline int __DEFAULT_FN_ATTRS_CONSTEXPR _mm256_testc_ps(__m256 __a, + __m256 __b) { return __builtin_ia32_vtestcps256((__v8sf)__a, (__v8sf)__b); } @@ -2862,9 +2847,8 @@ _mm256_testc_ps(__m256 __a, __m256 __b) /// \param __b /// A 256-bit vector of [8 x float]. /// \returns 1 if both the ZF and CF flags are set to 0, otherwise returns 0. -static __inline int __DEFAULT_FN_ATTRS -_mm256_testnzc_ps(__m256 __a, __m256 __b) -{ +static __inline int __DEFAULT_FN_ATTRS_CONSTEXPR _mm256_testnzc_ps(__m256 __a, + __m256 __b) { return __builtin_ia32_vtestnzcps256((__v8sf)__a, (__v8sf)__b); } @@ -2888,9 +2872,8 @@ _mm256_testnzc_ps(__m256 __a, __m256 __b) /// \param __b /// A 256-bit integer vector. /// \returns the ZF flag. -static __inline int __DEFAULT_FN_ATTRS -_mm256_testz_si256(__m256i __a, __m256i __b) -{ +static __inline int __DEFAULT_FN_ATTRS_CONSTEXPR +_mm256_testz_si256(__m256i __a, __m256i __b) { return __builtin_ia32_ptestz256((__v4di)__a, (__v4di)__b); } @@ -2914,9 +2897,8 @@ _mm256_testz_si256(__m256i __a, __m256i __b) /// \param __b /// A 256-bit integer vector. /// \returns the CF flag. -static __inline int __DEFAULT_FN_ATTRS -_mm256_testc_si256(__m256i __a, __m256i __b) -{ +static __inline int __DEFAULT_FN_ATTRS_CONSTEXPR +_mm256_testc_si256(__m256i __a, __m256i __b) { return __builtin_ia32_ptestc256((__v4di)__a, (__v4di)__b); } @@ -2941,9 +2923,8 @@ _mm256_testc_si256(__m256i __a, __m256i __b) /// \param __b /// A 256-bit integer vector. /// \returns 1 if both the ZF and CF flags are set to 0, otherwise returns 0. -static __inline int __DEFAULT_FN_ATTRS -_mm256_testnzc_si256(__m256i __a, __m256i __b) -{ +static __inline int __DEFAULT_FN_ATTRS_CONSTEXPR +_mm256_testnzc_si256(__m256i __a, __m256i __b) { return __builtin_ia32_ptestnzc256((__v4di)__a, (__v4di)__b); } diff --git a/clang/lib/Headers/float.h b/clang/lib/Headers/float.h index 84551af..30427c2 100644 --- a/clang/lib/Headers/float.h +++ b/clang/lib/Headers/float.h @@ -89,6 +89,9 @@ !defined(__STRICT_ANSI__) # undef INFINITY # undef NAN +# undef FLT_SNAN +# undef DBL_SNAN +# undef LDBL_SNAN #endif /* Characteristics of floating point types, C99 5.2.4.2.2 */ @@ -160,9 +163,15 @@ #if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 202311L) || \ !defined(__STRICT_ANSI__) + /* C23 5.2.5.3.2p28 */ +# define FLT_SNAN (__builtin_nansf("")) +# define DBL_SNAN (__builtin_nans("")) +# define LDBL_SNAN (__builtin_nansl("")) + /* C23 5.2.5.3.3p29-30 */ # define INFINITY (__builtin_inff()) # define NAN (__builtin_nanf("")) + /* C23 5.2.5.3.3p32 */ # define FLT_NORM_MAX __FLT_NORM_MAX__ # define DBL_NORM_MAX __DBL_NORM_MAX__ diff --git a/clang/lib/Headers/pmmintrin.h b/clang/lib/Headers/pmmintrin.h index f0c9b2b..42bd343 100644 --- a/clang/lib/Headers/pmmintrin.h +++ b/clang/lib/Headers/pmmintrin.h @@ -83,9 +83,8 @@ _mm_addsub_ps(__m128 __a, __m128 __b) /// destination. /// \returns A 128-bit vector of [4 x float] containing the horizontal sums of /// both operands. -static __inline__ __m128 __DEFAULT_FN_ATTRS -_mm_hadd_ps(__m128 __a, __m128 __b) -{ +static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR _mm_hadd_ps(__m128 __a, + __m128 __b) { return __builtin_ia32_haddps((__v4sf)__a, (__v4sf)__b); } @@ -106,9 +105,8 @@ _mm_hadd_ps(__m128 __a, __m128 __b) /// bits of the destination. /// \returns A 128-bit vector of [4 x float] containing the horizontal /// differences of both operands. -static __inline__ __m128 __DEFAULT_FN_ATTRS -_mm_hsub_ps(__m128 __a, __m128 __b) -{ +static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR _mm_hsub_ps(__m128 __a, + __m128 __b) { return __builtin_ia32_hsubps((__v4sf)__a, (__v4sf)__b); } @@ -168,9 +166,8 @@ _mm_moveldup_ps(__m128 __a) /// A 128-bit vector of [2 x double] containing the right source operand. /// \returns A 128-bit vector of [2 x double] containing the alternating sums /// and differences of both operands. -static __inline__ __m128d __DEFAULT_FN_ATTRS -_mm_addsub_pd(__m128d __a, __m128d __b) -{ +static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR +_mm_addsub_pd(__m128d __a, __m128d __b) { return __builtin_ia32_addsubpd((__v2df)__a, (__v2df)__b); } @@ -191,9 +188,8 @@ _mm_addsub_pd(__m128d __a, __m128d __b) /// destination. /// \returns A 128-bit vector of [2 x double] containing the horizontal sums of /// both operands. -static __inline__ __m128d __DEFAULT_FN_ATTRS -_mm_hadd_pd(__m128d __a, __m128d __b) -{ +static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR +_mm_hadd_pd(__m128d __a, __m128d __b) { return __builtin_ia32_haddpd((__v2df)__a, (__v2df)__b); } @@ -214,9 +210,8 @@ _mm_hadd_pd(__m128d __a, __m128d __b) /// the destination. /// \returns A 128-bit vector of [2 x double] containing the horizontal /// differences of both operands. -static __inline__ __m128d __DEFAULT_FN_ATTRS -_mm_hsub_pd(__m128d __a, __m128d __b) -{ +static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR +_mm_hsub_pd(__m128d __a, __m128d __b) { return __builtin_ia32_hsubpd((__v2df)__a, (__v2df)__b); } diff --git a/clang/lib/Headers/ptrauth.h b/clang/lib/Headers/ptrauth.h index f902ca1..ad28f06 100644 --- a/clang/lib/Headers/ptrauth.h +++ b/clang/lib/Headers/ptrauth.h @@ -241,6 +241,18 @@ typedef __UINTPTR_TYPE__ ptrauth_generic_signature_t; #define ptrauth_type_discriminator(__type) \ __builtin_ptrauth_type_discriminator(__type) +/* Compute the constant discriminator used by Clang to sign pointers with the + given C function pointer type. + + A call to this function is an integer constant expression. */ +#if __has_feature(ptrauth_function_pointer_type_discrimination) +#define ptrauth_function_pointer_type_discriminator(__type) \ + __builtin_ptrauth_type_discriminator(__type) +#else +#define ptrauth_function_pointer_type_discriminator(__type) \ + ((ptrauth_extra_data_t)0) +#endif + /* Compute a signature for the given pair of pointer-sized values. The order of the arguments is significant. @@ -372,6 +384,8 @@ typedef __UINTPTR_TYPE__ ptrauth_generic_signature_t; }) #define ptrauth_type_discriminator(__type) ((ptrauth_extra_data_t)0) +#define ptrauth_function_pointer_type_discriminator(__type) \ + ((ptrauth_extra_data_t)0) #define ptrauth_sign_generic_data(__value, __data) \ ({ \ diff --git a/clang/lib/Headers/smmintrin.h b/clang/lib/Headers/smmintrin.h index 5e63a1a..4f197d5 100644 --- a/clang/lib/Headers/smmintrin.h +++ b/clang/lib/Headers/smmintrin.h @@ -1093,8 +1093,8 @@ _mm_max_epu32(__m128i __V1, __m128i __V2) { /// \param __V /// A 128-bit integer vector selecting which bits to test in operand \a __M. /// \returns TRUE if the specified bits are all zeros; FALSE otherwise. -static __inline__ int __DEFAULT_FN_ATTRS _mm_testz_si128(__m128i __M, - __m128i __V) { +static __inline__ int __DEFAULT_FN_ATTRS_CONSTEXPR +_mm_testz_si128(__m128i __M, __m128i __V) { return __builtin_ia32_ptestz128((__v2di)__M, (__v2di)__V); } @@ -1110,8 +1110,8 @@ static __inline__ int __DEFAULT_FN_ATTRS _mm_testz_si128(__m128i __M, /// \param __V /// A 128-bit integer vector selecting which bits to test in operand \a __M. /// \returns TRUE if the specified bits are all ones; FALSE otherwise. -static __inline__ int __DEFAULT_FN_ATTRS _mm_testc_si128(__m128i __M, - __m128i __V) { +static __inline__ int __DEFAULT_FN_ATTRS_CONSTEXPR +_mm_testc_si128(__m128i __M, __m128i __V) { return __builtin_ia32_ptestc128((__v2di)__M, (__v2di)__V); } @@ -1128,8 +1128,8 @@ static __inline__ int __DEFAULT_FN_ATTRS _mm_testc_si128(__m128i __M, /// A 128-bit integer vector selecting which bits to test in operand \a __M. /// \returns TRUE if the specified bits are neither all zeros nor all ones; /// FALSE otherwise. -static __inline__ int __DEFAULT_FN_ATTRS _mm_testnzc_si128(__m128i __M, - __m128i __V) { +static __inline__ int __DEFAULT_FN_ATTRS_CONSTEXPR +_mm_testnzc_si128(__m128i __M, __m128i __V) { return __builtin_ia32_ptestnzc128((__v2di)__M, (__v2di)__V); } diff --git a/clang/lib/Headers/tmmintrin.h b/clang/lib/Headers/tmmintrin.h index 3fc9f98..ee96caa 100644 --- a/clang/lib/Headers/tmmintrin.h +++ b/clang/lib/Headers/tmmintrin.h @@ -202,10 +202,9 @@ _mm_abs_epi32(__m128i __a) { /// destination. /// \returns A 128-bit vector of [8 x i16] containing the horizontal sums of /// both operands. -static __inline__ __m128i __DEFAULT_FN_ATTRS -_mm_hadd_epi16(__m128i __a, __m128i __b) -{ - return (__m128i)__builtin_ia32_phaddw128((__v8hi)__a, (__v8hi)__b); +static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR +_mm_hadd_epi16(__m128i __a, __m128i __b) { + return (__m128i)__builtin_ia32_phaddw128((__v8hi)__a, (__v8hi)__b); } /// Horizontally adds the adjacent pairs of values contained in 2 packed @@ -225,10 +224,9 @@ _mm_hadd_epi16(__m128i __a, __m128i __b) /// destination. /// \returns A 128-bit vector of [4 x i32] containing the horizontal sums of /// both operands. -static __inline__ __m128i __DEFAULT_FN_ATTRS -_mm_hadd_epi32(__m128i __a, __m128i __b) -{ - return (__m128i)__builtin_ia32_phaddd128((__v4si)__a, (__v4si)__b); +static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR +_mm_hadd_epi32(__m128i __a, __m128i __b) { + return (__m128i)__builtin_ia32_phaddd128((__v4si)__a, (__v4si)__b); } /// Horizontally adds the adjacent pairs of values contained in 2 packed @@ -248,11 +246,10 @@ _mm_hadd_epi32(__m128i __a, __m128i __b) /// destination. /// \returns A 64-bit vector of [4 x i16] containing the horizontal sums of both /// operands. -static __inline__ __m64 __DEFAULT_FN_ATTRS -_mm_hadd_pi16(__m64 __a, __m64 __b) -{ - return __trunc64(__builtin_ia32_phaddw128( - (__v8hi)__builtin_shufflevector(__a, __b, 0, 1), (__v8hi){})); +static __inline__ __m64 __DEFAULT_FN_ATTRS_CONSTEXPR _mm_hadd_pi16(__m64 __a, + __m64 __b) { + return __trunc64(__builtin_ia32_phaddw128( + (__v8hi)__builtin_shufflevector(__a, __b, 0, 1), (__v8hi){})); } /// Horizontally adds the adjacent pairs of values contained in 2 packed @@ -272,11 +269,10 @@ _mm_hadd_pi16(__m64 __a, __m64 __b) /// destination. /// \returns A 64-bit vector of [2 x i32] containing the horizontal sums of both /// operands. -static __inline__ __m64 __DEFAULT_FN_ATTRS -_mm_hadd_pi32(__m64 __a, __m64 __b) -{ - return __trunc64(__builtin_ia32_phaddd128( - (__v4si)__builtin_shufflevector(__a, __b, 0, 1), (__v4si){})); +static __inline__ __m64 __DEFAULT_FN_ATTRS_CONSTEXPR _mm_hadd_pi32(__m64 __a, + __m64 __b) { + return __trunc64(__builtin_ia32_phaddd128( + (__v4si)__builtin_shufflevector(__a, __b, 0, 1), (__v4si){})); } /// Horizontally adds, with saturation, the adjacent pairs of values contained @@ -299,10 +295,9 @@ _mm_hadd_pi32(__m64 __a, __m64 __b) /// destination. /// \returns A 128-bit vector of [8 x i16] containing the horizontal saturated /// sums of both operands. -static __inline__ __m128i __DEFAULT_FN_ATTRS -_mm_hadds_epi16(__m128i __a, __m128i __b) -{ - return (__m128i)__builtin_ia32_phaddsw128((__v8hi)__a, (__v8hi)__b); +static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR +_mm_hadds_epi16(__m128i __a, __m128i __b) { + return (__m128i)__builtin_ia32_phaddsw128((__v8hi)__a, (__v8hi)__b); } /// Horizontally adds, with saturation, the adjacent pairs of values contained @@ -325,11 +320,10 @@ _mm_hadds_epi16(__m128i __a, __m128i __b) /// destination. /// \returns A 64-bit vector of [4 x i16] containing the horizontal saturated /// sums of both operands. -static __inline__ __m64 __DEFAULT_FN_ATTRS -_mm_hadds_pi16(__m64 __a, __m64 __b) -{ - return __trunc64(__builtin_ia32_phaddsw128( - (__v8hi)__builtin_shufflevector(__a, __b, 0, 1), (__v8hi){})); +static __inline__ __m64 __DEFAULT_FN_ATTRS_CONSTEXPR _mm_hadds_pi16(__m64 __a, + __m64 __b) { + return __trunc64(__builtin_ia32_phaddsw128( + (__v8hi)__builtin_shufflevector(__a, __b, 0, 1), (__v8hi){})); } /// Horizontally subtracts the adjacent pairs of values contained in 2 @@ -349,10 +343,9 @@ _mm_hadds_pi16(__m64 __a, __m64 __b) /// the destination. /// \returns A 128-bit vector of [8 x i16] containing the horizontal differences /// of both operands. -static __inline__ __m128i __DEFAULT_FN_ATTRS -_mm_hsub_epi16(__m128i __a, __m128i __b) -{ - return (__m128i)__builtin_ia32_phsubw128((__v8hi)__a, (__v8hi)__b); +static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR +_mm_hsub_epi16(__m128i __a, __m128i __b) { + return (__m128i)__builtin_ia32_phsubw128((__v8hi)__a, (__v8hi)__b); } /// Horizontally subtracts the adjacent pairs of values contained in 2 @@ -372,10 +365,9 @@ _mm_hsub_epi16(__m128i __a, __m128i __b) /// the destination. /// \returns A 128-bit vector of [4 x i32] containing the horizontal differences /// of both operands. -static __inline__ __m128i __DEFAULT_FN_ATTRS -_mm_hsub_epi32(__m128i __a, __m128i __b) -{ - return (__m128i)__builtin_ia32_phsubd128((__v4si)__a, (__v4si)__b); +static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR +_mm_hsub_epi32(__m128i __a, __m128i __b) { + return (__m128i)__builtin_ia32_phsubd128((__v4si)__a, (__v4si)__b); } /// Horizontally subtracts the adjacent pairs of values contained in 2 @@ -395,11 +387,10 @@ _mm_hsub_epi32(__m128i __a, __m128i __b) /// the destination. /// \returns A 64-bit vector of [4 x i16] containing the horizontal differences /// of both operands. -static __inline__ __m64 __DEFAULT_FN_ATTRS -_mm_hsub_pi16(__m64 __a, __m64 __b) -{ - return __trunc64(__builtin_ia32_phsubw128( - (__v8hi)__builtin_shufflevector(__a, __b, 0, 1), (__v8hi){})); +static __inline__ __m64 __DEFAULT_FN_ATTRS_CONSTEXPR _mm_hsub_pi16(__m64 __a, + __m64 __b) { + return __trunc64(__builtin_ia32_phsubw128( + (__v8hi)__builtin_shufflevector(__a, __b, 0, 1), (__v8hi){})); } /// Horizontally subtracts the adjacent pairs of values contained in 2 @@ -419,11 +410,10 @@ _mm_hsub_pi16(__m64 __a, __m64 __b) /// the destination. /// \returns A 64-bit vector of [2 x i32] containing the horizontal differences /// of both operands. -static __inline__ __m64 __DEFAULT_FN_ATTRS -_mm_hsub_pi32(__m64 __a, __m64 __b) -{ - return __trunc64(__builtin_ia32_phsubd128( - (__v4si)__builtin_shufflevector(__a, __b, 0, 1), (__v4si){})); +static __inline__ __m64 __DEFAULT_FN_ATTRS_CONSTEXPR _mm_hsub_pi32(__m64 __a, + __m64 __b) { + return __trunc64(__builtin_ia32_phsubd128( + (__v4si)__builtin_shufflevector(__a, __b, 0, 1), (__v4si){})); } /// Horizontally subtracts, with saturation, the adjacent pairs of values @@ -446,10 +436,9 @@ _mm_hsub_pi32(__m64 __a, __m64 __b) /// the destination. /// \returns A 128-bit vector of [8 x i16] containing the horizontal saturated /// differences of both operands. -static __inline__ __m128i __DEFAULT_FN_ATTRS -_mm_hsubs_epi16(__m128i __a, __m128i __b) -{ - return (__m128i)__builtin_ia32_phsubsw128((__v8hi)__a, (__v8hi)__b); +static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR +_mm_hsubs_epi16(__m128i __a, __m128i __b) { + return (__m128i)__builtin_ia32_phsubsw128((__v8hi)__a, (__v8hi)__b); } /// Horizontally subtracts, with saturation, the adjacent pairs of values @@ -472,11 +461,10 @@ _mm_hsubs_epi16(__m128i __a, __m128i __b) /// the destination. /// \returns A 64-bit vector of [4 x i16] containing the horizontal saturated /// differences of both operands. -static __inline__ __m64 __DEFAULT_FN_ATTRS -_mm_hsubs_pi16(__m64 __a, __m64 __b) -{ - return __trunc64(__builtin_ia32_phsubsw128( - (__v8hi)__builtin_shufflevector(__a, __b, 0, 1), (__v8hi){})); +static __inline__ __m64 __DEFAULT_FN_ATTRS_CONSTEXPR _mm_hsubs_pi16(__m64 __a, + __m64 __b) { + return __trunc64(__builtin_ia32_phsubsw128( + (__v8hi)__builtin_shufflevector(__a, __b, 0, 1), (__v8hi){})); } /// Multiplies corresponding pairs of packed 8-bit unsigned integer @@ -556,10 +544,9 @@ _mm_maddubs_pi16(__m64 __a, __m64 __b) { /// A 128-bit vector of [8 x i16] containing one of the source operands. /// \returns A 128-bit vector of [8 x i16] containing the rounded and scaled /// products of both operands. -static __inline__ __m128i __DEFAULT_FN_ATTRS -_mm_mulhrs_epi16(__m128i __a, __m128i __b) -{ - return (__m128i)__builtin_ia32_pmulhrsw128((__v8hi)__a, (__v8hi)__b); +static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_mulhrs_epi16(__m128i __a, + __m128i __b) { + return (__m128i)__builtin_ia32_pmulhrsw128((__v8hi)__a, (__v8hi)__b); } /// Multiplies packed 16-bit signed integer values, truncates the 32-bit @@ -603,10 +590,9 @@ _mm_mulhrs_pi16(__m64 __a, __m64 __b) /// Bits [6:4] Reserved. \n /// Bits [3:0] select the source byte to be copied. /// \returns A 128-bit integer vector containing the copied or cleared values. -static __inline__ __m128i __DEFAULT_FN_ATTRS -_mm_shuffle_epi8(__m128i __a, __m128i __b) -{ - return (__m128i)__builtin_ia32_pshufb128((__v16qi)__a, (__v16qi)__b); +static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR +_mm_shuffle_epi8(__m128i __a, __m128i __b) { + return (__m128i)__builtin_ia32_pshufb128((__v16qi)__a, (__v16qi)__b); } /// Copies the 8-bit integers from a 64-bit integer vector to the @@ -628,13 +614,12 @@ _mm_shuffle_epi8(__m128i __a, __m128i __b) /// destination. \n /// Bits [2:0] select the source byte to be copied. /// \returns A 64-bit integer vector containing the copied or cleared values. -static __inline__ __m64 __DEFAULT_FN_ATTRS -_mm_shuffle_pi8(__m64 __a, __m64 __b) -{ - return __trunc64(__builtin_ia32_pshufb128( - (__v16qi)__builtin_shufflevector( - (__v2si)(__a), __extension__ (__v2si){}, 0, 1, 0, 1), - (__v16qi)__anyext128(__b))); +static __inline__ __m64 __DEFAULT_FN_ATTRS_CONSTEXPR +_mm_shuffle_pi8(__m64 __a, __m64 __b) { + return __trunc64(__builtin_ia32_pshufb128( + (__v16qi)__builtin_shufflevector((__v2si)(__a), __extension__(__v2si){}, + 0, 1, 0, 1), + (__v16qi)__zext128(__b))); } /// For each 8-bit integer in the first source operand, perform one of diff --git a/clang/lib/Headers/xmmintrin.h b/clang/lib/Headers/xmmintrin.h index d876b47..605409c 100644 --- a/clang/lib/Headers/xmmintrin.h +++ b/clang/lib/Headers/xmmintrin.h @@ -2416,9 +2416,8 @@ _mm_min_pu8(__m64 __a, __m64 __b) { /// A 64-bit integer vector containing the values with bits to be extracted. /// \returns The most significant bit from each 8-bit element in \a __a, /// written to bits [7:0]. -static __inline__ int __DEFAULT_FN_ATTRS_SSE2 -_mm_movemask_pi8(__m64 __a) -{ +static __inline__ int __DEFAULT_FN_ATTRS_SSE2_CONSTEXPR +_mm_movemask_pi8(__m64 __a) { return __builtin_ia32_pmovmskb128((__v16qi)__zext128(__a)); } diff --git a/clang/lib/Index/IndexTypeSourceInfo.cpp b/clang/lib/Index/IndexTypeSourceInfo.cpp index 74c6c11..3c1e038 100644 --- a/clang/lib/Index/IndexTypeSourceInfo.cpp +++ b/clang/lib/Index/IndexTypeSourceInfo.cpp @@ -117,7 +117,7 @@ public: } bool VisitTagTypeLoc(TagTypeLoc TL) { - TagDecl *D = TL.getOriginalDecl(); + TagDecl *D = TL.getDecl(); if (!IndexCtx.shouldIndexFunctionLocalSymbols() && D->getParentFunctionOrMethod()) return true; diff --git a/clang/lib/Index/USRGeneration.cpp b/clang/lib/Index/USRGeneration.cpp index c78d66f..08835ea 100644 --- a/clang/lib/Index/USRGeneration.cpp +++ b/clang/lib/Index/USRGeneration.cpp @@ -911,11 +911,10 @@ void USRGenerator::VisitType(QualType T) { } if (const TagType *TT = T->getAs<TagType>()) { if (const auto *ICNT = dyn_cast<InjectedClassNameType>(TT)) { - T = ICNT->getOriginalDecl()->getCanonicalTemplateSpecializationType( - Ctx); + T = ICNT->getDecl()->getCanonicalTemplateSpecializationType(Ctx); } else { Out << '$'; - VisitTagDecl(TT->getOriginalDecl()); + VisitTagDecl(TT->getDecl()); return; } } diff --git a/clang/lib/InstallAPI/HeaderFile.cpp b/clang/lib/InstallAPI/HeaderFile.cpp index 0b7041e..d736a0a 100644 --- a/clang/lib/InstallAPI/HeaderFile.cpp +++ b/clang/lib/InstallAPI/HeaderFile.cpp @@ -38,7 +38,7 @@ std::optional<std::string> createIncludeHeaderName(const StringRef FullPath) { bool isHeaderFile(StringRef Path) { return StringSwitch<bool>(sys::path::extension(Path)) - .Cases(".h", ".H", ".hh", ".hpp", ".hxx", true) + .Cases({".h", ".H", ".hh", ".hpp", ".hxx"}, true) .Default(false); } diff --git a/clang/lib/InstallAPI/Visitor.cpp b/clang/lib/InstallAPI/Visitor.cpp index f12e040..53fbc36 100644 --- a/clang/lib/InstallAPI/Visitor.cpp +++ b/clang/lib/InstallAPI/Visitor.cpp @@ -543,8 +543,8 @@ void InstallAPIVisitor::emitVTableSymbols(const CXXRecordDecl *D, } for (const auto &It : D->bases()) { - const CXXRecordDecl *Base = cast<CXXRecordDecl>( - It.getType()->castAs<RecordType>()->getOriginalDecl()); + const auto *Base = + cast<CXXRecordDecl>(It.getType()->castAs<RecordType>()->getDecl()); const auto BaseAccess = getAccessForDecl(Base); if (!BaseAccess) continue; diff --git a/clang/lib/Interpreter/InterpreterValuePrinter.cpp b/clang/lib/Interpreter/InterpreterValuePrinter.cpp index 54abfa6..0ed02f3 100644 --- a/clang/lib/Interpreter/InterpreterValuePrinter.cpp +++ b/clang/lib/Interpreter/InterpreterValuePrinter.cpp @@ -66,10 +66,10 @@ static std::string QualTypeToString(ASTContext &Ctx, QualType QT) { const QualType NonRefTy = QT.getNonReferenceType(); if (const auto *TTy = llvm::dyn_cast<TagType>(NonRefTy)) - return DeclTypeToString(NonRefTy, TTy->getOriginalDecl()); + return DeclTypeToString(NonRefTy, TTy->getDecl()); if (const auto *TRy = dyn_cast<RecordType>(NonRefTy)) - return DeclTypeToString(NonRefTy, TRy->getOriginalDecl()); + return DeclTypeToString(NonRefTy, TRy->getDecl()); const QualType Canon = NonRefTy.getCanonicalType(); @@ -555,7 +555,7 @@ llvm::Expected<Expr *> Interpreter::convertExprToValue(Expr *E) { InterfaceKind Kind = V.computeInterfaceKind(DesugaredTy); switch (Kind) { case InterfaceKind::WithAlloc: - LLVM_FALLTHROUGH; + [[fallthrough]]; case InterfaceKind::CopyArray: { // __clang_Interpreter_SetValueWithAlloc. ExprResult AllocCall = diff --git a/clang/lib/Lex/PPDirectives.cpp b/clang/lib/Lex/PPDirectives.cpp index 360593d..6a5e5d4 100644 --- a/clang/lib/Lex/PPDirectives.cpp +++ b/clang/lib/Lex/PPDirectives.cpp @@ -248,50 +248,67 @@ static bool warnByDefaultOnWrongCase(StringRef Include) { // The standard C/C++ and Posix headers return llvm::StringSwitch<bool>(LowerInclude) - // C library headers - .Cases("assert.h", "complex.h", "ctype.h", "errno.h", "fenv.h", true) - .Cases("float.h", "inttypes.h", "iso646.h", "limits.h", "locale.h", true) - .Cases("math.h", "setjmp.h", "signal.h", "stdalign.h", "stdarg.h", true) - .Cases("stdatomic.h", "stdbool.h", "stdckdint.h", "stdcountof.h", true) - .Cases("stddef.h", "stdint.h", "stdio.h", "stdlib.h", "stdnoreturn.h", true) - .Cases("string.h", "tgmath.h", "threads.h", "time.h", "uchar.h", true) - .Cases("wchar.h", "wctype.h", true) - - // C++ headers for C library facilities - .Cases("cassert", "ccomplex", "cctype", "cerrno", "cfenv", true) - .Cases("cfloat", "cinttypes", "ciso646", "climits", "clocale", true) - .Cases("cmath", "csetjmp", "csignal", "cstdalign", "cstdarg", true) - .Cases("cstdbool", "cstddef", "cstdint", "cstdio", "cstdlib", true) - .Cases("cstring", "ctgmath", "ctime", "cuchar", "cwchar", true) - .Case("cwctype", true) - - // C++ library headers - .Cases("algorithm", "fstream", "list", "regex", "thread", true) - .Cases("array", "functional", "locale", "scoped_allocator", "tuple", true) - .Cases("atomic", "future", "map", "set", "type_traits", true) - .Cases("bitset", "initializer_list", "memory", "shared_mutex", "typeindex", true) - .Cases("chrono", "iomanip", "mutex", "sstream", "typeinfo", true) - .Cases("codecvt", "ios", "new", "stack", "unordered_map", true) - .Cases("complex", "iosfwd", "numeric", "stdexcept", "unordered_set", true) - .Cases("condition_variable", "iostream", "ostream", "streambuf", "utility", true) - .Cases("deque", "istream", "queue", "string", "valarray", true) - .Cases("exception", "iterator", "random", "strstream", "vector", true) - .Cases("forward_list", "limits", "ratio", "system_error", true) - - // POSIX headers (which aren't also C headers) - .Cases("aio.h", "arpa/inet.h", "cpio.h", "dirent.h", "dlfcn.h", true) - .Cases("fcntl.h", "fmtmsg.h", "fnmatch.h", "ftw.h", "glob.h", true) - .Cases("grp.h", "iconv.h", "langinfo.h", "libgen.h", "monetary.h", true) - .Cases("mqueue.h", "ndbm.h", "net/if.h", "netdb.h", "netinet/in.h", true) - .Cases("netinet/tcp.h", "nl_types.h", "poll.h", "pthread.h", "pwd.h", true) - .Cases("regex.h", "sched.h", "search.h", "semaphore.h", "spawn.h", true) - .Cases("strings.h", "stropts.h", "sys/ipc.h", "sys/mman.h", "sys/msg.h", true) - .Cases("sys/resource.h", "sys/select.h", "sys/sem.h", "sys/shm.h", "sys/socket.h", true) - .Cases("sys/stat.h", "sys/statvfs.h", "sys/time.h", "sys/times.h", "sys/types.h", true) - .Cases("sys/uio.h", "sys/un.h", "sys/utsname.h", "sys/wait.h", "syslog.h", true) - .Cases("tar.h", "termios.h", "trace.h", "ulimit.h", true) - .Cases("unistd.h", "utime.h", "utmpx.h", "wordexp.h", true) - .Default(false); + // C library headers + .Cases({"assert.h", "complex.h", "ctype.h", "errno.h", "fenv.h"}, true) + .Cases({"float.h", "inttypes.h", "iso646.h", "limits.h", "locale.h"}, + true) + .Cases({"math.h", "setjmp.h", "signal.h", "stdalign.h", "stdarg.h"}, true) + .Cases({"stdatomic.h", "stdbool.h", "stdckdint.h", "stdcountof.h"}, true) + .Cases({"stddef.h", "stdint.h", "stdio.h", "stdlib.h", "stdnoreturn.h"}, + true) + .Cases({"string.h", "tgmath.h", "threads.h", "time.h", "uchar.h"}, true) + .Cases({"wchar.h", "wctype.h"}, true) + + // C++ headers for C library facilities + .Cases({"cassert", "ccomplex", "cctype", "cerrno", "cfenv"}, true) + .Cases({"cfloat", "cinttypes", "ciso646", "climits", "clocale"}, true) + .Cases({"cmath", "csetjmp", "csignal", "cstdalign", "cstdarg"}, true) + .Cases({"cstdbool", "cstddef", "cstdint", "cstdio", "cstdlib"}, true) + .Cases({"cstring", "ctgmath", "ctime", "cuchar", "cwchar"}, true) + .Case("cwctype", true) + + // C++ library headers + .Cases({"algorithm", "fstream", "list", "regex", "thread"}, true) + .Cases({"array", "functional", "locale", "scoped_allocator", "tuple"}, + true) + .Cases({"atomic", "future", "map", "set", "type_traits"}, true) + .Cases( + {"bitset", "initializer_list", "memory", "shared_mutex", "typeindex"}, + true) + .Cases({"chrono", "iomanip", "mutex", "sstream", "typeinfo"}, true) + .Cases({"codecvt", "ios", "new", "stack", "unordered_map"}, true) + .Cases({"complex", "iosfwd", "numeric", "stdexcept", "unordered_set"}, + true) + .Cases( + {"condition_variable", "iostream", "ostream", "streambuf", "utility"}, + true) + .Cases({"deque", "istream", "queue", "string", "valarray"}, true) + .Cases({"exception", "iterator", "random", "strstream", "vector"}, true) + .Cases({"forward_list", "limits", "ratio", "system_error"}, true) + + // POSIX headers (which aren't also C headers) + .Cases({"aio.h", "arpa/inet.h", "cpio.h", "dirent.h", "dlfcn.h"}, true) + .Cases({"fcntl.h", "fmtmsg.h", "fnmatch.h", "ftw.h", "glob.h"}, true) + .Cases({"grp.h", "iconv.h", "langinfo.h", "libgen.h", "monetary.h"}, true) + .Cases({"mqueue.h", "ndbm.h", "net/if.h", "netdb.h", "netinet/in.h"}, + true) + .Cases({"netinet/tcp.h", "nl_types.h", "poll.h", "pthread.h", "pwd.h"}, + true) + .Cases({"regex.h", "sched.h", "search.h", "semaphore.h", "spawn.h"}, true) + .Cases({"strings.h", "stropts.h", "sys/ipc.h", "sys/mman.h", "sys/msg.h"}, + true) + .Cases({"sys/resource.h", "sys/select.h", "sys/sem.h", "sys/shm.h", + "sys/socket.h"}, + true) + .Cases({"sys/stat.h", "sys/statvfs.h", "sys/time.h", "sys/times.h", + "sys/types.h"}, + true) + .Cases( + {"sys/uio.h", "sys/un.h", "sys/utsname.h", "sys/wait.h", "syslog.h"}, + true) + .Cases({"tar.h", "termios.h", "trace.h", "ulimit.h"}, true) + .Cases({"unistd.h", "utime.h", "utmpx.h", "wordexp.h"}, true) + .Default(false); } /// Find a similar string in `Candidates`. @@ -3648,14 +3665,14 @@ Preprocessor::LexEmbedParameters(Token &CurTok, bool ForHasEmbed) { std::pair<tok::TokenKind, SourceLocation> Matches) { Diag(CurTok, diag::err_expected) << Expected; Diag(Matches.second, diag::note_matching) << Matches.first; - if (CurTok.isNot(tok::eod)) + if (CurTok.isNot(EndTokenKind)) DiscardUntilEndOfDirective(CurTok); }; auto ExpectOrDiagAndSkipToEOD = [&](tok::TokenKind Kind) { if (CurTok.isNot(Kind)) { Diag(CurTok, diag::err_expected) << Kind; - if (CurTok.isNot(tok::eod)) + if (CurTok.isNot(EndTokenKind)) DiscardUntilEndOfDirective(CurTok); return false; } @@ -3746,7 +3763,7 @@ Preprocessor::LexEmbedParameters(Token &CurTok, bool ForHasEmbed) { if (Result.isNegative()) { Diag(CurTok, diag::err_requires_positive_value) << toString(Result, 10) << /*positive*/ 0; - if (CurTok.isNot(tok::eod)) + if (CurTok.isNot(EndTokenKind)) DiscardUntilEndOfDirective(CurTok); return std::nullopt; } @@ -3889,7 +3906,7 @@ Preprocessor::LexEmbedParameters(Token &CurTok, bool ForHasEmbed) { } if (!ForHasEmbed) { Diag(ParamStartLoc, diag::err_pp_unknown_parameter) << 1 << Parameter; - if (CurTok.isNot(tok::eod)) + if (CurTok.isNot(EndTokenKind)) DiscardUntilEndOfDirective(CurTok); return std::nullopt; } @@ -3991,9 +4008,12 @@ void Preprocessor::HandleEmbedDirective(SourceLocation HashLoc, Token &EmbedTok, StringRef OriginalFilename = Filename; bool isAngled = GetIncludeFilenameSpelling(FilenameTok.getLocation(), Filename); + // If GetIncludeFilenameSpelling set the start ptr to null, there was an // error. - assert(!Filename.empty()); + if (Filename.empty()) + return; + OptionalFileEntryRef MaybeFileRef = this->LookupEmbedFile(Filename, isAngled, true, LookupFromFile); if (!MaybeFileRef) { diff --git a/clang/lib/Lex/PPMacroExpansion.cpp b/clang/lib/Lex/PPMacroExpansion.cpp index dec1956..dd80ae5 100644 --- a/clang/lib/Lex/PPMacroExpansion.cpp +++ b/clang/lib/Lex/PPMacroExpansion.cpp @@ -1262,16 +1262,11 @@ EmbedResult Preprocessor::EvaluateHasEmbed(Token &Tok, IdentifierInfo *II) { std::optional<LexEmbedParametersResult> Params = this->LexEmbedParameters(Tok, /*ForHasEmbed=*/true); - assert((Params || Tok.is(tok::eod)) && - "expected success or to be at the end of the directive"); if (!Params) return EmbedResult::Invalid; - if (Params->UnrecognizedParams > 0) - return EmbedResult::NotFound; - - if (!Tok.is(tok::r_paren)) { + if (Tok.isNot(tok::r_paren)) { Diag(this->getLocForEndOfToken(FilenameLoc), diag::err_pp_expected_after) << II << tok::r_paren; Diag(LParenLoc, diag::note_matching) << tok::l_paren; @@ -1280,6 +1275,9 @@ EmbedResult Preprocessor::EvaluateHasEmbed(Token &Tok, IdentifierInfo *II) { return EmbedResult::Invalid; } + if (Params->UnrecognizedParams > 0) + return EmbedResult::NotFound; + SmallString<128> FilenameBuffer; StringRef Filename = this->getSpelling(FilenameTok, FilenameBuffer); if (Filename.empty()) diff --git a/clang/lib/Parse/ParseOpenMP.cpp b/clang/lib/Parse/ParseOpenMP.cpp index 04f29c8..25199c7 100644 --- a/clang/lib/Parse/ParseOpenMP.cpp +++ b/clang/lib/Parse/ParseOpenMP.cpp @@ -3311,7 +3311,11 @@ OMPClause *Parser::ParseOpenMPClause(OpenMPDirectiveKind DKind, ErrorFound = true; } - Clause = ParseOpenMPClause(CKind, WrongDirective); + if (CKind == OMPC_nowait && PP.LookAhead(/*N=*/0).is(tok::l_paren) && + getLangOpts().OpenMP >= 60) + Clause = ParseOpenMPSingleExprClause(CKind, WrongDirective); + else + Clause = ParseOpenMPClause(CKind, WrongDirective); break; case OMPC_self_maps: // OpenMP [6.0, self_maps clause] diff --git a/clang/lib/Parse/Parser.cpp b/clang/lib/Parse/Parser.cpp index bbff627..ec01faf 100644 --- a/clang/lib/Parse/Parser.cpp +++ b/clang/lib/Parse/Parser.cpp @@ -1272,7 +1272,7 @@ Decl *Parser::ParseFunctionDefinition(ParsingDeclarator &D, // tokens and store them for late parsing at the end of the translation unit. if (getLangOpts().DelayedTemplateParsing && Tok.isNot(tok::equal) && TemplateInfo.Kind == ParsedTemplateKind::Template && - Actions.canDelayFunctionBody(D)) { + LateParsedAttrs->empty() && Actions.canDelayFunctionBody(D)) { MultiTemplateParamsArg TemplateParameterLists(*TemplateInfo.TemplateParams); ParseScope BodyScope(this, Scope::FnScope | Scope::DeclScope | @@ -1301,10 +1301,8 @@ Decl *Parser::ParseFunctionDefinition(ParsingDeclarator &D, } return DP; } - else if (CurParsedObjCImpl && - !TemplateInfo.TemplateParams && - (Tok.is(tok::l_brace) || Tok.is(tok::kw_try) || - Tok.is(tok::colon)) && + if (CurParsedObjCImpl && !TemplateInfo.TemplateParams && + (Tok.is(tok::l_brace) || Tok.is(tok::kw_try) || Tok.is(tok::colon)) && Actions.CurContext->isTranslationUnit()) { ParseScope BodyScope(this, Scope::FnScope | Scope::DeclScope | Scope::CompoundStmtScope); @@ -1420,7 +1418,8 @@ Decl *Parser::ParseFunctionDefinition(ParsingDeclarator &D, // Late attributes are parsed in the same scope as the function body. if (LateParsedAttrs) - ParseLexedAttributeList(*LateParsedAttrs, Res, false, true); + ParseLexedAttributeList(*LateParsedAttrs, Res, /*EnterScope=*/false, + /*OnDefinition=*/true); if (SkipFunctionBodies && (!Res || Actions.canSkipFunctionBody(Res)) && trySkippingFunctionBody()) { diff --git a/clang/lib/Sema/AnalysisBasedWarnings.cpp b/clang/lib/Sema/AnalysisBasedWarnings.cpp index 9abaf79..140b709 100644 --- a/clang/lib/Sema/AnalysisBasedWarnings.cpp +++ b/clang/lib/Sema/AnalysisBasedWarnings.cpp @@ -2609,7 +2609,6 @@ public: bool IsRelatedToDecl, ASTContext &Ctx) override { SourceLocation Loc; - std::string Message; Loc = Node.get<Stmt>()->getBeginLoc(); S.Diag(Loc, diag::warn_unsafe_buffer_usage_unique_ptr_array_access) diff --git a/clang/lib/Sema/CheckExprLifetime.cpp b/clang/lib/Sema/CheckExprLifetime.cpp index 8aebf53..e797400 100644 --- a/clang/lib/Sema/CheckExprLifetime.cpp +++ b/clang/lib/Sema/CheckExprLifetime.cpp @@ -361,11 +361,11 @@ static bool shouldTrackImplicitObjectArg(const CXXMethodDecl *Callee) { if (!Callee->getIdentifier()) return false; return llvm::StringSwitch<bool>(Callee->getName()) - .Cases("begin", "rbegin", "cbegin", "crbegin", true) - .Cases("end", "rend", "cend", "crend", true) - .Cases("c_str", "data", "get", true) + .Cases({"begin", "rbegin", "cbegin", "crbegin"}, true) + .Cases({"end", "rend", "cend", "crend"}, true) + .Cases({"c_str", "data", "get"}, true) // Map and set types. - .Cases("find", "equal_range", "lower_bound", "upper_bound", true) + .Cases({"find", "equal_range", "lower_bound", "upper_bound"}, true) .Default(false); } if (Callee->getReturnType()->isReferenceType()) { @@ -377,7 +377,7 @@ static bool shouldTrackImplicitObjectArg(const CXXMethodDecl *Callee) { OO == OverloadedOperatorKind::OO_Star; } return llvm::StringSwitch<bool>(Callee->getName()) - .Cases("front", "back", "at", "top", "value", true) + .Cases({"front", "back", "at", "top", "value"}, true) .Default(false); } return false; @@ -394,14 +394,14 @@ static bool shouldTrackFirstArgument(const FunctionDecl *FD) { if (FD->getReturnType()->isPointerType() || isRecordWithAttr<PointerAttr>(FD->getReturnType())) { return llvm::StringSwitch<bool>(FD->getName()) - .Cases("begin", "rbegin", "cbegin", "crbegin", true) - .Cases("end", "rend", "cend", "crend", true) + .Cases({"begin", "rbegin", "cbegin", "crbegin"}, true) + .Cases({"end", "rend", "cend", "crend"}, true) .Case("data", true) .Default(false); } if (FD->getReturnType()->isReferenceType()) { return llvm::StringSwitch<bool>(FD->getName()) - .Cases("get", "any_cast", true) + .Cases({"get", "any_cast"}, true) .Default(false); } return false; @@ -1265,7 +1265,7 @@ checkExprLifetimeImpl(Sema &SemaRef, const InitializedEntity *InitEntity, return true; case NotGSLPointer: IsGslPtrValueFromGslTempOwner = false; - LLVM_FALLTHROUGH; + [[fallthrough]]; case Report: break; } diff --git a/clang/lib/Sema/DeclSpec.cpp b/clang/lib/Sema/DeclSpec.cpp index 184d31e..9da3d0d 100644 --- a/clang/lib/Sema/DeclSpec.cpp +++ b/clang/lib/Sema/DeclSpec.cpp @@ -1369,7 +1369,8 @@ void DeclSpec::Finish(Sema &S, const PrintingPolicy &Policy) { if (S.getLangOpts().C23 && getConstexprSpecifier() == ConstexprSpecKind::Constexpr && - StorageClassSpec == SCS_extern) { + getTypeSpecType() != TST_unspecified && + (StorageClassSpec == SCS_extern || StorageClassSpec == SCS_auto)) { S.Diag(ConstexprLoc, diag::err_invalid_decl_spec_combination) << DeclSpec::getSpecifierName(getStorageClassSpec()) << SourceRange(getStorageClassSpecLoc()); diff --git a/clang/lib/Sema/HLSLExternalSemaSource.cpp b/clang/lib/Sema/HLSLExternalSemaSource.cpp index e118dda..f28a037 100644 --- a/clang/lib/Sema/HLSLExternalSemaSource.cpp +++ b/clang/lib/Sema/HLSLExternalSemaSource.cpp @@ -159,7 +159,8 @@ void HLSLExternalSemaSource::defineHLSLMatrixAlias() { SourceLocation(), ColsParam)); TemplateParams.emplace_back(ColsParam); - const unsigned MaxMatDim = 4; + const unsigned MaxMatDim = SemaPtr->getLangOpts().MaxMatrixDimension; + auto *MaxRow = IntegerLiteral::Create( AST, llvm::APInt(AST.getIntWidth(AST.IntTy), MaxMatDim), AST.IntTy, SourceLocation()); diff --git a/clang/lib/Sema/SemaAvailability.cpp b/clang/lib/Sema/SemaAvailability.cpp index f8d61d9..b09e168 100644 --- a/clang/lib/Sema/SemaAvailability.cpp +++ b/clang/lib/Sema/SemaAvailability.cpp @@ -102,7 +102,7 @@ Sema::ShouldDiagnoseAvailabilityOfDecl(const NamedDecl *D, std::string *Message, break; for (const Type *T = TD->getUnderlyingType().getTypePtr(); /**/; /**/) { if (auto *TT = dyn_cast<TagType>(T)) { - D = TT->getOriginalDecl()->getDefinitionOrSelf(); + D = TT->getDecl()->getDefinitionOrSelf(); } else if (isa<SubstTemplateTypeParmType>(T)) { // A Subst* node represents a use through a template. // Any uses of the underlying declaration happened through it's template @@ -1019,7 +1019,7 @@ bool DiagnoseUnguardedAvailability::VisitTypeLoc(TypeLoc Ty) { return true; if (const auto *TT = dyn_cast<TagType>(TyPtr)) { - TagDecl *TD = TT->getOriginalDecl()->getDefinitionOrSelf(); + TagDecl *TD = TT->getDecl()->getDefinitionOrSelf(); DiagnoseDeclAvailability(TD, Range); } else if (const auto *TD = dyn_cast<TypedefType>(TyPtr)) { diff --git a/clang/lib/Sema/SemaBPF.cpp b/clang/lib/Sema/SemaBPF.cpp index be890ab..a9761764 100644 --- a/clang/lib/Sema/SemaBPF.cpp +++ b/clang/lib/Sema/SemaBPF.cpp @@ -57,7 +57,7 @@ static bool isValidPreserveTypeInfoArg(Expr *Arg) { // Record type or Enum type. if (const auto *RT = ArgType->getAsCanonical<TagType>()) - if (!RT->getOriginalDecl()->getDeclName().isEmpty()) + if (!RT->getDecl()->getDeclName().isEmpty()) return true; return false; diff --git a/clang/lib/Sema/SemaCXXScopeSpec.cpp b/clang/lib/Sema/SemaCXXScopeSpec.cpp index 97ba1a51..c52fc5b 100644 --- a/clang/lib/Sema/SemaCXXScopeSpec.cpp +++ b/clang/lib/Sema/SemaCXXScopeSpec.cpp @@ -31,8 +31,7 @@ static CXXRecordDecl *getCurrentInstantiationOf(QualType T, const TagType *TagTy = dyn_cast<TagType>(T->getCanonicalTypeInternal()); if (!isa_and_present<RecordType, InjectedClassNameType>(TagTy)) return nullptr; - auto *RD = - cast<CXXRecordDecl>(TagTy->getOriginalDecl())->getDefinitionOrSelf(); + auto *RD = cast<CXXRecordDecl>(TagTy->getDecl())->getDefinitionOrSelf(); if (isa<InjectedClassNameType>(TagTy) || RD->isCurrentInstantiation(CurContext)) return RD; @@ -121,7 +120,7 @@ DeclContext *Sema::computeDeclContext(const CXXScopeSpec &SS, } } else if (const auto *RecordT = dyn_cast<RecordType>(NNSType)) { // The nested name specifier refers to a member of a class template. - return RecordT->getOriginalDecl()->getDefinitionOrSelf(); + return RecordT->getDecl()->getDefinitionOrSelf(); } } diff --git a/clang/lib/Sema/SemaCast.cpp b/clang/lib/Sema/SemaCast.cpp index d986e3b..5360f8a 100644 --- a/clang/lib/Sema/SemaCast.cpp +++ b/clang/lib/Sema/SemaCast.cpp @@ -962,7 +962,7 @@ void CastOperation::CheckDynamicCast() { } // C++ 5.2.7p6: Otherwise, v shall be [polymorphic]. - const RecordDecl *SrcDecl = SrcRecord->getOriginalDecl()->getDefinition(); + const RecordDecl *SrcDecl = SrcRecord->getDecl()->getDefinition(); assert(SrcDecl && "Definition missing"); if (!cast<CXXRecordDecl>(SrcDecl)->isPolymorphic()) { Self.Diag(OpRange.getBegin(), diag::err_bad_dynamic_cast_not_polymorphic) @@ -1453,7 +1453,7 @@ static TryCastResult TryStaticCast(Sema &Self, ExprResult &SrcExpr, // converted to an integral type. [...] A value of a scoped enumeration type // can also be explicitly converted to a floating-point type [...]. if (const EnumType *Enum = dyn_cast<EnumType>(SrcType)) { - if (Enum->getOriginalDecl()->isScoped()) { + if (Enum->getDecl()->isScoped()) { if (DestType->isBooleanType()) { Kind = CK_IntegralToBoolean; return TC_Success; @@ -2927,6 +2927,8 @@ bool CastOperation::CheckHLSLCStyleCast(CheckedConversionKind CCK) { SrcExpr = Self.ImpCastExprToType( SrcExpr.get(), Self.Context.getArrayParameterType(SrcTy), CK_HLSLArrayRValue, VK_PRValue, nullptr, CCK); + else + SrcExpr = Self.DefaultLvalueConversion(SrcExpr.get()); Kind = CK_HLSLElementwiseCast; return true; } @@ -2935,6 +2937,7 @@ bool CastOperation::CheckHLSLCStyleCast(CheckedConversionKind CCK) { // If the relative order of this and the HLSLElementWise cast checks // are changed, it might change which cast handles what in a few cases if (Self.HLSL().CanPerformAggregateSplatCast(SrcExpr.get(), DestType)) { + SrcExpr = Self.DefaultLvalueConversion(SrcExpr.get()); const VectorType *VT = SrcTy->getAs<VectorType>(); // change splat from vec1 case to splat from scalar if (VT && VT->getNumElements() == 1) @@ -3105,7 +3108,7 @@ void CastOperation::CheckCStyleCast() { } // GCC's cast to union extension. - if (RecordDecl *RD = DestRecordTy->getOriginalDecl(); RD->isUnion()) { + if (RecordDecl *RD = DestRecordTy->getDecl(); RD->isUnion()) { if (CastExpr::getTargetFieldForToUnionCast(RD->getDefinitionOrSelf(), SrcType)) { Self.Diag(OpRange.getBegin(), diag::ext_typecheck_cast_to_union) diff --git a/clang/lib/Sema/SemaChecking.cpp b/clang/lib/Sema/SemaChecking.cpp index 063db05..652527a 100644 --- a/clang/lib/Sema/SemaChecking.cpp +++ b/clang/lib/Sema/SemaChecking.cpp @@ -3580,9 +3580,8 @@ static bool CheckNonNullExpr(Sema &S, const Expr *Expr) { // As a special case, transparent unions initialized with zero are // considered null for the purposes of the nonnull attribute. if (const RecordType *UT = Expr->getType()->getAsUnionType(); - UT && UT->getOriginalDecl() - ->getMostRecentDecl() - ->hasAttr<TransparentUnionAttr>()) { + UT && + UT->getDecl()->getMostRecentDecl()->hasAttr<TransparentUnionAttr>()) { if (const auto *CLE = dyn_cast<CompoundLiteralExpr>(Expr)) if (const auto *ILE = dyn_cast<InitListExpr>(CLE->getInitializer())) Expr = ILE->getInit(0); @@ -12879,8 +12878,8 @@ void Sema::CheckImplicitConversion(Expr *E, QualType T, SourceLocation CC, if (const EnumType *SourceEnum = Source->getAsCanonical<EnumType>()) if (const EnumType *TargetEnum = Target->getAsCanonical<EnumType>()) - if (SourceEnum->getOriginalDecl()->hasNameForLinkage() && - TargetEnum->getOriginalDecl()->hasNameForLinkage() && + if (SourceEnum->getDecl()->hasNameForLinkage() && + TargetEnum->getDecl()->hasNameForLinkage() && SourceEnum != TargetEnum) { if (SourceMgr.isInSystemMacro(CC)) return; @@ -16240,9 +16239,9 @@ getAndVerifyMatrixDimension(Expr *Expr, StringRef Name, Sema &S) { return {}; } uint64_t Dim = Value->getZExtValue(); - if (!ConstantMatrixType::isDimensionValid(Dim)) { + if (Dim == 0 || Dim > S.Context.getLangOpts().MaxMatrixDimension) { S.Diag(Expr->getBeginLoc(), diag::err_builtin_matrix_invalid_dimension) - << Name << ConstantMatrixType::getMaxElementsPerDimension(); + << Name << S.Context.getLangOpts().MaxMatrixDimension; return {}; } return Dim; diff --git a/clang/lib/Sema/SemaCodeComplete.cpp b/clang/lib/Sema/SemaCodeComplete.cpp index 5dd4949..0514d10 100644 --- a/clang/lib/Sema/SemaCodeComplete.cpp +++ b/clang/lib/Sema/SemaCodeComplete.cpp @@ -2070,7 +2070,7 @@ static const char *GetCompletionTypeString(QualType T, ASTContext &Context, // Anonymous tag types are constant strings. if (const TagType *TagT = dyn_cast<TagType>(T)) - if (TagDecl *Tag = TagT->getOriginalDecl()) + if (TagDecl *Tag = TagT->getDecl()) if (!Tag->hasNameForLinkage()) { switch (Tag->getTagKind()) { case TagTypeKind::Struct: diff --git a/clang/lib/Sema/SemaConcept.cpp b/clang/lib/Sema/SemaConcept.cpp index 9cbd1bd..04a73181 100644 --- a/clang/lib/Sema/SemaConcept.cpp +++ b/clang/lib/Sema/SemaConcept.cpp @@ -606,10 +606,9 @@ ConstraintSatisfactionChecker::SubstitutionInTemplateArguments( Constraint.mappingOccurenceList(); // The empty MLTAL situation should only occur when evaluating non-dependent // constraints. - if (!MLTAL.getNumSubstitutedLevels()) - MLTAL.addOuterTemplateArguments(TD, {}, /*Final=*/false); - SubstitutedOuterMost = - llvm::to_vector_of<TemplateArgument>(MLTAL.getOutermost()); + if (MLTAL.getNumSubstitutedLevels()) + SubstitutedOuterMost = + llvm::to_vector_of<TemplateArgument>(MLTAL.getOutermost()); unsigned Offset = 0; for (unsigned I = 0, MappedIndex = 0; I < Used.size(); I++) { TemplateArgument Arg; @@ -627,8 +626,10 @@ ConstraintSatisfactionChecker::SubstitutionInTemplateArguments( if (Offset < SubstitutedOuterMost.size()) SubstitutedOuterMost.erase(SubstitutedOuterMost.begin() + Offset); - MLTAL.replaceOutermostTemplateArguments(TD, SubstitutedOuterMost); - return std::move(MLTAL); + MultiLevelTemplateArgumentList SubstitutedTemplateArgs; + SubstitutedTemplateArgs.addOuterTemplateArguments(TD, SubstitutedOuterMost, + /*Final=*/false); + return std::move(SubstitutedTemplateArgs); } ExprResult ConstraintSatisfactionChecker::EvaluateSlow( @@ -1216,13 +1217,51 @@ bool Sema::CheckConstraintSatisfaction( return false; } +static const ExprResult +SubstituteConceptsInConstrainExpression(Sema &S, const NamedDecl *D, + const ConceptSpecializationExpr *CSE, + UnsignedOrNone SubstIndex) { + + // [C++2c] [temp.constr.normal] + // Otherwise, to form CE, any non-dependent concept template argument Ai + // is substituted into the constraint-expression of C. + // If any such substitution results in an invalid concept-id, + // the program is ill-formed; no diagnostic is required. + + ConceptDecl *Concept = CSE->getNamedConcept()->getCanonicalDecl(); + Sema::ArgPackSubstIndexRAII _(S, SubstIndex); + + const ASTTemplateArgumentListInfo *ArgsAsWritten = + CSE->getTemplateArgsAsWritten(); + if (llvm::none_of( + ArgsAsWritten->arguments(), [&](const TemplateArgumentLoc &ArgLoc) { + return !ArgLoc.getArgument().isDependent() && + ArgLoc.getArgument().isConceptOrConceptTemplateParameter(); + })) { + return Concept->getConstraintExpr(); + } + + MultiLevelTemplateArgumentList MLTAL = S.getTemplateInstantiationArgs( + Concept, Concept->getLexicalDeclContext(), + /*Final=*/false, CSE->getTemplateArguments(), + /*RelativeToPrimary=*/true, + /*Pattern=*/nullptr, + /*ForConstraintInstantiation=*/true); + return S.SubstConceptTemplateArguments(CSE, Concept->getConstraintExpr(), + MLTAL); +} + bool Sema::CheckConstraintSatisfaction( const ConceptSpecializationExpr *ConstraintExpr, ConstraintSatisfaction &Satisfaction) { + ExprResult Res = SubstituteConceptsInConstrainExpression( + *this, nullptr, ConstraintExpr, ArgPackSubstIndex); + if (!Res.isUsable()) + return true; + llvm::SmallVector<AssociatedConstraint, 1> Constraints; - Constraints.emplace_back( - ConstraintExpr->getNamedConcept()->getConstraintExpr()); + Constraints.emplace_back(Res.get()); MultiLevelTemplateArgumentList MLTAL(ConstraintExpr->getNamedConcept(), ConstraintExpr->getTemplateArguments(), @@ -2248,8 +2287,14 @@ NormalizedConstraint *NormalizedConstraint::fromConstraintExpr( // Use canonical declarations to merge ConceptDecls across // different modules. ConceptDecl *CD = CSE->getNamedConcept()->getCanonicalDecl(); + + ExprResult Res = + SubstituteConceptsInConstrainExpression(S, D, CSE, SubstIndex); + if (!Res.isUsable()) + return nullptr; + SubNF = NormalizedConstraint::fromAssociatedConstraints( - S, CD, AssociatedConstraint(CD->getConstraintExpr(), SubstIndex)); + S, CD, AssociatedConstraint(Res.get(), SubstIndex)); if (!SubNF) return nullptr; diff --git a/clang/lib/Sema/SemaCoroutine.cpp b/clang/lib/Sema/SemaCoroutine.cpp index 229e91e..c0aba83 100644 --- a/clang/lib/Sema/SemaCoroutine.cpp +++ b/clang/lib/Sema/SemaCoroutine.cpp @@ -640,10 +640,9 @@ static void checkNoThrow(Sema &S, const Stmt *E, QualType::DestructionKind::DK_cxx_destructor) { const auto *T = cast<RecordType>(ReturnType.getCanonicalType().getTypePtr()); - checkDeclNoexcept(cast<CXXRecordDecl>(T->getOriginalDecl()) - ->getDefinition() - ->getDestructor(), - /*IsDtor=*/true); + checkDeclNoexcept( + cast<CXXRecordDecl>(T->getDecl())->getDefinition()->getDestructor(), + /*IsDtor=*/true); } } else for (const auto *Child : E->children()) { diff --git a/clang/lib/Sema/SemaDecl.cpp b/clang/lib/Sema/SemaDecl.cpp index 0e83c20..04d46d6 100644 --- a/clang/lib/Sema/SemaDecl.cpp +++ b/clang/lib/Sema/SemaDecl.cpp @@ -13816,13 +13816,20 @@ void Sema::AddInitializerToDecl(Decl *RealDecl, Expr *Init, bool DirectInit) { VDecl->setInvalidDecl(); } - // C++ [module.import/6] external definitions are not permitted in header - // units. + // C++ [module.import/6] + // ... + // A header unit shall not contain a definition of a non-inline function or + // variable whose name has external linkage. + // + // We choose to allow weak & selectany definitions, as they are common in + // headers, and have semantics similar to inline definitions which are allowed + // in header units. if (getLangOpts().CPlusPlusModules && currentModuleIsHeaderUnit() && !VDecl->isInvalidDecl() && VDecl->isThisDeclarationADefinition() && VDecl->getFormalLinkage() == Linkage::External && !VDecl->isInline() && !VDecl->isTemplated() && !isa<VarTemplateSpecializationDecl>(VDecl) && - !VDecl->getInstantiatedFromStaticDataMember()) { + !VDecl->getInstantiatedFromStaticDataMember() && + !(VDecl->hasAttr<SelectAnyAttr>() || VDecl->hasAttr<WeakAttr>())) { Diag(VDecl->getLocation(), diag::err_extern_def_in_header_unit); VDecl->setInvalidDecl(); } @@ -16153,16 +16160,24 @@ Decl *Sema::ActOnStartOfFunctionDef(Scope *FnBodyScope, Decl *D, } } - // C++ [module.import/6] external definitions are not permitted in header - // units. Deleted and Defaulted functions are implicitly inline (but the + // C++ [module.import/6] + // ... + // A header unit shall not contain a definition of a non-inline function or + // variable whose name has external linkage. + // + // Deleted and Defaulted functions are implicitly inline (but the // inline state is not set at this point, so check the BodyKind explicitly). + // We choose to allow weak & selectany definitions, as they are common in + // headers, and have semantics similar to inline definitions which are allowed + // in header units. // FIXME: Consider an alternate location for the test where the inlined() // state is complete. if (getLangOpts().CPlusPlusModules && currentModuleIsHeaderUnit() && !FD->isInvalidDecl() && !FD->isInlined() && BodyKind != FnBodyKind::Delete && BodyKind != FnBodyKind::Default && FD->getFormalLinkage() == Linkage::External && !FD->isTemplated() && - !FD->isTemplateInstantiation()) { + !FD->isTemplateInstantiation() && + !(FD->hasAttr<SelectAnyAttr>() || FD->hasAttr<WeakAttr>())) { assert(FD->isThisDeclarationADefinition()); Diag(FD->getLocation(), diag::err_extern_def_in_header_unit); FD->setInvalidDecl(); @@ -20797,7 +20812,7 @@ Sema::FunctionEmissionStatus Sema::getEmissionStatus(const FunctionDecl *FD, // SYCL functions can be template, so we check if they have appropriate // attribute prior to checking if it is a template. - if (LangOpts.SYCLIsDevice && FD->hasAttr<DeviceKernelAttr>()) + if (LangOpts.SYCLIsDevice && FD->hasAttr<SYCLKernelAttr>()) return FunctionEmissionStatus::Emitted; // Templates are emitted when they're instantiated. diff --git a/clang/lib/Sema/SemaDeclAttr.cpp b/clang/lib/Sema/SemaDeclAttr.cpp index 328ccf6..9475b8a 100644 --- a/clang/lib/Sema/SemaDeclAttr.cpp +++ b/clang/lib/Sema/SemaDeclAttr.cpp @@ -1255,7 +1255,7 @@ bool Sema::isValidPointerAttrType(QualType T, bool RefOkay) { // The nonnull attribute, and other similar attributes, can be applied to a // transparent union that contains a pointer type. if (const RecordType *UT = T->getAsUnionType()) { - RecordDecl *UD = UT->getOriginalDecl()->getDefinitionOrSelf(); + RecordDecl *UD = UT->getDecl()->getDefinitionOrSelf(); if (UD->hasAttr<TransparentUnionAttr>()) { for (const auto *I : UD->fields()) { QualType QT = I->getType(); @@ -3629,18 +3629,20 @@ static FormatAttrKind getFormatAttrKind(StringRef Format) { // Check for formats that get handled specially. .Case("NSString", NSStringFormat) .Case("CFString", CFStringFormat) - .Cases("gnu_strftime", "strftime", StrftimeFormat) + .Cases({"gnu_strftime", "strftime"}, StrftimeFormat) // Otherwise, check for supported formats. - .Cases("gnu_scanf", "scanf", "gnu_printf", "printf", "printf0", - "gnu_strfmon", "strfmon", SupportedFormat) - .Cases("cmn_err", "vcmn_err", "zcmn_err", SupportedFormat) - .Cases("kprintf", "syslog", SupportedFormat) // OpenBSD. - .Case("freebsd_kprintf", SupportedFormat) // FreeBSD. + .Cases({"gnu_scanf", "scanf", "gnu_printf", "printf", "printf0", + "gnu_strfmon", "strfmon"}, + SupportedFormat) + .Cases({"cmn_err", "vcmn_err", "zcmn_err"}, SupportedFormat) + .Cases({"kprintf", "syslog"}, SupportedFormat) // OpenBSD. + .Case("freebsd_kprintf", SupportedFormat) // FreeBSD. .Case("os_trace", SupportedFormat) .Case("os_log", SupportedFormat) - .Cases("gcc_diag", "gcc_cdiag", "gcc_cxxdiag", "gcc_tdiag", IgnoredFormat) + .Cases({"gcc_diag", "gcc_cdiag", "gcc_cxxdiag", "gcc_tdiag"}, + IgnoredFormat) .Default(InvalidFormat); } @@ -5204,16 +5206,7 @@ static void handleCallConvAttr(Sema &S, Decl *D, const ParsedAttr &AL) { static void handleDeviceKernelAttr(Sema &S, Decl *D, const ParsedAttr &AL) { const auto *FD = dyn_cast_or_null<FunctionDecl>(D); bool IsFunctionTemplate = FD && FD->getDescribedFunctionTemplate(); - if (S.getLangOpts().SYCLIsDevice) { - if (!IsFunctionTemplate) { - S.Diag(AL.getLoc(), diag::warn_attribute_wrong_decl_type_str) - << AL << AL.isRegularKeywordAttribute() << "function templates"; - } else { - S.SYCL().handleKernelAttr(D, AL); - } - } else if (DeviceKernelAttr::isSYCLSpelling(AL)) { - S.Diag(AL.getLoc(), diag::warn_attribute_ignored) << AL; - } else if (S.getASTContext().getTargetInfo().getTriple().isNVPTX()) { + if (S.getASTContext().getTargetInfo().getTriple().isNVPTX()) { handleGlobalAttr(S, D, AL); } else { // OpenCL C++ will throw a more specific error. @@ -5683,6 +5676,114 @@ static void handleLaunchBoundsAttr(Sema &S, Decl *D, const ParsedAttr &AL) { AL.getNumArgs() > 2 ? AL.getArgAsExpr(2) : nullptr); } +static std::pair<Expr *, int> +makeClusterDimsArgExpr(Sema &S, Expr *E, const CUDAClusterDimsAttr &AL, + const unsigned Idx) { + if (!E || S.DiagnoseUnexpandedParameterPack(E)) + return {}; + + // Accept template arguments for now as they depend on something else. + // We'll get to check them when they eventually get instantiated. + if (E->isInstantiationDependent()) + return {E, 1}; + + std::optional<llvm::APSInt> I = E->getIntegerConstantExpr(S.Context); + if (!I) { + S.Diag(E->getExprLoc(), diag::err_attribute_argument_n_type) + << &AL << Idx << AANT_ArgumentIntegerConstant << E->getSourceRange(); + return {}; + } + // Make sure we can fit it in 4 bits. + if (!I->isIntN(4)) { + S.Diag(E->getExprLoc(), diag::err_ice_too_large) + << toString(*I, 10, false) << 4 << /*Unsigned=*/1; + return {}; + } + if (*I < 0) { + S.Diag(E->getExprLoc(), diag::warn_attribute_argument_n_negative) + << &AL << Idx << E->getSourceRange(); + } + + return {ConstantExpr::Create(S.getASTContext(), E, APValue(*I)), + I->getZExtValue()}; +} + +CUDAClusterDimsAttr *Sema::createClusterDimsAttr(const AttributeCommonInfo &CI, + Expr *X, Expr *Y, Expr *Z) { + CUDAClusterDimsAttr TmpAttr(Context, CI, X, Y, Z); + + auto [NewX, ValX] = makeClusterDimsArgExpr(*this, X, TmpAttr, /*Idx=*/0); + auto [NewY, ValY] = makeClusterDimsArgExpr(*this, Y, TmpAttr, /*Idx=*/1); + auto [NewZ, ValZ] = makeClusterDimsArgExpr(*this, Z, TmpAttr, /*Idx=*/2); + + if (!NewX || (Y && !NewY) || (Z && !NewZ)) + return nullptr; + + int FlatDim = ValX * ValY * ValZ; + const llvm::Triple TT = + (!Context.getLangOpts().CUDAIsDevice && Context.getAuxTargetInfo()) + ? Context.getAuxTargetInfo()->getTriple() + : Context.getTargetInfo().getTriple(); + int MaxDim = 1; + if (TT.isNVPTX()) + MaxDim = 8; + else if (TT.isAMDGPU()) + MaxDim = 16; + else + return nullptr; + + // A maximum of 8 thread blocks in a cluster is supported as a portable + // cluster size in CUDA. The number is 16 for AMDGPU. + if (FlatDim > MaxDim) { + Diag(CI.getLoc(), diag::err_cluster_dims_too_large) << MaxDim << FlatDim; + return nullptr; + } + + return CUDAClusterDimsAttr::Create(Context, NewX, NewY, NewZ, CI); +} + +void Sema::addClusterDimsAttr(Decl *D, const AttributeCommonInfo &CI, Expr *X, + Expr *Y, Expr *Z) { + if (auto *Attr = createClusterDimsAttr(CI, X, Y, Z)) + D->addAttr(Attr); +} + +void Sema::addNoClusterAttr(Decl *D, const AttributeCommonInfo &CI) { + D->addAttr(CUDANoClusterAttr::Create(Context, CI)); +} + +static void handleClusterDimsAttr(Sema &S, Decl *D, const ParsedAttr &AL) { + const TargetInfo &TTI = S.Context.getTargetInfo(); + OffloadArch Arch = StringToOffloadArch(TTI.getTargetOpts().CPU); + if ((TTI.getTriple().isNVPTX() && Arch < clang::OffloadArch::SM_90) || + (TTI.getTriple().isAMDGPU() && + !TTI.hasFeatureEnabled(TTI.getTargetOpts().FeatureMap, "clusters"))) { + S.Diag(AL.getLoc(), diag::err_cluster_attr_not_supported) << AL; + return; + } + + if (!AL.checkAtLeastNumArgs(S, /*Num=*/1) || + !AL.checkAtMostNumArgs(S, /*Num=*/3)) + return; + + S.addClusterDimsAttr(D, AL, AL.getArgAsExpr(0), + AL.getNumArgs() > 1 ? AL.getArgAsExpr(1) : nullptr, + AL.getNumArgs() > 2 ? AL.getArgAsExpr(2) : nullptr); +} + +static void handleNoClusterAttr(Sema &S, Decl *D, const ParsedAttr &AL) { + const TargetInfo &TTI = S.Context.getTargetInfo(); + OffloadArch Arch = StringToOffloadArch(TTI.getTargetOpts().CPU); + if ((TTI.getTriple().isNVPTX() && Arch < clang::OffloadArch::SM_90) || + (TTI.getTriple().isAMDGPU() && + !TTI.hasFeatureEnabled(TTI.getTargetOpts().FeatureMap, "clusters"))) { + S.Diag(AL.getLoc(), diag::err_cluster_attr_not_supported) << AL; + return; + } + + S.addNoClusterAttr(D, AL); +} + static void handleArgumentWithTypeTagAttr(Sema &S, Decl *D, const ParsedAttr &AL) { if (!AL.isArgIdent(0)) { @@ -7100,6 +7201,9 @@ ProcessDeclAttribute(Sema &S, Scope *scope, Decl *D, const ParsedAttr &AL, case ParsedAttr::AT_EnumExtensibility: handleEnumExtensibilityAttr(S, D, AL); break; + case ParsedAttr::AT_SYCLKernel: + S.SYCL().handleKernelAttr(D, AL); + break; case ParsedAttr::AT_SYCLExternal: handleSimpleAttribute<SYCLExternalAttr>(S, D, AL); break; @@ -7145,6 +7249,12 @@ ProcessDeclAttribute(Sema &S, Scope *scope, Decl *D, const ParsedAttr &AL, case ParsedAttr::AT_CUDALaunchBounds: handleLaunchBoundsAttr(S, D, AL); break; + case ParsedAttr::AT_CUDAClusterDims: + handleClusterDimsAttr(S, D, AL); + break; + case ParsedAttr::AT_CUDANoCluster: + handleNoClusterAttr(S, D, AL); + break; case ParsedAttr::AT_Restrict: handleRestrictAttr(S, D, AL); break; diff --git a/clang/lib/Sema/SemaDeclCXX.cpp b/clang/lib/Sema/SemaDeclCXX.cpp index 215431c..d41ab12 100644 --- a/clang/lib/Sema/SemaDeclCXX.cpp +++ b/clang/lib/Sema/SemaDeclCXX.cpp @@ -5630,7 +5630,7 @@ bool Sema::SetCtorInitializers(CXXConstructorDecl *Constructor, bool AnyErrors, static void PopulateKeysForFields(FieldDecl *Field, SmallVectorImpl<const void*> &IdealInits) { if (const RecordType *RT = Field->getType()->getAsCanonical<RecordType>()) { - const RecordDecl *RD = RT->getOriginalDecl(); + const RecordDecl *RD = RT->getDecl(); if (RD->isAnonymousStructOrUnion()) { for (auto *Field : RD->getDefinitionOrSelf()->fields()) PopulateKeysForFields(Field, IdealInits); @@ -7630,9 +7630,8 @@ static bool defaultedSpecialMemberIsConstexpr( continue; QualType BaseType = S.Context.getBaseElementType(F->getType()); if (const RecordType *RecordTy = BaseType->getAsCanonical<RecordType>()) { - CXXRecordDecl *FieldRecDecl = - cast<CXXRecordDecl>(RecordTy->getOriginalDecl()) - ->getDefinitionOrSelf(); + auto *FieldRecDecl = + cast<CXXRecordDecl>(RecordTy->getDecl())->getDefinitionOrSelf(); if (!specialMemberIsConstexpr(S, FieldRecDecl, CSM, BaseType.getCVRQualifiers(), ConstArg && !F->isMutable())) @@ -10645,7 +10644,7 @@ void Sema::checkIllFormedTrivialABIStruct(CXXRecordDecl &RD) { if (const auto *RT = FT->getBaseElementTypeUnsafe()->getAsCanonical<RecordType>()) if (!RT->isDependentType() && - !cast<CXXRecordDecl>(RT->getOriginalDecl()->getDefinitionOrSelf()) + !cast<CXXRecordDecl>(RT->getDecl()->getDefinitionOrSelf()) ->canPassInRegisters()) { PrintDiagAndRemoveAttr(5); return; diff --git a/clang/lib/Sema/SemaDeclObjC.cpp b/clang/lib/Sema/SemaDeclObjC.cpp index 98eb5af..3df9f9c 100644 --- a/clang/lib/Sema/SemaDeclObjC.cpp +++ b/clang/lib/Sema/SemaDeclObjC.cpp @@ -3232,10 +3232,8 @@ static bool tryMatchRecordTypes(ASTContext &Context, assert(lt && rt && lt != rt); if (!isa<RecordType>(lt) || !isa<RecordType>(rt)) return false; - RecordDecl *left = - cast<RecordType>(lt)->getOriginalDecl()->getDefinitionOrSelf(); - RecordDecl *right = - cast<RecordType>(rt)->getOriginalDecl()->getDefinitionOrSelf(); + RecordDecl *left = cast<RecordType>(lt)->getDecl()->getDefinitionOrSelf(); + RecordDecl *right = cast<RecordType>(rt)->getDecl()->getDefinitionOrSelf(); // Require union-hood to match. if (left->isUnion() != right->isUnion()) return false; diff --git a/clang/lib/Sema/SemaExpr.cpp b/clang/lib/Sema/SemaExpr.cpp index 01abc1f..dca9d6e 100644 --- a/clang/lib/Sema/SemaExpr.cpp +++ b/clang/lib/Sema/SemaExpr.cpp @@ -1535,12 +1535,8 @@ void Sema::checkEnumArithmeticConversions(Expr *LHS, Expr *RHS, // are ill-formed. if (getLangOpts().CPlusPlus26) DiagID = diag::warn_conv_mixed_enum_types_cxx26; - else if (!L->castAsCanonical<EnumType>() - ->getOriginalDecl() - ->hasNameForLinkage() || - !R->castAsCanonical<EnumType>() - ->getOriginalDecl() - ->hasNameForLinkage()) { + else if (!L->castAsCanonical<EnumType>()->getDecl()->hasNameForLinkage() || + !R->castAsCanonical<EnumType>()->getDecl()->hasNameForLinkage()) { // If either enumeration type is unnamed, it's less likely that the // user cares about this, but this situation is still deprecated in // C++2a. Use a different warning group. @@ -7095,7 +7091,7 @@ ExprResult Sema::BuildResolvedCallExpr(Expr *Fn, NamedDecl *NDecl, for (unsigned i = 0, e = Args.size(); i != e; i++) { if (const auto *RT = dyn_cast<RecordType>(Args[i]->getType().getCanonicalType())) { - if (RT->getOriginalDecl()->isOrContainsUnion()) + if (RT->getDecl()->isOrContainsUnion()) Diag(Args[i]->getBeginLoc(), diag::warn_cmse_nonsecure_union) << 0 << i; } @@ -9748,7 +9744,7 @@ Sema::CheckTransparentUnionArgumentConstraints(QualType ArgType, if (!UT) return AssignConvertType::Incompatible; - RecordDecl *UD = UT->getOriginalDecl()->getDefinitionOrSelf(); + RecordDecl *UD = UT->getDecl()->getDefinitionOrSelf(); if (!UD->hasAttr<TransparentUnionAttr>()) return AssignConvertType::Incompatible; @@ -10844,7 +10840,7 @@ static void diagnoseScopedEnums(Sema &S, const SourceLocation Loc, auto DiagnosticHelper = [&S](const Expr *expr, const QualType type) { SourceLocation BeginLoc = expr->getBeginLoc(); QualType IntType = type->castAs<EnumType>() - ->getOriginalDecl() + ->getDecl() ->getDefinitionOrSelf() ->getIntegerType(); std::string InsertionString = "static_cast<" + IntType.getAsString() + ">("; @@ -11533,7 +11529,7 @@ QualType Sema::CheckSubtractionOperands(ExprResult &LHS, ExprResult &RHS, static bool isScopedEnumerationType(QualType T) { if (const EnumType *ET = T->getAsCanonical<EnumType>()) - return ET->getOriginalDecl()->isScoped(); + return ET->getDecl()->isScoped(); return false; } @@ -13839,7 +13835,7 @@ static void DiagnoseRecursiveConstFields(Sema &S, const ValueDecl *VD, while (RecordTypeList.size() > NextToCheckIndex) { bool IsNested = NextToCheckIndex > 0; for (const FieldDecl *Field : RecordTypeList[NextToCheckIndex] - ->getOriginalDecl() + ->getDecl() ->getDefinitionOrSelf() ->fields()) { // First, check every field for constness. @@ -15948,6 +15944,20 @@ ExprResult Sema::CreateBuiltinUnaryOp(SourceLocation OpLoc, return ExprError(Diag(OpLoc, diag::err_typecheck_unary_expr) << resultType << Input.get()->getSourceRange()); } + } else if (Context.getLangOpts().HLSL && resultType->isVectorType() && + !resultType->hasBooleanRepresentation()) { + // HLSL unary logical 'not' behaves like C++, which states that the + // operand is converted to bool and the result is bool, however HLSL + // extends this property to vectors. + const VectorType *VTy = resultType->castAs<VectorType>(); + resultType = + Context.getExtVectorType(Context.BoolTy, VTy->getNumElements()); + + Input = ImpCastExprToType( + Input.get(), resultType, + ScalarTypeToBooleanCastKind(VTy->getElementType())) + .get(); + break; } else if (resultType->isExtVectorType()) { if (Context.getLangOpts().OpenCL && Context.getLangOpts().getOpenCLCompatibleVersion() < 120) { diff --git a/clang/lib/Sema/SemaExprCXX.cpp b/clang/lib/Sema/SemaExprCXX.cpp index 0fe242dce..fe1f89b 100644 --- a/clang/lib/Sema/SemaExprCXX.cpp +++ b/clang/lib/Sema/SemaExprCXX.cpp @@ -1988,7 +1988,7 @@ static bool doesUsualArrayDeleteWantSize(Sema &S, SourceLocation loc, DeclarationName deleteName = S.Context.DeclarationNames.getCXXOperatorName(OO_Array_Delete); LookupResult ops(S, deleteName, loc, Sema::LookupOrdinaryName); - S.LookupQualifiedName(ops, record->getOriginalDecl()->getDefinitionOrSelf()); + S.LookupQualifiedName(ops, record->getDecl()->getDefinitionOrSelf()); // We're just doing this for information. ops.suppressDiagnostics(); @@ -6667,8 +6667,7 @@ ExprResult Sema::MaybeBindToTemporary(Expr *E) { // That should be enough to guarantee that this type is complete, if we're // not processing a decltype expression. - CXXRecordDecl *RD = - cast<CXXRecordDecl>(RT->getOriginalDecl())->getDefinitionOrSelf(); + auto *RD = cast<CXXRecordDecl>(RT->getDecl())->getDefinitionOrSelf(); if (RD->isInvalidDecl() || RD->isDependentContext()) return E; diff --git a/clang/lib/Sema/SemaExprObjC.cpp b/clang/lib/Sema/SemaExprObjC.cpp index 331f6e5..4daf0170 100644 --- a/clang/lib/Sema/SemaExprObjC.cpp +++ b/clang/lib/Sema/SemaExprObjC.cpp @@ -3846,8 +3846,7 @@ static inline T *getObjCBridgeAttr(const TypedefType *TD) { if (QT->isPointerType()) { QT = QT->getPointeeType(); if (const RecordType *RT = QT->getAsCanonical<RecordType>()) { - for (auto *Redecl : - RT->getOriginalDecl()->getMostRecentDecl()->redecls()) { + for (auto *Redecl : RT->getDecl()->getMostRecentDecl()->redecls()) { if (auto *attr = Redecl->getAttr<T>()) return attr; } diff --git a/clang/lib/Sema/SemaHLSL.cpp b/clang/lib/Sema/SemaHLSL.cpp index 17cb1e4..f347066 100644 --- a/clang/lib/Sema/SemaHLSL.cpp +++ b/clang/lib/Sema/SemaHLSL.cpp @@ -782,7 +782,7 @@ bool SemaHLSL::isSemanticValid(FunctionDecl *FD, DeclaratorDecl *D) { if (!RT) return false; - const RecordDecl *RD = RT->getOriginalDecl(); + const RecordDecl *RD = RT->getDecl(); for (FieldDecl *Field : RD->fields()) { if (!isSemanticValid(FD, Field)) return false; @@ -1986,7 +1986,7 @@ SemaHLSL::TakeLocForHLSLAttribute(const HLSLAttributedResourceType *RT) { // requirements and adds them to Bindings void SemaHLSL::collectResourceBindingsOnUserRecordDecl(const VarDecl *VD, const RecordType *RT) { - const RecordDecl *RD = RT->getOriginalDecl()->getDefinitionOrSelf(); + const RecordDecl *RD = RT->getDecl()->getDefinitionOrSelf(); for (FieldDecl *FD : RD->fields()) { const Type *Ty = FD->getType()->getUnqualifiedDesugaredType(); @@ -3544,40 +3544,6 @@ bool SemaHLSL::CanPerformScalarCast(QualType SrcTy, QualType DestTy) { llvm_unreachable("Unhandled scalar cast"); } -// Detect if a type contains a bitfield. Will be removed when -// bitfield support is added to HLSLElementwiseCast and HLSLAggregateSplatCast -bool SemaHLSL::ContainsBitField(QualType BaseTy) { - llvm::SmallVector<QualType, 16> WorkList; - WorkList.push_back(BaseTy); - while (!WorkList.empty()) { - QualType T = WorkList.pop_back_val(); - T = T.getCanonicalType().getUnqualifiedType(); - // only check aggregate types - if (const auto *AT = dyn_cast<ConstantArrayType>(T)) { - WorkList.push_back(AT->getElementType()); - continue; - } - if (const auto *RT = dyn_cast<RecordType>(T)) { - const RecordDecl *RD = RT->getOriginalDecl()->getDefinitionOrSelf(); - if (RD->isUnion()) - continue; - - const CXXRecordDecl *CXXD = dyn_cast<CXXRecordDecl>(RD); - - if (CXXD && CXXD->isStandardLayout()) - RD = CXXD->getStandardLayoutBaseWithFields(); - - for (const auto *FD : RD->fields()) { - if (FD->isBitField()) - return true; - WorkList.push_back(FD->getType()); - } - continue; - } - } - return false; -} - // Can perform an HLSL Aggregate splat cast if the Dest is an aggregate and the // Src is a scalar or a vector of length 1 // Or if Dest is a vector and Src is a vector of length 1 diff --git a/clang/lib/Sema/SemaInit.cpp b/clang/lib/Sema/SemaInit.cpp index 543db46..f7974eb 100644 --- a/clang/lib/Sema/SemaInit.cpp +++ b/clang/lib/Sema/SemaInit.cpp @@ -775,7 +775,7 @@ void InitListChecker::FillInEmptyInitForField(unsigned Init, FieldDecl *Field, if (Init >= NumInits || !ILE->getInit(Init)) { if (const RecordType *RType = ILE->getType()->getAsCanonical<RecordType>()) - if (!RType->getOriginalDecl()->isUnion()) + if (!RType->getDecl()->isUnion()) assert((Init < NumInits || VerifyOnly) && "This ILE should have been expanded"); @@ -9186,9 +9186,8 @@ bool InitializationSequence::Diagnose(Sema &S, diag::note_member_declared_at); if (const auto *Record = Entity.getType()->getAs<RecordType>()) - S.Diag(Record->getOriginalDecl()->getLocation(), - diag::note_previous_decl) - << S.Context.getCanonicalTagType(Record->getOriginalDecl()); + S.Diag(Record->getDecl()->getLocation(), diag::note_previous_decl) + << S.Context.getCanonicalTagType(Record->getDecl()); } break; } @@ -9974,8 +9973,8 @@ QualType Sema::DeduceTemplateSpecializationFromInitializer( // Cases where template arguments in the RHS of the alias are not // dependent. e.g. // using AliasFoo = Foo<bool>; - if (const auto *CTSD = llvm::dyn_cast<ClassTemplateSpecializationDecl>( - RT->getOriginalDecl())) + if (const auto *CTSD = + llvm::dyn_cast<ClassTemplateSpecializationDecl>(RT->getDecl())) Template = CTSD->getSpecializedTemplate(); } } diff --git a/clang/lib/Sema/SemaLookup.cpp b/clang/lib/Sema/SemaLookup.cpp index 25728de..5915d6e 100644 --- a/clang/lib/Sema/SemaLookup.cpp +++ b/clang/lib/Sema/SemaLookup.cpp @@ -2727,9 +2727,7 @@ bool Sema::LookupParsedName(LookupResult &R, Scope *S, CXXScopeSpec *SS, IsDependent = !DC && ObjectType->isDependentType(); assert(((!DC && ObjectType->isDependentType()) || !ObjectType->isIncompleteType() || !ObjectType->getAs<TagType>() || - ObjectType->castAs<TagType>() - ->getOriginalDecl() - ->isEntityBeingDefined()) && + ObjectType->castAs<TagType>()->getDecl()->isEntityBeingDefined()) && "Caller should have completed object type"); } else if (SS && SS->isNotEmpty()) { // This nested-name-specifier occurs after another nested-name-specifier, @@ -3191,9 +3189,8 @@ addAssociatedClassesAndNamespaces(AssociatedLookup &Result, QualType Ty) { // namespaces of its associated classes. case Type::Record: { // FIXME: This should use the original decl. - CXXRecordDecl *Class = - cast<CXXRecordDecl>(cast<RecordType>(T)->getOriginalDecl()) - ->getDefinitionOrSelf(); + auto *Class = cast<CXXRecordDecl>(cast<RecordType>(T)->getDecl()) + ->getDefinitionOrSelf(); addAssociatedClassesAndNamespaces(Result, Class); break; } @@ -4606,7 +4603,7 @@ static void getNestedNameSpecifierIdentifiers( case Type::InjectedClassName: { auto *TT = cast<TagType>(T); getNestedNameSpecifierIdentifiers(TT->getQualifier(), Identifiers); - Identifiers.push_back(TT->getOriginalDecl()->getIdentifier()); + Identifiers.push_back(TT->getDecl()->getIdentifier()); return; } case Type::Typedef: { diff --git a/clang/lib/Sema/SemaObjC.cpp b/clang/lib/Sema/SemaObjC.cpp index 4f9470a..7aaa56e 100644 --- a/clang/lib/Sema/SemaObjC.cpp +++ b/clang/lib/Sema/SemaObjC.cpp @@ -1407,7 +1407,7 @@ SemaObjC::ObjCSubscriptKind SemaObjC::CheckSubscriptingKind(Expr *FromE) { int NoIntegrals = 0, NoObjCIdPointers = 0; SmallVector<CXXConversionDecl *, 4> ConversionDecls; - for (NamedDecl *D : cast<CXXRecordDecl>(RecordTy->getOriginalDecl()) + for (NamedDecl *D : cast<CXXRecordDecl>(RecordTy->getDecl()) ->getDefinitionOrSelf() ->getVisibleConversionFunctions()) { if (CXXConversionDecl *Conversion = @@ -1511,7 +1511,7 @@ bool SemaObjC::isCFStringType(QualType T) { if (!RT) return false; - const RecordDecl *RD = RT->getOriginalDecl(); + const RecordDecl *RD = RT->getDecl(); if (RD->getTagKind() != TagTypeKind::Struct) return false; diff --git a/clang/lib/Sema/SemaOpenACC.cpp b/clang/lib/Sema/SemaOpenACC.cpp index f3969a9..ca99834 100644 --- a/clang/lib/Sema/SemaOpenACC.cpp +++ b/clang/lib/Sema/SemaOpenACC.cpp @@ -2883,12 +2883,12 @@ SemaOpenACC::CreateFirstPrivateInitRecipe(const Expr *VarExpr) { return OpenACCFirstPrivateRecipe(AllocaDecl, Temporary); } -OpenACCReductionRecipe SemaOpenACC::CreateReductionInitRecipe( +OpenACCReductionRecipeWithStorage SemaOpenACC::CreateReductionInitRecipe( OpenACCReductionOperator ReductionOperator, const Expr *VarExpr) { // We don't strip bounds here, so that we are doing our recipe init at the // 'lowest' possible level. Codegen is going to have to do its own 'looping'. if (!VarExpr || VarExpr->getType()->isDependentType()) - return OpenACCReductionRecipe::Empty(); + return OpenACCReductionRecipeWithStorage::Empty(); QualType VarTy = VarExpr->getType().getNonReferenceType().getUnqualifiedType(); @@ -2898,6 +2898,15 @@ OpenACCReductionRecipe SemaOpenACC::CreateReductionInitRecipe( dyn_cast<ArraySectionExpr>(VarExpr->IgnoreParenImpCasts())) VarTy = ASE->getElementType(); + llvm::SmallVector<OpenACCReductionRecipe::CombinerRecipe, 1> CombinerRecipes; + + // We use the 'set-ness' of the alloca-decl to determine whether the combiner + // is 'set' or not, so we can skip any attempts at it if we're going to fail + // at any of the combiners. + if (CreateReductionCombinerRecipe(VarExpr->getBeginLoc(), ReductionOperator, + VarTy, CombinerRecipes)) + return OpenACCReductionRecipeWithStorage::Empty(); + VarDecl *AllocaDecl = CreateAllocaDecl( getASTContext(), SemaRef.getCurContext(), VarExpr->getBeginLoc(), &getASTContext().Idents.get("openacc.reduction.init"), VarTy); @@ -2946,5 +2955,163 @@ OpenACCReductionRecipe SemaOpenACC::CreateReductionInitRecipe( AllocaDecl->setInit(Init.get()); AllocaDecl->setInitStyle(VarDecl::CallInit); } - return OpenACCReductionRecipe(AllocaDecl, {}); + + return OpenACCReductionRecipeWithStorage(AllocaDecl, CombinerRecipes); +} + +bool SemaOpenACC::CreateReductionCombinerRecipe( + SourceLocation Loc, OpenACCReductionOperator ReductionOperator, + QualType VarTy, + llvm::SmallVectorImpl<OpenACCReductionRecipe::CombinerRecipe> + &CombinerRecipes) { + // Now we can try to generate the 'combiner' recipe. This is a little + // complicated in that if the 'VarTy' is an array type, we want to take its + // element type so we can generate that. Additionally, if this is a struct, + // we have two options: If there is overloaded operators, we want to take + // THOSE, else we want to do the individual elements. + + BinaryOperatorKind BinOp; + switch (ReductionOperator) { + case OpenACCReductionOperator::Invalid: + // This can only happen when there is an error, and since these inits + // are used for code generation, we can just ignore/not bother doing any + // initialization here. + CombinerRecipes.push_back({nullptr, nullptr, nullptr}); + return false; + case OpenACCReductionOperator::Addition: + BinOp = BinaryOperatorKind::BO_AddAssign; + break; + case OpenACCReductionOperator::Multiplication: + BinOp = BinaryOperatorKind::BO_MulAssign; + break; + case OpenACCReductionOperator::BitwiseAnd: + BinOp = BinaryOperatorKind::BO_AndAssign; + break; + case OpenACCReductionOperator::BitwiseOr: + BinOp = BinaryOperatorKind::BO_OrAssign; + break; + case OpenACCReductionOperator::BitwiseXOr: + BinOp = BinaryOperatorKind::BO_XorAssign; + break; + + case OpenACCReductionOperator::Max: + case OpenACCReductionOperator::Min: + case OpenACCReductionOperator::And: + case OpenACCReductionOperator::Or: + // We just want a 'NYI' error in the backend, so leave an empty combiner + // recipe, and claim success. + CombinerRecipes.push_back({nullptr, nullptr, nullptr}); + return false; + } + + // If VarTy is an array type, at the top level only, we want to do our + // compares/decomp/etc at the element level. + if (auto *AT = getASTContext().getAsArrayType(VarTy)) + VarTy = AT->getElementType(); + + assert(!VarTy->isArrayType() && "Only 1 level of array allowed"); + + auto tryCombiner = [&, this](DeclRefExpr *LHSDRE, DeclRefExpr *RHSDRE, + bool IncludeTrap) { + // TODO: OpenACC: we have to figure out based on the bin-op how to do the + // ones that we can't just use compound operators for. So &&, ||, max, and + // min aren't really clear what we could do here. + if (IncludeTrap) { + // Trap all of the errors here, we'll emit our own at the end. + Sema::TentativeAnalysisScope Trap{SemaRef}; + + return SemaRef.BuildBinOp(SemaRef.getCurScope(), Loc, BinOp, LHSDRE, + RHSDRE, + /*ForFoldExpr=*/false); + } else { + return SemaRef.BuildBinOp(SemaRef.getCurScope(), Loc, BinOp, LHSDRE, + RHSDRE, + /*ForFoldExpr=*/false); + } + }; + + struct CombinerAttemptTy { + VarDecl *LHS; + DeclRefExpr *LHSDRE; + VarDecl *RHS; + DeclRefExpr *RHSDRE; + Expr *Op; + }; + + auto formCombiner = [&, this](QualType Ty) -> CombinerAttemptTy { + VarDecl *LHSDecl = CreateAllocaDecl( + getASTContext(), SemaRef.getCurContext(), Loc, + &getASTContext().Idents.get("openacc.reduction.combiner.lhs"), Ty); + auto *LHSDRE = DeclRefExpr::Create( + getASTContext(), NestedNameSpecifierLoc{}, SourceLocation{}, LHSDecl, + /*ReferstoEnclosingVariableOrCapture=*/false, + DeclarationNameInfo{DeclarationName{LHSDecl->getDeclName()}, + LHSDecl->getBeginLoc()}, + Ty, clang::VK_LValue, LHSDecl, nullptr, NOUR_None); + VarDecl *RHSDecl = CreateAllocaDecl( + getASTContext(), SemaRef.getCurContext(), Loc, + &getASTContext().Idents.get("openacc.reduction.combiner.lhs"), Ty); + auto *RHSDRE = DeclRefExpr::Create( + getASTContext(), NestedNameSpecifierLoc{}, SourceLocation{}, RHSDecl, + /*ReferstoEnclosingVariableOrCapture=*/false, + DeclarationNameInfo{DeclarationName{RHSDecl->getDeclName()}, + RHSDecl->getBeginLoc()}, + Ty, clang::VK_LValue, RHSDecl, nullptr, NOUR_None); + + ExprResult BinOpResult = tryCombiner(LHSDRE, RHSDRE, /*IncludeTrap=*/true); + + return {LHSDecl, LHSDRE, RHSDecl, RHSDRE, BinOpResult.get()}; + }; + + CombinerAttemptTy TopLevelCombinerInfo = formCombiner(VarTy); + + if (TopLevelCombinerInfo.Op) { + if (!TopLevelCombinerInfo.Op->containsErrors() && + TopLevelCombinerInfo.Op->isInstantiationDependent()) { + // If this is instantiation dependent, we're just going to 'give up' here + // and count on us to get it right during instantaition. + CombinerRecipes.push_back({nullptr, nullptr, nullptr}); + return false; + } else if (!TopLevelCombinerInfo.Op->containsErrors()) { + // Else, we succeeded, we can just return this combiner. + CombinerRecipes.push_back({TopLevelCombinerInfo.LHS, + TopLevelCombinerInfo.RHS, + TopLevelCombinerInfo.Op}); + return false; + } + } + + // Since the 'root' level didn't fail, the only thing that could be successful + // is a struct that we decompose on its individual fields. + + RecordDecl *RD = VarTy->getAsRecordDecl(); + if (!RD) { + Diag(Loc, diag::err_acc_reduction_recipe_no_op) << VarTy; + tryCombiner(TopLevelCombinerInfo.LHSDRE, TopLevelCombinerInfo.RHSDRE, + /*IncludeTrap=*/false); + return true; + } + + for (const FieldDecl *FD : RD->fields()) { + CombinerAttemptTy FieldCombinerInfo = formCombiner(FD->getType()); + + if (!FieldCombinerInfo.Op || FieldCombinerInfo.Op->containsErrors()) { + Diag(Loc, diag::err_acc_reduction_recipe_no_op) << FD->getType(); + Diag(FD->getBeginLoc(), diag::note_acc_reduction_recipe_noop_field) << RD; + tryCombiner(FieldCombinerInfo.LHSDRE, FieldCombinerInfo.RHSDRE, + /*IncludeTrap=*/false); + return true; + } + + if (FieldCombinerInfo.Op->isInstantiationDependent()) { + // If this is instantiation dependent, we're just going to 'give up' here + // and count on us to get it right during instantaition. + CombinerRecipes.push_back({nullptr, nullptr, nullptr}); + } else { + CombinerRecipes.push_back( + {FieldCombinerInfo.LHS, FieldCombinerInfo.RHS, FieldCombinerInfo.Op}); + } + } + + return false; } diff --git a/clang/lib/Sema/SemaOpenACCClause.cpp b/clang/lib/Sema/SemaOpenACCClause.cpp index 881e960..17078e8 100644 --- a/clang/lib/Sema/SemaOpenACCClause.cpp +++ b/clang/lib/Sema/SemaOpenACCClause.cpp @@ -1772,7 +1772,7 @@ OpenACCClause *SemaOpenACCClauseVisitor::VisitReductionClause( } SmallVector<Expr *> ValidVars; - SmallVector<OpenACCReductionRecipe> Recipes; + SmallVector<OpenACCReductionRecipeWithStorage> Recipes; for (Expr *Var : Clause.getVarList()) { ExprResult Res = SemaRef.CheckReductionVar(Clause.getDirectiveKind(), @@ -1924,7 +1924,7 @@ bool SemaOpenACC::CheckReductionVarType(Expr *VarExpr) { // off here. This will result in CurType being the actual 'type' of the // expression, which is what we are looking to check. QualType CurType = isa<ArraySectionExpr>(VarExpr) - ? ArraySectionExpr::getBaseOriginalType(VarExpr) + ? cast<ArraySectionExpr>(VarExpr)->getElementType() : VarExpr->getType(); // This can happen when we have a dependent type in an array element that the @@ -2196,7 +2196,7 @@ OpenACCClause *SemaOpenACC::CheckReductionClause( ArrayRef<const OpenACCClause *> ExistingClauses, OpenACCDirectiveKind DirectiveKind, SourceLocation BeginLoc, SourceLocation LParenLoc, OpenACCReductionOperator ReductionOp, - ArrayRef<Expr *> Vars, ArrayRef<OpenACCReductionRecipe> Recipes, + ArrayRef<Expr *> Vars, ArrayRef<OpenACCReductionRecipeWithStorage> Recipes, SourceLocation EndLoc) { if (DirectiveKind == OpenACCDirectiveKind::Loop || isOpenACCCombinedDirectiveKind(DirectiveKind)) { diff --git a/clang/lib/Sema/SemaOpenMP.cpp b/clang/lib/Sema/SemaOpenMP.cpp index 0fa21e8..5b5b1b6 100644 --- a/clang/lib/Sema/SemaOpenMP.cpp +++ b/clang/lib/Sema/SemaOpenMP.cpp @@ -16491,6 +16491,9 @@ OMPClause *SemaOpenMP::ActOnOpenMPSingleExprClause(OpenMPClauseKind Kind, case OMPC_ordered: Res = ActOnOpenMPOrderedClause(StartLoc, EndLoc, LParenLoc, Expr); break; + case OMPC_nowait: + Res = ActOnOpenMPNowaitClause(StartLoc, EndLoc, LParenLoc, Expr); + break; case OMPC_priority: Res = ActOnOpenMPPriorityClause(Expr, StartLoc, LParenLoc, EndLoc); break; @@ -16546,7 +16549,6 @@ OMPClause *SemaOpenMP::ActOnOpenMPSingleExprClause(OpenMPClauseKind Kind, case OMPC_aligned: case OMPC_copyin: case OMPC_copyprivate: - case OMPC_nowait: case OMPC_untied: case OMPC_mergeable: case OMPC_threadprivate: @@ -17955,7 +17957,9 @@ OMPClause *SemaOpenMP::ActOnOpenMPClause(OpenMPClauseKind Kind, Res = ActOnOpenMPOrderedClause(StartLoc, EndLoc); break; case OMPC_nowait: - Res = ActOnOpenMPNowaitClause(StartLoc, EndLoc); + Res = ActOnOpenMPNowaitClause(StartLoc, EndLoc, + /*LParenLoc=*/SourceLocation(), + /*Condition=*/nullptr); break; case OMPC_untied: Res = ActOnOpenMPUntiedClause(StartLoc, EndLoc); @@ -18107,9 +18111,24 @@ OMPClause *SemaOpenMP::ActOnOpenMPClause(OpenMPClauseKind Kind, } OMPClause *SemaOpenMP::ActOnOpenMPNowaitClause(SourceLocation StartLoc, - SourceLocation EndLoc) { + SourceLocation EndLoc, + SourceLocation LParenLoc, + Expr *Condition) { + Expr *ValExpr = Condition; + if (Condition && LParenLoc.isValid()) { + if (!Condition->isValueDependent() && !Condition->isTypeDependent() && + !Condition->isInstantiationDependent() && + !Condition->containsUnexpandedParameterPack()) { + ExprResult Val = SemaRef.CheckBooleanCondition(StartLoc, Condition); + if (Val.isInvalid()) + return nullptr; + + ValExpr = Val.get(); + } + } DSAStack->setNowaitRegion(); - return new (getASTContext()) OMPNowaitClause(StartLoc, EndLoc); + return new (getASTContext()) + OMPNowaitClause(ValExpr, StartLoc, LParenLoc, EndLoc); } OMPClause *SemaOpenMP::ActOnOpenMPUntiedClause(SourceLocation StartLoc, diff --git a/clang/lib/Sema/SemaOverload.cpp b/clang/lib/Sema/SemaOverload.cpp index 8339bb1..7da09e8 100644 --- a/clang/lib/Sema/SemaOverload.cpp +++ b/clang/lib/Sema/SemaOverload.cpp @@ -2594,7 +2594,7 @@ IsTransparentUnionStandardConversion(Sema &S, Expr* From, if (!UT) return false; // The field to initialize within the transparent union. - const RecordDecl *UD = UT->getOriginalDecl()->getDefinitionOrSelf(); + const RecordDecl *UD = UT->getDecl()->getDefinitionOrSelf(); if (!UD->hasAttr<TransparentUnionAttr>()) return false; // It's compatible if the expression matches any of the fields. @@ -3973,7 +3973,7 @@ IsUserDefinedConversion(Sema &S, Expr *From, QualType ToType, if (!S.isCompleteType(From->getExprLoc(), ToType)) { // We're not going to find any constructors. } else if (auto *ToRecordDecl = - dyn_cast<CXXRecordDecl>(ToRecordType->getOriginalDecl())) { + dyn_cast<CXXRecordDecl>(ToRecordType->getDecl())) { ToRecordDecl = ToRecordDecl->getDefinitionOrSelf(); Expr **Args = &From; @@ -4048,7 +4048,7 @@ IsUserDefinedConversion(Sema &S, Expr *From, QualType ToType, } else if (const RecordType *FromRecordType = From->getType()->getAsCanonical<RecordType>()) { if (auto *FromRecordDecl = - dyn_cast<CXXRecordDecl>(FromRecordType->getOriginalDecl())) { + dyn_cast<CXXRecordDecl>(FromRecordType->getDecl())) { FromRecordDecl = FromRecordDecl->getDefinitionOrSelf(); // Add all of the conversion functions as candidates. const auto &Conversions = FromRecordDecl->getVisibleConversionFunctions(); @@ -6840,7 +6840,7 @@ ExprResult Sema::PerformContextualImplicitConversion( UnresolvedSet<4> ViableConversions; // These are *potentially* viable in C++1y. UnresolvedSet<4> ExplicitConversions; - const auto &Conversions = cast<CXXRecordDecl>(RecordTy->getOriginalDecl()) + const auto &Conversions = cast<CXXRecordDecl>(RecordTy->getDecl()) ->getDefinitionOrSelf() ->getVisibleConversionFunctions(); @@ -10194,9 +10194,7 @@ public: if (S.getLangOpts().CPlusPlus11) { for (QualType EnumTy : CandidateTypes[ArgIdx].enumeration_types()) { - if (!EnumTy->castAsCanonical<EnumType>() - ->getOriginalDecl() - ->isScoped()) + if (!EnumTy->castAsCanonical<EnumType>()->getDecl()->isScoped()) continue; if (!AddedTypes.insert(S.Context.getCanonicalType(EnumTy)).second) diff --git a/clang/lib/Sema/SemaSYCL.cpp b/clang/lib/Sema/SemaSYCL.cpp index 2f97f62..67f3856 100644 --- a/clang/lib/Sema/SemaSYCL.cpp +++ b/clang/lib/Sema/SemaSYCL.cpp @@ -199,7 +199,7 @@ void SemaSYCL::handleKernelAttr(Decl *D, const ParsedAttr &AL) { return; } - handleSimpleAttribute<DeviceKernelAttr>(*this, D, AL); + handleSimpleAttribute<SYCLKernelAttr>(*this, D, AL); } void SemaSYCL::handleKernelEntryPointAttr(Decl *D, const ParsedAttr &AL) { @@ -221,8 +221,8 @@ static SourceLocation SourceLocationForUserDeclaredType(QualType QT) { SourceLocation Loc; const Type *T = QT->getUnqualifiedDesugaredType(); if (const TagType *TT = dyn_cast<TagType>(T)) - Loc = TT->getOriginalDecl()->getLocation(); - else if (const ObjCInterfaceType *ObjCIT = dyn_cast<ObjCInterfaceType>(T)) + Loc = TT->getDecl()->getLocation(); + else if (const auto *ObjCIT = dyn_cast<ObjCInterfaceType>(T)) Loc = ObjCIT->getDecl()->getLocation(); return Loc; } diff --git a/clang/lib/Sema/SemaStmt.cpp b/clang/lib/Sema/SemaStmt.cpp index ae0bb616..f398963 100644 --- a/clang/lib/Sema/SemaStmt.cpp +++ b/clang/lib/Sema/SemaStmt.cpp @@ -1277,11 +1277,11 @@ static void checkEnumTypesInSwitchStmt(Sema &S, const Expr *Cond, return; // Ignore anonymous enums. - if (!CondEnumType->getOriginalDecl()->getIdentifier() && - !CondEnumType->getOriginalDecl()->getTypedefNameForAnonDecl()) + if (!CondEnumType->getDecl()->getIdentifier() && + !CondEnumType->getDecl()->getTypedefNameForAnonDecl()) return; - if (!CaseEnumType->getOriginalDecl()->getIdentifier() && - !CaseEnumType->getOriginalDecl()->getTypedefNameForAnonDecl()) + if (!CaseEnumType->getDecl()->getIdentifier() && + !CaseEnumType->getDecl()->getTypedefNameForAnonDecl()) return; if (S.Context.hasSameUnqualifiedType(CondType, CaseType)) @@ -3760,7 +3760,7 @@ private: Sema &S; }; bool LocalTypedefNameReferencer::VisitRecordType(RecordType *RT) { - auto *R = dyn_cast<CXXRecordDecl>(RT->getOriginalDecl()); + auto *R = dyn_cast<CXXRecordDecl>(RT->getDecl()); if (!R || !R->isLocalClass() || !R->isLocalClass()->isExternallyVisible() || R->isDependentType()) return true; @@ -3979,7 +3979,7 @@ StmtResult Sema::BuildReturnStmt(SourceLocation ReturnLoc, Expr *RetValExp, << RetValExp->getSourceRange(); if (FD->hasAttr<CmseNSEntryAttr>() && RetValExp) { if (const auto *RT = dyn_cast<RecordType>(FnRetType.getCanonicalType())) { - if (RT->getOriginalDecl()->isOrContainsUnion()) + if (RT->getDecl()->isOrContainsUnion()) Diag(RetValExp->getBeginLoc(), diag::warn_cmse_nonsecure_union) << 1; } } diff --git a/clang/lib/Sema/SemaStmtAsm.cpp b/clang/lib/Sema/SemaStmtAsm.cpp index 0438af7..f957bdf 100644 --- a/clang/lib/Sema/SemaStmtAsm.cpp +++ b/clang/lib/Sema/SemaStmtAsm.cpp @@ -908,7 +908,7 @@ bool Sema::LookupInlineAsmField(StringRef Base, StringRef Member, LookupResult FieldResult(*this, &Context.Idents.get(NextMember), SourceLocation(), LookupMemberName); - RecordDecl *RD = RT->getOriginalDecl()->getDefinitionOrSelf(); + RecordDecl *RD = RT->getDecl()->getDefinitionOrSelf(); if (!LookupQualifiedName(FieldResult, RD)) return true; diff --git a/clang/lib/Sema/SemaTemplate.cpp b/clang/lib/Sema/SemaTemplate.cpp index 3a6ff99..2cc6593 100644 --- a/clang/lib/Sema/SemaTemplate.cpp +++ b/clang/lib/Sema/SemaTemplate.cpp @@ -408,9 +408,7 @@ bool Sema::LookupTemplateName(LookupResult &Found, Scope *S, CXXScopeSpec &SS, IsDependent = !LookupCtx && ObjectType->isDependentType(); assert((IsDependent || !ObjectType->isIncompleteType() || !ObjectType->getAs<TagType>() || - ObjectType->castAs<TagType>() - ->getOriginalDecl() - ->isEntityBeingDefined()) && + ObjectType->castAs<TagType>()->getDecl()->isEntityBeingDefined()) && "Caller should have completed object type"); // Template names cannot appear inside an Objective-C class or object type @@ -1819,7 +1817,7 @@ public: } bool VisitTagType(const TagType *T) override { - return TraverseDecl(T->getOriginalDecl()); + return TraverseDecl(T->getDecl()); } bool TraverseDecl(const Decl *D) override { @@ -2790,7 +2788,7 @@ struct DependencyChecker : DynamicRecursiveASTVisitor { // An InjectedClassNameType will never have a dependent template name, // so no need to traverse it. return TraverseTemplateArguments( - T->getTemplateArgs(T->getOriginalDecl()->getASTContext())); + T->getTemplateArgs(T->getDecl()->getASTContext())); } }; } // end anonymous namespace @@ -2914,7 +2912,7 @@ TemplateParameterList *Sema::MatchTemplateParametersToScopeSpecifier( if (const EnumType *EnumT = T->getAsCanonical<EnumType>()) { // FIXME: Forward-declared enums require a TSK_ExplicitSpecialization // check here. - EnumDecl *Enum = EnumT->getOriginalDecl(); + EnumDecl *Enum = EnumT->getDecl(); // Get to the parent type. if (TypeDecl *Parent = dyn_cast<TypeDecl>(Enum->getParent())) @@ -3352,7 +3350,7 @@ static QualType builtinCommonTypeImpl(Sema &S, ElaboratedTypeKeyword Keyword, } static bool isInVkNamespace(const RecordType *RT) { - DeclContext *DC = RT->getOriginalDecl()->getDeclContext(); + DeclContext *DC = RT->getDecl()->getDeclContext(); if (!DC) return false; @@ -3369,9 +3367,8 @@ static SpirvOperand checkHLSLSpirvTypeOperand(Sema &SemaRef, if (auto *RT = OperandArg->getAsCanonical<RecordType>()) { bool Literal = false; SourceLocation LiteralLoc; - if (isInVkNamespace(RT) && RT->getOriginalDecl()->getName() == "Literal") { - auto SpecDecl = - dyn_cast<ClassTemplateSpecializationDecl>(RT->getOriginalDecl()); + if (isInVkNamespace(RT) && RT->getDecl()->getName() == "Literal") { + auto SpecDecl = dyn_cast<ClassTemplateSpecializationDecl>(RT->getDecl()); assert(SpecDecl); const TemplateArgumentList &LiteralArgs = SpecDecl->getTemplateArgs(); @@ -3382,9 +3379,8 @@ static SpirvOperand checkHLSLSpirvTypeOperand(Sema &SemaRef, } if (RT && isInVkNamespace(RT) && - RT->getOriginalDecl()->getName() == "integral_constant") { - auto SpecDecl = - dyn_cast<ClassTemplateSpecializationDecl>(RT->getOriginalDecl()); + RT->getDecl()->getName() == "integral_constant") { + auto SpecDecl = dyn_cast<ClassTemplateSpecializationDecl>(RT->getDecl()); assert(SpecDecl); const TemplateArgumentList &ConstantArgs = SpecDecl->getTemplateArgs(); @@ -4110,7 +4106,7 @@ TypeResult Sema::ActOnTagTemplateIdType(TagUseKind TUK, // Check the tag kind if (const RecordType *RT = Result->getAs<RecordType>()) { - RecordDecl *D = RT->getOriginalDecl(); + RecordDecl *D = RT->getDecl(); IdentifierInfo *Id = D->getIdentifier(); assert(Id && "templated class must have an identifier"); @@ -6383,11 +6379,11 @@ bool UnnamedLocalNoLinkageFinder::VisitDeducedTemplateSpecializationType( } bool UnnamedLocalNoLinkageFinder::VisitRecordType(const RecordType* T) { - return VisitTagDecl(T->getOriginalDecl()->getDefinitionOrSelf()); + return VisitTagDecl(T->getDecl()->getDefinitionOrSelf()); } bool UnnamedLocalNoLinkageFinder::VisitEnumType(const EnumType* T) { - return VisitTagDecl(T->getOriginalDecl()->getDefinitionOrSelf()); + return VisitTagDecl(T->getDecl()->getDefinitionOrSelf()); } bool UnnamedLocalNoLinkageFinder::VisitTemplateTypeParmType( @@ -6412,7 +6408,7 @@ bool UnnamedLocalNoLinkageFinder::VisitTemplateSpecializationType( bool UnnamedLocalNoLinkageFinder::VisitInjectedClassNameType( const InjectedClassNameType* T) { - return VisitTagDecl(T->getOriginalDecl()->getDefinitionOrSelf()); + return VisitTagDecl(T->getDecl()->getDefinitionOrSelf()); } bool UnnamedLocalNoLinkageFinder::VisitDependentNameType( diff --git a/clang/lib/Sema/SemaTemplateDeduction.cpp b/clang/lib/Sema/SemaTemplateDeduction.cpp index 3baa977..6964242 100644 --- a/clang/lib/Sema/SemaTemplateDeduction.cpp +++ b/clang/lib/Sema/SemaTemplateDeduction.cpp @@ -5577,7 +5577,7 @@ static TemplateDeductionResult CheckDeductionConsistency( bool IsDeductionGuide = isa<CXXDeductionGuideDecl>(FTD->getTemplatedDecl()); if (IsDeductionGuide) { if (auto *Injected = P->getAsCanonical<InjectedClassNameType>()) - P = Injected->getOriginalDecl()->getCanonicalTemplateSpecializationType( + P = Injected->getDecl()->getCanonicalTemplateSpecializationType( S.Context); } QualType InstP = S.SubstType(P.getCanonicalType(), MLTAL, FTD->getLocation(), @@ -5598,10 +5598,10 @@ static TemplateDeductionResult CheckDeductionConsistency( auto T2 = S.Context.getUnqualifiedArrayType(A.getNonReferenceType()); if (IsDeductionGuide) { if (auto *Injected = T1->getAsCanonical<InjectedClassNameType>()) - T1 = Injected->getOriginalDecl()->getCanonicalTemplateSpecializationType( + T1 = Injected->getDecl()->getCanonicalTemplateSpecializationType( S.Context); if (auto *Injected = T2->getAsCanonical<InjectedClassNameType>()) - T2 = Injected->getOriginalDecl()->getCanonicalTemplateSpecializationType( + T2 = Injected->getDecl()->getCanonicalTemplateSpecializationType( S.Context); } if (!S.Context.hasSameType(T1, T2)) @@ -6973,7 +6973,7 @@ MarkUsedTemplateParameters(ASTContext &Ctx, QualType T, case Type::InjectedClassName: T = cast<InjectedClassNameType>(T) - ->getOriginalDecl() + ->getDecl() ->getCanonicalTemplateSpecializationType(Ctx); [[fallthrough]]; diff --git a/clang/lib/Sema/SemaTemplateDeductionGuide.cpp b/clang/lib/Sema/SemaTemplateDeductionGuide.cpp index 8ba23aa..ad50600 100644 --- a/clang/lib/Sema/SemaTemplateDeductionGuide.cpp +++ b/clang/lib/Sema/SemaTemplateDeductionGuide.cpp @@ -996,7 +996,7 @@ getRHSTemplateDeclAndArgs(Sema &SemaRef, TypeAliasTemplateDecl *AliasTemplate) { // dependent. e.g. // using AliasFoo = Foo<bool>; if (const auto *CTSD = - dyn_cast<ClassTemplateSpecializationDecl>(RT->getOriginalDecl())) { + dyn_cast<ClassTemplateSpecializationDecl>(RT->getDecl())) { Template = CTSD->getSpecializedTemplate(); AliasRhsTemplateArgs = CTSD->getTemplateArgs().asArray(); } @@ -1054,12 +1054,11 @@ BuildDeductionGuideForTypeAlias(Sema &SemaRef, // such that T can be deduced as U. auto RType = F->getTemplatedDecl()->getReturnType(); // The (trailing) return type of the deduction guide. - const TemplateSpecializationType *FReturnType = - RType->getAs<TemplateSpecializationType>(); + const auto *FReturnType = RType->getAs<TemplateSpecializationType>(); if (const auto *ICNT = RType->getAsCanonical<InjectedClassNameType>()) // implicitly-generated deduction guide. FReturnType = cast<TemplateSpecializationType>( - ICNT->getOriginalDecl()->getCanonicalTemplateSpecializationType( + ICNT->getDecl()->getCanonicalTemplateSpecializationType( SemaRef.Context)); assert(FReturnType && "expected to see a return type"); // Deduce template arguments of the deduction guide f from the RHS of diff --git a/clang/lib/Sema/SemaTemplateInstantiate.cpp b/clang/lib/Sema/SemaTemplateInstantiate.cpp index 7b05e4c..ca7e3b2 100644 --- a/clang/lib/Sema/SemaTemplateInstantiate.cpp +++ b/clang/lib/Sema/SemaTemplateInstantiate.cpp @@ -35,6 +35,7 @@ #include "clang/Sema/Template.h" #include "clang/Sema/TemplateDeduction.h" #include "clang/Sema/TemplateInstCallback.h" +#include "llvm/ADT/SmallVectorExtras.h" #include "llvm/ADT/StringExtras.h" #include "llvm/Support/ErrorHandling.h" #include "llvm/Support/SaveAndRestore.h" @@ -1681,7 +1682,7 @@ namespace { ICNT && SemaRef.CodeSynthesisContexts.back().Kind == Sema::CodeSynthesisContext::BuildingDeductionGuides) { Type = inherited::TransformType( - ICNT->getOriginalDecl()->getCanonicalTemplateSpecializationType( + ICNT->getDecl()->getCanonicalTemplateSpecializationType( SemaRef.Context)); TLB.pushTrivial(SemaRef.Context, Type, TL.getNameLoc()); } @@ -2105,7 +2106,7 @@ TemplateInstantiator::TransformFirstQualifierInScope(NamedDecl *D, return cast_or_null<NamedDecl>(TransformDecl(Loc, D)); if (const TagType *Tag = T->getAs<TagType>()) - return Tag->getOriginalDecl(); + return Tag->getDecl(); // The resulting type is not a tag; complain. getSema().Diag(Loc, diag::err_nested_name_spec_non_tag) << T; @@ -4487,6 +4488,119 @@ ExprResult Sema::SubstConstraintExprWithoutSatisfaction( return Instantiator.TransformExpr(E); } +ExprResult Sema::SubstConceptTemplateArguments( + const ConceptSpecializationExpr *CSE, const Expr *ConstraintExpr, + const MultiLevelTemplateArgumentList &MLTAL) { + TemplateInstantiator Instantiator(*this, MLTAL, SourceLocation(), + DeclarationName()); + const ASTTemplateArgumentListInfo *ArgsAsWritten = + CSE->getTemplateArgsAsWritten(); + TemplateArgumentListInfo SubstArgs(ArgsAsWritten->getLAngleLoc(), + ArgsAsWritten->getRAngleLoc()); + + Sema::InstantiatingTemplate Inst( + *this, ArgsAsWritten->arguments().front().getSourceRange().getBegin(), + Sema::InstantiatingTemplate::ConstraintNormalization{}, + CSE->getNamedConcept(), + ArgsAsWritten->arguments().front().getSourceRange()); + + if (Inst.isInvalid()) + return ExprError(); + + if (Instantiator.TransformConceptTemplateArguments( + ArgsAsWritten->getTemplateArgs(), + ArgsAsWritten->getTemplateArgs() + + ArgsAsWritten->getNumTemplateArgs(), + SubstArgs)) + return true; + + llvm::SmallVector<TemplateArgument, 4> NewArgList = llvm::map_to_vector( + SubstArgs.arguments(), + [](const TemplateArgumentLoc &Loc) { return Loc.getArgument(); }); + + MultiLevelTemplateArgumentList MLTALForConstraint = + getTemplateInstantiationArgs( + CSE->getNamedConcept(), + CSE->getNamedConcept()->getLexicalDeclContext(), + /*Final=*/false, + /*Innermost=*/NewArgList, + /*RelativeToPrimary=*/true, + /*Pattern=*/nullptr, + /*ForConstraintInstantiation=*/true); + + // Rebuild a constraint, only substituting non-dependent concept names + // and nothing else. + // Given C<SomeType, SomeValue, SomeConceptName, SomeDependentConceptName>. + // only SomeConceptName is substituted, in the constraint expression of C. + struct ConstraintExprTransformer : TreeTransform<ConstraintExprTransformer> { + using Base = TreeTransform<ConstraintExprTransformer>; + MultiLevelTemplateArgumentList &MLTAL; + + ConstraintExprTransformer(Sema &SemaRef, + MultiLevelTemplateArgumentList &MLTAL) + : TreeTransform(SemaRef), MLTAL(MLTAL) {} + + ExprResult TransformExpr(Expr *E) { + if (!E) + return E; + switch (E->getStmtClass()) { + case Stmt::BinaryOperatorClass: + case Stmt::ConceptSpecializationExprClass: + case Stmt::ParenExprClass: + case Stmt::UnresolvedLookupExprClass: + return Base::TransformExpr(E); + default: + break; + } + return E; + } + + // Rebuild both branches of a conjunction / disjunction + // even if there is a substitution failure in one of + // the branch. + ExprResult TransformBinaryOperator(BinaryOperator *E) { + if (!(E->getOpcode() == BinaryOperatorKind::BO_LAnd || + E->getOpcode() == BinaryOperatorKind::BO_LOr)) + return E; + + ExprResult LHS = TransformExpr(E->getLHS()); + ExprResult RHS = TransformExpr(E->getRHS()); + + if (LHS.get() == E->getLHS() && RHS.get() == E->getRHS()) + return E; + + return BinaryOperator::Create(SemaRef.Context, LHS.get(), RHS.get(), + E->getOpcode(), SemaRef.Context.BoolTy, + VK_PRValue, OK_Ordinary, + E->getOperatorLoc(), FPOptionsOverride{}); + } + + bool TransformTemplateArgument(const TemplateArgumentLoc &Input, + TemplateArgumentLoc &Output, + bool Uneval = false) { + if (Input.getArgument().isConceptOrConceptTemplateParameter()) + return Base::TransformTemplateArgument(Input, Output, Uneval); + + Output = Input; + return false; + } + + ExprResult TransformUnresolvedLookupExpr(UnresolvedLookupExpr *E, + bool IsAddressOfOperand = false) { + if (E->isConceptReference()) { + ExprResult Res = SemaRef.SubstExpr(E, MLTAL); + return Res; + } + return E; + } + }; + + ConstraintExprTransformer Transformer(*this, MLTALForConstraint); + ExprResult Res = + Transformer.TransformExpr(const_cast<Expr *>(ConstraintExpr)); + return Res; +} + ExprResult Sema::SubstInitializer(Expr *Init, const MultiLevelTemplateArgumentList &TemplateArgs, bool CXXDirectInit) { diff --git a/clang/lib/Sema/SemaTemplateInstantiateDecl.cpp b/clang/lib/Sema/SemaTemplateInstantiateDecl.cpp index 3819f77..4863b45 100644 --- a/clang/lib/Sema/SemaTemplateInstantiateDecl.cpp +++ b/clang/lib/Sema/SemaTemplateInstantiateDecl.cpp @@ -707,12 +707,29 @@ static void instantiateDependentAMDGPUMaxNumWorkGroupsAttr( S.AMDGPU().addAMDGPUMaxNumWorkGroupsAttr(New, Attr, XExpr, YExpr, ZExpr); } +static void instantiateDependentCUDAClusterDimsAttr( + Sema &S, const MultiLevelTemplateArgumentList &TemplateArgs, + const CUDAClusterDimsAttr &Attr, Decl *New) { + EnterExpressionEvaluationContext Unevaluated( + S, Sema::ExpressionEvaluationContext::ConstantEvaluated); + + auto SubstElt = [&S, &TemplateArgs](Expr *E) { + return E ? S.SubstExpr(E, TemplateArgs).get() : nullptr; + }; + + Expr *XExpr = SubstElt(Attr.getX()); + Expr *YExpr = SubstElt(Attr.getY()); + Expr *ZExpr = SubstElt(Attr.getZ()); + + S.addClusterDimsAttr(New, Attr, XExpr, YExpr, ZExpr); +} + // This doesn't take any template parameters, but we have a custom action that // needs to happen when the kernel itself is instantiated. We need to run the // ItaniumMangler to mark the names required to name this kernel. -static void instantiateDependentDeviceKernelAttr( +static void instantiateDependentSYCLKernelAttr( Sema &S, const MultiLevelTemplateArgumentList &TemplateArgs, - const DeviceKernelAttr &Attr, Decl *New) { + const SYCLKernelAttr &Attr, Decl *New) { New->addAttr(Attr.clone(S.getASTContext())); } @@ -765,10 +782,18 @@ static bool isRelevantAttr(Sema &S, const Decl *D, const Attr *A) { static void instantiateDependentHLSLParamModifierAttr( Sema &S, const MultiLevelTemplateArgumentList &TemplateArgs, - const HLSLParamModifierAttr *Attr, Decl *New) { - ParmVarDecl *P = cast<ParmVarDecl>(New); - P->addAttr(Attr->clone(S.getASTContext())); - P->setType(S.HLSL().getInoutParameterType(P->getType())); + const HLSLParamModifierAttr *Attr, const Decl *Old, Decl *New) { + ParmVarDecl *NewParm = cast<ParmVarDecl>(New); + NewParm->addAttr(Attr->clone(S.getASTContext())); + + const Type *OldParmTy = cast<ParmVarDecl>(Old)->getType().getTypePtr(); + if (OldParmTy->isDependentType() && Attr->isAnyOut()) + NewParm->setType(S.HLSL().getInoutParameterType(NewParm->getType())); + + assert( + (!Attr->isAnyOut() || (NewParm->getType().isRestrictQualified() && + NewParm->getType()->isReferenceType())) && + "out or inout parameter type must be a reference and restrict qualified"); } void Sema::InstantiateAttrsForDecl( @@ -921,9 +946,14 @@ void Sema::InstantiateAttrs(const MultiLevelTemplateArgumentList &TemplateArgs, *this, TemplateArgs, *AMDGPUMaxNumWorkGroups, New); } + if (const auto *CUDAClusterDims = dyn_cast<CUDAClusterDimsAttr>(TmplAttr)) { + instantiateDependentCUDAClusterDimsAttr(*this, TemplateArgs, + *CUDAClusterDims, New); + } + if (const auto *ParamAttr = dyn_cast<HLSLParamModifierAttr>(TmplAttr)) { instantiateDependentHLSLParamModifierAttr(*this, TemplateArgs, ParamAttr, - New); + Tmpl, New); continue; } @@ -966,8 +996,8 @@ void Sema::InstantiateAttrs(const MultiLevelTemplateArgumentList &TemplateArgs, continue; } - if (auto *A = dyn_cast<DeviceKernelAttr>(TmplAttr)) { - instantiateDependentDeviceKernelAttr(*this, TemplateArgs, *A, New); + if (auto *A = dyn_cast<SYCLKernelAttr>(TmplAttr)) { + instantiateDependentSYCLKernelAttr(*this, TemplateArgs, *A, New); continue; } @@ -1522,9 +1552,9 @@ Decl *TemplateDeclInstantiator::InstantiateTypedefNameDecl(TypedefNameDecl *D, // If the old typedef was the name for linkage purposes of an anonymous // tag decl, re-establish that relationship for the new typedef. if (const TagType *oldTagType = D->getUnderlyingType()->getAs<TagType>()) { - TagDecl *oldTag = oldTagType->getOriginalDecl(); + TagDecl *oldTag = oldTagType->getDecl(); if (oldTag->getTypedefNameForAnonDecl() == D && !Invalid) { - TagDecl *newTag = DI->getType()->castAs<TagType>()->getOriginalDecl(); + TagDecl *newTag = DI->getType()->castAs<TagType>()->getDecl(); assert(!newTag->hasNameForLinkage()); newTag->setTypedefNameForAnonDecl(Typedef); } @@ -5727,7 +5757,7 @@ void Sema::InstantiateFunctionDefinition(SourceLocation PointOfInstantiation, Function->setDeclarationNameLoc(NameLocPointsToPattern()); EnterExpressionEvaluationContextForFunction EvalContext( - *this, Sema::ExpressionEvaluationContext::PotentiallyEvaluated); + *this, Sema::ExpressionEvaluationContext::PotentiallyEvaluated, Function); Qualifiers ThisTypeQuals; CXXRecordDecl *ThisContext = nullptr; @@ -5791,7 +5821,7 @@ void Sema::InstantiateFunctionDefinition(SourceLocation PointOfInstantiation, QualType TransformRecordType(TypeLocBuilder &TLB, RecordTypeLoc TL) { const RecordType *T = TL.getTypePtr(); RecordDecl *Record = cast_or_null<RecordDecl>( - getDerived().TransformDecl(TL.getNameLoc(), T->getOriginalDecl())); + getDerived().TransformDecl(TL.getNameLoc(), T->getDecl())); if (Record != OldDecl) return Base::TransformRecordType(TLB, TL); diff --git a/clang/lib/Sema/SemaType.cpp b/clang/lib/Sema/SemaType.cpp index bee613a..7c1fb12 100644 --- a/clang/lib/Sema/SemaType.cpp +++ b/clang/lib/Sema/SemaType.cpp @@ -1238,8 +1238,8 @@ static QualType ConvertDeclSpecToType(TypeProcessingState &state) { Result = S.GetTypeFromParser(DS.getRepAsType()); assert(!Result.isNull() && "Didn't get a type for typeof?"); if (!Result->isDependentType()) - if (const TagType *TT = Result->getAs<TagType>()) - S.DiagnoseUseOfDecl(TT->getOriginalDecl(), DS.getTypeSpecTypeLoc()); + if (const auto *TT = Result->getAs<TagType>()) + S.DiagnoseUseOfDecl(TT->getDecl(), DS.getTypeSpecTypeLoc()); // TypeQuals handled by caller. Result = Context.getTypeOfType( Result, DS.getTypeSpecType() == DeclSpec::TST_typeof_unqualType @@ -2517,12 +2517,18 @@ QualType Sema::BuildMatrixType(QualType ElementTy, Expr *NumRows, Expr *NumCols, Diag(AttrLoc, diag::err_attribute_zero_size) << "matrix" << ColRange; return QualType(); } - if (!ConstantMatrixType::isDimensionValid(MatrixRows)) { + if (MatrixRows > Context.getLangOpts().MaxMatrixDimension && + MatrixColumns > Context.getLangOpts().MaxMatrixDimension) { + Diag(AttrLoc, diag::err_attribute_size_too_large) + << RowRange << ColRange << "matrix row and column"; + return QualType(); + } + if (MatrixRows > Context.getLangOpts().MaxMatrixDimension) { Diag(AttrLoc, diag::err_attribute_size_too_large) << RowRange << "matrix row"; return QualType(); } - if (!ConstantMatrixType::isDimensionValid(MatrixColumns)) { + if (MatrixColumns > Context.getLangOpts().MaxMatrixDimension) { Diag(AttrLoc, diag::err_attribute_size_too_large) << ColRange << "matrix column"; return QualType(); @@ -3780,12 +3786,10 @@ static CallingConv getCCForDeclaratorChunk( } } } - if (!S.getLangOpts().isSYCL()) { - for (const ParsedAttr &AL : D.getDeclSpec().getAttributes()) { - if (AL.getKind() == ParsedAttr::AT_DeviceKernel) { - CC = CC_DeviceKernel; - break; - } + for (const ParsedAttr &AL : D.getDeclSpec().getAttributes()) { + if (AL.getKind() == ParsedAttr::AT_DeviceKernel) { + CC = CC_DeviceKernel; + break; } } return CC; @@ -9701,7 +9705,7 @@ QualType Sema::BuildTypeofExprType(Expr *E, TypeOfKind Kind) { if (!E->isTypeDependent()) { QualType T = E->getType(); if (const TagType *TT = T->getAs<TagType>()) - DiagnoseUseOfDecl(TT->getOriginalDecl(), E->getExprLoc()); + DiagnoseUseOfDecl(TT->getDecl(), E->getExprLoc()); } return Context.getTypeOfExprType(E, Kind); } @@ -9867,7 +9871,7 @@ QualType Sema::BuildPackIndexingType(QualType Pattern, Expr *IndexExpr, static QualType GetEnumUnderlyingType(Sema &S, QualType BaseType, SourceLocation Loc) { assert(BaseType->isEnumeralType()); - EnumDecl *ED = BaseType->castAs<EnumType>()->getOriginalDecl(); + EnumDecl *ED = BaseType->castAs<EnumType>()->getDecl(); S.DiagnoseUseOfDecl(ED, Loc); diff --git a/clang/lib/Sema/SemaTypeTraits.cpp b/clang/lib/Sema/SemaTypeTraits.cpp index 3e34675..3887796 100644 --- a/clang/lib/Sema/SemaTypeTraits.cpp +++ b/clang/lib/Sema/SemaTypeTraits.cpp @@ -1076,8 +1076,7 @@ static bool EvaluateUnaryTypeTrait(Sema &Self, TypeTrait UTT, if (T.isPODType(C) || T->isObjCLifetimeType()) return true; if (CXXRecordDecl *RD = C.getBaseElementType(T)->getAsCXXRecordDecl()) { - if (RD->hasTrivialDefaultConstructor() && - !RD->hasNonTrivialDefaultConstructor()) + if (RD->hasTrivialDefaultConstructor()) return true; bool FoundConstructor = false; @@ -1165,14 +1164,26 @@ static bool EvaluateUnaryTypeTrait(Sema &Self, TypeTrait UTT, const CXXDestructorDecl *Dtor = RD->getDestructor(); if (UnqualT->isAggregateType() && (!Dtor || !Dtor->isUserProvided())) return true; - if (RD->hasTrivialDestructor() && (!Dtor || !Dtor->isDeleted())) { - for (CXXConstructorDecl *Ctr : RD->ctors()) { - if (Ctr->isIneligibleOrNotSelected() || Ctr->isDeleted()) - continue; - if (Ctr->isTrivial()) - return true; - } + bool HasTrivialNonDeletedDtr = + RD->hasTrivialDestructor() && (!Dtor || !Dtor->isDeleted()); + if (!HasTrivialNonDeletedDtr) + return false; + for (CXXConstructorDecl *Ctr : RD->ctors()) { + if (Ctr->isIneligibleOrNotSelected() || Ctr->isDeleted()) + continue; + if (Ctr->isTrivial()) + return true; } + if (RD->needsImplicitDefaultConstructor() && + RD->hasTrivialDefaultConstructor() && + !RD->hasNonTrivialDefaultConstructor()) + return true; + if (RD->needsImplicitCopyConstructor() && RD->hasTrivialCopyConstructor() && + !RD->defaultedCopyConstructorIsDeleted()) + return true; + if (RD->needsImplicitMoveConstructor() && RD->hasTrivialMoveConstructor() && + !RD->defaultedMoveConstructorIsDeleted()) + return true; return false; } case UTT_IsIntangibleType: @@ -1613,9 +1624,9 @@ bool Sema::BuiltinIsBaseOf(SourceLocation RhsTLoc, QualType LhsT, // Unions are never base classes, and never have base classes. // It doesn't matter if they are complete or not. See PR#41843 - if (lhsRecord && lhsRecord->getOriginalDecl()->isUnion()) + if (lhsRecord && lhsRecord->getDecl()->isUnion()) return false; - if (rhsRecord && rhsRecord->getOriginalDecl()->isUnion()) + if (rhsRecord && rhsRecord->getDecl()->isUnion()) return false; if (lhsRecord == rhsRecord) @@ -1629,8 +1640,8 @@ bool Sema::BuiltinIsBaseOf(SourceLocation RhsTLoc, QualType LhsT, diag::err_incomplete_type_used_in_type_trait_expr)) return false; - return cast<CXXRecordDecl>(rhsRecord->getOriginalDecl()) - ->isDerivedFrom(cast<CXXRecordDecl>(lhsRecord->getOriginalDecl())); + return cast<CXXRecordDecl>(rhsRecord->getDecl()) + ->isDerivedFrom(cast<CXXRecordDecl>(lhsRecord->getDecl())); } static bool EvaluateBinaryTypeTrait(Sema &Self, TypeTrait BTT, @@ -1670,9 +1681,8 @@ static bool EvaluateBinaryTypeTrait(Sema &Self, TypeTrait BTT, diag::err_incomplete_type)) return false; - return cast<CXXRecordDecl>(DerivedRecord->getOriginalDecl()) - ->isVirtuallyDerivedFrom( - cast<CXXRecordDecl>(BaseRecord->getOriginalDecl())); + return cast<CXXRecordDecl>(DerivedRecord->getDecl()) + ->isVirtuallyDerivedFrom(cast<CXXRecordDecl>(BaseRecord->getDecl())); } case BTT_IsSame: return Self.Context.hasSameType(LhsT, RhsT); diff --git a/clang/lib/Sema/TreeTransform.h b/clang/lib/Sema/TreeTransform.h index 940324b..29f0c30 100644 --- a/clang/lib/Sema/TreeTransform.h +++ b/clang/lib/Sema/TreeTransform.h @@ -694,6 +694,12 @@ public: TemplateArgumentListInfo &Outputs, bool Uneval = false); + template <typename InputIterator> + bool TransformConceptTemplateArguments(InputIterator First, + InputIterator Last, + TemplateArgumentListInfo &Outputs, + bool Uneval = false); + /// Checks if the argument pack from \p In will need to be expanded and does /// the necessary prework. /// Whether the expansion is needed is captured in Info.Expand. @@ -1865,6 +1871,17 @@ public: LParenLoc, Num); } + /// Build a new OpenMP 'nowait' clause. + /// + /// By default, performs semantic analysis to build the new OpenMP clause. + /// Subclasses may override this routine to provide different behavior. + OMPClause *RebuildOMPNowaitClause(Expr *Condition, SourceLocation StartLoc, + SourceLocation LParenLoc, + SourceLocation EndLoc) { + return getSema().OpenMP().ActOnOpenMPNowaitClause(StartLoc, EndLoc, + LParenLoc, Condition); + } + /// Build a new OpenMP 'private' clause. /// /// By default, performs semantic analysis to build the new OpenMP clause. @@ -5181,6 +5198,49 @@ bool TreeTransform<Derived>::TransformTemplateArguments( return false; } +template <typename Derived> +template <typename InputIterator> +bool TreeTransform<Derived>::TransformConceptTemplateArguments( + InputIterator First, InputIterator Last, TemplateArgumentListInfo &Outputs, + bool Uneval) { + + // [C++26][temp.constr.normal] + // any non-dependent concept template argument + // is substituted into the constraint-expression of C. + auto isNonDependentConceptArgument = [](const TemplateArgument &Arg) { + return !Arg.isDependent() && Arg.isConceptOrConceptTemplateParameter(); + }; + + for (; First != Last; ++First) { + TemplateArgumentLoc Out; + TemplateArgumentLoc In = *First; + + if (In.getArgument().getKind() == TemplateArgument::Pack) { + typedef TemplateArgumentLocInventIterator<Derived, + TemplateArgument::pack_iterator> + PackLocIterator; + if (TransformConceptTemplateArguments( + PackLocIterator(*this, In.getArgument().pack_begin()), + PackLocIterator(*this, In.getArgument().pack_end()), Outputs, + Uneval)) + return true; + continue; + } + + if (!isNonDependentConceptArgument(In.getArgument())) { + Outputs.addArgument(In); + continue; + } + + if (getDerived().TransformTemplateArgument(In, Out, Uneval)) + return true; + + Outputs.addArgument(Out); + } + + return false; +} + // FIXME: Find ways to reduce code duplication for pack expansions. template <typename Derived> bool TreeTransform<Derived>::PreparePackForExpansion(TemplateArgumentLoc In, @@ -7149,13 +7209,13 @@ QualType TreeTransform<Derived>::TransformTagType(TypeLocBuilder &TLB, } auto *TD = cast_or_null<TagDecl>( - getDerived().TransformDecl(TL.getNameLoc(), T->getOriginalDecl())); + getDerived().TransformDecl(TL.getNameLoc(), T->getDecl())); if (!TD) return QualType(); QualType Result = TL.getType(); if (getDerived().AlwaysRebuild() || QualifierLoc != TL.getQualifierLoc() || - TD != T->getOriginalDecl()) { + TD != T->getDecl()) { if (T->isCanonicalUnqualified()) Result = getDerived().RebuildCanonicalTagType(TD); else @@ -10612,8 +10672,14 @@ TreeTransform<Derived>::TransformOMPDetachClause(OMPDetachClause *C) { template <typename Derived> OMPClause * TreeTransform<Derived>::TransformOMPNowaitClause(OMPNowaitClause *C) { - // No need to rebuild this clause, no template-dependent parameters. - return C; + ExprResult Cond; + if (auto *Condition = C->getCondition()) { + Cond = getDerived().TransformExpr(Condition); + if (Cond.isInvalid()) + return nullptr; + } + return getDerived().RebuildOMPNowaitClause(Cond.get(), C->getBeginLoc(), + C->getLParenLoc(), C->getEndLoc()); } template <typename Derived> @@ -12374,7 +12440,7 @@ void OpenACCClauseTransform<Derived>::VisitReductionClause( const OpenACCReductionClause &C) { SmallVector<Expr *> TransformedVars = VisitVarList(C.getVarList()); SmallVector<Expr *> ValidVars; - llvm::SmallVector<OpenACCReductionRecipe> Recipes; + llvm::SmallVector<OpenACCReductionRecipeWithStorage> Recipes; for (const auto [Var, OrigRecipe] : llvm::zip(TransformedVars, C.getRecipes())) { @@ -12384,7 +12450,7 @@ void OpenACCClauseTransform<Derived>::VisitReductionClause( ValidVars.push_back(Res.get()); if (OrigRecipe.isSet()) - Recipes.push_back(OrigRecipe); + Recipes.emplace_back(OrigRecipe.AllocaDecl, OrigRecipe.CombinerRecipes); else Recipes.push_back(Self.getSema().OpenACC().CreateReductionInitRecipe( C.getReductionOp(), Res.get())); diff --git a/clang/lib/Serialization/ASTReader.cpp b/clang/lib/Serialization/ASTReader.cpp index 868f0cc..8b3fd41 100644 --- a/clang/lib/Serialization/ASTReader.cpp +++ b/clang/lib/Serialization/ASTReader.cpp @@ -3190,6 +3190,10 @@ ASTReader::ReadControlBlock(ModuleFile &F, F.Kind == MK_ImplicitModule) N = ForceValidateUserInputs ? NumUserInputs : 0; + if (N != 0) + Diag(diag::remark_module_validation) + << N << F.ModuleName << F.FileName; + for (unsigned I = 0; I < N; ++I) { InputFile IF = getInputFile(F, I+1, Complain); if (!IF.getFile() || IF.isOutOfDate()) @@ -5500,7 +5504,7 @@ void ASTReader::InitializeContext() { Error("Invalid FILE type in AST file"); return; } - Context.setFILEDecl(Tag->getOriginalDecl()); + Context.setFILEDecl(Tag->getDecl()); } } } @@ -5521,7 +5525,7 @@ void ASTReader::InitializeContext() { Error("Invalid jmp_buf type in AST file"); return; } - Context.setjmp_bufDecl(Tag->getOriginalDecl()); + Context.setjmp_bufDecl(Tag->getDecl()); } } } @@ -5539,7 +5543,7 @@ void ASTReader::InitializeContext() { else { const TagType *Tag = Sigjmp_bufType->getAs<TagType>(); assert(Tag && "Invalid sigjmp_buf type in AST file"); - Context.setsigjmp_bufDecl(Tag->getOriginalDecl()); + Context.setsigjmp_bufDecl(Tag->getDecl()); } } } @@ -5574,7 +5578,7 @@ void ASTReader::InitializeContext() { else { const TagType *Tag = Ucontext_tType->getAs<TagType>(); assert(Tag && "Invalid ucontext_t type in AST file"); - Context.setucontext_tDecl(Tag->getOriginalDecl()); + Context.setucontext_tDecl(Tag->getDecl()); } } } @@ -11684,7 +11688,10 @@ void OMPClauseReader::VisitOMPDetachClause(OMPDetachClause *C) { C->setLParenLoc(Record.readSourceLocation()); } -void OMPClauseReader::VisitOMPNowaitClause(OMPNowaitClause *) {} +void OMPClauseReader::VisitOMPNowaitClause(OMPNowaitClause *C) { + C->setCondition(Record.readSubExpr()); + C->setLParenLoc(Record.readSourceLocation()); +} void OMPClauseReader::VisitOMPUntiedClause(OMPUntiedClause *) {} @@ -13006,7 +13013,7 @@ OpenACCClause *ASTRecordReader::readOpenACCClause() { SourceLocation LParenLoc = readSourceLocation(); OpenACCReductionOperator Op = readEnum<OpenACCReductionOperator>(); llvm::SmallVector<Expr *> VarList = readOpenACCVarList(); - llvm::SmallVector<OpenACCReductionRecipe> RecipeList; + llvm::SmallVector<OpenACCReductionRecipeWithStorage> RecipeList; for (unsigned I = 0; I < VarList.size(); ++I) { VarDecl *Recipe = readDeclAs<VarDecl>(); diff --git a/clang/lib/Serialization/ASTWriter.cpp b/clang/lib/Serialization/ASTWriter.cpp index 82ccde8..377e396 100644 --- a/clang/lib/Serialization/ASTWriter.cpp +++ b/clang/lib/Serialization/ASTWriter.cpp @@ -7942,7 +7942,10 @@ void OMPClauseWriter::VisitOMPOrderedClause(OMPOrderedClause *C) { Record.AddSourceLocation(C->getLParenLoc()); } -void OMPClauseWriter::VisitOMPNowaitClause(OMPNowaitClause *) {} +void OMPClauseWriter::VisitOMPNowaitClause(OMPNowaitClause *C) { + Record.AddStmt(C->getCondition()); + Record.AddSourceLocation(C->getLParenLoc()); +} void OMPClauseWriter::VisitOMPUntiedClause(OMPUntiedClause *) {} diff --git a/clang/lib/Serialization/ModuleCache.cpp b/clang/lib/Serialization/ModuleCache.cpp index 9668727..9850956 100644 --- a/clang/lib/Serialization/ModuleCache.cpp +++ b/clang/lib/Serialization/ModuleCache.cpp @@ -115,8 +115,10 @@ public: } std::time_t getModuleTimestamp(StringRef ModuleFilename) override { + std::string TimestampFilename = + serialization::ModuleFile::getTimestampFilename(ModuleFilename); llvm::sys::fs::file_status Status; - if (llvm::sys::fs::status(ModuleFilename, Status) != std::error_code{}) + if (llvm::sys::fs::status(TimestampFilename, Status) != std::error_code{}) return 0; return llvm::sys::toTimeT(Status.getLastModificationTime()); } diff --git a/clang/lib/Serialization/TemplateArgumentHasher.cpp b/clang/lib/Serialization/TemplateArgumentHasher.cpp index 3e8ffea..353e8a2 100644 --- a/clang/lib/Serialization/TemplateArgumentHasher.cpp +++ b/clang/lib/Serialization/TemplateArgumentHasher.cpp @@ -358,7 +358,7 @@ public: AddQualType(T->getReplacementType()); } - void VisitTagType(const TagType *T) { AddDecl(T->getOriginalDecl()); } + void VisitTagType(const TagType *T) { AddDecl(T->getDecl()); } void VisitRecordType(const RecordType *T) { VisitTagType(T); } void VisitEnumType(const EnumType *T) { VisitTagType(T); } diff --git a/clang/lib/StaticAnalyzer/Checkers/CStringChecker.cpp b/clang/lib/StaticAnalyzer/Checkers/CStringChecker.cpp index 0ae784c..1444114 100644 --- a/clang/lib/StaticAnalyzer/Checkers/CStringChecker.cpp +++ b/clang/lib/StaticAnalyzer/Checkers/CStringChecker.cpp @@ -251,6 +251,8 @@ public: const Expr *Ex, const MemRegion *MR, bool hypothetical); + static const StringLiteral *getStringLiteralFromRegion(const MemRegion *MR); + SVal getCStringLength(CheckerContext &C, ProgramStateRef &state, const Expr *Ex, @@ -983,6 +985,21 @@ SVal CStringChecker::getCStringLengthForRegion(CheckerContext &C, return strLength; } +const StringLiteral * +CStringChecker::getStringLiteralFromRegion(const MemRegion *MR) { + switch (MR->getKind()) { + case MemRegion::StringRegionKind: + return cast<StringRegion>(MR)->getStringLiteral(); + case MemRegion::NonParamVarRegionKind: + if (const VarDecl *Decl = cast<NonParamVarRegion>(MR)->getDecl(); + Decl->getType().isConstQualified() && Decl->hasGlobalStorage()) + return dyn_cast_or_null<StringLiteral>(Decl->getInit()); + return nullptr; + default: + return nullptr; + } +} + SVal CStringChecker::getCStringLength(CheckerContext &C, ProgramStateRef &state, const Expr *Ex, SVal Buf, bool hypothetical) const { @@ -1013,30 +1030,19 @@ SVal CStringChecker::getCStringLength(CheckerContext &C, ProgramStateRef &state, // its length. For anything we can't figure out, just return UnknownVal. MR = MR->StripCasts(); - switch (MR->getKind()) { - case MemRegion::StringRegionKind: { - // Modifying the contents of string regions is undefined [C99 6.4.5p6], - // so we can assume that the byte length is the correct C string length. - SValBuilder &svalBuilder = C.getSValBuilder(); - QualType sizeTy = svalBuilder.getContext().getSizeType(); - const StringLiteral *strLit = cast<StringRegion>(MR)->getStringLiteral(); - return svalBuilder.makeIntVal(strLit->getLength(), sizeTy); - } - case MemRegion::NonParamVarRegionKind: { + if (const StringLiteral *StrLit = getStringLiteralFromRegion(MR)) { // If we have a global constant with a string literal initializer, // compute the initializer's length. - const VarDecl *Decl = cast<NonParamVarRegion>(MR)->getDecl(); - if (Decl->getType().isConstQualified() && Decl->hasGlobalStorage()) { - if (const Expr *Init = Decl->getInit()) { - if (auto *StrLit = dyn_cast<StringLiteral>(Init)) { - SValBuilder &SvalBuilder = C.getSValBuilder(); - QualType SizeTy = SvalBuilder.getContext().getSizeType(); - return SvalBuilder.makeIntVal(StrLit->getLength(), SizeTy); - } - } - } - [[fallthrough]]; + // Modifying the contents of string regions is undefined [C99 6.4.5p6], + // so we can assume that the byte length is the correct C string length. + // FIXME: Embedded null characters are not handled. + SValBuilder &SVB = C.getSValBuilder(); + return SVB.makeIntVal(StrLit->getLength(), SVB.getContext().getSizeType()); } + + switch (MR->getKind()) { + case MemRegion::StringRegionKind: + case MemRegion::NonParamVarRegionKind: case MemRegion::SymbolicRegionKind: case MemRegion::AllocaRegionKind: case MemRegion::ParamVarRegionKind: @@ -1046,10 +1052,28 @@ SVal CStringChecker::getCStringLength(CheckerContext &C, ProgramStateRef &state, case MemRegion::CompoundLiteralRegionKind: // FIXME: Can we track this? Is it necessary? return UnknownVal(); - case MemRegion::ElementRegionKind: - // FIXME: How can we handle this? It's not good enough to subtract the - // offset from the base string length; consider "123\x00567" and &a[5]. + case MemRegion::ElementRegionKind: { + // If an offset into the string literal is used, use the original length + // minus the offset. + // FIXME: Embedded null characters are not handled. + const ElementRegion *ER = cast<ElementRegion>(MR); + const SubRegion *SuperReg = + cast<SubRegion>(ER->getSuperRegion()->StripCasts()); + const StringLiteral *StrLit = getStringLiteralFromRegion(SuperReg); + if (!StrLit) + return UnknownVal(); + SValBuilder &SVB = C.getSValBuilder(); + NonLoc Idx = ER->getIndex(); + QualType SizeTy = SVB.getContext().getSizeType(); + NonLoc LengthVal = + SVB.makeIntVal(StrLit->getLength(), SizeTy).castAs<NonLoc>(); + if (state->assume(SVB.evalBinOpNN(state, BO_LE, Idx, LengthVal, + SVB.getConditionType()) + .castAs<DefinedOrUnknownSVal>(), + true)) + return SVB.evalBinOp(state, BO_Sub, LengthVal, Idx, SizeTy); return UnknownVal(); + } default: // Other regions (mostly non-data) can't have a reliable C string length. // In this case, an error is emitted and UndefinedVal is returned. @@ -1074,6 +1098,7 @@ SVal CStringChecker::getCStringLength(CheckerContext &C, ProgramStateRef &state, const StringLiteral *CStringChecker::getCStringLiteral(CheckerContext &C, ProgramStateRef &state, const Expr *expr, SVal val) const { + // FIXME: use getStringLiteralFromRegion (and remove unused parameters)? // Get the memory region pointed to by the val. const MemRegion *bufRegion = val.getAsRegion(); diff --git a/clang/lib/StaticAnalyzer/Checkers/CallAndMessageChecker.cpp b/clang/lib/StaticAnalyzer/Checkers/CallAndMessageChecker.cpp index b304350..7cc146e 100644 --- a/clang/lib/StaticAnalyzer/Checkers/CallAndMessageChecker.cpp +++ b/clang/lib/StaticAnalyzer/Checkers/CallAndMessageChecker.cpp @@ -246,7 +246,7 @@ public: bool Find(const TypedValueRegion *R) { QualType T = R->getValueType(); if (const RecordType *RT = T->getAsStructureType()) { - const RecordDecl *RD = RT->getOriginalDecl()->getDefinition(); + const RecordDecl *RD = RT->getDecl()->getDefinition(); assert(RD && "Referred record has no definition"); for (const auto *I : RD->fields()) { if (I->isUnnamedBitField()) diff --git a/clang/lib/StaticAnalyzer/Checkers/CheckSecuritySyntaxOnly.cpp b/clang/lib/StaticAnalyzer/Checkers/CheckSecuritySyntaxOnly.cpp index 17af1ae..5e75c1c 100644 --- a/clang/lib/StaticAnalyzer/Checkers/CheckSecuritySyntaxOnly.cpp +++ b/clang/lib/StaticAnalyzer/Checkers/CheckSecuritySyntaxOnly.cpp @@ -154,15 +154,15 @@ void WalkAST::VisitCallExpr(CallExpr *CE) { .Case("mkstemp", &WalkAST::checkCall_mkstemp) .Case("mkdtemp", &WalkAST::checkCall_mkstemp) .Case("mkstemps", &WalkAST::checkCall_mkstemp) - .Cases("strcpy", "__strcpy_chk", &WalkAST::checkCall_strcpy) - .Cases("strcat", "__strcat_chk", &WalkAST::checkCall_strcat) - .Cases("sprintf", "vsprintf", "scanf", "wscanf", "fscanf", "fwscanf", - "vscanf", "vwscanf", "vfscanf", "vfwscanf", + .Cases({"strcpy", "__strcpy_chk"}, &WalkAST::checkCall_strcpy) + .Cases({"strcat", "__strcat_chk"}, &WalkAST::checkCall_strcat) + .Cases({"sprintf", "vsprintf", "scanf", "wscanf", "fscanf", "fwscanf", + "vscanf", "vwscanf", "vfscanf", "vfwscanf"}, &WalkAST::checkDeprecatedOrUnsafeBufferHandling) - .Cases("sscanf", "swscanf", "vsscanf", "vswscanf", "swprintf", - "snprintf", "vswprintf", "vsnprintf", "memcpy", "memmove", + .Cases({"sscanf", "swscanf", "vsscanf", "vswscanf", "swprintf", + "snprintf", "vswprintf", "vsnprintf", "memcpy", "memmove"}, &WalkAST::checkDeprecatedOrUnsafeBufferHandling) - .Cases("strncpy", "strncat", "memset", "fprintf", + .Cases({"strncpy", "strncat", "memset", "fprintf"}, &WalkAST::checkDeprecatedOrUnsafeBufferHandling) .Case("drand48", &WalkAST::checkCall_rand) .Case("erand48", &WalkAST::checkCall_rand) @@ -766,12 +766,14 @@ void WalkAST::checkDeprecatedOrUnsafeBufferHandling(const CallExpr *CE, int ArgIndex = llvm::StringSwitch<int>(Name) - .Cases("scanf", "wscanf", "vscanf", "vwscanf", 0) - .Cases("fscanf", "fwscanf", "vfscanf", "vfwscanf", "sscanf", - "swscanf", "vsscanf", "vswscanf", 1) - .Cases("sprintf", "vsprintf", "fprintf", 1) - .Cases("swprintf", "snprintf", "vswprintf", "vsnprintf", "memcpy", - "memmove", "memset", "strncpy", "strncat", DEPR_ONLY) + .Cases({"scanf", "wscanf", "vscanf", "vwscanf"}, 0) + .Cases({"fscanf", "fwscanf", "vfscanf", "vfwscanf", "sscanf", + "swscanf", "vsscanf", "vswscanf"}, + 1) + .Cases({"sprintf", "vsprintf", "fprintf"}, 1) + .Cases({"swprintf", "snprintf", "vswprintf", "vsnprintf", "memcpy", + "memmove", "memset", "strncpy", "strncat"}, + DEPR_ONLY) .Default(UNKNOWN_CALL); assert(ArgIndex != UNKNOWN_CALL && "Unsupported function"); diff --git a/clang/lib/StaticAnalyzer/Checkers/DereferenceChecker.cpp b/clang/lib/StaticAnalyzer/Checkers/DereferenceChecker.cpp index 395d724..37f5ec3 100644 --- a/clang/lib/StaticAnalyzer/Checkers/DereferenceChecker.cpp +++ b/clang/lib/StaticAnalyzer/Checkers/DereferenceChecker.cpp @@ -19,6 +19,7 @@ #include "clang/StaticAnalyzer/Core/CheckerManager.h" #include "clang/StaticAnalyzer/Core/PathSensitive/CheckerContext.h" #include "clang/StaticAnalyzer/Core/PathSensitive/CheckerHelpers.h" +#include "llvm/Support/FormatVariadic.h" #include "llvm/Support/raw_ostream.h" using namespace clang; @@ -39,9 +40,10 @@ public: class DereferenceChecker : public CheckerFamily<check::Location, check::Bind, + check::PreStmt<BinaryOperator>, EventDispatcher<ImplicitNullDerefEvent>> { - void reportBug(const DerefBugType &BT, ProgramStateRef State, const Stmt *S, - CheckerContext &C) const; + void reportDerefBug(const DerefBugType &BT, ProgramStateRef State, + const Stmt *S, CheckerContext &C) const; bool suppressReport(CheckerContext &C, const Expr *E) const; @@ -50,6 +52,7 @@ public: CheckerContext &C) const; void checkBind(SVal L, SVal V, const Stmt *S, bool AtDeclInit, CheckerContext &C) const; + void checkPreStmt(const BinaryOperator *Op, CheckerContext &C) const; static void AddDerefSource(raw_ostream &os, SmallVectorImpl<SourceRange> &Ranges, @@ -57,7 +60,7 @@ public: const LocationContext *LCtx, bool loadedFrom = false); - CheckerFrontend NullDerefChecker, FixedDerefChecker; + CheckerFrontend NullDerefChecker, FixedDerefChecker, NullPointerArithmChecker; const DerefBugType NullBug{&NullDerefChecker, "Dereference of null pointer", "a null pointer dereference", "a dereference of a null pointer"}; @@ -72,9 +75,22 @@ public: const DerefBugType FixedAddressBug{&FixedDerefChecker, "Dereference of a fixed address", "a dereference of a fixed address"}; + const BugType NullPointerArithmBug{ + &NullPointerArithmChecker, + "Possibly undefined arithmetic operation involving a null pointer"}; StringRef getDebugTag() const override { return "DereferenceChecker"; } }; + +struct ValueDescStr { + SmallVectorImpl<SourceRange> &Ranges; + const Expr *Ex; + const ProgramState *State; + const LocationContext *LCtx; + bool IsPointer; + ConditionTruthVal IsNull; +}; + } // end anonymous namespace void @@ -173,9 +189,9 @@ static bool isDeclRefExprToReference(const Expr *E) { return false; } -void DereferenceChecker::reportBug(const DerefBugType &BT, - ProgramStateRef State, const Stmt *S, - CheckerContext &C) const { +void DereferenceChecker::reportDerefBug(const DerefBugType &BT, + ProgramStateRef State, const Stmt *S, + CheckerContext &C) const { if (&BT == &FixedAddressBug) { if (!FixedDerefChecker.isEnabled()) // Deliberately don't add a sink node if check is disabled. @@ -249,9 +265,8 @@ void DereferenceChecker::reportBug(const DerefBugType &BT, bugreporter::trackExpressionValue(N, bugreporter::getDerefExpr(S), *BR); - for (SmallVectorImpl<SourceRange>::iterator - I = Ranges.begin(), E = Ranges.end(); I!=E; ++I) - BR->addRange(*I); + for (const auto &R : Ranges) + BR->addRange(R); C.emitReport(std::move(BR)); } @@ -262,7 +277,7 @@ void DereferenceChecker::checkLocation(SVal l, bool isLoad, const Stmt* S, if (l.isUndef()) { const Expr *DerefExpr = getDereferenceExpr(S); if (!suppressReport(C, DerefExpr)) - reportBug(UndefBug, C.getState(), DerefExpr, C); + reportDerefBug(UndefBug, C.getState(), DerefExpr, C); return; } @@ -283,7 +298,7 @@ void DereferenceChecker::checkLocation(SVal l, bool isLoad, const Stmt* S, // we call an "explicit" null dereference. const Expr *expr = getDereferenceExpr(S); if (!suppressReport(C, expr)) { - reportBug(NullBug, nullState, expr, C); + reportDerefBug(NullBug, nullState, expr, C); return; } } @@ -301,7 +316,7 @@ void DereferenceChecker::checkLocation(SVal l, bool isLoad, const Stmt* S, if (location.isConstant()) { const Expr *DerefExpr = getDereferenceExpr(S, isLoad); if (!suppressReport(C, DerefExpr)) - reportBug(FixedAddressBug, notNullState, DerefExpr, C); + reportDerefBug(FixedAddressBug, notNullState, DerefExpr, C); return; } @@ -317,7 +332,7 @@ void DereferenceChecker::checkBind(SVal L, SVal V, const Stmt *S, // One should never write to label addresses. if (auto Label = L.getAs<loc::GotoLabel>()) { - reportBug(LabelBug, C.getState(), S, C); + reportDerefBug(LabelBug, C.getState(), S, C); return; } @@ -338,7 +353,7 @@ void DereferenceChecker::checkBind(SVal L, SVal V, const Stmt *S, if (!StNonNull) { const Expr *expr = getDereferenceExpr(S, /*IsBind=*/true); if (!suppressReport(C, expr)) { - reportBug(NullBug, StNull, expr, C); + reportDerefBug(NullBug, StNull, expr, C); return; } } @@ -356,7 +371,7 @@ void DereferenceChecker::checkBind(SVal L, SVal V, const Stmt *S, if (V.isConstant()) { const Expr *DerefExpr = getDereferenceExpr(S, true); if (!suppressReport(C, DerefExpr)) - reportBug(FixedAddressBug, State, DerefExpr, C); + reportDerefBug(FixedAddressBug, State, DerefExpr, C); return; } @@ -379,6 +394,96 @@ void DereferenceChecker::checkBind(SVal L, SVal V, const Stmt *S, C.addTransition(State, this); } +namespace llvm { +template <> struct format_provider<ValueDescStr> { + static void format(const ValueDescStr &V, raw_ostream &Stream, + StringRef Style) { + static const char *ValueStr[2][3] = { + {"zero", "nonzero integer value", "probably nonzero integer value"}, + {"null pointer", "non-null pointer", "probably non-null pointer"}, + }; + Stream + << ValueStr[V.IsPointer][V.IsNull.isConstrainedTrue() + ? 0 + : (V.IsNull.isConstrainedFalse() ? 1 : 2)]; + DereferenceChecker::AddDerefSource(Stream, V.Ranges, V.Ex, V.State, V.LCtx, + false); + } +}; +} // namespace llvm + +void DereferenceChecker::checkPreStmt(const BinaryOperator *Op, + CheckerContext &C) const { + if (!Op->isAdditiveOp() || !NullPointerArithmChecker.isEnabled()) + return; + const Expr *E1 = Op->getLHS(); + const Expr *E2 = Op->getRHS(); + QualType T1 = E1->getType().getCanonicalType(); + QualType T2 = E2->getType().getCanonicalType(); + bool T1IsPointer = T1->isPointerType(); + bool T2IsPointer = T2->isPointerType(); + if (T1->isIntegerType() && T2->isIntegerType()) + return; + if (!T1IsPointer && !T1->isIntegerType() && !T2IsPointer && + !T2->isIntegerType()) + return; + + ProgramStateRef State = C.getState(); + ConditionTruthVal V1IsNull = State->isNull(C.getSVal(E1)); + ConditionTruthVal V2IsNull = State->isNull(C.getSVal(E2)); + bool IsConstrained = true; + + // Check cases 'NULL + x' and 'NULL - x' + if (T1IsPointer && !T2IsPointer) { + if (!V1IsNull.isConstrainedTrue() || V2IsNull.isConstrainedTrue()) + return; + IsConstrained = V2IsNull.isConstrainedFalse(); + } + + // Check case 'x + NULL' + if (!T1IsPointer && T2IsPointer) { + if (V1IsNull.isConstrainedTrue() || !V2IsNull.isConstrainedTrue()) + return; + IsConstrained = V1IsNull.isConstrainedFalse(); + } + + // Check case 'NULL - p' or 'p - NULL' + if (T1IsPointer && T2IsPointer) { + if (!V1IsNull.isConstrainedTrue() && !V2IsNull.isConstrainedTrue()) + return; + if (V1IsNull.isConstrainedTrue() && V2IsNull.isConstrainedTrue()) + return; + IsConstrained = + V1IsNull.isConstrainedFalse() || V2IsNull.isConstrainedFalse(); + } + + SmallVector<SourceRange, 2> Ranges; + const char *OpcodeStr = + Op->getOpcode() == BO_Add ? "Addition" : "Subtraction"; + const char *ResultStr = IsConstrained ? "results" : "may result"; + ValueDescStr DerefArg1{ + Ranges, E1, State.get(), C.getLocationContext(), T1IsPointer, V1IsNull}; + ValueDescStr DerefArg2{ + Ranges, E2, State.get(), C.getLocationContext(), T2IsPointer, V2IsNull}; + std::string Msg = + llvm::formatv("{0} of a {1} and a {2} {3} in undefined behavior", + OpcodeStr, DerefArg1, DerefArg2, ResultStr); + + ExplodedNode *N = C.generateErrorNode(State); + if (!N) + return; + auto BR = + std::make_unique<PathSensitiveBugReport>(NullPointerArithmBug, Msg, N); + if (V1IsNull.isConstrainedTrue()) + bugreporter::trackExpressionValue(N, E1, *BR); + if (V2IsNull.isConstrainedTrue()) + bugreporter::trackExpressionValue(N, E2, *BR); + for (const auto &R : Ranges) + BR->addRange(R); + + C.emitReport(std::move(BR)); +} + void ento::registerNullDereferenceChecker(CheckerManager &Mgr) { Mgr.getChecker<DereferenceChecker>()->NullDerefChecker.enable(Mgr); } @@ -395,3 +500,11 @@ bool ento::shouldRegisterFixedAddressDereferenceChecker( const CheckerManager &) { return true; } + +void ento::registerNullPointerArithmChecker(CheckerManager &Mgr) { + Mgr.getChecker<DereferenceChecker>()->NullPointerArithmChecker.enable(Mgr); +} + +bool ento::shouldRegisterNullPointerArithmChecker(const CheckerManager &) { + return true; +} diff --git a/clang/lib/StaticAnalyzer/Checkers/NonNullParamChecker.cpp b/clang/lib/StaticAnalyzer/Checkers/NonNullParamChecker.cpp index b1a7cd7..bc67391 100644 --- a/clang/lib/StaticAnalyzer/Checkers/NonNullParamChecker.cpp +++ b/clang/lib/StaticAnalyzer/Checkers/NonNullParamChecker.cpp @@ -148,9 +148,8 @@ void NonNullParamChecker::checkPreCall(const CallEvent &Call, QualType T = ArgE->getType(); const RecordType *UT = T->getAsUnionType(); - if (!UT || !UT->getOriginalDecl() - ->getMostRecentDecl() - ->hasAttr<TransparentUnionAttr>()) + if (!UT || + !UT->getDecl()->getMostRecentDecl()->hasAttr<TransparentUnionAttr>()) continue; auto CSV = DV->getAs<nonloc::CompoundVal>(); diff --git a/clang/lib/StaticAnalyzer/Checkers/WebKit/ASTUtils.cpp b/clang/lib/StaticAnalyzer/Checkers/WebKit/ASTUtils.cpp index 66cfccb..84adbf3 100644 --- a/clang/lib/StaticAnalyzer/Checkers/WebKit/ASTUtils.cpp +++ b/clang/lib/StaticAnalyzer/Checkers/WebKit/ASTUtils.cpp @@ -26,6 +26,7 @@ bool tryToFindPtrOrigin( const Expr *E, bool StopAtFirstRefCountedObj, std::function<bool(const clang::CXXRecordDecl *)> isSafePtr, std::function<bool(const clang::QualType)> isSafePtrType, + std::function<bool(const clang::Decl *)> isSafeGlobalDecl, std::function<bool(const clang::Expr *, bool)> callback) { while (E) { if (auto *DRE = dyn_cast<DeclRefExpr>(E)) { @@ -34,6 +35,8 @@ bool tryToFindPtrOrigin( auto IsImmortal = safeGetName(VD) == "NSApp"; if (VD->hasGlobalStorage() && (IsImmortal || QT.isConstQualified())) return callback(E, true); + if (VD->hasGlobalStorage() && isSafeGlobalDecl(VD)) + return callback(E, true); } } if (auto *tempExpr = dyn_cast<MaterializeTemporaryExpr>(E)) { @@ -71,9 +74,11 @@ bool tryToFindPtrOrigin( } if (auto *Expr = dyn_cast<ConditionalOperator>(E)) { return tryToFindPtrOrigin(Expr->getTrueExpr(), StopAtFirstRefCountedObj, - isSafePtr, isSafePtrType, callback) && + isSafePtr, isSafePtrType, isSafeGlobalDecl, + callback) && tryToFindPtrOrigin(Expr->getFalseExpr(), StopAtFirstRefCountedObj, - isSafePtr, isSafePtrType, callback); + isSafePtr, isSafePtrType, isSafeGlobalDecl, + callback); } if (auto *cast = dyn_cast<CastExpr>(E)) { if (StopAtFirstRefCountedObj) { @@ -93,7 +98,8 @@ bool tryToFindPtrOrigin( if (auto *call = dyn_cast<CallExpr>(E)) { if (auto *Callee = call->getCalleeDecl()) { if (Callee->hasAttr<CFReturnsRetainedAttr>() || - Callee->hasAttr<NSReturnsRetainedAttr>()) { + Callee->hasAttr<NSReturnsRetainedAttr>() || + Callee->hasAttr<NSReturnsAutoreleasedAttr>()) { return callback(E, true); } } @@ -158,13 +164,23 @@ bool tryToFindPtrOrigin( auto Name = safeGetName(callee); if (Name == "__builtin___CFStringMakeConstantString" || - Name == "NSClassFromString") + Name == "NSStringFromSelector" || Name == "NSSelectorFromString" || + Name == "NSStringFromClass" || Name == "NSClassFromString" || + Name == "NSStringFromProtocol" || Name == "NSProtocolFromString") return callback(E, true); } else if (auto *CalleeE = call->getCallee()) { if (auto *E = dyn_cast<DeclRefExpr>(CalleeE->IgnoreParenCasts())) { if (isSingleton(E->getFoundDecl())) return callback(E, true); } + + if (auto *MemberExpr = dyn_cast<CXXDependentScopeMemberExpr>(CalleeE)) { + auto *Base = MemberExpr->getBase(); + auto MemberName = MemberExpr->getMember().getAsString(); + bool IsGetter = MemberName == "get" || MemberName == "ptr"; + if (Base && isSafePtrType(Base->getType()) && IsGetter) + return callback(E, true); + } } // Sometimes, canonical type erroneously turns Ref<T> into T. @@ -176,7 +192,7 @@ bool tryToFindPtrOrigin( if (auto *Subst = dyn_cast<SubstTemplateTypeParmType>(RetType)) { if (auto *SubstType = Subst->desugar().getTypePtr()) { if (auto *RD = dyn_cast<RecordType>(SubstType)) { - if (auto *CXX = dyn_cast<CXXRecordDecl>(RD->getOriginalDecl())) + if (auto *CXX = dyn_cast<CXXRecordDecl>(RD->getDecl())) if (isSafePtr(CXX)) return callback(E, true); } @@ -196,6 +212,8 @@ bool tryToFindPtrOrigin( !Selector.getNumArgs()) return callback(E, true); } + if (auto *ObjCProtocol = dyn_cast<ObjCProtocolExpr>(E)) + return callback(ObjCProtocol, true); if (auto *ObjCDict = dyn_cast<ObjCDictionaryLiteral>(E)) return callback(ObjCDict, true); if (auto *ObjCArray = dyn_cast<ObjCArrayLiteral>(E)) diff --git a/clang/lib/StaticAnalyzer/Checkers/WebKit/ASTUtils.h b/clang/lib/StaticAnalyzer/Checkers/WebKit/ASTUtils.h index 3a009d6..9fff456 100644 --- a/clang/lib/StaticAnalyzer/Checkers/WebKit/ASTUtils.h +++ b/clang/lib/StaticAnalyzer/Checkers/WebKit/ASTUtils.h @@ -56,6 +56,7 @@ bool tryToFindPtrOrigin( const clang::Expr *E, bool StopAtFirstRefCountedObj, std::function<bool(const clang::CXXRecordDecl *)> isSafePtr, std::function<bool(const clang::QualType)> isSafePtrType, + std::function<bool(const clang::Decl *)> isSafeGlobalDecl, std::function<bool(const clang::Expr *, bool)> callback); /// For \p E referring to a ref-countable/-counted pointer/reference we return diff --git a/clang/lib/StaticAnalyzer/Checkers/WebKit/ForwardDeclChecker.cpp b/clang/lib/StaticAnalyzer/Checkers/WebKit/ForwardDeclChecker.cpp index d8539ea..1d4e6dd 100644 --- a/clang/lib/StaticAnalyzer/Checkers/WebKit/ForwardDeclChecker.cpp +++ b/clang/lib/StaticAnalyzer/Checkers/WebKit/ForwardDeclChecker.cpp @@ -263,18 +263,43 @@ public: void visitCallArg(const Expr *Arg, const ParmVarDecl *Param, const Decl *DeclWithIssue) const { auto *ArgExpr = Arg->IgnoreParenCasts(); - if (auto *InnerCE = dyn_cast<CallExpr>(Arg)) { - auto *InnerCallee = InnerCE->getDirectCallee(); - if (InnerCallee && InnerCallee->isInStdNamespace() && - safeGetName(InnerCallee) == "move" && InnerCE->getNumArgs() == 1) { - ArgExpr = InnerCE->getArg(0); - if (ArgExpr) - ArgExpr = ArgExpr->IgnoreParenCasts(); + while (ArgExpr) { + ArgExpr = ArgExpr->IgnoreParenCasts(); + if (auto *InnerCE = dyn_cast<CallExpr>(ArgExpr)) { + auto *InnerCallee = InnerCE->getDirectCallee(); + if (InnerCallee && InnerCallee->isInStdNamespace() && + safeGetName(InnerCallee) == "move" && InnerCE->getNumArgs() == 1) { + ArgExpr = InnerCE->getArg(0); + continue; + } + } + if (auto *UO = dyn_cast<UnaryOperator>(ArgExpr)) { + auto OpCode = UO->getOpcode(); + if (OpCode == UO_Deref || OpCode == UO_AddrOf) { + ArgExpr = UO->getSubExpr(); + continue; + } } + break; } + + if (auto *MemberCallExpr = dyn_cast<CXXMemberCallExpr>(ArgExpr)) { + if (isOwnerPtrType(MemberCallExpr->getObjectType())) + return; + } + + if (auto *OpCE = dyn_cast<CXXOperatorCallExpr>(ArgExpr)) { + auto *Method = dyn_cast_or_null<CXXMethodDecl>(OpCE->getDirectCallee()); + if (Method && isOwnerPtr(safeGetName(Method->getParent()))) { + if (OpCE->getOperator() == OO_Star && OpCE->getNumArgs() == 1) + return; + } + } + if (isNullPtr(ArgExpr) || isa<IntegerLiteral>(ArgExpr) || isa<CXXDefaultArgExpr>(ArgExpr)) return; + if (auto *DRE = dyn_cast<DeclRefExpr>(ArgExpr)) { if (auto *ValDecl = DRE->getDecl()) { if (isa<ParmVarDecl>(ValDecl)) diff --git a/clang/lib/StaticAnalyzer/Checkers/WebKit/PtrTypesSemantics.cpp b/clang/lib/StaticAnalyzer/Checkers/WebKit/PtrTypesSemantics.cpp index e5c74bb..d3d1f13 100644 --- a/clang/lib/StaticAnalyzer/Checkers/WebKit/PtrTypesSemantics.cpp +++ b/clang/lib/StaticAnalyzer/Checkers/WebKit/PtrTypesSemantics.cpp @@ -138,6 +138,11 @@ bool isCheckedPtr(const std::string &Name) { return Name == "CheckedPtr" || Name == "CheckedRef"; } +bool isOwnerPtr(const std::string &Name) { + return isRefType(Name) || isCheckedPtr(Name) || Name == "unique_ptr" || + Name == "UniqueRef" || Name == "LazyUniqueRef"; +} + bool isSmartPtrClass(const std::string &Name) { return isRefType(Name) || isCheckedPtr(Name) || isRetainPtrOrOSPtr(Name) || Name == "WeakPtr" || Name == "WeakPtrFactory" || @@ -206,10 +211,7 @@ bool isRetainPtrOrOSPtrType(const clang::QualType T) { } bool isOwnerPtrType(const clang::QualType T) { - return isPtrOfType(T, [](auto Name) { - return isRefType(Name) || isCheckedPtr(Name) || Name == "unique_ptr" || - Name == "UniqueRef" || Name == "LazyUniqueRef"; - }); + return isPtrOfType(T, [](auto Name) { return isOwnerPtr(Name); }); } std::optional<bool> isUncounted(const QualType T) { @@ -255,7 +257,7 @@ void RetainTypeChecker::visitTypedef(const TypedefDecl *TD) { return; } - for (auto *Redecl : RT->getOriginalDecl()->getMostRecentDecl()->redecls()) { + for (auto *Redecl : RT->getDecl()->getMostRecentDecl()->redecls()) { if (Redecl->getAttr<ObjCBridgeAttr>() || Redecl->getAttr<ObjCBridgeMutableAttr>()) { CFPointees.insert(RT); @@ -296,7 +298,7 @@ std::optional<bool> isUnretained(const QualType T, bool IsARCEnabled) { auto *Record = PointeeType->getAsStructureType(); if (!Record) return false; - auto *Decl = Record->getOriginalDecl(); + auto *Decl = Record->getDecl(); if (!Decl) return false; auto TypeName = Decl->getName(); diff --git a/clang/lib/StaticAnalyzer/Checkers/WebKit/PtrTypesSemantics.h b/clang/lib/StaticAnalyzer/Checkers/WebKit/PtrTypesSemantics.h index 8300a6c..12e2e2d 100644 --- a/clang/lib/StaticAnalyzer/Checkers/WebKit/PtrTypesSemantics.h +++ b/clang/lib/StaticAnalyzer/Checkers/WebKit/PtrTypesSemantics.h @@ -143,6 +143,10 @@ bool isCheckedPtr(const std::string &Name); /// \returns true if \p Name is RetainPtr or its variant, false if not. bool isRetainPtrOrOSPtr(const std::string &Name); +/// \returns true if \p Name is an owning smar pointer such as Ref, CheckedPtr, +/// and unique_ptr. +bool isOwnerPtr(const std::string &Name); + /// \returns true if \p Name is a smart pointer type name, false if not. bool isSmartPtrClass(const std::string &Name); diff --git a/clang/lib/StaticAnalyzer/Checkers/WebKit/RawPtrRefCallArgsChecker.cpp b/clang/lib/StaticAnalyzer/Checkers/WebKit/RawPtrRefCallArgsChecker.cpp index 9585ceb..791e709 100644 --- a/clang/lib/StaticAnalyzer/Checkers/WebKit/RawPtrRefCallArgsChecker.cpp +++ b/clang/lib/StaticAnalyzer/Checkers/WebKit/RawPtrRefCallArgsChecker.cpp @@ -29,12 +29,12 @@ namespace { class RawPtrRefCallArgsChecker : public Checker<check::ASTDecl<TranslationUnitDecl>> { BugType Bug; - mutable BugReporter *BR; TrivialFunctionAnalysis TFA; EnsureFunctionAnalysis EFA; protected: + mutable BugReporter *BR; mutable std::optional<RetainTypeChecker> RTC; public: @@ -46,6 +46,7 @@ public: virtual bool isSafePtr(const CXXRecordDecl *Record) const = 0; virtual bool isSafePtrType(const QualType type) const = 0; virtual bool isSafeExpr(const Expr *) const { return false; } + virtual bool isSafeDecl(const Decl *) const { return false; } virtual const char *ptrKind() const = 0; void checkASTDecl(const TranslationUnitDecl *TUD, AnalysisManager &MGR, @@ -214,6 +215,7 @@ public: Arg, /*StopAtFirstRefCountedObj=*/true, [&](const clang::CXXRecordDecl *Record) { return isSafePtr(Record); }, [&](const clang::QualType T) { return isSafePtrType(T); }, + [&](const clang::Decl *D) { return isSafeDecl(D); }, [&](const clang::Expr *ArgOrigin, bool IsSafe) { if (IsSafe) return true; @@ -479,6 +481,11 @@ public: isa<ObjCMessageExpr>(E); } + bool isSafeDecl(const Decl *D) const final { + // Treat NS/CF globals in system header as immortal. + return BR->getSourceManager().isInSystemHeader(D->getLocation()); + } + const char *ptrKind() const final { return "unretained"; } }; diff --git a/clang/lib/StaticAnalyzer/Checkers/WebKit/RawPtrRefLocalVarsChecker.cpp b/clang/lib/StaticAnalyzer/Checkers/WebKit/RawPtrRefLocalVarsChecker.cpp index dd9701f..c13df479 100644 --- a/clang/lib/StaticAnalyzer/Checkers/WebKit/RawPtrRefLocalVarsChecker.cpp +++ b/clang/lib/StaticAnalyzer/Checkers/WebKit/RawPtrRefLocalVarsChecker.cpp @@ -166,10 +166,10 @@ bool isGuardedScopeEmbeddedInGuardianScope(const VarDecl *Guarded, class RawPtrRefLocalVarsChecker : public Checker<check::ASTDecl<TranslationUnitDecl>> { BugType Bug; - mutable BugReporter *BR; EnsureFunctionAnalysis EFA; protected: + mutable BugReporter *BR; mutable std::optional<RetainTypeChecker> RTC; public: @@ -180,6 +180,7 @@ public: virtual bool isSafePtr(const CXXRecordDecl *) const = 0; virtual bool isSafePtrType(const QualType) const = 0; virtual bool isSafeExpr(const Expr *) const { return false; } + virtual bool isSafeDecl(const Decl *) const { return false; } virtual const char *ptrKind() const = 0; void checkASTDecl(const TranslationUnitDecl *TUD, AnalysisManager &MGR, @@ -288,6 +289,7 @@ public: return isSafePtr(Record); }, [&](const clang::QualType Type) { return isSafePtrType(Type); }, + [&](const clang::Decl *D) { return isSafeDecl(D); }, [&](const clang::Expr *InitArgOrigin, bool IsSafe) { if (!InitArgOrigin || IsSafe) return true; @@ -443,6 +445,10 @@ public: return ento::cocoa::isCocoaObjectRef(E->getType()) && isa<ObjCMessageExpr>(E); } + bool isSafeDecl(const Decl *D) const final { + // Treat NS/CF globals in system header as immortal. + return BR->getSourceManager().isInSystemHeader(D->getLocation()); + } const char *ptrKind() const final { return "unretained"; } }; diff --git a/clang/lib/StaticAnalyzer/Checkers/WebKit/RefCntblBaseVirtualDtorChecker.cpp b/clang/lib/StaticAnalyzer/Checkers/WebKit/RefCntblBaseVirtualDtorChecker.cpp index 6f3a280..c6421f8 100644 --- a/clang/lib/StaticAnalyzer/Checkers/WebKit/RefCntblBaseVirtualDtorChecker.cpp +++ b/clang/lib/StaticAnalyzer/Checkers/WebKit/RefCntblBaseVirtualDtorChecker.cpp @@ -121,13 +121,13 @@ public: return true; } } else if (auto *RD = dyn_cast<RecordType>(PointeeType)) { - if (declaresSameEntity(RD->getOriginalDecl(), ClassDecl)) + if (declaresSameEntity(RD->getDecl(), ClassDecl)) return true; } else if (auto *ST = dyn_cast<SubstTemplateTypeParmType>(PointeeType)) { auto Type = ST->getReplacementType(); if (auto *RD = dyn_cast<RecordType>(Type)) { - if (declaresSameEntity(RD->getOriginalDecl(), ClassDecl)) + if (declaresSameEntity(RD->getDecl(), ClassDecl)) return true; } } diff --git a/clang/lib/StaticAnalyzer/Core/BasicValueFactory.cpp b/clang/lib/StaticAnalyzer/Core/BasicValueFactory.cpp index 02f34bc..c905ee6 100644 --- a/clang/lib/StaticAnalyzer/Core/BasicValueFactory.cpp +++ b/clang/lib/StaticAnalyzer/Core/BasicValueFactory.cpp @@ -173,7 +173,7 @@ const PointerToMemberData *BasicValueFactory::getPointerToMemberData( return D; } -LLVM_ATTRIBUTE_UNUSED static bool hasNoRepeatedElements( +[[maybe_unused]] static bool hasNoRepeatedElements( llvm::ImmutableList<const CXXBaseSpecifier *> BaseSpecList) { llvm::SmallPtrSet<QualType, 16> BaseSpecSeen; for (const CXXBaseSpecifier *BaseSpec : BaseSpecList) { diff --git a/clang/lib/StaticAnalyzer/Core/CallEvent.cpp b/clang/lib/StaticAnalyzer/Core/CallEvent.cpp index 06ba015..62460cc 100644 --- a/clang/lib/StaticAnalyzer/Core/CallEvent.cpp +++ b/clang/lib/StaticAnalyzer/Core/CallEvent.cpp @@ -89,7 +89,7 @@ static bool isCallback(QualType T) { T = T->getPointeeType(); if (const RecordType *RT = T->getAsStructureType()) { - const RecordDecl *RD = RT->getOriginalDecl()->getDefinitionOrSelf(); + const RecordDecl *RD = RT->getDecl()->getDefinitionOrSelf(); for (const auto *I : RD->fields()) { QualType FieldT = I->getType(); if (FieldT->isBlockPointerType() || FieldT->isFunctionPointerType()) @@ -391,9 +391,8 @@ bool CallEvent::isVariadic(const Decl *D) { static bool isTransparentUnion(QualType T) { const RecordType *UT = T->getAsUnionType(); - return UT && UT->getOriginalDecl() - ->getMostRecentDecl() - ->hasAttr<TransparentUnionAttr>(); + return UT && + UT->getDecl()->getMostRecentDecl()->hasAttr<TransparentUnionAttr>(); } // In some cases, symbolic cases should be transformed before we associate diff --git a/clang/lib/StaticAnalyzer/Core/EntryPointStats.cpp b/clang/lib/StaticAnalyzer/Core/EntryPointStats.cpp index abfb176..c207a7b 100644 --- a/clang/lib/StaticAnalyzer/Core/EntryPointStats.cpp +++ b/clang/lib/StaticAnalyzer/Core/EntryPointStats.cpp @@ -24,15 +24,21 @@ using namespace ento; namespace { struct Registry { + std::vector<UnsignedEPStat *> ExplicitlySetStats; + std::vector<UnsignedMaxEPStat *> MaxStats; std::vector<CounterEPStat *> CounterStats; - std::vector<UnsignedMaxEPStat *> UnsignedMaxStats; - std::vector<UnsignedEPStat *> UnsignedStats; bool IsLocked = false; struct Snapshot { const Decl *EntryPoint; - std::vector<unsigned> UnsignedStatValues; + // Explicitly set statistics may not have a value set, so they are separate + // from other unsigned statistics + std::vector<std::optional<unsigned>> ExplicitlySetStatValues; + // These are counting and maximizing statistics that initialize to 0, which + // is meaningful even if they are never updated, so their value is always + // present. + std::vector<unsigned> MaxOrCountStatValues; void dumpAsCSV(llvm::raw_ostream &OS) const; }; @@ -46,10 +52,16 @@ static llvm::ManagedStatic<Registry> StatsRegistry; namespace { template <typename Callback> void enumerateStatVectors(const Callback &Fn) { + // This order is important, it matches the order of the Snapshot fields: + // - ExplicitlySetStatValues + Fn(StatsRegistry->ExplicitlySetStats); + // - MaxOrCountStatValues + Fn(StatsRegistry->MaxStats); Fn(StatsRegistry->CounterStats); - Fn(StatsRegistry->UnsignedMaxStats); - Fn(StatsRegistry->UnsignedStats); } + +void clearSnapshots(void *) { StatsRegistry->Snapshots.clear(); } + } // namespace static void checkStatName(const EntryPointStat *M) { @@ -69,7 +81,8 @@ static void checkStatName(const EntryPointStat *M) { } } -void EntryPointStat::lockRegistry(llvm::StringRef CPPFileName) { +void EntryPointStat::lockRegistry(llvm::StringRef CPPFileName, + ASTContext &Ctx) { auto CmpByNames = [](const EntryPointStat *L, const EntryPointStat *R) { return L->name() < R->name(); }; @@ -80,6 +93,10 @@ void EntryPointStat::lockRegistry(llvm::StringRef CPPFileName) { StatsRegistry->IsLocked = true; llvm::raw_string_ostream OS(StatsRegistry->EscapedCPPFileName); llvm::printEscapedString(CPPFileName, OS); + // Make sure snapshots (that reference function Decl's) do not persist after + // the AST is destroyed. This is especially relevant in the context of unit + // tests that construct and destruct multiple ASTs in the same process. + Ctx.AddDeallocation(clearSnapshots, nullptr); } [[maybe_unused]] static bool isRegistered(llvm::StringLiteral Name) { @@ -101,30 +118,36 @@ UnsignedMaxEPStat::UnsignedMaxEPStat(llvm::StringLiteral Name) : EntryPointStat(Name) { assert(!StatsRegistry->IsLocked); assert(!isRegistered(Name)); - StatsRegistry->UnsignedMaxStats.push_back(this); + StatsRegistry->MaxStats.push_back(this); } UnsignedEPStat::UnsignedEPStat(llvm::StringLiteral Name) : EntryPointStat(Name) { assert(!StatsRegistry->IsLocked); assert(!isRegistered(Name)); - StatsRegistry->UnsignedStats.push_back(this); + StatsRegistry->ExplicitlySetStats.push_back(this); } -static std::vector<unsigned> consumeUnsignedStats() { - std::vector<unsigned> Result; - Result.reserve(StatsRegistry->CounterStats.size() + - StatsRegistry->UnsignedMaxStats.size() + - StatsRegistry->UnsignedStats.size()); - for (auto *M : StatsRegistry->CounterStats) { +static std::vector<std::optional<unsigned>> consumeExplicitlySetStats() { + std::vector<std::optional<unsigned>> Result; + Result.reserve(StatsRegistry->ExplicitlySetStats.size()); + for (auto *M : StatsRegistry->ExplicitlySetStats) { Result.push_back(M->value()); M->reset(); } - for (auto *M : StatsRegistry->UnsignedMaxStats) { + return Result; +} + +static std::vector<unsigned> consumeMaxAndCounterStats() { + std::vector<unsigned> Result; + Result.reserve(StatsRegistry->CounterStats.size() + + StatsRegistry->MaxStats.size()); + // Order is important, it must match the order in enumerateStatVectors + for (auto *M : StatsRegistry->MaxStats) { Result.push_back(M->value()); M->reset(); } - for (auto *M : StatsRegistry->UnsignedStats) { + for (auto *M : StatsRegistry->CounterStats) { Result.push_back(M->value()); M->reset(); } @@ -150,20 +173,33 @@ static std::string getUSR(const Decl *D) { } void Registry::Snapshot::dumpAsCSV(llvm::raw_ostream &OS) const { + auto PrintAsUnsignOpt = [&OS](std::optional<unsigned> U) { + OS << (U.has_value() ? std::to_string(*U) : ""); + }; + auto CommaIfNeeded = [&OS](const auto &Vec1, const auto &Vec2) { + if (!Vec1.empty() && !Vec2.empty()) + OS << ","; + }; + auto PrintAsUnsigned = [&OS](unsigned U) { OS << U; }; + OS << '"'; llvm::printEscapedString(getUSR(EntryPoint), OS); OS << "\",\""; OS << StatsRegistry->EscapedCPPFileName << "\",\""; llvm::printEscapedString( clang::AnalysisDeclContext::getFunctionName(EntryPoint), OS); - OS << "\""; - OS << (UnsignedStatValues.empty() ? "" : ","); - llvm::interleave(UnsignedStatValues, OS, [&OS](unsigned U) { OS << U; }, ","); + OS << "\","; + llvm::interleave(ExplicitlySetStatValues, OS, PrintAsUnsignOpt, ","); + CommaIfNeeded(ExplicitlySetStatValues, MaxOrCountStatValues); + llvm::interleave(MaxOrCountStatValues, OS, PrintAsUnsigned, ","); } void EntryPointStat::takeSnapshot(const Decl *EntryPoint) { - auto UnsignedValues = consumeUnsignedStats(); - StatsRegistry->Snapshots.push_back({EntryPoint, std::move(UnsignedValues)}); + auto ExplicitlySetValues = consumeExplicitlySetStats(); + auto MaxOrCounterValues = consumeMaxAndCounterStats(); + StatsRegistry->Snapshots.push_back({EntryPoint, + std::move(ExplicitlySetValues), + std::move(MaxOrCounterValues)}); } void EntryPointStat::dumpStatsAsCSV(llvm::StringRef FileName) { diff --git a/clang/lib/StaticAnalyzer/Core/RangeConstraintManager.cpp b/clang/lib/StaticAnalyzer/Core/RangeConstraintManager.cpp index ab45e67..245a730 100644 --- a/clang/lib/StaticAnalyzer/Core/RangeConstraintManager.cpp +++ b/clang/lib/StaticAnalyzer/Core/RangeConstraintManager.cpp @@ -983,7 +983,7 @@ public: } /// Check equivalence data for consistency. - [[nodiscard]] LLVM_ATTRIBUTE_UNUSED static bool + [[nodiscard]] [[maybe_unused]] static bool isClassDataConsistent(ProgramStateRef State); [[nodiscard]] QualType getType() const { @@ -1041,8 +1041,7 @@ private: // Constraint functions //===----------------------------------------------------------------------===// -[[nodiscard]] LLVM_ATTRIBUTE_UNUSED bool -areFeasible(ConstraintRangeTy Constraints) { +[[nodiscard]] [[maybe_unused]] bool areFeasible(ConstraintRangeTy Constraints) { return llvm::none_of( Constraints, [](const std::pair<EquivalenceClass, RangeSet> &ClassConstraint) { @@ -1134,7 +1133,7 @@ template <class EndTy> return End; } -[[nodiscard]] LLVM_ATTRIBUTE_UNUSED inline std::optional<RangeSet> +[[nodiscard]] [[maybe_unused]] inline std::optional<RangeSet> intersect(RangeSet::Factory &F, const RangeSet *End) { // This is an extraneous conversion from a raw pointer into // std::optional<RangeSet> diff --git a/clang/lib/StaticAnalyzer/Core/RegionStore.cpp b/clang/lib/StaticAnalyzer/Core/RegionStore.cpp index af0ef52..2838533 100644 --- a/clang/lib/StaticAnalyzer/Core/RegionStore.cpp +++ b/clang/lib/StaticAnalyzer/Core/RegionStore.cpp @@ -2457,7 +2457,7 @@ NonLoc RegionStoreManager::createLazyBinding(RegionBindingsConstRef B, SVal RegionStoreManager::getBindingForStruct(RegionBindingsConstRef B, const TypedValueRegion *R) { const RecordDecl *RD = - R->getValueType()->castAsCanonical<RecordType>()->getOriginalDecl(); + R->getValueType()->castAsCanonical<RecordType>()->getDecl(); if (!RD->getDefinition()) return UnknownVal(); diff --git a/clang/lib/StaticAnalyzer/Core/SimpleSValBuilder.cpp b/clang/lib/StaticAnalyzer/Core/SimpleSValBuilder.cpp index 84a9c43..6108931 100644 --- a/clang/lib/StaticAnalyzer/Core/SimpleSValBuilder.cpp +++ b/clang/lib/StaticAnalyzer/Core/SimpleSValBuilder.cpp @@ -1111,6 +1111,10 @@ SVal SimpleSValBuilder::evalBinOpLN(ProgramStateRef state, assert(!BinaryOperator::isComparisonOp(op) && "arguments to comparison ops must be of the same type"); + SVal simplifiedRhs = simplifySVal(state, rhs); + if (auto simplifiedRhsAsNonLoc = simplifiedRhs.getAs<NonLoc>()) + rhs = *simplifiedRhsAsNonLoc; + // Special case: rhs is a zero constant. if (rhs.isZeroConstant()) return lhs; diff --git a/clang/lib/StaticAnalyzer/Frontend/AnalysisConsumer.cpp b/clang/lib/StaticAnalyzer/Frontend/AnalysisConsumer.cpp index cf01e2f..f6a3e79 100644 --- a/clang/lib/StaticAnalyzer/Frontend/AnalysisConsumer.cpp +++ b/clang/lib/StaticAnalyzer/Frontend/AnalysisConsumer.cpp @@ -39,6 +39,7 @@ #include "llvm/Support/TimeProfiler.h" #include "llvm/Support/Timer.h" #include "llvm/Support/raw_ostream.h" +#include <cmath> #include <memory> #include <utility> @@ -61,7 +62,9 @@ ALWAYS_ENABLED_STATISTIC( "The # of visited basic blocks in the analyzed functions."); ALWAYS_ENABLED_STATISTIC(PercentReachableBlocks, "The % of reachable basic blocks."); -STAT_MAX(MaxCFGSize, "The maximum number of basic blocks in a function."); +ALWAYS_ENABLED_STATISTIC(MaxCFGSize, + "The maximum number of basic blocks in a function."); +static UnsignedEPStat CFGSize("CFGSize"); //===----------------------------------------------------------------------===// // AnalysisConsumer declaration. //===----------------------------------------------------------------------===// @@ -125,6 +128,7 @@ public: std::unique_ptr<llvm::Timer> SyntaxCheckTimer; std::unique_ptr<llvm::Timer> ExprEngineTimer; std::unique_ptr<llvm::Timer> BugReporterTimer; + bool ShouldClearTimersToPreventDisplayingThem; /// The information about analyzed functions shared throughout the /// translation unit. @@ -138,11 +142,12 @@ public: Injector(std::move(injector)), CTU(CI), MacroExpansions(CI.getLangOpts()) { - EntryPointStat::lockRegistry(getMainFileName(CI.getInvocation())); + EntryPointStat::lockRegistry(getMainFileName(CI.getInvocation()), + CI.getASTContext()); DigestAnalyzerOptions(); if (Opts.AnalyzerDisplayProgress || Opts.PrintStats || - Opts.ShouldSerializeStats) { + Opts.ShouldSerializeStats || !Opts.DumpEntryPointStatsToCSV.empty()) { AnalyzerTimers = std::make_unique<llvm::TimerGroup>( "analyzer", "Analyzer timers"); SyntaxCheckTimer = std::make_unique<llvm::Timer>( @@ -154,6 +159,12 @@ public: *AnalyzerTimers); } + // Avoid displaying the timers created above in case we only want to record + // per-entry-point stats. + ShouldClearTimersToPreventDisplayingThem = !Opts.AnalyzerDisplayProgress && + !Opts.PrintStats && + !Opts.ShouldSerializeStats; + if (Opts.PrintStats || Opts.ShouldSerializeStats) { llvm::EnableStatistics(/* DoPrintOnExit= */ false); } @@ -276,6 +287,9 @@ public: checkerMgr->runCheckersOnASTDecl(D, *Mgr, *RecVisitorBR); if (SyntaxCheckTimer) SyntaxCheckTimer->stopTimer(); + if (AnalyzerTimers && ShouldClearTimersToPreventDisplayingThem) { + AnalyzerTimers->clear(); + } } return true; } @@ -569,6 +583,9 @@ void AnalysisConsumer::runAnalysisOnTranslationUnit(ASTContext &C) { checkerMgr->runCheckersOnASTDecl(TU, *Mgr, BR); if (SyntaxCheckTimer) SyntaxCheckTimer->stopTimer(); + if (AnalyzerTimers && ShouldClearTimersToPreventDisplayingThem) { + AnalyzerTimers->clear(); + } // Run the AST-only checks using the order in which functions are defined. // If inlining is not turned on, use the simplest function order for path @@ -745,6 +762,9 @@ void AnalysisConsumer::HandleCode(Decl *D, AnalysisMode Mode, llvm::TimeRecord CheckerEndTime = SyntaxCheckTimer->getTotalTime(); CheckerEndTime -= CheckerStartTime; DisplayTime(CheckerEndTime); + if (AnalyzerTimers && ShouldClearTimersToPreventDisplayingThem) { + AnalyzerTimers->clear(); + } } } @@ -765,15 +785,19 @@ void AnalysisConsumer::HandleCode(Decl *D, AnalysisMode Mode, void AnalysisConsumer::RunPathSensitiveChecks(Decl *D, ExprEngine::InliningModes IMode, SetOfConstDecls *VisitedCallees) { + auto *CFG = Mgr->getCFG(D); + // Construct the analysis engine. First check if the CFG is valid. // FIXME: Inter-procedural analysis will need to handle invalid CFGs. - if (!Mgr->getCFG(D)) + if (!CFG) return; // See if the LiveVariables analysis scales. if (!Mgr->getAnalysisDeclContext(D)->getAnalysis<RelaxedLiveVariables>()) return; + CFGSize.set(CFG->size()); + ExprEngine Eng(CTU, *Mgr, VisitedCallees, &FunctionSummaries, IMode); // Execute the worklist algorithm. @@ -788,7 +812,12 @@ void AnalysisConsumer::RunPathSensitiveChecks(Decl *D, ExprEngineTimer->stopTimer(); llvm::TimeRecord ExprEngineEndTime = ExprEngineTimer->getTotalTime(); ExprEngineEndTime -= ExprEngineStartTime; + PathRunningTime.set(static_cast<unsigned>( + std::lround(ExprEngineEndTime.getWallTime() * 1000))); DisplayTime(ExprEngineEndTime); + if (AnalyzerTimers && ShouldClearTimersToPreventDisplayingThem) { + AnalyzerTimers->clear(); + } } if (!Mgr->options.DumpExplodedGraphTo.empty()) @@ -799,6 +828,9 @@ void AnalysisConsumer::RunPathSensitiveChecks(Decl *D, Eng.ViewGraph(Mgr->options.TrimGraph); flushReports(BugReporterTimer.get(), Eng.getBugReporter()); + if (AnalyzerTimers && ShouldClearTimersToPreventDisplayingThem) { + AnalyzerTimers->clear(); + } } //===----------------------------------------------------------------------===// diff --git a/clang/lib/Tooling/CompilationDatabase.cpp b/clang/lib/Tooling/CompilationDatabase.cpp index 860457a..4070bb8 100644 --- a/clang/lib/Tooling/CompilationDatabase.cpp +++ b/clang/lib/Tooling/CompilationDatabase.cpp @@ -403,7 +403,7 @@ namespace tooling { // This anchor is used to force the linker to link in the generated object file // and thus register the JSONCompilationDatabasePlugin. extern volatile int JSONAnchorSource; -static int LLVM_ATTRIBUTE_UNUSED JSONAnchorDest = JSONAnchorSource; +[[maybe_unused]] static int JSONAnchorDest = JSONAnchorSource; } // namespace tooling } // namespace clang diff --git a/clang/lib/Tooling/Execution.cpp b/clang/lib/Tooling/Execution.cpp index 247b260..d0499fa 100644 --- a/clang/lib/Tooling/Execution.cpp +++ b/clang/lib/Tooling/Execution.cpp @@ -96,9 +96,9 @@ createExecutorFromCommandLineArgs(int &argc, const char **argv, // and thus register the StandaloneToolExecutorPlugin etc. extern volatile int StandaloneToolExecutorAnchorSource; extern volatile int AllTUsToolExecutorAnchorSource; -static int LLVM_ATTRIBUTE_UNUSED StandaloneToolExecutorAnchorDest = +[[maybe_unused]] static int StandaloneToolExecutorAnchorDest = StandaloneToolExecutorAnchorSource; -static int LLVM_ATTRIBUTE_UNUSED AllTUsToolExecutorAnchorDest = +[[maybe_unused]] static int AllTUsToolExecutorAnchorDest = AllTUsToolExecutorAnchorSource; } // end namespace tooling diff --git a/clang/lib/Tooling/Syntax/BuildTree.cpp b/clang/lib/Tooling/Syntax/BuildTree.cpp index 90fd1f9..9d49d72 100644 --- a/clang/lib/Tooling/Syntax/BuildTree.cpp +++ b/clang/lib/Tooling/Syntax/BuildTree.cpp @@ -77,8 +77,10 @@ static Expr *IgnoreImplicit(Expr *E) { IgnoreCXXFunctionalCastExprWrappingConstructor); } -LLVM_ATTRIBUTE_UNUSED -static bool isImplicitExpr(Expr *E) { return IgnoreImplicit(E) != E; } +[[maybe_unused]] +static bool isImplicitExpr(Expr *E) { + return IgnoreImplicit(E) != E; +} namespace { /// Get start location of the Declarator from the TypeLoc. diff --git a/clang/test/AST/ByteCode/cxx11.cpp b/clang/test/AST/ByteCode/cxx11.cpp index 72bc762..8efd320 100644 --- a/clang/test/AST/ByteCode/cxx11.cpp +++ b/clang/test/AST/ByteCode/cxx11.cpp @@ -146,6 +146,14 @@ void testValueInRangeOfEnumerationValues() { const NumberType neg_one = (NumberType) ((NumberType) 0 - (NumberType) 1); // ok, not a constant expression context } +struct EnumTest { + enum type { + Type1, + BOUND + }; + static const type binding_completed = type(BOUND + 1); // both-error {{in-class initializer for static data member is not a constant expression}} \ + // both-note {{integer value 2 is outside the valid range of values}} +}; template<class T, unsigned size> struct Bitfield { static constexpr T max = static_cast<T>((1 << size) - 1); diff --git a/clang/test/AST/ByteCode/typeid.cpp b/clang/test/AST/ByteCode/typeid.cpp index 00b01c8..aca18d4 100644 --- a/clang/test/AST/ByteCode/typeid.cpp +++ b/clang/test/AST/ByteCode/typeid.cpp @@ -59,3 +59,16 @@ namespace TypeidPtrInEvaluationResult { consteval const std::type_info *ftype_info() { return &typeid(c); } const std::type_info *T1 = ftype_info(); } + +// Regression test for crash in ArrayElemPtrPop with typeid pointers. GH-163127 +namespace TypeidPtrRegression { + void dontcrash() { + constexpr auto res = ((void**)&typeid(int))[0]; // both-error {{must be initialized by a constant expression}} \ + // both-note {{cast that performs the conversions of a reinterpret_cast is not allowed in a constant expression}} + } + void dontcrash2() { + constexpr auto res = ((void**)&typeid(int))[1]; // both-error {{must be initialized by a constant expression}} \ + // both-note {{cast that performs the conversions of a reinterpret_cast is not allowed in a constant expression}} + } +} + diff --git a/clang/test/Analysis/Checkers/WebKit/forward-decl-checker.mm b/clang/test/Analysis/Checkers/WebKit/forward-decl-checker.mm index 104b555..8aad838 100644 --- a/clang/test/Analysis/Checkers/WebKit/forward-decl-checker.mm +++ b/clang/test/Analysis/Checkers/WebKit/forward-decl-checker.mm @@ -11,6 +11,8 @@ class Obj; Obj* provide_obj_ptr(); void receive_obj_ptr(Obj* p = nullptr); +void receive_obj_ref(Obj&); +void receive_obj_rref(Obj&&); sqlite3* open_db(); void close_db(sqlite3*); @@ -38,6 +40,16 @@ Obj& ref() { return obj; } +void opaque_call_arg(Obj* obj, Obj&& otherObj, const RefPtr<Obj>& safeObj, WeakPtr<Obj> weakObj, std::unique_ptr<Obj>& uniqObj) { + receive_obj_ref(*obj); + receive_obj_ptr(&*obj); + receive_obj_rref(std::move(otherObj)); + receive_obj_ref(*safeObj.get()); + receive_obj_ptr(weakObj.get()); + // expected-warning@-1{{Call argument for parameter 'p' uses a forward declared type 'Obj *'}} + receive_obj_ref(*uniqObj); +} + Obj&& provide_obj_rval(); void receive_obj_rval(Obj&& p); diff --git a/clang/test/Analysis/Checkers/WebKit/mock-system-header.h b/clang/test/Analysis/Checkers/WebKit/mock-system-header.h index 1e44de8..d55b3ab 100644 --- a/clang/test/Analysis/Checkers/WebKit/mock-system-header.h +++ b/clang/test/Analysis/Checkers/WebKit/mock-system-header.h @@ -34,6 +34,8 @@ void os_log_msg(os_log_t oslog, os_log_type_t type, const char *msg, ...); typedef const struct __attribute__((objc_bridge(NSString))) __CFString * CFStringRef; +extern CFStringRef const kCFURLTagNamesKey; + #ifdef __OBJC__ @class NSString; @interface SystemObject { @@ -41,4 +43,8 @@ typedef const struct __attribute__((objc_bridge(NSString))) __CFString * CFStrin CFStringRef cf_string; } @end + +typedef NSString *NSNotificationName; +extern "C" NSNotificationName NSApplicationDidBecomeActiveNotification; + #endif diff --git a/clang/test/Analysis/Checkers/WebKit/mock-types.h b/clang/test/Analysis/Checkers/WebKit/mock-types.h index a49faa1..7055a94 100644 --- a/clang/test/Analysis/Checkers/WebKit/mock-types.h +++ b/clang/test/Analysis/Checkers/WebKit/mock-types.h @@ -25,23 +25,23 @@ namespace std { template <typename T> class unique_ptr { private: - T *t; + void *t; public: unique_ptr() : t(nullptr) { } unique_ptr(T *t) : t(t) { } ~unique_ptr() { if (t) - delete t; + delete static_cast<T*>(t); } template <typename U> unique_ptr(unique_ptr<U>&& u) : t(u.t) { u.t = nullptr; } - T *get() const { return t; } - T *operator->() const { return t; } - T &operator*() const { return *t; } + T *get() const { return static_cast<T*>(t); } + T *operator->() const { return get(); } + T &operator*() const { return *get(); } unique_ptr &operator=(T *) { return *this; } explicit operator bool() const { return !!t; } }; @@ -313,4 +313,90 @@ public: UniqueRef &operator=(T &) { return *this; } }; +class WeakPtrImpl { +private: + void* ptr { nullptr }; + mutable unsigned m_refCount { 0 }; + + template <typename U> friend class CanMakeWeakPtr; + template <typename U> friend class WeakPtr; + +public: + template <typename T> + static Ref<WeakPtrImpl> create(T& t) + { + return adoptRef(*new WeakPtrImpl(t)); + } + + void ref() const { m_refCount++; } + void deref() const { + m_refCount--; + if (!m_refCount) + delete const_cast<WeakPtrImpl*>(this); + } + + template <typename T> + T* get() { return static_cast<T*>(ptr); } + operator bool() const { return !!ptr; } + void clear() { ptr = nullptr; } + +private: + template <typename T> + WeakPtrImpl(T* t) + : ptr(static_cast<void*>(t)) + { } +}; + +template <typename T> +class CanMakeWeakPtr { +private: + RefPtr<WeakPtrImpl> impl; + + template <typename U> friend class CanMakeWeakPtr; + template <typename U> friend class WeakPtr; + + Ref<WeakPtrImpl> createWeakPtrImpl() { + if (!impl) + impl = WeakPtrImpl::create(static_cast<T>(*this)); + return *impl; + } + +public: + ~CanMakeWeakPtr() { + if (!impl) + return; + impl->clear(); + impl = nullptr; + } +}; + +template <typename T> +class WeakPtr { +private: + RefPtr<WeakPtrImpl> impl; + +public: + WeakPtr(T& t) { + *this = t; + } + WeakPtr(T* t) { + *this = t; + } + + template <typename U> + WeakPtr<T> operator=(U& obj) { + impl = obj.createWeakPtrImpl(); + } + + template <typename U> + WeakPtr<T> operator=(U* obj) { + impl = obj ? obj->createWeakPtrImpl() : nullptr; + } + + T* get() { + return impl ? impl->get<T>() : nullptr; + } + +}; + #endif diff --git a/clang/test/Analysis/Checkers/WebKit/objc-mock-types.h b/clang/test/Analysis/Checkers/WebKit/objc-mock-types.h index a5fc3d7..edf4011 100644 --- a/clang/test/Analysis/Checkers/WebKit/objc-mock-types.h +++ b/clang/test/Analysis/Checkers/WebKit/objc-mock-types.h @@ -98,12 +98,20 @@ typedef CVImageBufferRef CVPixelBufferRef; typedef signed int CVReturn; CVReturn CVPixelBufferCreateWithIOSurface(CFAllocatorRef allocator, IOSurfaceRef surface, CFDictionaryRef pixelBufferAttributes, CF_RETURNS_RETAINED CVPixelBufferRef * pixelBufferOut); +extern "C" NSString *NSStringFromSelector(SEL aSelector); +extern "C" SEL NSSelectorFromString(NSString *aSelectorName); + +extern "C" NSString *NSStringFromClass(Class aClass); +extern "C" Class NSClassFromString(NSString *aClassName); + +extern "C" NSString *NSStringFromProtocol(Protocol *proto); +extern "C" Protocol * NSProtocolFromString(NSString *namestr); + CFRunLoopRef CFRunLoopGetCurrent(void); CFRunLoopRef CFRunLoopGetMain(void); extern CFTypeRef CFRetain(CFTypeRef cf); extern void CFRelease(CFTypeRef cf); #define CFSTR(cStr) ((CFStringRef) __builtin___CFStringMakeConstantString ("" cStr "")) -extern Class NSClassFromString(NSString *aClassName); #if __has_feature(objc_arc) id CFBridgingRelease(CFTypeRef X) { diff --git a/clang/test/Analysis/Checkers/WebKit/unretained-call-args.mm b/clang/test/Analysis/Checkers/WebKit/unretained-call-args.mm index a517dbc..8bef24f 100644 --- a/clang/test/Analysis/Checkers/WebKit/unretained-call-args.mm +++ b/clang/test/Analysis/Checkers/WebKit/unretained-call-args.mm @@ -567,6 +567,64 @@ struct Derived : Base { } // namespace ns_retained_return_value +namespace autoreleased { + +NSString *provideAutoreleased() __attribute__((ns_returns_autoreleased)); +void consume(NSString *); + +void foo() { + consume(provideAutoreleased()); +} + +} // autoreleased + +namespace sel_string { + +void consumeStr(NSString *); +void consumeSel(SEL); +void consumeClass(Class); +void consumeProtocol(Protocol *); + +void foo() { + consumeStr(NSStringFromSelector(@selector(mutableCopy))); + consumeSel(NSSelectorFromString(@"mutableCopy")); + consumeStr(NSStringFromClass(NSNumber.class)); + consumeClass(NSClassFromString(@"NSNumber")); + consumeStr(NSStringFromProtocol(@protocol(NSCopying))); + consumeProtocol(NSProtocolFromString(@"NSCopying")); +} + +} // namespace sel_string + +namespace template_function { + +class Base { +public: + virtual ~Base() = default; + void send(dispatch_queue_t) const; + void ref() const; + void deref() const; +}; + +template<typename Traits> +class Derived : public Base { +public: + virtual ~Derived() = default; + + void send(typename Traits::MessageType) const; + + virtual OSObjectPtr<dispatch_queue_t> msg(typename Traits::MessageType) const = 0; +}; + +template<typename Traits> +void Derived<Traits>::send(typename Traits::MessageType messageType) const +{ + OSObjectPtr dictionary = msg(messageType); + Base::send(dictionary.get()); +} + +} // namespace template_function + @interface TestObject : NSObject - (void)doWork:(NSString *)msg, ...; - (void)doWorkOnSelf; diff --git a/clang/test/Analysis/Checkers/WebKit/unretained-local-vars.mm b/clang/test/Analysis/Checkers/WebKit/unretained-local-vars.mm index 307a4d03..f49e7bd 100644 --- a/clang/test/Analysis/Checkers/WebKit/unretained-local-vars.mm +++ b/clang/test/Analysis/Checkers/WebKit/unretained-local-vars.mm @@ -1,8 +1,11 @@ // RUN: %clang_analyze_cc1 -analyzer-checker=alpha.webkit.UnretainedLocalVarsChecker -verify %s #import "objc-mock-types.h" +#import "mock-system-header.h" void someFunction(); +extern "C" CFStringRef LocalGlobalCFString; +extern "C" NSString *LocalGlobalNSString; namespace raw_ptr { void foo() { @@ -535,6 +538,41 @@ unsigned foo() { } // namespace ns_retained_return_value +namespace autoreleased { + +NSString *provideAutoreleased() __attribute__((ns_returns_autoreleased)); +void consume(NSString *); + +void foo() { + auto *string = provideAutoreleased(); + consume(string); +} + +} // autoreleased + +namespace ns_global { + +void consumeCFString(CFStringRef); +void consumeNSString(NSString *); + +void cf() { + auto *str = kCFURLTagNamesKey; + consumeCFString(str); + auto *localStr = LocalGlobalCFString; + // expected-warning@-1{{Local variable 'localStr' is unretained and unsafe [alpha.webkit.UnretainedLocalVarsChecker]}} + consumeCFString(localStr); +} + +void ns() { + auto *str = NSApplicationDidBecomeActiveNotification; + consumeNSString(str); + auto *localStr = LocalGlobalNSString; + // expected-warning@-1{{Local variable 'localStr' is unretained and unsafe [alpha.webkit.UnretainedLocalVarsChecker]}} + consumeNSString(localStr); +} + +} + bool doMoreWorkOpaque(OtherObj*); SomeObj* provide(); diff --git a/clang/test/Analysis/Checkers/WebKit/unretained-obj-arg.mm b/clang/test/Analysis/Checkers/WebKit/unretained-obj-arg.mm new file mode 100644 index 0000000..5c78b21 --- /dev/null +++ b/clang/test/Analysis/Checkers/WebKit/unretained-obj-arg.mm @@ -0,0 +1,18 @@ +// RUN: %clang_analyze_cc1 -analyzer-checker=alpha.webkit.UnretainedCallArgsChecker -verify %s + +#import "mock-types.h" +#import "mock-system-header.h" + +void consumeCFString(CFStringRef); +extern "C" CFStringRef LocalGlobalCFString; +void consumeNSString(NSString *); +extern "C" NSString *LocalGlobalNSString; + +void foo() { + consumeCFString(kCFURLTagNamesKey); + consumeCFString(LocalGlobalCFString); + // expected-warning@-1{{Call argument is unretained and unsafe}} + consumeNSString(NSApplicationDidBecomeActiveNotification); + consumeNSString(LocalGlobalNSString); + // expected-warning@-1{{Call argument is unretained and unsafe}} +} diff --git a/clang/test/Analysis/analyzer-enabled-checkers.c b/clang/test/Analysis/analyzer-enabled-checkers.c index 0092331..bfe418b 100644 --- a/clang/test/Analysis/analyzer-enabled-checkers.c +++ b/clang/test/Analysis/analyzer-enabled-checkers.c @@ -19,6 +19,7 @@ // CHECK-NEXT: core.NonNullParamChecker // CHECK-NEXT: core.NonnilStringConstants // CHECK-NEXT: core.NullDereference +// CHECK-NEXT: core.NullPointerArithm // CHECK-NEXT: core.StackAddressEscape // CHECK-NEXT: core.UndefinedBinaryOperatorResult // CHECK-NEXT: core.VLASize diff --git a/clang/test/Analysis/analyzer-stats/entry-point-stats.cpp b/clang/test/Analysis/analyzer-stats/entry-point-stats.cpp index 9cbe045..ebbc015 100644 --- a/clang/test/Analysis/analyzer-stats/entry-point-stats.cpp +++ b/clang/test/Analysis/analyzer-stats/entry-point-stats.cpp @@ -8,6 +8,13 @@ // CHECK-NEXT: "c:@F@fib#i#": { // CHECK-NEXT: "File": "{{.*}}entry-point-stats.cpp", // CHECK-NEXT: "DebugName": "fib(unsigned int)", +// CHECK-NEXT: "CFGSize": "5", +// CHECK-NEXT: "PathRunningTime": "{{[0-9]+}}", +// CHECK-NEXT: "MaxBugClassSize": "{{[0-9]+}}", +// CHECK-NEXT: "MaxQueueSize": "{{[0-9]+}}", +// CHECK-NEXT: "MaxReachableSize": "{{[0-9]+}}", +// CHECK-NEXT: "MaxTimeSpentSolvingZ3Queries": "{{[0-9]+}}", +// CHECK-NEXT: "MaxValidBugClassSize": "{{[0-9]+}}", // CHECK-NEXT: "NumBlocks": "{{[0-9]+}}", // CHECK-NEXT: "NumBlocksUnreachable": "{{[0-9]+}}", // CHECK-NEXT: "NumCTUSteps": "{{[0-9]+}}", @@ -33,18 +40,18 @@ // CHECK-NEXT: "NumTimesZ3SpendsTooMuchTimeOnASingleEQClass": "{{[0-9]+}}", // CHECK-NEXT: "NumTimesZ3TimedOut": "{{[0-9]+}}", // CHECK-NEXT: "NumZ3QueriesDone": "{{[0-9]+}}", -// CHECK-NEXT: "TimeSpentSolvingZ3Queries": "{{[0-9]+}}", +// CHECK-NEXT: "TimeSpentSolvingZ3Queries": "{{[0-9]+}}" +// CHECK-NEXT: }, +// CHECK-NEXT: "c:@F@main#I#**C#": { +// CHECK-NEXT: "File": "{{.*}}entry-point-stats.cpp", +// CHECK-NEXT: "DebugName": "main(int, char **)", +// CHECK-NEXT: "CFGSize": "3", +// CHECK-NEXT: "PathRunningTime": "{{[0-9]+}}", // CHECK-NEXT: "MaxBugClassSize": "{{[0-9]+}}", -// CHECK-NEXT: "MaxCFGSize": "{{[0-9]+}}", // CHECK-NEXT: "MaxQueueSize": "{{[0-9]+}}", // CHECK-NEXT: "MaxReachableSize": "{{[0-9]+}}", // CHECK-NEXT: "MaxTimeSpentSolvingZ3Queries": "{{[0-9]+}}", // CHECK-NEXT: "MaxValidBugClassSize": "{{[0-9]+}}", -// CHECK-NEXT: "PathRunningTime": "{{[0-9]+}}" -// CHECK-NEXT: }, -// CHECK-NEXT: "c:@F@main#I#**C#": { -// CHECK-NEXT: "File": "{{.*}}entry-point-stats.cpp", -// CHECK-NEXT: "DebugName": "main(int, char **)", // CHECK-NEXT: "NumBlocks": "{{[0-9]+}}", // CHECK-NEXT: "NumBlocksUnreachable": "{{[0-9]+}}", // CHECK-NEXT: "NumCTUSteps": "{{[0-9]+}}", @@ -70,14 +77,7 @@ // CHECK-NEXT: "NumTimesZ3SpendsTooMuchTimeOnASingleEQClass": "{{[0-9]+}}", // CHECK-NEXT: "NumTimesZ3TimedOut": "{{[0-9]+}}", // CHECK-NEXT: "NumZ3QueriesDone": "{{[0-9]+}}", -// CHECK-NEXT: "TimeSpentSolvingZ3Queries": "{{[0-9]+}}", -// CHECK-NEXT: "MaxBugClassSize": "{{[0-9]+}}", -// CHECK-NEXT: "MaxCFGSize": "{{[0-9]+}}", -// CHECK-NEXT: "MaxQueueSize": "{{[0-9]+}}", -// CHECK-NEXT: "MaxReachableSize": "{{[0-9]+}}", -// CHECK-NEXT: "MaxTimeSpentSolvingZ3Queries": "{{[0-9]+}}", -// CHECK-NEXT: "MaxValidBugClassSize": "{{[0-9]+}}", -// CHECK-NEXT: "PathRunningTime": "{{[0-9]+}}" +// CHECK-NEXT: "TimeSpentSolvingZ3Queries": "{{[0-9]+}}" // CHECK-NEXT: } // CHECK-NEXT: } // CHECK-NOT: non_entry_point diff --git a/clang/test/Analysis/loc-folding.cpp b/clang/test/Analysis/loc-folding.cpp new file mode 100644 index 0000000..1fcb066 --- /dev/null +++ b/clang/test/Analysis/loc-folding.cpp @@ -0,0 +1,61 @@ +// RUN: %clang_analyze_cc1 -verify %s -analyzer-config eagerly-assume=false \ +// RUN: -analyzer-checker=core,debug.ExprInspection + +void clang_analyzer_eval(bool); + +void element_constant() { + char arr[10]; + clang_analyzer_eval(arr + 1 > arr); // expected-warning{{TRUE}} +} + +void element_known() { + char arr[10]; + int off = 1; + clang_analyzer_eval(arr + off > arr); // expected-warning{{TRUE}} +} + +void element_constrained(int off) { + char arr[10]; + if (off == 1) { + clang_analyzer_eval(arr + off > arr); // expected-warning{{TRUE}} + } +} + +void element_unknown(int off) { + char arr[10]; + clang_analyzer_eval(arr + off > arr); // expected-warning{{UNKNOWN}} +} + +void element_complex(int off) { + char arr[10]; + int comp = off * 2; + if (off == 1) { + clang_analyzer_eval(arr + comp); // expected-warning{{TRUE}} + } +} + +void base_constant(int *arr) { + clang_analyzer_eval(arr + 1 > arr); // expected-warning{{TRUE}} +} + +void base_known(int *arr) { + int off = 1; + clang_analyzer_eval(arr + off > arr); // expected-warning{{TRUE}} +} + +void base_constrained(int *arr, int off) { + if (off == 1) { + clang_analyzer_eval(arr + off > arr); // expected-warning{{TRUE}} + } +} + +void base_unknown(int *arr, int off) { + clang_analyzer_eval(arr + off > arr); // expected-warning{{UNKNOWN}} +} + +void base_complex(int *arr, int off) { + int comp = off * 2; + if (off == 1) { + clang_analyzer_eval(arr + comp > arr); // expected-warning{{TRUE}} + } +} diff --git a/clang/test/Analysis/null-pointer-arithm.c b/clang/test/Analysis/null-pointer-arithm.c new file mode 100644 index 0000000..2288247 --- /dev/null +++ b/clang/test/Analysis/null-pointer-arithm.c @@ -0,0 +1,76 @@ +// RUN: %clang_analyze_cc1 -verify %s \ +// RUN: -analyzer-checker=core + +extern int *get_pointer(); + +int *test_add1(int offset) { + int *p = get_pointer(); + if (p) {} + return p + offset; // expected-warning{{Addition of a null pointer (from variable 'p') and a probably nonzero integer value (from variable 'offset') may result in undefined behavior}} +} + +int *test_add2(int offset) { + int *p = get_pointer(); + if (p) {} + if (offset) {} + return p + offset; // expected-warning{{Addition of a null pointer (from variable 'p') and a nonzero integer value (from variable 'offset') results in undefined behavior}} +} + +int *test_add3(int offset) { + int *p = get_pointer(); + if (p) {} + if (offset != 0) return 0; + return p + offset; +} + +int *test_add4(int offset) { + int *p = get_pointer(); + if (p) {} + if (offset == 0) return 0; + return p + offset; // expected-warning{{Addition of a null pointer (from variable 'p') and a nonzero integer value (from variable 'offset') results in undefined behavior}} +} + +int *test_add5(int offset) { + int *p = get_pointer(); + if (p) {} + return offset + p; // expected-warning{{Addition of a probably nonzero integer value (from variable 'offset') and a null pointer (from variable 'p') may result in undefined behavior}} +} + +int *test_sub1(int offset) { + int *p = get_pointer(); + if (p) {} + return p - offset; // expected-warning{{Subtraction of a null pointer (from variable 'p') and a probably nonzero integer value (from variable 'offset') may result in undefined behavior}} +} + +int test_sub_p1() { + int *p = get_pointer(); + if (p) {} + return p - p; +} + +int test_sub_p2() { + int *p1 = get_pointer(); + int *p2 = get_pointer(); + if (p1) {} + if (p2) {} + return p1 - p2; + // expected-warning@-1{{Subtraction of a non-null pointer (from variable 'p1') and a null pointer (from variable 'p2') results in undefined behavior}} + // expected-warning@-2{{Subtraction of a null pointer (from variable 'p1') and a non-null pointer (from variable 'p2') results in undefined behavior}} +} + +int test_sub_p3() { + int *p1 = get_pointer(); + int *p2 = get_pointer(); + if (p1) {} + return p1 - p2; // expected-warning{{Subtraction of a null pointer (from variable 'p1') and a probably non-null pointer (from variable 'p2') may result in undefined behavior}} +} + +struct S { + char *p; + int offset; +}; + +char *test_struct(struct S s) { + if (s.p) {} + return s.p + s.offset; // expected-warning{{Addition of a null pointer (via field 'p') and a probably nonzero integer value (via field 'offset') may result in undefined behavior}} +} diff --git a/clang/test/Analysis/std-c-library-functions-arg-enabled-checkers.c b/clang/test/Analysis/std-c-library-functions-arg-enabled-checkers.c index 7fae958..9b32960 100644 --- a/clang/test/Analysis/std-c-library-functions-arg-enabled-checkers.c +++ b/clang/test/Analysis/std-c-library-functions-arg-enabled-checkers.c @@ -27,6 +27,7 @@ // CHECK-NEXT: core.NonNullParamChecker // CHECK-NEXT: core.NonnilStringConstants // CHECK-NEXT: core.NullDereference +// CHECK-NEXT: core.NullPointerArithm // CHECK-NEXT: core.StackAddressEscape // CHECK-NEXT: core.UndefinedBinaryOperatorResult // CHECK-NEXT: core.VLASize diff --git a/clang/test/Analysis/string.c b/clang/test/Analysis/string.c index cdd3627..9d24583 100644 --- a/clang/test/Analysis/string.c +++ b/clang/test/Analysis/string.c @@ -82,16 +82,21 @@ size_t strlen(const char *s); void strlen_constant0(void) { clang_analyzer_eval(strlen("123") == 3); // expected-warning{{TRUE}} + clang_analyzer_eval(strlen(&("123"[1])) == 2); // expected-warning{{TRUE}} } void strlen_constant1(void) { const char *a = "123"; clang_analyzer_eval(strlen(a) == 3); // expected-warning{{TRUE}} + clang_analyzer_eval(strlen(a + 1) == 2); // expected-warning{{TRUE}} + clang_analyzer_eval(strlen(a + 3) == 0); // expected-warning{{TRUE}} + clang_analyzer_eval(strlen(a + 4)); // expected-warning{{UNKNOWN}} } void strlen_constant2(char x) { char a[] = "123"; clang_analyzer_eval(strlen(a) == 3); // expected-warning{{TRUE}} + clang_analyzer_eval(strlen(a + 1) == 2); // expected-warning{{UNKNOWN}} a[0] = x; clang_analyzer_eval(strlen(a) == 3); // expected-warning{{UNKNOWN}} @@ -105,10 +110,12 @@ char global_non_const_arr[] = "op"; void strlen_global_constant_ptr(void) { clang_analyzer_eval(strlen(global_str_ptr) == 4); // expected-warning{{TRUE}} + clang_analyzer_eval(strlen(global_str_ptr + 1) == 3); // expected-warning{{TRUE}} } void strlen_global_constant_arr(void) { clang_analyzer_eval(strlen(global_str_arr) == 4); // expected-warning{{TRUE}} + clang_analyzer_eval(strlen(global_str_arr + 1) == 3); // expected-warning{{TRUE}} } void strlen_global_non_const_ptr(void) { @@ -235,6 +242,17 @@ void testStrlenCallee(void) { clang_analyzer_eval(lenBefore == lenAfter); // expected-warning{{UNKNOWN}} } +void strlen_symbolic_offset(unsigned x) { + const char *str = "abcd"; + if (x < 1 || x > 3) + return; + // FIXME: these should be TRUE + clang_analyzer_eval(strlen(str + x) >= 1); // expected-warning{{UNKNOWN}} + clang_analyzer_eval(strlen(str + x) <= 3); // expected-warning{{UNKNOWN}} + if (x != 1) + return; + clang_analyzer_eval(strlen(str + x) == 3); // expected-warning{{TRUE}} +} //===----------------------------------------------------------------------=== // strnlen() @@ -244,32 +262,38 @@ size_t strnlen(const char *s, size_t maxlen); void strnlen_constant0(void) { clang_analyzer_eval(strnlen("123", 10) == 3); // expected-warning{{TRUE}} + clang_analyzer_eval(strnlen(&("123"[1]), 10) == 2); // expected-warning{{TRUE}} } void strnlen_constant1(void) { const char *a = "123"; clang_analyzer_eval(strnlen(a, 10) == 3); // expected-warning{{TRUE}} + clang_analyzer_eval(strnlen(a + 1, 10) == 2); // expected-warning{{TRUE}} } void strnlen_constant2(char x) { char a[] = "123"; clang_analyzer_eval(strnlen(a, 10) == 3); // expected-warning{{TRUE}} + clang_analyzer_eval(strnlen(a + 1, 10) == 2); // expected-warning{{UNKNOWN}} a[0] = x; clang_analyzer_eval(strnlen(a, 10) == 3); // expected-warning{{UNKNOWN}} } void strnlen_constant4(void) { clang_analyzer_eval(strnlen("123456", 3) == 3); // expected-warning{{TRUE}} + clang_analyzer_eval(strnlen(&("123456"[1]), 3) == 3); // expected-warning{{TRUE}} } void strnlen_constant5(void) { const char *a = "123456"; clang_analyzer_eval(strnlen(a, 3) == 3); // expected-warning{{TRUE}} + clang_analyzer_eval(strnlen(a + 1, 3) == 3); // expected-warning{{TRUE}} } void strnlen_constant6(char x) { char a[] = "123456"; clang_analyzer_eval(strnlen(a, 3) == 3); // expected-warning{{TRUE}} + clang_analyzer_eval(strnlen(a + 1, 3) == 3); // expected-warning{{UNKNOWN}} a[0] = x; clang_analyzer_eval(strnlen(a, 3) == 3); // expected-warning{{UNKNOWN}} } @@ -326,6 +350,19 @@ void strnlen_at_actual(size_t limit) { } } +void strnlen_at_actual_1(size_t limit) { + const char *str = "abc"; + size_t len = strnlen(str + 1, limit); + clang_analyzer_eval(len <= 2); // expected-warning{{TRUE}} + // This is due to eager assertion in strnlen. + if (limit == 0) { + clang_analyzer_eval(len == 0); // expected-warning{{TRUE}} + } else { + clang_analyzer_eval(len == 2); // expected-warning{{UNKNOWN}} + clang_analyzer_eval(len < 2); // expected-warning{{UNKNOWN}} + } +} + //===----------------------------------------------------------------------=== // strcpy() //===----------------------------------------------------------------------=== diff --git a/clang/test/Analysis/zero-size-non-pod-array.cpp b/clang/test/Analysis/zero-size-non-pod-array.cpp index 628be0d7..8a32a49 100644 --- a/clang/test/Analysis/zero-size-non-pod-array.cpp +++ b/clang/test/Analysis/zero-size-non-pod-array.cpp @@ -95,7 +95,7 @@ void zeroSizeArrayBinding() { // Note: This is an error in gcc but a warning in clang. // In MSVC the declaration of 'S arr[0]' is already an error // and it doesn't recognize this syntax as a structured binding. - auto [] = arr; //expected-warning{{ISO C++17 does not allow a decomposition group to be empty}} + auto [] = arr; //expected-warning{{ISO C++17 does not allow a structured binding group to be empty}} clang_analyzer_eval(S::CtorInvocationCount == 0); //expected-warning{{TRUE}} } diff --git a/clang/test/C/C23/n3037.c b/clang/test/C/C23/n3037.c index 3748375..113ecf7 100644 --- a/clang/test/C/C23/n3037.c +++ b/clang/test/C/C23/n3037.c @@ -30,11 +30,24 @@ void func2(PRODUCT(int, SUM(float, double)) y) { // c17-warning {{declaration of struct foop { struct { int x; }; }; // c17-note {{previous definition is here}} struct foop { struct { int x; }; }; // c17-error {{redefinition of 'foop'}} +// Test the field lookup compatibility isn't sufficient, the structure of types should be compatible. +struct AnonymousStructNotMatchingFields { // c17-note {{previous definition is here}} + struct { // c23-note {{field has name '' here}} + int x; + }; +}; +struct AnonymousStructNotMatchingFields { // c23-error {{type 'struct AnonymousStructNotMatchingFields' has incompatible definitions}} \ + c17-error {{redefinition of 'AnonymousStructNotMatchingFields'}} + int x; // c23-note {{field has name 'x' here}} +}; + union barp { int x; float y; }; // c17-note {{previous definition is here}} union barp { int x; float y; }; // c17-error {{redefinition of 'barp'}} typedef struct q { int x; } q_t; // c17-note 2 {{previous definition is here}} typedef struct q { int x; } q_t; // c17-error {{redefinition of 'q'}} \ c17-error-re {{typedef redefinition with different types ('struct (unnamed struct at {{.*}})' vs 'struct q')}} +typedef struct { int x; } untagged_q_t; // both-note {{previous definition is here}} +typedef struct { int x; } untagged_q_t; // both-error {{typedef redefinition with different types}} void func3(void) { struct S { int x; }; // c17-note {{previous definition is here}} struct T { struct S s; }; // c17-note {{previous definition is here}} @@ -389,13 +402,40 @@ void nontag_both_in_params(struct { int i; } Arg1, struct { int i; } Arg2) { _Static_assert(0 == _Generic(__typeof__(Arg1), __typeof__(Arg2) : 1, default : 0)); // both-warning {{passing a type argument as the first operand to '_Generic' is a C2y extension}} } -struct InnerAnonStruct { +struct InnerUnnamedStruct { struct { int i; } untagged; -} inner_anon_tagged; +} inner_unnamed_tagged; +_Static_assert(0 == _Generic(inner_unnamed_tagged.untagged, struct { int i; } : 1, default : 0)); -_Static_assert(0 == _Generic(inner_anon_tagged.untagged, struct { int i; } : 1, default : 0)); +struct InnerUnnamedStruct_same { + struct { + int i; + } untagged; +}; +struct InnerUnnamedStruct_differentNaming { + struct { + int i; + } untaggedDifferent; +}; +struct InnerUnnamedStruct_differentShape { + float x; + struct { + int i; + } untagged; + int y; +}; +void compare_unnamed_struct_from_different_outer_type( + struct InnerUnnamedStruct sameOuterType, + struct InnerUnnamedStruct_same matchingType, + struct InnerUnnamedStruct_differentNaming differentFieldName, + struct InnerUnnamedStruct_differentShape differentType) { + inner_unnamed_tagged.untagged = sameOuterType.untagged; + inner_unnamed_tagged.untagged = matchingType.untagged; // both-error-re {{assigning to 'struct (unnamed struct at {{.*}})' from incompatible type 'struct (unnamed struct at {{.*}})'}} + inner_unnamed_tagged.untagged = differentFieldName.untaggedDifferent; // both-error-re {{assigning to 'struct (unnamed struct at {{.*}})' from incompatible type 'struct (unnamed struct at {{.*}})'}} + inner_unnamed_tagged.untagged = differentType.untagged; // both-error-re {{assigning to 'struct (unnamed struct at {{.*}})' from incompatible type 'struct (unnamed struct at {{.*}})'}} +} // Test the same thing with enumerations (test for unions is omitted because // unions and structures are both RecordDecl objects, whereas EnumDecl is not). diff --git a/clang/test/C/C2y/n3364.c b/clang/test/C/C2y/n3364.c index 277b264..ccf7e8d 100644 --- a/clang/test/C/C2y/n3364.c +++ b/clang/test/C/C2y/n3364.c @@ -1,5 +1,5 @@ -// RUN: %clang_cc1 -verify -std=c2y -Wall -pedantic -emit-llvm -o - %s -// RUN: %clang_cc1 -verify -Wall -pedantic -emit-llvm -o - %s +// RUN: %clang_cc1 -verify -std=c2y -ffreestanding -Wall -pedantic -emit-llvm -o - %s | FileCheck %s +// RUN: %clang_cc1 -verify -ffreestanding -Wall -pedantic -emit-llvm -o - %s | FileCheck %s // expected-no-diagnostics /* WG14 N3364: Yes @@ -7,29 +7,36 @@ * * Ensure that initializing from a signaling NAN (optionally with a unary + or * -) at translation time behaves correctly at runtime. + * + * This also serves as a test for C23's WG14 N2710 which introduces these + * macros into float.h in Clang 22. */ -#define FLT_SNAN __builtin_nansf("1") -#define DBL_SNAN __builtin_nans("1") -#define LD_SNAN __builtin_nansl("1") +#if __STDC_VERSION__ >= 202311L +#include <float.h> +#else +#define FLT_SNAN __builtin_nansf("") +#define DBL_SNAN __builtin_nans("") +#define LDBL_SNAN __builtin_nansl("") +#endif float f1 = FLT_SNAN; float f2 = +FLT_SNAN; float f3 = -FLT_SNAN; -// CHECK: @f1 = {{.*}}global float 0x7FF0000020000000 -// CHECK: @f2 = {{.*}}global float 0x7FF0000020000000 -// CHECK: @f3 = {{.*}}global float 0xFFF0000020000000 +// CHECK: @f1 = {{.*}}global float 0x7FF4000000000000 +// CHECK: @f2 = {{.*}}global float 0x7FF4000000000000 +// CHECK: @f3 = {{.*}}global float 0xFFF4000000000000 double d1 = DBL_SNAN; double d2 = +DBL_SNAN; double d3 = -DBL_SNAN; -// CHECK: @d1 = {{.*}}global double 0x7FF0000000000001 -// CHECK: @d2 = {{.*}}global double 0x7FF0000000000001 -// CHECK: @d3 = {{.*}}global double 0xFFF0000000000001 +// CHECK: @d1 = {{.*}}global double 0x7FF4000000000000 +// CHECK: @d2 = {{.*}}global double 0x7FF4000000000000 +// CHECK: @d3 = {{.*}}global double 0xFFF4000000000000 -long double ld1 = LD_SNAN; -long double ld2 = +LD_SNAN; -long double ld3 = -LD_SNAN; -// CHECK: @ld1 = {{.*}}global {{double 0x7FF0000000000001|x86_fp80 0xK7FFF8000000000000001|fp128 0xL00000000000000017FFF000000000000}} -// CHECK: @ld2 = {{.*}}global {{double 0x7FF0000000000001|x86_fp80 0xK7FFF8000000000000001|fp128 0xL00000000000000017FFF000000000000}} -// CHECK: @ld3 = {{.*}}global {{double 0xFFF0000000000001|x86_fp80 0xKFFFF8000000000000001|fp128 0xL0000000000000001FFFF000000000000}} +long double ld1 = LDBL_SNAN; +long double ld2 = +LDBL_SNAN; +long double ld3 = -LDBL_SNAN; +// CHECK: @ld1 = {{.*}}global {{double 0x7FF4000000000000|x86_fp80 0xK7FFFA000000000000000|fp128 0xL00000000000000007FFF400000000000|ppc_fp128 0xM7FF40000000000000000000000000000}} +// CHECK: @ld2 = {{.*}}global {{double 0x7FF4000000000000|x86_fp80 0xK7FFFA000000000000000|fp128 0xL00000000000000007FFF400000000000|ppc_fp128 0xM7FF40000000000000000000000000000}} +// CHECK: @ld3 = {{.*}}global {{double 0xFFF4000000000000|x86_fp80 0xKFFFFA000000000000000|fp128 0xL0000000000000000FFFF400000000000|ppc_fp128 0xMFFF40000000000008000000000000000}} diff --git a/clang/test/CIR/CodeGen/array.cpp b/clang/test/CIR/CodeGen/array.cpp index d7488bf..82add4b 100644 --- a/clang/test/CIR/CodeGen/array.cpp +++ b/clang/test/CIR/CodeGen/array.cpp @@ -123,7 +123,7 @@ void func() { // CIR: %[[TMP:.*]] = cir.load{{.*}} %[[ELE_PTR]] : !cir.ptr<!s32i>, !s32i // CIR" cir.store %[[TMP]], %[[INIT_2]] : !s32i, !cir.ptr<!s32i> -// LLVM: define{{.*}} void @_Z4funcv() +// LLVM: define{{.*}} void @_Z4funcv(){{.*}} // LLVM-NEXT: %[[ARR:.*]] = alloca [10 x i32], i64 1, align 16 // LLVM-NEXT: %[[INIT:.*]] = alloca i32, i64 1, align 4 // LLVM-NEXT: %[[INIT_2:.*]] = alloca i32, i64 1, align 4 @@ -174,7 +174,7 @@ void func2() { // CIR: cir.condition(%[[CMP]]) // CIR: } -// LLVM: define{{.*}} void @_Z5func2v() +// LLVM: define{{.*}} void @_Z5func2v(){{.*}} // LLVM: %[[ARR:.*]] = alloca [2 x i32], i64 1, align 4 // LLVM: %[[TMP:.*]] = alloca ptr, i64 1, align 8 // LLVM: %[[ARR_PTR:.*]] = getelementptr i32, ptr %[[ARR]], i32 0 @@ -224,7 +224,7 @@ void func3() { // CIR: %[[ELE_TMP:.*]] = cir.load{{.*}} %[[ELE_PTR]] : !cir.ptr<!s32i>, !s32i // CIR: cir.store{{.*}} %[[ELE_TMP]], %[[INIT]] : !s32i, !cir.ptr<!s32i> -// LLVM: define{{.*}} void @_Z5func3v() +// LLVM: define{{.*}} void @_Z5func3v(){{.*}} // LLVM: %[[ARR:.*]] = alloca [2 x i32], i64 1, align 4 // LLVM: %[[IDX:.*]] = alloca i32, i64 1, align 4 // LLVM: %[[INIT:.*]] = alloca i32, i64 1, align 4 @@ -276,7 +276,7 @@ void func4() { // CIR: %[[TMP:.*]] = cir.load{{.*}} %[[ELE_0]] : !cir.ptr<!s32i>, !s32i // CIR: cir.store{{.*}} %[[TMP]], %[[INIT]] : !s32i, !cir.ptr<!s32i> -// LLVM: define{{.*}} void @_Z5func4v() +// LLVM: define{{.*}} void @_Z5func4v(){{.*}} // LLVM: %[[ARR:.*]] = alloca [2 x [1 x i32]], i64 1, align 4 // LLVM: %[[INIT:.*]] = alloca i32, i64 1, align 4 // LLVM: %[[ARR_PTR:.*]] = getelementptr [1 x i32], ptr %[[ARR]], i32 0 @@ -329,7 +329,7 @@ void func5() { // CIR: cir.condition(%[[CMP]]) // CIR: } -// LLVM: define{{.*}} void @_Z5func5v() +// LLVM: define{{.*}} void @_Z5func5v(){{.*}} // LLVM: %[[ARR:.*]] = alloca [2 x [1 x i32]], i64 1, align 4 // LLVM: %[[TMP:.*]] = alloca ptr, i64 1, align 8 // LLVM: %[[ARR_PTR:.*]] = getelementptr [1 x i32], ptr %[[ARR]], i32 0 @@ -372,7 +372,7 @@ void func6() { // CIR: %[[V1:.*]] = cir.const #cir.int<5> : !s32i // CIR: cir.store{{.*}} %[[V1]], %[[ELE_PTR]] : !s32i, !cir.ptr<!s32i> -// LLVM: define{{.*}} void @_Z5func6v() +// LLVM: define{{.*}} void @_Z5func6v(){{.*}} // LLVM: %[[VAR:.*]] = alloca i32, i64 1, align 4 // LLVM: %[[ARR:.*]] = alloca [2 x i32], i64 1, align 4 // LLVM: store i32 4, ptr %[[VAR]], align 4 @@ -414,7 +414,7 @@ void func7() { // CIR: cir.condition(%[[CMP]]) // CIR: } -// LLVM: define{{.*}} void @_Z5func7v() +// LLVM: define{{.*}} void @_Z5func7v(){{.*}} // LLVM: %[[ARR:.*]] = alloca [1 x ptr], i64 1, align 8 // LLVM: %[[TMP:.*]] = alloca ptr, i64 1, align 8 // LLVM: %[[ARR_PTR:.*]] = getelementptr ptr, ptr %[[ARR]], i32 0 @@ -458,7 +458,7 @@ void func8(int arr[10]) { // CIR: %[[TMP_4:.*]] = cir.load{{.*}} %[[ELE_1]] : !cir.ptr<!s32i>, !s32i // CIR: cir.store{{.*}} %[[TMP_4]], %[[INIT_2]] : !s32i, !cir.ptr<!s32i> -// LLVM: define{{.*}} void @_Z5func8Pi(ptr %[[ARG:.*]]) +// LLVM: define{{.*}} void @_Z5func8Pi(ptr %[[ARG:.*]]){{.*}} // LLVM: %[[ARR:.*]] = alloca ptr, i64 1, align 8 // LLVM: %[[INIT:.*]] = alloca i32, i64 1, align 4 // LLVM: %[[INIT_2:.*]] = alloca i32, i64 1, align 4 @@ -502,7 +502,7 @@ void func9(int arr[10][5]) { // CIR: %[[TMP_2:.*]] = cir.load{{.*}} %[[ARR_1_2]] : !cir.ptr<!s32i>, !s32i // CIR: cir.store{{.*}} %[[TMP_2]], %[[INIT]] : !s32i, !cir.ptr<!s32i> -// LLVM: define{{.*}} void @_Z5func9PA5_i(ptr %[[ARG:.*]]) +// LLVM: define{{.*}} void @_Z5func9PA5_i(ptr %[[ARG:.*]]){{.*}} // LLVM: %[[ARR:.*]] = alloca ptr, i64 1, align 8 // LLVM: %[[INIT:.*]] = alloca i32, i64 1, align 4 // LLVM: store ptr %[[ARG]], ptr %[[ARR]], align 8 @@ -536,7 +536,7 @@ void func10(int *a) { // CIR: %[[TMP_2:.*]] = cir.load{{.*}} %[[ELE]] : !cir.ptr<!s32i>, !s32i // CIR: cir.store{{.*}} %[[TMP_2]], %[[INIT]] : !s32i, !cir.ptr<!s32i> -// LLVM: define{{.*}} void @_Z6func10Pi(ptr %[[ARG:.*]]) { +// LLVM: define{{.*}} void @_Z6func10Pi(ptr %[[ARG:.*]]){{.*}} { // LLVM: %[[ARR:.*]] = alloca ptr, i64 1, align 8 // LLVM: %[[INIT:.*]] = alloca i32, i64 1, align 4 // LLVM: store ptr %[[ARG]], ptr %[[ARR]], align 8 diff --git a/clang/test/CIR/CodeGen/assign-operator.cpp b/clang/test/CIR/CodeGen/assign-operator.cpp index 1089d4b..66d4b48 100644 --- a/clang/test/CIR/CodeGen/assign-operator.cpp +++ b/clang/test/CIR/CodeGen/assign-operator.cpp @@ -20,7 +20,7 @@ void a() { // CIR: %[[ONE_CAST:.*]] = cir.cast integral %[[ONE]] : !u32i -> !s32i // CIR: %[[RET:.*]] = cir.call @_ZN1xaSEi(%[[A_ADDR]], %[[ONE_CAST]]) : (!cir.ptr<!rec_x>, !s32i) -> !s32i -// LLVM: define{{.*}} @_Z1av() +// LLVM: define{{.*}} @_Z1av(){{.*}} // OGCG: define{{.*}} @_Z1av() void f(int i, int j) { @@ -121,7 +121,7 @@ void copy_ref_to_ref(E &e1, E &e2) { // CIR: %[[D1_REF_2:.*]] = cir.call @_ZN1DaSERKS_(%[[D1_REF]], %[[D2_REF]]) // CIR: cir.return -// LLVM: define{{.*}} void @_Z15copy_ref_to_refR1ES0_(ptr %[[ARG0:.*]], ptr %[[ARG1:.*]]) { +// LLVM: define{{.*}} void @_Z15copy_ref_to_refR1ES0_(ptr %[[ARG0:.*]], ptr %[[ARG1:.*]]){{.*}} { // LLVM: %[[E1_ADDR:.*]] = alloca ptr // LLVM: %[[E2_ADDR:.*]] = alloca ptr // LLVM: store ptr %[[ARG0]], ptr %[[E1_ADDR]] diff --git a/clang/test/CIR/CodeGen/atomic.c b/clang/test/CIR/CodeGen/atomic.c index 76289c5..440010a 100644 --- a/clang/test/CIR/CodeGen/atomic.c +++ b/clang/test/CIR/CodeGen/atomic.c @@ -211,7 +211,7 @@ void c11_atomic_cmpxchg_strong(_Atomic(int) *ptr, int *expected, int desired) { __c11_atomic_compare_exchange_strong(ptr, expected, desired, __ATOMIC_SEQ_CST, __ATOMIC_ACQUIRE); - // CIR: %[[OLD:.+]], %[[SUCCESS:.+]] = cir.atomic.cmpxchg(%{{.+}} : !cir.ptr<!s32i>, %{{.+}} : !s32i, %{{.+}} : !s32i, success = seq_cst, failure = acquire) align(4) : (!s32i, !cir.bool) + // CIR: %[[OLD:.+]], %[[SUCCESS:.+]] = cir.atomic.cmpxchg success(seq_cst) failure(acquire) %{{.+}}, %{{.+}}, %{{.+}} align(4) : (!cir.ptr<!s32i>, !s32i, !s32i) -> (!s32i, !cir.bool) // CIR-NEXT: %[[FAILED:.+]] = cir.unary(not, %[[SUCCESS]]) : !cir.bool, !cir.bool // CIR-NEXT: cir.if %[[FAILED]] { // CIR-NEXT: cir.store align(4) %[[OLD]], %{{.+}} : !s32i, !cir.ptr<!s32i> @@ -249,7 +249,7 @@ void c11_atomic_cmpxchg_weak(_Atomic(int) *ptr, int *expected, int desired) { __c11_atomic_compare_exchange_weak(ptr, expected, desired, __ATOMIC_SEQ_CST, __ATOMIC_ACQUIRE); - // CIR: %[[OLD:.+]], %[[SUCCESS:.+]] = cir.atomic.cmpxchg(%{{.+}} : !cir.ptr<!s32i>, %{{.+}} : !s32i, %{{.+}} : !s32i, success = seq_cst, failure = acquire) align(4) weak : (!s32i, !cir.bool) + // CIR: %[[OLD:.+]], %[[SUCCESS:.+]] = cir.atomic.cmpxchg weak success(seq_cst) failure(acquire) %{{.+}}, %{{.+}}, %{{.+}} align(4) : (!cir.ptr<!s32i>, !s32i, !s32i) -> (!s32i, !cir.bool) // CIR-NEXT: %[[FAILED:.+]] = cir.unary(not, %[[SUCCESS]]) : !cir.bool, !cir.bool // CIR-NEXT: cir.if %[[FAILED]] { // CIR-NEXT: cir.store align(4) %[[OLD]], %{{.+}} : !s32i, !cir.ptr<!s32i> @@ -286,7 +286,7 @@ void atomic_cmpxchg(int *ptr, int *expected, int *desired) { // OGCG-LABEL: @atomic_cmpxchg __atomic_compare_exchange(ptr, expected, desired, /*weak=*/0, __ATOMIC_SEQ_CST, __ATOMIC_ACQUIRE); - // CIR: %[[OLD:.+]], %[[SUCCESS:.+]] = cir.atomic.cmpxchg(%{{.+}} : !cir.ptr<!s32i>, %{{.+}} : !s32i, %{{.+}} : !s32i, success = seq_cst, failure = acquire) align(4) : (!s32i, !cir.bool) + // CIR: %[[OLD:.+]], %[[SUCCESS:.+]] = cir.atomic.cmpxchg success(seq_cst) failure(acquire) %{{.+}}, %{{.+}}, %{{.+}} align(4) : (!cir.ptr<!s32i>, !s32i, !s32i) -> (!s32i, !cir.bool) // CIR-NEXT: %[[FAILED:.+]] = cir.unary(not, %[[SUCCESS]]) : !cir.bool, !cir.bool // CIR-NEXT: cir.if %[[FAILED]] { // CIR-NEXT: cir.store align(4) %[[OLD]], %{{.+}} : !s32i, !cir.ptr<!s32i> @@ -317,7 +317,7 @@ void atomic_cmpxchg(int *ptr, int *expected, int *desired) { // OGCG-NEXT: store i8 %[[SUCCESS_2]], ptr %{{.+}}, align 1 __atomic_compare_exchange(ptr, expected, desired, /*weak=*/1, __ATOMIC_SEQ_CST, __ATOMIC_ACQUIRE); - // CIR: %[[OLD:.+]], %[[SUCCESS:.+]] = cir.atomic.cmpxchg(%{{.+}} : !cir.ptr<!s32i>, %{{.+}} : !s32i, %{{.+}} : !s32i, success = seq_cst, failure = acquire) align(4) weak : (!s32i, !cir.bool) + // CIR: %[[OLD:.+]], %[[SUCCESS:.+]] = cir.atomic.cmpxchg weak success(seq_cst) failure(acquire) %{{.+}}, %{{.+}}, %{{.+}} align(4) : (!cir.ptr<!s32i>, !s32i, !s32i) -> (!s32i, !cir.bool) // CIR-NEXT: %[[FAILED:.+]] = cir.unary(not, %[[SUCCESS]]) : !cir.bool, !cir.bool // CIR-NEXT: cir.if %[[FAILED]] { // CIR-NEXT: cir.store align(4) %[[OLD]], %{{.+}} : !s32i, !cir.ptr<!s32i> @@ -354,7 +354,7 @@ void atomic_cmpxchg_n(int *ptr, int *expected, int desired) { // OGCG-LABEL: @atomic_cmpxchg_n __atomic_compare_exchange_n(ptr, expected, desired, /*weak=*/0, __ATOMIC_SEQ_CST, __ATOMIC_ACQUIRE); - // CIR: %[[OLD:.+]], %[[SUCCESS:.+]] = cir.atomic.cmpxchg(%{{.+}} : !cir.ptr<!s32i>, %{{.+}} : !s32i, %{{.+}} : !s32i, success = seq_cst, failure = acquire) align(4) : (!s32i, !cir.bool) + // CIR: %[[OLD:.+]], %[[SUCCESS:.+]] = cir.atomic.cmpxchg success(seq_cst) failure(acquire) %{{.+}}, %{{.+}}, %{{.+}} align(4) : (!cir.ptr<!s32i>, !s32i, !s32i) -> (!s32i, !cir.bool) // CIR-NEXT: %[[FAILED:.+]] = cir.unary(not, %[[SUCCESS]]) : !cir.bool, !cir.bool // CIR-NEXT: cir.if %[[FAILED]] { // CIR-NEXT: cir.store align(4) %[[OLD]], %{{.+}} : !s32i, !cir.ptr<!s32i> @@ -385,7 +385,7 @@ void atomic_cmpxchg_n(int *ptr, int *expected, int desired) { // OGCG-NEXT: store i8 %[[SUCCESS_2]], ptr %{{.+}}, align 1 __atomic_compare_exchange_n(ptr, expected, desired, /*weak=*/1, __ATOMIC_SEQ_CST, __ATOMIC_ACQUIRE); - // CIR: %[[OLD:.+]], %[[SUCCESS:.+]] = cir.atomic.cmpxchg(%{{.+}} : !cir.ptr<!s32i>, %{{.+}} : !s32i, %{{.+}} : !s32i, success = seq_cst, failure = acquire) align(4) weak : (!s32i, !cir.bool) + // CIR: %[[OLD:.+]], %[[SUCCESS:.+]] = cir.atomic.cmpxchg weak success(seq_cst) failure(acquire) %{{.+}}, %{{.+}}, %{{.+}} align(4) : (!cir.ptr<!s32i>, !s32i, !s32i) -> (!s32i, !cir.bool) // CIR-NEXT: %[[FAILED:.+]] = cir.unary(not, %[[SUCCESS]]) : !cir.bool, !cir.bool // CIR-NEXT: cir.if %[[FAILED]] { // CIR-NEXT: cir.store align(4) %[[OLD]], %{{.+}} : !s32i, !cir.ptr<!s32i> @@ -427,12 +427,12 @@ void c11_atomic_exchange(_Atomic(int) *ptr, int value) { __c11_atomic_exchange(ptr, value, __ATOMIC_RELEASE); __c11_atomic_exchange(ptr, value, __ATOMIC_ACQ_REL); __c11_atomic_exchange(ptr, value, __ATOMIC_SEQ_CST); - // CIR: %{{.+}} = cir.atomic.xchg relaxed %{{.+}}, %{{.+}} : !cir.ptr<!s32i> -> !s32i - // CIR: %{{.+}} = cir.atomic.xchg consume %{{.+}}, %{{.+}} : !cir.ptr<!s32i> -> !s32i - // CIR: %{{.+}} = cir.atomic.xchg acquire %{{.+}}, %{{.+}} : !cir.ptr<!s32i> -> !s32i - // CIR: %{{.+}} = cir.atomic.xchg release %{{.+}}, %{{.+}} : !cir.ptr<!s32i> -> !s32i - // CIR: %{{.+}} = cir.atomic.xchg acq_rel %{{.+}}, %{{.+}} : !cir.ptr<!s32i> -> !s32i - // CIR: %{{.+}} = cir.atomic.xchg seq_cst %{{.+}}, %{{.+}} : !cir.ptr<!s32i> -> !s32i + // CIR: %{{.+}} = cir.atomic.xchg relaxed %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i) -> !s32i + // CIR: %{{.+}} = cir.atomic.xchg consume %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i) -> !s32i + // CIR: %{{.+}} = cir.atomic.xchg acquire %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i) -> !s32i + // CIR: %{{.+}} = cir.atomic.xchg release %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i) -> !s32i + // CIR: %{{.+}} = cir.atomic.xchg acq_rel %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i) -> !s32i + // CIR: %{{.+}} = cir.atomic.xchg seq_cst %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i) -> !s32i // LLVM: %{{.+}} = atomicrmw xchg ptr %{{.+}}, i32 %{{.+}} monotonic, align 4 // LLVM: %{{.+}} = atomicrmw xchg ptr %{{.+}}, i32 %{{.+}} acquire, align 4 @@ -460,12 +460,12 @@ void atomic_exchange(int *ptr, int *value, int *old) { __atomic_exchange(ptr, value, old, __ATOMIC_RELEASE); __atomic_exchange(ptr, value, old, __ATOMIC_ACQ_REL); __atomic_exchange(ptr, value, old, __ATOMIC_SEQ_CST); - // CIR: %{{.+}} = cir.atomic.xchg relaxed %{{.+}}, %{{.+}} : !cir.ptr<!s32i> -> !s32i - // CIR: %{{.+}} = cir.atomic.xchg consume %{{.+}}, %{{.+}} : !cir.ptr<!s32i> -> !s32i - // CIR: %{{.+}} = cir.atomic.xchg acquire %{{.+}}, %{{.+}} : !cir.ptr<!s32i> -> !s32i - // CIR: %{{.+}} = cir.atomic.xchg release %{{.+}}, %{{.+}} : !cir.ptr<!s32i> -> !s32i - // CIR: %{{.+}} = cir.atomic.xchg acq_rel %{{.+}}, %{{.+}} : !cir.ptr<!s32i> -> !s32i - // CIR: %{{.+}} = cir.atomic.xchg seq_cst %{{.+}}, %{{.+}} : !cir.ptr<!s32i> -> !s32i + // CIR: %{{.+}} = cir.atomic.xchg relaxed %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i) -> !s32i + // CIR: %{{.+}} = cir.atomic.xchg consume %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i) -> !s32i + // CIR: %{{.+}} = cir.atomic.xchg acquire %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i) -> !s32i + // CIR: %{{.+}} = cir.atomic.xchg release %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i) -> !s32i + // CIR: %{{.+}} = cir.atomic.xchg acq_rel %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i) -> !s32i + // CIR: %{{.+}} = cir.atomic.xchg seq_cst %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i) -> !s32i // LLVM: %{{.+}} = atomicrmw xchg ptr %{{.+}}, i32 %{{.+}} monotonic, align 4 // LLVM: %{{.+}} = atomicrmw xchg ptr %{{.+}}, i32 %{{.+}} acquire, align 4 @@ -493,12 +493,12 @@ void atomic_exchange_n(int *ptr, int value) { __atomic_exchange_n(ptr, value, __ATOMIC_RELEASE); __atomic_exchange_n(ptr, value, __ATOMIC_ACQ_REL); __atomic_exchange_n(ptr, value, __ATOMIC_SEQ_CST); - // CIR: %{{.+}} = cir.atomic.xchg relaxed %{{.+}}, %{{.+}} : !cir.ptr<!s32i> -> !s32i - // CIR: %{{.+}} = cir.atomic.xchg consume %{{.+}}, %{{.+}} : !cir.ptr<!s32i> -> !s32i - // CIR: %{{.+}} = cir.atomic.xchg acquire %{{.+}}, %{{.+}} : !cir.ptr<!s32i> -> !s32i - // CIR: %{{.+}} = cir.atomic.xchg release %{{.+}}, %{{.+}} : !cir.ptr<!s32i> -> !s32i - // CIR: %{{.+}} = cir.atomic.xchg acq_rel %{{.+}}, %{{.+}} : !cir.ptr<!s32i> -> !s32i - // CIR: %{{.+}} = cir.atomic.xchg seq_cst %{{.+}}, %{{.+}} : !cir.ptr<!s32i> -> !s32i + // CIR: %{{.+}} = cir.atomic.xchg relaxed %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i) -> !s32i + // CIR: %{{.+}} = cir.atomic.xchg consume %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i) -> !s32i + // CIR: %{{.+}} = cir.atomic.xchg acquire %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i) -> !s32i + // CIR: %{{.+}} = cir.atomic.xchg release %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i) -> !s32i + // CIR: %{{.+}} = cir.atomic.xchg acq_rel %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i) -> !s32i + // CIR: %{{.+}} = cir.atomic.xchg seq_cst %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i) -> !s32i // LLVM: %{{.+}} = atomicrmw xchg ptr %{{.+}}, i32 %{{.+}} monotonic, align 4 // LLVM: %{{.+}} = atomicrmw xchg ptr %{{.+}}, i32 %{{.+}} acquire, align 4 diff --git a/clang/test/CIR/CodeGen/binassign.c b/clang/test/CIR/CodeGen/binassign.c index 65bea4d..dab9879 100644 --- a/clang/test/CIR/CodeGen/binassign.c +++ b/clang/test/CIR/CodeGen/binassign.c @@ -17,7 +17,7 @@ void binary_assign(void) { i = 42; } -// CIR-LABEL: cir.func{{.*}} @binary_assign() { +// CIR-LABEL: cir.func{{.*}} @binary_assign() // CIR: %[[B:.*]] = cir.alloca !cir.bool, !cir.ptr<!cir.bool>, ["b"] // CIR: %[[C:.*]] = cir.alloca !s8i, !cir.ptr<!s8i>, ["c"] // CIR: %[[F:.*]] = cir.alloca !cir.float, !cir.ptr<!cir.float>, ["f"] @@ -33,7 +33,7 @@ void binary_assign(void) { // CIR: cir.store{{.*}} %[[INT_VAL]], %[[I]] : !s32i, !cir.ptr<!s32i> // CIR: cir.return -// LLVM-LABEL: define {{.*}}void @binary_assign() { +// LLVM-LABEL: define {{.*}}void @binary_assign(){{.*}} { // LLVM: %[[B_PTR:.*]] = alloca i8 // LLVM: %[[C_PTR:.*]] = alloca i8 // LLVM: %[[F_PTR:.*]] = alloca float diff --git a/clang/test/CIR/CodeGen/bitfields_be.c b/clang/test/CIR/CodeGen/bitfields_be.c index 77741ba..3e1f054 100644 --- a/clang/test/CIR/CodeGen/bitfields_be.c +++ b/clang/test/CIR/CodeGen/bitfields_be.c @@ -27,7 +27,7 @@ int init(S* s) { //CIR: [[TMP2:%.*]] = cir.get_member [[TMP1]][0] {name = "c"} : !cir.ptr<!rec_S> -> !cir.ptr<!u32i> //CIR: [[TMP3:%.*]] = cir.get_bitfield align(4) (#bfi_c, [[TMP2]] : !cir.ptr<!u32i>) -> !s32i -//LLVM: define dso_local i32 @init(ptr %0) { +//LLVM: define dso_local i32 @init(ptr %0){{.*}} { //LLVM: [[TMP0:%.*]] = alloca ptr, i64 1, align 8 //LLVM: [[TMP1:%.*]] = alloca i32, i64 1, align 4 //LLVM: [[TMP2:%.*]] = load ptr, ptr [[TMP0]], align 8 @@ -59,7 +59,7 @@ void load(S* s) { // CIR: %[[GET0:.*]] = cir.get_member %[[VAL0]][0] {name = "a"} : !cir.ptr<!rec_S> -> !cir.ptr<!u32i> // CIR: %[[SET0:.*]] = cir.set_bitfield align(4) (#bfi_a, %[[GET0]] : !cir.ptr<!u32i>, %[[MIN1]] : !s32i) -> !s32i -// LLVM: define dso_local void @load +// LLVM: define dso_local void @load{{.*}}{{.*}} // LLVM: %[[PTR0:.*]] = load ptr // LLVM: %[[GET0:.*]] = getelementptr %struct.S, ptr %[[PTR0]], i32 0, i32 0 // LLVM: %[[VAL0:.*]] = load i32, ptr %[[GET0]], align 4 diff --git a/clang/test/CIR/CodeGen/builtin_call.cpp b/clang/test/CIR/CodeGen/builtin_call.cpp index a30df97..a08a784 100644 --- a/clang/test/CIR/CodeGen/builtin_call.cpp +++ b/clang/test/CIR/CodeGen/builtin_call.cpp @@ -82,7 +82,7 @@ void library_builtins() { __builtin_abort(); } -// CIR: cir.func{{.*}} @_Z16library_builtinsv() { +// CIR: cir.func{{.*}} @_Z16library_builtinsv() // CIR: %[[NULL:.+]] = cir.const #cir.ptr<null> : !cir.ptr<!s8i> // CIR: cir.call @printf(%[[NULL]]) nothrow : (!cir.ptr<!s8i>) -> !s32i // CIR: cir.call @abort() nothrow : () -> () diff --git a/clang/test/CIR/CodeGen/builtin_printf.cpp b/clang/test/CIR/CodeGen/builtin_printf.cpp index 898984a..7200df1 100644 --- a/clang/test/CIR/CodeGen/builtin_printf.cpp +++ b/clang/test/CIR/CodeGen/builtin_printf.cpp @@ -20,7 +20,7 @@ void func(char const * const str, int i) { // CIR: cir.func{{.*}} @printf(!cir.ptr<!s8i>, ...) -> !s32i -// CIR: cir.func{{.*}} @_Z4funcPKci(%[[arg0:.+]]: !cir.ptr<!s8i>{{.*}}, %[[arg1:.+]]: !s32i{{.*}}) { +// CIR: cir.func{{.*}} @_Z4funcPKci(%[[arg0:.+]]: !cir.ptr<!s8i>{{.*}}, %[[arg1:.+]]: !s32i // CIR: %[[str_ptr:.+]] = cir.alloca !cir.ptr<!s8i>, !cir.ptr<!cir.ptr<!s8i>>, ["str", init, const] // CIR: %[[i_ptr:.+]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["i", init] // CIR: cir.store %[[arg0]], %[[str_ptr]] : !cir.ptr<!s8i>, !cir.ptr<!cir.ptr<!s8i>> diff --git a/clang/test/CIR/CodeGen/call.c b/clang/test/CIR/CodeGen/call.c index 9d516c6..d780e37 100644 --- a/clang/test/CIR/CodeGen/call.c +++ b/clang/test/CIR/CodeGen/call.c @@ -16,11 +16,11 @@ void f2(void) { f1(s); } -// CIR-LABEL: cir.func{{.*}} @f2() +// CIR-LABEL: cir.func{{.*}} @f2(){{.*}} { // CIR: %[[S:.+]] = cir.load align(4) %{{.+}} : !cir.ptr<!rec_S>, !rec_S // CIR-NEXT: cir.call @f1(%[[S]]) : (!rec_S) -> () -// LLVM-LABEL: define{{.*}} void @f2() +// LLVM-LABEL: define{{.*}} void @f2(){{.*}} // LLVM: %[[S:.+]] = load %struct.S, ptr %{{.+}}, align 4 // LLVM-NEXT: call void @f1(%struct.S %[[S]]) @@ -33,11 +33,11 @@ void f4(void) { struct S s = f3(); } -// CIR-LABEL: cir.func{{.*}} @f4() { +// CIR-LABEL: cir.func{{.*}} @f4(){{.*}} { // CIR: %[[S:.+]] = cir.call @f3() : () -> !rec_S // CIR-NEXT: cir.store align(4) %[[S]], %{{.+}} : !rec_S, !cir.ptr<!rec_S> -// LLVM-LABEL: define{{.*}} void @f4() { +// LLVM-LABEL: define{{.*}} void @f4(){{.*}} { // LLVM: %[[S:.+]] = call %struct.S @f3() // LLVM-NEXT: store %struct.S %[[S]], ptr %{{.+}}, align 4 @@ -57,11 +57,11 @@ void f7(void) { f5(b); } -// CIR-LABEL: cir.func{{.*}} @f7() +// CIR-LABEL: cir.func{{.*}} @f7(){{.*}} { // CIR: %[[B:.+]] = cir.load align(4) %{{.+}} : !cir.ptr<!rec_Big>, !rec_Big // CIR-NEXT: cir.call @f5(%[[B]]) : (!rec_Big) -> () -// LLVM-LABEL: define{{.*}} void @f7() { +// LLVM-LABEL: define{{.*}} void @f7(){{.*}} { // LLVM: %[[B:.+]] = load %struct.Big, ptr %{{.+}}, align 4 // LLVM-NEXT: call void @f5(%struct.Big %[[B]]) @@ -73,11 +73,11 @@ void f8(void) { struct Big b = f6(); } -// CIR-LABEL: cir.func{{.*}} @f8() +// CIR-LABEL: cir.func{{.*}} @f8(){{.*}} { // CIR: %[[B:.+]] = cir.call @f6() : () -> !rec_Big // CIR: cir.store align(4) %[[B]], %{{.+}} : !rec_Big, !cir.ptr<!rec_Big> -// LLVM-LABEL: define{{.*}} void @f8() { +// LLVM-LABEL: define{{.*}} void @f8(){{.*}} { // LLVM: %[[B:.+]] = call %struct.Big @f6() // LLVM-NEXT: store %struct.Big %[[B]], ptr %{{.+}}, align 4 @@ -89,14 +89,14 @@ void f9(void) { f1(f3()); } -// CIR-LABEL: cir.func{{.*}} @f9() +// CIR-LABEL: cir.func{{.*}} @f9(){{.*}} { // CIR: %[[SLOT:.+]] = cir.alloca !rec_S, !cir.ptr<!rec_S>, ["agg.tmp0"] {alignment = 4 : i64} // CIR-NEXT: %[[RET:.+]] = cir.call @f3() : () -> !rec_S // CIR-NEXT: cir.store align(4) %[[RET]], %[[SLOT]] : !rec_S, !cir.ptr<!rec_S> // CIR-NEXT: %[[ARG:.+]] = cir.load align(4) %[[SLOT]] : !cir.ptr<!rec_S>, !rec_S // CIR-NEXT: cir.call @f1(%[[ARG]]) : (!rec_S) -> () -// LLVM-LABEL: define{{.*}} void @f9() { +// LLVM-LABEL: define{{.*}} void @f9(){{.*}} { // LLVM: %[[SLOT:.+]] = alloca %struct.S, i64 1, align 4 // LLVM-NEXT: %[[RET:.+]] = call %struct.S @f3() // LLVM-NEXT: store %struct.S %[[RET]], ptr %[[SLOT]], align 4 @@ -116,13 +116,13 @@ int f12(void) { return f10(1) + f11(2); } -// CIR-LABEL: cir.func{{.*}} @f12() -> !s32i +// CIR-LABEL: cir.func{{.*}} @f12() -> !s32i{{.*}} { // CIR: %[[A:.+]] = cir.const #cir.int<1> : !s32i // CIR-NEXT: %{{.+}} = cir.call @f10(%[[A]]) side_effect(pure) : (!s32i) -> !s32i // CIR-NEXT: %[[B:.+]] = cir.const #cir.int<2> : !s32i // CIR-NEXT: %{{.+}} = cir.call @f11(%[[B]]) side_effect(const) : (!s32i) -> !s32i -// LLVM-LABEL: define{{.*}} i32 @f12() +// LLVM-LABEL: define{{.*}} i32 @f12(){{.*}} // LLVM: %{{.+}} = call i32 @f10(i32 1) #[[ATTR0:.+]] // LLVM-NEXT: %{{.+}} = call i32 @f11(i32 2) #[[ATTR1:.+]] diff --git a/clang/test/CIR/CodeGen/call.cpp b/clang/test/CIR/CodeGen/call.cpp index 3e8cfc1..affa8af 100644 --- a/clang/test/CIR/CodeGen/call.cpp +++ b/clang/test/CIR/CodeGen/call.cpp @@ -12,7 +12,7 @@ void f2() { // CIR-LABEL: cir.func{{.*}} @_Z2f2v // CIR: cir.call @_Z2f1v() : () -> () -// LLVM-LABEL: define{{.*}} void @_Z2f2v() { +// LLVM-LABEL: define{{.*}} void @_Z2f2v(){{.*}} { // LLVM: call void @_Z2f1v() int f3() { return 2; } @@ -25,7 +25,7 @@ int f4() { // CIR-LABEL: cir.func{{.*}} @_Z2f4v() -> !s32i // CIR: cir.call @_Z2f3v() : () -> !s32i -// LLVM-LABEL: define{{.*}} i32 @_Z2f4v() { +// LLVM-LABEL: define{{.*}} i32 @_Z2f4v(){{.*}} { // LLVM: %{{.+}} = call i32 @_Z2f3v() int f5(int a, int *b, bool c); @@ -40,7 +40,7 @@ int f6() { // CIR-NEXT: %[[#c:]] = cir.const #false // CIR-NEXT: %{{.+}} = cir.call @_Z2f5iPib(%[[#a]], %[[#b:]], %[[#c]]) : (!s32i, !cir.ptr<!s32i>, !cir.bool) -> !s32i -// LLVM-LABEL: define{{.*}} i32 @_Z2f6v() { +// LLVM-LABEL: define{{.*}} i32 @_Z2f6v(){{.*}} { // LLVM: %{{.+}} = call i32 @_Z2f5iPib(i32 2, ptr %{{.+}}, i1 false) int f7(int (*ptr)(int, int)) { @@ -67,7 +67,7 @@ void f9() { // CIR: cir.call @_Z2f8iz(%{{.+}}) : (!s32i) -> () // CIR: cir.call @_Z2f8iz(%{{.+}}, %{{.+}}, %{{.+}}, %{{.+}}) : (!s32i, !s32i, !s32i, !s32i) -> () -// LLVM-LABEL: define{{.*}} void @_Z2f9v() +// LLVM-LABEL: define{{.*}} void @_Z2f9v(){{.*}} // LLVM: call void (i32, ...) @_Z2f8iz(i32 1) // LLVM: call void (i32, ...) @_Z2f8iz(i32 1, i32 2, i32 3, i32 4) @@ -85,7 +85,7 @@ void f11() { // CIR: %[[#s:]] = cir.call @_Z3f10v() : () -> !rec_S // CIR-NEXT: cir.store align(4) %[[#s]], %{{.+}} : !rec_S, !cir.ptr<!rec_S> -// LLVM-LABEL: define{{.*}} void @_Z3f11v() +// LLVM-LABEL: define{{.*}} void @_Z3f11v(){{.*}} // LLVM: %[[#s:]] = call %struct.S @_Z3f10v() // LLVM-NEXT: store %struct.S %[[#s]], ptr %{{.+}}, align 4 @@ -98,7 +98,7 @@ void f12() { // CIR-NEXT: %[[#ret:]] = cir.call @_Z3f10v() : () -> !rec_S // CIR-NEXT: cir.store align(4) %[[#ret]], %[[#slot]] : !rec_S, !cir.ptr<!rec_S> -// LLVM-LABEL: define{{.*}} void @_Z3f12v() { +// LLVM-LABEL: define{{.*}} void @_Z3f12v(){{.*}} { // LLVM: %[[#slot:]] = alloca %struct.S, i64 1, align 4 // LLVM-NEXT: %[[#ret:]] = call %struct.S @_Z3f10v() // LLVM-NEXT: store %struct.S %[[#ret]], ptr %[[#slot]], align 4 @@ -112,7 +112,7 @@ void f14() { // CIR: cir.call @_Z3f13v() nothrow : () -> () // CIR: } -// LLVM-LABEL: define{{.+}} void @_Z3f14v() +// LLVM-LABEL: define{{.+}} void @_Z3f14v(){{.*}} // LLVM: call void @_Z3f13v() #[[LLVM_ATTR_0:.+]] // LLVM: } @@ -126,7 +126,7 @@ void f16() { // CIR-NEXT: %{{.+}} = cir.call @_Z3f15v() : () -> !s32i // CIR: } -// LLVM-LABEL: define{{.+}} void @_Z3f16v() { +// LLVM-LABEL: define{{.+}} void @_Z3f16v(){{.*}} { // LLVM-NEXT: %{{.+}} = call i32 @_Z3f15v() // LLVM: } diff --git a/clang/test/CIR/CodeGen/cmp.cpp b/clang/test/CIR/CodeGen/cmp.cpp index 7e32d16..1871f94 100644 --- a/clang/test/CIR/CodeGen/cmp.cpp +++ b/clang/test/CIR/CodeGen/cmp.cpp @@ -45,7 +45,7 @@ void c0(int a, int b) { // CIR: %[[B6:.*]] = cir.load{{.*}} %[[B_PTR]] // CIR: %{{.*}} = cir.cmp(eq, %[[A6]], %[[B6]]) : !s32i, !cir.bool -// LLVM-LABEL: define{{.*}} void @_Z2c0ii(i32 %0, i32 %1) { +// LLVM-LABEL: define{{.*}} void @_Z2c0ii(i32 %0, i32 %1){{.*}} { // LLVM: %[[PTR1:.*]] = alloca i32, i64 1 // LLVM: %[[PTR2:.*]] = alloca i32, i64 1 // LLVM: %[[BOOL_PTR:.*]] = alloca i8, i64 1 @@ -170,7 +170,7 @@ void c0_unsigned(unsigned int a, unsigned int b) { // CIR: %[[UB6:.*]] = cir.load{{.*}} %[[U_B_PTR]] // CIR: %{{.*}} = cir.cmp(eq, %[[UA6]], %[[UB6]]) : !u32i, !cir.bool -// LLVM-LABEL: define{{.*}} void @_Z11c0_unsignedjj(i32 %0, i32 %1) { +// LLVM-LABEL: define{{.*}} void @_Z11c0_unsignedjj(i32 %0, i32 %1){{.*}} { // LLVM: %[[U_PTR1:.*]] = alloca i32, i64 1 // LLVM: %[[U_PTR2:.*]] = alloca i32, i64 1 // LLVM: %[[U_BOOL_PTR:.*]] = alloca i8, i64 1 @@ -265,7 +265,7 @@ void c0_float(float a, float b) { x = a == b; } -// CIR-LABEL: cir.func{{.*}} @_Z8c0_floatff(%arg0: !cir.float{{.*}}, %arg1: !cir.float{{.*}}) { +// CIR-LABEL: cir.func{{.*}} @_Z8c0_floatff(%arg0: !cir.float{{.*}}, %arg1: !cir.float{{.*}}) // CIR: %[[A_PTR:.*]] = cir.alloca !cir.float, !cir.ptr<!cir.float>, ["a", init] // CIR: %[[B_PTR:.*]] = cir.alloca !cir.float, !cir.ptr<!cir.float>, ["b", init] // CIR: %[[X_PTR:.*]] = cir.alloca !cir.bool, !cir.ptr<!cir.bool>, ["x", init] @@ -303,7 +303,7 @@ void c0_float(float a, float b) { // CIR: %[[CMP6:.*]] = cir.cmp(eq, %[[A6]], %[[B6]]) : !cir.float, !cir.bool // CIR: cir.store{{.*}} %[[CMP6]], %[[X_PTR]] : !cir.bool, !cir.ptr<!cir.bool> -// LLVM-LABEL: define{{.*}} void @_Z8c0_floatff(float %0, float %1) { +// LLVM-LABEL: define{{.*}} void @_Z8c0_floatff(float %0, float %1){{.*}} { // LLVM: %[[A_PTR:.*]] = alloca float // LLVM: %[[B_PTR:.*]] = alloca float // LLVM: store float %0, ptr %[[A_PTR]] @@ -346,7 +346,7 @@ void pointer_cmp(int *a, int *b) { x = a != b; } -// CIR-LABEL: cir.func{{.*}} @_Z11pointer_cmpPiS_(%arg0: !cir.ptr<!s32i>{{.*}}, %arg1: !cir.ptr<!s32i>{{.*}}) { +// CIR-LABEL: cir.func{{.*}} @_Z11pointer_cmpPiS_(%arg0: !cir.ptr<!s32i>{{.*}}, %arg1: !cir.ptr<!s32i>{{.*}}){{.*}} { // CIR: %[[A_PTR:.*]] = cir.alloca !cir.ptr<!s32i>, !cir.ptr<!cir.ptr<!s32i>>, ["a", init] // CIR: %[[B_PTR:.*]] = cir.alloca !cir.ptr<!s32i>, !cir.ptr<!cir.ptr<!s32i>>, ["b", init] @@ -360,7 +360,7 @@ void pointer_cmp(int *a, int *b) { // CIR: cir.cmp(eq, {{.*}}, {{.*}}) : !cir.ptr<!s32i>, !cir.bool // CIR: cir.cmp(ne, {{.*}}, {{.*}}) : !cir.ptr<!s32i>, !cir.bool -// LLVM-LABEL: define{{.*}} void @_Z11pointer_cmpPiS_(ptr %0, ptr %1) { +// LLVM-LABEL: define{{.*}} void @_Z11pointer_cmpPiS_(ptr %0, ptr %1){{.*}} { // LLVM: %[[A_PTR:.*]] = alloca ptr // LLVM: %[[B_PTR:.*]] = alloca ptr // LLVM: store ptr %0, ptr %[[A_PTR]] @@ -401,7 +401,7 @@ void bool_cmp(bool a, bool b) { x = a != b; } -// CIR-LABEL: cir.func{{.*}} @_Z8bool_cmpbb(%arg0: !cir.bool{{.*}}, %arg1: !cir.bool{{.*}}) { +// CIR-LABEL: cir.func{{.*}} @_Z8bool_cmpbb(%arg0: !cir.bool{{.*}}, %arg1: !cir.bool{{.*}}){{.*}} { // CIR: %[[A_PTR:.*]] = cir.alloca !cir.bool, !cir.ptr<!cir.bool>, ["a", init] // CIR: %[[B_PTR:.*]] = cir.alloca !cir.bool, !cir.ptr<!cir.bool>, ["b", init] // CIR: %[[X_PTR:.*]] = cir.alloca !cir.bool, !cir.ptr<!cir.bool>, ["x", init] @@ -419,7 +419,7 @@ void bool_cmp(bool a, bool b) { // CIR: cir.cmp(eq // CIR: cir.cmp(ne -// LLVM-LABEL: define{{.*}} void @_Z8bool_cmpbb(i1 %0, i1 %1) { +// LLVM-LABEL: define{{.*}} void @_Z8bool_cmpbb(i1 %0, i1 %1){{.*}} { // LLVM: %[[A_PTR:.*]] = alloca i8 // LLVM: %[[B_PTR:.*]] = alloca i8 // LLVM: %[[X_PTR:.*]] = alloca i8 diff --git a/clang/test/CIR/CodeGen/comma.c b/clang/test/CIR/CodeGen/comma.c index cc26a3f..c0bc442 100644 --- a/clang/test/CIR/CodeGen/comma.c +++ b/clang/test/CIR/CodeGen/comma.c @@ -16,7 +16,7 @@ void comma(void) { i = 100, 200; } -// CIR-LABEL: cir.func{{.*}} @comma() { +// CIR-LABEL: cir.func{{.*}} @comma() // CIR: %[[B:.*]] = cir.alloca !cir.bool, !cir.ptr<!cir.bool>, ["b"] // CIR: %[[C:.*]] = cir.alloca !s8i, !cir.ptr<!s8i>, ["c"] // CIR: %[[F:.*]] = cir.alloca !cir.float, !cir.ptr<!cir.float>, ["f"] @@ -34,7 +34,7 @@ void comma(void) { // CIR: cir.store{{.*}} %[[HUNDRED]], %[[I]] : !s32i, !cir.ptr<!s32i> // CIR: cir.return -// LLVM-LABEL: define {{.*}}void @comma() { +// LLVM-LABEL: define {{.*}}void @comma(){{.*}} { // LLVM: %[[B_PTR:.*]] = alloca i8 // LLVM: %[[C_PTR:.*]] = alloca i8 // LLVM: %[[F_PTR:.*]] = alloca float diff --git a/clang/test/CIR/CodeGen/ctor.cpp b/clang/test/CIR/CodeGen/ctor.cpp index 2b06bb0..238e415 100644 --- a/clang/test/CIR/CodeGen/ctor.cpp +++ b/clang/test/CIR/CodeGen/ctor.cpp @@ -49,7 +49,7 @@ void bar() { // CHECK: cir.func{{.*}} @_ZN13VariadicStrukC1Eiz(%arg0: !cir.ptr<!rec_VariadicStruk> // CHECK-SAME: %arg1: !s32i -// CHECK-SAME: ...) { +// CHECK-SAME: ...){{.*}} { // CHECK-NEXT: %[[THIS_ADDR:.*]] = cir.alloca {{.*}} ["this", init] // CHECK-NEXT: %[[N_ADDR:.*]] = cir.alloca {{.*}} ["n", init] // CHECK-NEXT: cir.store %arg0, %[[THIS_ADDR]] diff --git a/clang/test/CIR/CodeGen/dtors.cpp b/clang/test/CIR/CodeGen/dtors.cpp index 7fb0975..cb3886b 100644 --- a/clang/test/CIR/CodeGen/dtors.cpp +++ b/clang/test/CIR/CodeGen/dtors.cpp @@ -17,7 +17,7 @@ void test_temporary_dtor() { // CIR: %[[ALLOCA:.*]] = cir.alloca !rec_A, !cir.ptr<!rec_A>, ["agg.tmp0"] // CIR: cir.call @_ZN1AD1Ev(%[[ALLOCA]]) nothrow : (!cir.ptr<!rec_A>) -> () -// LLVM: define dso_local void @_Z19test_temporary_dtorv() +// LLVM: define dso_local void @_Z19test_temporary_dtorv(){{.*}} // LLVM: %[[ALLOCA:.*]] = alloca %struct.A, i64 1, align 1 // LLVM: call void @_ZN1AD1Ev(ptr %[[ALLOCA]]) @@ -55,7 +55,7 @@ bool test_temp_or() { return make_temp(1) || make_temp(2); } // CIR: cir.yield %[[TERNARY]] : !cir.bool // CIR: } : !cir.bool -// LLVM: define{{.*}} i1 @_Z12test_temp_orv() { +// LLVM: define{{.*}} i1 @_Z12test_temp_orv(){{.*}} { // LLVM: %[[REF_TMP0:.*]] = alloca %struct.B // LLVM: %[[REF_TMP1:.*]] = alloca %struct.B // LLVM: br label %[[LOR_BEGIN:.*]] @@ -125,7 +125,7 @@ bool test_temp_and() { return make_temp(1) && make_temp(2); } // CIR: cir.yield %[[TERNARY]] : !cir.bool // CIR: } : !cir.bool -// LLVM: define{{.*}} i1 @_Z13test_temp_andv() { +// LLVM: define{{.*}} i1 @_Z13test_temp_andv(){{.*}} { // LLVM: %[[REF_TMP0:.*]] = alloca %struct.B // LLVM: %[[REF_TMP1:.*]] = alloca %struct.B // LLVM: br label %[[LAND_BEGIN:.*]] @@ -199,7 +199,7 @@ void test_nested_dtor() { // CIR: cir.func{{.*}} @_Z16test_nested_dtorv() // CIR: cir.call @_ZN1DD2Ev(%{{.*}}) -// LLVM: define {{.*}} void @_Z16test_nested_dtorv() +// LLVM: define {{.*}} void @_Z16test_nested_dtorv(){{.*}} // LLVM: call void @_ZN1DD2Ev(ptr %{{.*}}) // OGCG: define {{.*}} void @_Z16test_nested_dtorv() @@ -236,7 +236,7 @@ void test_base_dtor_call() { // CIR: cir.func {{.*}} @_Z19test_base_dtor_callv() // cir.call @_ZN1FD2Ev(%{{.*}}) nothrow : (!cir.ptr<!rec_F>) -> () -// LLVM: define {{.*}} void @_Z19test_base_dtor_callv() +// LLVM: define {{.*}} void @_Z19test_base_dtor_callv(){{.*}} // LLVM: call void @_ZN1FD2Ev(ptr %{{.*}}) // OGCG: define {{.*}} void @_Z19test_base_dtor_callv() diff --git a/clang/test/CIR/CodeGen/dynamic-cast.cpp b/clang/test/CIR/CodeGen/dynamic-cast.cpp index b493840..5d010d2 100644 --- a/clang/test/CIR/CodeGen/dynamic-cast.cpp +++ b/clang/test/CIR/CodeGen/dynamic-cast.cpp @@ -101,3 +101,59 @@ Derived &ref_cast(Base &b) { // OGCG: br i1 %[[IS_NULL]], label %[[BAD_CAST:.*]], label %[[DONE:.*]] // OGCG: [[BAD_CAST]]: // OGCG: call void @__cxa_bad_cast() + +void *ptr_cast_to_complete(Base *ptr) { + return dynamic_cast<void *>(ptr); +} + +// CIR-BEFORE: cir.func dso_local @_Z20ptr_cast_to_completeP4Base +// CIR-BEFORE: %{{.+}} = cir.dyn_cast ptr %{{.+}} : !cir.ptr<!rec_Base> -> !cir.ptr<!void> +// CIR-BEFORE: } + +// CIR-AFTER: cir.func dso_local @_Z20ptr_cast_to_completeP4Base +// CIR-AFTER: %[[SRC:.*]] = cir.load{{.*}} %{{.+}} : !cir.ptr<!cir.ptr<!rec_Base>>, !cir.ptr<!rec_Base> +// CIR-AFTER-NEXT: %[[SRC_IS_NOT_NULL:.*]] = cir.cast ptr_to_bool %[[SRC]] : !cir.ptr<!rec_Base> -> !cir.bool +// CIR-AFTER-NEXT: %{{.+}} = cir.ternary(%[[SRC_IS_NOT_NULL]], true { +// CIR-AFTER-NEXT: %[[VPTR_PTR:.*]] = cir.vtable.get_vptr %[[SRC]] : !cir.ptr<!rec_Base> -> !cir.ptr<!cir.vptr> +// CIR-AFTER-NEXT: %[[VPTR:.*]] = cir.load %[[VPTR_PTR]] : !cir.ptr<!cir.vptr>, !cir.vptr +// CIR-AFTER-NEXT: %[[ELEM_PTR:.*]] = cir.cast bitcast %[[VPTR]] : !cir.vptr -> !cir.ptr<!s64i> +// CIR-AFTER-NEXT: %[[MINUS_TWO:.*]] = cir.const #cir.int<-2> : !s64i +// CIR-AFTER-NEXT: %[[BASE_OFFSET_PTR:.*]] = cir.ptr_stride %[[ELEM_PTR]], %[[MINUS_TWO]] : (!cir.ptr<!s64i>, !s64i) -> !cir.ptr<!s64i> +// CIR-AFTER-NEXT: %[[BASE_OFFSET:.*]] = cir.load{{.*}} %[[BASE_OFFSET_PTR]] : !cir.ptr<!s64i>, !s64i +// CIR-AFTER-NEXT: %[[SRC_BYTES_PTR:.*]] = cir.cast bitcast %[[SRC]] : !cir.ptr<!rec_Base> -> !cir.ptr<!u8i> +// CIR-AFTER-NEXT: %[[DST_BYTES_PTR:.*]] = cir.ptr_stride %[[SRC_BYTES_PTR]], %[[BASE_OFFSET]] : (!cir.ptr<!u8i>, !s64i) -> !cir.ptr<!u8i> +// CIR-AFTER-NEXT: %[[CASTED_PTR:.*]] = cir.cast bitcast %[[DST_BYTES_PTR]] : !cir.ptr<!u8i> -> !cir.ptr<!void> +// CIR-AFTER-NEXT: cir.yield %[[CASTED_PTR]] : !cir.ptr<!void> +// CIR-AFTER-NEXT: }, false { +// CIR-AFTER-NEXT: %[[NULL_PTR:.*]] = cir.const #cir.ptr<null> : !cir.ptr<!void> +// CIR-AFTER-NEXT: cir.yield %[[NULL_PTR]] : !cir.ptr<!void> +// CIR-AFTER-NEXT: }) : (!cir.bool) -> !cir.ptr<!void> +// CIR-AFTER: } + +// LLVM: define {{.*}} @_Z20ptr_cast_to_completeP4Base +// LLVM: %[[IS_NOT_NULL:.*]] = icmp ne ptr %[[PTR:.*]], null +// LLVM: br i1 %[[IS_NOT_NULL]], label %[[NOT_NULL:.*]], label %[[NULL:.*]] +// LLVM: [[NOT_NULL]]: +// LLVM: %[[VPTR:.*]] = load ptr, ptr %[[PTR]] +// LLVM: %[[BASE_OFFSET_PTR:.*]] = getelementptr i64, ptr %7, i64 -2 +// LLVM: %[[BASE_OFFSET:.*]] = load i64, ptr %[[BASE_OFFSET_PTR]] +// LLVM: %[[RESULT:.*]] = getelementptr i8, ptr %[[PTR]], i64 %[[BASE_OFFSET]] +// LLVM: br label %[[DONE:.*]] +// LLVM: [[NULL]]: +// LLVM: br label %[[DONE]] +// LLVM: [[DONE]]: +// LLVM: %[[RET:.*]] = phi ptr [ null, %[[NULL]] ], [ %[[RESULT]], %[[NOT_NULL]] ] + +// OGCG: define {{.*}} @_Z20ptr_cast_to_completeP4Base +// OGCG: %[[IS_NULL:.*]] = icmp eq ptr %[[PTR:.*]], null +// OGCG: br i1 %[[IS_NULL]], label %[[NULL:.*]], label %[[NOT_NULL:.*]] +// OGCG: [[NOT_NULL]]: +// OGCG: %[[VPTR:.*]] = load ptr, ptr %[[PTR]] +// OGCG: %[[BASE_OFFSET_PTR:.*]] = getelementptr inbounds i64, ptr %[[VPTR]], i64 -2 +// OGCG: %[[BASE_OFFSET:.*]] = load i64, ptr %[[BASE_OFFSET_PTR]] +// OGCG: %[[RESULT:.*]] = getelementptr inbounds i8, ptr %[[PTR]], i64 %[[BASE_OFFSET]] +// OGCG: br label %[[DONE:.*]] +// OGCG: [[NULL]]: +// OGCG: br label %[[DONE]] +// OGCG: [[DONE]]: +// OGCG: %[[RET:.*]] = phi ptr [ %[[RESULT]], %[[NOT_NULL]] ], [ null, %[[NULL]] ] diff --git a/clang/test/CIR/CodeGen/global-ctor-dtor.cpp b/clang/test/CIR/CodeGen/global-ctor-dtor.cpp new file mode 100644 index 0000000..2e03ff3 --- /dev/null +++ b/clang/test/CIR/CodeGen/global-ctor-dtor.cpp @@ -0,0 +1,45 @@ +// RUN: %clang_cc1 -std=c++20 -triple x86_64-unknown-linux-gnu -fclangir -emit-cir -mmlir --mlir-print-ir-before=cir-lowering-prepare %s -o %t.cir 2> %t-before.cir +// RUN: FileCheck --input-file=%t-before.cir %s --check-prefix=CIR-BEFORE-LPP +// RUN: FileCheck --input-file=%t.cir %s --check-prefix=CIR-AFTER +// RUN: FileCheck --check-prefix=CIR-AFTER --input-file=%t.cir %s +// RUN: %clang_cc1 -std=c++20 -triple x86_64-unknown-linux-gnu -fclangir -emit-llvm %s -o %t-cir.ll +// RUN: FileCheck --check-prefix=LLVM --input-file=%t-cir.ll %s +// RUN: %clang_cc1 -std=c++20 -triple x86_64-unknown-linux-gnu -emit-llvm %s -o %t.ll +// RUN: FileCheck --check-prefix=OGCG --input-file=%t.ll %s + +extern int bar(); +void foo(void) __attribute__((constructor)); +void foo(void) { + bar(); +} + +// CIR-BEFORE-LPP: cir.func dso_local @_Z3foov() global_ctor + +void foo2(void) __attribute__((constructor(777))); +void foo2(void) { + bar(); +} + +// CIR-BEFORE-LPP: cir.func dso_local @_Z4foo2v() global_ctor(777) + +void foo3(void) __attribute__((destructor)); +void foo3(void) { + bar(); +} + +// CIR-BEFORE-LPP: cir.func dso_local @_Z4foo3v() global_dtor + +void foo4(void) __attribute__((destructor(789))); +void foo4(void) { + bar(); +} + +// CIR-BEFORE-LPP: cir.func dso_local @_Z4foo4v() global_dtor(789) + +// CIR-AFTER: module @{{.*}} attributes {cir.global_ctors = [#cir.global_ctor<"_Z3foov", 65535>, #cir.global_ctor<"_Z4foo2v", 777>], cir.global_dtors = [#cir.global_dtor<"_Z4foo3v", 65535>, #cir.global_dtor<"_Z4foo4v", 789>] + +// LLVM: @llvm.global_ctors = appending global [2 x { i32, ptr, ptr }] [{ i32, ptr, ptr } { i32 65535, ptr @_Z3foov, ptr null }, { i32, ptr, ptr } { i32 777, ptr @_Z4foo2v, ptr null }] +// LLVM: @llvm.global_dtors = appending global [2 x { i32, ptr, ptr }] [{ i32, ptr, ptr } { i32 65535, ptr @_Z4foo3v, ptr null }, { i32, ptr, ptr } { i32 789, ptr @_Z4foo4v, ptr null }] + +// OGCG: @llvm.global_ctors = appending global [2 x { i32, ptr, ptr }] [{ i32, ptr, ptr } { i32 65535, ptr @_Z3foov, ptr null }, { i32, ptr, ptr } { i32 777, ptr @_Z4foo2v, ptr null }] +// OGCG: @llvm.global_dtors = appending global [2 x { i32, ptr, ptr }] [{ i32, ptr, ptr } { i32 65535, ptr @_Z4foo3v, ptr null }, { i32, ptr, ptr } { i32 789, ptr @_Z4foo4v, ptr null }] diff --git a/clang/test/CIR/CodeGen/goto.cpp b/clang/test/CIR/CodeGen/goto.cpp index 48cb44e..257c255 100644 --- a/clang/test/CIR/CodeGen/goto.cpp +++ b/clang/test/CIR/CodeGen/goto.cpp @@ -205,6 +205,8 @@ extern "C" void case_follow_label(int v) { // CIR: cir.func dso_local @case_follow_label // CIR: cir.switch // CIR: cir.case(equal, [#cir.int<1> : !s32i]) { +// CIR: cir.br ^bb1 +// CIR: ^bb1: // CIR: cir.label "label" // CIR: cir.case(equal, [#cir.int<2> : !s32i]) { // CIR: cir.call @action1() @@ -215,9 +217,11 @@ extern "C" void case_follow_label(int v) { // LLVM: define dso_local void @case_follow_label // LLVM: switch i32 {{.*}}, label %[[SWDEFAULT:.*]] [ -// LLVM: i32 1, label %[[LABEL:.*]] +// LLVM: i32 1, label %[[CASE1:.*]] // LLVM: i32 2, label %[[CASE2:.*]] // LLVM: ] +// LLVM: [[CASE1]]: +// LLVM: br label %[[LABEL:.*]] // LLVM: [[LABEL]]: // LLVM: br label %[[CASE2]] // LLVM: [[CASE2]]: @@ -303,3 +307,24 @@ extern "C" void default_follow_label(int v) { // OGCG: br label %label // OGCG: sw.epilog: // OGCG: ret void + +void g3() { +label: + goto label; +} + +// CIR: cir.func dso_local @_Z2g3v +// CIR: cir.br ^bb1 +// CIR: ^bb1: +// CIR: cir.label "label" +// CIR: cir.goto "label" + +// LLVM: define dso_local void @_Z2g3v() +// LLVM: br label %1 +// LLVM: 1: +// LLVM: br label %1 + +// OGCG: define dso_local void @_Z2g3v() +// OGCG: br label %label +// OGCG: label: +// OGCG: br label %label diff --git a/clang/test/CIR/CodeGen/inline-attributes.cpp b/clang/test/CIR/CodeGen/inline-attributes.cpp new file mode 100644 index 0000000..fab4010 --- /dev/null +++ b/clang/test/CIR/CodeGen/inline-attributes.cpp @@ -0,0 +1,75 @@ +// RUN: %clang_cc1 -std=c++11 -triple x86_64-unknown-linux-gnu -O1 -fclangir -emit-cir %s -o %t.cir +// RUN: FileCheck --check-prefix=CIR --input-file=%t.cir %s +// RUN: %clang_cc1 -std=c++11 -triple x86_64-unknown-linux-gnu -O1 -fclangir -emit-llvm %s -o %t-cir.ll +// RUN: FileCheck --check-prefix=LLVM --input-file=%t-cir.ll %s +// RUN: %clang_cc1 -std=c++11 -triple x86_64-unknown-linux-gnu -O1 -emit-llvm %s -o %t.ll +// RUN: FileCheck --check-prefix=OGCG --input-file=%t.ll %s + +extern int global_var; + +__attribute__((always_inline)) inline int always_inline_function(int x) { + return x * 2 + global_var; +} + +inline int inline_hint_function(int x) { + return x - 1 + global_var; +} + +__attribute__((noinline)) int noinline_function(int x) { + return x / 2 + global_var; +} + +int regular_function(int x) { + return x + 1 + global_var; +} + +// Force emission of all functions with function pointers +int (*always_inline_ptr)(int) = &always_inline_function; +int (*inline_hint_ptr)(int) = &inline_hint_function; +int (*noinline_ptr)(int) = &noinline_function; +int (*regular_ptr)(int) = ®ular_function; + +// CIR-LABEL: cir.func dso_local @_Z17noinline_functioni(%arg0: !s32i {{.*}}) -> !s32i inline(never) + +// CIR-LABEL: cir.func dso_local @_Z16regular_functioni(%arg0: !s32i {{.*}}) -> !s32i +// CIR-NOT: inline(never) +// CIR-NOT: inline(always) +// CIR-NOT: inline(hint) +// CIR-SAME: { + +// CIR-LABEL: cir.func {{.*}}@_Z22always_inline_functioni(%arg0: !s32i {{.*}}) -> !s32i inline(always) + +// CIR-LABEL: cir.func {{.*}}@_Z20inline_hint_functioni(%arg0: !s32i {{.*}}) -> !s32i inline(hint) + +// LLVM: ; Function Attrs:{{.*}} noinline +// LLVM: define{{.*}} i32 @_Z17noinline_functioni + +// LLVM: ; Function Attrs: +// LLVM-NOT: noinline +// LLVM-NOT: alwaysinline +// LLVM-NOT: inlinehint +// LLVM-SAME: {{$}} +// LLVM: define{{.*}} i32 @_Z16regular_functioni + +// LLVM: ; Function Attrs:{{.*}} alwaysinline +// LLVM: define{{.*}} i32 @_Z22always_inline_functioni + +// LLVM: ; Function Attrs:{{.*}} inlinehint +// LLVM: define{{.*}} i32 @_Z20inline_hint_functioni + +// OGCG: ; Function Attrs:{{.*}} noinline +// OGCG: define{{.*}} i32 @_Z17noinline_functioni + +// OGCG: ; Function Attrs: +// OGCG-NOT: noinline +// OGCG-NOT: alwaysinline +// OGCG-NOT: inlinehint +// OGCG-SAME: {{$}} +// OGCG: define{{.*}} i32 @_Z16regular_functioni + +// OGCG: ; Function Attrs:{{.*}} alwaysinline +// OGCG: define{{.*}} i32 @_Z22always_inline_functioni + +// OGCG: ; Function Attrs:{{.*}} inlinehint +// OGCG: define{{.*}} i32 @_Z20inline_hint_functioni + diff --git a/clang/test/CIR/CodeGen/label.c b/clang/test/CIR/CodeGen/label.c index a050094..fd3c7f2 100644 --- a/clang/test/CIR/CodeGen/label.c +++ b/clang/test/CIR/CodeGen/label.c @@ -11,10 +11,14 @@ labelA: } // CIR: cir.func no_proto dso_local @label +// CIR: cir.br ^bb1 +// CIR: ^bb1: // CIR: cir.label "labelA" // CIR: cir.return // LLVM:define dso_local void @label +// LLVM: br label %1 +// LLVM: 1: // LLVM: ret void // OGCG: define dso_local void @label @@ -29,15 +33,19 @@ labelC: } // CIR: cir.func no_proto dso_local @multiple_labels -// CIR: cir.label "labelB" // CIR: cir.br ^bb1 -// CIR: ^bb1: // pred: ^bb0 +// CIR: ^bb1: +// CIR: cir.label "labelB" +// CIR: cir.br ^bb2 +// CIR: ^bb2: // CIR: cir.label "labelC" // CIR: cir.return -// LLVM: define dso_local void @multiple_labels() +// LLVM: define dso_local void @multiple_labels(){{.*}} // LLVM: br label %1 // LLVM: 1: +// LLVM: br label %2 +// LLVM: 2: // LLVM: ret void // OGCG: define dso_local void @multiple_labels @@ -56,6 +64,8 @@ labelD: // CIR: cir.func dso_local @label_in_if // CIR: cir.if {{.*}} { +// CIR: cir.br ^bb1 +// CIR: ^bb1: // CIR: cir.label "labelD" // CIR: [[LOAD:%.*]] = cir.load align(4) [[COND:%.*]] : !cir.ptr<!s32i>, !s32i // CIR: [[INC:%.*]] = cir.unary(inc, %3) nsw : !s32i, !s32i @@ -63,20 +73,22 @@ labelD: // CIR: } // CIR: cir.return -// LLVM: define dso_local void @label_in_if +// LLVM: define dso_local void @label_in_if{{.*}} // LLVM: br label %3 // LLVM: 3: // LLVM: [[LOAD:%.*]] = load i32, ptr [[COND:%.*]], align 4 // LLVM: [[CMP:%.*]] = icmp ne i32 [[LOAD]], 0 -// LLVM: br i1 [[CMP]], label %6, label %9 +// LLVM: br i1 [[CMP]], label %6, label %10 // LLVM: 6: +// LLVM: br label %7 +// LLVM: 7: // LLVM: [[LOAD2:%.*]] = load i32, ptr [[COND]], align 4 // LLVM: [[ADD1:%.*]] = add nsw i32 [[LOAD2]], 1 // LLVM: store i32 [[ADD1]], ptr [[COND]], align 4 -// LLVM: br label %9 -// LLVM: 9: // LLVM: br label %10 // LLVM: 10: +// LLVM: br label %11 +// LLVM: 11: // LLVM: ret void // OGCG: define dso_local void @label_in_if @@ -103,7 +115,7 @@ void after_return() { // CIR: cir.label "label" // CIR: cir.br ^bb1 -// LLVM: define dso_local void @after_return +// LLVM: define dso_local void @after_return{{.*}} // LLVM: br label %1 // LLVM: 1: // LLVM: ret void @@ -127,7 +139,7 @@ void after_unreachable() { // CIR: cir.label "label" // CIR: cir.return -// LLVM: define dso_local void @after_unreachable +// LLVM: define dso_local void @after_unreachable{{.*}} // LLVM: unreachable // LLVM: 1: // LLVM: ret void @@ -142,11 +154,15 @@ end: return; } // CIR: cir.func no_proto dso_local @labelWithoutMatch +// CIR: cir.br ^bb1 +// CIR: ^bb1: // CIR: cir.label "end" // CIR: cir.return // CIR: } // LLVM: define dso_local void @labelWithoutMatch +// LLVM: br label %1 +// LLVM: 1: // LLVM: ret void // OGCG: define dso_local void @labelWithoutMatch @@ -167,13 +183,17 @@ void foo() { // CIR: cir.func no_proto dso_local @foo // CIR: cir.scope { -// CIR: cir.label "label" // CIR: %0 = cir.alloca !rec_S, !cir.ptr<!rec_S>, ["agg.tmp0"] +// CIR: cir.br ^bb1 +// CIR: ^bb1: +// CIR: cir.label "label" -// LLVM:define dso_local void @foo() { +// LLVM: define dso_local void @foo(){{.*}} { // LLVM: [[ALLOC:%.*]] = alloca %struct.S, i64 1, align 1 // LLVM: br label %2 // LLVM:2: +// LLVM: br label %3 +// LLVM:3: // LLVM: [[CALL:%.*]] = call %struct.S @get() // LLVM: store %struct.S [[CALL]], ptr [[ALLOC]], align 1 // LLVM: [[LOAD:%.*]] = load %struct.S, ptr [[ALLOC]], align 1 diff --git a/clang/test/CIR/CodeGen/lambda-static-invoker.cpp b/clang/test/CIR/CodeGen/lambda-static-invoker.cpp index 15d768e..e7d199b 100644 --- a/clang/test/CIR/CodeGen/lambda-static-invoker.cpp +++ b/clang/test/CIR/CodeGen/lambda-static-invoker.cpp @@ -50,7 +50,7 @@ int g3() { // CIR: %[[RET:.*]] = cir.load %[[RETVAL]] // CIR: cir.return %[[RET]] -// LLVM: define internal i32 @"_ZZ2g3vENK3$_0clERKi"(ptr %[[THIS_ARG:.*]], ptr %[[REF_I_ARG:.*]]) { +// LLVM: define internal i32 @"_ZZ2g3vENK3$_0clERKi"(ptr %[[THIS_ARG:.*]], ptr %[[REF_I_ARG:.*]]){{.*}} { // LLVM: %[[THIS_ALLOCA:.*]] = alloca ptr // LLVM: %[[REF_I_ALLOCA:.*]] = alloca ptr // LLVM: %[[RETVAL:.*]] = alloca i32 @@ -66,7 +66,7 @@ int g3() { // In OGCG, the _ZZ2g3vENK3$_0clERKi function is emitted after _ZZ2g3vEN3$_08__invokeERKi, see below. // lambda invoker -// CIR: cir.func internal private dso_local @_ZZ2g3vEN3$_08__invokeERKi(%[[REF_I_ARG:.*]]: !cir.ptr<!s32i> {{.*}}) -> !s32i { +// CIR: cir.func internal private dso_local @_ZZ2g3vEN3$_08__invokeERKi(%[[REF_I_ARG:.*]]: !cir.ptr<!s32i> {{.*}}) -> !s32i{{.*}} { // CIR: %[[REF_I_ALLOCA:.*]] = cir.alloca {{.*}} ["i", init, const] // CIR: %[[RETVAL:.*]] = cir.alloca {{.*}} ["__retval"] // CIR: %[[LAM_ALLOCA:.*]] = cir.alloca ![[REC_LAM_G3]], !cir.ptr<![[REC_LAM_G3]]>, ["unused.capture"] @@ -77,7 +77,7 @@ int g3() { // CIR: %[[RET:.*]] = cir.load %[[RETVAL]] // CIR: cir.return %[[RET]] -// LLVM: define internal i32 @"_ZZ2g3vEN3$_08__invokeERKi"(ptr %[[REF_I_ARG:.*]]) { +// LLVM: define internal i32 @"_ZZ2g3vEN3$_08__invokeERKi"(ptr %[[REF_I_ARG:.*]]){{.*}} { // LLVM: %[[REF_I_ALLOCA:.*]] = alloca ptr // LLVM: %[[RETVAL:.*]] = alloca i32 // LLVM: %[[LAM_ALLOCA:.*]] = alloca %[[REC_LAM_G3:.*]], @@ -91,7 +91,7 @@ int g3() { // In OGCG, the _ZZ2g3vEN3$_08__invokeERKi function is emitted after _ZN1A3barEv, see below. // lambda operator int (*)(int const&)() -// CIR: cir.func internal private dso_local @_ZZ2g3vENK3$_0cvPFiRKiEEv(%[[THIS_ARG:.*]]: !cir.ptr<![[REC_LAM_G3]]> {{.*}}) -> !cir.ptr<!cir.func<(!cir.ptr<!s32i>) -> !s32i>> { +// CIR: cir.func internal private dso_local @_ZZ2g3vENK3$_0cvPFiRKiEEv(%[[THIS_ARG:.*]]: !cir.ptr<![[REC_LAM_G3]]> {{.*}}) -> !cir.ptr<!cir.func<(!cir.ptr<!s32i>) -> !s32i>>{{.*}} { // CIR: %[[THIS_ALLOCA:.*]] = cir.alloca !cir.ptr<![[REC_LAM_G3]]>, !cir.ptr<!cir.ptr<![[REC_LAM_G3]]>>, ["this", init] // CIR: %[[RETVAL:.*]] = cir.alloca !cir.ptr<!cir.func<(!cir.ptr<!s32i>) -> !s32i>>, !cir.ptr<!cir.ptr<!cir.func<(!cir.ptr<!s32i>) -> !s32i>>>, ["__retval"] // CIR: cir.store %[[THIS_ARG]], %[[THIS_ALLOCA]] @@ -101,7 +101,7 @@ int g3() { // CIR: %[[RET:.*]] = cir.load %[[RETVAL]] // CIR: cir.return %[[RET]] -// LLVM: define internal ptr @"_ZZ2g3vENK3$_0cvPFiRKiEEv"(ptr %[[THIS_ARG:.*]]) { +// LLVM: define internal ptr @"_ZZ2g3vENK3$_0cvPFiRKiEEv"(ptr %[[THIS_ARG:.*]]){{.*}} { // LLVM: %[[THIS_ALLOCA:.*]] = alloca ptr // LLVM: %[[RETVAL:.*]] = alloca ptr // LLVM: store ptr %[[THIS_ARG]], ptr %[[THIS_ALLOCA]] @@ -112,7 +112,7 @@ int g3() { // In OGCG, the _ZZ2g3vENK3$_0cvPFiRKiEEv function is emitted just after the _Z2g3v function, see above. -// CIR: cir.func{{.*}} @_Z2g3v() -> !s32i { +// CIR: cir.func{{.*}} @_Z2g3v() -> !s32i{{.*}} { // CIR: %[[RETVAL:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["__retval"] // CIR: %[[FN_ADDR:.*]] = cir.alloca !cir.ptr<!cir.func<(!cir.ptr<!s32i>) -> !s32i>>, !cir.ptr<!cir.ptr<!cir.func<(!cir.ptr<!s32i>) -> !s32i>>>, ["fn", init] // CIR: %[[TASK:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["task", init] @@ -145,7 +145,7 @@ int g3() { // CIR: cir.return %[[RET]] // CIR: } -// LLVM: define dso_local i32 @_Z2g3v() { +// LLVM: define dso_local i32 @_Z2g3v(){{.*}} { // LLVM: %[[LAM_ALLOCA:.*]] = alloca %[[REC_LAM_G3]] // LLVM: %[[REF_TMP1:.*]] = alloca i32 // LLVM: %[[RETVAL:.*]] = alloca i32 diff --git a/clang/test/CIR/CodeGen/lambda.cpp b/clang/test/CIR/CodeGen/lambda.cpp index 033adc6..0c32ceb1 100644 --- a/clang/test/CIR/CodeGen/lambda.cpp +++ b/clang/test/CIR/CodeGen/lambda.cpp @@ -13,13 +13,13 @@ void fn() { a(); } -// CIR: cir.func lambda internal private dso_local @_ZZ2fnvENK3$_0clEv(%[[THIS_ARG:.*]]: !cir.ptr<![[REC_LAM_FN_A:.*]]> {{.*}}) +// CIR: cir.func lambda internal private dso_local @_ZZ2fnvENK3$_0clEv(%[[THIS_ARG:.*]]: !cir.ptr<![[REC_LAM_FN_A:.*]]> {{.*}}) {{.*}} { // CIR: %[[THIS:.*]] = cir.alloca !cir.ptr<![[REC_LAM_FN_A]]>, !cir.ptr<!cir.ptr<![[REC_LAM_FN_A]]>>, ["this", init] // CIR: cir.store %[[THIS_ARG]], %[[THIS]] // CIR: cir.load %[[THIS]] // CIR: cir.return -// CIR: cir.func dso_local @_Z2fnv() +// CIR: cir.func dso_local @_Z2fnv() {{.*}} { // CIR: %[[A:.*]] = cir.alloca ![[REC_LAM_FN_A]], !cir.ptr<![[REC_LAM_FN_A]]>, ["a"] // CIR: cir.call @_ZZ2fnvENK3$_0clEv(%[[A]]) @@ -52,7 +52,7 @@ void l0() { a(); } -// CIR: cir.func lambda internal private dso_local @_ZZ2l0vENK3$_0clEv(%[[THIS_ARG:.*]]: !cir.ptr<![[REC_LAM_L0_A:.*]]> {{.*}}) +// CIR: cir.func lambda internal private dso_local @_ZZ2l0vENK3$_0clEv(%[[THIS_ARG:.*]]: !cir.ptr<![[REC_LAM_L0_A:.*]]> {{.*}}) {{.*}} { // CIR: %[[THIS_ADDR:.*]] = cir.alloca !cir.ptr<![[REC_LAM_L0_A]]>, !cir.ptr<!cir.ptr<![[REC_LAM_L0_A]]>>, ["this", init] {alignment = 8 : i64} // CIR: cir.store %[[THIS_ARG]], %[[THIS_ADDR]] // CIR: %[[THIS:.*]] = cir.load %[[THIS_ADDR]] @@ -66,7 +66,7 @@ void l0() { // CIR: cir.store{{.*}} %[[I_PLUS_ONE]], %[[I_ADDR]] // CIR: cir.return -// CIR: cir.func {{.*}} @_Z2l0v() +// CIR: cir.func {{.*}} @_Z2l0v() {{.*}} { // CIR: %[[I:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["i"] // CIR: %[[A:.*]] = cir.alloca ![[REC_LAM_L0_A]], !cir.ptr<![[REC_LAM_L0_A]]>, ["a", init] // CIR: %[[I_ADDR:.*]] = cir.get_member %[[A]][0] {name = "i"} @@ -124,7 +124,7 @@ auto g() { }; } -// CIR: cir.func dso_local @_Z1gv() -> ![[REC_LAM_G:.*]] { +// CIR: cir.func dso_local @_Z1gv() -> ![[REC_LAM_G:.*]] {{.*}} { // CIR: %[[RETVAL:.*]] = cir.alloca ![[REC_LAM_G]], !cir.ptr<![[REC_LAM_G]]>, ["__retval"] // CIR: %[[I_ADDR:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["i", init] // CIR: %[[TWELVE:.*]] = cir.const #cir.int<12> : !s32i @@ -166,7 +166,7 @@ auto g2() { } // Should be same as above because of NRVO -// CIR: cir.func dso_local @_Z2g2v() -> ![[REC_LAM_G2:.*]] { +// CIR: cir.func dso_local @_Z2g2v() -> ![[REC_LAM_G2:.*]] {{.*}} { // CIR: %[[RETVAL:.*]] = cir.alloca ![[REC_LAM_G2]], !cir.ptr<![[REC_LAM_G2]]>, ["__retval", init] // CIR: %[[I_ADDR:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["i", init] // CIR: %[[TWELVE:.*]] = cir.const #cir.int<12> : !s32i @@ -199,7 +199,7 @@ int f() { return g2()(); } -// CIR:cir.func lambda internal private dso_local @_ZZ2g2vENK3$_0clEv(%[[THIS_ARG:.*]]: !cir.ptr<![[REC_LAM_G2]]> {{.*}}) -> !s32i +// CIR:cir.func lambda internal private dso_local @_ZZ2g2vENK3$_0clEv(%[[THIS_ARG:.*]]: !cir.ptr<![[REC_LAM_G2]]> {{.*}}) -> !s32i {{.*}} { // CIR: %[[THIS_ADDR:.*]] = cir.alloca !cir.ptr<![[REC_LAM_G2]]>, !cir.ptr<!cir.ptr<![[REC_LAM_G2]]>>, ["this", init] // CIR: %[[RETVAL:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["__retval"] // CIR: cir.store %[[THIS_ARG]], %[[THIS_ADDR]] @@ -217,7 +217,7 @@ int f() { // CIR: %[[RET:.*]] = cir.load %[[RETVAL]] // CIR: cir.return %[[RET]] -// CIR: cir.func dso_local @_Z1fv() -> !s32i +// CIR: cir.func dso_local @_Z1fv() -> !s32i {{.*}} { // CIR: %[[RETVAL:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["__retval"] // CIR: %[[SCOPE_RET:.*]] = cir.scope { // CIR: %[[TMP:.*]] = cir.alloca ![[REC_LAM_G2]], !cir.ptr<![[REC_LAM_G2]]>, ["ref.tmp0"] @@ -301,7 +301,7 @@ struct A { // OGCG: call noundef i32 @_ZN1A3barEv(ptr {{.*}} %[[A_THIS]]) // lambda operator() in foo() -// CIR: cir.func lambda comdat linkonce_odr @_ZZN1A3fooEvENKUlvE_clEv(%[[THIS_ARG:.*]]: !cir.ptr<![[REC_LAM_A:.*]]> {{.*}}) +// CIR: cir.func lambda comdat linkonce_odr @_ZZN1A3fooEvENKUlvE_clEv(%[[THIS_ARG:.*]]: !cir.ptr<![[REC_LAM_A:.*]]> {{.*}}) {{.*}} { // CIR: %[[THIS_ADDR:.*]] = cir.alloca !cir.ptr<![[REC_LAM_A]]>, !cir.ptr<!cir.ptr<![[REC_LAM_A]]>>, ["this", init] // CIR: %[[RETVAL:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["__retval"] // CIR: cir.store{{.*}} %[[THIS_ARG]], %[[THIS_ADDR]] @@ -328,7 +328,7 @@ struct A { // The function above is defined after _ZN1A3barEv in OGCG, see below. // A::foo() -// CIR: cir.func {{.*}} @_ZN1A3fooEv(%[[THIS_ARG:.*]]: !cir.ptr<!rec_A> {{.*}}) -> !s32i +// CIR: cir.func {{.*}} @_ZN1A3fooEv(%[[THIS_ARG:.*]]: !cir.ptr<!rec_A> {{.*}}) -> !s32i {{.*}} { // CIR: %[[THIS_ADDR:.*]] = cir.alloca !cir.ptr<!rec_A>, !cir.ptr<!cir.ptr<!rec_A>>, ["this", init] // CIR: %[[RETVAL:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["__retval"] // CIR: cir.store %[[THIS_ARG]], %[[THIS_ADDR]] @@ -373,7 +373,7 @@ struct A { // OGCG: ret i32 %[[LAM_RET]] // lambda operator() in bar() -// CIR: cir.func {{.*}} @_ZZN1A3barEvENKUlvE_clEv(%[[THIS_ARG2:.*]]: !cir.ptr<![[REC_LAM_PTR_A:.*]]> {{.*}}) -> !s32i +// CIR: cir.func {{.*}} @_ZZN1A3barEvENKUlvE_clEv(%[[THIS_ARG2:.*]]: !cir.ptr<![[REC_LAM_PTR_A:.*]]> {{.*}}) -> !s32i {{.*}} { // CIR: %[[THIS_ADDR:.*]] = cir.alloca !cir.ptr<![[REC_LAM_PTR_A]]>, !cir.ptr<!cir.ptr<![[REC_LAM_PTR_A]]>>, ["this", init] // CIR: %[[RETVAL:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["__retval"] // CIR: cir.store{{.*}} %[[THIS_ARG]], %[[THIS_ADDR]] @@ -402,7 +402,7 @@ struct A { // The function above is defined after _ZZN1A3fooEvENKUlvE_clEv in OGCG, see below. // A::bar() -// CIR: cir.func {{.*}} @_ZN1A3barEv(%[[THIS_ARG:.*]]: !cir.ptr<!rec_A> {{.*}}) -> !s32i +// CIR: cir.func {{.*}} @_ZN1A3barEv(%[[THIS_ARG:.*]]: !cir.ptr<!rec_A> {{.*}}) -> !s32i {{.*}} { // CIR: %[[THIS_ADDR:.*]] = cir.alloca !cir.ptr<!rec_A>, !cir.ptr<!cir.ptr<!rec_A>>, ["this", init] // CIR: %[[RETVAL:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["__retval"] // CIR: cir.store %[[THIS_ARG]], %[[THIS_ADDR]] @@ -472,7 +472,7 @@ int test_lambda_this1(){ return x+y; } -// CIR: cir.func {{.*}} @_Z17test_lambda_this1v +// CIR: cir.func {{.*}} @_Z17test_lambda_this1v{{.*}} { // CIR: cir.call @_ZN1AC1Ev(%[[A_THIS:.*]]){{.*}} : (!cir.ptr<!rec_A>) -> () // CIR: cir.call @_ZN1A3fooEv(%[[A_THIS]]){{.*}} : (!cir.ptr<!rec_A>) -> !s32i // CIR: cir.call @_ZN1A3barEv(%[[A_THIS]]){{.*}} : (!cir.ptr<!rec_A>) -> !s32i diff --git a/clang/test/CIR/CodeGen/linkage-spec.cpp b/clang/test/CIR/CodeGen/linkage-spec.cpp index eb6c7b0..1affecd 100644 --- a/clang/test/CIR/CodeGen/linkage-spec.cpp +++ b/clang/test/CIR/CodeGen/linkage-spec.cpp @@ -1,42 +1,42 @@ // RUN: %clang_cc1 -std=c++20 -triple x86_64-unknown-linux-gnu -fclangir -emit-cir %s -o - 2>&1 | FileCheck %s extern "C" void TopLevelC(){} -// CHECK: cir.func{{.*}} @TopLevelC() { +// CHECK: cir.func dso_local @TopLevelC() inline(never) { extern "C++" void TopLevelCpp(){} -// CHECK: cir.func{{.*}} @_Z11TopLevelCppv() { +// CHECK: cir.func dso_local @_Z11TopLevelCppv() inline(never) { extern "C++" { void ExternCppEmpty(){} - // CHECK: cir.func{{.*}} @_Z14ExternCppEmptyv() { + // CHECK: cir.func dso_local @_Z14ExternCppEmptyv() inline(never) { extern "C" void ExternCpp_C(){} - // CHECK: cir.func{{.*}} @ExternCpp_C() { + // CHECK: cir.func dso_local @ExternCpp_C() inline(never) { extern "C++" void ExternCpp_Cpp(){} - // CHECK: cir.func{{.*}} @_Z13ExternCpp_Cppv() { + // CHECK: cir.func dso_local @_Z13ExternCpp_Cppv() inline(never) { extern "C" { void ExternCpp_CEmpty(){} - // CHECK: cir.func{{.*}} @ExternCpp_CEmpty() { + // CHECK: cir.func dso_local @ExternCpp_CEmpty() inline(never) { extern "C" void ExternCpp_C_C(){} - // CHECK: cir.func{{.*}} @ExternCpp_C_C() { + // CHECK: cir.func dso_local @ExternCpp_C_C() inline(never) { extern "C++" void ExternCpp_C_Cpp(){} - // CHECK: cir.func{{.*}} @_Z15ExternCpp_C_Cppv() { + // CHECK: cir.func dso_local @_Z15ExternCpp_C_Cppv() inline(never) { } } extern "C" { void ExternCEmpty(){} - // CHECK: cir.func{{.*}} @ExternCEmpty() { + // CHECK: cir.func dso_local @ExternCEmpty() inline(never) { extern "C" void ExternC_C(){} - // CHECK: cir.func{{.*}} @ExternC_C() { + // CHECK: cir.func dso_local @ExternC_C() inline(never) { extern "C++" void ExternC_Cpp(){} - // CHECK: cir.func{{.*}} @_Z11ExternC_Cppv() { + // CHECK: cir.func dso_local @_Z11ExternC_Cppv() inline(never) { extern "C++" { void ExternC_CppEmpty(){} - // CHECK: cir.func{{.*}} @_Z16ExternC_CppEmptyv() { + // CHECK: cir.func dso_local @_Z16ExternC_CppEmptyv() inline(never) { extern "C" void ExternC_Cpp_C(){} - // CHECK: cir.func{{.*}} @ExternC_Cpp_C() { + // CHECK: cir.func dso_local @ExternC_Cpp_C() inline(never) { extern "C++" void ExternC_Cpp_Cpp(){} - // CHECK: cir.func{{.*}} @_Z15ExternC_Cpp_Cppv() { + // CHECK: cir.func dso_local @_Z15ExternC_Cpp_Cppv() inline(never) { } } diff --git a/clang/test/CIR/CodeGen/loop.cpp b/clang/test/CIR/CodeGen/loop.cpp index b932f9d..3d28666 100644 --- a/clang/test/CIR/CodeGen/loop.cpp +++ b/clang/test/CIR/CodeGen/loop.cpp @@ -24,7 +24,7 @@ void l0() { // CIR: cir.return // CIR: } -// LLVM: define{{.*}} void @_Z2l0v() +// LLVM: define{{.*}} void @_Z2l0v(){{.*}} // LLVM: br label %[[LABEL1:.*]] // LLVM: [[LABEL1]]: // LLVM: br label %[[LABEL2:.*]] @@ -67,7 +67,7 @@ void l1() { // CIR-NEXT: cir.return // CIR-NEXT: } -// LLVM: define{{.*}} void @_Z2l1v() +// LLVM: define{{.*}} void @_Z2l1v(){{.*}} // LLVM: %[[I:.*]] = alloca i32, i64 1, align 4 // LLVM: br label %[[LABEL1:.*]] // LLVM: [[LABEL1]]: @@ -117,7 +117,7 @@ void l2() { // CIR-NEXT: cir.return // CIR-NEXT: } -// LLVM: define{{.*}} void @_Z2l2v() +// LLVM: define{{.*}} void @_Z2l2v(){{.*}} // LLVM: %[[I:.*]] = alloca i32, i64 1, align 4 // LLVM: br label %[[LABEL1:.*]] // LLVM: [[LABEL1]]: @@ -165,7 +165,7 @@ void l3() { // CIR-NEXT: cir.return // CIR-NEXT: } -// LLVM: define{{.*}} void @_Z2l3v() +// LLVM: define{{.*}} void @_Z2l3v(){{.*}} // LLVM: %[[I:.*]] = alloca i32, i64 1, align 4 // LLVM: br label %[[LABEL1:.*]] // LLVM: [[LABEL1]]: @@ -231,7 +231,7 @@ void l4() { // CIR: } // CIR: } -// LLVM: define{{.*}} void @_Z2l4v() { +// LLVM: define{{.*}} void @_Z2l4v(){{.*}} { // LLVM: %[[RANGE_ADDR:.*]] = alloca ptr // LLVM: %[[BEGIN_ADDR:.*]] = alloca ptr // LLVM: %[[END_ADDR:.*]] = alloca ptr @@ -355,7 +355,7 @@ void l5() { // CIR: } // CIR: } -// LLVM: define{{.*}} void @_Z2l5v() { +// LLVM: define{{.*}} void @_Z2l5v(){{.*}} { // LLVM: %[[ARR_ADDR:.*]] = alloca [4 x i32] // LLVM: %[[RANGE_ADDR:.*]] = alloca ptr // LLVM: %[[BEGIN_ADDR:.*]] = alloca ptr @@ -448,7 +448,7 @@ void test_do_while_false() { // CIR-NEXT: %[[FALSE:.*]] = cir.cast int_to_bool %[[ZERO]] : !s32i -> !cir.bool // CIR-NEXT: cir.condition(%[[FALSE]]) -// LLVM: define{{.*}} void @_Z19test_do_while_falsev() +// LLVM: define{{.*}} void @_Z19test_do_while_falsev(){{.*}} // LLVM: br label %[[LABEL1:.*]] // LLVM: [[LABEL1]]: // LLVM: br label %[[LABEL3:.*]] @@ -486,7 +486,7 @@ void test_empty_while_true() { // CIR-NEXT: } // CIR-NEXT: cir.yield -// LLVM: define{{.*}} void @_Z21test_empty_while_truev() +// LLVM: define{{.*}} void @_Z21test_empty_while_truev(){{.*}} // LLVM: br label %[[LABEL1:.*]] // LLVM: [[LABEL1]]: // LLVM: br label %[[LABEL2:.*]] @@ -539,7 +539,7 @@ void unreachable_after_continue() { // CIR: cir.return // CIR: } -// LLVM: define{{.*}} void @_Z26unreachable_after_continuev() +// LLVM: define{{.*}} void @_Z26unreachable_after_continuev(){{.*}} // LLVM: %[[X:.*]] = alloca i32, i64 1, align 4 // LLVM: br label %[[LABEL1:.*]] // LLVM: [[LABEL1]]: @@ -599,7 +599,7 @@ void unreachable_after_break() { // CIR: cir.return // CIR: } -// LLVM: define{{.*}} void @_Z23unreachable_after_breakv() +// LLVM: define{{.*}} void @_Z23unreachable_after_breakv(){{.*}} // LLVM: %[[X:.*]] = alloca i32, i64 1, align 4 // LLVM: br label %[[LABEL1:.*]] // LLVM: [[LABEL1]]: diff --git a/clang/test/CIR/CodeGen/member-functions.cpp b/clang/test/CIR/CodeGen/member-functions.cpp index 8be2c7f..d46345db 100644 --- a/clang/test/CIR/CodeGen/member-functions.cpp +++ b/clang/test/CIR/CodeGen/member-functions.cpp @@ -19,7 +19,7 @@ void C::f() {} void C::f2(int a, int b) {} -// CIR: cir.func{{.*}} @_ZN1C2f2Eii(%[[THIS_ARG:.*]]: !cir.ptr<!rec_C> {{.*}}, %[[A_ARG:.*]]: !s32i {{.*}}, %[[B_ARG:.*]]: !s32i {{.*}}) { +// CIR: cir.func{{.*}} @_ZN1C2f2Eii(%[[THIS_ARG:.*]]: !cir.ptr<!rec_C> {{.*}}, %[[A_ARG:.*]]: !s32i {{.*}}, %[[B_ARG:.*]]: !s32i {{.*}}) // CIR-NEXT: %[[THIS_ADDR:.*]] = cir.alloca !cir.ptr<!rec_C>, !cir.ptr<!cir.ptr<!rec_C>>, ["this", init] // CIR-NEXT: %[[A_ADDR:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["a", init] // CIR-NEXT: %[[B_ADDR:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["b", init] @@ -36,7 +36,7 @@ void test1() { c.f2(1, 2); } -// CIR: cir.func{{.*}} @_Z5test1v() { +// CIR: cir.func{{.*}} @_Z5test1v() // CIR-NEXT: %[[C_ADDR:.*]] = cir.alloca !rec_C, !cir.ptr<!rec_C>, ["c"] // CIR-NEXT: cir.call @_ZN1C1fEv(%[[C_ADDR]]) : (!cir.ptr<!rec_C>) -> () // CIR-NEXT: %[[ONE:.*]] = cir.const #cir.int<1> : !s32i diff --git a/clang/test/CIR/CodeGen/nrvo.cpp b/clang/test/CIR/CodeGen/nrvo.cpp index 72c39d7..ce08c79 100644 --- a/clang/test/CIR/CodeGen/nrvo.cpp +++ b/clang/test/CIR/CodeGen/nrvo.cpp @@ -22,13 +22,13 @@ struct S f1() { return s; } -// CIR: cir.func{{.*}} @_Z2f1v() -> !rec_S { +// CIR: cir.func{{.*}} @_Z2f1v() -> !rec_S // CIR-NEXT: %[[RETVAL:.*]] = cir.alloca !rec_S, !cir.ptr<!rec_S>, ["__retval", init] // CIR-NEXT: cir.call @_ZN1SC1Ev(%[[RETVAL]]) : (!cir.ptr<!rec_S>) -> () // CIR-NEXT: %[[RET:.*]] = cir.load %[[RETVAL]] : !cir.ptr<!rec_S>, !rec_S // CIR-NEXT: cir.return %[[RET]] -// CIR-NOELIDE: cir.func{{.*}} @_Z2f1v() -> !rec_S { +// CIR-NOELIDE: cir.func{{.*}} @_Z2f1v() -> !rec_S // CIR-NOELIDE-NEXT: %[[RETVAL:.*]] = cir.alloca !rec_S, !cir.ptr<!rec_S>, ["__retval"] // CIR-NOELIDE-NEXT: %[[S:.*]] = cir.alloca !rec_S, !cir.ptr<!rec_S>, ["s", init] // CIR-NOELIDE-NEXT: cir.call @_ZN1SC1Ev(%[[S]]) : (!cir.ptr<!rec_S>) -> () diff --git a/clang/test/CIR/CodeGen/ptrdiff.c b/clang/test/CIR/CodeGen/ptrdiff.c new file mode 100644 index 0000000..e6a754e --- /dev/null +++ b/clang/test/CIR/CodeGen/ptrdiff.c @@ -0,0 +1,51 @@ +// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -fclangir -emit-cir %s -o %t.cir +// RUN: FileCheck --check-prefix=CIR --input-file=%t.cir %s +// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -fclangir -emit-llvm %s -o %t-cir.ll +// RUN: FileCheck --check-prefix=LLVM --input-file=%t-cir.ll %s +// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -emit-llvm %s -o %t.ll +// RUN: FileCheck --check-prefix=OGCG --input-file=%t.ll %s + +int addrcmp(const void* a, const void* b) { + // CIR-LABEL: addrcmp + // CIR: %[[R:.*]] = cir.ptr_diff + // CIR: cir.cast integral %[[R]] : !s64i -> !s32i + + // LLVM-LABEL: define dso_local i32 @addrcmp( + // LLVM: %[[PTR_A:.*]] = ptrtoint ptr {{.*}} to i64 + // LLVM: %[[PTR_B:.*]] = ptrtoint ptr {{.*}} to i64 + // LLVM: %[[SUB:.*]] = sub i64 %[[PTR_A]], %[[PTR_B]] + // LLVM-NOT: sdiv + // LLVM: trunc i64 %[[SUB]] to i32 + + // OGCG-LABEL: define dso_local i32 @addrcmp( + // OGCG: %[[PTR_A:.*]] = ptrtoint ptr {{.*}} to i64 + // OGCG: %[[PTR_B:.*]] = ptrtoint ptr {{.*}} to i64 + // OGCG: %[[SUB:.*]] = sub i64 %[[PTR_A]], %[[PTR_B]] + // OGCG-NOT: sdiv + // OGCG: trunc i64 %[[SUB]] to i32 + return *(const void**)a - *(const void**)b; +} + +unsigned long long test_ptr_diff(int *a, int* b) { + // CIR-LABEL: test_ptr_diff + // CIR: %[[D:.*]] = cir.ptr_diff {{.*}} : !cir.ptr<!s32i> -> !s64i + // CIR: %[[U:.*]] = cir.cast integral %[[D]] : !s64i -> !u64i + // CIR: cir.return {{.*}} : !u64i + + // LLVM-LABEL: define dso_local i64 @test_ptr_diff( + // LLVM: %[[IA:.*]] = ptrtoint ptr %{{.*}} to i64 + // LLVM: %[[IB:.*]] = ptrtoint ptr %{{.*}} to i64 + // LLVM: %[[SUB:.*]] = sub i64 %[[IA]], %[[IB]] + // LLVM: %[[Q:.*]] = sdiv exact i64 %[[SUB]], 4 + // LLVM: store i64 %[[Q]], ptr %[[RETADDR:.*]], align + // LLVM: %[[RETLOAD:.*]] = load i64, ptr %[[RETADDR]], align + // LLVM: ret i64 %[[RETLOAD]] + + // OGCG-LABEL: define dso_local i64 @test_ptr_diff( + // OGCG: %[[IA:.*]] = ptrtoint ptr %{{.*}} to i64 + // OGCG: %[[IB:.*]] = ptrtoint ptr %{{.*}} to i64 + // OGCG: %[[SUB:.*]] = sub i64 %[[IA]], %[[IB]] + // OGCG: %[[Q:.*]] = sdiv exact i64 %[[SUB]], 4 + // OGCG: ret i64 %[[Q]] + return a - b; +}
\ No newline at end of file diff --git a/clang/test/CIR/CodeGen/ptrdiff.cpp b/clang/test/CIR/CodeGen/ptrdiff.cpp new file mode 100644 index 0000000..34ba0ff --- /dev/null +++ b/clang/test/CIR/CodeGen/ptrdiff.cpp @@ -0,0 +1,33 @@ +// RUN: %clang_cc1 -std=c++20 -triple x86_64-unknown-linux-gnu -fclangir -emit-cir %s -o %t.cir +// RUN: FileCheck --input-file=%t.cir %s --check-prefix=CIR +// RUN: %clang_cc1 -std=c++20 -triple x86_64-unknown-linux-gnu -fclangir -emit-llvm %s -o %t-cir.ll +// RUN: FileCheck --input-file=%t-cir.ll %s --check-prefix=LLVM +// RUN: %clang_cc1 -std=c++20 -triple x86_64-unknown-linux-gnu -emit-llvm %s -o %t.ll +// RUN: FileCheck --input-file=%t.ll %s --check-prefix=OGCG + +typedef unsigned long size_type; + +size_type size(unsigned long *_start, unsigned long *_finish) { + // CIR-LABEL: cir.func dso_local @_Z4sizePmS_ + // CIR: %[[D:.*]] = cir.ptr_diff {{.*}} : !cir.ptr<!u64i> -> !s64i + // CIR: %[[U:.*]] = cir.cast integral %[[D]] : !s64i -> !u64i + // CIR: cir.return {{.*}} : !u64i + + // LLVM-LABEL: define dso_local {{.*}}i64 @_Z4sizePmS_( + // LLVM: %[[IA:.*]] = ptrtoint ptr %{{.*}} to i64 + // LLVM: %[[IB:.*]] = ptrtoint ptr %{{.*}} to i64 + // LLVM: %[[SUB:.*]] = sub i64 %[[IA]], %[[IB]] + // LLVM: %[[Q:.*]] = sdiv exact i64 %[[SUB]], 8 + // LLVM: store i64 %[[Q]], ptr %[[RETADDR:.*]], align + // LLVM: %[[RET:.*]] = load i64, ptr %[[RETADDR]], align + // LLVM: ret i64 %[[RET]] + + // OGCG-LABEL: define dso_local {{.*}}i64 @_Z4sizePmS_( + // OGCG: %[[IA:.*]] = ptrtoint ptr %{{.*}} to i64 + // OGCG: %[[IB:.*]] = ptrtoint ptr %{{.*}} to i64 + // OGCG: %[[SUB:.*]] = sub i64 %[[IA]], %[[IB]] + // OGCG: %[[Q:.*]] = sdiv exact i64 %[[SUB]], 8 + // OGCG: ret i64 %[[Q]] + + return static_cast<size_type>(_finish - _start); +}
\ No newline at end of file diff --git a/clang/test/CIR/CodeGen/ternary.cpp b/clang/test/CIR/CodeGen/ternary.cpp index eb38ee3..e7b7270 100644 --- a/clang/test/CIR/CodeGen/ternary.cpp +++ b/clang/test/CIR/CodeGen/ternary.cpp @@ -10,7 +10,7 @@ int x(int y) { } // CIR-LABEL: cir.func{{.*}} @_Z1xi( -// CIR-SAME: %[[ARG0:.*]]: !s32i {{.*}}) -> !s32i { +// CIR-SAME: %[[ARG0:.*]]: !s32i {{.*}}) -> !s32i // CIR: [[Y:%.+]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["y", init] {alignment = 4 : i64} // CIR: [[RETVAL:%.+]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["__retval"] {alignment = 4 : i64} // CIR: cir.store %[[ARG0]], [[Y]] : !s32i, !cir.ptr<!s32i> @@ -52,7 +52,7 @@ int foo(int a, int b) { } // CIR-LABEL: cir.func{{.*}} @_Z3fooii( -// CIR-SAME: %[[ARG0:.*]]: !s32i {{.*}}, %[[ARG1:.*]]: !s32i {{.*}}) -> !s32i { +// CIR-SAME: %[[ARG0:.*]]: !s32i {{.*}}, %[[ARG1:.*]]: !s32i {{.*}}) -> !s32i // CIR: [[A:%.+]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["a", init] {alignment = 4 : i64} // CIR: [[B:%.+]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["b", init] {alignment = 4 : i64} // CIR: [[RETVAL:%.+]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["__retval"] {alignment = 4 : i64} diff --git a/clang/test/CIR/CodeGen/throws.cpp b/clang/test/CIR/CodeGen/throws.cpp index ff6aa62..53af1ef 100644 --- a/clang/test/CIR/CodeGen/throws.cpp +++ b/clang/test/CIR/CodeGen/throws.cpp @@ -123,3 +123,124 @@ void paren_expr() { (throw 0, 1 + 2); } // OGCG: %[[EXCEPTION_ADDR:.*]] = call ptr @__cxa_allocate_exception(i64 4) // OGCG: store i32 0, ptr %[[EXCEPTION_ADDR]], align 16 // OGCG: call void @__cxa_throw(ptr %[[EXCEPTION_ADDR]], ptr @_ZTIi, ptr null) + +void throw_complex_expr() { + throw __builtin_complex(1.1f, 2.2f); +} + +// CIR: %[[EXCEPTION_ADDR:.*]] = cir.alloc.exception 8 -> !cir.ptr<!cir.complex<!cir.float>> +// CIR: %[[EXCEPTION_VALUE:.*]] = cir.const #cir.const_complex<#cir.fp<1.100000e+00> : !cir.float, #cir.fp<2.200000e+00> : !cir.float> : !cir.complex<!cir.float> +// CIR: cir.store{{.*}} %[[EXCEPTION_VALUE]], %[[EXCEPTION_ADDR]] : !cir.complex<!cir.float>, !cir.ptr<!cir.complex<!cir.float>> +// CIR: cir.throw %[[EXCEPTION_ADDR]] : !cir.ptr<!cir.complex<!cir.float>>, @_ZTICf +// CIR: cir.unreachable + +// LLVM: %[[EXCEPTION_ADDR:.*]] = call ptr @__cxa_allocate_exception(i64 8) +// LLVM: store { float, float } { float 0x3FF19999A0000000, float 0x40019999A0000000 }, ptr %[[EXCEPTION_ADDR]], align 16 +// LLVM: call void @__cxa_throw(ptr %[[EXCEPTION_ADDR]], ptr @_ZTICf, ptr null) + +// OGCG: %[[EXCEPTION_ADDR:.*]] = call ptr @__cxa_allocate_exception(i64 8) +// OGCG: %[[EXCEPTION_REAL:.*]] = getelementptr inbounds nuw { float, float }, ptr %[[EXCEPTION_ADDR]], i32 0, i32 0 +// OGCG: %[[EXCEPTION_IMAG:.*]] = getelementptr inbounds nuw { float, float }, ptr %[[EXCEPTION_ADDR]], i32 0, i32 1 +// OGCG: store float 0x3FF19999A0000000, ptr %[[EXCEPTION_REAL]], align 16 +// OGCG: store float 0x40019999A0000000, ptr %[[EXCEPTION_IMAG]], align 4 +// OGCG: call void @__cxa_throw(ptr %[[EXCEPTION_ADDR]], ptr @_ZTICf, ptr null) + +void throw_vector_type() { + typedef int vi4 __attribute__((vector_size(16))); + vi4 a; + throw a; +} + +// CIR: %[[A_ADDR:.*]] = cir.alloca !cir.vector<4 x !s32i>, !cir.ptr<!cir.vector<4 x !s32i>>, ["a"] +// CIR: %[[EXCEPTION_ADDR:.*]] = cir.alloc.exception 16 -> !cir.ptr<!cir.vector<4 x !s32i>> +// CIR: %[[TMP_A:.*]] = cir.load{{.*}} %[[A_ADDR]] : !cir.ptr<!cir.vector<4 x !s32i>>, !cir.vector<4 x !s32i> +// CIR: cir.store{{.*}} %[[TMP_A]], %[[EXCEPTION_ADDR]] : !cir.vector<4 x !s32i>, !cir.ptr<!cir.vector<4 x !s32i>> +// CIR: cir.throw %[[EXCEPTION_ADDR]] : !cir.ptr<!cir.vector<4 x !s32i>>, @_ZTIDv4_i +// CIR: cir.unreachable + +// LLVM: %[[A_ADDR:.*]] = alloca <4 x i32>, i64 1, align 16 +// LLVM: %[[EXCEPTION_ADDR:.*]] = call ptr @__cxa_allocate_exception(i64 16) +// LLVM: %[[TMP_A:.*]] = load <4 x i32>, ptr %[[A_ADDR]], align 16 +// LLVM: store <4 x i32> %[[TMP_A]], ptr %[[EXCEPTION_ADDR]], align 16 +// LLVM: call void @__cxa_throw(ptr %[[EXCEPTION_ADDR]], ptr @_ZTIDv4_i, ptr null) + +// OGCG: %[[A_ADDR:.*]] = alloca <4 x i32>, align 16 +// OGCG: %[[EXCEPTION_ADDR:.*]] = call ptr @__cxa_allocate_exception(i64 16) +// OGCG: %[[TMP_A:.*]] = load <4 x i32>, ptr %[[A_ADDR]], align 16 +// OGCG: store <4 x i32> %[[TMP_A]], ptr %[[EXCEPTION_ADDR]], align 16 +// OGCG: call void @__cxa_throw(ptr %[[EXCEPTION_ADDR]], ptr @_ZTIDv4_i, ptr null) +// OGCG: unreachable + +void throw_ext_vector_type() { + typedef int vi4 __attribute__((ext_vector_type(4))); + vi4 a; + throw a; +} + +// CIR: %[[A_ADDR:.*]] = cir.alloca !cir.vector<4 x !s32i>, !cir.ptr<!cir.vector<4 x !s32i>>, ["a"] +// CIR: %[[EXCEPTION_ADDR:.*]] = cir.alloc.exception 16 -> !cir.ptr<!cir.vector<4 x !s32i>> +// CIR: %[[TMP_A:.*]] = cir.load{{.*}} %[[A_ADDR]] : !cir.ptr<!cir.vector<4 x !s32i>>, !cir.vector<4 x !s32i> +// CIR: cir.store{{.*}} %[[TMP_A]], %[[EXCEPTION_ADDR]] : !cir.vector<4 x !s32i>, !cir.ptr<!cir.vector<4 x !s32i>> +// CIR: cir.throw %[[EXCEPTION_ADDR]] : !cir.ptr<!cir.vector<4 x !s32i>>, @_ZTIDv4_i +// CIR: cir.unreachable + +// LLVM: %[[A_ADDR:.*]] = alloca <4 x i32>, i64 1, align 16 +// LLVM: %[[EXCEPTION_ADDR:.*]] = call ptr @__cxa_allocate_exception(i64 16) +// LLVM: %[[TMP_A:.*]] = load <4 x i32>, ptr %[[A_ADDR]], align 16 +// LLVM: store <4 x i32> %[[TMP_A]], ptr %[[EXCEPTION_ADDR]], align 16 +// LLVM: call void @__cxa_throw(ptr %[[EXCEPTION_ADDR]], ptr @_ZTIDv4_i, ptr null) + +// OGCG: %[[A_ADDR:.*]] = alloca <4 x i32>, align 16 +// OGCG: %[[EXCEPTION_ADDR:.*]] = call ptr @__cxa_allocate_exception(i64 16) +// OGCG: %[[TMP_A:.*]] = load <4 x i32>, ptr %[[A_ADDR]], align 16 +// OGCG: store <4 x i32> %[[TMP_A]], ptr %[[EXCEPTION_ADDR]], align 16 +// OGCG: call void @__cxa_throw(ptr %[[EXCEPTION_ADDR]], ptr @_ZTIDv4_i, ptr null) +// OGCG: unreachable + +void throw_enum_expr() { + enum Test { + TestA, + TestB + }; + throw Test::TestA; +} + +// CIR: %[[EXCEPTION_ADDR:.*]] = cir.alloc.exception 4 -> !cir.ptr<!u32i> +// CIR: %[[EXCEPTION_VALUE:.*]] = cir.const #cir.int<0> : !u32i +// CIR: cir.store{{.*}} %[[EXCEPTION_VALUE]], %[[EXCEPTION_ADDR]] : !u32i, !cir.ptr<!u32i> +// CIR: cir.throw %[[EXCEPTION_ADDR]] : !cir.ptr<!u32i>, @_ZTIZ15throw_enum_exprvE4Test +// CIR: cir.unreachable + +// LLVM: %[[EXCEPTION_ADDR:.*]] = call ptr @__cxa_allocate_exception(i64 4) +// LLVM: store i32 0, ptr %[[EXCEPTION_ADDR]], align 16 +// LLVM: call void @__cxa_throw(ptr %[[EXCEPTION_ADDR]], ptr @_ZTIZ15throw_enum_exprvE4Test, ptr null) +// LLVM: unreachable + +// OGCG: %[[EXCEPTION_ADDR:.*]] = call ptr @__cxa_allocate_exception(i64 4) +// OGCG: store i32 0, ptr %[[EXCEPTION_ADDR]], align 16 +// OGCG: call void @__cxa_throw(ptr %[[EXCEPTION_ADDR]], ptr @_ZTIZ15throw_enum_exprvE4Test, ptr null) +// OGCG: unreachable + +void throw_enum_class_expr() { + enum class Test { + TestA, + TestB + }; + throw Test::TestA; +} + +// CIR: %[[EXCEPTION_ADDR:.*]] = cir.alloc.exception 4 -> !cir.ptr<!s32i> +// CIR: %[[EXCEPTION_VALUE:.*]] = cir.const #cir.int<0> : !s32i +// CIR: cir.store{{.*}} %[[EXCEPTION_VALUE]], %[[EXCEPTION_ADDR]] : !s32i, !cir.ptr<!s32i> +// CIR: cir.throw %[[EXCEPTION_ADDR]] : !cir.ptr<!s32i>, @_ZTIZ21throw_enum_class_exprvE4Test +// CIR: cir.unreachable + +// LLVM: %[[EXCEPTION_ADDR:.*]] = call ptr @__cxa_allocate_exception(i64 4) +// LLVM: store i32 0, ptr %[[EXCEPTION_ADDR]], align 16 +// LLVM: call void @__cxa_throw(ptr %[[EXCEPTION_ADDR]], ptr @_ZTIZ21throw_enum_class_exprvE4Test, ptr null) +// LLVM: unreachable + +// OGCG: %[[EXCEPTION_ADDR:.*]] = call ptr @__cxa_allocate_exception(i64 4) +// OGCG: store i32 0, ptr %[[EXCEPTION_ADDR]], align 16 +// OGCG: call void @__cxa_throw(ptr %[[EXCEPTION_ADDR]], ptr @_ZTIZ21throw_enum_class_exprvE4Test, ptr null) +// OGCG: unreachable diff --git a/clang/test/CIR/CodeGen/vbase.cpp b/clang/test/CIR/CodeGen/vbase.cpp index 9e42323..8fcb2a4 100644 --- a/clang/test/CIR/CodeGen/vbase.cpp +++ b/clang/test/CIR/CodeGen/vbase.cpp @@ -57,7 +57,7 @@ void ppp() { B b; } // OGCG: @_ZTV1B = linkonce_odr unnamed_addr constant { [3 x ptr] } { [3 x ptr] [ptr inttoptr (i64 12 to ptr), ptr null, ptr @_ZTI1B] }, comdat, align 8 -// CIR: cir.func {{.*}}@_Z1fv() { +// CIR: cir.func {{.*}}@_Z1fv() // CIR: %[[D:.+]] = cir.alloca !rec_Derived, !cir.ptr<!rec_Derived>, ["d", init] // CIR: cir.call @_ZN7DerivedC1Ev(%[[D]]) nothrow : (!cir.ptr<!rec_Derived>) -> () // CIR: %[[VPTR_PTR:.+]] = cir.vtable.get_vptr %[[D]] : !cir.ptr<!rec_Derived> -> !cir.ptr<!cir.vptr> @@ -78,7 +78,7 @@ void ppp() { B b; } // CIR: cir.call %[[FN]](%[[BASE_THIS]]) : (!cir.ptr<!cir.func<(!cir.ptr<!rec_Base>)>>, !cir.ptr<!rec_Base>) -> () // CIR: cir.return -// CIR: cir.func {{.*}}@_Z1gv() { +// CIR: cir.func {{.*}}@_Z1gv() // CIR: %[[DF:.+]] = cir.alloca !rec_DerivedFinal, !cir.ptr<!rec_DerivedFinal>, ["df", init] // CIR: cir.call @_ZN12DerivedFinalC1Ev(%[[DF]]) nothrow : (!cir.ptr<!rec_DerivedFinal>) -> () // CIR: %[[BASE_THIS_2:.+]] = cir.base_class_addr %[[DF]] : !cir.ptr<!rec_DerivedFinal> nonnull [0] -> !cir.ptr<!rec_Base> @@ -89,7 +89,7 @@ void ppp() { B b; } // CIR: cir.call %[[FN_2]](%[[BASE_THIS_2]]) : (!cir.ptr<!cir.func<(!cir.ptr<!rec_Base>)>>, !cir.ptr<!rec_Base>) -> () // CIR: cir.return -// LLVM: define {{.*}}void @_Z1fv() +// LLVM: define {{.*}}void @_Z1fv(){{.*}} // LLVM: %[[D:.+]] = alloca {{.*}} // LLVM: call void @_ZN7DerivedC1Ev(ptr %[[D]]) // LLVM: %[[VPTR_ADDR:.+]] = load ptr, ptr %[[D]] @@ -102,7 +102,7 @@ void ppp() { B b; } // LLVM: call void %[[VFN]](ptr %[[ADJ_THIS]]) // LLVM: ret void -// LLVM: define {{.*}}void @_Z1gv() +// LLVM: define {{.*}}void @_Z1gv(){{.*}} // LLVM: %[[DF:.+]] = alloca {{.*}} // LLVM: call void @_ZN12DerivedFinalC1Ev(ptr %[[DF]]) // LLVM: %[[VPTR2:.+]] = load ptr, ptr %[[DF]] @@ -138,7 +138,7 @@ void ppp() { B b; } // CIR: cir.store align(8) %[[VTABLE]], %[[B_VPTR]] : !cir.vptr, !cir.ptr<!cir.vptr> // CIR: cir.return -// LLVM: define{{.*}} void @_ZN1BC1Ev(ptr %[[THIS_ARG:.*]]) { +// LLVM: define{{.*}} void @_ZN1BC1Ev(ptr %[[THIS_ARG:.*]]){{.*}} { // LLVM: %[[THIS_ADDR:.*]] = alloca ptr // LLVM: store ptr %[[THIS_ARG]], ptr %[[THIS_ADDR]] // LLVM: %[[THIS:.*]] = load ptr, ptr %[[THIS_ADDR]] diff --git a/clang/test/CIR/CodeGen/virtual-destructor-calls.cpp b/clang/test/CIR/CodeGen/virtual-destructor-calls.cpp new file mode 100644 index 0000000..08a6b21 --- /dev/null +++ b/clang/test/CIR/CodeGen/virtual-destructor-calls.cpp @@ -0,0 +1,129 @@ +// RUN: %clang_cc1 -triple aarch64-none-linux-android21 -std=c++20 -mconstructor-aliases -O0 -fclangir -emit-cir %s -o %t.cir +// RUN: FileCheck --check-prefix=CIR --input-file=%t.cir %s +// RUN: %clang_cc1 -triple aarch64-none-linux-android21 -std=c++20 -mconstructor-aliases -O0 -fclangir -emit-llvm %s -o %t-cir.ll +// RUN: FileCheck --check-prefix=LLVM --input-file=%t-cir.ll %s +// RUN: %clang_cc1 -triple aarch64-none-linux-android21 -std=c++20 -mconstructor-aliases -O0 -emit-llvm %s -o %t.ll +// RUN: FileCheck --check-prefix=OGCG --input-file=%t.ll %s + +// TODO(cir): Try to emit base destructor as an alias at O1 or higher. + +// FIXME: LLVM IR dialect does not yet support function ptr globals, which precludes +// a lot of the proper semantics for properly representing alias functions in LLVM +// (see the note on LLVM_O1 below). + +struct Member { + ~Member(); +}; + +struct A { + virtual ~A(); +}; + +struct B : A { + Member m; + virtual ~B(); +}; + +B::~B() { } + +// Aliases are inserted before the function definitions in LLVM IR +// FIXME: These should have unnamed_addr set. +// LLVM: @_ZN1BD1Ev = alias void (ptr), ptr @_ZN1BD2Ev +// LLVM: @_ZN1CD1Ev = alias void (ptr), ptr @_ZN1CD2Ev + +// OGCG: @_ZN1BD1Ev = unnamed_addr alias void (ptr), ptr @_ZN1BD2Ev +// OGCG: @_ZN1CD1Ev = unnamed_addr alias void (ptr), ptr @_ZN1CD2Ev + + +// Base (D2) dtor for B: calls A's base dtor. + +// CIR: cir.func{{.*}} @_ZN1BD2Ev +// CIR: cir.call @_ZN6MemberD1Ev +// CIR: cir.call @_ZN1AD2Ev + +// LLVM: define{{.*}} void @_ZN1BD2Ev +// LLVM: call void @_ZN6MemberD1Ev +// LLVM: call void @_ZN1AD2Ev + +// OGCG: define{{.*}} @_ZN1BD2Ev +// OGCG: call void @_ZN6MemberD1Ev +// OGCG: call void @_ZN1AD2Ev + +// Complete (D1) dtor for B: just an alias because there are no virtual bases. + +// CIR: cir.func{{.*}} @_ZN1BD1Ev(!cir.ptr<!rec_B>) alias(@_ZN1BD2Ev) +// This is defined above for LLVM and OGCG. + +// Deleting (D0) dtor for B: defers to the complete dtor but also calls operator delete. + +// CIR: cir.func{{.*}} @_ZN1BD0Ev +// CIR: cir.call @_ZN1BD1Ev(%[[THIS:.*]]) nothrow : (!cir.ptr<!rec_B>) -> () +// CIR: %[[THIS_VOID:.*]] = cir.cast bitcast %[[THIS]] : !cir.ptr<!rec_B> -> !cir.ptr<!void> +// CIR: %[[SIZE:.*]] = cir.const #cir.int<16> +// CIR: cir.call @_ZdlPvm(%[[THIS_VOID]], %[[SIZE]]) + +// LLVM: define{{.*}} void @_ZN1BD0Ev +// LLVM: call void @_ZN1BD1Ev(ptr %[[THIS:.*]]) +// LLVM: call void @_ZdlPvm(ptr %[[THIS]], i64 16) + +// OGCG: define{{.*}} @_ZN1BD0Ev +// OGCG: call void @_ZN1BD1Ev(ptr{{.*}} %[[THIS:.*]]) +// OGCG: call void @_ZdlPvm(ptr{{.*}} %[[THIS]], i64{{.*}} 16) + +struct C : B { + ~C(); +}; + +C::~C() { } + +// Base (D2) dtor for C: calls B's base dtor. + +// CIR: cir.func{{.*}} @_ZN1CD2Ev +// CIR: %[[B:.*]] = cir.base_class_addr %[[THIS:.*]] : !cir.ptr<!rec_C> nonnull [0] -> !cir.ptr<!rec_B> +// CIR: cir.call @_ZN1BD2Ev(%[[B]]) + +// LLVM: define{{.*}} void @_ZN1CD2Ev +// LLVM: call void @_ZN1BD2Ev + +// OGCG: define{{.*}} @_ZN1CD2Ev +// OGCG: call void @_ZN1BD2Ev + +// Complete (D1) dtor for C: just an alias because there are no virtual bases. + +// CIR: cir.func{{.*}} @_ZN1CD1Ev(!cir.ptr<!rec_C>) alias(@_ZN1CD2Ev) +// This is defined above for LLVM and OGCG. + + +// Deleting (D0) dtor for C: defers to the complete dtor but also calls operator delete. + +// CIR: cir.func{{.*}} @_ZN1CD0Ev +// CIR: cir.call @_ZN1CD1Ev(%[[THIS:.*]]) nothrow : (!cir.ptr<!rec_C>) -> () +// CIR: %[[THIS_VOID:.*]] = cir.cast bitcast %[[THIS]] : !cir.ptr<!rec_C> -> !cir.ptr<!void> +// CIR: %[[SIZE:.*]] = cir.const #cir.int<16> +// CIR: cir.call @_ZdlPvm(%[[THIS_VOID]], %[[SIZE]]) + +// LLVM: define{{.*}} void @_ZN1CD0Ev +// LLVM: call void @_ZN1CD1Ev(ptr %[[THIS:.*]]) +// LLVM: call void @_ZdlPvm(ptr %[[THIS]], i64 16) + +// OGCG: define{{.*}} @_ZN1CD0Ev +// OGCG: call void @_ZN1CD1Ev(ptr{{.*}} %[[THIS:.*]]) +// OGCG: call void @_ZdlPvm(ptr{{.*}} %[[THIS]], i64{{.*}} 16) + +namespace PR12798 { + // A qualified call to a base class destructor should not undergo virtual + // dispatch. Template instantiation used to lose the qualifier. + struct A { virtual ~A(); }; + template<typename T> void f(T *p) { p->A::~A(); } + + // CIR: cir.func{{.*}} @_ZN7PR127981fINS_1AEEEvPT_ + // CIR: cir.call @_ZN7PR127981AD1Ev + + // LLVM: define{{.*}} @_ZN7PR127981fINS_1AEEEvPT_ + // LLVM: call void @_ZN7PR127981AD1Ev + + // OGCG: define{{.*}} @_ZN7PR127981fINS_1AEEEvPT_ + // OGCG: call void @_ZN7PR127981AD1Ev + + template void f(A*); +} diff --git a/clang/test/CIR/CodeGen/vla.c b/clang/test/CIR/CodeGen/vla.c new file mode 100644 index 0000000..b22c704 --- /dev/null +++ b/clang/test/CIR/CodeGen/vla.c @@ -0,0 +1,285 @@ +// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -Wno-unused-value -fclangir -emit-cir %s -o %t.cir +// RUN: FileCheck --input-file=%t.cir %s -check-prefix=CIR +// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -Wno-unused-value -fclangir -emit-llvm %s -o %t-cir.ll +// RUN: FileCheck --input-file=%t-cir.ll %s -check-prefix=LLVM +// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -Wno-unused-value -emit-llvm %s -o %t.ll +// RUN: FileCheck --input-file=%t.ll %s -check-prefix=OGCG + +void f0(int len) { + int arr[len]; +} + +// CIR: cir.func{{.*}} @f0(%[[LEN_ARG:.*]]: !s32i {{.*}}) +// CIR: %[[LEN_ADDR:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["len", init] +// CIR: %[[SAVED_STACK:.*]] = cir.alloca !cir.ptr<!u8i>, !cir.ptr<!cir.ptr<!u8i>>, ["saved_stack"] +// CIR: cir.store{{.*}} %[[LEN_ARG]], %[[LEN_ADDR]] +// CIR: %[[LEN:.*]] = cir.load{{.*}} %[[LEN_ADDR]] +// CIR: %[[LEN_SIZE_T:.*]] = cir.cast integral %[[LEN]] : !s32i -> !u64i +// CIR: %[[STACK_PTR:.*]] = cir.stacksave +// CIR: cir.store{{.*}} %[[STACK_PTR]], %[[SAVED_STACK]] +// CIR: %[[ARR:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, %[[LEN_SIZE_T]] : !u64i, ["arr"] +// CIR: %[[STACK_RESTORE_PTR:.*]] = cir.load{{.*}} %[[SAVED_STACK]] +// CIR: cir.stackrestore %[[STACK_RESTORE_PTR]] + +// LLVM: define{{.*}} void @f0(i32 %[[LEN_ARG:.*]]) +// LLVM: %[[LEN_ADDR:.*]] = alloca i32 +// LLVM: %[[SAVED_STACK:.*]] = alloca ptr +// LLVM: store i32 %[[LEN_ARG]], ptr %[[LEN_ADDR]] +// LLVM: %[[LEN:.*]] = load i32, ptr %[[LEN_ADDR]] +// LLVM: %[[LEN_SIZE_T:.*]] = sext i32 %[[LEN]] to i64 +// LLVM: %[[STACK_PTR:.*]] = call ptr @llvm.stacksave.p0() +// LLVM: store ptr %[[STACK_PTR]], ptr %[[SAVED_STACK]] +// LLVM: %[[ARR:.*]] = alloca i32, i64 %[[LEN_SIZE_T]] +// LLVM: %[[STACK_RESTORE_PTR:.*]] = load ptr, ptr %[[SAVED_STACK]] +// LLVM: call void @llvm.stackrestore.p0(ptr %[[STACK_RESTORE_PTR]]) + +// Note: VLA_EXPR0 below is emitted to capture debug info. + +// OGCG: define{{.*}} void @f0(i32 {{.*}} %[[LEN_ARG:.*]]) +// OGCG: %[[LEN_ADDR:.*]] = alloca i32 +// OGCG: %[[SAVED_STACK:.*]] = alloca ptr +// OGCG: %[[VLA_EXPR0:.*]] = alloca i64 +// OGCG: store i32 %[[LEN_ARG]], ptr %[[LEN_ADDR]] +// OGCG: %[[LEN:.*]] = load i32, ptr %[[LEN_ADDR]] +// OGCG: %[[LEN_SIZE_T:.*]] = zext i32 %[[LEN]] to i64 +// OGCG: %[[STACK_PTR:.*]] = call ptr @llvm.stacksave.p0() +// OGCG: store ptr %[[STACK_PTR]], ptr %[[SAVED_STACK]] +// OGCG: %[[ARR:.*]] = alloca i32, i64 %[[LEN_SIZE_T]] +// OGCG: store i64 %[[LEN_SIZE_T]], ptr %[[VLA_EXPR0]] +// OGCG: %[[STACK_RESTORE_PTR:.*]] = load ptr, ptr %[[SAVED_STACK]] +// OGCG: call void @llvm.stackrestore.p0(ptr %[[STACK_RESTORE_PTR]]) + +void f1(int len) { + int arr[16][len]; +} + +// CIR: cir.func{{.*}} @f1(%[[LEN_ARG:.*]]: !s32i {{.*}}) +// CIR: %[[LEN_ADDR:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["len", init] +// CIR: %[[SAVED_STACK:.*]] = cir.alloca !cir.ptr<!u8i>, !cir.ptr<!cir.ptr<!u8i>>, ["saved_stack"] +// CIR: cir.store{{.*}} %[[LEN_ARG]], %[[LEN_ADDR]] +// CIR: %[[SIXTEEN:.*]] = cir.const #cir.int<16> : !s32i +// CIR: %[[SIXTEEN_SIZE_T:.*]] = cir.cast integral %[[SIXTEEN]] : !s32i -> !u64i +// CIR: %[[LEN:.*]] = cir.load{{.*}} %[[LEN_ADDR]] +// CIR: %[[LEN_SIZE_T:.*]] = cir.cast integral %[[LEN]] : !s32i -> !u64i +// CIR: %[[STACK_PTR:.*]] = cir.stacksave +// CIR: cir.store{{.*}} %[[STACK_PTR]], %[[SAVED_STACK]] +// CIR: %[[TOTAL_LEN:.*]] = cir.binop(mul, %[[SIXTEEN_SIZE_T]], %[[LEN_SIZE_T]]) nuw +// CIR: %[[ARR:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, %[[TOTAL_LEN]] : !u64i, ["arr"] +// CIR: %[[STACK_RESTORE_PTR:.*]] = cir.load{{.*}} %[[SAVED_STACK]] +// CIR: cir.stackrestore %[[STACK_RESTORE_PTR]] + +// LLVM: define{{.*}} void @f1(i32 %[[LEN_ARG:.*]]) +// LLVM: %[[LEN_ADDR:.*]] = alloca i32 +// LLVM: %[[SAVED_STACK:.*]] = alloca ptr +// LLVM: store i32 %[[LEN_ARG]], ptr %[[LEN_ADDR]] +// LLVM: %[[LEN:.*]] = load i32, ptr %[[LEN_ADDR]] +// LLVM: %[[LEN_SIZE_T:.*]] = sext i32 %[[LEN]] to i64 +// LLVM: %[[STACK_PTR:.*]] = call ptr @llvm.stacksave.p0() +// LLVM: store ptr %[[STACK_PTR]], ptr %[[SAVED_STACK]] +// LLVM: %[[TOTAL_LEN:.*]] = mul nuw i64 16, %[[LEN_SIZE_T]] +// LLVM: %[[ARR:.*]] = alloca i32, i64 %[[TOTAL_LEN]] +// LLVM: %[[STACK_RESTORE_PTR:.*]] = load ptr, ptr %[[SAVED_STACK]] +// LLVM: call void @llvm.stackrestore.p0(ptr %[[STACK_RESTORE_PTR]]) + +// Note: VLA_EXPR0 below is emitted to capture debug info. + +// OGCG: define{{.*}} void @f1(i32 {{.*}} %[[LEN_ARG:.*]]) +// OGCG: %[[LEN_ADDR:.*]] = alloca i32 +// OGCG: %[[SAVED_STACK:.*]] = alloca ptr +// OGCG: %[[VLA_EXPR0:.*]] = alloca i64 +// OGCG: store i32 %[[LEN_ARG]], ptr %[[LEN_ADDR]] +// OGCG: %[[LEN:.*]] = load i32, ptr %[[LEN_ADDR]] +// OGCG: %[[LEN_SIZE_T:.*]] = zext i32 %[[LEN]] to i64 +// OGCG: %[[STACK_PTR:.*]] = call ptr @llvm.stacksave.p0() +// OGCG: store ptr %[[STACK_PTR]], ptr %[[SAVED_STACK]] +// OGCG: %[[TOTAL_LEN:.*]] = mul nuw i64 16, %[[LEN_SIZE_T]] +// OGCG: %[[ARR:.*]] = alloca i32, i64 %[[TOTAL_LEN]] +// OGCG: store i64 %[[LEN_SIZE_T]], ptr %[[VLA_EXPR0]] +// OGCG: %[[STACK_RESTORE_PTR:.*]] = load ptr, ptr %[[SAVED_STACK]] +// OGCG: call void @llvm.stackrestore.p0(ptr %[[STACK_RESTORE_PTR]]) + +void f2(int len) { + int arr[len + 4]; +} + +// CIR: cir.func{{.*}} @f2(%[[LEN_ARG:.*]]: !s32i {{.*}}) +// CIR: %[[LEN_ADDR:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["len", init] +// CIR: %[[SAVED_STACK:.*]] = cir.alloca !cir.ptr<!u8i>, !cir.ptr<!cir.ptr<!u8i>>, ["saved_stack"] +// CIR: cir.store{{.*}} %[[LEN_ARG]], %[[LEN_ADDR]] +// CIR: %[[LEN:.*]] = cir.load{{.*}} %[[LEN_ADDR]] +// CIR: %[[FOUR:.*]] = cir.const #cir.int<4> : !s32i +// CIR: %[[TOTAL_LEN:.*]] = cir.binop(add, %[[LEN]], %[[FOUR]]) nsw : !s32i +// CIR: %[[TOTAL_LEN_SIZE_T:.*]] = cir.cast integral %[[TOTAL_LEN]] : !s32i -> !u64i +// CIR: %[[STACK_PTR:.*]] = cir.stacksave +// CIR: cir.store{{.*}} %[[STACK_PTR]], %[[SAVED_STACK]] +// CIR: %[[ARR:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, %[[TOTAL_LEN_SIZE_T]] : !u64i, ["arr"] +// CIR: %[[STACK_RESTORE_PTR:.*]] = cir.load{{.*}} %[[SAVED_STACK]] +// CIR: cir.stackrestore %[[STACK_RESTORE_PTR]] + +// LLVM: define{{.*}} void @f2(i32 %[[LEN_ARG:.*]]) +// LLVM: %[[LEN_ADDR:.*]] = alloca i32 +// LLVM: %[[SAVED_STACK:.*]] = alloca ptr +// LLVM: store i32 %[[LEN_ARG]], ptr %[[LEN_ADDR]] +// LLVM: %[[LEN:.*]] = load i32, ptr %[[LEN_ADDR]] +// LLVM: %[[TOTAL_LEN:.*]] = add nsw i32 %[[LEN]], 4 +// LLVM: %[[TOTAL_LEN_SIZE_T:.*]] = sext i32 %[[TOTAL_LEN]] to i64 +// LLVM: %[[STACK_PTR:.*]] = call ptr @llvm.stacksave.p0() +// LLVM: store ptr %[[STACK_PTR]], ptr %[[SAVED_STACK]] +// LLVM: %[[ARR:.*]] = alloca i32, i64 %[[TOTAL_LEN_SIZE_T]] +// LLVM: %[[STACK_RESTORE_PTR:.*]] = load ptr, ptr %[[SAVED_STACK]] +// LLVM: call void @llvm.stackrestore.p0(ptr %[[STACK_RESTORE_PTR]]) + +// Note: VLA_EXPR0 below is emitted to capture debug info. + +// OGCG: define{{.*}} void @f2(i32 {{.*}} %[[LEN_ARG:.*]]) +// OGCG: %[[LEN_ADDR:.*]] = alloca i32 +// OGCG: %[[SAVED_STACK:.*]] = alloca ptr +// OGCG: %[[VLA_EXPR0:.*]] = alloca i64 +// OGCG: store i32 %[[LEN_ARG]], ptr %[[LEN_ADDR]] +// OGCG: %[[LEN:.*]] = load i32, ptr %[[LEN_ADDR]] +// OGCG: %[[TOTAL_LEN:.*]] = add nsw i32 %[[LEN]], 4 +// OGCG: %[[TOTAL_LEN_SIZE_T:.*]] = zext i32 %[[TOTAL_LEN]] to i64 +// OGCG: %[[STACK_PTR:.*]] = call ptr @llvm.stacksave.p0() +// OGCG: store ptr %[[STACK_PTR]], ptr %[[SAVED_STACK]] +// OGCG: %[[ARR:.*]] = alloca i32, i64 %[[TOTAL_LEN_SIZE_T]] +// OGCG: store i64 %[[TOTAL_LEN_SIZE_T]], ptr %[[VLA_EXPR0]] +// OGCG: %[[STACK_RESTORE_PTR:.*]] = load ptr, ptr %[[SAVED_STACK]] +// OGCG: call void @llvm.stackrestore.p0(ptr %[[STACK_RESTORE_PTR]]) + +void f3(unsigned len) { + char s1[len]; + unsigned i = 0u; + while (++i < len) { + char s2[i]; + } +} + +// CIR: cir.func{{.*}} @f3(%[[LEN_ARG:.*]]: !u32i {{.*}}) +// CIR: %[[LEN_ADDR:.*]] = cir.alloca !u32i, !cir.ptr<!u32i>, ["len", init] +// CIR: %[[SAVED_STACK:.*]] = cir.alloca !cir.ptr<!u8i>, !cir.ptr<!cir.ptr<!u8i>>, ["saved_stack"] +// CIR: cir.store{{.*}} %[[LEN_ARG]], %[[LEN_ADDR]] +// CIR: %[[LEN:.*]] = cir.load{{.*}} %[[LEN_ADDR]] +// CIR: %[[LEN_SIZE_T:.*]] = cir.cast integral %[[LEN]] : !u32i -> !u64i +// CIR: %[[STACK_PTR:.*]] = cir.stacksave +// CIR: cir.store{{.*}} %[[STACK_PTR]], %[[SAVED_STACK]] +// CIR: %[[S1:.*]] = cir.alloca !s8i, !cir.ptr<!s8i>, %[[LEN_SIZE_T]] : !u64i, ["s1"] +// CIR: %[[I:.*]] = cir.alloca !u32i, !cir.ptr<!u32i>, ["i", init] +// CIR: %[[ZERO:.*]] = cir.const #cir.int<0> : !u32i +// CIR: cir.store{{.*}} %[[ZERO]], %[[I]] +// CIR: cir.scope { +// CIR: cir.while { +// CIR: %[[CUR_I:.*]] = cir.load{{.*}} %[[I]] +// CIR: %[[NEXT:.*]] = cir.unary(inc, %[[CUR_I]]) +// CIR: cir.store{{.*}} %[[NEXT]], %[[I]] +// CIR: %[[LEN2:.*]] = cir.load{{.*}} %[[LEN_ADDR]] +// CIR: %[[CMP:.*]] = cir.cmp(lt, %[[NEXT]], %[[LEN2]]) +// CIR: cir.condition(%[[CMP]]) +// CIR: } do { +// CIR: cir.scope { +// CIR: %[[SAVED_STACK2:.*]] = cir.alloca !cir.ptr<!u8i>, !cir.ptr<!cir.ptr<!u8i>>, ["saved_stack"] +// CIR: %[[I_LEN:.*]] = cir.load{{.*}} %[[I]] +// CIR: %[[I_LEN_SIZE_T2:.*]] = cir.cast integral %[[I_LEN]] : !u32i -> !u64i +// CIR: %[[STACK_PTR2:.*]] = cir.stacksave +// CIR: cir.store{{.*}} %[[STACK_PTR2]], %[[SAVED_STACK2]] +// CIR: %[[S2:.*]] = cir.alloca !s8i, !cir.ptr<!s8i>, %[[I_LEN_SIZE_T2]] : !u64i, ["s2"] +// CIR: %[[SAVED_RESTORE_PTR2:.*]] = cir.load{{.*}} %[[SAVED_STACK2]] +// CIR: cir.stackrestore %[[SAVED_RESTORE_PTR2]] +// CIR: } +// CIR: cir.yield +// CIR: } +// CIR: } +// CIR: %[[STACK_RESTORE_PTR:.*]] = cir.load{{.*}} %[[SAVED_STACK]] +// CIR: cir.stackrestore %[[STACK_RESTORE_PTR]] + +// LLVM: define{{.*}} void @f3(i32 %[[LEN_ARG:.*]]) +// LLVM: %[[SAVED_STACK2:.*]] = alloca ptr +// LLVM: %[[LEN_ADDR:.*]] = alloca i32 +// LLVM: %[[SAVED_STACK:.*]] = alloca ptr +// LLVM: store i32 %[[LEN_ARG]], ptr %[[LEN_ADDR]] +// LLVM: %[[LEN:.*]] = load i32, ptr %[[LEN_ADDR]] +// LLVM: %[[LEN_SIZE_T:.*]] = zext i32 %[[LEN]] to i64 +// LLVM: %[[STACK_PTR:.*]] = call ptr @llvm.stacksave.p0() +// LLVM: store ptr %[[STACK_PTR]], ptr %[[SAVED_STACK]] +// LLVM: %[[S1:.*]] = alloca i8, i64 %[[LEN_SIZE_T]] +// LLVM: %[[I:.*]] = alloca i32 +// LLVM: store i32 0, ptr %[[I]] +// LLVM: br label %[[WHILE_START:.*]] +// LLVM: [[WHILE_START]]: +// LLVM: br label %[[WHILE_COND:.*]] +// LLVM: [[WHILE_COND]]: +// LLVM: %[[CUR_I:.*]] = load i32, ptr %[[I]] +// LLVM: %[[NEXT:.*]] = add i32 %[[CUR_I]], 1 +// LLVM: store i32 %[[NEXT]], ptr %[[I]] +// LLVM: %[[LEN2:.*]] = load i32, ptr %[[LEN_ADDR]] +// LLVM: %[[CMP:.*]] = icmp ult i32 %[[NEXT]], %[[LEN2]] +// LLVM: br i1 %[[CMP]], label %[[WHILE_BODY:.*]], label %[[WHILE_END:.*]] +// LLVM: [[WHILE_BODY]]: +// LLVM: br label %[[WHILE_BODY2:.*]] +// LLVM: [[WHILE_BODY2]]: +// LLVM: %[[I_LEN:.*]] = load i32, ptr %[[I]] +// LLVM: %[[I_LEN_SIZE_T2:.*]] = zext i32 %[[I_LEN]] to i64 +// LLVM: %[[STACK_PTR2:.*]] = call ptr @llvm.stacksave.p0() +// LLVM: store ptr %[[STACK_PTR2]], ptr %[[SAVED_STACK2]] +// LLVM: %[[S2:.*]] = alloca i8, i64 %[[I_LEN_SIZE_T2]] +// LLVM: %[[STACK_RESTORE_PTR2:.*]] = load ptr, ptr %[[SAVED_STACK2]] +// LLVM: call void @llvm.stackrestore.p0(ptr %[[STACK_RESTORE_PTR2]]) +// LLVM: br label %[[WHILE_BODY_END:.*]] +// LLVM: [[WHILE_BODY_END]]: +// LLVM: br label %[[WHILE_COND]] +// LLVM: [[WHILE_END]]: +// LLVM: br label %[[F3_END:.*]] +// LLVM: [[F3_END]]: +// LLVM: %[[STACK_RESTORE_PTR:.*]] = load ptr, ptr %[[SAVED_STACK]] +// LLVM: call void @llvm.stackrestore.p0(ptr %[[STACK_RESTORE_PTR]]) + +// Note: VLA_EXPR0 and VLA_EXPR1 below are emitted to capture debug info. + +// OGCG: define{{.*}} void @f3(i32 {{.*}} %[[LEN_ARG:.*]]) +// OGCG: %[[LEN_ADDR:.*]] = alloca i32 +// OGCG: %[[SAVED_STACK:.*]] = alloca ptr +// OGCG: %[[VLA_EXPR0:.*]] = alloca i64 +// OGCG: %[[I:.*]] = alloca i32 +// OGCG: %[[SAVED_STACK1:.*]] = alloca ptr +// OGCG: %[[VLA_EXPR1:.*]] = alloca i64 +// OGCG: store i32 %[[LEN_ARG]], ptr %[[LEN_ADDR]] +// OGCG: %[[LEN:.*]] = load i32, ptr %[[LEN_ADDR]] +// OGCG: %[[LEN_SIZE_T:.*]] = zext i32 %[[LEN]] to i64 +// OGCG: %[[STACK_PTR:.*]] = call ptr @llvm.stacksave.p0() +// OGCG: store ptr %[[STACK_PTR]], ptr %[[SAVED_STACK]] +// OGCG: %[[S1:.*]] = alloca i8, i64 %[[LEN_SIZE_T]] +// OGCG: store i64 %[[LEN_SIZE_T]], ptr %[[VLA_EXPR0]] +// OGCG: br label %[[WHILE_COND:.*]] +// OGCG: [[WHILE_COND]]: +// OGCG: %[[CUR_I:.*]] = load i32, ptr %[[I]] +// OGCG: %[[NEXT:.*]] = add i32 %[[CUR_I]], 1 +// OGCG: store i32 %[[NEXT]], ptr %[[I]] +// OGCG: %[[LEN2:.*]] = load i32, ptr %[[LEN_ADDR]] +// OGCG: %[[CMP:.*]] = icmp ult i32 %[[NEXT]], %[[LEN2]] +// OGCG: br i1 %[[CMP]], label %[[WHILE_BODY:.*]], label %[[WHILE_END:.*]] +// OGCG: [[WHILE_BODY]]: +// OGCG: %[[I_LEN:.*]] = load i32, ptr %[[I]] +// OGCG: %[[I_LEN_SIZE_T:.*]] = zext i32 %[[I_LEN]] to i64 +// OGCG: %[[STACK_PTR1:.*]] = call ptr @llvm.stacksave.p0() +// OGCG: store ptr %[[STACK_PTR1]], ptr %[[SAVED_STACK1]] +// OGCG: %[[S2:.*]] = alloca i8, i64 %[[I_LEN_SIZE_T]] +// OGCG: store i64 %[[I_LEN_SIZE_T]], ptr %[[VLA_EXPR1]] +// OGCG: %[[STACK_RESTORE_PTR1:.*]] = load ptr, ptr %[[SAVED_STACK1]] +// OGCG: call void @llvm.stackrestore.p0(ptr %[[STACK_RESTORE_PTR1]]) +// OGCG: br label %[[WHILE_COND]] +// OGCG: [[WHILE_END]]: +// OGCG: %[[STACK_RESTORE_PTR:.*]] = load ptr, ptr %[[SAVED_STACK]] +// OGCG: call void @llvm.stackrestore.p0(ptr %[[STACK_RESTORE_PTR]]) + + +// The following test case is disabled because it runs into a bug (unrelated +// to VLA) in the handling of cleanups in loops with break statements. +// +// void f4(unsigned len) { +// char s1[len]; +// while (1) { +// char s2[len]; +// if (1) +// break; +// } +// } +
\ No newline at end of file diff --git a/clang/test/CIR/CodeGen/vtt.cpp b/clang/test/CIR/CodeGen/vtt.cpp index dc30c32..f9a62e3 100644 --- a/clang/test/CIR/CodeGen/vtt.cpp +++ b/clang/test/CIR/CodeGen/vtt.cpp @@ -445,7 +445,7 @@ D::D() {} // CIR-COMMON: %[[C_VPTR_ADDR:.*]] = cir.vtable.get_vptr %[[C_ADDR]] : !cir.ptr<!rec_C> -> !cir.ptr<!cir.vptr> // CIR-COMMON: cir.store{{.*}} %[[C_VPTR]], %[[C_VPTR_ADDR]] : !cir.vptr, !cir.ptr<!cir.vptr> -// LLVM-COMMON: define {{.*}} void @_ZN1DC2Ev(ptr %[[THIS_ARG:.*]], ptr %[[VTT_ARG:.*]]) { +// LLVM-COMMON: define {{.*}} void @_ZN1DC2Ev(ptr %[[THIS_ARG:.*]], ptr %[[VTT_ARG:.*]]){{.*}} { // LLVM-COMMON: %[[THIS_ADDR:.*]] = alloca ptr // LLVM-COMMON: %[[VTT_ADDR:.*]] = alloca ptr // LLVM-COMMON: store ptr %[[THIS_ARG]], ptr %[[THIS_ADDR]] @@ -484,7 +484,7 @@ D::D() {} // CIR-COMMON: %[[VPTR_ADDR:.*]] = cir.vtable.get_vptr %[[THIS]] : !cir.ptr<!rec_A> -> !cir.ptr<!cir.vptr> // CIR-COMMON: cir.store{{.*}} %[[VPTR]], %[[VPTR_ADDR]] : !cir.vptr, !cir.ptr<!cir.vptr> -// LLVM-COMMON: define {{.*}} void @_ZN1AC2Ev(ptr %[[THIS_ARG:.*]]) { +// LLVM-COMMON: define {{.*}} void @_ZN1AC2Ev(ptr %[[THIS_ARG:.*]]){{.*}} { // LLVM-COMMON: %[[THIS_ADDR:.*]] = alloca ptr, i64 1, align 8 // LLVM-COMMON: store ptr %[[THIS_ARG]], ptr %[[THIS_ADDR]], align 8 // LLVM-COMMON: %[[THIS:.*]] = load ptr, ptr %[[THIS_ADDR]], align 8 diff --git a/clang/test/CIR/CodeGenOpenACC/cache.c b/clang/test/CIR/CodeGenOpenACC/cache.c index 76651c1..d82230a 100644 --- a/clang/test/CIR/CodeGenOpenACC/cache.c +++ b/clang/test/CIR/CodeGenOpenACC/cache.c @@ -1,7 +1,7 @@ // RUN: %clang_cc1 -fopenacc -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir %s -o - | FileCheck %s void acc_cache() { - // CHECK: cir.func{{.*}} @acc_cache() { + // CHECK: cir.func{{.*}} @acc_cache() int iArr[10]; // CHECK-NEXT: %[[IARR:.*]] = cir.alloca !cir.array<!s32i x 10>, !cir.ptr<!cir.array<!s32i x 10>>, ["iArr"] diff --git a/clang/test/CIR/CodeGenOpenACC/combined-copy.c b/clang/test/CIR/CodeGenOpenACC/combined-copy.c index 9afbab5..31956b3 100644 --- a/clang/test/CIR/CodeGenOpenACC/combined-copy.c +++ b/clang/test/CIR/CodeGenOpenACC/combined-copy.c @@ -2,7 +2,7 @@ int global; void acc_compute(int parmVar) { - // CHECK: cir.func{{.*}} @acc_compute(%[[ARG:.*]]: !s32i{{.*}}) { + // CHECK: cir.func{{.*}} @acc_compute(%[[ARG:.*]]: !s32i{{.*}}) // CHECK-NEXT: %[[PARM:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["parmVar", init] int localVar1; short localVar2; @@ -1082,7 +1082,7 @@ typedef struct OuterTy { } Outer; void copy_member_of_array_element_member() { - // CHECK: cir.func{{.*}} @copy_member_of_array_element_member() { + // CHECK: cir.func{{.*}} @copy_member_of_array_element_member() Outer outer; // CHECK-NEXT: %[[OUTER:.*]] = cir.alloca !rec_OuterTy, !cir.ptr<!rec_OuterTy>, ["outer"] @@ -1104,7 +1104,7 @@ void copy_member_of_array_element_member() { } void modifier_list() { - // CHECK: cir.func{{.*}} @modifier_list() { + // CHECK: cir.func{{.*}} @modifier_list() int localVar; // CHECK-NEXT: %[[LOCALVAR:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["localVar"] diff --git a/clang/test/CIR/CodeGenOpenACC/combined-firstprivate-clause.cpp b/clang/test/CIR/CodeGenOpenACC/combined-firstprivate-clause.cpp index aa1103d..94f3f1a 100644 --- a/clang/test/CIR/CodeGenOpenACC/combined-firstprivate-clause.cpp +++ b/clang/test/CIR/CodeGenOpenACC/combined-firstprivate-clause.cpp @@ -324,7 +324,7 @@ struct HasDtor { // extern "C" void acc_combined() { - // CHECK: cir.func{{.*}} @acc_combined() { + // CHECK: cir.func{{.*}} @acc_combined() int someInt; // CHECK-NEXT: %[[SOMEINT:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["someInt"] diff --git a/clang/test/CIR/CodeGenOpenACC/combined-private-clause.cpp b/clang/test/CIR/CodeGenOpenACC/combined-private-clause.cpp index 7a7338c..ee82757 100644 --- a/clang/test/CIR/CodeGenOpenACC/combined-private-clause.cpp +++ b/clang/test/CIR/CodeGenOpenACC/combined-private-clause.cpp @@ -158,7 +158,7 @@ struct HasDtor { // CHECK-NEXT: } extern "C" void acc_combined() { - // CHECK: cir.func{{.*}} @acc_combined() { + // CHECK: cir.func{{.*}} @acc_combined() int someInt; // CHECK-NEXT: %[[SOMEINT:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["someInt"] diff --git a/clang/test/CIR/CodeGenOpenACC/combined-reduction-clause-default-ops.cpp b/clang/test/CIR/CodeGenOpenACC/combined-reduction-clause-default-ops.cpp index 7b74b7c..ee4fffe 100644 --- a/clang/test/CIR/CodeGenOpenACC/combined-reduction-clause-default-ops.cpp +++ b/clang/test/CIR/CodeGenOpenACC/combined-reduction-clause-default-ops.cpp @@ -1,4 +1,4 @@ -// RUN: %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s +// RUN: not %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s struct DefaultOperators { int i; @@ -8,12 +8,19 @@ struct DefaultOperators { bool b; }; +struct DefaultOperatorsNoFloats { + int i; + unsigned int u; + bool b; +}; + template<typename T> void acc_combined() { T someVar; T someVarArr[5]; + struct DefaultOperatorsNoFloats someVarNoFloats; + struct DefaultOperatorsNoFloats someVarArrNoFloats[5]; #pragma acc parallel loop reduction(+:someVar) - for(int i=0;i < 5; ++i); // CHECK: acc.reduction.recipe @reduction_add__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <add> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperators, !cir.ptr<!rec_DefaultOperators>, ["openacc.reduction.init", init] @@ -36,11 +43,43 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.double +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !cir.double, !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) nsw : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[ADD]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> // CHECK-NEXT: } + for(int i = 0; i < 5; ++i); #pragma acc parallel loop reduction(*:someVar) - // CHECK-NEXT: acc.reduction.recipe @reduction_mul__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <mul> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperators, !cir.ptr<!rec_DefaultOperators>, ["openacc.reduction.init", init] @@ -63,10 +102,42 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.double +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !cir.double, !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) nsw : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[MUL]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc parallel loop reduction(max:someVar) // CHECK-NEXT: acc.reduction.recipe @reduction_max__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <max> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) @@ -93,7 +164,7 @@ void acc_combined() { // TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc parallel loop reduction(min:someVar) // CHECK-NEXT: acc.reduction.recipe @reduction_min__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <min> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) @@ -120,89 +191,130 @@ void acc_combined() { // TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(&:someVar) - -// CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <iand> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperators, !cir.ptr<!rec_DefaultOperators>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> + for(int i = 0; i < 5; ++i); +#pragma acc parallel loop reduction(&:someVarNoFloats) +// CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTS24DefaultOperatorsNoFloats : !cir.ptr<!rec_DefaultOperatorsNoFloats> reduction_operator <iand> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats>{{.*}}) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperatorsNoFloats, !cir.ptr<!rec_DefaultOperatorsNoFloats>, ["openacc.reduction.init", init] +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[ALLOCA]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[ALLOCA]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[ALLOCA]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[ALLOCA]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[ALLOCA]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[ALLOCA]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats> {{.*}}) +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[AND]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperatorsNoFloats> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(|:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <ior> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperators, !cir.ptr<!rec_DefaultOperators>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> + for(int i = 0; i < 5; ++i); +#pragma acc parallel loop reduction(|:someVarNoFloats) +// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTS24DefaultOperatorsNoFloats : !cir.ptr<!rec_DefaultOperatorsNoFloats> reduction_operator <ior> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats>{{.*}}) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperatorsNoFloats, !cir.ptr<!rec_DefaultOperatorsNoFloats>, ["openacc.reduction.init", init] +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[ALLOCA]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[ALLOCA]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[ALLOCA]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[ALLOCA]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[ALLOCA]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[ALLOCA]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #false // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats> {{.*}}) +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[OR]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperatorsNoFloats> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(^:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <xor> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperators, !cir.ptr<!rec_DefaultOperators>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> + for(int i = 0; i < 5; ++i); +#pragma acc parallel loop reduction(^:someVarNoFloats) +// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTS24DefaultOperatorsNoFloats : !cir.ptr<!rec_DefaultOperatorsNoFloats> reduction_operator <xor> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats>{{.*}}) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperatorsNoFloats, !cir.ptr<!rec_DefaultOperatorsNoFloats>, ["openacc.reduction.init", init] +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[ALLOCA]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[ALLOCA]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[ALLOCA]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[ALLOCA]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[ALLOCA]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[ALLOCA]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #false // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats> {{.*}}) +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[XOR]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperatorsNoFloats> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc parallel loop reduction(&&:someVar) // CHECK-NEXT: acc.reduction.recipe @reduction_land__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <land> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) @@ -229,7 +341,7 @@ void acc_combined() { // TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc parallel loop reduction(||:someVar) // CHECK-NEXT: acc.reduction.recipe @reduction_lor__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <lor> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) @@ -256,7 +368,7 @@ void acc_combined() { // TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc parallel loop reduction(+:someVarArr) // CHECK-NEXT: acc.reduction.recipe @reduction_add__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <add> init { @@ -297,10 +409,65 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> +// +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.double +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !cir.double, !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) nsw : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[ADD]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc parallel loop reduction(*:someVarArr) // CHECK-NEXT: acc.reduction.recipe @reduction_mul__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <mul> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}) @@ -399,11 +566,67 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> +// +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.double +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !cir.double, !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) nsw : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[MUL]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc parallel loop reduction(max:someVarArr) + // CHECK-NEXT: acc.reduction.recipe @reduction_max__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <max> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init", init] @@ -504,7 +727,7 @@ void acc_combined() { // TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc parallel loop reduction(min:someVarArr) // CHECK-NEXT: acc.reduction.recipe @reduction_min__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <min> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}) @@ -606,196 +829,283 @@ void acc_combined() { // TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(&:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <iand> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[DECAY]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> + for(int i = 0; i < 5; ++i); +#pragma acc parallel loop reduction(&:someVarArrNoFloats) +// CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTSA5_24DefaultOperatorsNoFloats : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> reduction_operator <iand> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>{{.*}}) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperatorsNoFloats x 5>, !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>, ["openacc.reduction.init", init] +// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[DECAY]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[DECAY]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[DECAY]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[DECAY]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[DECAY]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[DECAY]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[DECAY]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // // CHECK-NEXT: %[[ALL_ONES_IDX:.*]] = cir.const #cir.int<1> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[ALL_ONES_IDX]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[ALL_ONES_IDX]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[NEXT_ELT]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // // CHECK-NEXT: %[[TWO_IDX:.*]] = cir.const #cir.int<2> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[TWO_IDX]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[TWO_IDX]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[NEXT_ELT]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // // // CHECK-NEXT: %[[THREE_IDX:.*]] = cir.const #cir.int<3> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[THREE_IDX]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[THREE_IDX]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[NEXT_ELT]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // // CHECK-NEXT: %[[FOUR_IDX:.*]] = cir.const #cir.int<4> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[FOUR_IDX]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[FOUR_IDX]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[NEXT_ELT]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}) +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[AND]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield // CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(|:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <ior> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, ["arrayinit.temp"] -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: cir.store {{.*}} %[[DECAY]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>> +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> +// CHECK-NEXT: } + for(int i = 0; i < 5; ++i); +#pragma acc parallel loop reduction(|:someVarArrNoFloats) +// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTSA5_24DefaultOperatorsNoFloats : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> reduction_operator <ior> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>{{.*}}) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperatorsNoFloats x 5>, !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>, ["openacc.reduction.init", init] +// CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>>, ["arrayinit.temp"] +// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: cir.store {{.*}} %[[DECAY]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>> // CHECK-NEXT: %[[LAST_IDX:.*]] = cir.const #cir.int<5> : !s64i -// CHECK-NEXT: %[[END_ITR:.*]] = cir.ptr_stride %[[DECAY]], %[[LAST_IDX]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[END_ITR:.*]] = cir.ptr_stride %[[DECAY]], %[[LAST_IDX]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> // CHECK-NEXT: cir.do { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[TEMP_LOAD]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>>, !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[TEMP_LOAD]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[TEMP_LOAD]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[TEMP_LOAD]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[TEMP_LOAD]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[TEMP_LOAD]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[TEMP_LOAD]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[TEMP_LOAD]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #false // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // // CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.int<1> : !s64i -// CHECK-NEXT: %[[NEXT_ITEM:.*]] = cir.ptr_stride %[[TEMP_LOAD]], %[[ONE]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: cir.store {{.*}} %[[NEXT_ITEM]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>> +// CHECK-NEXT: %[[NEXT_ITEM:.*]] = cir.ptr_stride %[[TEMP_LOAD]], %[[ONE]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: cir.store {{.*}} %[[NEXT_ITEM]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>> // CHECK-NEXT: cir.yield // CHECK-NEXT: } while { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(ne, %[[TEMP_LOAD]], %[[END_ITR]]) : !cir.ptr<!rec_DefaultOperators>, !cir.bool +// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>>, !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(ne, %[[TEMP_LOAD]], %[[END_ITR]]) : !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.bool // CHECK-NEXT: cir.condition(%[[CMP]]) // CHECK-NEXT: } // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}) +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[OR]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield // CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(^:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <xor> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, ["arrayinit.temp"] -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: cir.store {{.*}} %[[DECAY]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>> +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> +// CHECK-NEXT: } + for(int i = 0; i < 5; ++i); +#pragma acc parallel loop reduction(^:someVarArrNoFloats) +// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTSA5_24DefaultOperatorsNoFloats : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> reduction_operator <xor> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>{{.*}}) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperatorsNoFloats x 5>, !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>, ["openacc.reduction.init", init] +// CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>>, ["arrayinit.temp"] +// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: cir.store {{.*}} %[[DECAY]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>> // CHECK-NEXT: %[[LAST_IDX:.*]] = cir.const #cir.int<5> : !s64i -// CHECK-NEXT: %[[END_ITR:.*]] = cir.ptr_stride %[[DECAY]], %[[LAST_IDX]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[END_ITR:.*]] = cir.ptr_stride %[[DECAY]], %[[LAST_IDX]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> // CHECK-NEXT: cir.do { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[TEMP_LOAD]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>>, !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[TEMP_LOAD]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[TEMP_LOAD]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[TEMP_LOAD]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[TEMP_LOAD]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[TEMP_LOAD]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[TEMP_LOAD]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[TEMP_LOAD]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #false // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.int<1> : !s64i -// CHECK-NEXT: %[[NEXT_ITEM:.*]] = cir.ptr_stride %[[TEMP_LOAD]], %[[ONE]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: cir.store {{.*}} %[[NEXT_ITEM]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>> +// CHECK-NEXT: %[[NEXT_ITEM:.*]] = cir.ptr_stride %[[TEMP_LOAD]], %[[ONE]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: cir.store {{.*}} %[[NEXT_ITEM]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>> // CHECK-NEXT: cir.yield // CHECK-NEXT: } while { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(ne, %[[TEMP_LOAD]], %[[END_ITR]]) : !cir.ptr<!rec_DefaultOperators>, !cir.bool +// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>>, !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(ne, %[[TEMP_LOAD]], %[[END_ITR]]) : !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.bool // CHECK-NEXT: cir.condition(%[[CMP]]) // CHECK-NEXT: } // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}) +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[XOR]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield // CHECK-NEXT: } - for(int i=0;i < 5; ++i); +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> +// CHECK-NEXT: } + for(int i = 0; i < 5; ++i); #pragma acc parallel loop reduction(&&:someVarArr) // CHECK-NEXT: acc.reduction.recipe @reduction_land__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <land> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}) @@ -897,9 +1207,9 @@ void acc_combined() { // TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc parallel loop reduction(||:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_lor__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <lor> init { +// CHECK: acc.reduction.recipe @reduction_lor__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <lor> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, ["arrayinit.temp"] @@ -941,7 +1251,7 @@ void acc_combined() { // TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc parallel loop reduction(+:someVarArr[2]) // CHECK-NEXT: acc.reduction.recipe @reduction_add__Bcnt1__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <add> init { @@ -988,9 +1298,67 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.double +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !cir.double, !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) nsw : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[ADD]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc parallel loop reduction(*:someVarArr[2]) // CHECK-NEXT: acc.reduction.recipe @reduction_mul__Bcnt1__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <mul> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) @@ -1036,9 +1404,67 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.double +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !cir.double, !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) nsw : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[MUL]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc parallel loop reduction(max:someVarArr[2]) // CHECK-NEXT: acc.reduction.recipe @reduction_max__Bcnt1__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <max> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) @@ -1086,7 +1512,7 @@ void acc_combined() { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc parallel loop reduction(min:someVarArr[2]) // CHECK-NEXT: acc.reduction.recipe @reduction_min__Bcnt1__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <min> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) @@ -1134,11 +1560,11 @@ void acc_combined() { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(&:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_iand__Bcnt1__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <iand> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init"] + for(int i = 0; i < 5; ++i); +#pragma acc parallel loop reduction(&:someVarArrNoFloats[2]) +// CHECK-NEXT: acc.reduction.recipe @reduction_iand__Bcnt1__ZTSA5_24DefaultOperatorsNoFloats : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> reduction_operator <iand> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperatorsNoFloats x 5>, !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { // CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index // CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i @@ -1152,21 +1578,15 @@ void acc_combined() { // CHECK-NEXT: cir.condition(%[[COND]]) // CHECK-NEXT: } body { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: cir.yield @@ -1179,14 +1599,60 @@ void acc_combined() { // CHECK-NEXT: } // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[AND]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield // CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(|:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__Bcnt1__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <ior> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init"] +// CHECK-NEXT: } +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> +// CHECK-NEXT: } + for(int i = 0; i < 5; ++i); +#pragma acc parallel loop reduction(|:someVarArrNoFloats[2]) +// CHECK-NEXT: acc.reduction.recipe @reduction_ior__Bcnt1__ZTSA5_24DefaultOperatorsNoFloats : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> reduction_operator <ior> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperatorsNoFloats x 5>, !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { // CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index // CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i @@ -1200,21 +1666,15 @@ void acc_combined() { // CHECK-NEXT: cir.condition(%[[COND]]) // CHECK-NEXT: } body { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #false // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: cir.yield @@ -1227,14 +1687,60 @@ void acc_combined() { // CHECK-NEXT: } // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[OR]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield // CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(^:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__Bcnt1__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <xor> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init"] +// CHECK-NEXT: } +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> +// CHECK-NEXT: } + for(int i = 0; i < 5; ++i); +#pragma acc parallel loop reduction(^:someVarArrNoFloats[2]) +// CHECK-NEXT: acc.reduction.recipe @reduction_xor__Bcnt1__ZTSA5_24DefaultOperatorsNoFloats : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> reduction_operator <xor> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperatorsNoFloats x 5>, !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { // CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index // CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i @@ -1248,21 +1754,15 @@ void acc_combined() { // CHECK-NEXT: cir.condition(%[[COND]]) // CHECK-NEXT: } body { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #false // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: cir.yield @@ -1275,10 +1775,56 @@ void acc_combined() { // CHECK-NEXT: } // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[XOR]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc parallel loop reduction(&&:someVarArr[2]) // CHECK-NEXT: acc.reduction.recipe @reduction_land__Bcnt1__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <land> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) @@ -1326,7 +1872,7 @@ void acc_combined() { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc parallel loop reduction(||:someVarArr[2]) // CHECK-NEXT: acc.reduction.recipe @reduction_lor__Bcnt1__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <lor> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) @@ -1374,26 +1920,26 @@ void acc_combined() { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc parallel loop reduction(+:someVarArr[1:1]) - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc parallel loop reduction(*:someVarArr[1:1]) - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc parallel loop reduction(max:someVarArr[1:1]) - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc parallel loop reduction(min:someVarArr[1:1]) - for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(&:someVarArr[1:1]) - for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(|:someVarArr[1:1]) - for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(^:someVarArr[1:1]) - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); +#pragma acc parallel loop reduction(&:someVarArrNoFloats[1:1]) + for(int i = 0; i < 5; ++i); +#pragma acc parallel loop reduction(|:someVarArrNoFloats[1:1]) + for(int i = 0; i < 5; ++i); +#pragma acc parallel loop reduction(^:someVarArrNoFloats[1:1]) + for(int i = 0; i < 5; ++i); #pragma acc parallel loop reduction(&&:someVarArr[1:1]) - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc parallel loop reduction(||:someVarArr[1:1]) - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); // CHECK-NEXT: cir.func {{.*}}@_Z12acc_combined } diff --git a/clang/test/CIR/CodeGenOpenACC/combined-reduction-clause-float.cpp b/clang/test/CIR/CodeGenOpenACC/combined-reduction-clause-float.cpp index 001c2fc..472e4ac 100644 --- a/clang/test/CIR/CodeGenOpenACC/combined-reduction-clause-float.cpp +++ b/clang/test/CIR/CodeGenOpenACC/combined-reduction-clause-float.cpp @@ -1,4 +1,4 @@ -// RUN: %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s +// RUN: not %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s template<typename T> void acc_combined() { T someVar; @@ -13,7 +13,10 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!cir.float> +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.float> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -26,7 +29,10 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!cir.float> +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.float> // CHECK-NEXT: } @@ -57,47 +63,6 @@ void acc_combined() { // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.float> // CHECK-NEXT: } for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(&:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTSf : !cir.ptr<!cir.float> reduction_operator <iand> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.float>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.float, !cir.ptr<!cir.float>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[ALLOCA]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: acc.yield -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.float> -// CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(|:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTSf : !cir.ptr<!cir.float> reduction_operator <ior> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.float>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.float, !cir.ptr<!cir.float>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[ALLOCA]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: acc.yield -// -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.float> -// CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(^:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTSf : !cir.ptr<!cir.float> reduction_operator <xor> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.float>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.float, !cir.ptr<!cir.float>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[ALLOCA]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: acc.yield -// -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.float> -// CHECK-NEXT: } - for(int i=0;i < 5; ++i); #pragma acc parallel loop reduction(&&:someVar) // CHECK-NEXT: acc.reduction.recipe @reduction_land__ZTSf : !cir.ptr<!cir.float> reduction_operator <land> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.float>{{.*}}) @@ -152,7 +117,31 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -183,7 +172,31 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -249,97 +262,6 @@ void acc_combined() { // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(&:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <iand> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ALL_ONES]], %[[DECAY]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ONE_IDX:.*]] = cir.const #cir.int<1> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[ONE_IDX]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ALL_ONES]], %[[NEXT_ELT]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[TWO_IDX:.*]] = cir.const #cir.int<2> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[TWO_IDX]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ALL_ONES]], %[[NEXT_ELT]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[THREE_IDX:.*]] = cir.const #cir.int<3> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[THREE_IDX]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ALL_ONES]], %[[NEXT_ELT]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[FOUR_IDX:.*]] = cir.const #cir.int<4> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[FOUR_IDX]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ALL_ONES]], %[[NEXT_ELT]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: acc.yield -// -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -// CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(|:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <ior> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!cir.float>, !cir.ptr<!cir.ptr<!cir.float>>, ["arrayinit.temp"] -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> -// CHECK-NEXT: cir.store {{.*}} %[[DECAY]], %[[TEMP_ITR]] : !cir.ptr<!cir.float>, !cir.ptr<!cir.ptr<!cir.float>> -// CHECK-NEXT: %[[LAST_IDX:.*]] = cir.const #cir.int<5> : !s64i -// CHECK-NEXT: %[[END_ITR:.*]] = cir.ptr_stride %[[DECAY]], %[[LAST_IDX]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: cir.do { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!cir.float>>, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[TEMP_LOAD]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.int<1> : !s64i -// CHECK-NEXT: %[[NEXT_ITEM:.*]] = cir.ptr_stride %[[TEMP_LOAD]], %[[ONE]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: cir.store {{.*}} %[[NEXT_ITEM]], %[[TEMP_ITR]] : !cir.ptr<!cir.float>, !cir.ptr<!cir.ptr<!cir.float>> -// CHECK-NEXT: cir.yield -// CHECK-NEXT: } while { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!cir.float>>, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(ne, %[[TEMP_LOAD]], %[[END_ITR]]) : !cir.ptr<!cir.float>, !cir.bool -// CHECK-NEXT: cir.condition(%[[CMP]]) -// CHECK-NEXT: } -// CHECK-NEXT: acc.yield -// -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -// CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(^:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <xor> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!cir.float>, !cir.ptr<!cir.ptr<!cir.float>>, ["arrayinit.temp"] -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> -// CHECK-NEXT: cir.store {{.*}} %[[DECAY]], %[[TEMP_ITR]] : !cir.ptr<!cir.float>, !cir.ptr<!cir.ptr<!cir.float>> -// CHECK-NEXT: %[[LAST_IDX:.*]] = cir.const #cir.int<5> : !s64i -// CHECK-NEXT: %[[END_ITR:.*]] = cir.ptr_stride %[[DECAY]], %[[LAST_IDX]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: cir.do { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!cir.float>>, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[TEMP_LOAD]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.int<1> : !s64i -// CHECK-NEXT: %[[NEXT_ITEM:.*]] = cir.ptr_stride %[[TEMP_LOAD]], %[[ONE]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: cir.store {{.*}} %[[NEXT_ITEM]], %[[TEMP_ITR]] : !cir.ptr<!cir.float>, !cir.ptr<!cir.ptr<!cir.float>> -// CHECK-NEXT: cir.yield -// CHECK-NEXT: } while { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!cir.float>>, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(ne, %[[TEMP_LOAD]], %[[END_ITR]]) : !cir.ptr<!cir.float>, !cir.bool -// CHECK-NEXT: cir.condition(%[[CMP]]) -// CHECK-NEXT: } -// CHECK-NEXT: acc.yield -// -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -// CHECK-NEXT: } - for(int i=0;i < 5; ++i); #pragma acc parallel loop reduction(&&:someVarArr) // CHECK-NEXT: acc.reduction.recipe @reduction_land__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <land> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}) @@ -434,13 +356,6 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -// CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(*:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_mul__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <mul> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { // CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index // CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i @@ -454,10 +369,14 @@ void acc_combined() { // CHECK-NEXT: cir.condition(%[[COND]]) // CHECK-NEXT: } body { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.fp<1{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ONE]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHS_STRIDE]] // CHECK-NEXT: cir.yield // CHECK-NEXT: } step { // CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i @@ -466,49 +385,11 @@ void acc_combined() { // CHECK-NEXT: cir.yield // CHECK-NEXT: } // CHECK-NEXT: } -// CHECK-NEXT: acc.yield -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(max:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_max__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <max> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init"] -// CHECK-NEXT: cir.scope { -// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index -// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i -// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index -// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i -// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} -// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> -// CHECK-NEXT: cir.for : cond { -// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool -// CHECK-NEXT: cir.condition(%[[COND]]) -// CHECK-NEXT: } body { -// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[LEAST:.*]] = cir.const #cir.fp<-3.4{{.*}}E+38> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[LEAST]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: cir.yield -// CHECK-NEXT: } step { -// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i -// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> -// CHECK-NEXT: cir.yield -// CHECK-NEXT: } -// CHECK-NEXT: } -// CHECK-NEXT: acc.yield -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -// CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(min:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_min__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <min> init { +#pragma acc parallel loop reduction(*:someVarArr[2]) +// CHECK-NEXT: acc.reduction.recipe @reduction_mul__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <mul> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { @@ -526,8 +407,8 @@ void acc_combined() { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i // CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> // CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[LARGEST:.*]] = cir.const #cir.fp<3.4{{.*}}E+38> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[LARGEST]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.fp<1{{.*}}> : !cir.float +// CHECK-NEXT: cir.store{{.*}} %[[ONE]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> // CHECK-NEXT: cir.yield // CHECK-NEXT: } step { // CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i @@ -539,13 +420,6 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -// CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(&:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_iand__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <iand> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { // CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index // CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i @@ -559,10 +433,14 @@ void acc_combined() { // CHECK-NEXT: cir.condition(%[[COND]]) // CHECK-NEXT: } body { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ALL_ONES]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHS_STRIDE]] // CHECK-NEXT: cir.yield // CHECK-NEXT: } step { // CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i @@ -571,14 +449,11 @@ void acc_combined() { // CHECK-NEXT: cir.yield // CHECK-NEXT: } // CHECK-NEXT: } -// CHECK-NEXT: acc.yield -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(|:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <ior> init { +#pragma acc parallel loop reduction(max:someVarArr[2]) +// CHECK-NEXT: acc.reduction.recipe @reduction_max__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <max> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { @@ -596,8 +471,8 @@ void acc_combined() { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i // CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> // CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ZERO]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LEAST:.*]] = cir.const #cir.fp<-3.4{{.*}}E+38> : !cir.float +// CHECK-NEXT: cir.store{{.*}} %[[LEAST]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> // CHECK-NEXT: cir.yield // CHECK-NEXT: } step { // CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i @@ -609,11 +484,12 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(^:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <xor> init { +#pragma acc parallel loop reduction(min:someVarArr[2]) +// CHECK-NEXT: acc.reduction.recipe @reduction_min__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <min> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { @@ -631,8 +507,8 @@ void acc_combined() { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i // CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> // CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ZERO]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LARGEST:.*]] = cir.const #cir.fp<3.4{{.*}}E+38> : !cir.float +// CHECK-NEXT: cir.store{{.*}} %[[LARGEST]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> // CHECK-NEXT: cir.yield // CHECK-NEXT: } step { // CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i @@ -644,6 +520,7 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -679,6 +556,7 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -714,6 +592,7 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -726,12 +605,6 @@ void acc_combined() { for(int i=0;i < 5; ++i); #pragma acc parallel loop reduction(min:someVarArr[1:1]) for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(&:someVarArr[1:1]) - for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(|:someVarArr[1:1]) - for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(^:someVarArr[1:1]) - for(int i=0;i < 5; ++i); #pragma acc parallel loop reduction(&&:someVarArr[1:1]) for(int i=0;i < 5; ++i); #pragma acc parallel loop reduction(||:someVarArr[1:1]) diff --git a/clang/test/CIR/CodeGenOpenACC/combined-reduction-clause-inline-ops.cpp b/clang/test/CIR/CodeGenOpenACC/combined-reduction-clause-inline-ops.cpp index 3d46ac7..112ff656 100644 --- a/clang/test/CIR/CodeGenOpenACC/combined-reduction-clause-inline-ops.cpp +++ b/clang/test/CIR/CodeGenOpenACC/combined-reduction-clause-inline-ops.cpp @@ -1,4 +1,4 @@ -// RUN: %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s +// RUN: not %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s struct HasOperatorsInline { int i; @@ -48,7 +48,7 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlinepLERS_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}): @@ -79,7 +79,7 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlinemLERS_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}): @@ -88,7 +88,7 @@ void acc_combined() { // CHECK-NEXT: } for(int i=0;i < 5; ++i); #pragma acc parallel loop reduction(max:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_max__ZTS18HasOperatorsInline : !cir.ptr<!rec_HasOperatorsInline> reduction_operator <max> init { +// CHECK: acc.reduction.recipe @reduction_max__ZTS18HasOperatorsInline : !cir.ptr<!rec_HasOperatorsInline> reduction_operator <max> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsInline, !cir.ptr<!rec_HasOperatorsInline>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_HasOperatorsInline> -> !cir.ptr<!s32i> @@ -172,7 +172,7 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineaNERS_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}): @@ -181,7 +181,7 @@ void acc_combined() { // CHECK-NEXT: } for(int i=0;i < 5; ++i); #pragma acc parallel loop reduction(|:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTS18HasOperatorsInline : !cir.ptr<!rec_HasOperatorsInline> reduction_operator <ior> init { +// CHECK: acc.reduction.recipe @reduction_ior__ZTS18HasOperatorsInline : !cir.ptr<!rec_HasOperatorsInline> reduction_operator <ior> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsInline, !cir.ptr<!rec_HasOperatorsInline>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_HasOperatorsInline> -> !cir.ptr<!s32i> @@ -203,7 +203,7 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineoRERS_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}): @@ -212,7 +212,7 @@ void acc_combined() { // CHECK-NEXT: } for(int i=0;i < 5; ++i); #pragma acc parallel loop reduction(^:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTS18HasOperatorsInline : !cir.ptr<!rec_HasOperatorsInline> reduction_operator <xor> init { +// CHECK: acc.reduction.recipe @reduction_xor__ZTS18HasOperatorsInline : !cir.ptr<!rec_HasOperatorsInline> reduction_operator <xor> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsInline, !cir.ptr<!rec_HasOperatorsInline>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_HasOperatorsInline> -> !cir.ptr<!s32i> @@ -234,7 +234,7 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineeOERS_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}): @@ -243,7 +243,7 @@ void acc_combined() { // CHECK-NEXT: } for(int i=0;i < 5; ++i); #pragma acc parallel loop reduction(&&:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_land__ZTS18HasOperatorsInline : !cir.ptr<!rec_HasOperatorsInline> reduction_operator <land> init { +// CHECK: acc.reduction.recipe @reduction_land__ZTS18HasOperatorsInline : !cir.ptr<!rec_HasOperatorsInline> reduction_operator <land> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsInline, !cir.ptr<!rec_HasOperatorsInline>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_HasOperatorsInline> -> !cir.ptr<!s32i> @@ -344,7 +344,29 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlinepLERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}): @@ -466,7 +488,28 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlinemLERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}): @@ -832,7 +875,28 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineaNERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}): @@ -896,7 +960,28 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineoRERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}): @@ -959,7 +1044,28 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineeOERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}): @@ -1216,6 +1322,32 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlinepLERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): @@ -1293,6 +1425,32 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlinemLERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): @@ -1524,6 +1682,32 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineaNERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): @@ -1601,6 +1785,32 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineoRERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): @@ -1678,6 +1888,32 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineeOERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): diff --git a/clang/test/CIR/CodeGenOpenACC/combined-reduction-clause-int.cpp b/clang/test/CIR/CodeGenOpenACC/combined-reduction-clause-int.cpp index e9669d3..7eaa822b 100644 --- a/clang/test/CIR/CodeGenOpenACC/combined-reduction-clause-int.cpp +++ b/clang/test/CIR/CodeGenOpenACC/combined-reduction-clause-int.cpp @@ -1,4 +1,4 @@ -// RUN: %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s +// RUN: not %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s template<typename T> void acc_combined() { @@ -14,7 +14,10 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!s32i> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!s32i> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!s32i> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -28,7 +31,10 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!s32i> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!s32i> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!s32i> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -67,7 +73,10 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!s32i> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!s32i> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!s32i> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -81,7 +90,10 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!s32i> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!s32i> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!s32i> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -95,7 +107,10 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!s32i> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!s32i> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!s32i> // CHECK-NEXT: } @@ -155,7 +170,31 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -186,7 +225,31 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -279,7 +342,31 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -309,7 +396,31 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -339,7 +450,31 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -437,6 +572,35 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -472,6 +636,35 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -507,6 +700,7 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -542,6 +736,7 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -577,6 +772,35 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -612,6 +836,35 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -647,6 +900,35 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -682,6 +964,7 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -717,6 +1000,7 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); diff --git a/clang/test/CIR/CodeGenOpenACC/combined-reduction-clause-outline-ops.cpp b/clang/test/CIR/CodeGenOpenACC/combined-reduction-clause-outline-ops.cpp index c99e3c1..c2c0c77 100644 --- a/clang/test/CIR/CodeGenOpenACC/combined-reduction-clause-outline-ops.cpp +++ b/clang/test/CIR/CodeGenOpenACC/combined-reduction-clause-outline-ops.cpp @@ -1,4 +1,4 @@ -// RUN: %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s +// RUN: not %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s struct HasOperatorsOutline { int i; unsigned u; @@ -47,7 +47,7 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZpLR19HasOperatorsOutlineS0_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}): @@ -55,7 +55,6 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } for(int i=0;i < 5; ++i); - #pragma acc parallel loop reduction(*:someVar) // CHECK: acc.reduction.recipe @reduction_mul__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <mul> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline>{{.*}}) @@ -78,17 +77,17 @@ void acc_combined() { // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { - for(int i=0;i < 5; ++i); -#pragma acc parallel loop reduction(max:someVar) // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZmLR19HasOperatorsOutlineS0_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}): // CHECK-NEXT: cir.call @_ZN19HasOperatorsOutlineD1Ev(%[[ARG]]) nothrow : (!cir.ptr<!rec_HasOperatorsOutline>) // CHECK-NEXT: acc.yield // CHECK-NEXT: } -// CHECK-NEXT: acc.reduction.recipe @reduction_max__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <max> init { + for(int i=0;i < 5; ++i); +#pragma acc parallel loop reduction(max:someVar) +// CHECK: acc.reduction.recipe @reduction_max__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <max> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsOutline, !cir.ptr<!rec_HasOperatorsOutline>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_HasOperatorsOutline> -> !cir.ptr<!s32i> @@ -172,7 +171,7 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZaNR19HasOperatorsOutlineS0_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}): @@ -181,7 +180,7 @@ void acc_combined() { // CHECK-NEXT: } for(int i=0;i < 5; ++i); #pragma acc parallel loop reduction(|:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <ior> init { +// CHECK: acc.reduction.recipe @reduction_ior__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <ior> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsOutline, !cir.ptr<!rec_HasOperatorsOutline>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_HasOperatorsOutline> -> !cir.ptr<!s32i> @@ -203,7 +202,7 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZoRR19HasOperatorsOutlineS0_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}): @@ -212,7 +211,7 @@ void acc_combined() { // CHECK-NEXT: } for(int i=0;i < 5; ++i); #pragma acc parallel loop reduction(^:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <xor> init { +// CHECK: acc.reduction.recipe @reduction_xor__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <xor> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsOutline, !cir.ptr<!rec_HasOperatorsOutline>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_HasOperatorsOutline> -> !cir.ptr<!s32i> @@ -234,7 +233,7 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZeOR19HasOperatorsOutlineS0_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}): @@ -243,7 +242,7 @@ void acc_combined() { // CHECK-NEXT: } for(int i=0;i < 5; ++i); #pragma acc parallel loop reduction(&&:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_land__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <land> init { +// CHECK: acc.reduction.recipe @reduction_land__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <land> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsOutline, !cir.ptr<!rec_HasOperatorsOutline>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_HasOperatorsOutline> -> !cir.ptr<!s32i> @@ -344,7 +343,29 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZpLR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}): @@ -466,7 +487,28 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZmLR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}): @@ -832,7 +874,28 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZaNR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}): @@ -896,7 +959,28 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZoRR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}): @@ -959,7 +1043,28 @@ void acc_combined() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZeOR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}): @@ -982,6 +1087,7 @@ void acc_combined() { // CHECK-NEXT: } // CHECK-NEXT: acc.yield // CHECK-NEXT: } + for(int i=0;i < 5; ++i); #pragma acc parallel loop reduction(&&:someVarArr) // CHECK-NEXT: acc.reduction.recipe @reduction_land__ZTSA5_19HasOperatorsOutline : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> reduction_operator <land> init { @@ -1106,8 +1212,7 @@ void acc_combined() { // CHECK-NEXT: } for(int i=0;i < 5; ++i); #pragma acc parallel loop reduction(||:someVarArr) - -// CHECK-NEXT: acc.reduction.recipe @reduction_lor__ZTSA5_19HasOperatorsOutline : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> reduction_operator <lor> init { +// CHECK: acc.reduction.recipe @reduction_lor__ZTSA5_19HasOperatorsOutline : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> reduction_operator <lor> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_HasOperatorsOutline x 5>, !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!cir.ptr<!rec_HasOperatorsOutline>>, ["arrayinit.temp"] @@ -1216,6 +1321,32 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZpLR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): @@ -1293,6 +1424,32 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZmLR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): @@ -1524,6 +1681,32 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZaNR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): @@ -1601,6 +1784,32 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZoRR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): @@ -1678,6 +1887,32 @@ void acc_combined() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZeOR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): diff --git a/clang/test/CIR/CodeGenOpenACC/compute-copy.c b/clang/test/CIR/CodeGenOpenACC/compute-copy.c index d7676d6..41e594e 100644 --- a/clang/test/CIR/CodeGenOpenACC/compute-copy.c +++ b/clang/test/CIR/CodeGenOpenACC/compute-copy.c @@ -2,7 +2,7 @@ int global; void acc_compute(int parmVar) { - // CHECK: cir.func{{.*}} @acc_compute(%[[ARG:.*]]: !s32i{{.*}}) { + // CHECK: cir.func{{.*}} @acc_compute(%[[ARG:.*]]: !s32i{{.*}}) // CHECK-NEXT: %[[PARM:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["parmVar", init] int localVar1; short localVar2; @@ -899,7 +899,7 @@ void acc_compute_members() { } void modifier_list() { - // CHECK: cir.func{{.*}} @modifier_list() { + // CHECK: cir.func{{.*}} @modifier_list() int localVar; // CHECK-NEXT: %[[LOCALVAR:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["localVar"] diff --git a/clang/test/CIR/CodeGenOpenACC/compute-firstprivate-clause.c b/clang/test/CIR/CodeGenOpenACC/compute-firstprivate-clause.c index 184f904..52342e7 100644 --- a/clang/test/CIR/CodeGenOpenACC/compute-firstprivate-clause.c +++ b/clang/test/CIR/CodeGenOpenACC/compute-firstprivate-clause.c @@ -142,7 +142,7 @@ struct NoCopyConstruct {}; // CHECK-NEXT: } void acc_compute() { - // CHECK: cir.func{{.*}} @acc_compute() { + // CHECK: cir.func{{.*}} @acc_compute() int someInt; // CHECK-NEXT: %[[SOMEINT:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["someInt"] diff --git a/clang/test/CIR/CodeGenOpenACC/compute-firstprivate-clause.cpp b/clang/test/CIR/CodeGenOpenACC/compute-firstprivate-clause.cpp index e3f091a..a2c6c38 100644 --- a/clang/test/CIR/CodeGenOpenACC/compute-firstprivate-clause.cpp +++ b/clang/test/CIR/CodeGenOpenACC/compute-firstprivate-clause.cpp @@ -324,7 +324,7 @@ struct HasDtor { // extern "C" void acc_compute() { - // CHECK: cir.func{{.*}} @acc_compute() { + // CHECK: cir.func{{.*}} @acc_compute() int someInt; // CHECK-NEXT: %[[SOMEINT:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["someInt"] diff --git a/clang/test/CIR/CodeGenOpenACC/compute-private-clause.c b/clang/test/CIR/CodeGenOpenACC/compute-private-clause.c index 34b8b69..943539e 100644 --- a/clang/test/CIR/CodeGenOpenACC/compute-private-clause.c +++ b/clang/test/CIR/CodeGenOpenACC/compute-private-clause.c @@ -45,7 +45,7 @@ struct NoCopyConstruct {}; // CHECK-NEXT: } void acc_compute() { - // CHECK: cir.func{{.*}} @acc_compute() { + // CHECK: cir.func{{.*}} @acc_compute() int someInt; // CHECK-NEXT: %[[SOMEINT:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["someInt"] diff --git a/clang/test/CIR/CodeGenOpenACC/compute-private-clause.cpp b/clang/test/CIR/CodeGenOpenACC/compute-private-clause.cpp index 8262a31..f0bd98c 100644 --- a/clang/test/CIR/CodeGenOpenACC/compute-private-clause.cpp +++ b/clang/test/CIR/CodeGenOpenACC/compute-private-clause.cpp @@ -146,7 +146,7 @@ struct HasDtor { // CHECK-NEXT: } extern "C" void acc_compute() { - // CHECK: cir.func{{.*}} @acc_compute() { + // CHECK: cir.func{{.*}} @acc_compute() int someInt; // CHECK-NEXT: %[[SOMEINT:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["someInt"] diff --git a/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-default-ops.c b/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-default-ops.c index f70113e..b439623 100644 --- a/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-default-ops.c +++ b/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-default-ops.c @@ -1,4 +1,4 @@ -// RUN: %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -std=c23 -triple x86_64-linux-pc %s -o - | FileCheck %s +// RUN: not %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -std=c23 -triple x86_64-linux-pc %s -o - | FileCheck %s struct DefaultOperators { int i; @@ -8,9 +8,17 @@ struct DefaultOperators { bool b; }; +struct DefaultOperatorsNoFloats { + int i; + unsigned int u; + bool b; +}; + void acc_compute() { struct DefaultOperators someVar; struct DefaultOperators someVarArr[5]; + struct DefaultOperatorsNoFloats someVarNoFloats; + struct DefaultOperatorsNoFloats someVarArrNoFloats[5]; #pragma acc parallel reduction(+:someVar) // CHECK: acc.reduction.recipe @reduction_add__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <add> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) @@ -34,7 +42,39 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.double +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !cir.double, !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) nsw : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[ADD]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> // CHECK-NEXT: } ; @@ -61,7 +101,39 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.double +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !cir.double, !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) nsw : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[MUL]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> // CHECK-NEXT: } ; @@ -119,85 +191,127 @@ void acc_compute() { // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> // CHECK-NEXT: } ; -#pragma acc parallel reduction(&:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <iand> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperators, !cir.ptr<!rec_DefaultOperators>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +#pragma acc parallel reduction(&:someVarNoFloats) +// CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTS24DefaultOperatorsNoFloats : !cir.ptr<!rec_DefaultOperatorsNoFloats> reduction_operator <iand> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats>{{.*}}) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperatorsNoFloats, !cir.ptr<!rec_DefaultOperatorsNoFloats>, ["openacc.reduction.init", init] +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[ALLOCA]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[ALLOCA]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[ALLOCA]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[ALLOCA]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[ALLOCA]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[ALLOCA]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats> {{.*}}) +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[AND]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperatorsNoFloats> // CHECK-NEXT: } ; -#pragma acc parallel reduction(|:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <ior> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperators, !cir.ptr<!rec_DefaultOperators>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +#pragma acc parallel reduction(|:someVarNoFloats) +// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTS24DefaultOperatorsNoFloats : !cir.ptr<!rec_DefaultOperatorsNoFloats> reduction_operator <ior> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats>{{.*}}) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperatorsNoFloats, !cir.ptr<!rec_DefaultOperatorsNoFloats>, ["openacc.reduction.init", init] +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[ALLOCA]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[ALLOCA]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[ALLOCA]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[ALLOCA]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[ALLOCA]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[ALLOCA]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #false // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats> {{.*}}) +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[OR]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperatorsNoFloats> // CHECK-NEXT: } ; -#pragma acc parallel reduction(^:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <xor> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperators, !cir.ptr<!rec_DefaultOperators>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +#pragma acc parallel reduction(^:someVarNoFloats) +// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTS24DefaultOperatorsNoFloats : !cir.ptr<!rec_DefaultOperatorsNoFloats> reduction_operator <xor> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats>{{.*}}) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperatorsNoFloats, !cir.ptr<!rec_DefaultOperatorsNoFloats>, ["openacc.reduction.init", init] +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[ALLOCA]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[ALLOCA]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[ALLOCA]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[ALLOCA]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[ALLOCA]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[ALLOCA]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #false // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats> {{.*}}) +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[XOR]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperatorsNoFloats> // CHECK-NEXT: } ; #pragma acc parallel reduction(&&:someVar) @@ -281,7 +395,62 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> +// +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.double +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !cir.double, !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) nsw : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[ADD]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } ; @@ -383,7 +552,62 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> +// +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.double +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !cir.double, !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) nsw : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[MUL]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } ; @@ -592,166 +816,265 @@ void acc_compute() { // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } ; -#pragma acc parallel reduction(&:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <iand> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[DECAY]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +#pragma acc parallel reduction(&:someVarArrNoFloats) +// CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTSA5_24DefaultOperatorsNoFloats : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> reduction_operator <iand> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>{{.*}}) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperatorsNoFloats x 5>, !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>, ["openacc.reduction.init", init] +// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[DECAY]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[DECAY]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[DECAY]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[DECAY]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[DECAY]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[DECAY]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[DECAY]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // // CHECK-NEXT: %[[ALL_ONES_IDX:.*]] = cir.const #cir.int<1> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[ALL_ONES_IDX]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[ALL_ONES_IDX]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[NEXT_ELT]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // // CHECK-NEXT: %[[TWO_IDX:.*]] = cir.const #cir.int<2> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[TWO_IDX]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[TWO_IDX]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[NEXT_ELT]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // // // CHECK-NEXT: %[[THREE_IDX:.*]] = cir.const #cir.int<3> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[THREE_IDX]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[THREE_IDX]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[NEXT_ELT]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // // CHECK-NEXT: %[[FOUR_IDX:.*]] = cir.const #cir.int<4> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[FOUR_IDX]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[FOUR_IDX]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[NEXT_ELT]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}) +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[AND]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> // CHECK-NEXT: } ; -#pragma acc parallel reduction(|:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <ior> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, ["arrayinit.temp"] -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: cir.store {{.*}} %[[DECAY]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>> +#pragma acc parallel reduction(|:someVarArrNoFloats) +// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTSA5_24DefaultOperatorsNoFloats : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> reduction_operator <ior> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>{{.*}}) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperatorsNoFloats x 5>, !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>, ["openacc.reduction.init", init] +// CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>>, ["arrayinit.temp"] +// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: cir.store {{.*}} %[[DECAY]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>> // CHECK-NEXT: %[[LAST_IDX:.*]] = cir.const #cir.int<5> : !s64i -// CHECK-NEXT: %[[END_ITR:.*]] = cir.ptr_stride %[[DECAY]], %[[LAST_IDX]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[END_ITR:.*]] = cir.ptr_stride %[[DECAY]], %[[LAST_IDX]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> // CHECK-NEXT: cir.do { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.zero : !rec_DefaultOperators +// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>>, !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.zero : !rec_DefaultOperatorsNoFloats // CHECK-NEXT: cir.store{{.*}} %[[ZERO]], %[[TEMP_LOAD]] // CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.int<1> : !s64i -// CHECK-NEXT: %[[NEXT_ITEM:.*]] = cir.ptr_stride %[[TEMP_LOAD]], %[[ONE]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: cir.store {{.*}} %[[NEXT_ITEM]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>> +// CHECK-NEXT: %[[NEXT_ITEM:.*]] = cir.ptr_stride %[[TEMP_LOAD]], %[[ONE]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: cir.store {{.*}} %[[NEXT_ITEM]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>> // CHECK-NEXT: cir.yield // CHECK-NEXT: } while { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(ne, %[[TEMP_LOAD]], %[[END_ITR]]) : !cir.ptr<!rec_DefaultOperators>, !cir.bool +// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>>, !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(ne, %[[TEMP_LOAD]], %[[END_ITR]]) : !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.bool // CHECK-NEXT: cir.condition(%[[CMP]]) // CHECK-NEXT: } // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}) +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[OR]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> // CHECK-NEXT: } ; -#pragma acc parallel reduction(^:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <xor> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, ["arrayinit.temp"] -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: cir.store {{.*}} %[[DECAY]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>> +#pragma acc parallel reduction(^:someVarArrNoFloats) +// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTSA5_24DefaultOperatorsNoFloats : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> reduction_operator <xor> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>{{.*}}) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperatorsNoFloats x 5>, !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>, ["openacc.reduction.init", init] +// CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>>, ["arrayinit.temp"] +// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: cir.store {{.*}} %[[DECAY]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>> // CHECK-NEXT: %[[LAST_IDX:.*]] = cir.const #cir.int<5> : !s64i -// CHECK-NEXT: %[[END_ITR:.*]] = cir.ptr_stride %[[DECAY]], %[[LAST_IDX]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[END_ITR:.*]] = cir.ptr_stride %[[DECAY]], %[[LAST_IDX]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> // CHECK-NEXT: cir.do { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.zero : !rec_DefaultOperators +// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>>, !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.zero : !rec_DefaultOperatorsNoFloats // CHECK-NEXT: cir.store{{.*}} %[[ZERO]], %[[TEMP_LOAD]] // CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.int<1> : !s64i -// CHECK-NEXT: %[[NEXT_ITEM:.*]] = cir.ptr_stride %[[TEMP_LOAD]], %[[ONE]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: cir.store {{.*}} %[[NEXT_ITEM]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>> +// CHECK-NEXT: %[[NEXT_ITEM:.*]] = cir.ptr_stride %[[TEMP_LOAD]], %[[ONE]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: cir.store {{.*}} %[[NEXT_ITEM]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>> // CHECK-NEXT: cir.yield // CHECK-NEXT: } while { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(ne, %[[TEMP_LOAD]], %[[END_ITR]]) : !cir.ptr<!rec_DefaultOperators>, !cir.bool +// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>>, !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(ne, %[[TEMP_LOAD]], %[[END_ITR]]) : !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.bool // CHECK-NEXT: cir.condition(%[[CMP]]) // CHECK-NEXT: } // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}) +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[XOR]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> // CHECK-NEXT: } ; #pragma acc parallel reduction(&&:someVarArr) @@ -932,6 +1255,64 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.double +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !cir.double, !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) nsw : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[ADD]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } ; @@ -980,6 +1361,64 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.double +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !cir.double, !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) nsw : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[MUL]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } ; @@ -1079,10 +1518,10 @@ void acc_compute() { // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } ; -#pragma acc parallel reduction(&:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_iand__Bcnt1__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <iand> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init"] +#pragma acc parallel reduction(&:someVarArrNoFloats[2]) +// CHECK-NEXT: acc.reduction.recipe @reduction_iand__Bcnt1__ZTSA5_24DefaultOperatorsNoFloats : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> reduction_operator <iand> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperatorsNoFloats x 5>, !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { // CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index // CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i @@ -1096,21 +1535,15 @@ void acc_compute() { // CHECK-NEXT: cir.condition(%[[COND]]) // CHECK-NEXT: } body { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: cir.yield @@ -1123,14 +1556,60 @@ void acc_compute() { // CHECK-NEXT: } // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[AND]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> // CHECK-NEXT: } ; -#pragma acc parallel reduction(|:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__Bcnt1__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <ior> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init"] +#pragma acc parallel reduction(|:someVarArrNoFloats[2]) +// CHECK-NEXT: acc.reduction.recipe @reduction_ior__Bcnt1__ZTSA5_24DefaultOperatorsNoFloats : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> reduction_operator <ior> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperatorsNoFloats x 5>, !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { // CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index // CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i @@ -1144,21 +1623,15 @@ void acc_compute() { // CHECK-NEXT: cir.condition(%[[COND]]) // CHECK-NEXT: } body { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #false // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: cir.yield @@ -1171,14 +1644,60 @@ void acc_compute() { // CHECK-NEXT: } // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[OR]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> // CHECK-NEXT: } ; -#pragma acc parallel reduction(^:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__Bcnt1__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <xor> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init"] +#pragma acc parallel reduction(^:someVarArrNoFloats[2]) +// CHECK-NEXT: acc.reduction.recipe @reduction_xor__Bcnt1__ZTSA5_24DefaultOperatorsNoFloats : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> reduction_operator <xor> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperatorsNoFloats x 5>, !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { // CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index // CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i @@ -1192,21 +1711,15 @@ void acc_compute() { // CHECK-NEXT: cir.condition(%[[COND]]) // CHECK-NEXT: } body { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #false // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: cir.yield @@ -1219,8 +1732,54 @@ void acc_compute() { // CHECK-NEXT: } // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[XOR]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> // CHECK-NEXT: } ; #pragma acc parallel reduction(&&:someVarArr[2]) @@ -1328,11 +1887,11 @@ void acc_compute() { ; #pragma acc parallel reduction(min:someVarArr[1:1]) ; -#pragma acc parallel reduction(&:someVarArr[1:1]) +#pragma acc parallel reduction(&:someVarArrNoFloats[1:1]) ; -#pragma acc parallel reduction(|:someVarArr[1:1]) +#pragma acc parallel reduction(|:someVarArrNoFloats[1:1]) ; -#pragma acc parallel reduction(^:someVarArr[1:1]) +#pragma acc parallel reduction(^:someVarArrNoFloats[1:1]) ; #pragma acc parallel reduction(&&:someVarArr[1:1]) ; diff --git a/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-default-ops.cpp b/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-default-ops.cpp index 6a987c2..f997902 100644 --- a/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-default-ops.cpp +++ b/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-default-ops.cpp @@ -1,4 +1,4 @@ -// RUN: %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s +// RUN: not %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s struct DefaultOperators { int i; @@ -8,12 +8,19 @@ struct DefaultOperators { bool b; }; +struct DefaultOperatorsNoFloats { + int i; + unsigned int u; + bool b; +}; + template<typename T> void acc_compute() { T someVar; T someVarArr[5]; + struct DefaultOperatorsNoFloats someVarNoFloats; + struct DefaultOperatorsNoFloats someVarArrNoFloats[5]; #pragma acc parallel reduction(+:someVar) - ; // CHECK: acc.reduction.recipe @reduction_add__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <add> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperators, !cir.ptr<!rec_DefaultOperators>, ["openacc.reduction.init", init] @@ -36,11 +43,43 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.double +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !cir.double, !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) nsw : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[ADD]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> // CHECK-NEXT: } + ; #pragma acc parallel reduction(*:someVar) - // CHECK-NEXT: acc.reduction.recipe @reduction_mul__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <mul> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperators, !cir.ptr<!rec_DefaultOperators>, ["openacc.reduction.init", init] @@ -63,7 +102,39 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.double +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !cir.double, !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) nsw : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[MUL]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> // CHECK-NEXT: } ; @@ -121,86 +192,127 @@ void acc_compute() { // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> // CHECK-NEXT: } ; -#pragma acc parallel reduction(&:someVar) - -// CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <iand> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperators, !cir.ptr<!rec_DefaultOperators>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +#pragma acc parallel reduction(&:someVarNoFloats) +// CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTS24DefaultOperatorsNoFloats : !cir.ptr<!rec_DefaultOperatorsNoFloats> reduction_operator <iand> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats>{{.*}}) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperatorsNoFloats, !cir.ptr<!rec_DefaultOperatorsNoFloats>, ["openacc.reduction.init", init] +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[ALLOCA]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[ALLOCA]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[ALLOCA]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[ALLOCA]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[ALLOCA]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[ALLOCA]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats> {{.*}}) +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[AND]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperatorsNoFloats> // CHECK-NEXT: } ; -#pragma acc parallel reduction(|:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <ior> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperators, !cir.ptr<!rec_DefaultOperators>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +#pragma acc parallel reduction(|:someVarNoFloats) +// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTS24DefaultOperatorsNoFloats : !cir.ptr<!rec_DefaultOperatorsNoFloats> reduction_operator <ior> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats>{{.*}}) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperatorsNoFloats, !cir.ptr<!rec_DefaultOperatorsNoFloats>, ["openacc.reduction.init", init] +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[ALLOCA]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[ALLOCA]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[ALLOCA]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[ALLOCA]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[ALLOCA]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[ALLOCA]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #false // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats> {{.*}}) +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[OR]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperatorsNoFloats> // CHECK-NEXT: } ; -#pragma acc parallel reduction(^:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <xor> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperators, !cir.ptr<!rec_DefaultOperators>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +#pragma acc parallel reduction(^:someVarNoFloats) +// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTS24DefaultOperatorsNoFloats : !cir.ptr<!rec_DefaultOperatorsNoFloats> reduction_operator <xor> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats>{{.*}}) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperatorsNoFloats, !cir.ptr<!rec_DefaultOperatorsNoFloats>, ["openacc.reduction.init", init] +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[ALLOCA]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[ALLOCA]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[ALLOCA]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[ALLOCA]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[ALLOCA]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[ALLOCA]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #false // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats> {{.*}}) +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[XOR]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperatorsNoFloats> // CHECK-NEXT: } ; #pragma acc parallel reduction(&&:someVar) @@ -297,7 +409,62 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> +// +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.double +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !cir.double, !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) nsw : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[ADD]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } ; @@ -399,11 +566,67 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> +// +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.double +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !cir.double, !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) nsw : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[MUL]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } ; #pragma acc parallel reduction(max:someVarArr) + // CHECK-NEXT: acc.reduction.recipe @reduction_max__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <max> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init", init] @@ -607,193 +830,280 @@ void acc_compute() { // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } ; -#pragma acc parallel reduction(&:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <iand> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[DECAY]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +#pragma acc parallel reduction(&:someVarArrNoFloats) +// CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTSA5_24DefaultOperatorsNoFloats : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> reduction_operator <iand> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>{{.*}}) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperatorsNoFloats x 5>, !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>, ["openacc.reduction.init", init] +// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[DECAY]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[DECAY]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[DECAY]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[DECAY]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[DECAY]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[DECAY]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[DECAY]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // // CHECK-NEXT: %[[ALL_ONES_IDX:.*]] = cir.const #cir.int<1> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[ALL_ONES_IDX]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[ALL_ONES_IDX]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[NEXT_ELT]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // // CHECK-NEXT: %[[TWO_IDX:.*]] = cir.const #cir.int<2> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[TWO_IDX]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[TWO_IDX]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[NEXT_ELT]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // // // CHECK-NEXT: %[[THREE_IDX:.*]] = cir.const #cir.int<3> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[THREE_IDX]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[THREE_IDX]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[NEXT_ELT]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // // CHECK-NEXT: %[[FOUR_IDX:.*]] = cir.const #cir.int<4> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[FOUR_IDX]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[FOUR_IDX]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[NEXT_ELT]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}) +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[AND]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> // CHECK-NEXT: } ; -#pragma acc parallel reduction(|:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <ior> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, ["arrayinit.temp"] -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: cir.store {{.*}} %[[DECAY]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>> +#pragma acc parallel reduction(|:someVarArrNoFloats) +// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTSA5_24DefaultOperatorsNoFloats : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> reduction_operator <ior> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>{{.*}}) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperatorsNoFloats x 5>, !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>, ["openacc.reduction.init", init] +// CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>>, ["arrayinit.temp"] +// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: cir.store {{.*}} %[[DECAY]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>> // CHECK-NEXT: %[[LAST_IDX:.*]] = cir.const #cir.int<5> : !s64i -// CHECK-NEXT: %[[END_ITR:.*]] = cir.ptr_stride %[[DECAY]], %[[LAST_IDX]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[END_ITR:.*]] = cir.ptr_stride %[[DECAY]], %[[LAST_IDX]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> // CHECK-NEXT: cir.do { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[TEMP_LOAD]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>>, !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[TEMP_LOAD]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[TEMP_LOAD]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[TEMP_LOAD]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[TEMP_LOAD]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[TEMP_LOAD]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[TEMP_LOAD]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[TEMP_LOAD]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #false // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // // CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.int<1> : !s64i -// CHECK-NEXT: %[[NEXT_ITEM:.*]] = cir.ptr_stride %[[TEMP_LOAD]], %[[ONE]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: cir.store {{.*}} %[[NEXT_ITEM]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>> +// CHECK-NEXT: %[[NEXT_ITEM:.*]] = cir.ptr_stride %[[TEMP_LOAD]], %[[ONE]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: cir.store {{.*}} %[[NEXT_ITEM]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>> // CHECK-NEXT: cir.yield // CHECK-NEXT: } while { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(ne, %[[TEMP_LOAD]], %[[END_ITR]]) : !cir.ptr<!rec_DefaultOperators>, !cir.bool +// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>>, !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(ne, %[[TEMP_LOAD]], %[[END_ITR]]) : !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.bool // CHECK-NEXT: cir.condition(%[[CMP]]) // CHECK-NEXT: } // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}) +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[OR]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> // CHECK-NEXT: } ; -#pragma acc parallel reduction(^:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <xor> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, ["arrayinit.temp"] -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: cir.store {{.*}} %[[DECAY]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>> +#pragma acc parallel reduction(^:someVarArrNoFloats) +// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTSA5_24DefaultOperatorsNoFloats : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> reduction_operator <xor> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>{{.*}}) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperatorsNoFloats x 5>, !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>, ["openacc.reduction.init", init] +// CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>>, ["arrayinit.temp"] +// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: cir.store {{.*}} %[[DECAY]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>> // CHECK-NEXT: %[[LAST_IDX:.*]] = cir.const #cir.int<5> : !s64i -// CHECK-NEXT: %[[END_ITR:.*]] = cir.ptr_stride %[[DECAY]], %[[LAST_IDX]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[END_ITR:.*]] = cir.ptr_stride %[[DECAY]], %[[LAST_IDX]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> // CHECK-NEXT: cir.do { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[TEMP_LOAD]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>>, !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[TEMP_LOAD]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[TEMP_LOAD]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[TEMP_LOAD]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[TEMP_LOAD]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[TEMP_LOAD]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[TEMP_LOAD]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[TEMP_LOAD]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #false // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.int<1> : !s64i -// CHECK-NEXT: %[[NEXT_ITEM:.*]] = cir.ptr_stride %[[TEMP_LOAD]], %[[ONE]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: cir.store {{.*}} %[[NEXT_ITEM]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>> +// CHECK-NEXT: %[[NEXT_ITEM:.*]] = cir.ptr_stride %[[TEMP_LOAD]], %[[ONE]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: cir.store {{.*}} %[[NEXT_ITEM]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>> // CHECK-NEXT: cir.yield // CHECK-NEXT: } while { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(ne, %[[TEMP_LOAD]], %[[END_ITR]]) : !cir.ptr<!rec_DefaultOperators>, !cir.bool +// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>>, !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(ne, %[[TEMP_LOAD]], %[[END_ITR]]) : !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.bool // CHECK-NEXT: cir.condition(%[[CMP]]) // CHECK-NEXT: } // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}) +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[XOR]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> // CHECK-NEXT: } ; #pragma acc parallel reduction(&&:someVarArr) @@ -899,7 +1209,7 @@ void acc_compute() { // CHECK-NEXT: } ; #pragma acc parallel reduction(||:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_lor__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <lor> init { +// CHECK: acc.reduction.recipe @reduction_lor__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <lor> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, ["arrayinit.temp"] @@ -988,6 +1298,64 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.double +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !cir.double, !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) nsw : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[ADD]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } ; @@ -1036,6 +1404,64 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.double +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !cir.double, !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) nsw : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[MUL]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } ; @@ -1135,10 +1561,10 @@ void acc_compute() { // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } ; -#pragma acc parallel reduction(&:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_iand__Bcnt1__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <iand> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init"] +#pragma acc parallel reduction(&:someVarArrNoFloats[2]) +// CHECK-NEXT: acc.reduction.recipe @reduction_iand__Bcnt1__ZTSA5_24DefaultOperatorsNoFloats : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> reduction_operator <iand> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperatorsNoFloats x 5>, !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { // CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index // CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i @@ -1152,21 +1578,15 @@ void acc_compute() { // CHECK-NEXT: cir.condition(%[[COND]]) // CHECK-NEXT: } body { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: cir.yield @@ -1179,14 +1599,60 @@ void acc_compute() { // CHECK-NEXT: } // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[AND]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> // CHECK-NEXT: } ; -#pragma acc parallel reduction(|:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__Bcnt1__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <ior> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init"] +#pragma acc parallel reduction(|:someVarArrNoFloats[2]) +// CHECK-NEXT: acc.reduction.recipe @reduction_ior__Bcnt1__ZTSA5_24DefaultOperatorsNoFloats : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> reduction_operator <ior> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperatorsNoFloats x 5>, !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { // CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index // CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i @@ -1200,21 +1666,15 @@ void acc_compute() { // CHECK-NEXT: cir.condition(%[[COND]]) // CHECK-NEXT: } body { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #false // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: cir.yield @@ -1227,14 +1687,60 @@ void acc_compute() { // CHECK-NEXT: } // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[OR]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> // CHECK-NEXT: } ; -#pragma acc parallel reduction(^:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__Bcnt1__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <xor> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init"] +#pragma acc parallel reduction(^:someVarArrNoFloats[2]) +// CHECK-NEXT: acc.reduction.recipe @reduction_xor__Bcnt1__ZTSA5_24DefaultOperatorsNoFloats : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> reduction_operator <xor> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperatorsNoFloats x 5>, !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { // CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index // CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i @@ -1248,21 +1754,15 @@ void acc_compute() { // CHECK-NEXT: cir.condition(%[[COND]]) // CHECK-NEXT: } body { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #false // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: cir.yield @@ -1275,8 +1775,54 @@ void acc_compute() { // CHECK-NEXT: } // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[XOR]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> // CHECK-NEXT: } ; #pragma acc parallel reduction(&&:someVarArr[2]) @@ -1384,19 +1930,21 @@ void acc_compute() { ; #pragma acc parallel reduction(min:someVarArr[1:1]) ; -#pragma acc parallel reduction(&:someVarArr[1:1]) +#pragma acc parallel reduction(&:someVarArrNoFloats[1:1]) ; -#pragma acc parallel reduction(|:someVarArr[1:1]) +#pragma acc parallel reduction(|:someVarArrNoFloats[1:1]) ; -#pragma acc parallel reduction(^:someVarArr[1:1]) +#pragma acc parallel reduction(^:someVarArrNoFloats[1:1]) ; #pragma acc parallel reduction(&&:someVarArr[1:1]) ; #pragma acc parallel reduction(||:someVarArr[1:1]) ; + // CHECK-NEXT: cir.func {{.*}}@_Z11acc_compute } void uses() { acc_compute<DefaultOperators>(); } + diff --git a/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-float.c b/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-float.c index 1b969ac..3e4583f 100644 --- a/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-float.c +++ b/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-float.c @@ -1,4 +1,4 @@ -// RUN: %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s +// RUN: not %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s void acc_compute() { float someVar; @@ -13,7 +13,10 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!cir.float> +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.float> // CHECK-NEXT: } ; @@ -26,7 +29,10 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!cir.float> +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.float> // CHECK-NEXT: } ; @@ -56,48 +62,6 @@ void acc_compute() { // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.float> // CHECK-NEXT: } ; -#pragma acc parallel reduction(&:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTSf : !cir.ptr<!cir.float> reduction_operator <iand> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.float>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.float, !cir.ptr<!cir.float>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[ALLOCA]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: acc.yield -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.float> -// CHECK-NEXT: } - ; -#pragma acc parallel reduction(|:someVar) - -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTSf : !cir.ptr<!cir.float> reduction_operator <ior> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.float>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.float, !cir.ptr<!cir.float>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[ALLOCA]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: acc.yield -// -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.float> -// CHECK-NEXT: } - ; -#pragma acc parallel reduction(^:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTSf : !cir.ptr<!cir.float> reduction_operator <xor> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.float>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.float, !cir.ptr<!cir.float>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[ALLOCA]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: acc.yield -// -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.float> -// CHECK-NEXT: } - ; #pragma acc parallel reduction(&&:someVar) // CHECK-NEXT: acc.reduction.recipe @reduction_land__ZTSf : !cir.ptr<!cir.float> reduction_operator <land> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.float>{{.*}}) @@ -152,7 +116,31 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } ; @@ -183,7 +171,31 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } ; @@ -249,97 +261,6 @@ void acc_compute() { // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } ; -#pragma acc parallel reduction(&:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <iand> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ALL_ONES]], %[[DECAY]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ONE_IDX:.*]] = cir.const #cir.int<1> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[ONE_IDX]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ALL_ONES]], %[[NEXT_ELT]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[TWO_IDX:.*]] = cir.const #cir.int<2> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[TWO_IDX]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ALL_ONES]], %[[NEXT_ELT]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[THREE_IDX:.*]] = cir.const #cir.int<3> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[THREE_IDX]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ALL_ONES]], %[[NEXT_ELT]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[FOUR_IDX:.*]] = cir.const #cir.int<4> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[FOUR_IDX]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ALL_ONES]], %[[NEXT_ELT]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: acc.yield -// -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -// CHECK-NEXT: } - ; -#pragma acc parallel reduction(|:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <ior> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!cir.float>, !cir.ptr<!cir.ptr<!cir.float>>, ["arrayinit.temp"] -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> -// CHECK-NEXT: cir.store {{.*}} %[[DECAY]], %[[TEMP_ITR]] : !cir.ptr<!cir.float>, !cir.ptr<!cir.ptr<!cir.float>> -// CHECK-NEXT: %[[LAST_IDX:.*]] = cir.const #cir.int<5> : !s64i -// CHECK-NEXT: %[[END_ITR:.*]] = cir.ptr_stride %[[DECAY]], %[[LAST_IDX]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: cir.do { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!cir.float>>, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[TEMP_LOAD]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.int<1> : !s64i -// CHECK-NEXT: %[[NEXT_ITEM:.*]] = cir.ptr_stride %[[TEMP_LOAD]], %[[ONE]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: cir.store {{.*}} %[[NEXT_ITEM]], %[[TEMP_ITR]] : !cir.ptr<!cir.float>, !cir.ptr<!cir.ptr<!cir.float>> -// CHECK-NEXT: cir.yield -// CHECK-NEXT: } while { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!cir.float>>, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(ne, %[[TEMP_LOAD]], %[[END_ITR]]) : !cir.ptr<!cir.float>, !cir.bool -// CHECK-NEXT: cir.condition(%[[CMP]]) -// CHECK-NEXT: } -// CHECK-NEXT: acc.yield -// -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -// CHECK-NEXT: } - ; -#pragma acc parallel reduction(^:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <xor> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!cir.float>, !cir.ptr<!cir.ptr<!cir.float>>, ["arrayinit.temp"] -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> -// CHECK-NEXT: cir.store {{.*}} %[[DECAY]], %[[TEMP_ITR]] : !cir.ptr<!cir.float>, !cir.ptr<!cir.ptr<!cir.float>> -// CHECK-NEXT: %[[LAST_IDX:.*]] = cir.const #cir.int<5> : !s64i -// CHECK-NEXT: %[[END_ITR:.*]] = cir.ptr_stride %[[DECAY]], %[[LAST_IDX]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: cir.do { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!cir.float>>, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[TEMP_LOAD]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.int<1> : !s64i -// CHECK-NEXT: %[[NEXT_ITEM:.*]] = cir.ptr_stride %[[TEMP_LOAD]], %[[ONE]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: cir.store {{.*}} %[[NEXT_ITEM]], %[[TEMP_ITR]] : !cir.ptr<!cir.float>, !cir.ptr<!cir.ptr<!cir.float>> -// CHECK-NEXT: cir.yield -// CHECK-NEXT: } while { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!cir.float>>, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(ne, %[[TEMP_LOAD]], %[[END_ITR]]) : !cir.ptr<!cir.float>, !cir.bool -// CHECK-NEXT: cir.condition(%[[CMP]]) -// CHECK-NEXT: } -// CHECK-NEXT: acc.yield -// -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -// CHECK-NEXT: } - ; #pragma acc parallel reduction(&&:someVarArr) // CHECK-NEXT: acc.reduction.recipe @reduction_land__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <land> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}) @@ -434,13 +355,6 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -// CHECK-NEXT: } - ; -#pragma acc parallel reduction(*:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_mul__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <mul> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { // CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index // CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i @@ -454,10 +368,14 @@ void acc_compute() { // CHECK-NEXT: cir.condition(%[[COND]]) // CHECK-NEXT: } body { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.fp<1{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ONE]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHS_STRIDE]] // CHECK-NEXT: cir.yield // CHECK-NEXT: } step { // CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i @@ -466,49 +384,11 @@ void acc_compute() { // CHECK-NEXT: cir.yield // CHECK-NEXT: } // CHECK-NEXT: } -// CHECK-NEXT: acc.yield -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } ; -#pragma acc parallel reduction(max:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_max__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <max> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init"] -// CHECK-NEXT: cir.scope { -// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index -// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i -// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index -// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i -// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} -// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> -// CHECK-NEXT: cir.for : cond { -// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool -// CHECK-NEXT: cir.condition(%[[COND]]) -// CHECK-NEXT: } body { -// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[LEAST:.*]] = cir.const #cir.fp<-3.4{{.*}}E+38> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[LEAST]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: cir.yield -// CHECK-NEXT: } step { -// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i -// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> -// CHECK-NEXT: cir.yield -// CHECK-NEXT: } -// CHECK-NEXT: } -// CHECK-NEXT: acc.yield -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -// CHECK-NEXT: } - ; -#pragma acc parallel reduction(min:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_min__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <min> init { +#pragma acc parallel reduction(*:someVarArr[2]) +// CHECK-NEXT: acc.reduction.recipe @reduction_mul__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <mul> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { @@ -526,8 +406,8 @@ void acc_compute() { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i // CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> // CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[LARGEST:.*]] = cir.const #cir.fp<3.4{{.*}}E+38> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[LARGEST]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.fp<1{{.*}}> : !cir.float +// CHECK-NEXT: cir.store{{.*}} %[[ONE]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> // CHECK-NEXT: cir.yield // CHECK-NEXT: } step { // CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i @@ -539,13 +419,6 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -// CHECK-NEXT: } - ; -#pragma acc parallel reduction(&:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_iand__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <iand> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { // CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index // CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i @@ -559,10 +432,14 @@ void acc_compute() { // CHECK-NEXT: cir.condition(%[[COND]]) // CHECK-NEXT: } body { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ALL_ONES]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHS_STRIDE]] // CHECK-NEXT: cir.yield // CHECK-NEXT: } step { // CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i @@ -571,14 +448,11 @@ void acc_compute() { // CHECK-NEXT: cir.yield // CHECK-NEXT: } // CHECK-NEXT: } -// CHECK-NEXT: acc.yield -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } ; -#pragma acc parallel reduction(|:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <ior> init { +#pragma acc parallel reduction(max:someVarArr[2]) +// CHECK-NEXT: acc.reduction.recipe @reduction_max__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <max> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { @@ -596,8 +470,8 @@ void acc_compute() { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i // CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> // CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ZERO]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LEAST:.*]] = cir.const #cir.fp<-3.4{{.*}}E+38> : !cir.float +// CHECK-NEXT: cir.store{{.*}} %[[LEAST]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> // CHECK-NEXT: cir.yield // CHECK-NEXT: } step { // CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i @@ -609,11 +483,12 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } ; -#pragma acc parallel reduction(^:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <xor> init { +#pragma acc parallel reduction(min:someVarArr[2]) +// CHECK-NEXT: acc.reduction.recipe @reduction_min__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <min> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { @@ -631,8 +506,8 @@ void acc_compute() { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i // CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> // CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ZERO]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LARGEST:.*]] = cir.const #cir.fp<3.4{{.*}}E+38> : !cir.float +// CHECK-NEXT: cir.store{{.*}} %[[LARGEST]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> // CHECK-NEXT: cir.yield // CHECK-NEXT: } step { // CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i @@ -644,6 +519,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } ; @@ -679,6 +555,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } ; @@ -714,6 +591,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } ; @@ -726,12 +604,6 @@ void acc_compute() { ; #pragma acc parallel reduction(min:someVarArr[1:1]) ; -#pragma acc parallel reduction(&:someVarArr[1:1]) - ; -#pragma acc parallel reduction(|:someVarArr[1:1]) - ; -#pragma acc parallel reduction(^:someVarArr[1:1]) - ; #pragma acc parallel reduction(&&:someVarArr[1:1]) ; #pragma acc parallel reduction(||:someVarArr[1:1]) diff --git a/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-float.cpp b/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-float.cpp index b406bd5..833cfad 100644 --- a/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-float.cpp +++ b/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-float.cpp @@ -1,4 +1,4 @@ -// RUN: %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s +// RUN: not %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s template<typename T> void acc_compute() { @@ -14,7 +14,10 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!cir.float> +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.float> // CHECK-NEXT: } ; @@ -27,7 +30,10 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!cir.float> +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.float> // CHECK-NEXT: } @@ -58,47 +64,6 @@ void acc_compute() { // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.float> // CHECK-NEXT: } ; -#pragma acc parallel reduction(&:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTSf : !cir.ptr<!cir.float> reduction_operator <iand> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.float>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.float, !cir.ptr<!cir.float>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[ALLOCA]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: acc.yield -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.float> -// CHECK-NEXT: } - ; -#pragma acc parallel reduction(|:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTSf : !cir.ptr<!cir.float> reduction_operator <ior> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.float>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.float, !cir.ptr<!cir.float>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[ALLOCA]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: acc.yield -// -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.float> -// CHECK-NEXT: } - ; -#pragma acc parallel reduction(^:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTSf : !cir.ptr<!cir.float> reduction_operator <xor> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.float>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.float, !cir.ptr<!cir.float>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[ALLOCA]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: acc.yield -// -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.float> -// CHECK-NEXT: } - ; #pragma acc parallel reduction(&&:someVar) // CHECK-NEXT: acc.reduction.recipe @reduction_land__ZTSf : !cir.ptr<!cir.float> reduction_operator <land> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.float>{{.*}}) @@ -153,7 +118,31 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } ; @@ -184,7 +173,31 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } ; @@ -250,97 +263,6 @@ void acc_compute() { // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } ; -#pragma acc parallel reduction(&:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <iand> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ALL_ONES]], %[[DECAY]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ONE_IDX:.*]] = cir.const #cir.int<1> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[ONE_IDX]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ALL_ONES]], %[[NEXT_ELT]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[TWO_IDX:.*]] = cir.const #cir.int<2> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[TWO_IDX]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ALL_ONES]], %[[NEXT_ELT]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[THREE_IDX:.*]] = cir.const #cir.int<3> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[THREE_IDX]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ALL_ONES]], %[[NEXT_ELT]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[FOUR_IDX:.*]] = cir.const #cir.int<4> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[FOUR_IDX]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ALL_ONES]], %[[NEXT_ELT]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: acc.yield -// -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -// CHECK-NEXT: } - ; -#pragma acc parallel reduction(|:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <ior> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!cir.float>, !cir.ptr<!cir.ptr<!cir.float>>, ["arrayinit.temp"] -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> -// CHECK-NEXT: cir.store {{.*}} %[[DECAY]], %[[TEMP_ITR]] : !cir.ptr<!cir.float>, !cir.ptr<!cir.ptr<!cir.float>> -// CHECK-NEXT: %[[LAST_IDX:.*]] = cir.const #cir.int<5> : !s64i -// CHECK-NEXT: %[[END_ITR:.*]] = cir.ptr_stride %[[DECAY]], %[[LAST_IDX]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: cir.do { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!cir.float>>, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[TEMP_LOAD]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.int<1> : !s64i -// CHECK-NEXT: %[[NEXT_ITEM:.*]] = cir.ptr_stride %[[TEMP_LOAD]], %[[ONE]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: cir.store {{.*}} %[[NEXT_ITEM]], %[[TEMP_ITR]] : !cir.ptr<!cir.float>, !cir.ptr<!cir.ptr<!cir.float>> -// CHECK-NEXT: cir.yield -// CHECK-NEXT: } while { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!cir.float>>, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(ne, %[[TEMP_LOAD]], %[[END_ITR]]) : !cir.ptr<!cir.float>, !cir.bool -// CHECK-NEXT: cir.condition(%[[CMP]]) -// CHECK-NEXT: } -// CHECK-NEXT: acc.yield -// -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -// CHECK-NEXT: } - ; -#pragma acc parallel reduction(^:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <xor> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!cir.float>, !cir.ptr<!cir.ptr<!cir.float>>, ["arrayinit.temp"] -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> -// CHECK-NEXT: cir.store {{.*}} %[[DECAY]], %[[TEMP_ITR]] : !cir.ptr<!cir.float>, !cir.ptr<!cir.ptr<!cir.float>> -// CHECK-NEXT: %[[LAST_IDX:.*]] = cir.const #cir.int<5> : !s64i -// CHECK-NEXT: %[[END_ITR:.*]] = cir.ptr_stride %[[DECAY]], %[[LAST_IDX]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: cir.do { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!cir.float>>, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[TEMP_LOAD]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.int<1> : !s64i -// CHECK-NEXT: %[[NEXT_ITEM:.*]] = cir.ptr_stride %[[TEMP_LOAD]], %[[ONE]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: cir.store {{.*}} %[[NEXT_ITEM]], %[[TEMP_ITR]] : !cir.ptr<!cir.float>, !cir.ptr<!cir.ptr<!cir.float>> -// CHECK-NEXT: cir.yield -// CHECK-NEXT: } while { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!cir.float>>, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(ne, %[[TEMP_LOAD]], %[[END_ITR]]) : !cir.ptr<!cir.float>, !cir.bool -// CHECK-NEXT: cir.condition(%[[CMP]]) -// CHECK-NEXT: } -// CHECK-NEXT: acc.yield -// -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -// CHECK-NEXT: } - ; #pragma acc parallel reduction(&&:someVarArr) // CHECK-NEXT: acc.reduction.recipe @reduction_land__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <land> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}) @@ -435,48 +357,6 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -// CHECK-NEXT: } - ; -#pragma acc parallel reduction(*:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_mul__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <mul> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init"] -// CHECK-NEXT: cir.scope { -// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index -// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i -// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index -// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i -// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} -// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> -// CHECK-NEXT: cir.for : cond { -// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool -// CHECK-NEXT: cir.condition(%[[COND]]) -// CHECK-NEXT: } body { -// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.fp<1{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ONE]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: cir.yield -// CHECK-NEXT: } step { -// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i -// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> -// CHECK-NEXT: cir.yield -// CHECK-NEXT: } -// CHECK-NEXT: } -// CHECK-NEXT: acc.yield -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -// CHECK-NEXT: } - ; -#pragma acc parallel reduction(max:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_max__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <max> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { // CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index // CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i @@ -490,10 +370,14 @@ void acc_compute() { // CHECK-NEXT: cir.condition(%[[COND]]) // CHECK-NEXT: } body { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[LEAST:.*]] = cir.const #cir.fp<-3.4{{.*}}E+38> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[LEAST]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHS_STRIDE]] // CHECK-NEXT: cir.yield // CHECK-NEXT: } step { // CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i @@ -502,14 +386,11 @@ void acc_compute() { // CHECK-NEXT: cir.yield // CHECK-NEXT: } // CHECK-NEXT: } -// CHECK-NEXT: acc.yield -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } ; -#pragma acc parallel reduction(min:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_min__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <min> init { +#pragma acc parallel reduction(*:someVarArr[2]) +// CHECK-NEXT: acc.reduction.recipe @reduction_mul__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <mul> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { @@ -527,8 +408,8 @@ void acc_compute() { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i // CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> // CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[LARGEST:.*]] = cir.const #cir.fp<3.4{{.*}}E+38> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[LARGEST]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.fp<1{{.*}}> : !cir.float +// CHECK-NEXT: cir.store{{.*}} %[[ONE]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> // CHECK-NEXT: cir.yield // CHECK-NEXT: } step { // CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i @@ -540,13 +421,6 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -// CHECK-NEXT: } - ; -#pragma acc parallel reduction(&:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_iand__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <iand> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { // CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index // CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i @@ -560,10 +434,14 @@ void acc_compute() { // CHECK-NEXT: cir.condition(%[[COND]]) // CHECK-NEXT: } body { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ALL_ONES]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHS_STRIDE]] // CHECK-NEXT: cir.yield // CHECK-NEXT: } step { // CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i @@ -572,14 +450,11 @@ void acc_compute() { // CHECK-NEXT: cir.yield // CHECK-NEXT: } // CHECK-NEXT: } -// CHECK-NEXT: acc.yield -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } ; -#pragma acc parallel reduction(|:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <ior> init { +#pragma acc parallel reduction(max:someVarArr[2]) +// CHECK-NEXT: acc.reduction.recipe @reduction_max__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <max> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { @@ -597,8 +472,8 @@ void acc_compute() { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i // CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> // CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ZERO]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LEAST:.*]] = cir.const #cir.fp<-3.4{{.*}}E+38> : !cir.float +// CHECK-NEXT: cir.store{{.*}} %[[LEAST]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> // CHECK-NEXT: cir.yield // CHECK-NEXT: } step { // CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i @@ -610,11 +485,12 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } ; -#pragma acc parallel reduction(^:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <xor> init { +#pragma acc parallel reduction(min:someVarArr[2]) +// CHECK-NEXT: acc.reduction.recipe @reduction_min__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <min> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { @@ -632,8 +508,8 @@ void acc_compute() { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i // CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> // CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ZERO]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LARGEST:.*]] = cir.const #cir.fp<3.4{{.*}}E+38> : !cir.float +// CHECK-NEXT: cir.store{{.*}} %[[LARGEST]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> // CHECK-NEXT: cir.yield // CHECK-NEXT: } step { // CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i @@ -680,6 +556,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } ; @@ -715,6 +592,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } ; @@ -727,12 +605,6 @@ void acc_compute() { ; #pragma acc parallel reduction(min:someVarArr[1:1]) ; -#pragma acc parallel reduction(&:someVarArr[1:1]) - ; -#pragma acc parallel reduction(|:someVarArr[1:1]) - ; -#pragma acc parallel reduction(^:someVarArr[1:1]) - ; #pragma acc parallel reduction(&&:someVarArr[1:1]) ; #pragma acc parallel reduction(||:someVarArr[1:1]) diff --git a/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-inline-ops.cpp b/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-inline-ops.cpp index ea00c07..ec4372d 100644 --- a/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-inline-ops.cpp +++ b/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-inline-ops.cpp @@ -1,4 +1,4 @@ -// RUN: %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s +// RUN: not %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s struct HasOperatorsInline { int i; @@ -48,7 +48,7 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlinepLERS_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}): @@ -79,7 +79,7 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlinemLERS_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}): @@ -88,7 +88,7 @@ void acc_compute() { // CHECK-NEXT: } ; #pragma acc parallel reduction(max:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_max__ZTS18HasOperatorsInline : !cir.ptr<!rec_HasOperatorsInline> reduction_operator <max> init { +// CHECK: acc.reduction.recipe @reduction_max__ZTS18HasOperatorsInline : !cir.ptr<!rec_HasOperatorsInline> reduction_operator <max> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsInline, !cir.ptr<!rec_HasOperatorsInline>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_HasOperatorsInline> -> !cir.ptr<!s32i> @@ -172,7 +172,7 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineaNERS_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}): @@ -181,7 +181,7 @@ void acc_compute() { // CHECK-NEXT: } ; #pragma acc parallel reduction(|:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTS18HasOperatorsInline : !cir.ptr<!rec_HasOperatorsInline> reduction_operator <ior> init { +// CHECK: acc.reduction.recipe @reduction_ior__ZTS18HasOperatorsInline : !cir.ptr<!rec_HasOperatorsInline> reduction_operator <ior> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsInline, !cir.ptr<!rec_HasOperatorsInline>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_HasOperatorsInline> -> !cir.ptr<!s32i> @@ -203,7 +203,7 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineoRERS_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}): @@ -212,7 +212,7 @@ void acc_compute() { // CHECK-NEXT: } ; #pragma acc parallel reduction(^:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTS18HasOperatorsInline : !cir.ptr<!rec_HasOperatorsInline> reduction_operator <xor> init { +// CHECK: acc.reduction.recipe @reduction_xor__ZTS18HasOperatorsInline : !cir.ptr<!rec_HasOperatorsInline> reduction_operator <xor> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsInline, !cir.ptr<!rec_HasOperatorsInline>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_HasOperatorsInline> -> !cir.ptr<!s32i> @@ -234,7 +234,7 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineeOERS_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}): @@ -243,7 +243,7 @@ void acc_compute() { // CHECK-NEXT: } ; #pragma acc parallel reduction(&&:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_land__ZTS18HasOperatorsInline : !cir.ptr<!rec_HasOperatorsInline> reduction_operator <land> init { +// CHECK: acc.reduction.recipe @reduction_land__ZTS18HasOperatorsInline : !cir.ptr<!rec_HasOperatorsInline> reduction_operator <land> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsInline, !cir.ptr<!rec_HasOperatorsInline>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_HasOperatorsInline> -> !cir.ptr<!s32i> @@ -344,7 +344,29 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlinepLERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}): @@ -466,7 +488,28 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlinemLERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}): @@ -832,7 +875,28 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineaNERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}): @@ -896,7 +960,28 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineoRERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}): @@ -959,7 +1044,28 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineeOERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}): @@ -1216,6 +1322,32 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlinepLERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): @@ -1293,6 +1425,32 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlinemLERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): @@ -1524,6 +1682,32 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineaNERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): @@ -1601,6 +1785,32 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineoRERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): @@ -1678,6 +1888,32 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineeOERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): diff --git a/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-int.c b/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-int.c index b170ed0bf..0cee5c6 100644 --- a/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-int.c +++ b/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-int.c @@ -1,4 +1,4 @@ -// RUN: %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s +// RUN: not %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s void acc_compute() { int someVar; @@ -13,7 +13,10 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!s32i> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!s32i> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!s32i> // CHECK-NEXT: } ; @@ -27,7 +30,10 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!s32i> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!s32i> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!s32i> // CHECK-NEXT: } ; @@ -66,7 +72,10 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!s32i> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!s32i> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!s32i> // CHECK-NEXT: } ; @@ -80,7 +89,10 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!s32i> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!s32i> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!s32i> // CHECK-NEXT: } ; @@ -94,9 +106,13 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!s32i> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!s32i> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!s32i> // CHECK-NEXT: } + ; #pragma acc parallel reduction(&&:someVar) // CHECK-NEXT: acc.reduction.recipe @reduction_land__ZTSi : !cir.ptr<!s32i> reduction_operator <land> init { @@ -153,7 +169,31 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -184,7 +224,31 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -277,7 +341,31 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -307,7 +395,31 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -337,7 +449,31 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -435,6 +571,35 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -470,6 +635,35 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -505,6 +699,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -540,6 +735,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -575,6 +771,35 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -610,6 +835,35 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -645,6 +899,35 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -680,6 +963,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -715,6 +999,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; diff --git a/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-int.cpp b/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-int.cpp index c678eae..822dd9f6 100644 --- a/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-int.cpp +++ b/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-int.cpp @@ -1,4 +1,4 @@ -// RUN: %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s +// RUN: not %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s template<typename T> void acc_compute() { @@ -14,7 +14,10 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!s32i> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!s32i> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!s32i> // CHECK-NEXT: } ; @@ -28,7 +31,10 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!s32i> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!s32i> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!s32i> // CHECK-NEXT: } ; @@ -67,7 +73,10 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!s32i> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!s32i> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!s32i> // CHECK-NEXT: } ; @@ -81,7 +90,10 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!s32i> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!s32i> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!s32i> // CHECK-NEXT: } ; @@ -95,7 +107,10 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!s32i> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!s32i> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!s32i> // CHECK-NEXT: } @@ -155,7 +170,31 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -186,7 +225,31 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -279,7 +342,31 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -309,7 +396,31 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -339,7 +450,31 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -437,6 +572,35 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -472,6 +636,35 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -507,6 +700,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -542,6 +736,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -577,6 +772,35 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -612,6 +836,35 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -647,6 +900,35 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -682,6 +964,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; @@ -717,6 +1000,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } ; diff --git a/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-outline-ops.cpp b/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-outline-ops.cpp index 9ccaea2..873bf51 100644 --- a/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-outline-ops.cpp +++ b/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-outline-ops.cpp @@ -1,4 +1,4 @@ -// RUN: %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s +// RUN: not %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s struct HasOperatorsOutline { int i; unsigned u; @@ -24,7 +24,7 @@ template<typename T> void acc_compute() { T someVar; T someVarArr[5]; -#pragma acc parallel reduction(+:someVar) +#pragma acc parallel reduction(+:someVar) // CHECK: acc.reduction.recipe @reduction_add__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <add> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsOutline, !cir.ptr<!rec_HasOperatorsOutline>, ["openacc.reduction.init", init] @@ -47,7 +47,7 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZpLR19HasOperatorsOutlineS0_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}): @@ -55,8 +55,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } ; - -#pragma acc parallel reduction(*:someVar) +#pragma acc parallel reduction(*:someVar) // CHECK: acc.reduction.recipe @reduction_mul__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <mul> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsOutline, !cir.ptr<!rec_HasOperatorsOutline>, ["openacc.reduction.init", init] @@ -78,17 +77,17 @@ void acc_compute() { // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { - ; -#pragma acc parallel reduction(max:someVar) // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZmLR19HasOperatorsOutlineS0_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}): // CHECK-NEXT: cir.call @_ZN19HasOperatorsOutlineD1Ev(%[[ARG]]) nothrow : (!cir.ptr<!rec_HasOperatorsOutline>) // CHECK-NEXT: acc.yield // CHECK-NEXT: } -// CHECK-NEXT: acc.reduction.recipe @reduction_max__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <max> init { + ; +#pragma acc parallel reduction(max:someVar) +// CHECK: acc.reduction.recipe @reduction_max__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <max> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsOutline, !cir.ptr<!rec_HasOperatorsOutline>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_HasOperatorsOutline> -> !cir.ptr<!s32i> @@ -118,7 +117,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } ; -#pragma acc parallel reduction(min:someVar) +#pragma acc parallel reduction(min:someVar) // CHECK-NEXT: acc.reduction.recipe @reduction_min__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <min> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsOutline, !cir.ptr<!rec_HasOperatorsOutline>, ["openacc.reduction.init", init] @@ -149,7 +148,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } ; -#pragma acc parallel reduction(&:someVar) +#pragma acc parallel reduction(&:someVar) // CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <iand> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsOutline, !cir.ptr<!rec_HasOperatorsOutline>, ["openacc.reduction.init", init] @@ -172,7 +171,7 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZaNR19HasOperatorsOutlineS0_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}): @@ -180,8 +179,8 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } ; -#pragma acc parallel reduction(|:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <ior> init { +#pragma acc parallel reduction(|:someVar) +// CHECK: acc.reduction.recipe @reduction_ior__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <ior> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsOutline, !cir.ptr<!rec_HasOperatorsOutline>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_HasOperatorsOutline> -> !cir.ptr<!s32i> @@ -203,7 +202,7 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZoRR19HasOperatorsOutlineS0_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}): @@ -211,8 +210,8 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } ; -#pragma acc parallel reduction(^:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <xor> init { +#pragma acc parallel reduction(^:someVar) +// CHECK: acc.reduction.recipe @reduction_xor__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <xor> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsOutline, !cir.ptr<!rec_HasOperatorsOutline>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_HasOperatorsOutline> -> !cir.ptr<!s32i> @@ -234,7 +233,7 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZeOR19HasOperatorsOutlineS0_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}): @@ -242,8 +241,8 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } ; -#pragma acc parallel reduction(&&:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_land__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <land> init { +#pragma acc parallel reduction(&&:someVar) +// CHECK: acc.reduction.recipe @reduction_land__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <land> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsOutline, !cir.ptr<!rec_HasOperatorsOutline>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_HasOperatorsOutline> -> !cir.ptr<!s32i> @@ -273,7 +272,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } ; -#pragma acc parallel reduction(||:someVar) +#pragma acc parallel reduction(||:someVar) // CHECK-NEXT: acc.reduction.recipe @reduction_lor__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <lor> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsOutline, !cir.ptr<!rec_HasOperatorsOutline>, ["openacc.reduction.init", init] @@ -305,7 +304,7 @@ void acc_compute() { // CHECK-NEXT: } ; -#pragma acc parallel reduction(+:someVarArr) +#pragma acc parallel reduction(+:someVarArr) // CHECK-NEXT: acc.reduction.recipe @reduction_add__ZTSA5_19HasOperatorsOutline : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> reduction_operator <add> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_HasOperatorsOutline x 5>, !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>, ["openacc.reduction.init", init] @@ -344,7 +343,29 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZpLR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}): @@ -368,7 +389,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } ; -#pragma acc parallel reduction(*:someVarArr) +#pragma acc parallel reduction(*:someVarArr) // CHECK-NEXT: acc.reduction.recipe @reduction_mul__ZTSA5_19HasOperatorsOutline : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> reduction_operator <mul> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_HasOperatorsOutline x 5>, !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>, ["openacc.reduction.init", init] @@ -466,7 +487,28 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZmLR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}): @@ -490,7 +532,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } ; -#pragma acc parallel reduction(max:someVarArr) +#pragma acc parallel reduction(max:someVarArr) // CHECK-NEXT: acc.reduction.recipe @reduction_max__ZTSA5_19HasOperatorsOutline : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> reduction_operator <max> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_HasOperatorsOutline x 5>, !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>, ["openacc.reduction.init", init] @@ -612,7 +654,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } ; -#pragma acc parallel reduction(min:someVarArr) +#pragma acc parallel reduction(min:someVarArr) // CHECK-NEXT: acc.reduction.recipe @reduction_min__ZTSA5_19HasOperatorsOutline : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> reduction_operator <min> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_HasOperatorsOutline x 5>, !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>, ["openacc.reduction.init", init] @@ -734,7 +776,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } ; -#pragma acc parallel reduction(&:someVarArr) +#pragma acc parallel reduction(&:someVarArr) // CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTSA5_19HasOperatorsOutline : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> reduction_operator <iand> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_HasOperatorsOutline x 5>, !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>, ["openacc.reduction.init", init] @@ -832,7 +874,28 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZaNR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}): @@ -856,7 +919,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } ; -#pragma acc parallel reduction(|:someVarArr) +#pragma acc parallel reduction(|:someVarArr) // CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTSA5_19HasOperatorsOutline : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> reduction_operator <ior> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_HasOperatorsOutline x 5>, !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>, ["openacc.reduction.init", init] @@ -896,7 +959,28 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZoRR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}): @@ -920,7 +1004,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } ; -#pragma acc parallel reduction(^:someVarArr) +#pragma acc parallel reduction(^:someVarArr) // CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTSA5_19HasOperatorsOutline : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> reduction_operator <xor> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_HasOperatorsOutline x 5>, !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>, ["openacc.reduction.init", init] @@ -959,7 +1043,28 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZeOR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}): @@ -982,8 +1087,9 @@ void acc_compute() { // CHECK-NEXT: } // CHECK-NEXT: acc.yield // CHECK-NEXT: } + ; -#pragma acc parallel reduction(&&:someVarArr) +#pragma acc parallel reduction(&&:someVarArr) // CHECK-NEXT: acc.reduction.recipe @reduction_land__ZTSA5_19HasOperatorsOutline : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> reduction_operator <land> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_HasOperatorsOutline x 5>, !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>, ["openacc.reduction.init", init] @@ -1105,9 +1211,8 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } ; -#pragma acc parallel reduction(||:someVarArr) - -// CHECK-NEXT: acc.reduction.recipe @reduction_lor__ZTSA5_19HasOperatorsOutline : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> reduction_operator <lor> init { +#pragma acc parallel reduction(||:someVarArr) +// CHECK: acc.reduction.recipe @reduction_lor__ZTSA5_19HasOperatorsOutline : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> reduction_operator <lor> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_HasOperatorsOutline x 5>, !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!cir.ptr<!rec_HasOperatorsOutline>>, ["arrayinit.temp"] @@ -1171,7 +1276,7 @@ void acc_compute() { // CHECK-NEXT: } ; -#pragma acc parallel reduction(+:someVarArr[2]) +#pragma acc parallel reduction(+:someVarArr[2]) // CHECK-NEXT: acc.reduction.recipe @reduction_add__Bcnt1__ZTSA5_19HasOperatorsOutline : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> reduction_operator <add> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_HasOperatorsOutline x 5>, !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>, ["openacc.reduction.init"] @@ -1216,6 +1321,32 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZpLR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): @@ -1248,7 +1379,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } ; -#pragma acc parallel reduction(*:someVarArr[2]) +#pragma acc parallel reduction(*:someVarArr[2]) // CHECK-NEXT: acc.reduction.recipe @reduction_mul__Bcnt1__ZTSA5_19HasOperatorsOutline : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> reduction_operator <mul> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_HasOperatorsOutline x 5>, !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>, ["openacc.reduction.init"] @@ -1293,6 +1424,32 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZmLR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): @@ -1325,7 +1482,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } ; -#pragma acc parallel reduction(max:someVarArr[2]) +#pragma acc parallel reduction(max:someVarArr[2]) // CHECK-NEXT: acc.reduction.recipe @reduction_max__Bcnt1__ZTSA5_19HasOperatorsOutline : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> reduction_operator <max> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_HasOperatorsOutline x 5>, !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>, ["openacc.reduction.init"] @@ -1402,7 +1559,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } ; -#pragma acc parallel reduction(min:someVarArr[2]) +#pragma acc parallel reduction(min:someVarArr[2]) // CHECK-NEXT: acc.reduction.recipe @reduction_min__Bcnt1__ZTSA5_19HasOperatorsOutline : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> reduction_operator <min> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_HasOperatorsOutline x 5>, !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>, ["openacc.reduction.init"] @@ -1479,7 +1636,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } ; -#pragma acc parallel reduction(&:someVarArr[2]) +#pragma acc parallel reduction(&:someVarArr[2]) // CHECK-NEXT: acc.reduction.recipe @reduction_iand__Bcnt1__ZTSA5_19HasOperatorsOutline : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> reduction_operator <iand> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_HasOperatorsOutline x 5>, !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>, ["openacc.reduction.init"] @@ -1524,6 +1681,32 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZaNR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): @@ -1556,7 +1739,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } ; -#pragma acc parallel reduction(|:someVarArr[2]) +#pragma acc parallel reduction(|:someVarArr[2]) // CHECK-NEXT: acc.reduction.recipe @reduction_ior__Bcnt1__ZTSA5_19HasOperatorsOutline : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> reduction_operator <ior> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_HasOperatorsOutline x 5>, !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>, ["openacc.reduction.init"] @@ -1601,6 +1784,32 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZoRR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): @@ -1633,7 +1842,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } ; -#pragma acc parallel reduction(^:someVarArr[2]) +#pragma acc parallel reduction(^:someVarArr[2]) // CHECK-NEXT: acc.reduction.recipe @reduction_xor__Bcnt1__ZTSA5_19HasOperatorsOutline : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> reduction_operator <xor> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_HasOperatorsOutline x 5>, !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>, ["openacc.reduction.init"] @@ -1678,6 +1887,32 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZeOR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): @@ -1710,7 +1945,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } ; -#pragma acc parallel reduction(&&:someVarArr[2]) +#pragma acc parallel reduction(&&:someVarArr[2]) // CHECK-NEXT: acc.reduction.recipe @reduction_land__Bcnt1__ZTSA5_19HasOperatorsOutline : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> reduction_operator <land> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_HasOperatorsOutline x 5>, !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>, ["openacc.reduction.init"] @@ -1787,7 +2022,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } ; -#pragma acc parallel reduction(||:someVarArr[2]) +#pragma acc parallel reduction(||:someVarArr[2]) // CHECK-NEXT: acc.reduction.recipe @reduction_lor__Bcnt1__ZTSA5_19HasOperatorsOutline : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> reduction_operator <lor> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_HasOperatorsOutline x 5>, !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>, ["openacc.reduction.init"] diff --git a/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-unsigned-int.c b/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-unsigned-int.c index 783aa9a..b2d1362 100644 --- a/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-unsigned-int.c +++ b/clang/test/CIR/CodeGenOpenACC/compute-reduction-clause-unsigned-int.c @@ -1,4 +1,4 @@ -// RUN: %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s +// RUN: not %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s void acc_compute() { unsigned int someVar; @@ -13,7 +13,10 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!u32i> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!u32i> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!u32i> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!u32i> +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!u32i> // CHECK-NEXT: } ; @@ -27,7 +30,10 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!u32i> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!u32i> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!u32i> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!u32i> +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!u32i> // CHECK-NEXT: } ; @@ -66,7 +72,10 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!u32i> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!u32i> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!u32i> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!u32i> +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!u32i> // CHECK-NEXT: } ; @@ -80,7 +89,10 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!u32i> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!u32i> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!u32i> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!u32i> +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!u32i> // CHECK-NEXT: } ; @@ -94,7 +106,10 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!u32i> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!u32i> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!u32i> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!u32i> +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!u32i> // CHECK-NEXT: } ; @@ -153,7 +168,31 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!u32i>, !s64i) -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!u32i>, !s64i) -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> // CHECK-NEXT: } ; @@ -184,7 +223,31 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!u32i>, !s64i) -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!u32i>, !s64i) -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> // CHECK-NEXT: } ; @@ -277,7 +340,31 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!u32i>, !s64i) -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!u32i>, !s64i) -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> // CHECK-NEXT: } ; @@ -307,7 +394,31 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!u32i>, !s64i) -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!u32i>, !s64i) -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> // CHECK-NEXT: } ; @@ -337,7 +448,31 @@ void acc_compute() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!u32i>, !s64i) -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!u32i>, !s64i) -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> // CHECK-NEXT: } ; @@ -435,6 +570,35 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!u32i>, !u64i) -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!u32i>, !u64i) -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> // CHECK-NEXT: } ; @@ -470,6 +634,35 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!u32i>, !u64i) -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!u32i>, !u64i) -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> // CHECK-NEXT: } ; @@ -505,6 +698,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> // CHECK-NEXT: } ; @@ -540,6 +734,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> // CHECK-NEXT: } ; @@ -575,6 +770,35 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!u32i>, !u64i) -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!u32i>, !u64i) -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> // CHECK-NEXT: } ; @@ -610,6 +834,35 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!u32i>, !u64i) -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!u32i>, !u64i) -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> // CHECK-NEXT: } ; @@ -645,6 +898,35 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!u32i>, !u64i) -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!u32i>, !u64i) -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> // CHECK-NEXT: } ; @@ -680,6 +962,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> // CHECK-NEXT: } ; @@ -715,6 +998,7 @@ void acc_compute() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!u32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!u32i x 5>> // CHECK-NEXT: } ; diff --git a/clang/test/CIR/CodeGenOpenACC/loop-private-clause.cpp b/clang/test/CIR/CodeGenOpenACC/loop-private-clause.cpp index 7bbc58109..423b42b 100644 --- a/clang/test/CIR/CodeGenOpenACC/loop-private-clause.cpp +++ b/clang/test/CIR/CodeGenOpenACC/loop-private-clause.cpp @@ -158,7 +158,7 @@ struct HasDtor { // CHECK-NEXT: } extern "C" void acc_loop() { - // CHECK: cir.func{{.*}} @acc_loop() { + // CHECK: cir.func{{.*}} @acc_loop() int someInt; // CHECK-NEXT: %[[SOMEINT:.*]] = cir.alloca !s32i, !cir.ptr<!s32i>, ["someInt"] diff --git a/clang/test/CIR/CodeGenOpenACC/loop-reduction-clause-default-ops.cpp b/clang/test/CIR/CodeGenOpenACC/loop-reduction-clause-default-ops.cpp index 750c7b4..349e0fb 100644 --- a/clang/test/CIR/CodeGenOpenACC/loop-reduction-clause-default-ops.cpp +++ b/clang/test/CIR/CodeGenOpenACC/loop-reduction-clause-default-ops.cpp @@ -1,4 +1,4 @@ -// RUN: %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s +// RUN: not %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s struct DefaultOperators { int i; @@ -8,12 +8,19 @@ struct DefaultOperators { bool b; }; +struct DefaultOperatorsNoFloats { + int i; + unsigned int u; + bool b; +}; + template<typename T> void acc_loop() { T someVar; T someVarArr[5]; + struct DefaultOperatorsNoFloats someVarNoFloats; + struct DefaultOperatorsNoFloats someVarArrNoFloats[5]; #pragma acc loop reduction(+:someVar) - for(int i=0;i < 5; ++i); // CHECK: acc.reduction.recipe @reduction_add__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <add> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperators, !cir.ptr<!rec_DefaultOperators>, ["openacc.reduction.init", init] @@ -36,11 +43,43 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.double +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !cir.double, !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) nsw : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[ADD]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> // CHECK-NEXT: } + for(int i = 0; i < 5; ++i); #pragma acc loop reduction(*:someVar) - // CHECK-NEXT: acc.reduction.recipe @reduction_mul__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <mul> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperators, !cir.ptr<!rec_DefaultOperators>, ["openacc.reduction.init", init] @@ -63,10 +102,42 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.double +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !cir.double, !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) nsw : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[MUL]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc loop reduction(max:someVar) // CHECK-NEXT: acc.reduction.recipe @reduction_max__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <max> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) @@ -93,7 +164,7 @@ void acc_loop() { // TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc loop reduction(min:someVar) // CHECK-NEXT: acc.reduction.recipe @reduction_min__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <min> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) @@ -120,89 +191,130 @@ void acc_loop() { // TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc loop reduction(&:someVar) - -// CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <iand> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperators, !cir.ptr<!rec_DefaultOperators>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> + for(int i = 0; i < 5; ++i); +#pragma acc loop reduction(&:someVarNoFloats) +// CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTS24DefaultOperatorsNoFloats : !cir.ptr<!rec_DefaultOperatorsNoFloats> reduction_operator <iand> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats>{{.*}}) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperatorsNoFloats, !cir.ptr<!rec_DefaultOperatorsNoFloats>, ["openacc.reduction.init", init] +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[ALLOCA]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[ALLOCA]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[ALLOCA]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[ALLOCA]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[ALLOCA]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[ALLOCA]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats> {{.*}}) +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[AND]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperatorsNoFloats> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc loop reduction(|:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <ior> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperators, !cir.ptr<!rec_DefaultOperators>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> + for(int i = 0; i < 5; ++i); +#pragma acc loop reduction(|:someVarNoFloats) +// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTS24DefaultOperatorsNoFloats : !cir.ptr<!rec_DefaultOperatorsNoFloats> reduction_operator <ior> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats>{{.*}}) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperatorsNoFloats, !cir.ptr<!rec_DefaultOperatorsNoFloats>, ["openacc.reduction.init", init] +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[ALLOCA]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[ALLOCA]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[ALLOCA]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[ALLOCA]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[ALLOCA]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[ALLOCA]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #false // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats> {{.*}}) +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[OR]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperatorsNoFloats> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc loop reduction(^:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <xor> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperators, !cir.ptr<!rec_DefaultOperators>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> + for(int i = 0; i < 5; ++i); +#pragma acc loop reduction(^:someVarNoFloats) +// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTS24DefaultOperatorsNoFloats : !cir.ptr<!rec_DefaultOperatorsNoFloats> reduction_operator <xor> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats>{{.*}}) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_DefaultOperatorsNoFloats, !cir.ptr<!rec_DefaultOperatorsNoFloats>, ["openacc.reduction.init", init] +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[ALLOCA]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[ALLOCA]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[ALLOCA]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[ALLOCA]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[ALLOCA]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[ALLOCA]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #false // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperators> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_DefaultOperatorsNoFloats> {{.*}}) +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHSARG]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHSARG]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[XOR]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperatorsNoFloats> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc loop reduction(&&:someVar) // CHECK-NEXT: acc.reduction.recipe @reduction_land__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <land> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) @@ -229,7 +341,7 @@ void acc_loop() { // TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc loop reduction(||:someVar) // CHECK-NEXT: acc.reduction.recipe @reduction_lor__ZTS16DefaultOperators : !cir.ptr<!rec_DefaultOperators> reduction_operator <lor> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_DefaultOperators>{{.*}}) @@ -256,7 +368,7 @@ void acc_loop() { // TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_DefaultOperators> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc loop reduction(+:someVarArr) // CHECK-NEXT: acc.reduction.recipe @reduction_add__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <add> init { @@ -297,10 +409,65 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> +// +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.double +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !cir.double, !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) nsw : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[ADD]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc loop reduction(*:someVarArr) // CHECK-NEXT: acc.reduction.recipe @reduction_mul__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <mul> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}) @@ -399,11 +566,67 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> +// +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.double +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !cir.double, !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) nsw : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[MUL]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc loop reduction(max:someVarArr) + // CHECK-NEXT: acc.reduction.recipe @reduction_max__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <max> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init", init] @@ -504,7 +727,7 @@ void acc_loop() { // TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc loop reduction(min:someVarArr) // CHECK-NEXT: acc.reduction.recipe @reduction_min__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <min> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}) @@ -606,196 +829,283 @@ void acc_loop() { // TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc loop reduction(&:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <iand> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[DECAY]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> + for(int i = 0; i < 5; ++i); +#pragma acc loop reduction(&:someVarArrNoFloats) +// CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTSA5_24DefaultOperatorsNoFloats : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> reduction_operator <iand> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>{{.*}}) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperatorsNoFloats x 5>, !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>, ["openacc.reduction.init", init] +// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[DECAY]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[DECAY]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[DECAY]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[DECAY]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[DECAY]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[DECAY]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[DECAY]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // // CHECK-NEXT: %[[ALL_ONES_IDX:.*]] = cir.const #cir.int<1> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[ALL_ONES_IDX]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[ALL_ONES_IDX]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[NEXT_ELT]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // // CHECK-NEXT: %[[TWO_IDX:.*]] = cir.const #cir.int<2> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[TWO_IDX]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[TWO_IDX]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[NEXT_ELT]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // // // CHECK-NEXT: %[[THREE_IDX:.*]] = cir.const #cir.int<3> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[THREE_IDX]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[THREE_IDX]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[NEXT_ELT]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // // CHECK-NEXT: %[[FOUR_IDX:.*]] = cir.const #cir.int<4> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[FOUR_IDX]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[FOUR_IDX]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[NEXT_ELT]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[NEXT_ELT]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[NEXT_ELT]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[NEXT_ELT]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}) +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[AND]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield // CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc loop reduction(|:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <ior> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, ["arrayinit.temp"] -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: cir.store {{.*}} %[[DECAY]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>> +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> +// CHECK-NEXT: } + for(int i = 0; i < 5; ++i); +#pragma acc loop reduction(|:someVarArrNoFloats) +// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTSA5_24DefaultOperatorsNoFloats : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> reduction_operator <ior> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>{{.*}}) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperatorsNoFloats x 5>, !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>, ["openacc.reduction.init", init] +// CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>>, ["arrayinit.temp"] +// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: cir.store {{.*}} %[[DECAY]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>> // CHECK-NEXT: %[[LAST_IDX:.*]] = cir.const #cir.int<5> : !s64i -// CHECK-NEXT: %[[END_ITR:.*]] = cir.ptr_stride %[[DECAY]], %[[LAST_IDX]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[END_ITR:.*]] = cir.ptr_stride %[[DECAY]], %[[LAST_IDX]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> // CHECK-NEXT: cir.do { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[TEMP_LOAD]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>>, !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[TEMP_LOAD]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[TEMP_LOAD]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[TEMP_LOAD]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[TEMP_LOAD]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[TEMP_LOAD]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[TEMP_LOAD]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[TEMP_LOAD]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #false // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // // CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.int<1> : !s64i -// CHECK-NEXT: %[[NEXT_ITEM:.*]] = cir.ptr_stride %[[TEMP_LOAD]], %[[ONE]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: cir.store {{.*}} %[[NEXT_ITEM]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>> +// CHECK-NEXT: %[[NEXT_ITEM:.*]] = cir.ptr_stride %[[TEMP_LOAD]], %[[ONE]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: cir.store {{.*}} %[[NEXT_ITEM]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>> // CHECK-NEXT: cir.yield // CHECK-NEXT: } while { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(ne, %[[TEMP_LOAD]], %[[END_ITR]]) : !cir.ptr<!rec_DefaultOperators>, !cir.bool +// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>>, !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(ne, %[[TEMP_LOAD]], %[[END_ITR]]) : !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.bool // CHECK-NEXT: cir.condition(%[[CMP]]) // CHECK-NEXT: } // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}) +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[OR]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield // CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc loop reduction(^:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <xor> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, ["arrayinit.temp"] -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: cir.store {{.*}} %[[DECAY]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>> +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> +// CHECK-NEXT: } + for(int i = 0; i < 5; ++i); +#pragma acc loop reduction(^:someVarArrNoFloats) +// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTSA5_24DefaultOperatorsNoFloats : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> reduction_operator <xor> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>{{.*}}) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperatorsNoFloats x 5>, !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>, ["openacc.reduction.init", init] +// CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>>, ["arrayinit.temp"] +// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: cir.store {{.*}} %[[DECAY]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>> // CHECK-NEXT: %[[LAST_IDX:.*]] = cir.const #cir.int<5> : !s64i -// CHECK-NEXT: %[[END_ITR:.*]] = cir.ptr_stride %[[DECAY]], %[[LAST_IDX]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[END_ITR:.*]] = cir.ptr_stride %[[DECAY]], %[[LAST_IDX]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> // CHECK-NEXT: cir.do { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[TEMP_LOAD]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>>, !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[TEMP_LOAD]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[TEMP_LOAD]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[TEMP_LOAD]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[TEMP_LOAD]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[TEMP_LOAD]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[TEMP_LOAD]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[TEMP_LOAD]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #false // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.int<1> : !s64i -// CHECK-NEXT: %[[NEXT_ITEM:.*]] = cir.ptr_stride %[[TEMP_LOAD]], %[[ONE]] : (!cir.ptr<!rec_DefaultOperators>, !s64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: cir.store {{.*}} %[[NEXT_ITEM]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>> +// CHECK-NEXT: %[[NEXT_ITEM:.*]] = cir.ptr_stride %[[TEMP_LOAD]], %[[ONE]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: cir.store {{.*}} %[[NEXT_ITEM]], %[[TEMP_ITR]] : !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>> // CHECK-NEXT: cir.yield // CHECK-NEXT: } while { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(ne, %[[TEMP_LOAD]], %[[END_ITR]]) : !cir.ptr<!rec_DefaultOperators>, !cir.bool +// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!rec_DefaultOperatorsNoFloats>>, !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(ne, %[[TEMP_LOAD]], %[[END_ITR]]) : !cir.ptr<!rec_DefaultOperatorsNoFloats>, !cir.bool // CHECK-NEXT: cir.condition(%[[CMP]]) // CHECK-NEXT: } // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}) +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !s64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[XOR]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield // CHECK-NEXT: } - for(int i=0;i < 5; ++i); +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> +// CHECK-NEXT: } + for(int i = 0; i < 5; ++i); #pragma acc loop reduction(&&:someVarArr) // CHECK-NEXT: acc.reduction.recipe @reduction_land__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <land> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}) @@ -897,9 +1207,9 @@ void acc_loop() { // TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc loop reduction(||:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_lor__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <lor> init { +// CHECK: acc.reduction.recipe @reduction_lor__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <lor> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!rec_DefaultOperators>, !cir.ptr<!cir.ptr<!rec_DefaultOperators>>, ["arrayinit.temp"] @@ -941,7 +1251,7 @@ void acc_loop() { // TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc loop reduction(+:someVarArr[2]) // CHECK-NEXT: acc.reduction.recipe @reduction_add__Bcnt1__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <add> init { @@ -988,9 +1298,67 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.double +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[GET_MEM_LHS]] : !cir.double, !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) nsw : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[ADD]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc loop reduction(*:someVarArr[2]) // CHECK-NEXT: acc.reduction.recipe @reduction_mul__Bcnt1__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <mul> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) @@ -1036,9 +1404,67 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.double>, !cir.double +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.double +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[GET_MEM_LHS]] : !cir.double, !cir.ptr<!cir.double> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) nsw : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[MUL]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc loop reduction(max:someVarArr[2]) // CHECK-NEXT: acc.reduction.recipe @reduction_max__Bcnt1__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <max> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) @@ -1086,7 +1512,7 @@ void acc_loop() { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc loop reduction(min:someVarArr[2]) // CHECK-NEXT: acc.reduction.recipe @reduction_min__Bcnt1__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <min> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) @@ -1134,11 +1560,11 @@ void acc_loop() { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc loop reduction(&:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_iand__Bcnt1__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <iand> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init"] + for(int i = 0; i < 5; ++i); +#pragma acc loop reduction(&:someVarArrNoFloats[2]) +// CHECK-NEXT: acc.reduction.recipe @reduction_iand__Bcnt1__ZTSA5_24DefaultOperatorsNoFloats : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> reduction_operator <iand> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperatorsNoFloats x 5>, !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { // CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index // CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i @@ -1152,21 +1578,15 @@ void acc_loop() { // CHECK-NEXT: cir.condition(%[[COND]]) // CHECK-NEXT: } body { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<-1> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.int<4294967295> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xFF{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #true // CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: cir.yield @@ -1179,14 +1599,60 @@ void acc_loop() { // CHECK-NEXT: } // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[AND]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield // CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc loop reduction(|:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__Bcnt1__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <ior> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init"] +// CHECK-NEXT: } +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> +// CHECK-NEXT: } + for(int i = 0; i < 5; ++i); +#pragma acc loop reduction(|:someVarArrNoFloats[2]) +// CHECK-NEXT: acc.reduction.recipe @reduction_ior__Bcnt1__ZTSA5_24DefaultOperatorsNoFloats : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> reduction_operator <ior> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperatorsNoFloats x 5>, !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { // CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index // CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i @@ -1200,21 +1666,15 @@ void acc_loop() { // CHECK-NEXT: cir.condition(%[[COND]]) // CHECK-NEXT: } body { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #false // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: cir.yield @@ -1227,14 +1687,60 @@ void acc_loop() { // CHECK-NEXT: } // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[OR]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield // CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc loop reduction(^:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__Bcnt1__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <xor> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperators x 5>, !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>, ["openacc.reduction.init"] +// CHECK-NEXT: } +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> +// CHECK-NEXT: } + for(int i = 0; i < 5; ++i); +#pragma acc loop reduction(^:someVarArrNoFloats[2]) +// CHECK-NEXT: acc.reduction.recipe @reduction_xor__Bcnt1__ZTSA5_24DefaultOperatorsNoFloats : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> reduction_operator <xor> init { +// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_DefaultOperatorsNoFloats x 5>, !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { // CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index // CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i @@ -1248,21 +1754,15 @@ void acc_loop() { // CHECK-NEXT: cir.condition(%[[COND]]) // CHECK-NEXT: } body { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperators>, !u64i) -> !cir.ptr<!rec_DefaultOperators> -// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_I]] : !s32i, !cir.ptr<!s32i> -// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_U:.*]] = cir.get_member %[[STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !u32i // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_U]] : !u32i, !cir.ptr<!u32i> -// CHECK-NEXT: %[[GET_F:.*]] = cir.get_member %[[STRIDE]][2] {name = "f"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_F]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[GET_D:.*]] = cir.get_member %[[STRIDE]][3] {name = "d"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.double> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.double -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_D]] : !cir.double, !cir.ptr<!cir.double> -// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[STRIDE]][4] {name = "b"} : !cir.ptr<!rec_DefaultOperators> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_B:.*]] = cir.get_member %[[STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> // CHECK-NEXT: %[[ZERO:.*]] = cir.const #false // CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[GET_B]] : !cir.bool, !cir.ptr<!cir.bool> // CHECK-NEXT: cir.yield @@ -1275,10 +1775,56 @@ void acc_loop() { // CHECK-NEXT: } // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_DefaultOperatorsNoFloats>, !u64i) -> !cir.ptr<!rec_DefaultOperatorsNoFloats> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[GET_MEM_LHS]] : !s32i, !cir.ptr<!s32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][1] {name = "u"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!u32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!u32i>, !u32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !u32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[GET_MEM_LHS]] : !u32i, !cir.ptr<!u32i> +// CHECK-NEXT: %[[GET_MEM_LHS:.*]] = cir.get_member %[[LHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[GET_MEM_RHS:.*]] = cir.get_member %[[RHS_STRIDE]][2] {name = "b"} : !cir.ptr<!rec_DefaultOperatorsNoFloats> -> !cir.ptr<!cir.bool> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_RHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[RHS_INT_CAST:.*]] = cir.cast bool_to_int %[[RHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load{{.*}} %[[GET_MEM_LHS]] : !cir.ptr<!cir.bool>, !cir.bool +// CHECK-NEXT: %[[LHS_INT_CAST:.*]] = cir.cast bool_to_int %[[LHS_LOAD]] : !cir.bool -> !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_INT_CAST]], %[[RHS_INT_CAST]]) : !s32i +// CHECK-NEXT: %[[RES_TO_BOOL_CAST:.*]] = cir.cast int_to_bool %[[XOR]] : !s32i -> !cir.bool +// CHECK-NEXT: cir.store {{.*}} %[[RES_TO_BOOL_CAST]], %[[GET_MEM_LHS]] : !cir.bool, !cir.ptr<!cir.bool> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } +// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperatorsNoFloats x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc loop reduction(&&:someVarArr[2]) // CHECK-NEXT: acc.reduction.recipe @reduction_land__Bcnt1__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <land> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) @@ -1326,7 +1872,7 @@ void acc_loop() { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc loop reduction(||:someVarArr[2]) // CHECK-NEXT: acc.reduction.recipe @reduction_lor__Bcnt1__ZTSA5_16DefaultOperators : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> reduction_operator <lor> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) @@ -1374,26 +1920,26 @@ void acc_loop() { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_DefaultOperators x 5>> // CHECK-NEXT: } - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc loop reduction(+:someVarArr[1:1]) - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc loop reduction(*:someVarArr[1:1]) - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc loop reduction(max:someVarArr[1:1]) - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc loop reduction(min:someVarArr[1:1]) - for(int i=0;i < 5; ++i); -#pragma acc loop reduction(&:someVarArr[1:1]) - for(int i=0;i < 5; ++i); -#pragma acc loop reduction(|:someVarArr[1:1]) - for(int i=0;i < 5; ++i); -#pragma acc loop reduction(^:someVarArr[1:1]) - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); +#pragma acc loop reduction(&:someVarArrNoFloats[1:1]) + for(int i = 0; i < 5; ++i); +#pragma acc loop reduction(|:someVarArrNoFloats[1:1]) + for(int i = 0; i < 5; ++i); +#pragma acc loop reduction(^:someVarArrNoFloats[1:1]) + for(int i = 0; i < 5; ++i); #pragma acc loop reduction(&&:someVarArr[1:1]) - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); #pragma acc loop reduction(||:someVarArr[1:1]) - for(int i=0;i < 5; ++i); + for(int i = 0; i < 5; ++i); // CHECK-NEXT: cir.func {{.*}}@_Z8acc_loop } diff --git a/clang/test/CIR/CodeGenOpenACC/loop-reduction-clause-float.cpp b/clang/test/CIR/CodeGenOpenACC/loop-reduction-clause-float.cpp index 52406fd..8d9269b 100644 --- a/clang/test/CIR/CodeGenOpenACC/loop-reduction-clause-float.cpp +++ b/clang/test/CIR/CodeGenOpenACC/loop-reduction-clause-float.cpp @@ -1,4 +1,4 @@ -// RUN: %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s +// RUN: not %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s template<typename T> void acc_loop() { @@ -14,7 +14,10 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!cir.float> +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.float> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -27,7 +30,10 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!cir.float> +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.float> // CHECK-NEXT: } @@ -58,47 +64,6 @@ void acc_loop() { // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.float> // CHECK-NEXT: } for(int i=0;i < 5; ++i); -#pragma acc loop reduction(&:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTSf : !cir.ptr<!cir.float> reduction_operator <iand> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.float>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.float, !cir.ptr<!cir.float>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ALL_ONES]], %[[ALLOCA]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: acc.yield -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.float> -// CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc loop reduction(|:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTSf : !cir.ptr<!cir.float> reduction_operator <ior> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.float>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.float, !cir.ptr<!cir.float>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[ALLOCA]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: acc.yield -// -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.float> -// CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc loop reduction(^:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTSf : !cir.ptr<!cir.float> reduction_operator <xor> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.float>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.float, !cir.ptr<!cir.float>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[ALLOCA]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: acc.yield -// -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.float> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.float> -// CHECK-NEXT: } - for(int i=0;i < 5; ++i); #pragma acc loop reduction(&&:someVar) // CHECK-NEXT: acc.reduction.recipe @reduction_land__ZTSf : !cir.ptr<!cir.float> reduction_operator <land> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.float>{{.*}}) @@ -153,7 +118,31 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -184,7 +173,31 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -250,97 +263,6 @@ void acc_loop() { // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); -#pragma acc loop reduction(&:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_iand__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <iand> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ALL_ONES]], %[[DECAY]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ONE_IDX:.*]] = cir.const #cir.int<1> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[ONE_IDX]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ALL_ONES]], %[[NEXT_ELT]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[TWO_IDX:.*]] = cir.const #cir.int<2> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[TWO_IDX]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ALL_ONES]], %[[NEXT_ELT]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[THREE_IDX:.*]] = cir.const #cir.int<3> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[THREE_IDX]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ALL_ONES]], %[[NEXT_ELT]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[FOUR_IDX:.*]] = cir.const #cir.int<4> : !s64i -// CHECK-NEXT: %[[NEXT_ELT:.*]] = cir.ptr_stride %[[DECAY]], %[[FOUR_IDX]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ALL_ONES]], %[[NEXT_ELT]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: acc.yield -// -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -// CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc loop reduction(|:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <ior> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!cir.float>, !cir.ptr<!cir.ptr<!cir.float>>, ["arrayinit.temp"] -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> -// CHECK-NEXT: cir.store {{.*}} %[[DECAY]], %[[TEMP_ITR]] : !cir.ptr<!cir.float>, !cir.ptr<!cir.ptr<!cir.float>> -// CHECK-NEXT: %[[LAST_IDX:.*]] = cir.const #cir.int<5> : !s64i -// CHECK-NEXT: %[[END_ITR:.*]] = cir.ptr_stride %[[DECAY]], %[[LAST_IDX]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: cir.do { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!cir.float>>, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[TEMP_LOAD]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.int<1> : !s64i -// CHECK-NEXT: %[[NEXT_ITEM:.*]] = cir.ptr_stride %[[TEMP_LOAD]], %[[ONE]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: cir.store {{.*}} %[[NEXT_ITEM]], %[[TEMP_ITR]] : !cir.ptr<!cir.float>, !cir.ptr<!cir.ptr<!cir.float>> -// CHECK-NEXT: cir.yield -// CHECK-NEXT: } while { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!cir.float>>, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(ne, %[[TEMP_LOAD]], %[[END_ITR]]) : !cir.ptr<!cir.float>, !cir.bool -// CHECK-NEXT: cir.condition(%[[CMP]]) -// CHECK-NEXT: } -// CHECK-NEXT: acc.yield -// -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -// CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc loop reduction(^:someVarArr) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <xor> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init", init] -// CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!cir.float>, !cir.ptr<!cir.ptr<!cir.float>>, ["arrayinit.temp"] -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> -// CHECK-NEXT: cir.store {{.*}} %[[DECAY]], %[[TEMP_ITR]] : !cir.ptr<!cir.float>, !cir.ptr<!cir.ptr<!cir.float>> -// CHECK-NEXT: %[[LAST_IDX:.*]] = cir.const #cir.int<5> : !s64i -// CHECK-NEXT: %[[END_ITR:.*]] = cir.ptr_stride %[[DECAY]], %[[LAST_IDX]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: cir.do { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!cir.float>>, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store {{.*}} %[[ZERO]], %[[TEMP_LOAD]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.int<1> : !s64i -// CHECK-NEXT: %[[NEXT_ITEM:.*]] = cir.ptr_stride %[[TEMP_LOAD]], %[[ONE]] : (!cir.ptr<!cir.float>, !s64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: cir.store {{.*}} %[[NEXT_ITEM]], %[[TEMP_ITR]] : !cir.ptr<!cir.float>, !cir.ptr<!cir.ptr<!cir.float>> -// CHECK-NEXT: cir.yield -// CHECK-NEXT: } while { -// CHECK-NEXT: %[[TEMP_LOAD:.*]] = cir.load {{.*}} %[[TEMP_ITR]] : !cir.ptr<!cir.ptr<!cir.float>>, !cir.ptr<!cir.float> -// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(ne, %[[TEMP_LOAD]], %[[END_ITR]]) : !cir.ptr<!cir.float>, !cir.bool -// CHECK-NEXT: cir.condition(%[[CMP]]) -// CHECK-NEXT: } -// CHECK-NEXT: acc.yield -// -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -// CHECK-NEXT: } - for(int i=0;i < 5; ++i); #pragma acc loop reduction(&&:someVarArr) // CHECK-NEXT: acc.reduction.recipe @reduction_land__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <land> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}) @@ -435,13 +357,6 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -// CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc loop reduction(*:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_mul__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <mul> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { // CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index // CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i @@ -455,10 +370,14 @@ void acc_loop() { // CHECK-NEXT: cir.condition(%[[COND]]) // CHECK-NEXT: } body { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.fp<1{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ONE]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHS_STRIDE]] // CHECK-NEXT: cir.yield // CHECK-NEXT: } step { // CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i @@ -467,49 +386,11 @@ void acc_loop() { // CHECK-NEXT: cir.yield // CHECK-NEXT: } // CHECK-NEXT: } -// CHECK-NEXT: acc.yield -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); -#pragma acc loop reduction(max:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_max__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <max> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init"] -// CHECK-NEXT: cir.scope { -// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index -// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i -// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index -// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i -// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} -// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> -// CHECK-NEXT: cir.for : cond { -// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool -// CHECK-NEXT: cir.condition(%[[COND]]) -// CHECK-NEXT: } body { -// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[LEAST:.*]] = cir.const #cir.fp<-3.4{{.*}}E+38> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[LEAST]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> -// CHECK-NEXT: cir.yield -// CHECK-NEXT: } step { -// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i -// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> -// CHECK-NEXT: cir.yield -// CHECK-NEXT: } -// CHECK-NEXT: } -// CHECK-NEXT: acc.yield -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -// CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc loop reduction(min:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_min__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <min> init { +#pragma acc loop reduction(*:someVarArr[2]) +// CHECK-NEXT: acc.reduction.recipe @reduction_mul__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <mul> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { @@ -527,8 +408,8 @@ void acc_loop() { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i // CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> // CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[LARGEST:.*]] = cir.const #cir.fp<3.4{{.*}}E+38> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[LARGEST]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[ONE:.*]] = cir.const #cir.fp<1{{.*}}> : !cir.float +// CHECK-NEXT: cir.store{{.*}} %[[ONE]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> // CHECK-NEXT: cir.yield // CHECK-NEXT: } step { // CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i @@ -540,13 +421,6 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -// CHECK-NEXT: } - for(int i=0;i < 5; ++i); -#pragma acc loop reduction(&:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_iand__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <iand> init { -// CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) -// CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { // CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index // CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i @@ -560,10 +434,14 @@ void acc_loop() { // CHECK-NEXT: cir.condition(%[[COND]]) // CHECK-NEXT: } body { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i -// CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ALL_ONES:.*]] = cir.const #cir.fp<0xF{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ALL_ONES]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!cir.float>, !cir.float +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !cir.float +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHS_STRIDE]] // CHECK-NEXT: cir.yield // CHECK-NEXT: } step { // CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i @@ -572,14 +450,11 @@ void acc_loop() { // CHECK-NEXT: cir.yield // CHECK-NEXT: } // CHECK-NEXT: } -// CHECK-NEXT: acc.yield -// CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); -#pragma acc loop reduction(|:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <ior> init { +#pragma acc loop reduction(max:someVarArr[2]) +// CHECK-NEXT: acc.reduction.recipe @reduction_max__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <max> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { @@ -597,8 +472,8 @@ void acc_loop() { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i // CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> // CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ZERO]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LEAST:.*]] = cir.const #cir.fp<-3.4{{.*}}E+38> : !cir.float +// CHECK-NEXT: cir.store{{.*}} %[[LEAST]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> // CHECK-NEXT: cir.yield // CHECK-NEXT: } step { // CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i @@ -610,11 +485,12 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); -#pragma acc loop reduction(^:someVarArr[2]) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <xor> init { +#pragma acc loop reduction(min:someVarArr[2]) +// CHECK-NEXT: acc.reduction.recipe @reduction_min__Bcnt1__ZTSA5_f : !cir.ptr<!cir.array<!cir.float x 5>> reduction_operator <min> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>>{{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!cir.float x 5>, !cir.ptr<!cir.array<!cir.float x 5>>, ["openacc.reduction.init"] // CHECK-NEXT: cir.scope { @@ -632,8 +508,8 @@ void acc_loop() { // CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i // CHECK-NEXT: %[[DECAY:.*]] = cir.cast array_to_ptrdecay %[[ALLOCA]] : !cir.ptr<!cir.array<!cir.float x 5>> -> !cir.ptr<!cir.float> // CHECK-NEXT: %[[STRIDE:.*]] = cir.ptr_stride %[[DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!cir.float>, !u64i) -> !cir.ptr<!cir.float> -// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.fp<0{{.*}}> : !cir.float -// CHECK-NEXT: cir.store{{.*}} %[[ZERO]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> +// CHECK-NEXT: %[[LARGEST:.*]] = cir.const #cir.fp<3.4{{.*}}E+38> : !cir.float +// CHECK-NEXT: cir.store{{.*}} %[[LARGEST]], %[[STRIDE]] : !cir.float, !cir.ptr<!cir.float> // CHECK-NEXT: cir.yield // CHECK-NEXT: } step { // CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i @@ -645,6 +521,7 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -680,6 +557,7 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -715,6 +593,7 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.float x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!cir.float x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -727,12 +606,6 @@ void acc_loop() { for(int i=0;i < 5; ++i); #pragma acc loop reduction(min:someVarArr[1:1]) for(int i=0;i < 5; ++i); -#pragma acc loop reduction(&:someVarArr[1:1]) - for(int i=0;i < 5; ++i); -#pragma acc loop reduction(|:someVarArr[1:1]) - for(int i=0;i < 5; ++i); -#pragma acc loop reduction(^:someVarArr[1:1]) - for(int i=0;i < 5; ++i); #pragma acc loop reduction(&&:someVarArr[1:1]) for(int i=0;i < 5; ++i); #pragma acc loop reduction(||:someVarArr[1:1]) diff --git a/clang/test/CIR/CodeGenOpenACC/loop-reduction-clause-inline-ops.cpp b/clang/test/CIR/CodeGenOpenACC/loop-reduction-clause-inline-ops.cpp index d95da8c..1c89515 100644 --- a/clang/test/CIR/CodeGenOpenACC/loop-reduction-clause-inline-ops.cpp +++ b/clang/test/CIR/CodeGenOpenACC/loop-reduction-clause-inline-ops.cpp @@ -1,4 +1,4 @@ -// RUN: %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s +// RUN: not %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s struct HasOperatorsInline { int i; @@ -48,7 +48,7 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlinepLERS_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}): @@ -79,7 +79,7 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlinemLERS_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}): @@ -88,7 +88,7 @@ void acc_loop() { // CHECK-NEXT: } for(int i=0;i < 5; ++i); #pragma acc loop reduction(max:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_max__ZTS18HasOperatorsInline : !cir.ptr<!rec_HasOperatorsInline> reduction_operator <max> init { +// CHECK: acc.reduction.recipe @reduction_max__ZTS18HasOperatorsInline : !cir.ptr<!rec_HasOperatorsInline> reduction_operator <max> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsInline, !cir.ptr<!rec_HasOperatorsInline>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_HasOperatorsInline> -> !cir.ptr<!s32i> @@ -172,7 +172,7 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineaNERS_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}): @@ -181,7 +181,7 @@ void acc_loop() { // CHECK-NEXT: } for(int i=0;i < 5; ++i); #pragma acc loop reduction(|:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTS18HasOperatorsInline : !cir.ptr<!rec_HasOperatorsInline> reduction_operator <ior> init { +// CHECK: acc.reduction.recipe @reduction_ior__ZTS18HasOperatorsInline : !cir.ptr<!rec_HasOperatorsInline> reduction_operator <ior> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsInline, !cir.ptr<!rec_HasOperatorsInline>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_HasOperatorsInline> -> !cir.ptr<!s32i> @@ -203,7 +203,7 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineoRERS_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}): @@ -212,7 +212,7 @@ void acc_loop() { // CHECK-NEXT: } for(int i=0;i < 5; ++i); #pragma acc loop reduction(^:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTS18HasOperatorsInline : !cir.ptr<!rec_HasOperatorsInline> reduction_operator <xor> init { +// CHECK: acc.reduction.recipe @reduction_xor__ZTS18HasOperatorsInline : !cir.ptr<!rec_HasOperatorsInline> reduction_operator <xor> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsInline, !cir.ptr<!rec_HasOperatorsInline>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_HasOperatorsInline> -> !cir.ptr<!s32i> @@ -234,7 +234,7 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineeOERS_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsInline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline> {{.*}}): @@ -243,7 +243,7 @@ void acc_loop() { // CHECK-NEXT: } for(int i=0;i < 5; ++i); #pragma acc loop reduction(&&:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_land__ZTS18HasOperatorsInline : !cir.ptr<!rec_HasOperatorsInline> reduction_operator <land> init { +// CHECK: acc.reduction.recipe @reduction_land__ZTS18HasOperatorsInline : !cir.ptr<!rec_HasOperatorsInline> reduction_operator <land> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsInline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsInline, !cir.ptr<!rec_HasOperatorsInline>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_HasOperatorsInline> -> !cir.ptr<!s32i> @@ -344,7 +344,29 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlinepLERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}): @@ -466,7 +488,28 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlinemLERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}): @@ -832,7 +875,28 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineaNERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}): @@ -896,7 +960,28 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineoRERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}): @@ -959,7 +1044,28 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !s64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineeOERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}): @@ -1216,6 +1322,32 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlinepLERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): @@ -1293,6 +1425,32 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlinemLERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): @@ -1524,6 +1682,32 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineaNERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): @@ -1601,6 +1785,32 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineoRERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): @@ -1678,6 +1888,32 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsInline>, !u64i) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.call @_ZN18HasOperatorsInlineeOERS_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsInline>, !cir.ptr<!rec_HasOperatorsInline>) -> !cir.ptr<!rec_HasOperatorsInline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsInline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): diff --git a/clang/test/CIR/CodeGenOpenACC/loop-reduction-clause-int.cpp b/clang/test/CIR/CodeGenOpenACC/loop-reduction-clause-int.cpp index d207e07b..72e9d1f 100644 --- a/clang/test/CIR/CodeGenOpenACC/loop-reduction-clause-int.cpp +++ b/clang/test/CIR/CodeGenOpenACC/loop-reduction-clause-int.cpp @@ -1,4 +1,4 @@ -// RUN: %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s +// RUN: not %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s template<typename T> void acc_loop() { @@ -14,7 +14,10 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!s32i> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!s32i> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!s32i> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -28,7 +31,10 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!s32i> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!s32i> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!s32i> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -67,7 +73,10 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!s32i> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!s32i> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!s32i> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -81,7 +90,10 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!s32i> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!s32i> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!s32i> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -95,7 +107,10 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!s32i> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!s32i> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHSARG]] : !cir.ptr<!s32i> +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[LHSARG]] // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!s32i> // CHECK-NEXT: } @@ -155,7 +170,31 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -186,7 +225,31 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -279,7 +342,31 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -309,7 +396,31 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -339,7 +450,31 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !s64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -437,6 +572,35 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[ADD:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[ADD]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -472,6 +636,35 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[MUL:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[MUL]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -507,6 +700,7 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -542,6 +736,7 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -577,6 +772,35 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[AND:.*]] = cir.binop(and, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[AND]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -612,6 +836,35 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[OR:.*]] = cir.binop(or, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[OR]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -647,6 +900,35 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!s32i>, !u64i) -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_STRIDE]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[XOR:.*]] = cir.binop(xor, %[[LHS_LOAD]], %[[RHS_LOAD]]) : !s32i +// CHECK-NEXT: cir.store {{.*}} %[[XOR]], %[[LHS_STRIDE]] +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -682,6 +964,7 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); @@ -717,6 +1000,7 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!s32i x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// TODO OpenACC: Expecting combination operation here // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!s32i x 5>> // CHECK-NEXT: } for(int i=0;i < 5; ++i); diff --git a/clang/test/CIR/CodeGenOpenACC/loop-reduction-clause-outline-ops.cpp b/clang/test/CIR/CodeGenOpenACC/loop-reduction-clause-outline-ops.cpp index a33c25a..a36d41c1 100644 --- a/clang/test/CIR/CodeGenOpenACC/loop-reduction-clause-outline-ops.cpp +++ b/clang/test/CIR/CodeGenOpenACC/loop-reduction-clause-outline-ops.cpp @@ -1,4 +1,4 @@ -// RUN: %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s +// RUN: not %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s struct HasOperatorsOutline { int i; unsigned u; @@ -47,7 +47,7 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZpLR19HasOperatorsOutlineS0_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}): @@ -55,7 +55,6 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } for(int i=0;i < 5; ++i); - #pragma acc loop reduction(*:someVar) // CHECK: acc.reduction.recipe @reduction_mul__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <mul> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline>{{.*}}) @@ -78,17 +77,17 @@ void acc_loop() { // CHECK-NEXT: acc.yield // // CHECK-NEXT: } combiner { - for(int i=0;i < 5; ++i); -#pragma acc loop reduction(max:someVar) // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZmLR19HasOperatorsOutlineS0_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}): // CHECK-NEXT: cir.call @_ZN19HasOperatorsOutlineD1Ev(%[[ARG]]) nothrow : (!cir.ptr<!rec_HasOperatorsOutline>) // CHECK-NEXT: acc.yield // CHECK-NEXT: } -// CHECK-NEXT: acc.reduction.recipe @reduction_max__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <max> init { + for(int i=0;i < 5; ++i); +#pragma acc loop reduction(max:someVar) +// CHECK: acc.reduction.recipe @reduction_max__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <max> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsOutline, !cir.ptr<!rec_HasOperatorsOutline>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_HasOperatorsOutline> -> !cir.ptr<!s32i> @@ -172,7 +171,7 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZaNR19HasOperatorsOutlineS0_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}): @@ -181,7 +180,7 @@ void acc_loop() { // CHECK-NEXT: } for(int i=0;i < 5; ++i); #pragma acc loop reduction(|:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_ior__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <ior> init { +// CHECK: acc.reduction.recipe @reduction_ior__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <ior> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsOutline, !cir.ptr<!rec_HasOperatorsOutline>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_HasOperatorsOutline> -> !cir.ptr<!s32i> @@ -203,7 +202,7 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZoRR19HasOperatorsOutlineS0_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}): @@ -212,7 +211,7 @@ void acc_loop() { // CHECK-NEXT: } for(int i=0;i < 5; ++i); #pragma acc loop reduction(^:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_xor__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <xor> init { +// CHECK: acc.reduction.recipe @reduction_xor__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <xor> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsOutline, !cir.ptr<!rec_HasOperatorsOutline>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_HasOperatorsOutline> -> !cir.ptr<!s32i> @@ -234,7 +233,7 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: cir.call @_ZeOR19HasOperatorsOutlineS0_(%[[LHSARG]], %[[RHSARG]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!rec_HasOperatorsOutline> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}, %[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline> {{.*}}): @@ -243,7 +242,7 @@ void acc_loop() { // CHECK-NEXT: } for(int i=0;i < 5; ++i); #pragma acc loop reduction(&&:someVar) -// CHECK-NEXT: acc.reduction.recipe @reduction_land__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <land> init { +// CHECK: acc.reduction.recipe @reduction_land__ZTS19HasOperatorsOutline : !cir.ptr<!rec_HasOperatorsOutline> reduction_operator <land> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!rec_HasOperatorsOutline>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !rec_HasOperatorsOutline, !cir.ptr<!rec_HasOperatorsOutline>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[GET_I:.*]] = cir.get_member %[[ALLOCA]][0] {name = "i"} : !cir.ptr<!rec_HasOperatorsOutline> -> !cir.ptr<!s32i> @@ -344,7 +343,29 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZpLR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}): @@ -466,7 +487,28 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZmLR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}): @@ -832,7 +874,28 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZaNR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}): @@ -896,7 +959,28 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZoRR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}): @@ -959,7 +1043,28 @@ void acc_loop() { // // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}) -// TODO OpenACC: Expecting combination operation here +// CHECK-NEXT: %[[ZERO:.*]] = cir.const #cir.int<0> : !s64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !s64i, !cir.ptr<!s64i>, ["itr"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[ZERO]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[END_VAL:.*]] = cir.const #cir.int<5> : !s64i +// CHECK-NEXT: %[[CMP:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[END_VAL]]) : !s64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[CMP]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !s64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZeOR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!s64i>, !s64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !s64i, !s64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !s64i, !cir.ptr<!s64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}): @@ -982,6 +1087,7 @@ void acc_loop() { // CHECK-NEXT: } // CHECK-NEXT: acc.yield // CHECK-NEXT: } + for(int i=0;i < 5; ++i); #pragma acc loop reduction(&&:someVarArr) // CHECK-NEXT: acc.reduction.recipe @reduction_land__ZTSA5_19HasOperatorsOutline : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> reduction_operator <land> init { @@ -1106,8 +1212,7 @@ void acc_loop() { // CHECK-NEXT: } for(int i=0;i < 5; ++i); #pragma acc loop reduction(||:someVarArr) - -// CHECK-NEXT: acc.reduction.recipe @reduction_lor__ZTSA5_19HasOperatorsOutline : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> reduction_operator <lor> init { +// CHECK: acc.reduction.recipe @reduction_lor__ZTSA5_19HasOperatorsOutline : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> reduction_operator <lor> init { // CHECK-NEXT: ^bb0(%[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>{{.*}}) // CHECK-NEXT: %[[ALLOCA:.*]] = cir.alloca !cir.array<!rec_HasOperatorsOutline x 5>, !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>>, ["openacc.reduction.init", init] // CHECK-NEXT: %[[TEMP_ITR:.*]] = cir.alloca !cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!cir.ptr<!rec_HasOperatorsOutline>>, ["arrayinit.temp"] @@ -1216,6 +1321,32 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZpLR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): @@ -1293,6 +1424,32 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZmLR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): @@ -1524,6 +1681,32 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZaNR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): @@ -1601,6 +1784,32 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZoRR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): @@ -1678,6 +1887,32 @@ void acc_loop() { // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { // CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}})) +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB]] : index to !u64i +// CHECK-NEXT: %[[UB:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB]] : index to !u64i +// CHECK-NEXT: %[[ITR:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB_CAST]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR_LOAD]], %[[UB_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR_LOAD:.*]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_DECAY]], %[[ITR_LOAD]] : (!cir.ptr<!rec_HasOperatorsOutline>, !u64i) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.call @_ZeOR19HasOperatorsOutlineS0_(%[[LHS_STRIDE]], %[[RHS_STRIDE]]) : (!cir.ptr<!rec_HasOperatorsOutline>, !cir.ptr<!rec_HasOperatorsOutline>) -> !cir.ptr<!rec_HasOperatorsOutline> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR_LOAD]] = cir.load %[[ITR]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield %[[LHSARG]] : !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[ORIG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[ARG:.*]]: !cir.ptr<!cir.array<!rec_HasOperatorsOutline x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty{{.*}}): diff --git a/clang/test/CIR/CodeGenOpenACC/reduction-clause-recipes.cpp b/clang/test/CIR/CodeGenOpenACC/reduction-clause-recipes.cpp index fc696ff..20ad7a3 100644 --- a/clang/test/CIR/CodeGenOpenACC/reduction-clause-recipes.cpp +++ b/clang/test/CIR/CodeGenOpenACC/reduction-clause-recipes.cpp @@ -1,4 +1,4 @@ -// RUN: %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s +// RUN: not %clang_cc1 -fopenacc -triple x86_64-linux-gnu -Wno-openacc-self-if-potential-conflict -emit-cir -fclangir -triple x86_64-linux-pc %s -o - | FileCheck %s // Note: unlike the 'private' recipe checks, this is just for spot-checking, // so this test isn't as comprehensive. The same code paths are used for @@ -90,7 +90,88 @@ void do_things(unsigned A, unsigned B) { // CHECK-NEXT: } // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[REF:.*]]: !cir.ptr<!cir.array<!cir.array<!cir.array<!rec_NoOps x 5> x 5> x 5>> {{.*}}, %[[PRIVATE:.*]]: !cir.ptr<!cir.array<!cir.array<!cir.array<!rec_NoOps x 5> x 5> x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND2:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND3:.*]]: !acc.data_bounds_ty {{.*}}): +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.array<!cir.array<!cir.array<!rec_NoOps x 5> x 5> x 5>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.array<!cir.array<!cir.array<!rec_NoOps x 5> x 5> x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND2:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND3:.*]]: !acc.data_bounds_ty {{.*}}): +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB3:.*]] = acc.get_lowerbound %[[BOUND3]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB3_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB3]] : index to !u64i +// CHECK-NEXT: %[[UB3:.*]] = acc.get_upperbound %[[BOUND3]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB3_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB3]] : index to !u64i +// CHECK-NEXT: %[[ITR3:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB3_CAST]], %[[ITR3]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR3_LOAD:.*]] = cir.load %[[ITR3]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR3_LOAD]], %[[UB3_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR3_LOAD:.*]] = cir.load %[[ITR3]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_TLA_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHSARG]] : !cir.ptr<!cir.array<!cir.array<!cir.array<!rec_NoOps x 5> x 5> x 5>> -> !cir.ptr<!cir.array<!cir.array<!rec_NoOps x 5> x 5>> +// CHECK-NEXT: %[[LHS_BOUND3_STRIDE:.*]] = cir.ptr_stride %[[LHS_TLA_DECAY]], %[[ITR3_LOAD]] : (!cir.ptr<!cir.array<!cir.array<!rec_NoOps x 5> x 5>>, !u64i) -> !cir.ptr<!cir.array<!cir.array<!rec_NoOps x 5> x 5>> +// CHECK-NEXT: %[[RHS_TLA_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHSARG]] : !cir.ptr<!cir.array<!cir.array<!cir.array<!rec_NoOps x 5> x 5> x 5>> -> !cir.ptr<!cir.array<!cir.array<!rec_NoOps x 5> x 5>> +// CHECK-NEXT: %[[RHS_BOUND3_STRIDE:.*]] = cir.ptr_stride %[[RHS_TLA_DECAY]], %[[ITR3_LOAD]] : (!cir.ptr<!cir.array<!cir.array<!rec_NoOps x 5> x 5>>, !u64i) -> !cir.ptr<!cir.array<!cir.array<!rec_NoOps x 5> x 5>> +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB2:.*]] = acc.get_lowerbound %[[BOUND2]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB2_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB2]] : index to !u64i +// CHECK-NEXT: %[[UB2:.*]] = acc.get_upperbound %[[BOUND2]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB2_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB2]] : index to !u64i +// CHECK-NEXT: %[[ITR2:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB2_CAST]], %[[ITR2]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR2_LOAD:.*]] = cir.load %[[ITR2]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR2_LOAD]], %[[UB2_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR2_LOAD:.*]] = cir.load %[[ITR2]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_BOUND3_STRIDE_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHS_BOUND3_STRIDE]] : !cir.ptr<!cir.array<!cir.array<!rec_NoOps x 5> x 5>> -> !cir.ptr<!cir.array<!rec_NoOps x 5>> +// CHECK-NEXT: %[[LHS_BOUND2_STRIDE:.*]] = cir.ptr_stride %[[LHS_BOUND3_STRIDE_DECAY]], %[[ITR2_LOAD]] : (!cir.ptr<!cir.array<!rec_NoOps x 5>>, !u64i) -> !cir.ptr<!cir.array<!rec_NoOps x 5>> +// CHECK-NEXT: %[[RHS_BOUND3_STRIDE_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHS_BOUND3_STRIDE]] : !cir.ptr<!cir.array<!cir.array<!rec_NoOps x 5> x 5>> -> !cir.ptr<!cir.array<!rec_NoOps x 5>> +// CHECK-NEXT: %[[RHS_BOUND2_STRIDE:.*]] = cir.ptr_stride %[[RHS_BOUND3_STRIDE_DECAY]], %[[ITR2_LOAD]] : (!cir.ptr<!cir.array<!rec_NoOps x 5>>, !u64i) -> !cir.ptr<!cir.array<!rec_NoOps x 5>> +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB1:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB1_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB1]] : index to !u64i +// CHECK-NEXT: %[[UB1:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB1_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB1]] : index to !u64i +// CHECK-NEXT: %[[ITR1:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB1_CAST]], %[[ITR1]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR1_LOAD:.*]] = cir.load %[[ITR1]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR1_LOAD]], %[[UB1_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR1_LOAD:.*]] = cir.load %[[ITR1]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_BOUND2_STRIDE_DECAY:.*]] = cir.cast array_to_ptrdecay %[[LHS_BOUND2_STRIDE]] : !cir.ptr<!cir.array<!rec_NoOps x 5>> -> !cir.ptr<!rec_NoOps> +// CHECK-NEXT: %[[LHS_BOUND1_STRIDE:.*]] = cir.ptr_stride %[[LHS_BOUND2_STRIDE_DECAY]], %[[ITR1_LOAD]] : (!cir.ptr<!rec_NoOps>, !u64i) -> !cir.ptr<!rec_NoOps> +// CHECK-NEXT: %[[RHS_BOUND2_STRIDE_DECAY:.*]] = cir.cast array_to_ptrdecay %[[RHS_BOUND2_STRIDE]] : !cir.ptr<!cir.array<!rec_NoOps x 5>> -> !cir.ptr<!rec_NoOps> +// CHECK-NEXT: %[[RHS_BOUND1_STRIDE:.*]] = cir.ptr_stride %[[RHS_BOUND2_STRIDE_DECAY]], %[[ITR1_LOAD]] : (!cir.ptr<!rec_NoOps>, !u64i) -> !cir.ptr<!rec_NoOps> +// CHECK-NEXT: %[[LHS_GET_I:.*]] = cir.get_member %[[LHS_BOUND1_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_NoOps> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_GET_I:.*]] = cir.get_member %[[RHS_BOUND1_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_NoOps> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_GET_I]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_GET_I]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[OP:.*]] = cir.binop(add, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw +// CHECK-NEXT: cir.store{{.*}} %[[OP]], %[[LHS_GET_I]] : !s32i +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR1_LOAD]] = cir.load %[[ITR1]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR1_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR1]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR2_LOAD]] = cir.load %[[ITR2]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR2_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR2]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR3_LOAD]] = cir.load %[[ITR3]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR3_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR3]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield // CHECK-NEXT:} destroy { // CHECK-NEXT: ^bb0(%[[REF:.*]]: !cir.ptr<!cir.array<!cir.array<!cir.array<!rec_NoOps x 5> x 5> x 5>> {{.*}}, %[[PRIVATE:.*]]: !cir.ptr<!cir.array<!cir.array<!cir.array<!rec_NoOps x 5> x 5> x 5>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND2:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND3:.*]]: !acc.data_bounds_ty {{.*}}): @@ -355,7 +436,89 @@ void do_things(unsigned A, unsigned B) { // // CHECK-NEXT: acc.yield // CHECK-NEXT: } combiner { -// CHECK-NEXT: ^bb0(%[[REF:.*]]: !cir.ptr<!cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>> {{.*}}, %[[PRIVATE:.*]]: !cir.ptr<!cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND2:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND3:.*]]: !acc.data_bounds_ty {{.*}}): +// CHECK-NEXT: ^bb0(%[[LHSARG:.*]]: !cir.ptr<!cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>> {{.*}}, %[[RHSARG:.*]]: !cir.ptr<!cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND2:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND3:.*]]: !acc.data_bounds_ty {{.*}}): +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB3:.*]] = acc.get_lowerbound %[[BOUND3]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB3_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB3]] : index to !u64i +// CHECK-NEXT: %[[UB3:.*]] = acc.get_upperbound %[[BOUND3]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB3_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB3]] : index to !u64i +// CHECK-NEXT: %[[ITR3:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB3_CAST]], %[[ITR3]] : !u64i, !cir.ptr<!u64i> + +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR3_LOAD:.*]] = cir.load %[[ITR3]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR3_LOAD]], %[[UB3_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR3_LOAD:.*]] = cir.load %[[ITR3]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_TLA_LOAD:.*]] = cir.load %[[LHSARG]] : !cir.ptr<!cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>>, !cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>> +// CHECK-NEXT: %[[LHS_BOUND3_STRIDE:.*]] = cir.ptr_stride %[[LHS_TLA_LOAD]], %[[ITR3_LOAD]] : (!cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>, !u64i) -> !cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>> +// CHECK-NEXT: %[[RHS_TLA_LOAD:.*]] = cir.load %[[RHSARG]] : !cir.ptr<!cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>>, !cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>> +// CHECK-NEXT: %[[RHS_BOUND3_STRIDE:.*]] = cir.ptr_stride %[[RHS_TLA_LOAD]], %[[ITR3_LOAD]] : (!cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>, !u64i) -> !cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>> +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB2:.*]] = acc.get_lowerbound %[[BOUND2]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB2_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB2]] : index to !u64i +// CHECK-NEXT: %[[UB2:.*]] = acc.get_upperbound %[[BOUND2]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB2_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB2]] : index to !u64i +// CHECK-NEXT: %[[ITR2:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB2_CAST]], %[[ITR2]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR2_LOAD:.*]] = cir.load %[[ITR2]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR2_LOAD]], %[[UB2_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR2_LOAD:.*]] = cir.load %[[ITR2]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_BOUND3_STRIDE_LOAD:.*]] = cir.load %[[LHS_BOUND3_STRIDE]] : !cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>, !cir.ptr<!cir.ptr<!rec_NoOps>> +// CHECK-NEXT: %[[LHS_BOUND2_STRIDE:.*]] = cir.ptr_stride %[[LHS_BOUND3_STRIDE_LOAD]], %[[ITR2_LOAD]] : (!cir.ptr<!cir.ptr<!rec_NoOps>>, !u64i) -> !cir.ptr<!cir.ptr<!rec_NoOps>> +// CHECK-NEXT: %[[RHS_BOUND3_STRIDE_LOAD:.*]] = cir.load %[[RHS_BOUND3_STRIDE]] : !cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>, !cir.ptr<!cir.ptr<!rec_NoOps>> +// CHECK-NEXT: %[[RHS_BOUND2_STRIDE:.*]] = cir.ptr_stride %[[RHS_BOUND3_STRIDE_LOAD]], %[[ITR2_LOAD]] : (!cir.ptr<!cir.ptr<!rec_NoOps>>, !u64i) -> !cir.ptr<!cir.ptr<!rec_NoOps>> +// CHECK-NEXT: cir.scope { +// CHECK-NEXT: %[[LB1:.*]] = acc.get_lowerbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[LB1_CAST:.*]] = builtin.unrealized_conversion_cast %[[LB1]] : index to !u64i +// CHECK-NEXT: %[[UB1:.*]] = acc.get_upperbound %[[BOUND1]] : (!acc.data_bounds_ty) -> index +// CHECK-NEXT: %[[UB1_CAST:.*]] = builtin.unrealized_conversion_cast %[[UB1]] : index to !u64i +// CHECK-NEXT: %[[ITR1:.*]] = cir.alloca !u64i, !cir.ptr<!u64i>, ["iter"] {alignment = 8 : i64} +// CHECK-NEXT: cir.store %[[LB1_CAST]], %[[ITR1]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.for : cond { +// CHECK-NEXT: %[[ITR1_LOAD:.*]] = cir.load %[[ITR1]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[COND:.*]] = cir.cmp(lt, %[[ITR1_LOAD]], %[[UB1_CAST]]) : !u64i, !cir.bool +// CHECK-NEXT: cir.condition(%[[COND]]) +// CHECK-NEXT: } body { +// CHECK-NEXT: %[[ITR1_LOAD:.*]] = cir.load %[[ITR1]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[LHS_BOUND2_STRIDE_LOAD:.*]] = cir.load %[[LHS_BOUND2_STRIDE]] : !cir.ptr<!cir.ptr<!rec_NoOps>>, !cir.ptr<!rec_NoOps> +// CHECK-NEXT: %[[LHS_STRIDE:.*]] = cir.ptr_stride %[[LHS_BOUND2_STRIDE_LOAD]], %[[ITR1_LOAD]] : (!cir.ptr<!rec_NoOps>, !u64i) -> !cir.ptr<!rec_NoOps> +// CHECK-NEXT: %[[RHS_BOUND2_STRIDE_LOAD:.*]] = cir.load %[[RHS_BOUND2_STRIDE]] : !cir.ptr<!cir.ptr<!rec_NoOps>>, !cir.ptr<!rec_NoOps> +// CHECK-NEXT: %[[RHS_STRIDE:.*]] = cir.ptr_stride %[[RHS_BOUND2_STRIDE_LOAD]], %[[ITR1_LOAD]] : (!cir.ptr<!rec_NoOps>, !u64i) -> !cir.ptr<!rec_NoOps> +// CHECK-NEXT: %[[LHS_GET_I:.*]] = cir.get_member %[[LHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_NoOps> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_GET_I:.*]] = cir.get_member %[[RHS_STRIDE]][0] {name = "i"} : !cir.ptr<!rec_NoOps> -> !cir.ptr<!s32i> +// CHECK-NEXT: %[[RHS_LOAD:.*]] = cir.load {{.*}} %[[RHS_GET_I]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[LHS_LOAD:.*]] = cir.load {{.*}} %[[LHS_GET_I]] : !cir.ptr<!s32i>, !s32i +// CHECK-NEXT: %[[OP:.*]] = cir.binop(mul, %[[LHS_LOAD]], %[[RHS_LOAD]]) nsw +// CHECK-NEXT: cir.store{{.*}} %[[OP]], %[[LHS_GET_I]] : !s32i +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR1_LOAD]] = cir.load %[[ITR1]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR1_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR1]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR2_LOAD]] = cir.load %[[ITR2]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR2_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR2]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } step { +// CHECK-NEXT: %[[ITR3_LOAD]] = cir.load %[[ITR3]] : !cir.ptr<!u64i>, !u64i +// CHECK-NEXT: %[[INC:.*]] = cir.unary(inc, %[[ITR3_LOAD]]) : !u64i, !u64i +// CHECK-NEXT: cir.store %[[INC]], %[[ITR3]] : !u64i, !cir.ptr<!u64i> +// CHECK-NEXT: cir.yield +// CHECK-NEXT: } +// CHECK-NEXT: } // CHECK-NEXT: acc.yield // CHECK-NEXT: } destroy { // CHECK-NEXT: ^bb0(%[[REF:.*]]: !cir.ptr<!cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>> {{.*}}, %[[PRIVATE:.*]]: !cir.ptr<!cir.ptr<!cir.ptr<!cir.ptr<!rec_NoOps>>>> {{.*}}, %[[BOUND1:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND2:.*]]: !acc.data_bounds_ty {{.*}}, %[[BOUND3:.*]]: !acc.data_bounds_ty {{.*}}): diff --git a/clang/test/CIR/IR/atomic.cir b/clang/test/CIR/IR/atomic.cir index 8520763..790297f 100644 --- a/clang/test/CIR/IR/atomic.cir +++ b/clang/test/CIR/IR/atomic.cir @@ -5,17 +5,30 @@ cir.func @atomic_xchg(%ptr: !cir.ptr<!s32i>, %val: !s32i) { // CHECK-LABEL: @atomic_xchg - %0 = cir.atomic.xchg relaxed %ptr, %val : !cir.ptr<!s32i> -> !s32i - // CHECK: cir.atomic.xchg relaxed %{{.+}}, %{{.+}} : !cir.ptr<!s32i> -> !s32i - %1 = cir.atomic.xchg consume %ptr, %val : !cir.ptr<!s32i> -> !s32i - // CHECK: cir.atomic.xchg consume %{{.+}}, %{{.+}} : !cir.ptr<!s32i> -> !s32i - %2 = cir.atomic.xchg acquire %ptr, %val : !cir.ptr<!s32i> -> !s32i - // CHECK: cir.atomic.xchg acquire %{{.+}}, %{{.+}} : !cir.ptr<!s32i> -> !s32i - %3 = cir.atomic.xchg release %ptr, %val : !cir.ptr<!s32i> -> !s32i - // CHECK: cir.atomic.xchg release %{{.+}}, %{{.+}} : !cir.ptr<!s32i> -> !s32i - %4 = cir.atomic.xchg acq_rel %ptr, %val : !cir.ptr<!s32i> -> !s32i - // CHECK: cir.atomic.xchg acq_rel %{{.+}}, %{{.+}} : !cir.ptr<!s32i> -> !s32i - %5 = cir.atomic.xchg seq_cst %ptr, %val : !cir.ptr<!s32i> -> !s32i - // CHECK: cir.atomic.xchg seq_cst %{{.+}}, %{{.+}} : !cir.ptr<!s32i> -> !s32i + %0 = cir.atomic.xchg relaxed %ptr, %val : (!cir.ptr<!s32i>, !s32i) -> !s32i + // CHECK: cir.atomic.xchg relaxed %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i) -> !s32i + %1 = cir.atomic.xchg consume %ptr, %val : (!cir.ptr<!s32i>, !s32i) -> !s32i + // CHECK: cir.atomic.xchg consume %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i) -> !s32i + %2 = cir.atomic.xchg acquire %ptr, %val : (!cir.ptr<!s32i>, !s32i) -> !s32i + // CHECK: cir.atomic.xchg acquire %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i) -> !s32i + %3 = cir.atomic.xchg release %ptr, %val : (!cir.ptr<!s32i>, !s32i) -> !s32i + // CHECK: cir.atomic.xchg release %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i) -> !s32i + %4 = cir.atomic.xchg acq_rel %ptr, %val : (!cir.ptr<!s32i>, !s32i) -> !s32i + // CHECK: cir.atomic.xchg acq_rel %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i) -> !s32i + %5 = cir.atomic.xchg seq_cst %ptr, %val : (!cir.ptr<!s32i>, !s32i) -> !s32i + // CHECK: cir.atomic.xchg seq_cst %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i) -> !s32i + cir.return +} + +cir.func @atomic_cmpxchg(%ptr: !cir.ptr<!s32i>, %expected: !s32i, %desired: !s32i) { + // CHECK-LABEL: @atomic_cmpxchg + %0, %1 = cir.atomic.cmpxchg success(relaxed) failure(relaxed) %ptr, %expected, %desired : (!cir.ptr<!s32i>, !s32i, !s32i) -> (!s32i, !cir.bool) + // CHECK: cir.atomic.cmpxchg success(relaxed) failure(relaxed) %{{.+}}, %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i, !s32i) -> (!s32i, !cir.bool) + %2, %3 = cir.atomic.cmpxchg weak success(relaxed) failure(relaxed) %ptr, %expected, %desired : (!cir.ptr<!s32i>, !s32i, !s32i) -> (!s32i, !cir.bool) + // CHECK: cir.atomic.cmpxchg weak success(relaxed) failure(relaxed) %{{.+}}, %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i, !s32i) -> (!s32i, !cir.bool) + %4, %5 = cir.atomic.cmpxchg success(seq_cst) failure(acquire) %ptr, %expected, %desired : (!cir.ptr<!s32i>, !s32i, !s32i) -> (!s32i, !cir.bool) + // CHECK: cir.atomic.cmpxchg success(seq_cst) failure(acquire) %{{.+}}, %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i, !s32i) -> (!s32i, !cir.bool) + %6, %7 = cir.atomic.cmpxchg weak success(seq_cst) failure(acquire) %ptr, %expected, %desired : (!cir.ptr<!s32i>, !s32i, !s32i) -> (!s32i, !cir.bool) + // CHECK: cir.atomic.cmpxchg weak success(seq_cst) failure(acquire) %{{.+}}, %{{.+}}, %{{.+}} : (!cir.ptr<!s32i>, !s32i, !s32i) -> (!s32i, !cir.bool) cir.return } diff --git a/clang/test/CIR/IR/func.cir b/clang/test/CIR/IR/func.cir index d7e8184..6e91898 100644 --- a/clang/test/CIR/IR/func.cir +++ b/clang/test/CIR/IR/func.cir @@ -110,4 +110,36 @@ cir.func builtin @builtin() { } // CHECK: cir.func{{.*}} builtin @builtin() +cir.func @global_ctor_func() global_ctor { + cir.return +} + +// CHECK: cir.func @global_ctor_func() global_ctor { +// CHECK: cir.return +// CHECK: } + +cir.func @global_ctor_with_priority() global_ctor(200) { + cir.return +} + +// CHECK: cir.func @global_ctor_with_priority() global_ctor(200) { +// CHECK: cir.return +// CHECK: } + +cir.func @global_dtor_func() global_dtor { + cir.return +} + +// CHECK: cir.func @global_dtor_func() global_dtor { +// CHECK: cir.return +// CHECK: } + +cir.func @global_dtor_with_priority() global_dtor(201) { + cir.return +} + +// CHECK: cir.func @global_dtor_with_priority() global_dtor(201) { +// CHECK: cir.return +// CHECK: } + } diff --git a/clang/test/CIR/IR/inline-attrs.cir b/clang/test/CIR/IR/inline-attrs.cir new file mode 100644 index 0000000..f525abe --- /dev/null +++ b/clang/test/CIR/IR/inline-attrs.cir @@ -0,0 +1,33 @@ +// RUN: cir-opt %s --verify-roundtrip | FileCheck %s + +!s32i = !cir.int<s, 32> + +module { + cir.func @noinline_func(%arg0: !s32i) -> !s32i inline(never) { + cir.return %arg0 : !s32i + } + cir.func @always_inline_func(%arg0: !s32i) -> !s32i inline(always) { + cir.return %arg0 : !s32i + } + cir.func @inline_hint_func(%arg0: !s32i) -> !s32i inline(hint) { + cir.return %arg0 : !s32i + } + cir.func @regular_func(%arg0: !s32i) -> !s32i { + cir.return %arg0 : !s32i + } + cir.func dso_local @noinline_with_attrs(%arg0: !s32i) -> !s32i inline(never) { + cir.return %arg0 : !s32i + } + cir.func private @noinline_decl(!s32i) -> !s32i inline(never) + cir.func private @always_inline_decl(!s32i) -> !s32i inline(always) + cir.func private @inline_hint_decl(!s32i) -> !s32i inline(hint) +} + +// CHECK: cir.func @noinline_func(%arg0: !s32i) -> !s32i inline(never) +// CHECK: cir.func @always_inline_func(%arg0: !s32i) -> !s32i inline(always) +// CHECK: cir.func @inline_hint_func(%arg0: !s32i) -> !s32i inline(hint) +// CHECK: cir.func @regular_func(%arg0: !s32i) -> !s32i { +// CHECK: cir.func dso_local @noinline_with_attrs(%arg0: !s32i) -> !s32i inline(never) +// CHECK: cir.func private @noinline_decl(!s32i) -> !s32i inline(never) +// CHECK: cir.func private @always_inline_decl(!s32i) -> !s32i inline(always) +// CHECK: cir.func private @inline_hint_decl(!s32i) -> !s32i inline(hint) diff --git a/clang/test/CIR/IR/invalid-try-catch.cir b/clang/test/CIR/IR/invalid-try-catch.cir new file mode 100644 index 0000000..94df4b6 --- /dev/null +++ b/clang/test/CIR/IR/invalid-try-catch.cir @@ -0,0 +1,157 @@ +// RUN: cir-opt %s -verify-diagnostics -split-input-file + +module { + +cir.func dso_local @invalid_catch_without_all_or_type() { + cir.scope { + cir.try { + cir.yield + // expected-error @below {{'cir.try' expected 'all' or 'type' keyword}} + } catch [invalid_keyword { + cir.yield + } + } + cir.return +} + +} + +// ----- + +module { + +cir.func dso_local @invalid_catch_rtti_type() { + cir.scope { + // expected-error @below {{'cir.try' op attribute 'handler_types' failed to satisfy constraint: catch all or unwind or global view array attribute}} + cir.try { + cir.yield + } catch [type #cir.undef] { + cir.yield + } + } + cir.return +} + +} + +// ----- + +module { + +cir.func dso_local @invalid_catch_empty_block() { + cir.scope { + cir.try { + cir.yield + } + // expected-error @below {{'cir.try' handler region shall not be empty}} + catch all { + } + } + cir.return +} + +} + +// ----- + +!s32i = !cir.int<s, 32> + +module { + +cir.func dso_local @invalid_catch_not_terminated() { + %a = cir.alloca !s32i, !cir.ptr<!s32i>, ["a", init] + cir.scope { + cir.try { + cir.yield + } + // expected-error @below {{'cir.try' blocks are expected to be explicitly terminated}} + catch all { + %tmp_a = cir.load %a : !cir.ptr<!s32i>, !s32i + } + } + cir.return +} + +} + +// ----- + +module { + +cir.func dso_local @invalid_catch_multiple_catch_all() { + cir.scope { + cir.try { + cir.yield + } catch all { + cir.yield + } + // expected-error @below {{op 'cir.try' can't have more than one catch all}} + catch all { + cir.yield + } + } + cir.return +} + +} + +// ----- + +module { + +cir.func dso_local @invalid_catch_without_type_info() { + cir.scope { + cir.try { + cir.yield + } + // expected-error @below {{expected attribute value}} + // expected-error @below {{op 'cir.try' expected valid RTTI info attribute}} + catch [type] { + cir.yield + } + } + cir.return +} + +} + +// ----- + +module { + +cir.func dso_local @invalid_catch_all_with_type_info() { + cir.scope { + cir.try { + cir.yield + } + // expected-error @below {{op 'cir.try' catch all dosen't need RTTI info attribute}} + catch [all] { + cir.yield + } + } + cir.return +} + +} + +// ----- + +module { + +cir.func dso_local @invalid_unwind_with_catch_all() { + cir.scope { + cir.try { + cir.yield + } + catch all { + cir.yield + } + // expected-error @below {{op 'cir.try' unwind can't be used with catch all}} + unwind { + + } + } + cir.return +} + +} diff --git a/clang/test/CIR/IR/try-catch.cir b/clang/test/CIR/IR/try-catch.cir new file mode 100644 index 0000000..7becd0b --- /dev/null +++ b/clang/test/CIR/IR/try-catch.cir @@ -0,0 +1,84 @@ +// RUN: cir-opt %s --verify-roundtrip | FileCheck %s + +!u8i = !cir.int<u, 8> + +module { + +cir.global "private" constant external @_ZTIi : !cir.ptr<!u8i> +cir.global "private" constant external @_ZTIPKc : !cir.ptr<!u8i> + +cir.func dso_local @empty_try_block_with_catch_all() { + cir.scope { + cir.try { + cir.yield + } catch all { + cir.yield + } + } + cir.return +} + +// CHECK: cir.func dso_local @empty_try_block_with_catch_all() { +// CHECK: cir.scope { +// CHECK: cir.try { +// CHECK: cir.yield +// CHECK: } catch all { +// CHECK: cir.yield +// CHECK: } +// CHECK: } +// CHECK: cir.return +// CHECK: } + +cir.func dso_local @empty_try_block_with_catch_unwind() { + cir.scope { + cir.try { + cir.yield + } unwind { + cir.yield + } + } + cir.return +} + +// CHECK: cir.func dso_local @empty_try_block_with_catch_unwind() { +// CHECK: cir.scope { +// CHECK: cir.try { +// CHECK: cir.yield +// CHECK: } unwind { +// CHECK: cir.yield +// CHECK: } +// CHECK: } +// CHECK: cir.return +// CHECK: } + +cir.func dso_local @empty_try_block_with_catch_ist() { + cir.scope { + cir.try { + cir.yield + } catch [type #cir.global_view<@_ZTIi> : !cir.ptr<!u8i>] { + cir.yield + } catch [type #cir.global_view<@_ZTIPKc> : !cir.ptr<!u8i>] { + cir.yield + } unwind { + cir.yield + } + } + cir.return +} + +// CHECK: cir.func dso_local @empty_try_block_with_catch_ist() { +// CHECK: cir.scope { +// CHECK: cir.try { +// CHECK: cir.yield +// CHECK: } catch [type #cir.global_view<@_ZTIi> : !cir.ptr<!u8i>] { +// CHECK: cir.yield +// CHECK: } catch [type #cir.global_view<@_ZTIPKc> : !cir.ptr<!u8i>] { +// CHECK: cir.yield +// CHECK: } unwind { +// CHECK: cir.yield +// CHECK: } +// CHECK: } +// CHECK: cir.return +// CHECK: } + +} diff --git a/clang/test/CIR/Lowering/basic.cpp b/clang/test/CIR/Lowering/basic.cpp index 5642cef..63beb0a 100644 --- a/clang/test/CIR/Lowering/basic.cpp +++ b/clang/test/CIR/Lowering/basic.cpp @@ -5,7 +5,7 @@ int f1() { return i; } -// CHECK: define{{.*}} i32 @_Z2f1v() { +// CHECK: define{{.*}} i32 @_Z2f1v(){{.*}} { // CHECK: %[[RV:.*]] = alloca i32, i64 1, align 4 // CHECK: %[[I_PTR:.*]] = alloca i32, i64 1, align 4 // CHECK: %[[I:.*]] = load i32, ptr %[[I_PTR]], align 4 @@ -18,7 +18,7 @@ int f2() { return i; } -// CHECK: define{{.*}} i32 @_Z2f2v() { +// CHECK: define{{.*}} i32 @_Z2f2v(){{.*}} { // CHECK: %[[RV:.*]] = alloca i32, i64 1, align 4 // CHECK: %[[I_PTR:.*]] = alloca i32, i64 1, align 4 // CHECK: store i32 2, ptr %[[I_PTR]], align 4 diff --git a/clang/test/CIR/Lowering/func-simple.cpp b/clang/test/CIR/Lowering/func-simple.cpp index 96306bab..df5b007 100644 --- a/clang/test/CIR/Lowering/func-simple.cpp +++ b/clang/test/CIR/Lowering/func-simple.cpp @@ -23,7 +23,7 @@ int scopes() { } } } -// CHECK: define{{.*}} i32 @_Z6scopesv() { +// CHECK: define{{.*}} i32 @_Z6scopesv(){{.*}} { // CHECK: %[[RV:.*]] = alloca i32, i64 1, align 4 // CHECK: br label %[[LABEL1:.*]] // CHECK: [[LABEL1]]: @@ -40,7 +40,7 @@ int scopes() { // CHECK: } long longfunc() { return 42l; } -// CHECK: define{{.*}} i64 @_Z8longfuncv() { +// CHECK: define{{.*}} i64 @_Z8longfuncv(){{.*}} { // CHECK: %[[RV:.*]] = alloca i64, i64 1, align 8 // CHECK: store i64 42, ptr %[[RV]], align 8 // CHECK: %[[R:.*]] = load i64, ptr %[[RV]], align 8 @@ -48,7 +48,7 @@ long longfunc() { return 42l; } // CHECK: } unsigned unsignedfunc() { return 42u; } -// CHECK: define{{.*}} i32 @_Z12unsignedfuncv() { +// CHECK: define{{.*}} i32 @_Z12unsignedfuncv(){{.*}} { // CHECK: %[[RV:.*]] = alloca i32, i64 1, align 4 // CHECK: store i32 42, ptr %[[RV]], align 4 // CHECK: %[[R:.*]] = load i32, ptr %[[RV]], align 4 @@ -56,7 +56,7 @@ unsigned unsignedfunc() { return 42u; } // CHECK: } unsigned long long ullfunc() { return 42ull; } -// CHECK: define{{.*}} i64 @_Z7ullfuncv() { +// CHECK: define{{.*}} i64 @_Z7ullfuncv(){{.*}} { // CHECK: %[[RV:.*]] = alloca i64, i64 1, align 8 // CHECK: store i64 42, ptr %[[RV]], align 8 // CHECK: %[[R:.*]] = load i64, ptr %[[RV]], align 8 @@ -64,7 +64,7 @@ unsigned long long ullfunc() { return 42ull; } // CHECK: } bool boolfunc() { return true; } -// CHECK: define{{.*}} i1 @_Z8boolfuncv() { +// CHECK: define{{.*}} i1 @_Z8boolfuncv(){{.*}} { // CHECK: %[[RV:.*]] = alloca i8, i64 1, align 1 // CHECK: store i8 1, ptr %[[RV]], align 1 // CHECK: %[[R8:.*]] = load i8, ptr %[[RV]], align 1 diff --git a/clang/test/CIR/func-simple.cpp b/clang/test/CIR/func-simple.cpp index c9cb5c5..8894757 100644 --- a/clang/test/CIR/func-simple.cpp +++ b/clang/test/CIR/func-simple.cpp @@ -2,17 +2,17 @@ // RUN: %clang_cc1 -std=c++20 -triple x86_64-unknown-linux-gnu -fclangir -emit-cir %s -o - | FileCheck %s void empty() { } -// CHECK: cir.func{{.*}} @_Z5emptyv() { +// CHECK: cir.func{{.*}} @_Z5emptyv() // CHECK: cir.return // CHECK: } void voidret() { return; } -// CHECK: cir.func{{.*}} @_Z7voidretv() { +// CHECK: cir.func{{.*}} @_Z7voidretv() // CHECK: cir.return // CHECK: } int intfunc() { return 42; } -// CHECK: cir.func{{.*}} @_Z7intfuncv() -> !s32i { +// CHECK: cir.func{{.*}} @_Z7intfuncv() -> !s32i // CHECK: %0 = cir.alloca !s32i, !cir.ptr<!s32i>, ["__retval"] {alignment = 4 : i64} // CHECK: %1 = cir.const #cir.int<42> : !s32i // CHECK: cir.store %1, %0 : !s32i, !cir.ptr<!s32i> @@ -27,7 +27,7 @@ int scopes() { } } } -// CHECK: cir.func{{.*}} @_Z6scopesv() -> !s32i { +// CHECK: cir.func{{.*}} @_Z6scopesv() -> !s32i // CHECK: %0 = cir.alloca !s32i, !cir.ptr<!s32i>, ["__retval"] {alignment = 4 : i64} // CHECK: cir.scope { // CHECK: cir.scope { @@ -68,7 +68,7 @@ unsigned long long ullfunc() { return 42ull; } // CHECK: } bool boolfunc() { return true; } -// CHECK: cir.func{{.*}} @_Z8boolfuncv() -> !cir.bool { +// CHECK: cir.func{{.*}} @_Z8boolfuncv() -> !cir.bool // CHECK: %0 = cir.alloca !cir.bool, !cir.ptr<!cir.bool>, ["__retval"] {alignment = 1 : i64} // CHECK: %1 = cir.const #true // CHECK: cir.store %1, %0 : !cir.bool, !cir.ptr<!cir.bool> @@ -77,7 +77,7 @@ bool boolfunc() { return true; } // CHECK: } float floatfunc() { return 42.42f; } -// CHECK: cir.func{{.*}} @_Z9floatfuncv() -> !cir.float { +// CHECK: cir.func{{.*}} @_Z9floatfuncv() -> !cir.float // CHECK: %0 = cir.alloca !cir.float, !cir.ptr<!cir.float>, ["__retval"] {alignment = 4 : i64} // CHECK: %1 = cir.const #cir.fp<4.242 // CHECK: cir.store %1, %0 : !cir.float, !cir.ptr<!cir.float> @@ -86,7 +86,7 @@ float floatfunc() { return 42.42f; } // CHECK: } double doublefunc() { return 42.42; } -// CHECK: cir.func{{.*}} @_Z10doublefuncv() -> !cir.double { +// CHECK: cir.func{{.*}} @_Z10doublefuncv() -> !cir.double // CHECK: %0 = cir.alloca !cir.double, !cir.ptr<!cir.double>, ["__retval"] {alignment = 8 : i64} // CHECK: %1 = cir.const #cir.fp<4.242 // CHECK: cir.store %1, %0 : !cir.double, !cir.ptr<!cir.double> diff --git a/clang/test/CXX/dcl.decl/dcl.decomp/p2.cpp b/clang/test/CXX/dcl.decl/dcl.decomp/p2.cpp index cad7d8e..717ea3aa 100644 --- a/clang/test/CXX/dcl.decl/dcl.decomp/p2.cpp +++ b/clang/test/CXX/dcl.decl/dcl.decomp/p2.cpp @@ -16,8 +16,8 @@ int array() { using X3 = X[3]; auto [a3, b3, c3] = X3{1, 2, 3}; - auto &[d, e] = arr; // expected-error {{type 'int[3]' decomposes into 3 elements, but only 2 names were provided}} - auto &[f, g, h, i] = arr; // expected-error {{type 'int[3]' decomposes into 3 elements, but 4 names were provided}} + auto &[d, e] = arr; // expected-error {{type 'int[3]' binds to 3 elements, but only 2 names were provided}} + auto &[f, g, h, i] = arr; // expected-error {{type 'int[3]' binds to 3 elements, but 4 names were provided}} auto &[r0, r1, r2] = arr; const auto &[cr0, cr1, cr2] = arr; diff --git a/clang/test/CXX/dcl.decl/dcl.decomp/p3.cpp b/clang/test/CXX/dcl.decl/dcl.decomp/p3.cpp index ce5eefc..b7fef12 100644 --- a/clang/test/CXX/dcl.decl/dcl.decomp/p3.cpp +++ b/clang/test/CXX/dcl.decl/dcl.decomp/p3.cpp @@ -5,10 +5,10 @@ using size_t = decltype(sizeof(0)); struct A { int x, y; }; struct B { int x, y; }; -void no_tuple_size_1() { auto [x, y] = A(); } // ok, decompose elementwise +void no_tuple_size_1() { auto [x, y] = A(); } // ok, bind elementwise namespace std { template<typename T> struct tuple_size; } -void no_tuple_size_2() { auto [x, y] = A(); } // ok, decompose elementwise +void no_tuple_size_2() { auto [x, y] = A(); } // ok, bind elementwise struct Bad1 { int a, b; }; template<> struct std::tuple_size<Bad1> {}; @@ -16,15 +16,15 @@ void no_tuple_size_3() { auto [x, y] = Bad1(); } // ok, omitting value is valid struct Bad2 {}; template<> struct std::tuple_size<Bad2> { const int value = 5; }; -void no_tuple_size_4() { auto [x, y] = Bad2(); } // expected-error {{cannot decompose this type; 'std::tuple_size<Bad2>::value' is not a valid integral constant expression}} +void no_tuple_size_4() { auto [x, y] = Bad2(); } // expected-error {{cannot bind this type; 'std::tuple_size<Bad2>::value' is not a valid integral constant expression}} template<> struct std::tuple_size<A> { static const int value = 3; }; template<> struct std::tuple_size<B> { enum { value = 3 }; }; void no_get_1() { { - auto [a0, a1] = A(); // expected-error {{decomposes into 3 elements}} - auto [b0, b1] = B(); // expected-error {{decomposes into 3 elements}} + auto [a0, a1] = A(); // expected-error {{binds to 3 elements}} + auto [b0, b1] = B(); // expected-error {{binds to 3 elements}} } auto [a0, a1, a2] = A(); // expected-error {{undeclared identifier 'get'}} expected-note {{in implicit initialization of binding declaration 'a0'}} } diff --git a/clang/test/CXX/dcl.decl/dcl.decomp/p4.cpp b/clang/test/CXX/dcl.decl/dcl.decomp/p4.cpp index 7141124..532a967 100644 --- a/clang/test/CXX/dcl.decl/dcl.decomp/p4.cpp +++ b/clang/test/CXX/dcl.decl/dcl.decomp/p4.cpp @@ -25,10 +25,10 @@ namespace NonPublicMembers { struct NonPublic4 : NonPublic2 {}; void test() { - auto [a1] = NonPublic1(); // expected-error {{cannot decompose protected member 'a' of 'NonPublicMembers::NonPublic1'}} - auto [a2] = NonPublic2(); // expected-error {{cannot decompose private member 'a' of 'NonPublicMembers::NonPublic2'}} - auto [a3] = NonPublic3(); // expected-error {{cannot decompose members of inaccessible base class 'A' of 'NonPublicMembers::NonPublic3'}} - auto [a4] = NonPublic4(); // expected-error {{cannot decompose private member 'a' of 'NonPublicMembers::NonPublic2'}} + auto [a1] = NonPublic1(); // expected-error {{cannot bind protected member 'a' of 'NonPublicMembers::NonPublic1'}} + auto [a2] = NonPublic2(); // expected-error {{cannot bind private member 'a' of 'NonPublicMembers::NonPublic2'}} + auto [a3] = NonPublic3(); // expected-error {{cannot bind members of inaccessible base class 'A' of 'NonPublicMembers::NonPublic3'}} + auto [a4] = NonPublic4(); // expected-error {{cannot bind private member 'a' of 'NonPublicMembers::NonPublic2'}} } } @@ -46,8 +46,8 @@ namespace AnonymousMember { }; void test() { - auto [a1] = Struct(); // expected-error {{cannot decompose class type 'Struct' because it has an anonymous struct member}} - auto [a2] = Union(); // expected-error {{cannot decompose class type 'Union' because it has an anonymous union member}} + auto [a1] = Struct(); // expected-error {{cannot bind class type 'Struct' because it has an anonymous struct member}} + auto [a2] = Union(); // expected-error {{cannot bind class type 'Union' because it has an anonymous union member}} } } @@ -73,12 +73,12 @@ namespace MultipleClasses { struct M : virtual J, L {}; void test() { - auto [b] = B(); // expected-error {{cannot decompose class type 'B': both it and its base class 'A' have non-static data members}} - auto [d] = D(); // expected-error {{cannot decompose class type 'D': its base classes 'A' and 'C' have non-static data members}} + auto [b] = B(); // expected-error {{cannot bind class type 'B': both it and its base class 'A' have non-static data members}} + auto [d] = D(); // expected-error {{cannot bind class type 'D': its base classes 'A' and 'C' have non-static data members}} auto [e] = E(); - auto [f] = F(); // expected-error-re {{cannot decompose members of ambiguous base class 'A' of 'F':{{.*}}struct MultipleClasses::F -> A{{.*}}struct MultipleClasses::F -> E -> A}} + auto [f] = F(); // expected-error-re {{cannot bind members of ambiguous base class 'A' of 'F':{{.*}}struct MultipleClasses::F -> A{{.*}}struct MultipleClasses::F -> E -> A}} auto [h] = H(); // ok, only one (virtual) base subobject even though there are two paths to it - auto [k] = K(); // expected-error {{cannot decompose members of ambiguous base class 'I'}} + auto [k] = K(); // expected-error {{cannot bind members of ambiguous base class 'I'}} auto [m] = M(); // ok, all paths to I are through the same virtual base subobject J same<decltype(m), int>(); @@ -214,7 +214,7 @@ namespace p0969r0 { auto &[x, y] = b; } void test_external(B b) { - auto &[x, y] = b; // expected-error {{cannot decompose members of inaccessible base class 'A' of 'p0969r0::B'}} + auto &[x, y] = b; // expected-error {{cannot bind members of inaccessible base class 'A' of 'p0969r0::B'}} } struct C { @@ -229,13 +229,13 @@ namespace p0969r0 { struct D : C { static void test_member(D d, C c) { auto &[x1, y1] = d; - auto &[x2, y2] = c; // expected-error {{cannot decompose protected member 'y' of 'p0969r0::C'}} + auto &[x2, y2] = c; // expected-error {{cannot bind protected member 'y' of 'p0969r0::C'}} } }; void test_friend(D d) { auto &[x, y] = d; } void test_external(D d) { - auto &[x, y] = d; // expected-error {{cannot decompose protected member 'y' of 'p0969r0::C'}} + auto &[x, y] = d; // expected-error {{cannot bind protected member 'y' of 'p0969r0::C'}} } } diff --git a/clang/test/CXX/drs/cwg22xx.cpp b/clang/test/CXX/drs/cwg22xx.cpp index 8c8ad9f..34119a1 100644 --- a/clang/test/CXX/drs/cwg22xx.cpp +++ b/clang/test/CXX/drs/cwg22xx.cpp @@ -202,7 +202,7 @@ namespace cwg2285 { // cwg2285: 4 void test() { using T = int[1]; auto [a] = T{a}; - // since-cxx17-error@-1 {{binding 'a' cannot appear in the initializer of its own decomposition declaration}} + // since-cxx17-error@-1 {{binding 'a' cannot appear in the initializer of its own structured binding declaration}} } #endif } // namespace cwg2285 diff --git a/clang/test/CXX/drs/cwg23xx.cpp b/clang/test/CXX/drs/cwg23xx.cpp index 1285665..72cf249 100644 --- a/clang/test/CXX/drs/cwg23xx.cpp +++ b/clang/test/CXX/drs/cwg23xx.cpp @@ -440,7 +440,7 @@ template <> struct tuple_size<cwg2386::Bad2> { namespace cwg2386 { void no_value() { auto [x, y] = Bad1(); } void wrong_value() { auto [x, y] = Bad2(); } -// since-cxx17-error@-1 {{type 'Bad2' decomposes into 42 elements, but only 2 names were provided}} +// since-cxx17-error@-1 {{type 'Bad2' binds to 42 elements, but only 2 names were provided}} #endif } // namespace cwg2386 diff --git a/clang/test/CXX/drs/cwg26xx.cpp b/clang/test/CXX/drs/cwg26xx.cpp index bceef64..aa87f5a 100644 --- a/clang/test/CXX/drs/cwg26xx.cpp +++ b/clang/test/CXX/drs/cwg26xx.cpp @@ -183,17 +183,17 @@ T get_T(); void use() { UnaryC auto [a, b] = get_S(); - // since-cxx20-error@-1 {{decomposition declaration cannot be declared with constrained 'auto'}} + // since-cxx20-error@-1 {{structured binding declaration cannot be declared with constrained 'auto'}} BinaryC<int> auto [c, d] = get_S(); - // since-cxx20-error@-1 {{decomposition declaration cannot be declared with constrained 'auto'}} + // since-cxx20-error@-1 {{structured binding declaration cannot be declared with constrained 'auto'}} } template<typename T> void TemplUse() { UnaryC auto [a, b] = get_T<T>(); - // since-cxx20-error@-1 {{decomposition declaration cannot be declared with constrained 'auto'}} + // since-cxx20-error@-1 {{structured binding declaration cannot be declared with constrained 'auto'}} BinaryC<T> auto [c, d] = get_T<T>(); - // since-cxx20-error@-1 {{decomposition declaration cannot be declared with constrained 'auto'}} + // since-cxx20-error@-1 {{structured binding declaration cannot be declared with constrained 'auto'}} } #endif } // namespace cwg2635 diff --git a/clang/test/CXX/module/module.import/p6.cpp b/clang/test/CXX/module/module.import/p6.cpp index cb2d799..9e378a5 100644 --- a/clang/test/CXX/module/module.import/p6.cpp +++ b/clang/test/CXX/module/module.import/p6.cpp @@ -3,6 +3,9 @@ // RUN: %clang_cc1 -std=c++20 -x c++-header %t/bad-header-unit.h \ // RUN: -emit-header-unit -o %t/bad-header-unit.pcm -verify +// RUN: %clang_cc1 -std=c++20 -x c++-header %t/bad-header-unit-declspec.h \ +// RUN: -emit-header-unit -o %t/bad-header-unit.pcm -verify \ +// RUN: -fdeclspec //--- bad-header-unit.h @@ -77,3 +80,13 @@ template <typename T> bool b() { } inline bool B = b<int>(); + +__attribute__((weak)) int weak_fun_definition() { return 42; } + +__attribute__((weak)) int weak_var_definition = 42; + +//--- bad-header-unit-declspec.h + +/* The cases below should compile without diagnostics. */ + +__declspec(selectany) int selectany_var_definition = 42; // expected-no-diagnostics diff --git a/clang/test/CXX/stmt.stmt/stmt.iter/stmt.ranged/p1.cpp b/clang/test/CXX/stmt.stmt/stmt.iter/stmt.ranged/p1.cpp index 93c2beb..938a3d0 100644 --- a/clang/test/CXX/stmt.stmt/stmt.iter/stmt.ranged/p1.cpp +++ b/clang/test/CXX/stmt.stmt/stmt.iter/stmt.ranged/p1.cpp @@ -75,7 +75,7 @@ namespace X { void test_D() { #if __cplusplus >= 201703L for (extern auto [x, y] : D()) { - } // expected-error@-1 {{decomposition declaration cannot be declared 'extern'}} + } // expected-error@-1 {{structured binding declaration cannot be declared 'extern'}} // expected-error@-2 {{loop variable '[x, y]' may not be declared 'extern'}} #endif } diff --git a/clang/test/CXX/temp/temp.res/temp.local/p6.cpp b/clang/test/CXX/temp/temp.res/temp.local/p6.cpp index 205df5f..beed7bf 100644 --- a/clang/test/CXX/temp/temp.res/temp.local/p6.cpp +++ b/clang/test/CXX/temp/temp.res/temp.local/p6.cpp @@ -165,7 +165,7 @@ A<0>::B a; template <typename T> int shadow() { // expected-note{{template parameter is declared here}} using arr = int[1]; - // expected-warning@+1 {{decomposition declarations are a C++17 extension}} + // expected-warning@+1 {{structured binding declarations are a C++17 extension}} auto [ T // expected-error {{declaration of 'T' shadows template parameter}} ] = arr{}; diff --git a/clang/test/CodeGen/AArch64/atomic-ops-float-check-minmax.c b/clang/test/CodeGen/AArch64/atomic-ops-float-check-minmax.c new file mode 100644 index 0000000..4d9b29b --- /dev/null +++ b/clang/test/CodeGen/AArch64/atomic-ops-float-check-minmax.c @@ -0,0 +1,117 @@ +// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 6 +// RUN: %clang_cc1 %s -emit-llvm -o - -ffreestanding -triple=aarch64-linux-gnu -pthread | FileCheck %s + +#include <stdint.h> +#include <stdatomic.h> + +// CHECK-LABEL: define dso_local void @test_minmax_postop( +// CHECK-SAME: ptr noundef [[F32:%.*]], ptr noundef [[F16:%.*]], ptr noundef [[BF16:%.*]], ptr noundef [[F64:%.*]]) #[[ATTR0:[0-9]+]] { +// CHECK-NEXT: [[ENTRY:.*:]] +// CHECK-NEXT: [[F32_ADDR:%.*]] = alloca ptr, align 8 +// CHECK-NEXT: [[F16_ADDR:%.*]] = alloca ptr, align 8 +// CHECK-NEXT: [[BF16_ADDR:%.*]] = alloca ptr, align 8 +// CHECK-NEXT: [[F64_ADDR:%.*]] = alloca ptr, align 8 +// CHECK-NEXT: [[DOTATOMICTMP:%.*]] = alloca double, align 8 +// CHECK-NEXT: [[ATOMIC_TEMP:%.*]] = alloca double, align 8 +// CHECK-NEXT: [[DOTATOMICTMP1:%.*]] = alloca float, align 4 +// CHECK-NEXT: [[ATOMIC_TEMP2:%.*]] = alloca float, align 4 +// CHECK-NEXT: [[DOTATOMICTMP4:%.*]] = alloca half, align 2 +// CHECK-NEXT: [[ATOMIC_TEMP5:%.*]] = alloca half, align 2 +// CHECK-NEXT: [[DOTATOMICTMP7:%.*]] = alloca bfloat, align 2 +// CHECK-NEXT: [[ATOMIC_TEMP8:%.*]] = alloca bfloat, align 2 +// CHECK-NEXT: [[DOTATOMICTMP10:%.*]] = alloca double, align 8 +// CHECK-NEXT: [[ATOMIC_TEMP11:%.*]] = alloca double, align 8 +// CHECK-NEXT: [[DOTATOMICTMP13:%.*]] = alloca float, align 4 +// CHECK-NEXT: [[ATOMIC_TEMP14:%.*]] = alloca float, align 4 +// CHECK-NEXT: [[DOTATOMICTMP16:%.*]] = alloca half, align 2 +// CHECK-NEXT: [[ATOMIC_TEMP17:%.*]] = alloca half, align 2 +// CHECK-NEXT: [[DOTATOMICTMP19:%.*]] = alloca bfloat, align 2 +// CHECK-NEXT: [[ATOMIC_TEMP20:%.*]] = alloca bfloat, align 2 +// CHECK-NEXT: store ptr [[F32]], ptr [[F32_ADDR]], align 8 +// CHECK-NEXT: store ptr [[F16]], ptr [[F16_ADDR]], align 8 +// CHECK-NEXT: store ptr [[BF16]], ptr [[BF16_ADDR]], align 8 +// CHECK-NEXT: store ptr [[F64]], ptr [[F64_ADDR]], align 8 +// CHECK-NEXT: [[TMP0:%.*]] = load ptr, ptr [[F64_ADDR]], align 8 +// CHECK-NEXT: store double 4.210000e+01, ptr [[DOTATOMICTMP]], align 8 +// CHECK-NEXT: [[TMP1:%.*]] = load double, ptr [[DOTATOMICTMP]], align 8 +// CHECK-NEXT: [[TMP2:%.*]] = atomicrmw fmax ptr [[TMP0]], double [[TMP1]] release, align 8 +// CHECK-NEXT: [[NEWVAL:%.*]] = call double @llvm.maxnum.f64(double [[TMP2]], double [[TMP1]]) +// CHECK-NEXT: store double [[NEWVAL]], ptr [[ATOMIC_TEMP]], align 8 +// CHECK-NEXT: [[TMP3:%.*]] = load double, ptr [[ATOMIC_TEMP]], align 8 +// CHECK-NEXT: [[TMP4:%.*]] = load ptr, ptr [[F64_ADDR]], align 8 +// CHECK-NEXT: store double [[TMP3]], ptr [[TMP4]], align 8 +// CHECK-NEXT: [[TMP5:%.*]] = load ptr, ptr [[F32_ADDR]], align 8 +// CHECK-NEXT: store float 0x40450CCCC0000000, ptr [[DOTATOMICTMP1]], align 4 +// CHECK-NEXT: [[TMP6:%.*]] = load float, ptr [[DOTATOMICTMP1]], align 4 +// CHECK-NEXT: [[TMP7:%.*]] = atomicrmw fmax ptr [[TMP5]], float [[TMP6]] release, align 4 +// CHECK-NEXT: [[NEWVAL3:%.*]] = call float @llvm.maxnum.f32(float [[TMP7]], float [[TMP6]]) +// CHECK-NEXT: store float [[NEWVAL3]], ptr [[ATOMIC_TEMP2]], align 4 +// CHECK-NEXT: [[TMP8:%.*]] = load float, ptr [[ATOMIC_TEMP2]], align 4 +// CHECK-NEXT: [[TMP9:%.*]] = load ptr, ptr [[F32_ADDR]], align 8 +// CHECK-NEXT: store float [[TMP8]], ptr [[TMP9]], align 4 +// CHECK-NEXT: [[TMP10:%.*]] = load ptr, ptr [[F16_ADDR]], align 8 +// CHECK-NEXT: store half 0xH5143, ptr [[DOTATOMICTMP4]], align 2 +// CHECK-NEXT: [[TMP11:%.*]] = load half, ptr [[DOTATOMICTMP4]], align 2 +// CHECK-NEXT: [[TMP12:%.*]] = atomicrmw fmax ptr [[TMP10]], half [[TMP11]] release, align 2 +// CHECK-NEXT: [[NEWVAL6:%.*]] = call half @llvm.maxnum.f16(half [[TMP12]], half [[TMP11]]) +// CHECK-NEXT: store half [[NEWVAL6]], ptr [[ATOMIC_TEMP5]], align 2 +// CHECK-NEXT: [[TMP13:%.*]] = load half, ptr [[ATOMIC_TEMP5]], align 2 +// CHECK-NEXT: [[TMP14:%.*]] = load ptr, ptr [[F16_ADDR]], align 8 +// CHECK-NEXT: store half [[TMP13]], ptr [[TMP14]], align 2 +// CHECK-NEXT: [[TMP15:%.*]] = load ptr, ptr [[BF16_ADDR]], align 8 +// CHECK-NEXT: store bfloat 0xR4228, ptr [[DOTATOMICTMP7]], align 2 +// CHECK-NEXT: [[TMP16:%.*]] = load bfloat, ptr [[DOTATOMICTMP7]], align 2 +// CHECK-NEXT: [[TMP17:%.*]] = atomicrmw fmax ptr [[TMP15]], bfloat [[TMP16]] release, align 2 +// CHECK-NEXT: [[NEWVAL9:%.*]] = call bfloat @llvm.maxnum.bf16(bfloat [[TMP17]], bfloat [[TMP16]]) +// CHECK-NEXT: store bfloat [[NEWVAL9]], ptr [[ATOMIC_TEMP8]], align 2 +// CHECK-NEXT: [[TMP18:%.*]] = load bfloat, ptr [[ATOMIC_TEMP8]], align 2 +// CHECK-NEXT: [[TMP19:%.*]] = load ptr, ptr [[BF16_ADDR]], align 8 +// CHECK-NEXT: store bfloat [[TMP18]], ptr [[TMP19]], align 2 +// CHECK-NEXT: [[TMP20:%.*]] = load ptr, ptr [[F64_ADDR]], align 8 +// CHECK-NEXT: store double 4.210000e+01, ptr [[DOTATOMICTMP10]], align 8 +// CHECK-NEXT: [[TMP21:%.*]] = load double, ptr [[DOTATOMICTMP10]], align 8 +// CHECK-NEXT: [[TMP22:%.*]] = atomicrmw fmin ptr [[TMP20]], double [[TMP21]] release, align 8 +// CHECK-NEXT: [[NEWVAL12:%.*]] = call double @llvm.minnum.f64(double [[TMP22]], double [[TMP21]]) +// CHECK-NEXT: store double [[NEWVAL12]], ptr [[ATOMIC_TEMP11]], align 8 +// CHECK-NEXT: [[TMP23:%.*]] = load double, ptr [[ATOMIC_TEMP11]], align 8 +// CHECK-NEXT: [[TMP24:%.*]] = load ptr, ptr [[F64_ADDR]], align 8 +// CHECK-NEXT: store double [[TMP23]], ptr [[TMP24]], align 8 +// CHECK-NEXT: [[TMP25:%.*]] = load ptr, ptr [[F32_ADDR]], align 8 +// CHECK-NEXT: store float 0x40450CCCC0000000, ptr [[DOTATOMICTMP13]], align 4 +// CHECK-NEXT: [[TMP26:%.*]] = load float, ptr [[DOTATOMICTMP13]], align 4 +// CHECK-NEXT: [[TMP27:%.*]] = atomicrmw fmin ptr [[TMP25]], float [[TMP26]] release, align 4 +// CHECK-NEXT: [[NEWVAL15:%.*]] = call float @llvm.minnum.f32(float [[TMP27]], float [[TMP26]]) +// CHECK-NEXT: store float [[NEWVAL15]], ptr [[ATOMIC_TEMP14]], align 4 +// CHECK-NEXT: [[TMP28:%.*]] = load float, ptr [[ATOMIC_TEMP14]], align 4 +// CHECK-NEXT: [[TMP29:%.*]] = load ptr, ptr [[F32_ADDR]], align 8 +// CHECK-NEXT: store float [[TMP28]], ptr [[TMP29]], align 4 +// CHECK-NEXT: [[TMP30:%.*]] = load ptr, ptr [[F16_ADDR]], align 8 +// CHECK-NEXT: store half 0xH5143, ptr [[DOTATOMICTMP16]], align 2 +// CHECK-NEXT: [[TMP31:%.*]] = load half, ptr [[DOTATOMICTMP16]], align 2 +// CHECK-NEXT: [[TMP32:%.*]] = atomicrmw fmin ptr [[TMP30]], half [[TMP31]] release, align 2 +// CHECK-NEXT: [[NEWVAL18:%.*]] = call half @llvm.minnum.f16(half [[TMP32]], half [[TMP31]]) +// CHECK-NEXT: store half [[NEWVAL18]], ptr [[ATOMIC_TEMP17]], align 2 +// CHECK-NEXT: [[TMP33:%.*]] = load half, ptr [[ATOMIC_TEMP17]], align 2 +// CHECK-NEXT: [[TMP34:%.*]] = load ptr, ptr [[F16_ADDR]], align 8 +// CHECK-NEXT: store half [[TMP33]], ptr [[TMP34]], align 2 +// CHECK-NEXT: [[TMP35:%.*]] = load ptr, ptr [[BF16_ADDR]], align 8 +// CHECK-NEXT: store bfloat 0xR4228, ptr [[DOTATOMICTMP19]], align 2 +// CHECK-NEXT: [[TMP36:%.*]] = load bfloat, ptr [[DOTATOMICTMP19]], align 2 +// CHECK-NEXT: [[TMP37:%.*]] = atomicrmw fmin ptr [[TMP35]], bfloat [[TMP36]] release, align 2 +// CHECK-NEXT: [[NEWVAL21:%.*]] = call bfloat @llvm.minnum.bf16(bfloat [[TMP37]], bfloat [[TMP36]]) +// CHECK-NEXT: store bfloat [[NEWVAL21]], ptr [[ATOMIC_TEMP20]], align 2 +// CHECK-NEXT: [[TMP38:%.*]] = load bfloat, ptr [[ATOMIC_TEMP20]], align 2 +// CHECK-NEXT: [[TMP39:%.*]] = load ptr, ptr [[BF16_ADDR]], align 8 +// CHECK-NEXT: store bfloat [[TMP38]], ptr [[TMP39]], align 2 +// CHECK-NEXT: ret void +// +void test_minmax_postop(float *f32, _Float16 *f16, __bf16 *bf16, double *f64) { + *f64 = __atomic_max_fetch(f64, 42.1, memory_order_release); + *f32 = __atomic_max_fetch(f32, 42.1, memory_order_release); + *f16 = __atomic_max_fetch(f16, 42.1, memory_order_release); + *bf16 = __atomic_max_fetch(bf16, 42.1, memory_order_release); + *f64 = __atomic_min_fetch(f64, 42.1, memory_order_release); + *f32 = __atomic_min_fetch(f32, 42.1, memory_order_release); + *f16 = __atomic_min_fetch(f16, 42.1, memory_order_release); + *bf16 = __atomic_min_fetch(bf16, 42.1, memory_order_release); +} diff --git a/clang/test/CodeGen/RISCV/bitint.c b/clang/test/CodeGen/RISCV/bitint.c new file mode 100644 index 0000000..1ad43af --- /dev/null +++ b/clang/test/CodeGen/RISCV/bitint.c @@ -0,0 +1,342 @@ +// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --function-signature +// RUN: %clang_cc1 -triple riscv64 -O2 -emit-llvm -o - %s | FileCheck %s --check-prefix=RISCV64 +// RUN: %clang_cc1 -triple riscv32 -O2 -emit-llvm -o - %s | FileCheck %s --check-prefix=RISCV32 +// RUN: %clang_cc1 -triple riscv32 -fforce-enable-int128 -O2 -emit-llvm -o - %s | FileCheck %s --check-prefix=RISCV32_INT128 + +// RISCV64-LABEL: define {{[^@]+}}@test_bitint_17_add_unsigned +// RISCV64-SAME: (i17 noundef zeroext [[A:%.*]], i17 noundef zeroext [[B:%.*]]) local_unnamed_addr #[[ATTR0:[0-9]+]] { +// RISCV64-NEXT: entry: +// RISCV64-NEXT: [[ADD:%.*]] = add i17 [[B]], [[A]] +// RISCV64-NEXT: ret i17 [[ADD]] +// +// RISCV32-LABEL: define {{[^@]+}}@test_bitint_17_add_unsigned +// RISCV32-SAME: (i17 noundef zeroext [[A:%.*]], i17 noundef zeroext [[B:%.*]]) local_unnamed_addr #[[ATTR0:[0-9]+]] { +// RISCV32-NEXT: entry: +// RISCV32-NEXT: [[ADD:%.*]] = add i17 [[B]], [[A]] +// RISCV32-NEXT: ret i17 [[ADD]] +// +// RISCV32_INT128-LABEL: define {{[^@]+}}@test_bitint_17_add_unsigned +// RISCV32_INT128-SAME: (i17 noundef zeroext [[A:%.*]], i17 noundef zeroext [[B:%.*]]) local_unnamed_addr #[[ATTR0:[0-9]+]] { +// RISCV32_INT128-NEXT: entry: +// RISCV32_INT128-NEXT: [[ADD:%.*]] = add i17 [[B]], [[A]] +// RISCV32_INT128-NEXT: ret i17 [[ADD]] +// +unsigned _BitInt(17) test_bitint_17_add_unsigned(unsigned _BitInt(17) a, unsigned _BitInt(17) b) { + return a + b; +} + +// RISCV64-LABEL: define {{[^@]+}}@test_bitint_17_add_signed +// RISCV64-SAME: (i17 noundef signext [[A:%.*]], i17 noundef signext [[B:%.*]]) local_unnamed_addr #[[ATTR0]] { +// RISCV64-NEXT: entry: +// RISCV64-NEXT: [[ADD:%.*]] = add nsw i17 [[B]], [[A]] +// RISCV64-NEXT: ret i17 [[ADD]] +// +// RISCV32-LABEL: define {{[^@]+}}@test_bitint_17_add_signed +// RISCV32-SAME: (i17 noundef signext [[A:%.*]], i17 noundef signext [[B:%.*]]) local_unnamed_addr #[[ATTR0]] { +// RISCV32-NEXT: entry: +// RISCV32-NEXT: [[ADD:%.*]] = add nsw i17 [[B]], [[A]] +// RISCV32-NEXT: ret i17 [[ADD]] +// +// RISCV32_INT128-LABEL: define {{[^@]+}}@test_bitint_17_add_signed +// RISCV32_INT128-SAME: (i17 noundef signext [[A:%.*]], i17 noundef signext [[B:%.*]]) local_unnamed_addr #[[ATTR0]] { +// RISCV32_INT128-NEXT: entry: +// RISCV32_INT128-NEXT: [[ADD:%.*]] = add nsw i17 [[B]], [[A]] +// RISCV32_INT128-NEXT: ret i17 [[ADD]] +// +signed _BitInt(17) test_bitint_17_add_signed(signed _BitInt(17) a, signed _BitInt(17) b) { + return a + b; +} + +// RISCV64-LABEL: define {{[^@]+}}@test_bitint_17_add_default +// RISCV64-SAME: (i17 noundef signext [[A:%.*]], i17 noundef signext [[B:%.*]]) local_unnamed_addr #[[ATTR0]] { +// RISCV64-NEXT: entry: +// RISCV64-NEXT: [[ADD:%.*]] = add nsw i17 [[B]], [[A]] +// RISCV64-NEXT: ret i17 [[ADD]] +// +// RISCV32-LABEL: define {{[^@]+}}@test_bitint_17_add_default +// RISCV32-SAME: (i17 noundef signext [[A:%.*]], i17 noundef signext [[B:%.*]]) local_unnamed_addr #[[ATTR0]] { +// RISCV32-NEXT: entry: +// RISCV32-NEXT: [[ADD:%.*]] = add nsw i17 [[B]], [[A]] +// RISCV32-NEXT: ret i17 [[ADD]] +// +// RISCV32_INT128-LABEL: define {{[^@]+}}@test_bitint_17_add_default +// RISCV32_INT128-SAME: (i17 noundef signext [[A:%.*]], i17 noundef signext [[B:%.*]]) local_unnamed_addr #[[ATTR0]] { +// RISCV32_INT128-NEXT: entry: +// RISCV32_INT128-NEXT: [[ADD:%.*]] = add nsw i17 [[B]], [[A]] +// RISCV32_INT128-NEXT: ret i17 [[ADD]] +// +_BitInt(17) test_bitint_17_add_default(_BitInt(17) a, _BitInt(17) b) { + return a + b; +} + +// RISCV64-LABEL: define {{[^@]+}}@test_bitint_32_add_unsigned +// RISCV64-SAME: (i32 noundef signext [[A:%.*]], i32 noundef signext [[B:%.*]]) local_unnamed_addr #[[ATTR0]] { +// RISCV64-NEXT: entry: +// RISCV64-NEXT: [[ADD:%.*]] = add i32 [[B]], [[A]] +// RISCV64-NEXT: ret i32 [[ADD]] +// +// RISCV32-LABEL: define {{[^@]+}}@test_bitint_32_add_unsigned +// RISCV32-SAME: (i32 noundef zeroext [[A:%.*]], i32 noundef zeroext [[B:%.*]]) local_unnamed_addr #[[ATTR0]] { +// RISCV32-NEXT: entry: +// RISCV32-NEXT: [[ADD:%.*]] = add i32 [[B]], [[A]] +// RISCV32-NEXT: ret i32 [[ADD]] +// +// RISCV32_INT128-LABEL: define {{[^@]+}}@test_bitint_32_add_unsigned +// RISCV32_INT128-SAME: (i32 noundef zeroext [[A:%.*]], i32 noundef zeroext [[B:%.*]]) local_unnamed_addr #[[ATTR0]] { +// RISCV32_INT128-NEXT: entry: +// RISCV32_INT128-NEXT: [[ADD:%.*]] = add i32 [[B]], [[A]] +// RISCV32_INT128-NEXT: ret i32 [[ADD]] +// +unsigned _BitInt(32) test_bitint_32_add_unsigned(unsigned _BitInt(32) a, unsigned _BitInt(32) b) { + return a + b; +} + +// RISCV64-LABEL: define {{[^@]+}}@test_bitint_32_add_signed +// RISCV64-SAME: (i32 noundef signext [[A:%.*]], i32 noundef signext [[B:%.*]]) local_unnamed_addr #[[ATTR0]] { +// RISCV64-NEXT: entry: +// RISCV64-NEXT: [[ADD:%.*]] = add nsw i32 [[B]], [[A]] +// RISCV64-NEXT: ret i32 [[ADD]] +// +// RISCV32-LABEL: define {{[^@]+}}@test_bitint_32_add_signed +// RISCV32-SAME: (i32 noundef signext [[A:%.*]], i32 noundef signext [[B:%.*]]) local_unnamed_addr #[[ATTR0]] { +// RISCV32-NEXT: entry: +// RISCV32-NEXT: [[ADD:%.*]] = add nsw i32 [[B]], [[A]] +// RISCV32-NEXT: ret i32 [[ADD]] +// +// RISCV32_INT128-LABEL: define {{[^@]+}}@test_bitint_32_add_signed +// RISCV32_INT128-SAME: (i32 noundef signext [[A:%.*]], i32 noundef signext [[B:%.*]]) local_unnamed_addr #[[ATTR0]] { +// RISCV32_INT128-NEXT: entry: +// RISCV32_INT128-NEXT: [[ADD:%.*]] = add nsw i32 [[B]], [[A]] +// RISCV32_INT128-NEXT: ret i32 [[ADD]] +// +signed _BitInt(32) test_bitint_32_add_signed(signed _BitInt(32) a, signed _BitInt(32) b) { + return a + b; +} + +// RISCV64-LABEL: define {{[^@]+}}@test_bitint_32_add_default +// RISCV64-SAME: (i32 noundef signext [[A:%.*]], i32 noundef signext [[B:%.*]]) local_unnamed_addr #[[ATTR0]] { +// RISCV64-NEXT: entry: +// RISCV64-NEXT: [[ADD:%.*]] = add nsw i32 [[B]], [[A]] +// RISCV64-NEXT: ret i32 [[ADD]] +// +// RISCV32-LABEL: define {{[^@]+}}@test_bitint_32_add_default +// RISCV32-SAME: (i32 noundef signext [[A:%.*]], i32 noundef signext [[B:%.*]]) local_unnamed_addr #[[ATTR0]] { +// RISCV32-NEXT: entry: +// RISCV32-NEXT: [[ADD:%.*]] = add nsw i32 [[B]], [[A]] +// RISCV32-NEXT: ret i32 [[ADD]] +// +// RISCV32_INT128-LABEL: define {{[^@]+}}@test_bitint_32_add_default +// RISCV32_INT128-SAME: (i32 noundef signext [[A:%.*]], i32 noundef signext [[B:%.*]]) local_unnamed_addr #[[ATTR0]] { +// RISCV32_INT128-NEXT: entry: +// RISCV32_INT128-NEXT: [[ADD:%.*]] = add nsw i32 [[B]], [[A]] +// RISCV32_INT128-NEXT: ret i32 [[ADD]] +// +_BitInt(32) test_bitint_32_add_default(_BitInt(32) a, _BitInt(32) b) { + return a + b; +} + + +// RISCV64-LABEL: define {{[^@]+}}@test_bitint_65_add_unsigned +// RISCV64-SAME: (i65 noundef zeroext [[A:%.*]], i65 noundef zeroext [[B:%.*]]) local_unnamed_addr #[[ATTR0]] { +// RISCV64-NEXT: entry: +// RISCV64-NEXT: [[ADD:%.*]] = add i65 [[B]], [[A]] +// RISCV64-NEXT: ret i65 [[ADD]] +// +// RISCV32-LABEL: define {{[^@]+}}@test_bitint_65_add_unsigned +// RISCV32-SAME: (ptr dead_on_unwind noalias writable writeonly sret(i128) align 8 captures(none) initializes((0, 16)) [[AGG_RESULT:%.*]], ptr dead_on_return noundef readonly captures(none) [[TMP0:%.*]], ptr dead_on_return noundef readonly captures(none) [[TMP1:%.*]]) local_unnamed_addr #[[ATTR1:[0-9]+]] { +// RISCV32-NEXT: entry: +// RISCV32-NEXT: [[TMP2:%.*]] = load i128, ptr [[TMP0]], align 8, !tbaa [[TBAA6:![0-9]+]] +// RISCV32-NEXT: [[A:%.*]] = trunc i128 [[TMP2]] to i65 +// RISCV32-NEXT: [[TMP3:%.*]] = load i128, ptr [[TMP1]], align 8, !tbaa [[TBAA6]] +// RISCV32-NEXT: [[B:%.*]] = trunc i128 [[TMP3]] to i65 +// RISCV32-NEXT: [[ADD:%.*]] = add i65 [[B]], [[A]] +// RISCV32-NEXT: [[STOREDV4:%.*]] = zext i65 [[ADD]] to i128 +// RISCV32-NEXT: store i128 [[STOREDV4]], ptr [[AGG_RESULT]], align 8, !tbaa [[TBAA6]] +// RISCV32-NEXT: ret void +// +// RISCV32_INT128-LABEL: define {{[^@]+}}@test_bitint_65_add_unsigned +// RISCV32_INT128-SAME: (ptr dead_on_unwind noalias writable writeonly sret(i128) align 8 captures(none) initializes((0, 16)) [[AGG_RESULT:%.*]], ptr dead_on_return noundef readonly captures(none) [[TMP0:%.*]], ptr dead_on_return noundef readonly captures(none) [[TMP1:%.*]]) local_unnamed_addr #[[ATTR1:[0-9]+]] { +// RISCV32_INT128-NEXT: entry: +// RISCV32_INT128-NEXT: [[TMP2:%.*]] = load i128, ptr [[TMP0]], align 8, !tbaa [[TBAA6:![0-9]+]] +// RISCV32_INT128-NEXT: [[A:%.*]] = trunc i128 [[TMP2]] to i65 +// RISCV32_INT128-NEXT: [[TMP3:%.*]] = load i128, ptr [[TMP1]], align 8, !tbaa [[TBAA6]] +// RISCV32_INT128-NEXT: [[B:%.*]] = trunc i128 [[TMP3]] to i65 +// RISCV32_INT128-NEXT: [[ADD:%.*]] = add i65 [[B]], [[A]] +// RISCV32_INT128-NEXT: [[STOREDV4:%.*]] = zext i65 [[ADD]] to i128 +// RISCV32_INT128-NEXT: store i128 [[STOREDV4]], ptr [[AGG_RESULT]], align 8, !tbaa [[TBAA6]] +// RISCV32_INT128-NEXT: ret void +// +unsigned _BitInt(65) test_bitint_65_add_unsigned(unsigned _BitInt(65) a, unsigned _BitInt(65) b) { + return a + b; +} + +// RISCV64-LABEL: define {{[^@]+}}@test_bitint_65_add_signed +// RISCV64-SAME: (i65 noundef signext [[A:%.*]], i65 noundef signext [[B:%.*]]) local_unnamed_addr #[[ATTR0]] { +// RISCV64-NEXT: entry: +// RISCV64-NEXT: [[ADD:%.*]] = add nsw i65 [[B]], [[A]] +// RISCV64-NEXT: ret i65 [[ADD]] +// +// RISCV32-LABEL: define {{[^@]+}}@test_bitint_65_add_signed +// RISCV32-SAME: (ptr dead_on_unwind noalias writable writeonly sret(i128) align 8 captures(none) initializes((0, 16)) [[AGG_RESULT:%.*]], ptr dead_on_return noundef readonly captures(none) [[TMP0:%.*]], ptr dead_on_return noundef readonly captures(none) [[TMP1:%.*]]) local_unnamed_addr #[[ATTR1]] { +// RISCV32-NEXT: entry: +// RISCV32-NEXT: [[TMP2:%.*]] = load i128, ptr [[TMP0]], align 8, !tbaa [[TBAA6]] +// RISCV32-NEXT: [[A:%.*]] = trunc i128 [[TMP2]] to i65 +// RISCV32-NEXT: [[TMP3:%.*]] = load i128, ptr [[TMP1]], align 8, !tbaa [[TBAA6]] +// RISCV32-NEXT: [[B:%.*]] = trunc i128 [[TMP3]] to i65 +// RISCV32-NEXT: [[ADD:%.*]] = add nsw i65 [[B]], [[A]] +// RISCV32-NEXT: [[STOREDV4:%.*]] = sext i65 [[ADD]] to i128 +// RISCV32-NEXT: store i128 [[STOREDV4]], ptr [[AGG_RESULT]], align 8, !tbaa [[TBAA6]] +// RISCV32-NEXT: ret void +// +// RISCV32_INT128-LABEL: define {{[^@]+}}@test_bitint_65_add_signed +// RISCV32_INT128-SAME: (ptr dead_on_unwind noalias writable writeonly sret(i128) align 8 captures(none) initializes((0, 16)) [[AGG_RESULT:%.*]], ptr dead_on_return noundef readonly captures(none) [[TMP0:%.*]], ptr dead_on_return noundef readonly captures(none) [[TMP1:%.*]]) local_unnamed_addr #[[ATTR1]] { +// RISCV32_INT128-NEXT: entry: +// RISCV32_INT128-NEXT: [[TMP2:%.*]] = load i128, ptr [[TMP0]], align 8, !tbaa [[TBAA6]] +// RISCV32_INT128-NEXT: [[A:%.*]] = trunc i128 [[TMP2]] to i65 +// RISCV32_INT128-NEXT: [[TMP3:%.*]] = load i128, ptr [[TMP1]], align 8, !tbaa [[TBAA6]] +// RISCV32_INT128-NEXT: [[B:%.*]] = trunc i128 [[TMP3]] to i65 +// RISCV32_INT128-NEXT: [[ADD:%.*]] = add nsw i65 [[B]], [[A]] +// RISCV32_INT128-NEXT: [[STOREDV4:%.*]] = sext i65 [[ADD]] to i128 +// RISCV32_INT128-NEXT: store i128 [[STOREDV4]], ptr [[AGG_RESULT]], align 8, !tbaa [[TBAA6]] +// RISCV32_INT128-NEXT: ret void +// +signed _BitInt(65) test_bitint_65_add_signed(signed _BitInt(65) a, signed _BitInt(65) b) { + return a + b; +} + +// RISCV64-LABEL: define {{[^@]+}}@test_bitint_65_add_default +// RISCV64-SAME: (i65 noundef signext [[A:%.*]], i65 noundef signext [[B:%.*]]) local_unnamed_addr #[[ATTR0]] { +// RISCV64-NEXT: entry: +// RISCV64-NEXT: [[ADD:%.*]] = add nsw i65 [[B]], [[A]] +// RISCV64-NEXT: ret i65 [[ADD]] +// +// RISCV32-LABEL: define {{[^@]+}}@test_bitint_65_add_default +// RISCV32-SAME: (ptr dead_on_unwind noalias writable writeonly sret(i128) align 8 captures(none) initializes((0, 16)) [[AGG_RESULT:%.*]], ptr dead_on_return noundef readonly captures(none) [[TMP0:%.*]], ptr dead_on_return noundef readonly captures(none) [[TMP1:%.*]]) local_unnamed_addr #[[ATTR1]] { +// RISCV32-NEXT: entry: +// RISCV32-NEXT: [[TMP2:%.*]] = load i128, ptr [[TMP0]], align 8, !tbaa [[TBAA6]] +// RISCV32-NEXT: [[A:%.*]] = trunc i128 [[TMP2]] to i65 +// RISCV32-NEXT: [[TMP3:%.*]] = load i128, ptr [[TMP1]], align 8, !tbaa [[TBAA6]] +// RISCV32-NEXT: [[B:%.*]] = trunc i128 [[TMP3]] to i65 +// RISCV32-NEXT: [[ADD:%.*]] = add nsw i65 [[B]], [[A]] +// RISCV32-NEXT: [[STOREDV4:%.*]] = sext i65 [[ADD]] to i128 +// RISCV32-NEXT: store i128 [[STOREDV4]], ptr [[AGG_RESULT]], align 8, !tbaa [[TBAA6]] +// RISCV32-NEXT: ret void +// +// RISCV32_INT128-LABEL: define {{[^@]+}}@test_bitint_65_add_default +// RISCV32_INT128-SAME: (ptr dead_on_unwind noalias writable writeonly sret(i128) align 8 captures(none) initializes((0, 16)) [[AGG_RESULT:%.*]], ptr dead_on_return noundef readonly captures(none) [[TMP0:%.*]], ptr dead_on_return noundef readonly captures(none) [[TMP1:%.*]]) local_unnamed_addr #[[ATTR1]] { +// RISCV32_INT128-NEXT: entry: +// RISCV32_INT128-NEXT: [[TMP2:%.*]] = load i128, ptr [[TMP0]], align 8, !tbaa [[TBAA6]] +// RISCV32_INT128-NEXT: [[A:%.*]] = trunc i128 [[TMP2]] to i65 +// RISCV32_INT128-NEXT: [[TMP3:%.*]] = load i128, ptr [[TMP1]], align 8, !tbaa [[TBAA6]] +// RISCV32_INT128-NEXT: [[B:%.*]] = trunc i128 [[TMP3]] to i65 +// RISCV32_INT128-NEXT: [[ADD:%.*]] = add nsw i65 [[B]], [[A]] +// RISCV32_INT128-NEXT: [[STOREDV4:%.*]] = sext i65 [[ADD]] to i128 +// RISCV32_INT128-NEXT: store i128 [[STOREDV4]], ptr [[AGG_RESULT]], align 8, !tbaa [[TBAA6]] +// RISCV32_INT128-NEXT: ret void +// +_BitInt(65) test_bitint_65_add_default(_BitInt(65) a, _BitInt(65) b) { + return a + b; +} + + +// RISCV64-LABEL: define {{[^@]+}}@test_bitint_77_add_unsigned +// RISCV64-SAME: (i77 noundef zeroext [[A:%.*]], i77 noundef zeroext [[B:%.*]]) local_unnamed_addr #[[ATTR0]] { +// RISCV64-NEXT: entry: +// RISCV64-NEXT: [[ADD:%.*]] = add i77 [[B]], [[A]] +// RISCV64-NEXT: ret i77 [[ADD]] +// +// RISCV32-LABEL: define {{[^@]+}}@test_bitint_77_add_unsigned +// RISCV32-SAME: (ptr dead_on_unwind noalias writable writeonly sret(i128) align 8 captures(none) initializes((0, 16)) [[AGG_RESULT:%.*]], ptr dead_on_return noundef readonly captures(none) [[TMP0:%.*]], ptr dead_on_return noundef readonly captures(none) [[TMP1:%.*]]) local_unnamed_addr #[[ATTR1]] { +// RISCV32-NEXT: entry: +// RISCV32-NEXT: [[TMP2:%.*]] = load i128, ptr [[TMP0]], align 8, !tbaa [[TBAA10:![0-9]+]] +// RISCV32-NEXT: [[A:%.*]] = trunc i128 [[TMP2]] to i77 +// RISCV32-NEXT: [[TMP3:%.*]] = load i128, ptr [[TMP1]], align 8, !tbaa [[TBAA10]] +// RISCV32-NEXT: [[B:%.*]] = trunc i128 [[TMP3]] to i77 +// RISCV32-NEXT: [[ADD:%.*]] = add i77 [[B]], [[A]] +// RISCV32-NEXT: [[STOREDV4:%.*]] = zext i77 [[ADD]] to i128 +// RISCV32-NEXT: store i128 [[STOREDV4]], ptr [[AGG_RESULT]], align 8, !tbaa [[TBAA10]] +// RISCV32-NEXT: ret void +// +// RISCV32_INT128-LABEL: define {{[^@]+}}@test_bitint_77_add_unsigned +// RISCV32_INT128-SAME: (ptr dead_on_unwind noalias writable writeonly sret(i128) align 8 captures(none) initializes((0, 16)) [[AGG_RESULT:%.*]], ptr dead_on_return noundef readonly captures(none) [[TMP0:%.*]], ptr dead_on_return noundef readonly captures(none) [[TMP1:%.*]]) local_unnamed_addr #[[ATTR1]] { +// RISCV32_INT128-NEXT: entry: +// RISCV32_INT128-NEXT: [[TMP2:%.*]] = load i128, ptr [[TMP0]], align 8, !tbaa [[TBAA10:![0-9]+]] +// RISCV32_INT128-NEXT: [[A:%.*]] = trunc i128 [[TMP2]] to i77 +// RISCV32_INT128-NEXT: [[TMP3:%.*]] = load i128, ptr [[TMP1]], align 8, !tbaa [[TBAA10]] +// RISCV32_INT128-NEXT: [[B:%.*]] = trunc i128 [[TMP3]] to i77 +// RISCV32_INT128-NEXT: [[ADD:%.*]] = add i77 [[B]], [[A]] +// RISCV32_INT128-NEXT: [[STOREDV4:%.*]] = zext i77 [[ADD]] to i128 +// RISCV32_INT128-NEXT: store i128 [[STOREDV4]], ptr [[AGG_RESULT]], align 8, !tbaa [[TBAA10]] +// RISCV32_INT128-NEXT: ret void +// +unsigned _BitInt(77) test_bitint_77_add_unsigned(unsigned _BitInt(77) a, unsigned _BitInt(77) b) { + return a + b; +} + +// RISCV64-LABEL: define {{[^@]+}}@test_bitint_77_add_signed +// RISCV64-SAME: (i77 noundef signext [[A:%.*]], i77 noundef signext [[B:%.*]]) local_unnamed_addr #[[ATTR0]] { +// RISCV64-NEXT: entry: +// RISCV64-NEXT: [[ADD:%.*]] = add nsw i77 [[B]], [[A]] +// RISCV64-NEXT: ret i77 [[ADD]] +// +// RISCV32-LABEL: define {{[^@]+}}@test_bitint_77_add_signed +// RISCV32-SAME: (ptr dead_on_unwind noalias writable writeonly sret(i128) align 8 captures(none) initializes((0, 16)) [[AGG_RESULT:%.*]], ptr dead_on_return noundef readonly captures(none) [[TMP0:%.*]], ptr dead_on_return noundef readonly captures(none) [[TMP1:%.*]]) local_unnamed_addr #[[ATTR1]] { +// RISCV32-NEXT: entry: +// RISCV32-NEXT: [[TMP2:%.*]] = load i128, ptr [[TMP0]], align 8, !tbaa [[TBAA10]] +// RISCV32-NEXT: [[A:%.*]] = trunc i128 [[TMP2]] to i77 +// RISCV32-NEXT: [[TMP3:%.*]] = load i128, ptr [[TMP1]], align 8, !tbaa [[TBAA10]] +// RISCV32-NEXT: [[B:%.*]] = trunc i128 [[TMP3]] to i77 +// RISCV32-NEXT: [[ADD:%.*]] = add nsw i77 [[B]], [[A]] +// RISCV32-NEXT: [[STOREDV4:%.*]] = sext i77 [[ADD]] to i128 +// RISCV32-NEXT: store i128 [[STOREDV4]], ptr [[AGG_RESULT]], align 8, !tbaa [[TBAA10]] +// RISCV32-NEXT: ret void +// +// RISCV32_INT128-LABEL: define {{[^@]+}}@test_bitint_77_add_signed +// RISCV32_INT128-SAME: (ptr dead_on_unwind noalias writable writeonly sret(i128) align 8 captures(none) initializes((0, 16)) [[AGG_RESULT:%.*]], ptr dead_on_return noundef readonly captures(none) [[TMP0:%.*]], ptr dead_on_return noundef readonly captures(none) [[TMP1:%.*]]) local_unnamed_addr #[[ATTR1]] { +// RISCV32_INT128-NEXT: entry: +// RISCV32_INT128-NEXT: [[TMP2:%.*]] = load i128, ptr [[TMP0]], align 8, !tbaa [[TBAA10]] +// RISCV32_INT128-NEXT: [[A:%.*]] = trunc i128 [[TMP2]] to i77 +// RISCV32_INT128-NEXT: [[TMP3:%.*]] = load i128, ptr [[TMP1]], align 8, !tbaa [[TBAA10]] +// RISCV32_INT128-NEXT: [[B:%.*]] = trunc i128 [[TMP3]] to i77 +// RISCV32_INT128-NEXT: [[ADD:%.*]] = add nsw i77 [[B]], [[A]] +// RISCV32_INT128-NEXT: [[STOREDV4:%.*]] = sext i77 [[ADD]] to i128 +// RISCV32_INT128-NEXT: store i128 [[STOREDV4]], ptr [[AGG_RESULT]], align 8, !tbaa [[TBAA10]] +// RISCV32_INT128-NEXT: ret void +// +signed _BitInt(77) test_bitint_77_add_signed(signed _BitInt(77) a, signed _BitInt(77) b) { + return a + b; +} + +// RISCV64-LABEL: define {{[^@]+}}@test_bitint_77_add_default +// RISCV64-SAME: (i77 noundef signext [[A:%.*]], i77 noundef signext [[B:%.*]]) local_unnamed_addr #[[ATTR0]] { +// RISCV64-NEXT: entry: +// RISCV64-NEXT: [[ADD:%.*]] = add nsw i77 [[B]], [[A]] +// RISCV64-NEXT: ret i77 [[ADD]] +// +// RISCV32-LABEL: define {{[^@]+}}@test_bitint_77_add_default +// RISCV32-SAME: (ptr dead_on_unwind noalias writable writeonly sret(i128) align 8 captures(none) initializes((0, 16)) [[AGG_RESULT:%.*]], ptr dead_on_return noundef readonly captures(none) [[TMP0:%.*]], ptr dead_on_return noundef readonly captures(none) [[TMP1:%.*]]) local_unnamed_addr #[[ATTR1]] { +// RISCV32-NEXT: entry: +// RISCV32-NEXT: [[TMP2:%.*]] = load i128, ptr [[TMP0]], align 8, !tbaa [[TBAA10]] +// RISCV32-NEXT: [[A:%.*]] = trunc i128 [[TMP2]] to i77 +// RISCV32-NEXT: [[TMP3:%.*]] = load i128, ptr [[TMP1]], align 8, !tbaa [[TBAA10]] +// RISCV32-NEXT: [[B:%.*]] = trunc i128 [[TMP3]] to i77 +// RISCV32-NEXT: [[ADD:%.*]] = add nsw i77 [[B]], [[A]] +// RISCV32-NEXT: [[STOREDV4:%.*]] = sext i77 [[ADD]] to i128 +// RISCV32-NEXT: store i128 [[STOREDV4]], ptr [[AGG_RESULT]], align 8, !tbaa [[TBAA10]] +// RISCV32-NEXT: ret void +// +// RISCV32_INT128-LABEL: define {{[^@]+}}@test_bitint_77_add_default +// RISCV32_INT128-SAME: (ptr dead_on_unwind noalias writable writeonly sret(i128) align 8 captures(none) initializes((0, 16)) [[AGG_RESULT:%.*]], ptr dead_on_return noundef readonly captures(none) [[TMP0:%.*]], ptr dead_on_return noundef readonly captures(none) [[TMP1:%.*]]) local_unnamed_addr #[[ATTR1]] { +// RISCV32_INT128-NEXT: entry: +// RISCV32_INT128-NEXT: [[TMP2:%.*]] = load i128, ptr [[TMP0]], align 8, !tbaa [[TBAA10]] +// RISCV32_INT128-NEXT: [[A:%.*]] = trunc i128 [[TMP2]] to i77 +// RISCV32_INT128-NEXT: [[TMP3:%.*]] = load i128, ptr [[TMP1]], align 8, !tbaa [[TBAA10]] +// RISCV32_INT128-NEXT: [[B:%.*]] = trunc i128 [[TMP3]] to i77 +// RISCV32_INT128-NEXT: [[ADD:%.*]] = add nsw i77 [[B]], [[A]] +// RISCV32_INT128-NEXT: [[STOREDV4:%.*]] = sext i77 [[ADD]] to i128 +// RISCV32_INT128-NEXT: store i128 [[STOREDV4]], ptr [[AGG_RESULT]], align 8, !tbaa [[TBAA10]] +// RISCV32_INT128-NEXT: ret void +// +_BitInt(77) test_bitint_77_add_default(_BitInt(77) a, _BitInt(77) b) { + return a + b; +} diff --git a/clang/test/CodeGen/X86/avx-builtins.c b/clang/test/CodeGen/X86/avx-builtins.c index 5f08b6b..bcffd861 100644 --- a/clang/test/CodeGen/X86/avx-builtins.c +++ b/clang/test/CodeGen/X86/avx-builtins.c @@ -1100,6 +1100,7 @@ __m256d test_mm256_hadd_pd(__m256d A, __m256d B) { // CHECK: call {{.*}}<4 x double> @llvm.x86.avx.hadd.pd.256(<4 x double> %{{.*}}, <4 x double> %{{.*}}) return _mm256_hadd_pd(A, B); } +TEST_CONSTEXPR(match_m256d(_mm256_hadd_pd((__m256d){+1.0, +2.0, +3.0, +4.0}, (__m256d){+5.0, +6.0, +7.0, +8.0}), +3.0, +11.0, +7.0, +15.0)); __m256 test_mm256_hadd_ps(__m256 A, __m256 B) { // CHECK-LABEL: test_mm256_hadd_ps @@ -1107,17 +1108,27 @@ __m256 test_mm256_hadd_ps(__m256 A, __m256 B) { return _mm256_hadd_ps(A, B); } +TEST_CONSTEXPR(match_m256(_mm256_hadd_ps( + (__m256){+1.0f, +2.0f, +3.0f, +4.0f, +5.0f, +6.0f, +7.0f, +8.0f}, + (__m256){+9.0f, +10.0f, +11.0f, +12.0f, +13.0f, +14.0f, +15.0f, +16.0f}), + +3.0f, +7.0f, +19.0f, +23.0f, +11.0f, +15.0f, +27.0f, +31.0f)); + __m256d test_mm256_hsub_pd(__m256d A, __m256d B) { // CHECK-LABEL: test_mm256_hsub_pd // CHECK: call {{.*}}<4 x double> @llvm.x86.avx.hsub.pd.256(<4 x double> %{{.*}}, <4 x double> %{{.*}}) return _mm256_hsub_pd(A, B); } +TEST_CONSTEXPR(match_m256d(_mm256_hsub_pd((__m256d){+1.0, +2.0, +4.0, +3.0}, (__m256d){+10.0, +6.0, +16.0, +8.0}), -1.0,+4.0,+1.0,+8.0)); __m256 test_mm256_hsub_ps(__m256 A, __m256 B) { // CHECK-LABEL: test_mm256_hsub_ps // CHECK: call {{.*}}<8 x float> @llvm.x86.avx.hsub.ps.256(<8 x float> %{{.*}}, <8 x float> %{{.*}}) return _mm256_hsub_ps(A, B); } +TEST_CONSTEXPR(match_m256(_mm256_hsub_ps( + (__m256){1.0f, 2.0f, 4.0f, 3.0f, 5.0f, 7.0f, 7.0f, 5.0f}, + (__m256){6.0f, 9.0f, 11.0f, 8.0f, 13.0f, 17.0f, 15.0f, 11.0f}), + -1.0f, 1.0f, -3.0f, 3.0f, -2.0f, 2.0f, -4.0f, 4.0f)); __m256i test_mm256_insert_epi8(__m256i x, char b) { // CHECK-LABEL: test_mm256_insert_epi8 @@ -2009,90 +2020,136 @@ int test_mm_testc_pd(__m128d A, __m128d B) { // CHECK: call {{.*}}i32 @llvm.x86.avx.vtestc.pd(<2 x double> %{{.*}}, <2 x double> %{{.*}}) return _mm_testc_pd(A, B); } +TEST_CONSTEXPR(_mm_testc_pd((__m128d)(__v2df){-1.0, -2.0},(__m128d)(__v2df){-3.0, 4.0}) == 1); +TEST_CONSTEXPR(_mm_testc_pd((__m128d)(__v2df){ 1.0, -2.0},(__m128d)(__v2df){-3.0, 4.0}) == 0); +TEST_CONSTEXPR(_mm_testc_pd((__m128d)(__v2df){ 1.0, -2.0},(__m128d)(__v2df){ 0.0, 5.0}) == 1); int test_mm256_testc_pd(__m256d A, __m256d B) { // CHECK-LABEL: test_mm256_testc_pd // CHECK: call {{.*}}i32 @llvm.x86.avx.vtestc.pd.256(<4 x double> %{{.*}}, <4 x double> %{{.*}}) return _mm256_testc_pd(A, B); } +TEST_CONSTEXPR(_mm256_testc_pd((__m256d)(__v4df){-1.0, 2.0, -3.0, 4.0},(__m256d)(__v4df){-5.0, 6.0, 7.0, 8.0}) == 1); +TEST_CONSTEXPR(_mm256_testc_pd((__m256d)(__v4df){ 1.0, 2.0, -3.0, 4.0},(__m256d)(__v4df){-5.0, 6.0, 7.0, 8.0}) == 0); +TEST_CONSTEXPR(_mm256_testc_pd((__m256d)(__v4df){-1.0, -2.0, -3.0, -4.0},(__m256d)(__v4df){ 5.0, 6.0, 7.0, 8.0}) == 1); int test_mm_testc_ps(__m128 A, __m128 B) { // CHECK-LABEL: test_mm_testc_ps // CHECK: call {{.*}}i32 @llvm.x86.avx.vtestc.ps(<4 x float> %{{.*}}, <4 x float> %{{.*}}) return _mm_testc_ps(A, B); } +TEST_CONSTEXPR(_mm_testc_ps((__m128)(__v4sf){-1,-1,-9001.1009,0x0FA1L}, (__m128)(__v4sf){-1.0,-9001,9001,9000}) == 1); +TEST_CONSTEXPR(_mm_testc_ps((__m128)(__v4sf){-1,2384.23,-9001.1009,0x0FA1L}, (__m128)(__v4sf){-1.0,-9001,9001,9000}) == 0); +TEST_CONSTEXPR(_mm_testc_ps((__m128)(__v4sf){-1,-2,-9001.1009,-93}, (__m128)(__v4sf){-1.0,-9001,-0.9001,-1000}) == 1); int test_mm256_testc_ps(__m256 A, __m256 B) { // CHECK-LABEL: test_mm256_testc_ps // CHECK: call {{.*}}i32 @llvm.x86.avx.vtestc.ps.256(<8 x float> %{{.*}}, <8 x float> %{{.*}}) return _mm256_testc_ps(A, B); } +TEST_CONSTEXPR(_mm256_testc_ps((__m256)(__v8sf){-1, -2, -3, -4, -5, -6, -7, 8},(__m256)(__v8sf){1, 2, 3, 4, 5, 6, 7, -8}) == 0); +TEST_CONSTEXPR(_mm256_testc_ps((__m256)(__v8sf){0,0,0,0,0,0,0,-1.00001},(__m256)(__v8sf){1, 2, 3, 4, 5, 6, 7, -8}) == 1); +TEST_CONSTEXPR(_mm256_testc_ps((__m256)(__v8sf){0,-0.00002,0,0,0,0,0,-1.00001},(__m256)(__v8sf){1, 2, 3, -4, 5, -6, -7, 8}) == 0); int test_mm256_testc_si256(__m256i A, __m256i B) { // CHECK-LABEL: test_mm256_testc_si256 // CHECK: call {{.*}}i32 @llvm.x86.avx.ptestc.256(<4 x i64> %{{.*}}, <4 x i64> %{{.*}}) return _mm256_testc_si256(A, B); } +TEST_CONSTEXPR(_mm256_testc_si256((__m256i)(__v4di){0,0,0,0}, (__m256i)(__v4di){0,0,0,0}) == 1); +TEST_CONSTEXPR(_mm256_testc_si256((__m256i)(__v4di){0,0,-1,0}, (__m256i)(__v4di){0,0,1,0}) == 1); +TEST_CONSTEXPR(_mm256_testc_si256((__m256i)(__v4di){-1,-2,1,3}, (__m256i)(__v4di){0,-1,1,1}) == 0); int test_mm_testnzc_pd(__m128d A, __m128d B) { // CHECK-LABEL: test_mm_testnzc_pd // CHECK: call {{.*}}i32 @llvm.x86.avx.vtestnzc.pd(<2 x double> %{{.*}}, <2 x double> %{{.*}}) return _mm_testnzc_pd(A, B); } +TEST_CONSTEXPR(_mm_testnzc_pd((__m128d)(__v2df){-1.0, +2.0},(__m128d)(__v2df){-3.0, -4.0}) == 1); +TEST_CONSTEXPR(_mm_testnzc_pd((__m128d)(__v2df){+1.0, +2.0},(__m128d)(__v2df){+3.0, -4.0}) == 0); +TEST_CONSTEXPR(_mm_testnzc_pd((__m128d)(__v2df){-1.0, -2.0},(__m128d)(__v2df){-3.0, +4.0}) == 0); int test_mm256_testnzc_pd(__m256d A, __m256d B) { // CHECK-LABEL: test_mm256_testnzc_pd // CHECK: call {{.*}}i32 @llvm.x86.avx.vtestnzc.pd.256(<4 x double> %{{.*}}, <4 x double> %{{.*}}) return _mm256_testnzc_pd(A, B); } +TEST_CONSTEXPR(_mm256_testnzc_pd((__m256d)(__v4df){-1.0, 2.0, 3.0, -4.0}, (__m256d)(__v4df){-5.0, -6.0, 7.0, 8.0}) == 1); +TEST_CONSTEXPR(_mm256_testnzc_pd((__m256d)(__v4df){1.0, 2.0, 3.0, 4.0}, (__m256d)(__v4df){-1.0, 6.0, 7.0, 8.0}) == 0); +TEST_CONSTEXPR(_mm256_testnzc_pd((__m256d)(__v4df){-1.0, -2.0, -3.0, -4.0}, (__m256d)(__v4df){-5.0, 6.0, 7.0, 8.0}) == 0); int test_mm_testnzc_ps(__m128 A, __m128 B) { // CHECK-LABEL: test_mm_testnzc_ps // CHECK: call {{.*}}i32 @llvm.x86.avx.vtestnzc.ps(<4 x float> %{{.*}}, <4 x float> %{{.*}}) return _mm_testnzc_ps(A, B); } +TEST_CONSTEXPR(_mm_testnzc_ps((__m128)(__v4sf){-9.9,987,-67,0}, (__m128)(__v4sf){10.0,-1.12,-29.29,0}) == 1); +TEST_CONSTEXPR(_mm_testnzc_ps((__m128)(__v4sf){-810.0,-1.0,-1.0,-3.0}, (__m128)(__v4sf){-10.0,-1.0,-1.0,-2.0}) == 0); +TEST_CONSTEXPR(_mm_testnzc_ps((__m128)(__v4sf){0,0,0,0}, (__m128)(__v4sf){0,-1,0,-1}) == 0); int test_mm256_testnzc_ps(__m256 A, __m256 B) { // CHECK-LABEL: test_mm256_testnzc_ps // CHECK: call {{.*}}i32 @llvm.x86.avx.vtestnzc.ps.256(<8 x float> %{{.*}}, <8 x float> %{{.*}}) return _mm256_testnzc_ps(A, B); } +TEST_CONSTEXPR(_mm256_testnzc_ps((__m256)(__v8sf){-1, -2, -3, -4, -5, -6, -7, 8},(__m256)(__v8sf){1, -2, 3, 4, 5, 6, 7, -8}) == 1); +TEST_CONSTEXPR(_mm256_testnzc_ps((__m256)(__v8sf){0,0,0,0,0,0,0,-1.00001},(__m256)(__v8sf){1, 2, 3, 4, 5, 6, 7, -8}) == 0); +TEST_CONSTEXPR(_mm256_testnzc_ps((__m256)(__v8sf){0,-0.00002,0,0,0,0,0,-1.00001},(__m256)(__v8sf){1, 2, 3, -4, 5, -6, -7, 8}) == 0); int test_mm256_testnzc_si256(__m256i A, __m256i B) { // CHECK-LABEL: test_mm256_testnzc_si256 // CHECK: call {{.*}}i32 @llvm.x86.avx.ptestnzc.256(<4 x i64> %{{.*}}, <4 x i64> %{{.*}}) return _mm256_testnzc_si256(A, B); } +TEST_CONSTEXPR(_mm256_testnzc_si256((__m256i)(__v4di){-1,0,0,0}, (__m256i)(__v4di){478329848,23438,2343,-3483}) == 1); +TEST_CONSTEXPR(_mm256_testnzc_si256((__m256i)(__v4di){1,0,0,0}, (__m256i)(__v4di){3,0,0,0}) == 1); +TEST_CONSTEXPR(_mm256_testnzc_si256((__m256i)(__v4di){-1,239483,-1,0}, (__m256i)(__v4di){3849234,0,-2,0}) == 0); +TEST_CONSTEXPR(_mm256_testnzc_si256((__m256i)(__v4di){-1,0,-1,3}, (__m256i)(__v4di){1,0,9999999,1}) == 0); int test_mm_testz_pd(__m128d A, __m128d B) { // CHECK-LABEL: test_mm_testz_pd // CHECK: call {{.*}}i32 @llvm.x86.avx.vtestz.pd(<2 x double> %{{.*}}, <2 x double> %{{.*}}) return _mm_testz_pd(A, B); } +TEST_CONSTEXPR(_mm_testz_pd((__m128d)(__v2df){-1,0}, (__m128d)(__v2df){0,-1}) == 1); +TEST_CONSTEXPR(_mm_testz_pd((__m128d)(__v2df){0,-13.13}, (__m128d)(__v2df){0,-11.1}) == 0); +TEST_CONSTEXPR(_mm_testz_pd((__m128d)(__v2df){0,5.13}, (__m128d)(__v2df){0,-113.1324823}) == 1); int test_mm256_testz_pd(__m256d A, __m256d B) { // CHECK-LABEL: test_mm256_testz_pd // CHECK: call {{.*}}i32 @llvm.x86.avx.vtestz.pd.256(<4 x double> %{{.*}}, <4 x double> %{{.*}}) return _mm256_testz_pd(A, B); } +TEST_CONSTEXPR(_mm256_testz_pd((__m256d)(__v4df){-1,0,-47.47,0.00002}, (__m256d)(__v4df){0,-1,74.0101,-1}) == 1); +TEST_CONSTEXPR(_mm256_testz_pd((__m256d)(__v4df){-1,3249.9,-47.47,-0.00002}, (__m256d)(__v4df){0,-1,74.0101,-9999900}) == 0); +TEST_CONSTEXPR(_mm256_testz_pd((__m256d)(__v4df){0,0,-8,0}, (__m256d)(__v4df){0,-1,-101,-123}) == 0); int test_mm_testz_ps(__m128 A, __m128 B) { // CHECK-LABEL: test_mm_testz_ps // CHECK: call {{.*}}i32 @llvm.x86.avx.vtestz.ps(<4 x float> %{{.*}}, <4 x float> %{{.*}}) return _mm_testz_ps(A, B); } +TEST_CONSTEXPR(_mm_testz_ps((__m128)(__v4sf){-9.9,987,67,0}, (__m128)(__v4sf){10.0,-1.12,-29.29,0}) == 1); +TEST_CONSTEXPR(_mm_testz_ps((__m128)(__v4sf){10.0,1.0,-1.0,-3.0}, (__m128)(__v4sf){-10.0,-1.0,-1.0,-2.0}) == 0); +TEST_CONSTEXPR(_mm_testz_ps((__m128)(__v4sf){0,0,0,0}, (__m128)(__v4sf){0,-1,0,-1}) == 1); int test_mm256_testz_ps(__m256 A, __m256 B) { // CHECK-LABEL: test_mm256_testz_ps // CHECK: call {{.*}}i32 @llvm.x86.avx.vtestz.ps.256(<8 x float> %{{.*}}, <8 x float> %{{.*}}) return _mm256_testz_ps(A, B); } +TEST_CONSTEXPR(_mm256_testz_ps((__m256)(__v8sf){-1, -2, -3, -4, -5, -6, -7, 8},(__m256)(__v8sf){1, 2, 3, 4, 5, 6, 7, -8}) == 1); +TEST_CONSTEXPR(_mm256_testz_ps((__m256)(__v8sf){0,0,0,0,0,0,0,-1.00001},(__m256)(__v8sf){1, 2, 3, 4, 5, 6, 7, -8}) == 0); +TEST_CONSTEXPR(_mm256_testz_ps((__m256)(__v8sf){0,-0.00002,0,0,0,0,0,-1.00001},(__m256)(__v8sf){1, 2, 3, -4, 5, -6, -7, 8}) == 1); int test_mm256_testz_si256(__m256i A, __m256i B) { // CHECK-LABEL: test_mm256_testz_si256 // CHECK: call {{.*}}i32 @llvm.x86.avx.ptestz.256(<4 x i64> %{{.*}}, <4 x i64> %{{.*}}) return _mm256_testz_si256(A, B); } +TEST_CONSTEXPR(_mm256_testz_si256((__m256i)(__v4di){0,0,0,0}, (__m256i)(__v4di){0,0,0,0}) == 1); +TEST_CONSTEXPR(_mm256_testz_si256((__m256i)(__v4di){0,0,-1,0}, (__m256i)(__v4di){0,0,-1,0}) == 0); +TEST_CONSTEXPR(_mm256_testz_si256((__m256i)(__v4di){-1,0,1,0}, (__m256i)(__v4di){0,-1,0,1}) == 1); __m256 test_mm256_undefined_ps(void) { // X64-LABEL: test_mm256_undefined_ps diff --git a/clang/test/CodeGen/X86/avx2-builtins.c b/clang/test/CodeGen/X86/avx2-builtins.c index 55f18f9..b798618 100644 --- a/clang/test/CodeGen/X86/avx2-builtins.c +++ b/clang/test/CodeGen/X86/avx2-builtins.c @@ -485,36 +485,60 @@ __m256i test_mm256_hadd_epi16(__m256i a, __m256i b) { // CHECK: call <16 x i16> @llvm.x86.avx2.phadd.w(<16 x i16> %{{.*}}, <16 x i16> %{{.*}}) return _mm256_hadd_epi16(a, b); } +TEST_CONSTEXPR(match_v16hi(_mm256_hadd_epi16( + (__m256i)(__v16hi){1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16}, + (__m256i)(__v16hi){17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32}), + 3,7,11,15,35,39,43,47,19,23,27,31,51,55,59,63)); __m256i test_mm256_hadd_epi32(__m256i a, __m256i b) { // CHECK-LABEL: test_mm256_hadd_epi32 // CHECK: call <8 x i32> @llvm.x86.avx2.phadd.d(<8 x i32> %{{.*}}, <8 x i32> %{{.*}}) return _mm256_hadd_epi32(a, b); } +TEST_CONSTEXPR(match_v8si(_mm256_hadd_epi32( + (__m256i)(__v8si){10, 20, 30, 40, 50, 60, 70, 80}, + (__m256i)(__v8si){5, 15, 25, 35, 45, 55, 65, 75}), + 30,70,20,60,110,150,100,140)); __m256i test_mm256_hadds_epi16(__m256i a, __m256i b) { // CHECK-LABEL: test_mm256_hadds_epi16 // CHECK:call <16 x i16> @llvm.x86.avx2.phadd.sw(<16 x i16> %{{.*}}, <16 x i16> %{{.*}}) return _mm256_hadds_epi16(a, b); } +TEST_CONSTEXPR(match_v16hi( _mm256_hadds_epi16( + (__m256i)(__v16hi){32767, 32767, 1,2,3,4,5,6,7,8,9,10,11,12,13,14}, + (__m256i)(__v16hi){19,20,21,22,23,24,25,26,27,28,29,30,31,32, 32767, 5}), + 32767, 3,7,11, 39,43,47,51,15,19,23,27, 55,59,63, 32767)); __m256i test_mm256_hsub_epi16(__m256i a, __m256i b) { // CHECK-LABEL: test_mm256_hsub_epi16 // CHECK: call <16 x i16> @llvm.x86.avx2.phsub.w(<16 x i16> %{{.*}}, <16 x i16> %{{.*}}) return _mm256_hsub_epi16(a, b); } +TEST_CONSTEXPR(match_v16hi(_mm256_hsub_epi16( + (__m256i)(__v16hi){2,1,1,2,5,3,3,5,7,4,4,7,9,5,5,9}, + (__m256i)(__v16hi){10,5,5,10,12,6,6,12,21,14,14,21,24,16,16,24}), + 1,-1,2,-2,5,-5,6,-6,3,-3,4,-4, 7,-7,8,-8)); __m256i test_mm256_hsub_epi32(__m256i a, __m256i b) { // CHECK-LABEL: test_mm256_hsub_epi32 // CHECK: call <8 x i32> @llvm.x86.avx2.phsub.d(<8 x i32> %{{.*}}, <8 x i32> %{{.*}}) return _mm256_hsub_epi32(a, b); } +TEST_CONSTEXPR(match_v8si(_mm256_hsub_epi32( + (__m256i)(__v8si){10, 20, 30,50,60,90,100,140}, + (__m256i)(__v8si){200,150,260,200,420,350,800,720}), + -10,-20,50,60, -30,-40, 70,80)); __m256i test_mm256_hsubs_epi16(__m256i a, __m256i b) { // CHECK-LABEL: test_mm256_hsubs_epi16 // CHECK:call <16 x i16> @llvm.x86.avx2.phsub.sw(<16 x i16> %{{.*}}, <16 x i16> %{{.*}}) return _mm256_hsubs_epi16(a, b); } +TEST_CONSTEXPR(match_v16hi(_mm256_hsubs_epi16( + (__m256i)(__v16hi){32726, -100, 3, 2, 6, 4, 8, 5,15,10 ,21, 14, 27, 18, 100, 90}, + (__m256i)(__v16hi){40, 20, 100, 70, 200,150, 100,40, 1000,900,300,150, 500,300, 1, 1}), + 32767, 1, 2, 3, 20, 30, 50, 60, 5, 7, 9, 10, 100, 150, 200, 0)); __m128i test_mm_i32gather_epi32(int const *b, __m128i c) { // CHECK-LABEL: test_mm_i32gather_epi32 @@ -1106,6 +1130,8 @@ __m256i test_mm256_shuffle_epi8(__m256i a, __m256i b) { return _mm256_shuffle_epi8(a, b); } +TEST_CONSTEXPR(match_v32qi(_mm256_shuffle_epi8((__m256i)(__v32qi){0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31}, (__m256i)(__v32qs){0,33,2,35,4,37,6,-39,8,41,10,43,12,45,14,-47,16,49,18,51,20,53,22,-55,24,57,26,59,28,61,30,-63}), 0,1,2,3,4,5,6,0,8,9,10,11,12,13,14,0,16,17,18,19,20,21,22,0,24,25,26,27,28,29,30,0)); + __m256i test_mm256_shuffle_epi32(__m256i a) { // CHECK-LABEL: test_mm256_shuffle_epi32 // CHECK: shufflevector <8 x i32> %{{.*}}, <8 x i32> poison, <8 x i32> <i32 3, i32 3, i32 0, i32 0, i32 7, i32 7, i32 4, i32 4> diff --git a/clang/test/CodeGen/X86/avx512bw-builtins.c b/clang/test/CodeGen/X86/avx512bw-builtins.c index af1c904..fddf17d 100644 --- a/clang/test/CodeGen/X86/avx512bw-builtins.c +++ b/clang/test/CodeGen/X86/avx512bw-builtins.c @@ -1466,18 +1466,27 @@ __m512i test_mm512_shuffle_epi8(__m512i __A, __m512i __B) { // CHECK: @llvm.x86.avx512.pshuf.b.512 return _mm512_shuffle_epi8(__A,__B); } + +TEST_CONSTEXPR(match_v64qi(_mm512_shuffle_epi8((__m512i)(__v64qi){0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63}, (__m512i)(__v64qs){0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,-15,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,-15,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,-79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,-95}), 0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,0,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,0,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,0,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,0)); + __m512i test_mm512_mask_shuffle_epi8(__m512i __W, __mmask64 __U, __m512i __A, __m512i __B) { // CHECK-LABEL: test_mm512_mask_shuffle_epi8 // CHECK: @llvm.x86.avx512.pshuf.b.512 // CHECK: select <64 x i1> %{{.*}}, <64 x i8> %{{.*}}, <64 x i8> %{{.*}} return _mm512_mask_shuffle_epi8(__W,__U,__A,__B); } + +TEST_CONSTEXPR(match_v64qi(_mm512_mask_shuffle_epi8((__m512i)(__v64qi){1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,3,3,3,3,3,3,3,3,4,4,4,4,4,4,4,4,5,5,5,5,5,5,5,5,6,6,6,6,6,6,6,6,7,7,7,7,7,7,7,7,8,8,8,8,8,8,8,8}, 0xFFFFFFFF00000000, (__m512i)(__v64qi){0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63}, (__m512i)(__v64qi){63,62,61,60,59,58,57,56,55,54,53,52,51,50,49,48,47,46,45,44,43,42,41,40,39,38,37,36,35,34,33,32,31,30,29,28,27,26,25,24,23,22,21,20,19,18,17,16,15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0}), 1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,3,3,3,3,3,3,3,3,4,4,4,4,4,4,4,4,47,46,45,44,43,42,41,40,39,38,37,36,35,34,33,32,63,62,61,60,59,58,57,56,55,54,53,52,51,50,49,48)); + __m512i test_mm512_maskz_shuffle_epi8(__mmask64 __U, __m512i __A, __m512i __B) { // CHECK-LABEL: test_mm512_maskz_shuffle_epi8 // CHECK: @llvm.x86.avx512.pshuf.b.512 // CHECK: select <64 x i1> %{{.*}}, <64 x i8> %{{.*}}, <64 x i8> %{{.*}} return _mm512_maskz_shuffle_epi8(__U,__A,__B); } + +TEST_CONSTEXPR(match_v64qi(_mm512_maskz_shuffle_epi8(0x8888888888888888,(__m512i)(__v64qi){0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63}, (__m512i)(__v64qi){127,126,125,124,123,122,121,120,119,118,117,116,115,114,113,112,111,110,109,108,107,106,105,104,103,102,101,100,99,98,97,96,95,94,93,92,91,90,89,88,87,86,85,84,83,82,81,80,79,78,77,76,75,74,73,72,71,70,69,68,67,66,65,64}), 0,0,0,12,0,0,0,8,0,0,0,4,0,0,0,0,0,0,0,28,0,0,0,24,0,0,0,20,0,0,0,16,0,0,0,44,0,0,0,40,0,0,0,36,0,0,0,32,0,0,0,60,0,0,0,56,0,0,0,52,0,0,0,48)); + __m512i test_mm512_subs_epi8(__m512i __A, __m512i __B) { // CHECK-LABEL: test_mm512_subs_epi8 // CHECK: @llvm.ssub.sat.v64i8 diff --git a/clang/test/CodeGen/X86/avx512cd-builtins.c b/clang/test/CodeGen/X86/avx512cd-builtins.c index b9d42b7..80a20b1 100644 --- a/clang/test/CodeGen/X86/avx512cd-builtins.c +++ b/clang/test/CodeGen/X86/avx512cd-builtins.c @@ -14,37 +14,53 @@ __m512i test_mm512_conflict_epi64(__m512i __A) { // CHECK-LABEL: test_mm512_conflict_epi64 // CHECK: call {{.*}}<8 x i64> @llvm.x86.avx512.conflict.q.512(<8 x i64> %{{.*}}) - return _mm512_conflict_epi64(__A); + return _mm512_conflict_epi64(__A); } + +TEST_CONSTEXPR(match_v8di(_mm512_conflict_epi64((__m512i)(__v8di){1, 2, 1, 3, 2, 4, 1, 5}), 0, 0, 1, 0, 2, 0, 5, 0)); +TEST_CONSTEXPR(match_v8di(_mm512_conflict_epi64((__m512i)(__v8di){5, 5, 5, 5, 5, 5, 5, 5}), 0, 1, 3, 7, 15, 31, 63, 127)); +TEST_CONSTEXPR(match_v8di(_mm512_conflict_epi64((__m512i)(__v8di){1, 2, 3, 4, 5, 6, 7, 8}), 0, 0, 0, 0, 0, 0, 0, 0)); __m512i test_mm512_mask_conflict_epi64(__m512i __W, __mmask8 __U, __m512i __A) { // CHECK-LABEL: test_mm512_mask_conflict_epi64 // CHECK: call {{.*}}<8 x i64> @llvm.x86.avx512.conflict.q.512(<8 x i64> %{{.*}}) // CHECK: select <8 x i1> %{{.*}}, <8 x i64> %{{.*}}, <8 x i64> %{{.*}} - return _mm512_mask_conflict_epi64(__W,__U,__A); + return _mm512_mask_conflict_epi64(__W,__U,__A); } + +TEST_CONSTEXPR(match_v8di(_mm512_mask_conflict_epi64((__m512i)(__v8di){0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF}, 0x55, (__m512i)(__v8di){1, 2, 1, 3, 2, 4, 1, 5}), 0, 0xFF, 1, 0xFF, 2, 0xFF, 5, 0xFF)); __m512i test_mm512_maskz_conflict_epi64(__mmask8 __U, __m512i __A) { // CHECK-LABEL: test_mm512_maskz_conflict_epi64 // CHECK: call {{.*}}<8 x i64> @llvm.x86.avx512.conflict.q.512(<8 x i64> %{{.*}}) // CHECK: select <8 x i1> %{{.*}}, <8 x i64> %{{.*}}, <8 x i64> %{{.*}} - return _mm512_maskz_conflict_epi64(__U,__A); + return _mm512_maskz_conflict_epi64(__U,__A); } + +TEST_CONSTEXPR(match_v8di(_mm512_maskz_conflict_epi64(0x55, (__m512i)(__v8di){1, 2, 1, 3, 2, 4, 1, 5}), 0, 0, 1, 0, 2, 0, 5, 0)); __m512i test_mm512_conflict_epi32(__m512i __A) { // CHECK-LABEL: test_mm512_conflict_epi32 // CHECK: call <16 x i32> @llvm.x86.avx512.conflict.d.512(<16 x i32> %{{.*}}) - return _mm512_conflict_epi32(__A); + return _mm512_conflict_epi32(__A); } + +TEST_CONSTEXPR(match_v16si(_mm512_conflict_epi32((__m512i)(__v16si){1, 2, 1, 3, 2, 4, 1, 5, 6, 7, 6, 8, 7, 9, 6, 10}), 0, 0, 1, 0, 2, 0, 5, 0, 0, 0, 256, 0, 512, 0, 1280, 0)); +TEST_CONSTEXPR(match_v16si(_mm512_conflict_epi32((__m512i)(__v16si){9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9}), 0, 1, 3, 7, 15, 31, 63, 127, 255, 511, 1023, 2047, 4095, 8191, 16383, 32767)); +TEST_CONSTEXPR(match_v16si(_mm512_conflict_epi32((__m512i)(__v16si){1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16}), 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0)); __m512i test_mm512_mask_conflict_epi32(__m512i __W, __mmask16 __U, __m512i __A) { // CHECK-LABEL: test_mm512_mask_conflict_epi32 // CHECK: call <16 x i32> @llvm.x86.avx512.conflict.d.512(<16 x i32> %{{.*}}) // CHECK: select <16 x i1> %{{.*}}, <16 x i32> %{{.*}}, <16 x i32> %{{.*}} - return _mm512_mask_conflict_epi32(__W,__U,__A); + return _mm512_mask_conflict_epi32(__W,__U,__A); } + +TEST_CONSTEXPR(match_v16si(_mm512_mask_conflict_epi32((__m512i)(__v16si){0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF}, 0x5555, (__m512i)(__v16si){1, 2, 1, 3, 2, 4, 1, 5, 6, 7, 6, 8, 7, 9, 6, 10}), 0, 0xFF, 1, 0xFF, 2, 0xFF, 5, 0xFF, 0, 0xFF, 256, 0xFF, 512, 0xFF, 1280, 0xFF)); __m512i test_mm512_maskz_conflict_epi32(__mmask16 __U, __m512i __A) { // CHECK-LABEL: test_mm512_maskz_conflict_epi32 // CHECK: call <16 x i32> @llvm.x86.avx512.conflict.d.512(<16 x i32> %{{.*}}) // CHECK: select <16 x i1> %{{.*}}, <16 x i32> %{{.*}}, <16 x i32> %{{.*}} - return _mm512_maskz_conflict_epi32(__U,__A); + return _mm512_maskz_conflict_epi32(__U,__A); } + +TEST_CONSTEXPR(match_v16si(_mm512_maskz_conflict_epi32(0x5555, (__m512i)(__v16si){1, 2, 1, 3, 2, 4, 1, 5, 6, 7, 6, 8, 7, 9, 6, 10}), 0, 0, 1, 0, 2, 0, 5, 0, 0, 0, 256, 0, 512, 0, 1280, 0)); __m512i test_mm512_lzcnt_epi32(__m512i __A) { // CHECK-LABEL: test_mm512_lzcnt_epi32 // CHECK: call <16 x i32> @llvm.ctlz.v16i32(<16 x i32> %{{.*}}, i1 true) @@ -125,6 +141,8 @@ __m512i test_mm512_broadcastmb_epi64(__m512i a, __m512i b) { // CHECK: insertelement <8 x i64> %{{.*}}, i64 %{{.*}}, i32 7 return _mm512_broadcastmb_epi64(_mm512_cmpeq_epu64_mask ( a, b)); } +TEST_CONSTEXPR(match_v8di(_mm512_broadcastmb_epi64((__mmask8)(0)), 0,0,0,0, 0,0,0,0)); +TEST_CONSTEXPR(match_v8di(_mm512_broadcastmb_epi64((__mmask8)(0xab)), 0xab,0xab,0xab,0xab, 0xab,0xab,0xab,0xab)); __m512i test_mm512_broadcastmw_epi32(__m512i a, __m512i b) { // CHECK-LABEL: test_mm512_broadcastmw_epi32 @@ -148,3 +166,5 @@ __m512i test_mm512_broadcastmw_epi32(__m512i a, __m512i b) { // CHECK: insertelement <16 x i32> %{{.*}}, i32 %{{.*}} return _mm512_broadcastmw_epi32(_mm512_cmpeq_epi32_mask ( a, b)); } +TEST_CONSTEXPR(match_v16si(_mm512_broadcastmw_epi32((__mmask16)(0xff)), 0xff,0xff,0xff,0xff, 0xff,0xff,0xff,0xff, 0xff,0xff,0xff,0xff, 0xff,0xff,0xff,0xff)); +TEST_CONSTEXPR(match_v16si(_mm512_broadcastmw_epi32((__mmask16)(0x0FA1L)), 0x0FA1L,0x0FA1L,0x0FA1L,0x0FA1L, 0x0FA1L,0x0FA1L,0x0FA1L,0x0FA1L, 0x0FA1L,0x0FA1L,0x0FA1L,0x0FA1L, 0x0FA1L,0x0FA1L,0x0FA1L,0x0FA1L)); diff --git a/clang/test/CodeGen/X86/avx512dq-builtins.c b/clang/test/CodeGen/X86/avx512dq-builtins.c index 4112561..1b09959 100644 --- a/clang/test/CodeGen/X86/avx512dq-builtins.c +++ b/clang/test/CodeGen/X86/avx512dq-builtins.c @@ -1305,6 +1305,7 @@ __m512 test_mm512_mask_broadcast_f32x8(__m512 __O, __mmask16 __M, float const* _ // CHECK: select <16 x i1> %{{.*}}, <16 x float> %{{.*}}, <16 x float> %{{.*}} return _mm512_mask_broadcast_f32x8(__O, __M, _mm256_loadu_ps(__A)); } +TEST_CONSTEXPR(match_m512(_mm512_mask_broadcast_f32x8(_mm512_setzero_ps(), 0xAAAA, (__m256)(__v8sf){5.0f,5.0f,5.0f,5.0f,5.0f,5.0f,5.0f,5.0f}), 0,5,0,5,0,5,0,5,0,5,0,5,0,5,0,5)); __m512 test_mm512_maskz_broadcast_f32x8(__mmask16 __M, float const* __A) { // CHECK-LABEL: test_mm512_maskz_broadcast_f32x8 @@ -1312,6 +1313,7 @@ __m512 test_mm512_maskz_broadcast_f32x8(__mmask16 __M, float const* __A) { // CHECK: select <16 x i1> %{{.*}}, <16 x float> %{{.*}}, <16 x float> %{{.*}} return _mm512_maskz_broadcast_f32x8(__M, _mm256_loadu_ps(__A)); } +TEST_CONSTEXPR(match_m512(_mm512_maskz_broadcast_f32x8(0xAAAA, _mm256_set1_ps(7.0f)), 0,7,0,7,0,7,0,7,0,7,0,7,0,7,0,7)); __m512d test_mm512_broadcast_f64x2(double const* __A) { // CHECK-LABEL: test_mm512_broadcast_f64x2 @@ -1327,6 +1329,8 @@ __m512d test_mm512_mask_broadcast_f64x2(__m512d __O, __mmask8 __M, double const* return _mm512_mask_broadcast_f64x2(__O, __M, _mm_loadu_pd(__A)); } +TEST_CONSTEXPR(match_m512d(_mm512_mask_broadcast_f64x2(_mm512_setzero_pd(), 0xAA, (__m128d)(__v2df){1,2}), 0,2,0,2,0,2,0,2)); + __m512d test_mm512_maskz_broadcast_f64x2(__mmask8 __M, double const* __A) { // CHECK-LABEL: test_mm512_maskz_broadcast_f64x2 // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1> @@ -1334,6 +1338,8 @@ __m512d test_mm512_maskz_broadcast_f64x2(__mmask8 __M, double const* __A) { return _mm512_maskz_broadcast_f64x2(__M, _mm_loadu_pd(__A)); } +TEST_CONSTEXPR(match_m512d(_mm512_maskz_broadcast_f64x2(0xAA, (__m128d)(__v2df){1,2}), 0,2,0,2,0,2,0,2)); + __m512i test_mm512_broadcast_i32x2(__m128i __A) { // CHECK-LABEL: test_mm512_broadcast_i32x2 // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <16 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1> @@ -1348,6 +1354,8 @@ __m512i test_mm512_mask_broadcast_i32x2(__m512i __O, __mmask16 __M, __m128i __A) return _mm512_mask_broadcast_i32x2(__O, __M, __A); } +TEST_CONSTEXPR(match_v16si(_mm512_mask_broadcast_i32x2(_mm512_setzero_si512(), 0xAAAA, (__m128i)(__v4si){0,1,2,3}), 0,1,0,1,0,1,0,1,0,1,0,1,0,1,0,1)); + __m512i test_mm512_maskz_broadcast_i32x2(__mmask16 __M, __m128i __A) { // CHECK-LABEL: test_mm512_maskz_broadcast_i32x2 // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <16 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1> @@ -1355,6 +1363,8 @@ __m512i test_mm512_maskz_broadcast_i32x2(__mmask16 __M, __m128i __A) { return _mm512_maskz_broadcast_i32x2(__M, __A); } +TEST_CONSTEXPR(match_v16si(_mm512_maskz_broadcast_i32x2(0xAAAA, (__m128i)(__v4si){0,1,2,3}), 0,1,0,1,0,1,0,1,0,1,0,1,0,1,0,1)); + __m512i test_mm512_broadcast_i32x8(__m256i const* __A) { // CHECK-LABEL: test_mm512_broadcast_i32x8 // CHECK: shufflevector <8 x i32> %{{.*}}, <8 x i32> %{{.*}}, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7> @@ -1368,6 +1378,7 @@ __m512i test_mm512_mask_broadcast_i32x8(__m512i __O, __mmask16 __M, __m256i cons // CHECK: select <16 x i1> %{{.*}}, <16 x i32> %{{.*}}, <16 x i32> %{{.*}} return _mm512_mask_broadcast_i32x8(__O, __M, _mm256_loadu_si256(__A)); } +TEST_CONSTEXPR(match_v16si(_mm512_mask_broadcast_i32x8(_mm512_setzero_si512(), 0xAAAA, _mm256_set1_epi32(8)), 0,8,0,8,0,8,0,8,0,8,0,8,0,8,0,8)); __m512i test_mm512_maskz_broadcast_i32x8(__mmask16 __M, __m256i const* __A) { // CHECK-LABEL: test_mm512_maskz_broadcast_i32x8 @@ -1376,6 +1387,8 @@ __m512i test_mm512_maskz_broadcast_i32x8(__mmask16 __M, __m256i const* __A) { return _mm512_maskz_broadcast_i32x8(__M, _mm256_loadu_si256(__A)); } +TEST_CONSTEXPR(match_v16si(_mm512_maskz_broadcast_i32x8(0xAAAA, _mm256_set1_epi32(9)), 0,9,0,9,0,9,0,9,0,9,0,9,0,9,0,9)); + __m512i test_mm512_broadcast_i64x2(__m128i const* __A) { // CHECK-LABEL: test_mm512_broadcast_i64x2 // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1> diff --git a/clang/test/CodeGen/X86/avx512f-builtins.c b/clang/test/CodeGen/X86/avx512f-builtins.c index 7756f0d..3deaf8e 100644 --- a/clang/test/CodeGen/X86/avx512f-builtins.c +++ b/clang/test/CodeGen/X86/avx512f-builtins.c @@ -6836,6 +6836,8 @@ __m512 test_mm512_mask_broadcast_f32x4(__m512 __O, __mmask16 __M, float const* _ return _mm512_mask_broadcast_f32x4(__O, __M, _mm_loadu_ps(__A)); } +TEST_CONSTEXPR(match_m512(_mm512_mask_broadcast_f32x4(_mm512_setzero_ps(), 0xAAAA, (__m128)(__v4sf){1,2,3,4}), 0,2,0,4,0,2,0,4,0,2,0,4,0,2,0,4)); + __m512 test_mm512_maskz_broadcast_f32x4(__mmask16 __M, float const* __A) { // CHECK-LABEL: test_mm512_maskz_broadcast_f32x4 // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3> @@ -6843,6 +6845,8 @@ __m512 test_mm512_maskz_broadcast_f32x4(__mmask16 __M, float const* __A) { return _mm512_maskz_broadcast_f32x4(__M, _mm_loadu_ps(__A)); } +TEST_CONSTEXPR(match_m512(_mm512_maskz_broadcast_f32x4(0xAAAA, (__m128)(__v4sf){1,2,3,4}), 0,2,0,4,0,2,0,4,0,2,0,4,0,2,0,4)); + __m512d test_mm512_broadcast_f64x4(double const* __A) { // CHECK-LABEL: test_mm512_broadcast_f64x4 // CHECK: shufflevector <4 x double> %{{.*}}, <4 x double> %{{.*}}, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3> @@ -6885,6 +6889,8 @@ __m512i test_mm512_maskz_broadcast_i32x4(__mmask16 __M, __m128i const* __A) { return _mm512_maskz_broadcast_i32x4(__M, _mm_loadu_si128(__A)); } +TEST_CONSTEXPR(match_v16si(_mm512_maskz_broadcast_i32x4(0xAAAA, (__m128i)(__v4si){0,1,2,3}), 0,1,0,3,0,1,0,3,0,1,0,3,0,1,0,3)); + __m512i test_mm512_broadcast_i64x4(__m256i const* __A) { // CHECK-LABEL: test_mm512_broadcast_i64x4 // CHECK: shufflevector <4 x i64> %{{.*}}, <4 x i64> %{{.*}}, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3> @@ -10903,6 +10909,8 @@ __m512i test_mm512_mask_set1_epi32 (__m512i __O, __mmask16 __M, int __A) return _mm512_mask_set1_epi32 ( __O, __M, __A); } +TEST_CONSTEXPR(match_v16si(_mm512_mask_set1_epi32(_mm512_setzero_si512(), 0xAAAA, 13), 0,13,0,13,0,13,0,13,0,13,0,13,0,13,0,13)); + __m512i test_mm512_maskz_set1_epi32(__mmask16 __M, int __A) { // CHECK-LABEL: test_mm512_maskz_set1_epi32 @@ -10926,6 +10934,8 @@ __m512i test_mm512_maskz_set1_epi32(__mmask16 __M, int __A) return _mm512_maskz_set1_epi32(__M, __A); } +TEST_CONSTEXPR(match_v16si(_mm512_maskz_set1_epi32(0xAAAA, 19), 0,19,0,19,0,19,0,19,0,19,0,19,0,19,0,19)); + __m512i test_mm512_set_epi8(char e63, char e62, char e61, char e60, char e59, char e58, char e57, char e56, char e55, char e54, char e53, char e52, @@ -11145,6 +11155,8 @@ __m512i test_mm512_mask_set1_epi64 (__m512i __O, __mmask8 __M, long long __A) return _mm512_mask_set1_epi64 (__O, __M, __A); } +TEST_CONSTEXPR(match_v8di(_mm512_mask_set1_epi64(_mm512_setzero_si512(), 0xAA, 21), 0,21,0,21,0,21,0,21)); + __m512i test_mm512_maskz_set1_epi64 (__mmask8 __M, long long __A) { // CHECK-LABEL: test_mm512_maskz_set1_epi64 @@ -11160,6 +11172,8 @@ __m512i test_mm512_maskz_set1_epi64 (__mmask8 __M, long long __A) return _mm512_maskz_set1_epi64 (__M, __A); } +TEST_CONSTEXPR(match_v8di(_mm512_maskz_set1_epi64(0xAA, 23), 0, 23, 0, 23, 0, 23, 0, 23)); + __m512i test_mm512_set_epi64 (long long __A, long long __B, long long __C, long long __D, long long __E, long long __F, diff --git a/clang/test/CodeGen/X86/avx512ifma-builtins.c b/clang/test/CodeGen/X86/avx512ifma-builtins.c index eebefb0..f90697e 100644 --- a/clang/test/CodeGen/X86/avx512ifma-builtins.c +++ b/clang/test/CodeGen/X86/avx512ifma-builtins.c @@ -8,45 +8,230 @@ // RUN: %clang_cc1 -x c++ -flax-vector-conversions=none -ffreestanding %s -triple=x86_64-apple-darwin -target-feature +avx512ifma -emit-llvm -o - -Wall -Werror -fexperimental-new-constant-interpreter | FileCheck %s // RUN: %clang_cc1 -x c++ -flax-vector-conversions=none -ffreestanding %s -triple=i386-apple-darwin -target-feature +avx512ifma -emit-llvm -o - -Wall -Werror -fexperimental-new-constant-interpreter | FileCheck %s - #include <immintrin.h> +#include "builtin_test_helpers.h" __m512i test_mm512_madd52hi_epu64(__m512i __X, __m512i __Y, __m512i __Z) { // CHECK-LABEL: test_mm512_madd52hi_epu64 // CHECK: call {{.*}}<8 x i64> @llvm.x86.avx512.vpmadd52h.uq.512(<8 x i64> %{{.*}}, <8 x i64> %{{.*}}, <8 x i64> %{{.*}}) - return _mm512_madd52hi_epu64(__X, __Y, __Z); + return _mm512_madd52hi_epu64(__X, __Y, __Z); } +TEST_CONSTEXPR(match_v8di(_mm512_madd52hi_epu64( + (__m512i)(__v8du){100, 0, 0, 0, 0, 0, 0, 0}, + (__m512i)(__v8du){10, 0, 0, 0, 0, 0, 0, 0}, + (__m512i)(__v8du){5, 0, 0, 0, 0, 0, 0, 0}), + 100, 0, 0, 0, 0, 0, 0, 0)); + +TEST_CONSTEXPR(match_v8di(_mm512_madd52hi_epu64( + (__m512i)(__v8du){0, 0, 0, 0, 0, 0, 0, 0}, + (__m512i)(__v8du){0xFFFFFFFFFFFFFull, 0, 0, 0, + 0, 0, 0, 0}, + (__m512i)(__v8du){0xFFFFFFFFFFFFFull, 0, 0, 0, + 0, 0, 0, 0}), + 0xFFFFFFFFFFFFEull, 0, 0, 0, 0, 0, 0, 0)); + +TEST_CONSTEXPR(match_v8di(_mm512_madd52hi_epu64( + (__m512i)(__v8du){1, 2, 3, 4, 5, 6, 7, 8}, + (__m512i)(__v8du){0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull}, + (__m512i)(__v8du){0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull}), + 4503599627370495ull, 4503599627370496ull, + 4503599627370497ull, 4503599627370498ull, + 4503599627370499ull, 4503599627370500ull, + 4503599627370501ull, 4503599627370502ull)); + __m512i test_mm512_mask_madd52hi_epu64(__m512i __W, __mmask8 __M, __m512i __X, __m512i __Y) { // CHECK-LABEL: test_mm512_mask_madd52hi_epu64 // CHECK: call {{.*}}<8 x i64> @llvm.x86.avx512.vpmadd52h.uq.512(<8 x i64> %{{.*}}, <8 x i64> %{{.*}}, <8 x i64> %{{.*}}) // CHECK: select <8 x i1> %{{.*}}, <8 x i64> %{{.*}}, <8 x i64> %{{.*}} - return _mm512_mask_madd52hi_epu64(__W, __M, __X, __Y); + return _mm512_mask_madd52hi_epu64(__W, __M, __X, __Y); } +TEST_CONSTEXPR(match_v8di(_mm512_mask_madd52hi_epu64( + (__m512i)(__v8du){111, 222, 333, 444, 555, 666, + 777, 888}, + 0x00, + (__m512i)(__v8du){1, 2, 3, 4, 5, 6, 7, 8}, + (__m512i)(__v8du){10, 20, 30, 40, 50, 60, 70, + 80}), + 111, 222, 333, 444, 555, 666, 777, 888)); + +TEST_CONSTEXPR(match_v8di(_mm512_mask_madd52hi_epu64( + (__m512i)(__v8du){10, 20, 30, 40, 50, 60, 70, + 80}, + 0xFF, + (__m512i)(__v8du){100, 200, 300, 400, 500, 600, + 700, 800}, + (__m512i)(__v8du){10, 20, 30, 40, 50, 60, 70, + 80}), + 10, 20, 30, 40, 50, 60, 70, 80)); + __m512i test_mm512_maskz_madd52hi_epu64(__mmask8 __M, __m512i __X, __m512i __Y, __m512i __Z) { // CHECK-LABEL: test_mm512_maskz_madd52hi_epu64 // CHECK: call {{.*}}<8 x i64> @llvm.x86.avx512.vpmadd52h.uq.512(<8 x i64> %{{.*}}, <8 x i64> %{{.*}}, <8 x i64> %{{.*}}) // CHECK: select <8 x i1> %{{.*}}, <8 x i64> %{{.*}}, <8 x i64> %{{.*}} - return _mm512_maskz_madd52hi_epu64(__M, __X, __Y, __Z); + return _mm512_maskz_madd52hi_epu64(__M, __X, __Y, __Z); } +TEST_CONSTEXPR(match_v8di(_mm512_maskz_madd52hi_epu64( + 0x00, + (__m512i)(__v8du){1, 2, 3, 4, 5, 6, 7, 8}, + (__m512i)(__v8du){10, 20, 30, 40, 50, 60, 70, + 80}, + (__m512i)(__v8du){100, 200, 300, 400, 500, 600, + 700, 800}), + 0, 0, 0, 0, 0, 0, 0, 0)); + +TEST_CONSTEXPR(match_v8di(_mm512_maskz_madd52hi_epu64( + 0xFF, + (__m512i)(__v8du){1, 2, 3, 4, 5, 6, 7, 8}, + (__m512i)(__v8du){10, 20, 30, 40, 50, 60, 70, + 80}, + (__m512i)(__v8du){100, 200, 300, 400, 500, 600, + 700, 800}), + 1, 2, 3, 4, 5, 6, 7, 8)); + __m512i test_mm512_madd52lo_epu64(__m512i __X, __m512i __Y, __m512i __Z) { // CHECK-LABEL: test_mm512_madd52lo_epu64 // CHECK: call {{.*}}<8 x i64> @llvm.x86.avx512.vpmadd52l.uq.512(<8 x i64> %{{.*}}, <8 x i64> %{{.*}}, <8 x i64> %{{.*}}) - return _mm512_madd52lo_epu64(__X, __Y, __Z); + return _mm512_madd52lo_epu64(__X, __Y, __Z); } +TEST_CONSTEXPR(match_v8di(_mm512_madd52lo_epu64( + (__m512i)(__v8du){0, 0, 0, 0, 0, 0, 0, 0}, + (__m512i)(__v8du){10, 0, 0, 0, 0, 0, 0, 0}, + (__m512i)(__v8du){5, 0, 0, 0, 0, 0, 0, 0}), + 50, 0, 0, 0, 0, 0, 0, 0)); + +TEST_CONSTEXPR(match_v8di(_mm512_madd52lo_epu64( + (__m512i)(__v8du){100, 0, 0, 0, 0, 0, 0, 0}, + (__m512i)(__v8du){20, 0, 0, 0, 0, 0, 0, 0}, + (__m512i)(__v8du){30, 0, 0, 0, 0, 0, 0, 0}), + 700, 0, 0, 0, 0, 0, 0, 0)); + +TEST_CONSTEXPR(match_v8di(_mm512_madd52lo_epu64( + (__m512i)(__v8du){0, 0, 0, 0, 0, 0, 0, 0}, + (__m512i)(__v8du){0xFFFFFFFFFFFFFull, 0, 0, 0, + 0, 0, 0, 0}, + (__m512i)(__v8du){1, 0, 0, 0, 0, 0, 0, 0}), + 0xFFFFFFFFFFFFFull, 0, 0, 0, 0, 0, 0, 0)); + +TEST_CONSTEXPR(match_v8di(_mm512_madd52lo_epu64( + (__m512i)(__v8du){0, 0, 0, 0, 0, 0, 0, 0}, + (__m512i)(__v8du){0x1F000000000000ull, 0, 0, 0, + 0, 0, 0, 0}, + (__m512i)(__v8du){2, 0, 0, 0, 0, 0, 0, 0}), + 0xE000000000000ull, 0, 0, 0, 0, 0, 0, 0)); + +TEST_CONSTEXPR(match_v8di(_mm512_madd52lo_epu64( + (__m512i)(__v8du){1, 2, 3, 4, 5, 6, 7, 8}, + (__m512i)(__v8du){10, 20, 30, 40, 50, 60, 70, + 80}, + (__m512i)(__v8du){2, 3, 4, 5, 6, 7, 8, 9}), + 21, 62, 123, 204, 305, 426, 567, 728)); + +TEST_CONSTEXPR(match_v8di(_mm512_madd52lo_epu64( + (__m512i)(__v8du){0xFFFFFFFFFFFFFull, 0, 0, 0, + 0, 0, 0, 0}, + (__m512i)(__v8du){10, 0, 0, 0, 0, 0, 0, 0}, + (__m512i)(__v8du){5, 0, 0, 0, 0, 0, 0, 0}), + 4503599627370545ull, 0, 0, 0, 0, 0, 0, 0)); + +TEST_CONSTEXPR(match_v8di(_mm512_madd52lo_epu64( + (__m512i)(__v8du){10, 20, 30, 40, 50, 60, 70, + 80}, + (__m512i)(__v8du){100, 200, 300, 400, 500, 600, + 700, 800}, + (__m512i)(__v8du){2, 3, 4, 5, 6, 7, 8, 9}), + 210, 620, 1230, 2040, 3050, 4260, 5670, 7280)); + +TEST_CONSTEXPR(match_v8di(_mm512_madd52lo_epu64( + (__m512i)(__v8du){0, 0, 0, 0, 0, 0, 0, 0}, + (__m512i)(__v8du){0x1F000000000000ull, + 0x1F000000000000ull, 0, 0, 0, + 0, 0, 0}, + (__m512i)(__v8du){2, 3, 0, 0, 0, 0, 0, 0}), + 0xE000000000000ull, 0xD000000000000ull, 0, 0, 0, 0, + 0, 0)); + +TEST_CONSTEXPR(match_v8di(_mm512_madd52lo_epu64( + (__m512i)(__v8du){0, 0, 0, 0, 0, 0, 0, 0}, + (__m512i)(__v8du){0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull}, + (__m512i)(__v8du){1, 1, 1, 1, 1, 1, 1, 1}), + 0xFFFFFFFFFFFFFull, 0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull, 0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull, 0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull, 0xFFFFFFFFFFFFFull)); + __m512i test_mm512_mask_madd52lo_epu64(__m512i __W, __mmask8 __M, __m512i __X, __m512i __Y) { // CHECK-LABEL: test_mm512_mask_madd52lo_epu64 // CHECK: call {{.*}}<8 x i64> @llvm.x86.avx512.vpmadd52l.uq.512(<8 x i64> %{{.*}}, <8 x i64> %{{.*}}, <8 x i64> %{{.*}}) // CHECK: select <8 x i1> %{{.*}}, <8 x i64> %{{.*}}, <8 x i64> %{{.*}} - return _mm512_mask_madd52lo_epu64(__W, __M, __X, __Y); + return _mm512_mask_madd52lo_epu64(__W, __M, __X, __Y); } +TEST_CONSTEXPR(match_v8di(_mm512_mask_madd52lo_epu64( + (__m512i)(__v8du){111, 222, 333, 444, 555, 666, + 777, 888}, + 0x00, + (__m512i)(__v8du){1, 2, 3, 4, 5, 6, 7, 8}, + (__m512i)(__v8du){10, 20, 30, 40, 50, 60, 70, + 80}), + 111, 222, 333, 444, 555, 666, 777, 888)); + +TEST_CONSTEXPR(match_v8di(_mm512_mask_madd52lo_epu64( + (__m512i)(__v8du){1000, 2000, 3000, 4000, 5000, + 6000, 7000, 8000}, + 0xFF, + (__m512i)(__v8du){100, 200, 300, 400, 500, 600, + 700, 800}, + (__m512i)(__v8du){20, 30, 40, 50, 60, 70, 80, + 90}), + 3000, 8000, 15000, 24000, 35000, 48000, 63000, + 80000)); + __m512i test_mm512_maskz_madd52lo_epu64(__mmask8 __M, __m512i __X, __m512i __Y, __m512i __Z) { // CHECK-LABEL: test_mm512_maskz_madd52lo_epu64 // CHECK: call {{.*}}<8 x i64> @llvm.x86.avx512.vpmadd52l.uq.512(<8 x i64> %{{.*}}, <8 x i64> %{{.*}}, <8 x i64> %{{.*}}) // CHECK: select <8 x i1> %{{.*}}, <8 x i64> %{{.*}}, <8 x i64> %{{.*}} - return _mm512_maskz_madd52lo_epu64(__M, __X, __Y, __Z); + return _mm512_maskz_madd52lo_epu64(__M, __X, __Y, __Z); } + +TEST_CONSTEXPR(match_v8di(_mm512_maskz_madd52lo_epu64( + 0x00, + (__m512i)(__v8du){1, 2, 3, 4, 5, 6, 7, 8}, + (__m512i)(__v8du){10, 20, 30, 40, 50, 60, 70, + 80}, + (__m512i)(__v8du){2, 3, 4, 5, 6, 7, 8, 9}), + 0, 0, 0, 0, 0, 0, 0, 0)); + +TEST_CONSTEXPR(match_v8di(_mm512_maskz_madd52lo_epu64( + 0xFF, + (__m512i)(__v8du){100, 200, 300, 400, 500, 600, + 700, 800}, + (__m512i)(__v8du){20, 30, 40, 50, 60, 70, 80, + 90}, + (__m512i)(__v8du){30, 40, 50, 60, 70, 80, 90, + 100}), + 700, 1400, 2300, 3400, 4700, 6200, 7900, 9800)); diff --git a/clang/test/CodeGen/X86/avx512ifmavl-builtins.c b/clang/test/CodeGen/X86/avx512ifmavl-builtins.c index 89108fc..1cbb580 100644 --- a/clang/test/CodeGen/X86/avx512ifmavl-builtins.c +++ b/clang/test/CodeGen/X86/avx512ifmavl-builtins.c @@ -8,85 +8,241 @@ // RUN: %clang_cc1 -x c++ %s -flax-vector-conversions=none -ffreestanding -triple=x86_64-apple-darwin -target-feature +avx512ifma -target-feature +avx512vl -emit-llvm -o - -Wall -Werror -fexperimental-new-constant-interpreter | FileCheck %s // RUN: %clang_cc1 -x c++ %s -flax-vector-conversions=none -ffreestanding -triple=i386-apple-darwin -target-feature +avx512ifma -target-feature +avx512vl -emit-llvm -o - -Wall -Werror -fexperimental-new-constant-interpreter | FileCheck %s - #include <immintrin.h> +#include "builtin_test_helpers.h" __m128i test_mm_madd52hi_epu64(__m128i __X, __m128i __Y, __m128i __Z) { // CHECK-LABEL: test_mm_madd52hi_epu64 // CHECK: call {{.*}}<2 x i64> @llvm.x86.avx512.vpmadd52h.uq.128(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <2 x i64> %{{.*}}) - return _mm_madd52hi_epu64(__X, __Y, __Z); + return _mm_madd52hi_epu64(__X, __Y, __Z); } +TEST_CONSTEXPR(match_v2di(_mm_madd52hi_epu64( + (__m128i)((__v2du){100, 0}), + (__m128i)((__v2du){10, 0}), + (__m128i)((__v2du){5, 0})), + 100, 0)); + +TEST_CONSTEXPR(match_v2di(_mm_madd52hi_epu64( + (__m128i)((__v2du){0, 0}), + (__m128i)((__v2du){0xFFFFFFFFFFFFFull, 0}), + (__m128i)((__v2du){0xFFFFFFFFFFFFFull, 0})), + 0xFFFFFFFFFFFFEull, 0)); + __m128i test_mm_mask_madd52hi_epu64(__m128i __W, __mmask8 __M, __m128i __X, __m128i __Y) { // CHECK-LABEL: test_mm_mask_madd52hi_epu64 // CHECK: call {{.*}}<2 x i64> @llvm.x86.avx512.vpmadd52h.uq.128(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <2 x i64> %{{.*}}) // CHECK: select <2 x i1> %{{.*}}, <2 x i64> %{{.*}}, <2 x i64> %{{.*}} - return _mm_mask_madd52hi_epu64(__W, __M, __X, __Y); + return _mm_mask_madd52hi_epu64(__W, __M, __X, __Y); } +TEST_CONSTEXPR(match_v2di(_mm_mask_madd52hi_epu64((__m128i)((__v2du){111, 222}), + 0x0, + (__m128i)((__v2du){1, 2}), + (__m128i)((__v2du){10, 20})), + 111, 222)); + +TEST_CONSTEXPR(match_v2di(_mm_mask_madd52hi_epu64((__m128i)((__v2du){10, 20}), + 0x2, + (__m128i)((__v2du){0x1000000000000ULL, 0x1000000000000ULL}), + (__m128i)((__v2du){0x1000000000000ULL, 0x1000000000000ULL})), + 10, 0x100000000014ULL)); + __m128i test_mm_maskz_madd52hi_epu64(__mmask8 __M, __m128i __X, __m128i __Y, __m128i __Z) { // CHECK-LABEL: test_mm_maskz_madd52hi_epu64 // CHECK: call {{.*}}<2 x i64> @llvm.x86.avx512.vpmadd52h.uq.128(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <2 x i64> %{{.*}}) // CHECK: select <2 x i1> %{{.*}}, <2 x i64> %{{.*}}, <2 x i64> %{{.*}} - return _mm_maskz_madd52hi_epu64(__M, __X, __Y, __Z); + return _mm_maskz_madd52hi_epu64(__M, __X, __Y, __Z); } +TEST_CONSTEXPR(match_v2di(_mm_maskz_madd52hi_epu64(0x3, + (__m128i)((__v2du){1, 2}), + (__m128i)((__v2du){10, 20}), + (__m128i)((__v2du){100, 200})), + 1, 2)); + +TEST_CONSTEXPR(match_v2di(_mm_maskz_madd52hi_epu64(0x1, + (__m128i)((__v2du){0x1000000000000ULL, 0x1000000000000ULL}), + (__m128i)((__v2du){0x1000000000000ULL, 0x1000000000000ULL}), + (__m128i)((__v2du){0, 0})), + 0x1000000000000ULL, 0)); + __m256i test_mm256_madd52hi_epu64(__m256i __X, __m256i __Y, __m256i __Z) { // CHECK-LABEL: test_mm256_madd52hi_epu64 // CHECK: call {{.*}}<4 x i64> @llvm.x86.avx512.vpmadd52h.uq.256(<4 x i64> %{{.*}}, <4 x i64> %{{.*}}, <4 x i64> %{{.*}}) - return _mm256_madd52hi_epu64(__X, __Y, __Z); + return _mm256_madd52hi_epu64(__X, __Y, __Z); } +TEST_CONSTEXPR(match_v4di(_mm256_madd52hi_epu64( + (__m256i)((__v4du){100, 200, 300, 400}), + (__m256i)((__v4du){10, 20, 30, 40}), + (__m256i)((__v4du){5, 6, 7, 8})), + 100, 200, 300, 400)); + +TEST_CONSTEXPR(match_v4di(_mm256_madd52hi_epu64( + (__m256i)((__v4du){0, 0, 0, 0}), + (__m256i)((__v4du){0xFFFFFFFFFFFFFull, 0, 0, + 0}), + (__m256i)((__v4du){0xFFFFFFFFFFFFFull, 0, 0, + 0})), + 0xFFFFFFFFFFFFEull, 0, 0, 0)); + __m256i test_mm256_mask_madd52hi_epu64(__m256i __W, __mmask8 __M, __m256i __X, __m256i __Y) { // CHECK-LABEL: test_mm256_mask_madd52hi_epu64 // CHECK: call {{.*}}<4 x i64> @llvm.x86.avx512.vpmadd52h.uq.256(<4 x i64> %{{.*}}, <4 x i64> %{{.*}}, <4 x i64> %{{.*}}) // CHECK: select <4 x i1> %{{.*}}, <4 x i64> %{{.*}}, <4 x i64> %{{.*}} - return _mm256_mask_madd52hi_epu64(__W, __M, __X, __Y); + return _mm256_mask_madd52hi_epu64(__W, __M, __X, __Y); } +TEST_CONSTEXPR(match_v4di(_mm256_mask_madd52hi_epu64((__m256i)((__v4du){111, 222, 333, 444}), + 0x0, + (__m256i)((__v4du){1, 2, 3, 4}), + (__m256i)((__v4du){10, 20, 30, 40})), + 111, 222, 333, 444)); + +TEST_CONSTEXPR(match_v4di(_mm256_mask_madd52hi_epu64((__m256i)((__v4du){10, 20, 30, 40}), + 0xA, + (__m256i)((__v4du){0x1000000000000ULL, 0x1000000000000ULL, + 0x1000000000000ULL, 0x1000000000000ULL}), + (__m256i)((__v4du){0x1000000000000ULL, 0x1000000000000ULL, + 0x1000000000000ULL, 0x1000000000000ULL})), + 10, 0x100000000014ULL, 30, 0x100000000028ULL)); + __m256i test_mm256_maskz_madd52hi_epu64(__mmask8 __M, __m256i __X, __m256i __Y, __m256i __Z) { // CHECK-LABEL: test_mm256_maskz_madd52hi_epu64 // CHECK: call {{.*}}<4 x i64> @llvm.x86.avx512.vpmadd52h.uq.256(<4 x i64> %{{.*}}, <4 x i64> %{{.*}}, <4 x i64> %{{.*}}) // CHECK: select <4 x i1> %{{.*}}, <4 x i64> %{{.*}}, <4 x i64> %{{.*}} - return _mm256_maskz_madd52hi_epu64(__M, __X, __Y, __Z); + return _mm256_maskz_madd52hi_epu64(__M, __X, __Y, __Z); } +TEST_CONSTEXPR(match_v4di(_mm256_maskz_madd52hi_epu64(0xF, + (__m256i)((__v4du){1, 2, 3, 4}), + (__m256i)((__v4du){10, 20, 30, 40}), + (__m256i)((__v4du){100, 200, 300, 400})), + 1, 2, 3, 4)); + +TEST_CONSTEXPR(match_v4di(_mm256_maskz_madd52hi_epu64(0x5, + (__m256i)((__v4du){0x1000000000000ULL, 0x1000000000000ULL, + 0x1000000000000ULL, 0x1000000000000ULL}), + (__m256i)((__v4du){0x1000000000000ULL, 0x1000000000000ULL, + 0x1000000000000ULL, 0x1000000000000ULL}), + (__m256i)((__v4du){0, 0, 0, 0})), + 0x1000000000000ULL, 0, 0x1000000000000ULL, 0)); + __m128i test_mm_madd52lo_epu64(__m128i __X, __m128i __Y, __m128i __Z) { // CHECK-LABEL: test_mm_madd52lo_epu64 // CHECK: call {{.*}}<2 x i64> @llvm.x86.avx512.vpmadd52l.uq.128(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <2 x i64> %{{.*}}) - return _mm_madd52lo_epu64(__X, __Y, __Z); + return _mm_madd52lo_epu64(__X, __Y, __Z); } +TEST_CONSTEXPR(match_v2di(_mm_madd52lo_epu64( + (__m128i)((__v2du){0, 0}), + (__m128i)((__v2du){10, 0}), + (__m128i)((__v2du){5, 0})), + 50, 0)); + +TEST_CONSTEXPR(match_v2di(_mm_madd52lo_epu64( + (__m128i)((__v2du){100, 0}), + (__m128i)((__v2du){20, 0}), + (__m128i)((__v2du){30, 0})), + 700, 0)); + +TEST_CONSTEXPR(match_v2di(_mm_madd52lo_epu64( + (__m128i)((__v2du){1, 2}), + (__m128i)((__v2du){10, 20}), + (__m128i)((__v2du){2, 3})), + 21, 62)); + __m128i test_mm_mask_madd52lo_epu64(__m128i __W, __mmask8 __M, __m128i __X, __m128i __Y) { // CHECK-LABEL: test_mm_mask_madd52lo_epu64 // CHECK: call {{.*}}<2 x i64> @llvm.x86.avx512.vpmadd52l.uq.128(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <2 x i64> %{{.*}}) // CHECK: select <2 x i1> %{{.*}}, <2 x i64> %{{.*}}, <2 x i64> %{{.*}} - return _mm_mask_madd52lo_epu64(__W, __M, __X, __Y); + return _mm_mask_madd52lo_epu64(__W, __M, __X, __Y); } +TEST_CONSTEXPR(match_v2di(_mm_mask_madd52lo_epu64((__m128i)((__v2du){1000, 2000}), + 0x3, + (__m128i)((__v2du){100, 200}), + (__m128i)((__v2du){20, 30})), + 3000, 8000)); + +TEST_CONSTEXPR(match_v2di(_mm_mask_madd52lo_epu64((__m128i)((__v2du){111, 222}), + 0x0, + (__m128i)((__v2du){1, 2}), + (__m128i)((__v2du){10, 20})), + 111, 222)); + __m128i test_mm_maskz_madd52lo_epu64(__mmask8 __M, __m128i __X, __m128i __Y, __m128i __Z) { // CHECK-LABEL: test_mm_maskz_madd52lo_epu64 // CHECK: call {{.*}}<2 x i64> @llvm.x86.avx512.vpmadd52l.uq.128(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <2 x i64> %{{.*}}) // CHECK: select <2 x i1> %{{.*}}, <2 x i64> %{{.*}}, <2 x i64> %{{.*}} - return _mm_maskz_madd52lo_epu64(__M, __X, __Y, __Z); + return _mm_maskz_madd52lo_epu64(__M, __X, __Y, __Z); } +TEST_CONSTEXPR(match_v2di(_mm_maskz_madd52lo_epu64(0x3, + (__m128i)((__v2du){100, 200}), + (__m128i)((__v2du){20, 30}), + (__m128i)((__v2du){30, 40})), + 700, 1400)); + +TEST_CONSTEXPR(match_v2di(_mm_maskz_madd52lo_epu64(0x1, + (__m128i)((__v2du){100, 0}), + (__m128i)((__v2du){20, 0}), + (__m128i)((__v2du){30, 0})), + 700, 0)); + __m256i test_mm256_madd52lo_epu64(__m256i __X, __m256i __Y, __m256i __Z) { // CHECK-LABEL: test_mm256_madd52lo_epu64 // CHECK: call {{.*}}<4 x i64> @llvm.x86.avx512.vpmadd52l.uq.256(<4 x i64> %{{.*}}, <4 x i64> %{{.*}}, <4 x i64> %{{.*}}) - return _mm256_madd52lo_epu64(__X, __Y, __Z); + return _mm256_madd52lo_epu64(__X, __Y, __Z); } +TEST_CONSTEXPR(match_v4di(_mm256_madd52lo_epu64( + (__m256i)((__v4du){1, 2, 3, 4}), + (__m256i)((__v4du){10, 20, 30, 40}), + (__m256i)((__v4du){2, 3, 4, 5})), + 21, 62, 123, 204)); + __m256i test_mm256_mask_madd52lo_epu64(__m256i __W, __mmask8 __M, __m256i __X, __m256i __Y) { // CHECK-LABEL: test_mm256_mask_madd52lo_epu64 // CHECK: call {{.*}}<4 x i64> @llvm.x86.avx512.vpmadd52l.uq.256(<4 x i64> %{{.*}}, <4 x i64> %{{.*}}, <4 x i64> %{{.*}}) // CHECK: select <4 x i1> %{{.*}}, <4 x i64> %{{.*}}, <4 x i64> %{{.*}} - return _mm256_mask_madd52lo_epu64(__W, __M, __X, __Y); + return _mm256_mask_madd52lo_epu64(__W, __M, __X, __Y); } +TEST_CONSTEXPR(match_v4di(_mm256_mask_madd52lo_epu64((__m256i)((__v4du){1000, 2000, 3000, 4000}), + 0xF, + (__m256i)((__v4du){100, 200, 300, 400}), + (__m256i)((__v4du){20, 30, 40, 50})), + 3000, 8000, 15000, 24000)); + +TEST_CONSTEXPR(match_v4di(_mm256_mask_madd52lo_epu64((__m256i)((__v4du){111, 222, 333, 444}), + 0x0, + (__m256i)((__v4du){1, 2, 3, 4}), + (__m256i)((__v4du){10, 20, 30, 40})), + 111, 222, 333, 444)); + +TEST_CONSTEXPR(match_v4di(_mm256_mask_madd52lo_epu64((__m256i)((__v4du){11, 22, 33, 44}), + 0x5, + (__m256i)((__v4du){100, 200, 300, 400}), + (__m256i)((__v4du){10, 20, 30, 40})), + 1011, 22, 9033, 44)); + __m256i test_mm256_maskz_madd52lo_epu64(__mmask8 __M, __m256i __X, __m256i __Y, __m256i __Z) { // CHECK-LABEL: test_mm256_maskz_madd52lo_epu64 // CHECK: call {{.*}}<4 x i64> @llvm.x86.avx512.vpmadd52l.uq.256(<4 x i64> %{{.*}}, <4 x i64> %{{.*}}, <4 x i64> %{{.*}}) // CHECK: select <4 x i1> %{{.*}}, <4 x i64> %{{.*}}, <4 x i64> %{{.*}} - return _mm256_maskz_madd52lo_epu64(__M, __X, __Y, __Z); + return _mm256_maskz_madd52lo_epu64(__M, __X, __Y, __Z); } + +TEST_CONSTEXPR(match_v4di(_mm256_maskz_madd52lo_epu64(0xF, + (__m256i)((__v4du){100, 200, 300, 400}), + (__m256i)((__v4du){20, 30, 40, 50}), + (__m256i)((__v4du){30, 40, 50, 60})), + 700, 1400, 2300, 3400)); + +TEST_CONSTEXPR(match_v4di(_mm256_maskz_madd52lo_epu64(0x9, + (__m256i)((__v4du){100, 200, 300, 400}), + (__m256i)((__v4du){10, 20, 30, 40}), + (__m256i)((__v4du){5, 10, 15, 20})), + 150, 0, 0, 1200)); diff --git a/clang/test/CodeGen/X86/avx512vl-builtins.c b/clang/test/CodeGen/X86/avx512vl-builtins.c index 51385d5..9b6bfea9 100644 --- a/clang/test/CodeGen/X86/avx512vl-builtins.c +++ b/clang/test/CodeGen/X86/avx512vl-builtins.c @@ -7201,6 +7201,8 @@ __m128i test_mm_mask_set1_epi32(__m128i __O, __mmask8 __M) { return _mm_mask_set1_epi32(__O, __M, 5); } +TEST_CONSTEXPR(match_v4si(_mm_mask_set1_epi32(_mm_setzero_si128(), 0xF, 7), 7, 7, 7, 7)); + __m128i test_mm_maskz_set1_epi32(__mmask8 __M) { // CHECK-LABEL: test_mm_maskz_set1_epi32 // CHECK: insertelement <4 x i32> poison, i32 %{{.*}}, i32 0 @@ -7212,6 +7214,8 @@ __m128i test_mm_maskz_set1_epi32(__mmask8 __M) { return _mm_maskz_set1_epi32(__M, 5); } +TEST_CONSTEXPR(match_v4si(_mm_maskz_set1_epi32(0xA, 11), 0, 11, 0, 11)); + __m256i test_mm256_mask_set1_epi32(__m256i __O, __mmask8 __M) { // CHECK-LABEL: test_mm256_mask_set1_epi32 // CHECK: insertelement <8 x i32> poison, i32 %{{.*}}, i32 0 @@ -7226,6 +7230,8 @@ __m256i test_mm256_mask_set1_epi32(__m256i __O, __mmask8 __M) { return _mm256_mask_set1_epi32(__O, __M, 5); } +TEST_CONSTEXPR(match_v8si(_mm256_mask_set1_epi32(_mm256_setzero_si256(), 0xAA, 5), 0, 5, 0, 5, 0, 5, 0, 5)); + __m256i test_mm256_maskz_set1_epi32(__mmask8 __M) { // CHECK-LABEL: test_mm256_maskz_set1_epi32 // CHECK: insertelement <8 x i32> poison, i32 %{{.*}}, i32 0 @@ -7240,6 +7246,8 @@ __m256i test_mm256_maskz_set1_epi32(__mmask8 __M) { return _mm256_maskz_set1_epi32(__M, 5); } +TEST_CONSTEXPR(match_v8si(_mm256_maskz_set1_epi32(0xAA, 13), 0, 13, 0, 13, 0, 13, 0, 13)); + __m128i test_mm_mask_set1_epi64(__m128i __O, __mmask8 __M, long long __A) { // CHECK-LABEL: test_mm_mask_set1_epi64 // CHECK: insertelement <2 x i64> poison, i64 %{{.*}}, i32 0 @@ -7249,6 +7257,8 @@ __m128i test_mm_mask_set1_epi64(__m128i __O, __mmask8 __M, long long __A) { return _mm_mask_set1_epi64(__O, __M, __A); } +TEST_CONSTEXPR(match_v2di(_mm_mask_set1_epi64(_mm_setzero_si128(), 0x3, 9), 9, 9)); + __m128i test_mm_maskz_set1_epi64(__mmask8 __M, long long __A) { // CHECK-LABEL: test_mm_maskz_set1_epi64 // CHECK: insertelement <2 x i64> poison, i64 %{{.*}}, i32 0 @@ -7258,6 +7268,8 @@ __m128i test_mm_maskz_set1_epi64(__mmask8 __M, long long __A) { return _mm_maskz_set1_epi64(__M, __A); } +TEST_CONSTEXPR(match_v2di(_mm_maskz_set1_epi64(0x2, 15), 0, 15)); + __m256i test_mm256_mask_set1_epi64(__m256i __O, __mmask8 __M, long long __A) { // CHECK-LABEL: test_mm256_mask_set1_epi64 // CHECK: insertelement <4 x i64> poison, i64 %{{.*}}, i32 0 @@ -7269,6 +7281,8 @@ __m256i test_mm256_mask_set1_epi64(__m256i __O, __mmask8 __M, long long __A) { return _mm256_mask_set1_epi64(__O, __M, __A); } +TEST_CONSTEXPR(match_v4di(_mm256_mask_set1_epi64(_mm256_setzero_si256(), 0xF, 11), 11, 11, 11, 11)); + __m256i test_mm256_maskz_set1_epi64(__mmask8 __M, long long __A) { // CHECK-LABEL: test_mm256_maskz_set1_epi64 // CHECK: insertelement <4 x i64> poison, i64 %{{.*}}, i32 0 @@ -7280,6 +7294,8 @@ __m256i test_mm256_maskz_set1_epi64(__mmask8 __M, long long __A) { return _mm256_maskz_set1_epi64(__M, __A); } +TEST_CONSTEXPR(match_v4di(_mm256_maskz_set1_epi64(0xA, 17), 0, 17, 0, 17)); + __m128d test_mm_fixupimm_pd(__m128d __A, __m128d __B, __m128i __C) { // CHECK-LABEL: test_mm_fixupimm_pd // CHECK: @llvm.x86.avx512.mask.fixupimm.pd.128 @@ -7623,6 +7639,8 @@ __m128d test_mm_mask_unpackhi_pd(__m128d __W, __mmask8 __U, __m128d __A, __m128d return _mm_mask_unpackhi_pd(__W, __U, __A, __B); } +TEST_CONSTEXPR(match_m128d(_mm_mask_unpackhi_pd(_mm_setzero_pd(), 0x3, (__m128d)(__v2df){1.0,2.0}, (__m128d)(__v2df){3.0,4.0}), 2.0,4.0)); + __m128d test_mm_maskz_unpackhi_pd(__mmask8 __U, __m128d __A, __m128d __B) { // CHECK-LABEL: test_mm_maskz_unpackhi_pd // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <2 x i32> <i32 1, i32 3> @@ -7637,6 +7655,8 @@ __m256d test_mm256_mask_unpackhi_pd(__m256d __W, __mmask8 __U, __m256d __A, __m2 return _mm256_mask_unpackhi_pd(__W, __U, __A, __B); } +TEST_CONSTEXPR(match_m256d(_mm256_mask_unpackhi_pd(_mm256_setzero_pd(), 0xAA, (__m256d)(__v4df){1.0,2.0,3.0,4.0}, (__m256d)(__v4df){5.0,6.0,7.0,8.0}), 0,6.0,0,8.0)); + __m256d test_mm256_maskz_unpackhi_pd(__mmask8 __U, __m256d __A, __m256d __B) { // CHECK-LABEL: test_mm256_maskz_unpackhi_pd // CHECK: shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> @@ -7679,6 +7699,8 @@ __m128d test_mm_mask_unpacklo_pd(__m128d __W, __mmask8 __U, __m128d __A, __m128d return _mm_mask_unpacklo_pd(__W, __U, __A, __B); } +TEST_CONSTEXPR(match_m128d(_mm_mask_unpacklo_pd(_mm_setzero_pd(), 0x3, (__m128d)(__v2df){1.0,2.0}, (__m128d)(__v2df){3.0,4.0}), 1.0,3.0)); + __m128d test_mm_maskz_unpacklo_pd(__mmask8 __U, __m128d __A, __m128d __B) { // CHECK-LABEL: test_mm_maskz_unpacklo_pd // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <2 x i32> <i32 0, i32 2> @@ -7686,6 +7708,8 @@ __m128d test_mm_maskz_unpacklo_pd(__mmask8 __U, __m128d __A, __m128d __B) { return _mm_maskz_unpacklo_pd(__U, __A, __B); } +TEST_CONSTEXPR(match_m128d(_mm_maskz_unpacklo_pd(0x2, (__m128d)(__v2df){1.0,2.0}, (__m128d)(__v2df){3.0,4.0}), 0.0,3.0)); + __m256d test_mm256_mask_unpacklo_pd(__m256d __W, __mmask8 __U, __m256d __A, __m256d __B) { // CHECK-LABEL: test_mm256_mask_unpacklo_pd // CHECK: shufflevector <4 x double> %{{.*}}, <4 x double> %{{.*}}, <4 x i32> <i32 0, i32 4, i32 2, i32 6> @@ -7693,6 +7717,8 @@ __m256d test_mm256_mask_unpacklo_pd(__m256d __W, __mmask8 __U, __m256d __A, __m2 return _mm256_mask_unpacklo_pd(__W, __U, __A, __B); } +TEST_CONSTEXPR(match_m256d(_mm256_mask_unpacklo_pd(_mm256_setzero_pd(), 0xAA, (__m256d)(__v4df){1.0,2.0,3.0,4.0}, (__m256d)(__v4df){5.0,6.0,7.0,8.0}), 0,5.0,0,7.0)); + __m256d test_mm256_maskz_unpacklo_pd(__mmask8 __U, __m256d __A, __m256d __B) { // CHECK-LABEL: test_mm256_maskz_unpacklo_pd // CHECK: shufflevector <4 x double> %{{.*}}, <4 x double> %{{.*}}, <4 x i32> <i32 0, i32 4, i32 2, i32 6> @@ -7700,6 +7726,8 @@ __m256d test_mm256_maskz_unpacklo_pd(__mmask8 __U, __m256d __A, __m256d __B) { return _mm256_maskz_unpacklo_pd(__U, __A, __B); } +TEST_CONSTEXPR(match_m256d(_mm256_maskz_unpacklo_pd(0x0A, (__m256d)(__v4df){1.0,2.0,3.0,4.0}, (__m256d)(__v4df){5.0,6.0,7.0,8.0}), 0.0,5.0,0.0,7.0)); + __m128 test_mm_mask_unpacklo_ps(__m128 __W, __mmask8 __U, __m128 __A, __m128 __B) { // CHECK-LABEL: test_mm_mask_unpacklo_ps // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x i32> <i32 0, i32 4, i32 1, i32 5> @@ -8039,6 +8067,8 @@ __m128i test_mm_mask_unpackhi_epi32(__m128i __W, __mmask8 __U, __m128i __A, __m1 return _mm_mask_unpackhi_epi32(__W, __U, __A, __B); } +TEST_CONSTEXPR(match_v4si(_mm_mask_unpackhi_epi32(_mm_setzero_si128(), 0xA, (__m128i)(__v4si){0,1,2,3}, (__m128i)(__v4si){4,5,6,7}), 0,6,0,7)); + __m128i test_mm_maskz_unpackhi_epi32(__mmask8 __U, __m128i __A, __m128i __B) { // CHECK-LABEL: test_mm_maskz_unpackhi_epi32 // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> <i32 2, i32 6, i32 3, i32 7> @@ -8046,6 +8076,8 @@ __m128i test_mm_maskz_unpackhi_epi32(__mmask8 __U, __m128i __A, __m128i __B) { return _mm_maskz_unpackhi_epi32(__U, __A, __B); } +TEST_CONSTEXPR(match_v4si(_mm_maskz_unpackhi_epi32(0x5, (__m128i)(__v4si){0,1,2,3}, (__m128i)(__v4si){4,5,6,7}), 2,0,3,0)); + __m256i test_mm256_mask_unpackhi_epi32(__m256i __W, __mmask8 __U, __m256i __A, __m256i __B) { // CHECK-LABEL: test_mm256_mask_unpackhi_epi32 // CHECK: shufflevector <8 x i32> %{{.*}}, <8 x i32> %{{.*}}, <8 x i32> <i32 2, i32 10, i32 3, i32 11, i32 6, i32 14, i32 7, i32 15> @@ -8997,6 +9029,8 @@ __m256 test_mm256_maskz_broadcast_f32x4(__mmask8 __M, __m128 __A) { return _mm256_maskz_broadcast_f32x4(__M, __A); } +TEST_CONSTEXPR(match_m256(_mm256_maskz_broadcast_f32x4(0xAA, (__m128)(__v4sf){0,1,2,3}), 0,1,0,3,0,1,0,3)); + __m256i test_mm256_broadcast_i32x4(__m128i const* __A) { // CHECK-LABEL: test_mm256_broadcast_i32x4 // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3> @@ -9018,6 +9052,8 @@ __m256i test_mm256_maskz_broadcast_i32x4(__mmask8 __M, __m128i const* __A) { return _mm256_maskz_broadcast_i32x4(__M, _mm_loadu_si128(__A)); } +TEST_CONSTEXPR(match_v8si(_mm256_maskz_broadcast_i32x4(0xAA, (__m128i)(__v4si){0,1,2,3}), 0,1,0,3,0,1,0,3)); + __m256d test_mm256_mask_broadcastsd_pd(__m256d __O, __mmask8 __M, __m128d __A) { // CHECK-LABEL: test_mm256_mask_broadcastsd_pd // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <4 x i32> zeroinitializer @@ -10324,6 +10360,8 @@ __m128 test_mm_mask_movehdup_ps(__m128 __W, __mmask8 __U, __m128 __A) { return _mm_mask_movehdup_ps(__W, __U, __A); } +TEST_CONSTEXPR(match_m128(_mm_mask_movehdup_ps(_mm_setzero_ps(), 0xF, (__m128)(__v4sf){1.f,2.f,3.f,4.f}), 2.f,2.f,4.f,4.f)); + __m128 test_mm_maskz_movehdup_ps(__mmask8 __U, __m128 __A) { // CHECK-LABEL: test_mm_maskz_movehdup_ps // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x i32> <i32 1, i32 1, i32 3, i32 3> @@ -10338,6 +10376,8 @@ __m256 test_mm256_mask_movehdup_ps(__m256 __W, __mmask8 __U, __m256 __A) { return _mm256_mask_movehdup_ps(__W, __U, __A); } +TEST_CONSTEXPR(match_m256(_mm256_mask_movehdup_ps(_mm256_setzero_ps(), 0xAA, (__m256)(__v8sf){1,2,3,4,5,6,7,8}), 0,2,0,4,0,6,0,8)); + __m256 test_mm256_maskz_movehdup_ps(__mmask8 __U, __m256 __A) { // CHECK-LABEL: test_mm256_maskz_movehdup_ps // CHECK: shufflevector <8 x float> %{{.*}}, <8 x float> %{{.*}}, <8 x i32> <i32 1, i32 1, i32 3, i32 3, i32 5, i32 5, i32 7, i32 7> @@ -10352,6 +10392,8 @@ __m128 test_mm_mask_moveldup_ps(__m128 __W, __mmask8 __U, __m128 __A) { return _mm_mask_moveldup_ps(__W, __U, __A); } +TEST_CONSTEXPR(match_m128(_mm_mask_moveldup_ps(_mm_setzero_ps(), 0xF, (__m128)(__v4sf){1.f,2.f,3.f,4.f}), 1.f,1.f,3.f,3.f)); + __m128 test_mm_maskz_moveldup_ps(__mmask8 __U, __m128 __A) { // CHECK-LABEL: test_mm_maskz_moveldup_ps // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x i32> <i32 0, i32 0, i32 2, i32 2> @@ -10366,6 +10408,8 @@ __m256 test_mm256_mask_moveldup_ps(__m256 __W, __mmask8 __U, __m256 __A) { return _mm256_mask_moveldup_ps(__W, __U, __A); } +TEST_CONSTEXPR(match_m256(_mm256_mask_moveldup_ps(_mm256_setzero_ps(), 0xAA, (__m256)(__v8sf){1,2,3,4,5,6,7,8}), 0,1,0,3,0,5,0,7)); + __m256 test_mm256_maskz_moveldup_ps(__mmask8 __U, __m256 __A) { // CHECK-LABEL: test_mm256_maskz_moveldup_ps // CHECK: shufflevector <8 x float> %{{.*}}, <8 x float> %{{.*}}, <8 x i32> <i32 0, i32 0, i32 2, i32 2, i32 4, i32 4, i32 6, i32 6> diff --git a/clang/test/CodeGen/X86/avx512vlbw-builtins.c b/clang/test/CodeGen/X86/avx512vlbw-builtins.c index c0e46de..d569283 100644 --- a/clang/test/CodeGen/X86/avx512vlbw-builtins.c +++ b/clang/test/CodeGen/X86/avx512vlbw-builtins.c @@ -1688,24 +1688,37 @@ __m128i test_mm_mask_shuffle_epi8(__m128i __W, __mmask16 __U, __m128i __A, __m12 // CHECK: select <16 x i1> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}} return _mm_mask_shuffle_epi8(__W,__U,__A,__B); } + +TEST_CONSTEXPR(match_v16qi(_mm_mask_shuffle_epi8((__m128i)(__v16qi){1,1,1,1,1,1,1,1,2,2,4,4,6,6,8,8}, 0x00FF, (__m128i)(__v16qi){0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15}, (__m128i)(__v16qi){15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0}), 15,14,13,12,11,10,9,8,2,2,4,4,6,6,8,8)); + __m128i test_mm_maskz_shuffle_epi8(__mmask16 __U, __m128i __A, __m128i __B) { // CHECK-LABEL: test_mm_maskz_shuffle_epi8 // CHECK: @llvm.x86.ssse3.pshuf.b // CHECK: select <16 x i1> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}} return _mm_maskz_shuffle_epi8(__U,__A,__B); } + +TEST_CONSTEXPR(match_v16qi(_mm_maskz_shuffle_epi8(0xAAAA, (__m128i)(__v16qi){0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15}, (__m128i)(__v16qi){15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0}), 0,14,0,12,0,10,0,8,0,6,0,4,0,2,0,0)); + __m256i test_mm256_mask_shuffle_epi8(__m256i __W, __mmask32 __U, __m256i __A, __m256i __B) { // CHECK-LABEL: test_mm256_mask_shuffle_epi8 // CHECK: @llvm.x86.avx2.pshuf.b // CHECK: select <32 x i1> %{{.*}}, <32 x i8> %{{.*}}, <32 x i8> %{{.*}} return _mm256_mask_shuffle_epi8(__W,__U,__A,__B); } + +TEST_CONSTEXPR(match_v32qi(_mm256_mask_shuffle_epi8((__m256i)(__v32qi){1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,3,3,3,3,3,3,3,3,4,4,4,4,4,4,4,4}, 0x80808080, (__m256i)(__v32qi){0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31}, (__m256i)(__v32qi){31,30,29,28,27,26,25,24,23,22,21,20,19,18,17,16,15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0}), 1,1,1,1,1,1,1,8,2,2,2,2,2,2,2,0,3,3,3,3,3,3,3,24,4,4,4,4,4,4,4,16)); + + __m256i test_mm256_maskz_shuffle_epi8(__mmask32 __U, __m256i __A, __m256i __B) { // CHECK-LABEL: test_mm256_maskz_shuffle_epi8 // CHECK: @llvm.x86.avx2.pshuf.b // CHECK: select <32 x i1> %{{.*}}, <32 x i8> %{{.*}}, <32 x i8> %{{.*}} return _mm256_maskz_shuffle_epi8(__U,__A,__B); } + +TEST_CONSTEXPR(match_v32qi(_mm256_maskz_shuffle_epi8(0x0000FFFF, (__m256i)(__v32qi){0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31}, (__m256i)(__v32qi){31,30,29,28,27,26,25,24,23,22,21,20,19,18,17,16,15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0}), 15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0)); + __m128i test_mm_mask_subs_epi8(__m128i __W, __mmask16 __U, __m128i __A, __m128i __B) { // CHECK-LABEL: test_mm_mask_subs_epi8 // CHECK: @llvm.ssub.sat.v16i8 diff --git a/clang/test/CodeGen/X86/avx512vlcd-builtins.c b/clang/test/CodeGen/X86/avx512vlcd-builtins.c index 1619305..29fc6fd 100644 --- a/clang/test/CodeGen/X86/avx512vlcd-builtins.c +++ b/clang/test/CodeGen/X86/avx512vlcd-builtins.c @@ -20,6 +20,7 @@ __m128i test_mm_broadcastmb_epi64(__m128i a,__m128i b) { // CHECK: insertelement <2 x i64> %{{.*}}, i64 %{{.*}}, i32 1 return _mm_broadcastmb_epi64(_mm_cmpeq_epi32_mask (a, b)); } +TEST_CONSTEXPR(match_v2du(_mm_broadcastmb_epi64((__mmask8)(76)), 76, 76)); __m256i test_mm256_broadcastmb_epi64(__m256i a, __m256i b) { // CHECK-LABEL: test_mm256_broadcastmb_epi64 @@ -32,6 +33,7 @@ __m256i test_mm256_broadcastmb_epi64(__m256i a, __m256i b) { // CHECK: insertelement <4 x i64> %{{.*}}, i64 %{{.*}}, i32 3 return _mm256_broadcastmb_epi64(_mm256_cmpeq_epi64_mask ( a, b)); } +TEST_CONSTEXPR(match_v4di(_mm256_broadcastmb_epi64((__mmask8)(67)), 67, 67, 67, 67)); __m128i test_mm_broadcastmw_epi32(__m512i a, __m512i b) { // CHECK-LABEL: test_mm_broadcastmw_epi32 @@ -43,6 +45,7 @@ __m128i test_mm_broadcastmw_epi32(__m512i a, __m512i b) { // CHECK: insertelement <4 x i32> %{{.*}}, i32 %{{.*}}, i32 3 return _mm_broadcastmw_epi32(_mm512_cmpeq_epi32_mask ( a, b)); } +TEST_CONSTEXPR(match_v4su(_mm_broadcastmw_epi32((__mmask16)(0xbabe)), 0xbabe, 0xbabe, 0xbabe, 0xbabe)); __m256i test_mm256_broadcastmw_epi32(__m512i a, __m512i b) { // CHECK-LABEL: test_mm256_broadcastmw_epi32 @@ -58,87 +61,119 @@ __m256i test_mm256_broadcastmw_epi32(__m512i a, __m512i b) { // CHECK: insertelement <8 x i32> %{{.*}}, i32 %{{.*}}, i32 7 return _mm256_broadcastmw_epi32(_mm512_cmpeq_epi32_mask ( a, b)); } +TEST_CONSTEXPR(match_v8si(_mm256_broadcastmw_epi32((__mmask16)(0xcafe)), 0xcafe,0xcafe,0xcafe,0xcafe, 0xcafe,0xcafe,0xcafe,0xcafe)); __m128i test_mm_conflict_epi64(__m128i __A) { // CHECK-LABEL: test_mm_conflict_epi64 // CHECK: call {{.*}}<2 x i64> @llvm.x86.avx512.conflict.q.128(<2 x i64> %{{.*}}) - return _mm_conflict_epi64(__A); + return _mm_conflict_epi64(__A); } +TEST_CONSTEXPR(match_v2di(_mm_conflict_epi64((__m128i)(__v2di){1, 2}), 0, 0)); +TEST_CONSTEXPR(match_v2di(_mm_conflict_epi64((__m128i)(__v2di){5, 5}), 0, 1)); + __m128i test_mm_mask_conflict_epi64(__m128i __W, __mmask8 __U, __m128i __A) { // CHECK-LABEL: test_mm_mask_conflict_epi64 // CHECK: call {{.*}}<2 x i64> @llvm.x86.avx512.conflict.q.128(<2 x i64> %{{.*}}) // CHECK: select <2 x i1> %{{.*}}, <2 x i64> %{{.*}}, <2 x i64> %{{.*}} - return _mm_mask_conflict_epi64(__W, __U, __A); + return _mm_mask_conflict_epi64(__W, __U, __A); } +TEST_CONSTEXPR(match_v2di(_mm_mask_conflict_epi64((__m128i)(__v2di){0xFF, 0xFF}, 0x2, (__m128i)(__v2di){5, 5}), 0xFF, 1)); + __m128i test_mm_maskz_conflict_epi64(__mmask8 __U, __m128i __A) { // CHECK-LABEL: test_mm_maskz_conflict_epi64 // CHECK: call {{.*}}<2 x i64> @llvm.x86.avx512.conflict.q.128(<2 x i64> %{{.*}}) // CHECK: select <2 x i1> %{{.*}}, <2 x i64> %{{.*}}, <2 x i64> %{{.*}} - return _mm_maskz_conflict_epi64(__U, __A); + return _mm_maskz_conflict_epi64(__U, __A); } +TEST_CONSTEXPR(match_v2di(_mm_maskz_conflict_epi64(0x2, (__m128i)(__v2di){5, 5}), 0, 1)); + __m256i test_mm256_conflict_epi64(__m256i __A) { // CHECK-LABEL: test_mm256_conflict_epi64 // CHECK: call {{.*}}<4 x i64> @llvm.x86.avx512.conflict.q.256(<4 x i64> %{{.*}}) - return _mm256_conflict_epi64(__A); + return _mm256_conflict_epi64(__A); } +TEST_CONSTEXPR(match_v4di(_mm256_conflict_epi64((__m256i)(__v4di){1, 2, 1, 3}), 0, 0, 1, 0)); +TEST_CONSTEXPR(match_v4di(_mm256_conflict_epi64((__m256i)(__v4di){7, 7, 7, 7}), 0, 1, 3, 7)); +TEST_CONSTEXPR(match_v4di(_mm256_conflict_epi64((__m256i)(__v4di){1, 2, 3, 4}), 0, 0, 0, 0)); + __m256i test_mm256_mask_conflict_epi64(__m256i __W, __mmask8 __U, __m256i __A) { // CHECK-LABEL: test_mm256_mask_conflict_epi64 // CHECK: call {{.*}}<4 x i64> @llvm.x86.avx512.conflict.q.256(<4 x i64> %{{.*}}) // CHECK: select <4 x i1> %{{.*}}, <4 x i64> %{{.*}}, <4 x i64> %{{.*}} - return _mm256_mask_conflict_epi64(__W, __U, __A); + return _mm256_mask_conflict_epi64(__W, __U, __A); } +TEST_CONSTEXPR(match_v4di(_mm256_mask_conflict_epi64((__m256i)(__v4di){0xFF, 0xFF, 0xFF, 0xFF}, 0x5, (__m256i)(__v4di){1, 2, 1, 3}), 0, 0xFF, 1, 0xFF)); + __m256i test_mm256_maskz_conflict_epi64(__mmask8 __U, __m256i __A) { // CHECK-LABEL: test_mm256_maskz_conflict_epi64 // CHECK: call {{.*}}<4 x i64> @llvm.x86.avx512.conflict.q.256(<4 x i64> %{{.*}}) // CHECK: select <4 x i1> %{{.*}}, <4 x i64> %{{.*}}, <4 x i64> %{{.*}} - return _mm256_maskz_conflict_epi64(__U, __A); + return _mm256_maskz_conflict_epi64(__U, __A); } +TEST_CONSTEXPR(match_v4di(_mm256_maskz_conflict_epi64(0x5, (__m256i)(__v4di){1, 2, 1, 3}), 0, 0, 1, 0)); + __m128i test_mm_conflict_epi32(__m128i __A) { // CHECK-LABEL: test_mm_conflict_epi32 // CHECK: call <4 x i32> @llvm.x86.avx512.conflict.d.128(<4 x i32> %{{.*}}) - return _mm_conflict_epi32(__A); + return _mm_conflict_epi32(__A); } +TEST_CONSTEXPR(match_v4si(_mm_conflict_epi32((__m128i)(__v4si){1, 2, 1, 3}), 0, 0, 1, 0)); +TEST_CONSTEXPR(match_v4si(_mm_conflict_epi32((__m128i)(__v4si){3, 3, 3, 3}), 0, 1, 3, 7)); +TEST_CONSTEXPR(match_v4si(_mm_conflict_epi32((__m128i)(__v4si){1, 2, 3, 4}), 0, 0, 0, 0)); + __m128i test_mm_mask_conflict_epi32(__m128i __W, __mmask8 __U, __m128i __A) { // CHECK-LABEL: test_mm_mask_conflict_epi32 // CHECK: call <4 x i32> @llvm.x86.avx512.conflict.d.128(<4 x i32> %{{.*}}) // CHECK: select <4 x i1> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}} - return _mm_mask_conflict_epi32(__W, __U, __A); + return _mm_mask_conflict_epi32(__W, __U, __A); } +TEST_CONSTEXPR(match_v4si(_mm_mask_conflict_epi32((__m128i)(__v4si){0xFF, 0xFF, 0xFF, 0xFF}, 0x5, (__m128i)(__v4si){1, 2, 1, 3}), 0, 0xFF, 1, 0xFF)); + __m128i test_mm_maskz_conflict_epi32(__mmask8 __U, __m128i __A) { // CHECK-LABEL: test_mm_maskz_conflict_epi32 // CHECK: call <4 x i32> @llvm.x86.avx512.conflict.d.128(<4 x i32> %{{.*}}) // CHECK: select <4 x i1> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}} - return _mm_maskz_conflict_epi32(__U, __A); + return _mm_maskz_conflict_epi32(__U, __A); } +TEST_CONSTEXPR(match_v4si(_mm_maskz_conflict_epi32(0x5, (__m128i)(__v4si){1, 2, 1, 3}), 0, 0, 1, 0)); + __m256i test_mm256_conflict_epi32(__m256i __A) { // CHECK-LABEL: test_mm256_conflict_epi32 // CHECK: call <8 x i32> @llvm.x86.avx512.conflict.d.256(<8 x i32> %{{.*}}) - return _mm256_conflict_epi32(__A); + return _mm256_conflict_epi32(__A); } +TEST_CONSTEXPR(match_v8si(_mm256_conflict_epi32((__m256i)(__v8si){1, 2, 1, 3, 2, 4, 1, 5}), 0, 0, 1, 0, 2, 0, 5, 0)); +TEST_CONSTEXPR(match_v8si(_mm256_conflict_epi32((__m256i)(__v8si){4, 4, 4, 4, 4, 4, 4, 4}), 0, 1, 3, 7, 15, 31, 63, 127)); +TEST_CONSTEXPR(match_v8si(_mm256_conflict_epi32((__m256i)(__v8si){1, 2, 3, 4, 5, 6, 7, 8}), 0, 0, 0, 0, 0, 0, 0, 0)); + __m256i test_mm256_mask_conflict_epi32(__m256i __W, __mmask8 __U, __m256i __A) { // CHECK-LABEL: test_mm256_mask_conflict_epi32 // CHECK: call <8 x i32> @llvm.x86.avx512.conflict.d.256(<8 x i32> %{{.*}}) // CHECK: select <8 x i1> %{{.*}}, <8 x i32> %{{.*}}, <8 x i32> %{{.*}} - return _mm256_mask_conflict_epi32(__W, __U, __A); + return _mm256_mask_conflict_epi32(__W, __U, __A); } +TEST_CONSTEXPR(match_v8si(_mm256_mask_conflict_epi32((__m256i)(__v8si){0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF}, /*0101 0101=*/0x55, (__m256i)(__v8si){1, 2, 1, 3, 2, 4, 1, 5}), 0, 0xFF, 1, 0xFF, 2, 0xFF, 5, 0xFF)); + __m256i test_mm256_maskz_conflict_epi32(__mmask8 __U, __m256i __A) { // CHECK-LABEL: test_mm256_maskz_conflict_epi32 // CHECK: call <8 x i32> @llvm.x86.avx512.conflict.d.256(<8 x i32> %{{.*}}) // CHECK: select <8 x i1> %{{.*}}, <8 x i32> %{{.*}}, <8 x i32> %{{.*}} - return _mm256_maskz_conflict_epi32(__U, __A); + return _mm256_maskz_conflict_epi32(__U, __A); } +TEST_CONSTEXPR(match_v8si(_mm256_maskz_conflict_epi32(0x55, (__m256i)(__v8si){1, 2, 1, 3, 2, 4, 1, 5}), 0, 0, 1, 0, 2, 0, 5, 0)); + __m128i test_mm_lzcnt_epi32(__m128i __A) { // CHECK-LABEL: test_mm_lzcnt_epi32 // CHECK: call <4 x i32> @llvm.ctlz.v4i32(<4 x i32> %{{.*}}, i1 true) diff --git a/clang/test/CodeGen/X86/avx512vldq-builtins.c b/clang/test/CodeGen/X86/avx512vldq-builtins.c index 9388457..4773b60 100644 --- a/clang/test/CodeGen/X86/avx512vldq-builtins.c +++ b/clang/test/CodeGen/X86/avx512vldq-builtins.c @@ -987,6 +987,8 @@ __m256 test_mm256_mask_broadcast_f32x2(__m256 __O, __mmask8 __M, __m128 __A) { return _mm256_mask_broadcast_f32x2(__O, __M, __A); } +TEST_CONSTEXPR(match_m256(_mm256_mask_broadcast_f32x2(_mm256_setzero_ps(), 0xAA, (__m128)(__v4sf){1.f,2.f,3.f,4.f}), 0,2.f,0,2.f,0,2.f,0,2.f)); + __m256 test_mm256_maskz_broadcast_f32x2(__mmask8 __M, __m128 __A) { // CHECK-LABEL: test_mm256_maskz_broadcast_f32x2 // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1> @@ -994,6 +996,8 @@ __m256 test_mm256_maskz_broadcast_f32x2(__mmask8 __M, __m128 __A) { return _mm256_maskz_broadcast_f32x2(__M, __A); } +TEST_CONSTEXPR(match_m256(_mm256_maskz_broadcast_f32x2(0xAA, (__m128)(__v4sf){1.f,2.f,3.f,4.f}), 0,2.f,0,2.f,0,2.f,0,2.f)); + __m256d test_mm256_broadcast_f64x2(double const* __A) { // CHECK-LABEL: test_mm256_broadcast_f64x2 // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 0, i32 1> @@ -1008,6 +1012,8 @@ __m256d test_mm256_mask_broadcast_f64x2(__m256d __O, __mmask8 __M, double const* return _mm256_mask_broadcast_f64x2(__O, __M, _mm_loadu_pd(__A)); } +TEST_CONSTEXPR(match_m256d(_mm256_mask_broadcast_f64x2(_mm256_setzero_pd(), 0xA, (__m128d)(__v2df){1.0,2.0}), 0,2.0,0,2.0)); + __m256d test_mm256_maskz_broadcast_f64x2(__mmask8 __M, double const* __A) { // CHECK-LABEL: test_mm256_maskz_broadcast_f64x2 // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 0, i32 1> @@ -1015,6 +1021,8 @@ __m256d test_mm256_maskz_broadcast_f64x2(__mmask8 __M, double const* __A) { return _mm256_maskz_broadcast_f64x2(__M, _mm_loadu_pd(__A)); } +TEST_CONSTEXPR(match_m256d(_mm256_maskz_broadcast_f64x2(0xA, (__m128d)(__v2df){1.0,2.0}), 0,2.0,0,2.0)); + __m128i test_mm_broadcast_i32x2(__m128i __A) { // CHECK-LABEL: test_mm_broadcast_i32x2 // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 0, i32 1> @@ -1029,6 +1037,8 @@ __m128i test_mm_mask_broadcast_i32x2(__m128i __O, __mmask8 __M, __m128i __A) { return _mm_mask_broadcast_i32x2(__O, __M, __A); } +TEST_CONSTEXPR(match_v4si(_mm_mask_broadcast_i32x2(_mm_setzero_si128(), 0xF, (__m128i)(__v4si){0,1,2,3}), 0,1,0,1)); + __m128i test_mm_maskz_broadcast_i32x2(__mmask8 __M, __m128i __A) { // CHECK-LABEL: test_mm_maskz_broadcast_i32x2 // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 0, i32 1> @@ -1036,6 +1046,8 @@ __m128i test_mm_maskz_broadcast_i32x2(__mmask8 __M, __m128i __A) { return _mm_maskz_broadcast_i32x2(__M, __A); } +TEST_CONSTEXPR(match_v4si(_mm_maskz_broadcast_i32x2(0xF, (__m128i)(__v4si){0,1,2,3}), 0,1,0,1)); + __m256i test_mm256_broadcast_i32x2(__m128i __A) { // CHECK-LABEL: test_mm256_broadcast_i32x2 // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1> @@ -1050,6 +1062,8 @@ __m256i test_mm256_mask_broadcast_i32x2(__m256i __O, __mmask8 __M, __m128i __A) return _mm256_mask_broadcast_i32x2(__O, __M, __A); } +TEST_CONSTEXPR(match_v8si(_mm256_mask_broadcast_i32x2(_mm256_setzero_si256(), 0xAA, (__m128i)(__v4si){0,1,2,3}), 0,1,0,1,0,1,0,1)); + __m256i test_mm256_maskz_broadcast_i32x2(__mmask8 __M, __m128i __A) { // CHECK-LABEL: test_mm256_maskz_broadcast_i32x2 // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1> @@ -1057,6 +1071,8 @@ __m256i test_mm256_maskz_broadcast_i32x2(__mmask8 __M, __m128i __A) { return _mm256_maskz_broadcast_i32x2(__M, __A); } +TEST_CONSTEXPR(match_v8si(_mm256_maskz_broadcast_i32x2(0xAA, (__m128i)(__v4si){0,1,2,3}), 0,1,0,1,0,1,0,1)); + __m256i test_mm256_broadcast_i64x2(__m128i const* __A) { // CHECK-LABEL: test_mm256_broadcast_i64x2 // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 0, i32 1> @@ -1078,6 +1094,8 @@ __m256i test_mm256_maskz_broadcast_i64x2(__mmask8 __M, __m128i const* __A) { return _mm256_maskz_broadcast_i64x2(__M, _mm_loadu_si128(__A)); } +TEST_CONSTEXPR(match_v4di(_mm256_maskz_broadcast_i64x2(0xF, (__m128i)(__v2di){1,2}), 1,2,1,2)); + __m128d test_mm256_extractf64x2_pd(__m256d __A) { // CHECK-LABEL: test_mm256_extractf64x2_pd // CHECK: shufflevector <4 x double> %{{.*}}, <4 x double> poison, <2 x i32> <i32 2, i32 3> diff --git a/clang/test/CodeGen/X86/avxifma-builtins.c b/clang/test/CodeGen/X86/avxifma-builtins.c index aa15159..70531da 100644 --- a/clang/test/CodeGen/X86/avxifma-builtins.c +++ b/clang/test/CodeGen/X86/avxifma-builtins.c @@ -8,8 +8,9 @@ // RUN: %clang_cc1 -x c++ -ffreestanding %s -triple=x86_64-apple-darwin -target-feature +avxifma -emit-llvm -o - -Wall -Werror -fexperimental-new-constant-interpreter | FileCheck %s // RUN: %clang_cc1 -x c++ -ffreestanding %s -triple=i386-apple-darwin -target-feature +avxifma -emit-llvm -o - -Wall -Werror -fexperimental-new-constant-interpreter | FileCheck %s - #include <immintrin.h> +#include "builtin_test_helpers.h" + __m128i test_mm_madd52hi_epu64(__m128i __X, __m128i __Y, __m128i __Z) { // CHECK-LABEL: test_mm_madd52hi_epu64 @@ -17,44 +18,207 @@ __m128i test_mm_madd52hi_epu64(__m128i __X, __m128i __Y, __m128i __Z) { return _mm_madd52hi_epu64(__X, __Y, __Z); } +TEST_CONSTEXPR(match_v2di(_mm_madd52hi_epu64( + (__m128i)((__v2du){50, 100}), + (__m128i)((__v2du){10, 20}), + (__m128i)((__v2du){5, 6})), + 50, 100)); + +TEST_CONSTEXPR(match_v2di(_mm_madd52hi_epu64( + (__m128i)((__v2du){0, 0}), + (__m128i)((__v2du){0xFFFFFFFFFFFFFull, 0}), + (__m128i)((__v2du){0xFFFFFFFFFFFFFull, 0})), + 0xFFFFFFFFFFFFEull, 0)); + __m256i test_mm256_madd52hi_epu64(__m256i __X, __m256i __Y, __m256i __Z) { // CHECK-LABEL: test_mm256_madd52hi_epu64 // CHECK: call {{.*}}<4 x i64> @llvm.x86.avx512.vpmadd52h.uq.256(<4 x i64> %{{.*}}, <4 x i64> %{{.*}}, <4 x i64> %{{.*}}) return _mm256_madd52hi_epu64(__X, __Y, __Z); } +TEST_CONSTEXPR(match_v4di(_mm256_madd52hi_epu64( + (__m256i)((__v4du){100, 200, 300, 400}), + (__m256i)((__v4du){10, 20, 30, 40}), + (__m256i)((__v4du){5, 6, 7, 8})), + 100, 200, 300, 400)); + +TEST_CONSTEXPR(match_v4di(_mm256_madd52hi_epu64( + (__m256i)((__v4du){0, 0, 0, 0}), + (__m256i)((__v4du){0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull, 0, 0}), + (__m256i)((__v4du){0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull, 0, 0})), + 0xFFFFFFFFFFFFEull, 0xFFFFFFFFFFFFEull, 0, 0)); + __m128i test_mm_madd52lo_epu64(__m128i __X, __m128i __Y, __m128i __Z) { // CHECK-LABEL: test_mm_madd52lo_epu64 // CHECK: call {{.*}}<2 x i64> @llvm.x86.avx512.vpmadd52l.uq.128(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <2 x i64> %{{.*}}) return _mm_madd52lo_epu64(__X, __Y, __Z); } +TEST_CONSTEXPR(match_v2di(_mm_madd52lo_epu64( + (__m128i)((__v2du){0, 0}), + (__m128i)((__v2du){10, 0}), + (__m128i)((__v2du){5, 0})), + 50, 0)); + +TEST_CONSTEXPR(match_v2di(_mm_madd52lo_epu64( + (__m128i)((__v2du){1, 2}), + (__m128i)((__v2du){10, 20}), + (__m128i)((__v2du){2, 3})), + 21, 62)); + +TEST_CONSTEXPR(match_v2di(_mm_madd52lo_epu64( + (__m128i)((__v2du){0, 0}), + (__m128i)((__v2du){0xFFFFFFFFFFFFFull, 0}), + (__m128i)((__v2du){1, 0})), + 0xFFFFFFFFFFFFFull, 0)); + __m256i test_mm256_madd52lo_epu64(__m256i __X, __m256i __Y, __m256i __Z) { // CHECK-LABEL: test_mm256_madd52lo_epu64 // CHECK: call {{.*}}<4 x i64> @llvm.x86.avx512.vpmadd52l.uq.256(<4 x i64> %{{.*}}, <4 x i64> %{{.*}}, <4 x i64> %{{.*}}) return _mm256_madd52lo_epu64(__X, __Y, __Z); } +TEST_CONSTEXPR(match_v4di(_mm256_madd52lo_epu64( + (__m256i)((__v4du){1, 2, 3, 4}), + (__m256i)((__v4du){10, 20, 30, 40}), + (__m256i)((__v4du){2, 3, 4, 5})), + 21, 62, 123, 204)); + +TEST_CONSTEXPR(match_v4di(_mm256_madd52lo_epu64( + (__m256i)((__v4du){0, 0, 0, 0}), + (__m256i)((__v4du){0xFFFFFFFFFFFFFull, 0, 0, + 0}), + (__m256i)((__v4du){1, 0, 0, 0})), + 0xFFFFFFFFFFFFFull, 0, 0, 0)); + +TEST_CONSTEXPR(match_v4di(_mm256_madd52lo_epu64( + (__m256i)((__v4du){0, 0, 0, 0}), + (__m256i)((__v4du){0x1F000000000000ull, 0, 0, + 0}), + (__m256i)((__v4du){2, 0, 0, 0})), + 0xE000000000000ull, 0, 0, 0)); + __m128i test_mm_madd52hi_avx_epu64(__m128i __X, __m128i __Y, __m128i __Z) { -// CHECK-LABEL: test_mm_madd52hi_avx_epu64 -// CHECK: call {{.*}}<2 x i64> @llvm.x86.avx512.vpmadd52h.uq.128(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <2 x i64> %{{.*}}) + // CHECK-LABEL: test_mm_madd52hi_avx_epu64 + // CHECK: call {{.*}}<2 x i64> @llvm.x86.avx512.vpmadd52h.uq.128(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <2 x i64> %{{.*}}) return _mm_madd52hi_avx_epu64(__X, __Y, __Z); } +TEST_CONSTEXPR(match_v2di(_mm_madd52hi_avx_epu64( + (__m128i)((__v2du){50, 100}), + (__m128i)((__v2du){10, 20}), + (__m128i)((__v2du){5, 6})), + 50, 100)); + +TEST_CONSTEXPR(match_v2di(_mm_madd52hi_avx_epu64( + (__m128i)((__v2du){100, 0}), + (__m128i)((__v2du){10, 0}), + (__m128i)((__v2du){5, 0})), + 100, 0)); + +TEST_CONSTEXPR(match_v2di(_mm_madd52hi_avx_epu64( + (__m128i)((__v2du){0, 0}), + (__m128i)((__v2du){0xFFFFFFFFFFFFFull, 0}), + (__m128i)((__v2du){0xFFFFFFFFFFFFFull, 0})), + 0xFFFFFFFFFFFFEull, 0)); + __m256i test_mm256_madd52hi_avx_epu64(__m256i __X, __m256i __Y, __m256i __Z) { -// CHECK-LABEL: test_mm256_madd52hi_avx_epu64 -// CHECK: call {{.*}}<4 x i64> @llvm.x86.avx512.vpmadd52h.uq.256(<4 x i64> %{{.*}}, <4 x i64> %{{.*}}, <4 x i64> %{{.*}}) + // CHECK-LABEL: test_mm256_madd52hi_avx_epu64 + // CHECK: call {{.*}}<4 x i64> @llvm.x86.avx512.vpmadd52h.uq.256(<4 x i64> %{{.*}}, <4 x i64> %{{.*}}, <4 x i64> %{{.*}}) return _mm256_madd52hi_avx_epu64(__X, __Y, __Z); } +TEST_CONSTEXPR(match_v4di(_mm256_madd52hi_avx_epu64( + (__m256i)((__v4du){0, 0, 0, 0}), + (__m256i)((__v4du){0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull, 0, 0}), + (__m256i)((__v4du){0xFFFFFFFFFFFFFull, + 0xFFFFFFFFFFFFFull, 0, 0})), + 0xFFFFFFFFFFFFEull, 0xFFFFFFFFFFFFEull, 0, 0)); + +TEST_CONSTEXPR(match_v4di(_mm256_madd52hi_avx_epu64( + (__m256i)((__v4du){100, 200, 300, 400}), + (__m256i)((__v4du){10, 20, 30, 40}), + (__m256i)((__v4du){5, 6, 7, 8})), + 100, 200, 300, 400)); + +TEST_CONSTEXPR(match_v4di(_mm256_madd52hi_avx_epu64( + (__m256i)((__v4du){0, 0, 0, 0}), + (__m256i)((__v4du){0xFFFFFFFFFFFFFull, 0, 0, + 0}), + (__m256i)((__v4du){0xFFFFFFFFFFFFFull, 0, 0, + 0})), + 0xFFFFFFFFFFFFEull, 0, 0, 0)); + __m128i test_mm_madd52lo_avx_epu64(__m128i __X, __m128i __Y, __m128i __Z) { -// CHECK-LABEL: test_mm_madd52lo_avx_epu64 -// CHECK: call {{.*}}<2 x i64> @llvm.x86.avx512.vpmadd52l.uq.128(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <2 x i64> %{{.*}}) + // CHECK-LABEL: test_mm_madd52lo_avx_epu64 + // CHECK: call {{.*}}<2 x i64> @llvm.x86.avx512.vpmadd52l.uq.128(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <2 x i64> %{{.*}}) return _mm_madd52lo_avx_epu64(__X, __Y, __Z); } +TEST_CONSTEXPR(match_v2di(_mm_madd52lo_avx_epu64( + (__m128i)((__v2du){0, 0}), + (__m128i)((__v2du){10, 0}), + (__m128i)((__v2du){5, 0})), + 50, 0)); + +TEST_CONSTEXPR(match_v2di(_mm_madd52lo_avx_epu64( + (__m128i)((__v2du){100, 0}), + (__m128i)((__v2du){20, 0}), + (__m128i)((__v2du){30, 0})), + 700, 0)); + +TEST_CONSTEXPR(match_v2di(_mm_madd52lo_avx_epu64( + (__m128i)((__v2du){1, 2}), + (__m128i)((__v2du){10, 20}), + (__m128i)((__v2du){2, 3})), + 21, 62)); + +TEST_CONSTEXPR(match_v2di(_mm_madd52lo_avx_epu64( + (__m128i)((__v2du){0, 0}), + (__m128i)((__v2du){0xFFFFFFFFFFFFFull, 0}), + (__m128i)((__v2du){1, 0})), + 0xFFFFFFFFFFFFFull, 0)); + __m256i test_mm256_madd52lo_avx_epu64(__m256i __X, __m256i __Y, __m256i __Z) { -// CHECK-LABEL: test_mm256_madd52lo_avx_epu64 -// CHECK: call {{.*}}<4 x i64> @llvm.x86.avx512.vpmadd52l.uq.256(<4 x i64> %{{.*}}, <4 x i64> %{{.*}}, <4 x i64> %{{.*}}) + // CHECK-LABEL: test_mm256_madd52lo_avx_epu64 + // CHECK: call {{.*}}<4 x i64> @llvm.x86.avx512.vpmadd52l.uq.256(<4 x i64> %{{.*}}, <4 x i64> %{{.*}}, <4 x i64> %{{.*}}) return _mm256_madd52lo_avx_epu64(__X, __Y, __Z); } + +TEST_CONSTEXPR(match_v4di(_mm256_madd52lo_avx_epu64( + (__m256i)((__v4du){1, 2, 3, 4}), + (__m256i)((__v4du){10, 20, 30, 40}), + (__m256i)((__v4du){2, 3, 4, 5})), + 21, 62, 123, 204)); + + + +TEST_CONSTEXPR(match_v4di(_mm256_madd52lo_avx_epu64( + (__m256i)((__v4du){0, 0, 0, 0}), + (__m256i)((__v4du){0xFFFFFFFFFFFFFull, 0, 0, + 0}), + (__m256i)((__v4du){1, 0, 0, 0})), + 0xFFFFFFFFFFFFFull, 0, 0, 0)); + +TEST_CONSTEXPR(match_v4di(_mm256_madd52lo_avx_epu64( + (__m256i)((__v4du){0, 0, 0, 0}), + (__m256i)((__v4du){0x1F000000000000ull, 0, 0, + 0}), + (__m256i)((__v4du){2, 0, 0, 0})), + 0xE000000000000ull, 0, 0, 0)); + +TEST_CONSTEXPR(match_v2di(_mm_madd52lo_avx_epu64( + (__m128i)((__v2du){5, 10}), + (__m128i)((__v2du){100, 200}), + (__m128i)((__v2du){7, 8})), + 705, 1610)); + +TEST_CONSTEXPR(match_v4di(_mm256_madd52lo_avx_epu64( + (__m256i)((__v4du){1, 2, 3, 4}), + (__m256i)((__v4du){10, 20, 30, 40}), + (__m256i)((__v4du){2, 3, 4, 5})), + 21, 62, 123, 204)); + diff --git a/clang/test/CodeGen/X86/mmx-builtins.c b/clang/test/CodeGen/X86/mmx-builtins.c index 2b45b92..d9041d4 100644 --- a/clang/test/CodeGen/X86/mmx-builtins.c +++ b/clang/test/CodeGen/X86/mmx-builtins.c @@ -312,36 +312,42 @@ __m64 test_mm_hadd_pi16(__m64 a, __m64 b) { // CHECK: call <8 x i16> @llvm.x86.ssse3.phadd.w.128( return _mm_hadd_pi16(a, b); } +TEST_CONSTEXPR(match_v4hi(_mm_hadd_pi16((__m64)(__v4hi){1,2,3,4},(__m64)(__v4hi){5,6,7,8}),3,7,11,15)); __m64 test_mm_hadd_pi32(__m64 a, __m64 b) { // CHECK-LABEL: test_mm_hadd_pi32 // CHECK: call <4 x i32> @llvm.x86.ssse3.phadd.d.128( return _mm_hadd_pi32(a, b); } +TEST_CONSTEXPR(match_v2si(_mm_hadd_pi32((__m64)(__v2si){1,2},(__m64)(__v2si){3,4}),3,7)); __m64 test_mm_hadds_pi16(__m64 a, __m64 b) { // CHECK-LABEL: test_mm_hadds_pi16 // CHECK: call <8 x i16> @llvm.x86.ssse3.phadd.sw.128( return _mm_hadds_pi16(a, b); } +TEST_CONSTEXPR(match_v4hi(_mm_hadds_pi16((__m64)(__v4hi){32767, 32767, 1,3},(__m64)(__v4hi){-1,3, 40, 60}),32767, 4, 2,100)); __m64 test_mm_hsub_pi16(__m64 a, __m64 b) { // CHECK-LABEL: test_mm_hsub_pi16 // CHECK: call <8 x i16> @llvm.x86.ssse3.phsub.w.128( return _mm_hsub_pi16(a, b); } +TEST_CONSTEXPR(match_v4hi(_mm_hsub_pi16((__m64)(__v4hi){1,2,4,3},(__m64)(__v4hi){10,5,0,-10}),-1,1,5,10)); __m64 test_mm_hsub_pi32(__m64 a, __m64 b) { // CHECK-LABEL: test_mm_hsub_pi32 // CHECK: call <4 x i32> @llvm.x86.ssse3.phsub.d.128( return _mm_hsub_pi32(a, b); } +TEST_CONSTEXPR(match_v2si(_mm_hsub_pi32((__m64)(__v2si){1,2},(__m64)(__v2si){4,3}),-1,1)); __m64 test_mm_hsubs_pi16(__m64 a, __m64 b) { // CHECK-LABEL: test_mm_hsubs_pi16 // CHECK: call <8 x i16> @llvm.x86.ssse3.phsub.sw.128( return _mm_hsubs_pi16(a, b); } +TEST_CONSTEXPR(match_v4hi(_mm_hsubs_pi16((__m64)(__v4hi){32767, 32767, 5, -32767},(__m64)(__v4hi){4,5,10,5}),0,32767,-1,5)); __m64 test_mm_insert_pi16(__m64 a, int d) { // CHECK-LABEL: test_mm_insert_pi16 @@ -583,6 +589,8 @@ __m64 test_mm_shuffle_pi8(__m64 a, __m64 b) { return _mm_shuffle_pi8(a, b); } +TEST_CONSTEXPR(match_v8qi(_mm_shuffle_pi8((__m64)(__v8qi){0,1,2,3,4,5,6,7}, (__m64)(__v8qi){10,20,30,40,50,60,70,80}), 2,4,6,0,2,4,6,0)); + __m64 test_mm_shuffle_pi16(__m64 a) { // CHECK-LABEL: test_mm_shuffle_pi16 // CHECK: shufflevector <4 x i16> {{%.*}}, <4 x i16> {{%.*}}, <4 x i32> <i32 3, i32 0, i32 0, i32 0> diff --git a/clang/test/CodeGen/X86/sse3-builtins.c b/clang/test/CodeGen/X86/sse3-builtins.c index c53afc5..a82dd40 100644 --- a/clang/test/CodeGen/X86/sse3-builtins.c +++ b/clang/test/CodeGen/X86/sse3-builtins.c @@ -31,24 +31,28 @@ __m128d test_mm_hadd_pd(__m128d A, __m128d B) { // CHECK: call {{.*}}<2 x double> @llvm.x86.sse3.hadd.pd(<2 x double> %{{.*}}, <2 x double> %{{.*}}) return _mm_hadd_pd(A, B); } +TEST_CONSTEXPR(match_m128d(_mm_hadd_pd((__m128d){+1.0, +2.0}, (__m128d){+3.0, +4.0}), +3.0, +7.0)); __m128 test_mm_hadd_ps(__m128 A, __m128 B) { // CHECK-LABEL: test_mm_hadd_ps // CHECK: call {{.*}}<4 x float> @llvm.x86.sse3.hadd.ps(<4 x float> %{{.*}}, <4 x float> %{{.*}}) return _mm_hadd_ps(A, B); } +TEST_CONSTEXPR(match_m128(_mm_hadd_ps((__m128){+1.0f, +2.0f, +3.0f, +4.0f}, (__m128){+5.0f,+6.0f,+7.0f,+8.0f}), +3.0f, +7.0f, +11.0f, +15.0f)); __m128d test_mm_hsub_pd(__m128d A, __m128d B) { // CHECK-LABEL: test_mm_hsub_pd // CHECK: call {{.*}}<2 x double> @llvm.x86.sse3.hsub.pd(<2 x double> %{{.*}}, <2 x double> %{{.*}}) return _mm_hsub_pd(A, B); } +TEST_CONSTEXPR(match_m128d(_mm_hsub_pd((__m128d){+1.0, +2.0}, (__m128d){+4.0, +3.0}), -1.0, +1.0)); __m128 test_mm_hsub_ps(__m128 A, __m128 B) { // CHECK-LABEL: test_mm_hsub_ps // CHECK: call {{.*}}<4 x float> @llvm.x86.sse3.hsub.ps(<4 x float> %{{.*}}, <4 x float> %{{.*}}) return _mm_hsub_ps(A, B); } +TEST_CONSTEXPR(match_m128(_mm_hsub_ps((__m128){+1.0f, +2.0f, +4.0f, +3.0f}, (__m128){+5.0f,+7.0f,+10.0f,+8.0f}), -1.0f, +1.0f, -2.0f, +2.0f)); __m128i test_mm_lddqu_si128(__m128i const* P) { // CHECK-LABEL: test_mm_lddqu_si128 diff --git a/clang/test/CodeGen/X86/sse41-builtins.c b/clang/test/CodeGen/X86/sse41-builtins.c index eee479a..89a7ac2 100644 --- a/clang/test/CodeGen/X86/sse41-builtins.c +++ b/clang/test/CodeGen/X86/sse41-builtins.c @@ -449,33 +449,53 @@ int test_mm_test_all_ones(__m128i x) { // CHECK: call {{.*}}i32 @llvm.x86.sse41.ptestc(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}) return _mm_test_all_ones(x); } +TEST_CONSTEXPR(_mm_test_all_ones(((__m128i)(__v2di){-1, -1})) == 1); +TEST_CONSTEXPR(_mm_test_all_ones(((__m128i)(__v2di){-1, 0})) == 0); +TEST_CONSTEXPR(_mm_test_all_ones(((__m128i)(__v4si){-1, -1, -1, 0x7FFFFFFF})) == 0); int test_mm_test_all_zeros(__m128i x, __m128i y) { // CHECK-LABEL: test_mm_test_all_zeros // CHECK: call {{.*}}i32 @llvm.x86.sse41.ptestz(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}) return _mm_test_all_zeros(x, y); } +TEST_CONSTEXPR(_mm_test_all_zeros(((__m128i)(__v2di){0,0}), ((__m128i)(__v2di){0,0})) == 1); +TEST_CONSTEXPR(_mm_test_all_zeros(((__m128i)(__v2di){0xFF00,0}), ((__m128i)(__v2di){0x00FF,0})) == 1); +TEST_CONSTEXPR(_mm_test_all_zeros(((__m128i)(__v2di){1,0}), ((__m128i)(__v2di){-1,0})) == 0); +TEST_CONSTEXPR(_mm_test_all_zeros(((__m128i)(__v2di){0,1}), ((__m128i)(__v2di){0,-1})) == 0); int test_mm_test_mix_ones_zeros(__m128i x, __m128i y) { // CHECK-LABEL: test_mm_test_mix_ones_zeros // CHECK: call {{.*}}i32 @llvm.x86.sse41.ptestnzc(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}) return _mm_test_mix_ones_zeros(x, y); } +TEST_CONSTEXPR(_mm_test_mix_ones_zeros(((__m128i)(__v2di){0xFF, 0}), ((__m128i)(__v2di){0xF0, 1})) == 1); +TEST_CONSTEXPR(_mm_test_mix_ones_zeros(((__m128i)(__v2di){0xF0, 0}), ((__m128i)(__v2di){0x0F, 0})) == 0); +TEST_CONSTEXPR(_mm_test_mix_ones_zeros(((__m128i)(__v2di){-1, -1}), ((__m128i)(__v2di){1, 0})) == 0); +TEST_CONSTEXPR(_mm_test_mix_ones_zeros(((__m128i)(__v2di){0, 0}), ((__m128i)(__v2di){0, 0})) == 0); int test_mm_testc_si128(__m128i x, __m128i y) { // CHECK-LABEL: test_mm_testc_si128 // CHECK: call {{.*}}i32 @llvm.x86.sse41.ptestc(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}) return _mm_testc_si128(x, y); } +TEST_CONSTEXPR(_mm_testc_si128((__m128i)(__v2di){0,0}, (__m128i)(__v2di){0,0}) == 1); +TEST_CONSTEXPR(_mm_testc_si128((__m128i)(__v2di){1,0}, (__m128i)(__v2di){-1,0}) == 0); +TEST_CONSTEXPR(_mm_testc_si128((__m128i)(__v2di){0,-1}, (__m128i)(__v2di){0,1}) == 1); int test_mm_testnzc_si128(__m128i x, __m128i y) { // CHECK-LABEL: test_mm_testnzc_si128 // CHECK: call {{.*}}i32 @llvm.x86.sse41.ptestnzc(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}) return _mm_testnzc_si128(x, y); } +TEST_CONSTEXPR(_mm_testnzc_si128((__m128i)(__v2di){3,0}, (__m128i)(__v2di){1,1}) == 1); +TEST_CONSTEXPR(_mm_testnzc_si128((__m128i)(__v2di){32,-1}, (__m128i)(__v2di){15,0}) == 0); +TEST_CONSTEXPR(_mm_testnzc_si128((__m128i)(__v2di){0,999}, (__m128i)(__v2di){0,999}) == 0); int test_mm_testz_si128(__m128i x, __m128i y) { // CHECK-LABEL: test_mm_testz_si128 // CHECK: call {{.*}}i32 @llvm.x86.sse41.ptestz(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}) return _mm_testz_si128(x, y); } +TEST_CONSTEXPR(_mm_testz_si128((__m128i)(__v2di){0,0}, (__m128i)(__v2di){0,0}) == 1); +TEST_CONSTEXPR(_mm_testz_si128((__m128i)(__v2di){1,0}, (__m128i)(__v2di){-1,0}) == 0); +TEST_CONSTEXPR(_mm_testz_si128((__m128i)(__v2di){1,0}, (__m128i)(__v2di){0,1}) == 1); diff --git a/clang/test/CodeGen/X86/ssse3-builtins.c b/clang/test/CodeGen/X86/ssse3-builtins.c index 5885768..32abd9d 100644 --- a/clang/test/CodeGen/X86/ssse3-builtins.c +++ b/clang/test/CodeGen/X86/ssse3-builtins.c @@ -60,36 +60,43 @@ __m128i test_mm_hadd_epi16(__m128i a, __m128i b) { // CHECK: call <8 x i16> @llvm.x86.ssse3.phadd.w.128(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}) return _mm_hadd_epi16(a, b); } +TEST_CONSTEXPR(match_v8hi(_mm_hadd_epi16((__m128i)(__v8hi){1,2,3,4,5,6,7,8}, (__m128i)(__v8hi){17,18,19,20,21,22,23,24}), 3,7,11,15,35,39,43,47)); __m128i test_mm_hadd_epi32(__m128i a, __m128i b) { // CHECK-LABEL: test_mm_hadd_epi32 // CHECK: call <4 x i32> @llvm.x86.ssse3.phadd.d.128(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}) return _mm_hadd_epi32(a, b); } +TEST_CONSTEXPR(match_v4si(_mm_hadd_epi32((__m128i)(__v4si){1,2,3,4}, (__m128i)(__v4si){5,6,7,8}), 3,7,11,15)); __m128i test_mm_hadds_epi16(__m128i a, __m128i b) { // CHECK-LABEL: test_mm_hadds_epi16 // CHECK: call <8 x i16> @llvm.x86.ssse3.phadd.sw.128(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}) return _mm_hadds_epi16(a, b); } +TEST_CONSTEXPR(match_v8hi(_mm_hadds_epi16((__m128i)(__v8hi){30000,30000,-1,2,-3,3,1,4}, (__m128i)(__v8hi){2,6,1,9,-4,16,7,8}), 32767, 1,0,5,8,10,12,15)); + __m128i test_mm_hsub_epi16(__m128i a, __m128i b) { // CHECK-LABEL: test_mm_hsub_epi16 // CHECK: call <8 x i16> @llvm.x86.ssse3.phsub.w.128(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}) return _mm_hsub_epi16(a, b); } +TEST_CONSTEXPR(match_v8hi(_mm_hsub_epi16((__m128i)(__v8hi){20,15,16,12,9,6,4,2}, (__m128i)(__v8hi){3,2,1,1,4,5,0,2}), 5,4,3,2,1,0,-1,-2)); __m128i test_mm_hsub_epi32(__m128i a, __m128i b) { // CHECK-LABEL: test_mm_hsub_epi32 // CHECK: call <4 x i32> @llvm.x86.ssse3.phsub.d.128(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}) return _mm_hsub_epi32(a, b); } +TEST_CONSTEXPR(match_v4si(_mm_hsub_epi32((__m128i)(__v4si){4,3,1,1}, (__m128i)(__v4si){7,5,10,5}), 1,0,2,5)); __m128i test_mm_hsubs_epi16(__m128i a, __m128i b) { // CHECK-LABEL: test_mm_hsubs_epi16 // CHECK: call <8 x i16> @llvm.x86.ssse3.phsub.sw.128(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}) return _mm_hsubs_epi16(a, b); } +TEST_CONSTEXPR(match_v8hi(_mm_hsubs_epi16((__m128i)(__v8hi){32767, -15,16,12,9,6,4,2},(__m128i)(__v8hi){3,2,1,1,4,5,0,2}), 32767,4,3,2,1,0,-1,-2)); __m128i test_mm_maddubs_epi16(__m128i a, __m128i b) { // CHECK-LABEL: test_mm_maddubs_epi16 @@ -110,6 +117,8 @@ __m128i test_mm_shuffle_epi8(__m128i a, __m128i b) { return _mm_shuffle_epi8(a, b); } +TEST_CONSTEXPR(match_v16qi(_mm_shuffle_epi8((__m128i)(__v16qs){0,-1,-2,-3,-4,-5,-6,-7,-8,-9,-10,-11,-12,-13,-14,-15}, (__m128i)(__v16qs){15,-14,13,-12,11,-10,9,-8,7,-6,5,-4,3,-2,1,0}), -15,0,-13,0,-11,0,-9,0,-7,0,-5,0,-3,0,-1,0)); + __m128i test_mm_sign_epi8(__m128i a, __m128i b) { // CHECK-LABEL: test_mm_sign_epi8 // CHECK: call <16 x i8> @llvm.x86.ssse3.psign.b.128(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}) diff --git a/clang/test/CodeGen/arm-mve-intrinsics/load-store.c b/clang/test/CodeGen/arm-mve-intrinsics/load-store.c index 2dde75f..50f7011 100644 --- a/clang/test/CodeGen/arm-mve-intrinsics/load-store.c +++ b/clang/test/CodeGen/arm-mve-intrinsics/load-store.c @@ -8,8 +8,8 @@ // CHECK-LABEL: @test_vld1q_f16( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = load <8 x half>, ptr [[BASE:%.*]], align 2 -// CHECK-NEXT: ret <8 x half> [[TMP1]] +// CHECK-NEXT: [[TMP0:%.*]] = load <8 x half>, ptr [[BASE:%.*]], align 2 +// CHECK-NEXT: ret <8 x half> [[TMP0]] // float16x8_t test_vld1q_f16(const float16_t *base) { @@ -22,8 +22,8 @@ float16x8_t test_vld1q_f16(const float16_t *base) // CHECK-LABEL: @test_vld1q_f32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = load <4 x float>, ptr [[BASE:%.*]], align 4 -// CHECK-NEXT: ret <4 x float> [[TMP1]] +// CHECK-NEXT: [[TMP0:%.*]] = load <4 x float>, ptr [[BASE:%.*]], align 4 +// CHECK-NEXT: ret <4 x float> [[TMP0]] // float32x4_t test_vld1q_f32(const float32_t *base) { @@ -36,8 +36,8 @@ float32x4_t test_vld1q_f32(const float32_t *base) // CHECK-LABEL: @test_vld1q_s8( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = load <16 x i8>, ptr [[BASE:%.*]], align 1 -// CHECK-NEXT: ret <16 x i8> [[TMP1]] +// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i8>, ptr [[BASE:%.*]], align 1 +// CHECK-NEXT: ret <16 x i8> [[TMP0]] // int8x16_t test_vld1q_s8(const int8_t *base) { @@ -50,8 +50,8 @@ int8x16_t test_vld1q_s8(const int8_t *base) // CHECK-LABEL: @test_vld1q_s16( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = load <8 x i16>, ptr [[BASE:%.*]], align 2 -// CHECK-NEXT: ret <8 x i16> [[TMP1]] +// CHECK-NEXT: [[TMP0:%.*]] = load <8 x i16>, ptr [[BASE:%.*]], align 2 +// CHECK-NEXT: ret <8 x i16> [[TMP0]] // int16x8_t test_vld1q_s16(const int16_t *base) { @@ -64,8 +64,8 @@ int16x8_t test_vld1q_s16(const int16_t *base) // CHECK-LABEL: @test_vld1q_s32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = load <4 x i32>, ptr [[BASE:%.*]], align 4 -// CHECK-NEXT: ret <4 x i32> [[TMP1]] +// CHECK-NEXT: [[TMP0:%.*]] = load <4 x i32>, ptr [[BASE:%.*]], align 4 +// CHECK-NEXT: ret <4 x i32> [[TMP0]] // int32x4_t test_vld1q_s32(const int32_t *base) { @@ -78,8 +78,8 @@ int32x4_t test_vld1q_s32(const int32_t *base) // CHECK-LABEL: @test_vld1q_u8( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = load <16 x i8>, ptr [[BASE:%.*]], align 1 -// CHECK-NEXT: ret <16 x i8> [[TMP1]] +// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i8>, ptr [[BASE:%.*]], align 1 +// CHECK-NEXT: ret <16 x i8> [[TMP0]] // uint8x16_t test_vld1q_u8(const uint8_t *base) { @@ -92,8 +92,8 @@ uint8x16_t test_vld1q_u8(const uint8_t *base) // CHECK-LABEL: @test_vld1q_u16( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = load <8 x i16>, ptr [[BASE:%.*]], align 2 -// CHECK-NEXT: ret <8 x i16> [[TMP1]] +// CHECK-NEXT: [[TMP0:%.*]] = load <8 x i16>, ptr [[BASE:%.*]], align 2 +// CHECK-NEXT: ret <8 x i16> [[TMP0]] // uint16x8_t test_vld1q_u16(const uint16_t *base) { @@ -106,8 +106,8 @@ uint16x8_t test_vld1q_u16(const uint16_t *base) // CHECK-LABEL: @test_vld1q_u32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = load <4 x i32>, ptr [[BASE:%.*]], align 4 -// CHECK-NEXT: ret <4 x i32> [[TMP1]] +// CHECK-NEXT: [[TMP0:%.*]] = load <4 x i32>, ptr [[BASE:%.*]], align 4 +// CHECK-NEXT: ret <4 x i32> [[TMP0]] // uint32x4_t test_vld1q_u32(const uint32_t *base) { @@ -120,10 +120,10 @@ uint32x4_t test_vld1q_u32(const uint32_t *base) // CHECK-LABEL: @test_vld1q_z_f16( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP1]]) -// CHECK-NEXT: [[TMP3:%.*]] = call <8 x half> @llvm.masked.load.v8f16.p0(ptr [[BASE:%.*]], i32 2, <8 x i1> [[TMP2]], <8 x half> zeroinitializer) -// CHECK-NEXT: ret <8 x half> [[TMP3]] +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]]) +// CHECK-NEXT: [[TMP2:%.*]] = call <8 x half> @llvm.masked.load.v8f16.p0(ptr [[BASE:%.*]], i32 2, <8 x i1> [[TMP1]], <8 x half> zeroinitializer) +// CHECK-NEXT: ret <8 x half> [[TMP2]] // float16x8_t test_vld1q_z_f16(const float16_t *base, mve_pred16_t p) { @@ -136,10 +136,10 @@ float16x8_t test_vld1q_z_f16(const float16_t *base, mve_pred16_t p) // CHECK-LABEL: @test_vld1q_z_f32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP1]]) -// CHECK-NEXT: [[TMP3:%.*]] = call <4 x float> @llvm.masked.load.v4f32.p0(ptr [[BASE:%.*]], i32 4, <4 x i1> [[TMP2]], <4 x float> zeroinitializer) -// CHECK-NEXT: ret <4 x float> [[TMP3]] +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]]) +// CHECK-NEXT: [[TMP2:%.*]] = call <4 x float> @llvm.masked.load.v4f32.p0(ptr [[BASE:%.*]], i32 4, <4 x i1> [[TMP1]], <4 x float> zeroinitializer) +// CHECK-NEXT: ret <4 x float> [[TMP2]] // float32x4_t test_vld1q_z_f32(const float32_t *base, mve_pred16_t p) { @@ -152,10 +152,10 @@ float32x4_t test_vld1q_z_f32(const float32_t *base, mve_pred16_t p) // CHECK-LABEL: @test_vld1q_z_s8( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP1]]) -// CHECK-NEXT: [[TMP3:%.*]] = call <16 x i8> @llvm.masked.load.v16i8.p0(ptr [[BASE:%.*]], i32 1, <16 x i1> [[TMP2]], <16 x i8> zeroinitializer) -// CHECK-NEXT: ret <16 x i8> [[TMP3]] +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]]) +// CHECK-NEXT: [[TMP2:%.*]] = call <16 x i8> @llvm.masked.load.v16i8.p0(ptr [[BASE:%.*]], i32 1, <16 x i1> [[TMP1]], <16 x i8> zeroinitializer) +// CHECK-NEXT: ret <16 x i8> [[TMP2]] // int8x16_t test_vld1q_z_s8(const int8_t *base, mve_pred16_t p) { @@ -168,10 +168,10 @@ int8x16_t test_vld1q_z_s8(const int8_t *base, mve_pred16_t p) // CHECK-LABEL: @test_vld1q_z_s16( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP1]]) -// CHECK-NEXT: [[TMP3:%.*]] = call <8 x i16> @llvm.masked.load.v8i16.p0(ptr [[BASE:%.*]], i32 2, <8 x i1> [[TMP2]], <8 x i16> zeroinitializer) -// CHECK-NEXT: ret <8 x i16> [[TMP3]] +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]]) +// CHECK-NEXT: [[TMP2:%.*]] = call <8 x i16> @llvm.masked.load.v8i16.p0(ptr [[BASE:%.*]], i32 2, <8 x i1> [[TMP1]], <8 x i16> zeroinitializer) +// CHECK-NEXT: ret <8 x i16> [[TMP2]] // int16x8_t test_vld1q_z_s16(const int16_t *base, mve_pred16_t p) { @@ -184,10 +184,10 @@ int16x8_t test_vld1q_z_s16(const int16_t *base, mve_pred16_t p) // CHECK-LABEL: @test_vld1q_z_s32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP1]]) -// CHECK-NEXT: [[TMP3:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0(ptr [[BASE:%.*]], i32 4, <4 x i1> [[TMP2]], <4 x i32> zeroinitializer) -// CHECK-NEXT: ret <4 x i32> [[TMP3]] +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]]) +// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0(ptr [[BASE:%.*]], i32 4, <4 x i1> [[TMP1]], <4 x i32> zeroinitializer) +// CHECK-NEXT: ret <4 x i32> [[TMP2]] // int32x4_t test_vld1q_z_s32(const int32_t *base, mve_pred16_t p) { @@ -200,10 +200,10 @@ int32x4_t test_vld1q_z_s32(const int32_t *base, mve_pred16_t p) // CHECK-LABEL: @test_vld1q_z_u8( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP1]]) -// CHECK-NEXT: [[TMP3:%.*]] = call <16 x i8> @llvm.masked.load.v16i8.p0(ptr [[BASE:%.*]], i32 1, <16 x i1> [[TMP2]], <16 x i8> zeroinitializer) -// CHECK-NEXT: ret <16 x i8> [[TMP3]] +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]]) +// CHECK-NEXT: [[TMP2:%.*]] = call <16 x i8> @llvm.masked.load.v16i8.p0(ptr [[BASE:%.*]], i32 1, <16 x i1> [[TMP1]], <16 x i8> zeroinitializer) +// CHECK-NEXT: ret <16 x i8> [[TMP2]] // uint8x16_t test_vld1q_z_u8(const uint8_t *base, mve_pred16_t p) { @@ -216,10 +216,10 @@ uint8x16_t test_vld1q_z_u8(const uint8_t *base, mve_pred16_t p) // CHECK-LABEL: @test_vld1q_z_u16( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP1]]) -// CHECK-NEXT: [[TMP3:%.*]] = call <8 x i16> @llvm.masked.load.v8i16.p0(ptr [[BASE:%.*]], i32 2, <8 x i1> [[TMP2]], <8 x i16> zeroinitializer) -// CHECK-NEXT: ret <8 x i16> [[TMP3]] +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]]) +// CHECK-NEXT: [[TMP2:%.*]] = call <8 x i16> @llvm.masked.load.v8i16.p0(ptr [[BASE:%.*]], i32 2, <8 x i1> [[TMP1]], <8 x i16> zeroinitializer) +// CHECK-NEXT: ret <8 x i16> [[TMP2]] // uint16x8_t test_vld1q_z_u16(const uint16_t *base, mve_pred16_t p) { @@ -232,10 +232,10 @@ uint16x8_t test_vld1q_z_u16(const uint16_t *base, mve_pred16_t p) // CHECK-LABEL: @test_vld1q_z_u32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP1]]) -// CHECK-NEXT: [[TMP3:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0(ptr [[BASE:%.*]], i32 4, <4 x i1> [[TMP2]], <4 x i32> zeroinitializer) -// CHECK-NEXT: ret <4 x i32> [[TMP3]] +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]]) +// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0(ptr [[BASE:%.*]], i32 4, <4 x i1> [[TMP1]], <4 x i32> zeroinitializer) +// CHECK-NEXT: ret <4 x i32> [[TMP2]] // uint32x4_t test_vld1q_z_u32(const uint32_t *base, mve_pred16_t p) { @@ -248,8 +248,8 @@ uint32x4_t test_vld1q_z_u32(const uint32_t *base, mve_pred16_t p) // CHECK-LABEL: @test_vldrbq_s8( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = load <16 x i8>, ptr [[BASE:%.*]], align 1 -// CHECK-NEXT: ret <16 x i8> [[TMP1]] +// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i8>, ptr [[BASE:%.*]], align 1 +// CHECK-NEXT: ret <16 x i8> [[TMP0]] // int8x16_t test_vldrbq_s8(const int8_t *base) { @@ -258,9 +258,9 @@ int8x16_t test_vldrbq_s8(const int8_t *base) // CHECK-LABEL: @test_vldrbq_s16( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = load <8 x i8>, ptr [[BASE:%.*]], align 1 -// CHECK-NEXT: [[TMP2:%.*]] = sext <8 x i8> [[TMP1]] to <8 x i16> -// CHECK-NEXT: ret <8 x i16> [[TMP2]] +// CHECK-NEXT: [[TMP0:%.*]] = load <8 x i8>, ptr [[BASE:%.*]], align 1 +// CHECK-NEXT: [[TMP1:%.*]] = sext <8 x i8> [[TMP0]] to <8 x i16> +// CHECK-NEXT: ret <8 x i16> [[TMP1]] // int16x8_t test_vldrbq_s16(const int8_t *base) { @@ -269,9 +269,9 @@ int16x8_t test_vldrbq_s16(const int8_t *base) // CHECK-LABEL: @test_vldrbq_s32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = load <4 x i8>, ptr [[BASE:%.*]], align 1 -// CHECK-NEXT: [[TMP2:%.*]] = sext <4 x i8> [[TMP1]] to <4 x i32> -// CHECK-NEXT: ret <4 x i32> [[TMP2]] +// CHECK-NEXT: [[TMP0:%.*]] = load <4 x i8>, ptr [[BASE:%.*]], align 1 +// CHECK-NEXT: [[TMP1:%.*]] = sext <4 x i8> [[TMP0]] to <4 x i32> +// CHECK-NEXT: ret <4 x i32> [[TMP1]] // int32x4_t test_vldrbq_s32(const int8_t *base) { @@ -280,8 +280,8 @@ int32x4_t test_vldrbq_s32(const int8_t *base) // CHECK-LABEL: @test_vldrbq_u8( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = load <16 x i8>, ptr [[BASE:%.*]], align 1 -// CHECK-NEXT: ret <16 x i8> [[TMP1]] +// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i8>, ptr [[BASE:%.*]], align 1 +// CHECK-NEXT: ret <16 x i8> [[TMP0]] // uint8x16_t test_vldrbq_u8(const uint8_t *base) { @@ -290,9 +290,9 @@ uint8x16_t test_vldrbq_u8(const uint8_t *base) // CHECK-LABEL: @test_vldrbq_u16( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = load <8 x i8>, ptr [[BASE:%.*]], align 1 -// CHECK-NEXT: [[TMP2:%.*]] = zext <8 x i8> [[TMP1]] to <8 x i16> -// CHECK-NEXT: ret <8 x i16> [[TMP2]] +// CHECK-NEXT: [[TMP0:%.*]] = load <8 x i8>, ptr [[BASE:%.*]], align 1 +// CHECK-NEXT: [[TMP1:%.*]] = zext <8 x i8> [[TMP0]] to <8 x i16> +// CHECK-NEXT: ret <8 x i16> [[TMP1]] // uint16x8_t test_vldrbq_u16(const uint8_t *base) { @@ -301,9 +301,9 @@ uint16x8_t test_vldrbq_u16(const uint8_t *base) // CHECK-LABEL: @test_vldrbq_u32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = load <4 x i8>, ptr [[BASE:%.*]], align 1 -// CHECK-NEXT: [[TMP2:%.*]] = zext <4 x i8> [[TMP1]] to <4 x i32> -// CHECK-NEXT: ret <4 x i32> [[TMP2]] +// CHECK-NEXT: [[TMP0:%.*]] = load <4 x i8>, ptr [[BASE:%.*]], align 1 +// CHECK-NEXT: [[TMP1:%.*]] = zext <4 x i8> [[TMP0]] to <4 x i32> +// CHECK-NEXT: ret <4 x i32> [[TMP1]] // uint32x4_t test_vldrbq_u32(const uint8_t *base) { @@ -312,10 +312,10 @@ uint32x4_t test_vldrbq_u32(const uint8_t *base) // CHECK-LABEL: @test_vldrbq_z_s8( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP1]]) -// CHECK-NEXT: [[TMP3:%.*]] = call <16 x i8> @llvm.masked.load.v16i8.p0(ptr [[BASE:%.*]], i32 1, <16 x i1> [[TMP2]], <16 x i8> zeroinitializer) -// CHECK-NEXT: ret <16 x i8> [[TMP3]] +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]]) +// CHECK-NEXT: [[TMP2:%.*]] = call <16 x i8> @llvm.masked.load.v16i8.p0(ptr [[BASE:%.*]], i32 1, <16 x i1> [[TMP1]], <16 x i8> zeroinitializer) +// CHECK-NEXT: ret <16 x i8> [[TMP2]] // int8x16_t test_vldrbq_z_s8(const int8_t *base, mve_pred16_t p) { @@ -324,11 +324,11 @@ int8x16_t test_vldrbq_z_s8(const int8_t *base, mve_pred16_t p) // CHECK-LABEL: @test_vldrbq_z_s16( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP1]]) -// CHECK-NEXT: [[TMP3:%.*]] = call <8 x i8> @llvm.masked.load.v8i8.p0(ptr [[BASE:%.*]], i32 1, <8 x i1> [[TMP2]], <8 x i8> zeroinitializer) -// CHECK-NEXT: [[TMP4:%.*]] = sext <8 x i8> [[TMP3]] to <8 x i16> -// CHECK-NEXT: ret <8 x i16> [[TMP4]] +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]]) +// CHECK-NEXT: [[TMP2:%.*]] = call <8 x i8> @llvm.masked.load.v8i8.p0(ptr [[BASE:%.*]], i32 1, <8 x i1> [[TMP1]], <8 x i8> zeroinitializer) +// CHECK-NEXT: [[TMP3:%.*]] = sext <8 x i8> [[TMP2]] to <8 x i16> +// CHECK-NEXT: ret <8 x i16> [[TMP3]] // int16x8_t test_vldrbq_z_s16(const int8_t *base, mve_pred16_t p) { @@ -337,11 +337,11 @@ int16x8_t test_vldrbq_z_s16(const int8_t *base, mve_pred16_t p) // CHECK-LABEL: @test_vldrbq_z_s32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP1]]) -// CHECK-NEXT: [[TMP3:%.*]] = call <4 x i8> @llvm.masked.load.v4i8.p0(ptr [[BASE:%.*]], i32 1, <4 x i1> [[TMP2]], <4 x i8> zeroinitializer) -// CHECK-NEXT: [[TMP4:%.*]] = sext <4 x i8> [[TMP3]] to <4 x i32> -// CHECK-NEXT: ret <4 x i32> [[TMP4]] +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]]) +// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i8> @llvm.masked.load.v4i8.p0(ptr [[BASE:%.*]], i32 1, <4 x i1> [[TMP1]], <4 x i8> zeroinitializer) +// CHECK-NEXT: [[TMP3:%.*]] = sext <4 x i8> [[TMP2]] to <4 x i32> +// CHECK-NEXT: ret <4 x i32> [[TMP3]] // int32x4_t test_vldrbq_z_s32(const int8_t *base, mve_pred16_t p) { @@ -350,10 +350,10 @@ int32x4_t test_vldrbq_z_s32(const int8_t *base, mve_pred16_t p) // CHECK-LABEL: @test_vldrbq_z_u8( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP1]]) -// CHECK-NEXT: [[TMP3:%.*]] = call <16 x i8> @llvm.masked.load.v16i8.p0(ptr [[BASE:%.*]], i32 1, <16 x i1> [[TMP2]], <16 x i8> zeroinitializer) -// CHECK-NEXT: ret <16 x i8> [[TMP3]] +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]]) +// CHECK-NEXT: [[TMP2:%.*]] = call <16 x i8> @llvm.masked.load.v16i8.p0(ptr [[BASE:%.*]], i32 1, <16 x i1> [[TMP1]], <16 x i8> zeroinitializer) +// CHECK-NEXT: ret <16 x i8> [[TMP2]] // uint8x16_t test_vldrbq_z_u8(const uint8_t *base, mve_pred16_t p) { @@ -362,11 +362,11 @@ uint8x16_t test_vldrbq_z_u8(const uint8_t *base, mve_pred16_t p) // CHECK-LABEL: @test_vldrbq_z_u16( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP1]]) -// CHECK-NEXT: [[TMP3:%.*]] = call <8 x i8> @llvm.masked.load.v8i8.p0(ptr [[BASE:%.*]], i32 1, <8 x i1> [[TMP2]], <8 x i8> zeroinitializer) -// CHECK-NEXT: [[TMP4:%.*]] = zext <8 x i8> [[TMP3]] to <8 x i16> -// CHECK-NEXT: ret <8 x i16> [[TMP4]] +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]]) +// CHECK-NEXT: [[TMP2:%.*]] = call <8 x i8> @llvm.masked.load.v8i8.p0(ptr [[BASE:%.*]], i32 1, <8 x i1> [[TMP1]], <8 x i8> zeroinitializer) +// CHECK-NEXT: [[TMP3:%.*]] = zext <8 x i8> [[TMP2]] to <8 x i16> +// CHECK-NEXT: ret <8 x i16> [[TMP3]] // uint16x8_t test_vldrbq_z_u16(const uint8_t *base, mve_pred16_t p) { @@ -375,11 +375,11 @@ uint16x8_t test_vldrbq_z_u16(const uint8_t *base, mve_pred16_t p) // CHECK-LABEL: @test_vldrbq_z_u32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP1]]) -// CHECK-NEXT: [[TMP3:%.*]] = call <4 x i8> @llvm.masked.load.v4i8.p0(ptr [[BASE:%.*]], i32 1, <4 x i1> [[TMP2]], <4 x i8> zeroinitializer) -// CHECK-NEXT: [[TMP4:%.*]] = zext <4 x i8> [[TMP3]] to <4 x i32> -// CHECK-NEXT: ret <4 x i32> [[TMP4]] +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]]) +// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i8> @llvm.masked.load.v4i8.p0(ptr [[BASE:%.*]], i32 1, <4 x i1> [[TMP1]], <4 x i8> zeroinitializer) +// CHECK-NEXT: [[TMP3:%.*]] = zext <4 x i8> [[TMP2]] to <4 x i32> +// CHECK-NEXT: ret <4 x i32> [[TMP3]] // uint32x4_t test_vldrbq_z_u32(const uint8_t *base, mve_pred16_t p) { @@ -388,8 +388,8 @@ uint32x4_t test_vldrbq_z_u32(const uint8_t *base, mve_pred16_t p) // CHECK-LABEL: @test_vldrhq_f16( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = load <8 x half>, ptr [[BASE:%.*]], align 2 -// CHECK-NEXT: ret <8 x half> [[TMP1]] +// CHECK-NEXT: [[TMP0:%.*]] = load <8 x half>, ptr [[BASE:%.*]], align 2 +// CHECK-NEXT: ret <8 x half> [[TMP0]] // float16x8_t test_vldrhq_f16(const float16_t *base) { @@ -398,8 +398,8 @@ float16x8_t test_vldrhq_f16(const float16_t *base) // CHECK-LABEL: @test_vldrhq_s16( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = load <8 x i16>, ptr [[BASE:%.*]], align 2 -// CHECK-NEXT: ret <8 x i16> [[TMP1]] +// CHECK-NEXT: [[TMP0:%.*]] = load <8 x i16>, ptr [[BASE:%.*]], align 2 +// CHECK-NEXT: ret <8 x i16> [[TMP0]] // int16x8_t test_vldrhq_s16(const int16_t *base) { @@ -408,9 +408,9 @@ int16x8_t test_vldrhq_s16(const int16_t *base) // CHECK-LABEL: @test_vldrhq_s32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = load <4 x i16>, ptr [[BASE:%.*]], align 2 -// CHECK-NEXT: [[TMP2:%.*]] = sext <4 x i16> [[TMP1]] to <4 x i32> -// CHECK-NEXT: ret <4 x i32> [[TMP2]] +// CHECK-NEXT: [[TMP0:%.*]] = load <4 x i16>, ptr [[BASE:%.*]], align 2 +// CHECK-NEXT: [[TMP1:%.*]] = sext <4 x i16> [[TMP0]] to <4 x i32> +// CHECK-NEXT: ret <4 x i32> [[TMP1]] // int32x4_t test_vldrhq_s32(const int16_t *base) { @@ -419,8 +419,8 @@ int32x4_t test_vldrhq_s32(const int16_t *base) // CHECK-LABEL: @test_vldrhq_u16( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = load <8 x i16>, ptr [[BASE:%.*]], align 2 -// CHECK-NEXT: ret <8 x i16> [[TMP1]] +// CHECK-NEXT: [[TMP0:%.*]] = load <8 x i16>, ptr [[BASE:%.*]], align 2 +// CHECK-NEXT: ret <8 x i16> [[TMP0]] // uint16x8_t test_vldrhq_u16(const uint16_t *base) { @@ -429,9 +429,9 @@ uint16x8_t test_vldrhq_u16(const uint16_t *base) // CHECK-LABEL: @test_vldrhq_u32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = load <4 x i16>, ptr [[BASE:%.*]], align 2 -// CHECK-NEXT: [[TMP2:%.*]] = zext <4 x i16> [[TMP1]] to <4 x i32> -// CHECK-NEXT: ret <4 x i32> [[TMP2]] +// CHECK-NEXT: [[TMP0:%.*]] = load <4 x i16>, ptr [[BASE:%.*]], align 2 +// CHECK-NEXT: [[TMP1:%.*]] = zext <4 x i16> [[TMP0]] to <4 x i32> +// CHECK-NEXT: ret <4 x i32> [[TMP1]] // uint32x4_t test_vldrhq_u32(const uint16_t *base) { @@ -440,10 +440,10 @@ uint32x4_t test_vldrhq_u32(const uint16_t *base) // CHECK-LABEL: @test_vldrhq_z_f16( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP1]]) -// CHECK-NEXT: [[TMP3:%.*]] = call <8 x half> @llvm.masked.load.v8f16.p0(ptr [[BASE:%.*]], i32 2, <8 x i1> [[TMP2]], <8 x half> zeroinitializer) -// CHECK-NEXT: ret <8 x half> [[TMP3]] +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]]) +// CHECK-NEXT: [[TMP2:%.*]] = call <8 x half> @llvm.masked.load.v8f16.p0(ptr [[BASE:%.*]], i32 2, <8 x i1> [[TMP1]], <8 x half> zeroinitializer) +// CHECK-NEXT: ret <8 x half> [[TMP2]] // float16x8_t test_vldrhq_z_f16(const float16_t *base, mve_pred16_t p) { @@ -452,10 +452,10 @@ float16x8_t test_vldrhq_z_f16(const float16_t *base, mve_pred16_t p) // CHECK-LABEL: @test_vldrhq_z_s16( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP1]]) -// CHECK-NEXT: [[TMP3:%.*]] = call <8 x i16> @llvm.masked.load.v8i16.p0(ptr [[BASE:%.*]], i32 2, <8 x i1> [[TMP2]], <8 x i16> zeroinitializer) -// CHECK-NEXT: ret <8 x i16> [[TMP3]] +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]]) +// CHECK-NEXT: [[TMP2:%.*]] = call <8 x i16> @llvm.masked.load.v8i16.p0(ptr [[BASE:%.*]], i32 2, <8 x i1> [[TMP1]], <8 x i16> zeroinitializer) +// CHECK-NEXT: ret <8 x i16> [[TMP2]] // int16x8_t test_vldrhq_z_s16(const int16_t *base, mve_pred16_t p) { @@ -464,11 +464,11 @@ int16x8_t test_vldrhq_z_s16(const int16_t *base, mve_pred16_t p) // CHECK-LABEL: @test_vldrhq_z_s32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP1]]) -// CHECK-NEXT: [[TMP3:%.*]] = call <4 x i16> @llvm.masked.load.v4i16.p0(ptr [[BASE:%.*]], i32 2, <4 x i1> [[TMP2]], <4 x i16> zeroinitializer) -// CHECK-NEXT: [[TMP4:%.*]] = sext <4 x i16> [[TMP3]] to <4 x i32> -// CHECK-NEXT: ret <4 x i32> [[TMP4]] +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]]) +// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i16> @llvm.masked.load.v4i16.p0(ptr [[BASE:%.*]], i32 2, <4 x i1> [[TMP1]], <4 x i16> zeroinitializer) +// CHECK-NEXT: [[TMP3:%.*]] = sext <4 x i16> [[TMP2]] to <4 x i32> +// CHECK-NEXT: ret <4 x i32> [[TMP3]] // int32x4_t test_vldrhq_z_s32(const int16_t *base, mve_pred16_t p) { @@ -477,10 +477,10 @@ int32x4_t test_vldrhq_z_s32(const int16_t *base, mve_pred16_t p) // CHECK-LABEL: @test_vldrhq_z_u16( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP1]]) -// CHECK-NEXT: [[TMP3:%.*]] = call <8 x i16> @llvm.masked.load.v8i16.p0(ptr [[BASE:%.*]], i32 2, <8 x i1> [[TMP2]], <8 x i16> zeroinitializer) -// CHECK-NEXT: ret <8 x i16> [[TMP3]] +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]]) +// CHECK-NEXT: [[TMP2:%.*]] = call <8 x i16> @llvm.masked.load.v8i16.p0(ptr [[BASE:%.*]], i32 2, <8 x i1> [[TMP1]], <8 x i16> zeroinitializer) +// CHECK-NEXT: ret <8 x i16> [[TMP2]] // uint16x8_t test_vldrhq_z_u16(const uint16_t *base, mve_pred16_t p) { @@ -489,11 +489,11 @@ uint16x8_t test_vldrhq_z_u16(const uint16_t *base, mve_pred16_t p) // CHECK-LABEL: @test_vldrhq_z_u32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP1]]) -// CHECK-NEXT: [[TMP3:%.*]] = call <4 x i16> @llvm.masked.load.v4i16.p0(ptr [[BASE:%.*]], i32 2, <4 x i1> [[TMP2]], <4 x i16> zeroinitializer) -// CHECK-NEXT: [[TMP4:%.*]] = zext <4 x i16> [[TMP3]] to <4 x i32> -// CHECK-NEXT: ret <4 x i32> [[TMP4]] +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]]) +// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i16> @llvm.masked.load.v4i16.p0(ptr [[BASE:%.*]], i32 2, <4 x i1> [[TMP1]], <4 x i16> zeroinitializer) +// CHECK-NEXT: [[TMP3:%.*]] = zext <4 x i16> [[TMP2]] to <4 x i32> +// CHECK-NEXT: ret <4 x i32> [[TMP3]] // uint32x4_t test_vldrhq_z_u32(const uint16_t *base, mve_pred16_t p) { @@ -502,8 +502,8 @@ uint32x4_t test_vldrhq_z_u32(const uint16_t *base, mve_pred16_t p) // CHECK-LABEL: @test_vldrwq_f32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = load <4 x float>, ptr [[BASE:%.*]], align 4 -// CHECK-NEXT: ret <4 x float> [[TMP1]] +// CHECK-NEXT: [[TMP0:%.*]] = load <4 x float>, ptr [[BASE:%.*]], align 4 +// CHECK-NEXT: ret <4 x float> [[TMP0]] // float32x4_t test_vldrwq_f32(const float32_t *base) { @@ -512,8 +512,8 @@ float32x4_t test_vldrwq_f32(const float32_t *base) // CHECK-LABEL: @test_vldrwq_s32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = load <4 x i32>, ptr [[BASE:%.*]], align 4 -// CHECK-NEXT: ret <4 x i32> [[TMP1]] +// CHECK-NEXT: [[TMP0:%.*]] = load <4 x i32>, ptr [[BASE:%.*]], align 4 +// CHECK-NEXT: ret <4 x i32> [[TMP0]] // int32x4_t test_vldrwq_s32(const int32_t *base) { @@ -522,8 +522,8 @@ int32x4_t test_vldrwq_s32(const int32_t *base) // CHECK-LABEL: @test_vldrwq_u32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = load <4 x i32>, ptr [[BASE:%.*]], align 4 -// CHECK-NEXT: ret <4 x i32> [[TMP1]] +// CHECK-NEXT: [[TMP0:%.*]] = load <4 x i32>, ptr [[BASE:%.*]], align 4 +// CHECK-NEXT: ret <4 x i32> [[TMP0]] // uint32x4_t test_vldrwq_u32(const uint32_t *base) { @@ -532,10 +532,10 @@ uint32x4_t test_vldrwq_u32(const uint32_t *base) // CHECK-LABEL: @test_vldrwq_z_f32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP1]]) -// CHECK-NEXT: [[TMP3:%.*]] = call <4 x float> @llvm.masked.load.v4f32.p0(ptr [[BASE:%.*]], i32 4, <4 x i1> [[TMP2]], <4 x float> zeroinitializer) -// CHECK-NEXT: ret <4 x float> [[TMP3]] +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]]) +// CHECK-NEXT: [[TMP2:%.*]] = call <4 x float> @llvm.masked.load.v4f32.p0(ptr [[BASE:%.*]], i32 4, <4 x i1> [[TMP1]], <4 x float> zeroinitializer) +// CHECK-NEXT: ret <4 x float> [[TMP2]] // float32x4_t test_vldrwq_z_f32(const float32_t *base, mve_pred16_t p) { @@ -544,10 +544,10 @@ float32x4_t test_vldrwq_z_f32(const float32_t *base, mve_pred16_t p) // CHECK-LABEL: @test_vldrwq_z_s32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP1]]) -// CHECK-NEXT: [[TMP3:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0(ptr [[BASE:%.*]], i32 4, <4 x i1> [[TMP2]], <4 x i32> zeroinitializer) -// CHECK-NEXT: ret <4 x i32> [[TMP3]] +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]]) +// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0(ptr [[BASE:%.*]], i32 4, <4 x i1> [[TMP1]], <4 x i32> zeroinitializer) +// CHECK-NEXT: ret <4 x i32> [[TMP2]] // int32x4_t test_vldrwq_z_s32(const int32_t *base, mve_pred16_t p) { @@ -556,10 +556,10 @@ int32x4_t test_vldrwq_z_s32(const int32_t *base, mve_pred16_t p) // CHECK-LABEL: @test_vldrwq_z_u32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP1]]) -// CHECK-NEXT: [[TMP3:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0(ptr [[BASE:%.*]], i32 4, <4 x i1> [[TMP2]], <4 x i32> zeroinitializer) -// CHECK-NEXT: ret <4 x i32> [[TMP3]] +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]]) +// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0(ptr [[BASE:%.*]], i32 4, <4 x i1> [[TMP1]], <4 x i32> zeroinitializer) +// CHECK-NEXT: ret <4 x i32> [[TMP2]] // uint32x4_t test_vldrwq_z_u32(const uint32_t *base, mve_pred16_t p) { @@ -680,9 +680,9 @@ void test_vst1q_u32(uint32_t *base, uint32x4_t value) // CHECK-LABEL: @test_vst1q_p_f16( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP1]]) -// CHECK-NEXT: call void @llvm.masked.store.v8f16.p0(<8 x half> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 2, <8 x i1> [[TMP2]]) +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]]) +// CHECK-NEXT: call void @llvm.masked.store.v8f16.p0(<8 x half> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 2, <8 x i1> [[TMP1]]) // CHECK-NEXT: ret void // void test_vst1q_p_f16(float16_t *base, float16x8_t value, mve_pred16_t p) @@ -696,9 +696,9 @@ void test_vst1q_p_f16(float16_t *base, float16x8_t value, mve_pred16_t p) // CHECK-LABEL: @test_vst1q_p_f32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP1]]) -// CHECK-NEXT: call void @llvm.masked.store.v4f32.p0(<4 x float> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 4, <4 x i1> [[TMP2]]) +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]]) +// CHECK-NEXT: call void @llvm.masked.store.v4f32.p0(<4 x float> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 4, <4 x i1> [[TMP1]]) // CHECK-NEXT: ret void // void test_vst1q_p_f32(float32_t *base, float32x4_t value, mve_pred16_t p) @@ -712,9 +712,9 @@ void test_vst1q_p_f32(float32_t *base, float32x4_t value, mve_pred16_t p) // CHECK-LABEL: @test_vst1q_p_s8( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP1]]) -// CHECK-NEXT: call void @llvm.masked.store.v16i8.p0(<16 x i8> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 1, <16 x i1> [[TMP2]]) +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]]) +// CHECK-NEXT: call void @llvm.masked.store.v16i8.p0(<16 x i8> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 1, <16 x i1> [[TMP1]]) // CHECK-NEXT: ret void // void test_vst1q_p_s8(int8_t *base, int8x16_t value, mve_pred16_t p) @@ -728,9 +728,9 @@ void test_vst1q_p_s8(int8_t *base, int8x16_t value, mve_pred16_t p) // CHECK-LABEL: @test_vst1q_p_s16( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP1]]) -// CHECK-NEXT: call void @llvm.masked.store.v8i16.p0(<8 x i16> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 2, <8 x i1> [[TMP2]]) +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]]) +// CHECK-NEXT: call void @llvm.masked.store.v8i16.p0(<8 x i16> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 2, <8 x i1> [[TMP1]]) // CHECK-NEXT: ret void // void test_vst1q_p_s16(int16_t *base, int16x8_t value, mve_pred16_t p) @@ -744,9 +744,9 @@ void test_vst1q_p_s16(int16_t *base, int16x8_t value, mve_pred16_t p) // CHECK-LABEL: @test_vst1q_p_s32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP1]]) -// CHECK-NEXT: call void @llvm.masked.store.v4i32.p0(<4 x i32> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 4, <4 x i1> [[TMP2]]) +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]]) +// CHECK-NEXT: call void @llvm.masked.store.v4i32.p0(<4 x i32> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 4, <4 x i1> [[TMP1]]) // CHECK-NEXT: ret void // void test_vst1q_p_s32(int32_t *base, int32x4_t value, mve_pred16_t p) @@ -760,9 +760,9 @@ void test_vst1q_p_s32(int32_t *base, int32x4_t value, mve_pred16_t p) // CHECK-LABEL: @test_vst1q_p_u8( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP1]]) -// CHECK-NEXT: call void @llvm.masked.store.v16i8.p0(<16 x i8> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 1, <16 x i1> [[TMP2]]) +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]]) +// CHECK-NEXT: call void @llvm.masked.store.v16i8.p0(<16 x i8> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 1, <16 x i1> [[TMP1]]) // CHECK-NEXT: ret void // void test_vst1q_p_u8(uint8_t *base, uint8x16_t value, mve_pred16_t p) @@ -776,9 +776,9 @@ void test_vst1q_p_u8(uint8_t *base, uint8x16_t value, mve_pred16_t p) // CHECK-LABEL: @test_vst1q_p_u16( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP1]]) -// CHECK-NEXT: call void @llvm.masked.store.v8i16.p0(<8 x i16> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 2, <8 x i1> [[TMP2]]) +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]]) +// CHECK-NEXT: call void @llvm.masked.store.v8i16.p0(<8 x i16> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 2, <8 x i1> [[TMP1]]) // CHECK-NEXT: ret void // void test_vst1q_p_u16(uint16_t *base, uint16x8_t value, mve_pred16_t p) @@ -792,9 +792,9 @@ void test_vst1q_p_u16(uint16_t *base, uint16x8_t value, mve_pred16_t p) // CHECK-LABEL: @test_vst1q_p_u32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP1]]) -// CHECK-NEXT: call void @llvm.masked.store.v4i32.p0(<4 x i32> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 4, <4 x i1> [[TMP2]]) +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]]) +// CHECK-NEXT: call void @llvm.masked.store.v4i32.p0(<4 x i32> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 4, <4 x i1> [[TMP1]]) // CHECK-NEXT: ret void // void test_vst1q_p_u32(uint32_t *base, uint32x4_t value, mve_pred16_t p) @@ -896,9 +896,9 @@ void test_vstrbq_u32(uint8_t *base, uint32x4_t value) // CHECK-LABEL: @test_vstrbq_p_s8( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP1]]) -// CHECK-NEXT: call void @llvm.masked.store.v16i8.p0(<16 x i8> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 1, <16 x i1> [[TMP2]]) +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]]) +// CHECK-NEXT: call void @llvm.masked.store.v16i8.p0(<16 x i8> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 1, <16 x i1> [[TMP1]]) // CHECK-NEXT: ret void // void test_vstrbq_p_s8(int8_t *base, int8x16_t value, mve_pred16_t p) @@ -913,9 +913,9 @@ void test_vstrbq_p_s8(int8_t *base, int8x16_t value, mve_pred16_t p) // CHECK-LABEL: @test_vstrbq_p_s16( // CHECK-NEXT: entry: // CHECK-NEXT: [[TMP0:%.*]] = trunc <8 x i16> [[VALUE:%.*]] to <8 x i8> -// CHECK-NEXT: [[TMP2:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP3:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP2]]) -// CHECK-NEXT: call void @llvm.masked.store.v8i8.p0(<8 x i8> [[TMP0]], ptr [[BASE:%.*]], i32 1, <8 x i1> [[TMP3]]) +// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP2:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP1]]) +// CHECK-NEXT: call void @llvm.masked.store.v8i8.p0(<8 x i8> [[TMP0]], ptr [[BASE:%.*]], i32 1, <8 x i1> [[TMP2]]) // CHECK-NEXT: ret void // void test_vstrbq_p_s16(int8_t *base, int16x8_t value, mve_pred16_t p) @@ -930,9 +930,9 @@ void test_vstrbq_p_s16(int8_t *base, int16x8_t value, mve_pred16_t p) // CHECK-LABEL: @test_vstrbq_p_s32( // CHECK-NEXT: entry: // CHECK-NEXT: [[TMP0:%.*]] = trunc <4 x i32> [[VALUE:%.*]] to <4 x i8> -// CHECK-NEXT: [[TMP2:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP3:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP2]]) -// CHECK-NEXT: call void @llvm.masked.store.v4i8.p0(<4 x i8> [[TMP0]], ptr [[BASE:%.*]], i32 1, <4 x i1> [[TMP3]]) +// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP1]]) +// CHECK-NEXT: call void @llvm.masked.store.v4i8.p0(<4 x i8> [[TMP0]], ptr [[BASE:%.*]], i32 1, <4 x i1> [[TMP2]]) // CHECK-NEXT: ret void // void test_vstrbq_p_s32(int8_t *base, int32x4_t value, mve_pred16_t p) @@ -946,9 +946,9 @@ void test_vstrbq_p_s32(int8_t *base, int32x4_t value, mve_pred16_t p) // CHECK-LABEL: @test_vstrbq_p_u8( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP1]]) -// CHECK-NEXT: call void @llvm.masked.store.v16i8.p0(<16 x i8> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 1, <16 x i1> [[TMP2]]) +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <16 x i1> @llvm.arm.mve.pred.i2v.v16i1(i32 [[TMP0]]) +// CHECK-NEXT: call void @llvm.masked.store.v16i8.p0(<16 x i8> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 1, <16 x i1> [[TMP1]]) // CHECK-NEXT: ret void // void test_vstrbq_p_u8(uint8_t *base, uint8x16_t value, mve_pred16_t p) @@ -963,9 +963,9 @@ void test_vstrbq_p_u8(uint8_t *base, uint8x16_t value, mve_pred16_t p) // CHECK-LABEL: @test_vstrbq_p_u16( // CHECK-NEXT: entry: // CHECK-NEXT: [[TMP0:%.*]] = trunc <8 x i16> [[VALUE:%.*]] to <8 x i8> -// CHECK-NEXT: [[TMP2:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP3:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP2]]) -// CHECK-NEXT: call void @llvm.masked.store.v8i8.p0(<8 x i8> [[TMP0]], ptr [[BASE:%.*]], i32 1, <8 x i1> [[TMP3]]) +// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP2:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP1]]) +// CHECK-NEXT: call void @llvm.masked.store.v8i8.p0(<8 x i8> [[TMP0]], ptr [[BASE:%.*]], i32 1, <8 x i1> [[TMP2]]) // CHECK-NEXT: ret void // void test_vstrbq_p_u16(uint8_t *base, uint16x8_t value, mve_pred16_t p) @@ -980,9 +980,9 @@ void test_vstrbq_p_u16(uint8_t *base, uint16x8_t value, mve_pred16_t p) // CHECK-LABEL: @test_vstrbq_p_u32( // CHECK-NEXT: entry: // CHECK-NEXT: [[TMP0:%.*]] = trunc <4 x i32> [[VALUE:%.*]] to <4 x i8> -// CHECK-NEXT: [[TMP2:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP3:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP2]]) -// CHECK-NEXT: call void @llvm.masked.store.v4i8.p0(<4 x i8> [[TMP0]], ptr [[BASE:%.*]], i32 1, <4 x i1> [[TMP3]]) +// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP1]]) +// CHECK-NEXT: call void @llvm.masked.store.v4i8.p0(<4 x i8> [[TMP0]], ptr [[BASE:%.*]], i32 1, <4 x i1> [[TMP2]]) // CHECK-NEXT: ret void // void test_vstrbq_p_u32(uint8_t *base, uint32x4_t value, mve_pred16_t p) @@ -1068,9 +1068,9 @@ void test_vstrhq_u32(uint16_t *base, uint32x4_t value) // CHECK-LABEL: @test_vstrhq_p_f16( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP1]]) -// CHECK-NEXT: call void @llvm.masked.store.v8f16.p0(<8 x half> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 2, <8 x i1> [[TMP2]]) +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]]) +// CHECK-NEXT: call void @llvm.masked.store.v8f16.p0(<8 x half> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 2, <8 x i1> [[TMP1]]) // CHECK-NEXT: ret void // void test_vstrhq_p_f16(float16_t *base, float16x8_t value, mve_pred16_t p) @@ -1084,9 +1084,9 @@ void test_vstrhq_p_f16(float16_t *base, float16x8_t value, mve_pred16_t p) // CHECK-LABEL: @test_vstrhq_p_s16( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP1]]) -// CHECK-NEXT: call void @llvm.masked.store.v8i16.p0(<8 x i16> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 2, <8 x i1> [[TMP2]]) +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]]) +// CHECK-NEXT: call void @llvm.masked.store.v8i16.p0(<8 x i16> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 2, <8 x i1> [[TMP1]]) // CHECK-NEXT: ret void // void test_vstrhq_p_s16(int16_t *base, int16x8_t value, mve_pred16_t p) @@ -1101,9 +1101,9 @@ void test_vstrhq_p_s16(int16_t *base, int16x8_t value, mve_pred16_t p) // CHECK-LABEL: @test_vstrhq_p_s32( // CHECK-NEXT: entry: // CHECK-NEXT: [[TMP0:%.*]] = trunc <4 x i32> [[VALUE:%.*]] to <4 x i16> -// CHECK-NEXT: [[TMP2:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP3:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP2]]) -// CHECK-NEXT: call void @llvm.masked.store.v4i16.p0(<4 x i16> [[TMP0]], ptr [[BASE:%.*]], i32 2, <4 x i1> [[TMP3]]) +// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP1]]) +// CHECK-NEXT: call void @llvm.masked.store.v4i16.p0(<4 x i16> [[TMP0]], ptr [[BASE:%.*]], i32 2, <4 x i1> [[TMP2]]) // CHECK-NEXT: ret void // void test_vstrhq_p_s32(int16_t *base, int32x4_t value, mve_pred16_t p) @@ -1117,9 +1117,9 @@ void test_vstrhq_p_s32(int16_t *base, int32x4_t value, mve_pred16_t p) // CHECK-LABEL: @test_vstrhq_p_u16( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP1]]) -// CHECK-NEXT: call void @llvm.masked.store.v8i16.p0(<8 x i16> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 2, <8 x i1> [[TMP2]]) +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <8 x i1> @llvm.arm.mve.pred.i2v.v8i1(i32 [[TMP0]]) +// CHECK-NEXT: call void @llvm.masked.store.v8i16.p0(<8 x i16> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 2, <8 x i1> [[TMP1]]) // CHECK-NEXT: ret void // void test_vstrhq_p_u16(uint16_t *base, uint16x8_t value, mve_pred16_t p) @@ -1134,9 +1134,9 @@ void test_vstrhq_p_u16(uint16_t *base, uint16x8_t value, mve_pred16_t p) // CHECK-LABEL: @test_vstrhq_p_u32( // CHECK-NEXT: entry: // CHECK-NEXT: [[TMP0:%.*]] = trunc <4 x i32> [[VALUE:%.*]] to <4 x i16> -// CHECK-NEXT: [[TMP2:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP3:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP2]]) -// CHECK-NEXT: call void @llvm.masked.store.v4i16.p0(<4 x i16> [[TMP0]], ptr [[BASE:%.*]], i32 2, <4 x i1> [[TMP3]]) +// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP1]]) +// CHECK-NEXT: call void @llvm.masked.store.v4i16.p0(<4 x i16> [[TMP0]], ptr [[BASE:%.*]], i32 2, <4 x i1> [[TMP2]]) // CHECK-NEXT: ret void // void test_vstrhq_p_u32(uint16_t *base, uint32x4_t value, mve_pred16_t p) @@ -1192,9 +1192,9 @@ void test_vstrwq_u32(uint32_t *base, uint32x4_t value) // CHECK-LABEL: @test_vstrwq_p_f32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP1]]) -// CHECK-NEXT: call void @llvm.masked.store.v4f32.p0(<4 x float> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 4, <4 x i1> [[TMP2]]) +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]]) +// CHECK-NEXT: call void @llvm.masked.store.v4f32.p0(<4 x float> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 4, <4 x i1> [[TMP1]]) // CHECK-NEXT: ret void // void test_vstrwq_p_f32(float32_t *base, float32x4_t value, mve_pred16_t p) @@ -1208,9 +1208,9 @@ void test_vstrwq_p_f32(float32_t *base, float32x4_t value, mve_pred16_t p) // CHECK-LABEL: @test_vstrwq_p_s32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP1]]) -// CHECK-NEXT: call void @llvm.masked.store.v4i32.p0(<4 x i32> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 4, <4 x i1> [[TMP2]]) +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]]) +// CHECK-NEXT: call void @llvm.masked.store.v4i32.p0(<4 x i32> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 4, <4 x i1> [[TMP1]]) // CHECK-NEXT: ret void // void test_vstrwq_p_s32(int32_t *base, int32x4_t value, mve_pred16_t p) @@ -1224,9 +1224,9 @@ void test_vstrwq_p_s32(int32_t *base, int32x4_t value, mve_pred16_t p) // CHECK-LABEL: @test_vstrwq_p_u32( // CHECK-NEXT: entry: -// CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[P:%.*]] to i32 -// CHECK-NEXT: [[TMP2:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP1]]) -// CHECK-NEXT: call void @llvm.masked.store.v4i32.p0(<4 x i32> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 4, <4 x i1> [[TMP2]]) +// CHECK-NEXT: [[TMP0:%.*]] = zext i16 [[P:%.*]] to i32 +// CHECK-NEXT: [[TMP1:%.*]] = call <4 x i1> @llvm.arm.mve.pred.i2v.v4i1(i32 [[TMP0]]) +// CHECK-NEXT: call void @llvm.masked.store.v4i32.p0(<4 x i32> [[VALUE:%.*]], ptr [[BASE:%.*]], i32 4, <4 x i1> [[TMP1]]) // CHECK-NEXT: ret void // void test_vstrwq_p_u32(uint32_t *base, uint32x4_t value, mve_pred16_t p) diff --git a/clang/test/CodeGen/attr-target-mv.c b/clang/test/CodeGen/attr-target-mv.c index 07f47d9..b8807dd 100644 --- a/clang/test/CodeGen/attr-target-mv.c +++ b/clang/test/CodeGen/attr-target-mv.c @@ -30,6 +30,8 @@ int __attribute__((target("arch=gracemont"))) foo(void) {return 24;} int __attribute__((target("arch=pantherlake"))) foo(void) {return 25;} int __attribute__((target("arch=clearwaterforest"))) foo(void) {return 26;} int __attribute__((target("arch=diamondrapids"))) foo(void) {return 27;} +int __attribute__((target("arch=wildcatlake"))) foo(void) {return 28;} +int __attribute__((target("arch=novalake"))) foo(void) {return 29;} int __attribute__((target("default"))) foo(void) { return 2; } int bar(void) { @@ -203,6 +205,10 @@ void calls_pr50025c(void) { pr50025c(); } // ITANIUM: ret i32 26 // ITANIUM: define{{.*}} i32 @foo.arch_diamondrapids() // ITANIUM: ret i32 27 +// ITANIUM: define{{.*}} i32 @foo.arch_wildcatlake() +// ITANIUM: ret i32 28 +// ITANIUM: define{{.*}} i32 @foo.arch_novalake() +// ITANIUM: ret i32 29 // ITANIUM: define{{.*}} i32 @foo() // ITANIUM: ret i32 2 // ITANIUM: define{{.*}} i32 @bar() @@ -262,6 +268,10 @@ void calls_pr50025c(void) { pr50025c(); } // WINDOWS: ret i32 26 // WINDOWS: define dso_local i32 @foo.arch_diamondrapids() // WINDOWS: ret i32 27 +// WINDOWS: define dso_local i32 @foo.arch_wildcatlake() +// WINDOWS: ret i32 28 +// WINDOWS: define dso_local i32 @foo.arch_novalake() +// WINDOWS: ret i32 29 // WINDOWS: define dso_local i32 @foo() // WINDOWS: ret i32 2 // WINDOWS: define dso_local i32 @bar() diff --git a/clang/test/CodeGen/distributed-thin-lto/supports-hot-cold-new.ll b/clang/test/CodeGen/distributed-thin-lto/supports-hot-cold-new.ll index 08c1a29..90cda3e 100644 --- a/clang/test/CodeGen/distributed-thin-lto/supports-hot-cold-new.ll +++ b/clang/test/CodeGen/distributed-thin-lto/supports-hot-cold-new.ll @@ -22,7 +22,7 @@ ; RUN: %clang -target x86_64-unknown-linux-gnu -O2 -o %t1.o -x ir %t.o -c -fthinlto-index=%t.o.thinlto.bc -save-temps=obj -; RUN: llvm-dis %t.s.3.import.bc -o - | FileCheck %s --check-prefix=CHECK-IR +; RUN: llvm-dis %t.s.4.opt.bc -o - | FileCheck %s --check-prefix=CHECK-IR ; CHECK-IR: !memprof {{.*}} !callsite ; CHECK-IR: "memprof"="cold" @@ -42,10 +42,15 @@ ; RUN: %clang -target x86_64-unknown-linux-gnu -O2 -o %t1.o -x ir %t.o -c -fthinlto-index=%t.o.thinlto.bc -save-temps=obj -; RUN: llvm-dis %t.s.3.import.bc -o - | FileCheck %s \ +; RUN: llvm-dis %t.s.4.opt.bc -o - | FileCheck %s \ ; RUN: --implicit-check-not "!memprof" --implicit-check-not "!callsite" \ ; RUN: --implicit-check-not "memprof"="cold" +;; Ensure the attributes and metadata are stripped when running a non-LTO pipeline. +; RUN: %clang -target x86_64-unknown-linux-gnu -O2 -x ir %t.o -S -emit-llvm -o - | FileCheck %s \ +; RUN: --implicit-check-not "!memprof" --implicit-check-not "!callsite" \ +; RUN: --implicit-check-not "memprof"="cold" + source_filename = "thinlto-distributed-supports-hot-cold-new.ll" target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128" target triple = "x86_64-unknown-linux-gnu" diff --git a/clang/test/CodeGen/ext-int-cc.c b/clang/test/CodeGen/ext-int-cc.c index 7cfd992..f845afc 100644 --- a/clang/test/CodeGen/ext-int-cc.c +++ b/clang/test/CodeGen/ext-int-cc.c @@ -49,8 +49,8 @@ void ParamPassing(_BitInt(128) b, _BitInt(64) c) {} // R600: define{{.*}} void @ParamPassing(ptr addrspace(5) byval(i128) align 8 %{{.+}}, i64 %{{.+}}) // ARC: define{{.*}} void @ParamPassing(ptr byval(i128) align 4 %{{.+}}, i64 inreg %{{.+}}) // XCORE: define{{.*}} void @ParamPassing(ptr byval(i128) align 4 %{{.+}}, i64 %{{.+}}) -// RISCV64: define{{.*}} void @ParamPassing(i128 %{{.+}}, i64 %{{.+}}) -// RISCV32: define{{.*}} void @ParamPassing(ptr dead_on_return %{{.+}}, i64 %{{.+}}) +// RISCV64: define{{.*}} void @ParamPassing(i128 signext %{{.+}}, i64 signext %{{.+}}) +// RISCV32: define{{.*}} void @ParamPassing(ptr dead_on_return %{{.+}}, i64 signext %{{.+}}) // WASM: define{{.*}} void @ParamPassing(i128 %{{.+}}, i64 %{{.+}}) // SYSTEMZ: define{{.*}} void @ParamPassing(ptr dead_on_return %{{.+}}, i64 %{{.+}}) // PPC64: define{{.*}} void @ParamPassing(i128 %{{.+}}, i64 %{{.+}}) @@ -79,8 +79,8 @@ void ParamPassing2(_BitInt(127) b, _BitInt(63) c) {} // R600: define{{.*}} void @ParamPassing2(ptr addrspace(5) byval(i128) align 8 %{{.+}}, i63 %{{.+}}) // ARC: define{{.*}} void @ParamPassing2(ptr byval(i128) align 4 %{{.+}}, i63 inreg %{{.+}}) // XCORE: define{{.*}} void @ParamPassing2(ptr byval(i128) align 4 %{{.+}}, i63 %{{.+}}) -// RISCV64: define{{.*}} void @ParamPassing2(i127 %{{.+}}, i63 signext %{{.+}}) -// RISCV32: define{{.*}} void @ParamPassing2(ptr dead_on_return %{{.+}}, i63 %{{.+}}) +// RISCV64: define{{.*}} void @ParamPassing2(i127 signext %{{.+}}, i63 signext %{{.+}}) +// RISCV32: define{{.*}} void @ParamPassing2(ptr dead_on_return %{{.+}}, i63 signext %{{.+}}) // WASM: define{{.*}} void @ParamPassing2(i127 %{{.+}}, i63 %{{.+}}) // SYSTEMZ: define{{.*}} void @ParamPassing2(ptr dead_on_return %{{.+}}, i63 signext %{{.+}}) // PPC64: define{{.*}} void @ParamPassing2(i127 %{{.+}}, i63 signext %{{.+}}) diff --git a/clang/test/CodeGen/inline-asm-systemz-flag-output.c b/clang/test/CodeGen/inline-asm-systemz-flag-output.c new file mode 100644 index 0000000..041797b --- /dev/null +++ b/clang/test/CodeGen/inline-asm-systemz-flag-output.c @@ -0,0 +1,57 @@ +// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 6 +// RUN: %clang_cc1 -O2 -triple s390x-linux -emit-llvm -o - %s | FileCheck %s + +// CHECK-LABEL: define dso_local signext range(i32 0, 4) i32 @test( +// CHECK-SAME: i32 noundef signext [[X:%.*]]) local_unnamed_addr #[[ATTR0:[0-9]+]] { +// CHECK-NEXT: [[ENTRY:.*:]] +// CHECK-NEXT: [[TMP0:%.*]] = tail call { i32, i32 } asm "ahi $0,42\0A", "=d,={@cc},0"(i32 [[X]]) #[[ATTR2:[0-9]+]], !srcloc [[META2:![0-9]+]] +// CHECK-NEXT: [[ASMRESULT1:%.*]] = extractvalue { i32, i32 } [[TMP0]], 1 +// CHECK-NEXT: [[TMP1:%.*]] = icmp ult i32 [[ASMRESULT1]], 4 +// CHECK-NEXT: tail call void @llvm.assume(i1 [[TMP1]]) +// CHECK-NEXT: ret i32 [[ASMRESULT1]] +// +int test(int x) { + int cc; + asm ("ahi %[x],42\n" : [x] "+d"(x), "=@cc" (cc)); + return cc; +} + +// CHECK-LABEL: define dso_local signext range(i32 0, 2) i32 @test_low_high_transformation( +// CHECK-SAME: i32 noundef signext [[X:%.*]]) local_unnamed_addr #[[ATTR0]] { +// CHECK-NEXT: [[ENTRY:.*:]] +// CHECK-NEXT: [[TMP0:%.*]] = tail call { i32, i32 } asm "ahi $0,42\0A", "=d,={@cc},0"(i32 [[X]]) #[[ATTR2]], !srcloc [[META3:![0-9]+]] +// CHECK-NEXT: [[ASMRESULT1:%.*]] = extractvalue { i32, i32 } [[TMP0]], 1 +// CHECK-NEXT: [[TMP1:%.*]] = icmp ult i32 [[ASMRESULT1]], 4 +// CHECK-NEXT: tail call void @llvm.assume(i1 [[TMP1]]) +// CHECK-NEXT: [[TMP2:%.*]] = add nsw i32 [[ASMRESULT1]], -1 +// CHECK-NEXT: [[TMP3:%.*]] = icmp ult i32 [[TMP2]], 2 +// CHECK-NEXT: [[LOR_EXT:%.*]] = zext i1 [[TMP3]] to i32 +// CHECK-NEXT: ret i32 [[LOR_EXT]] +// +int test_low_high_transformation(int x) { + int cc; + asm ("ahi %[x],42\n" : [x] "+d"(x), "=@cc" (cc)); + return cc == 1 || cc == 2; +} + +// CHECK-LABEL: define dso_local signext range(i32 0, 2) i32 @test_equal_high_transformation( +// CHECK-SAME: i32 noundef signext [[X:%.*]]) local_unnamed_addr #[[ATTR0]] { +// CHECK-NEXT: [[ENTRY:.*:]] +// CHECK-NEXT: [[TMP0:%.*]] = tail call { i32, i32 } asm "ahi $0,42\0A", "=d,={@cc},0"(i32 [[X]]) #[[ATTR2]], !srcloc [[META4:![0-9]+]] +// CHECK-NEXT: [[ASMRESULT1:%.*]] = extractvalue { i32, i32 } [[TMP0]], 1 +// CHECK-NEXT: [[TMP1:%.*]] = icmp ult i32 [[ASMRESULT1]], 4 +// CHECK-NEXT: tail call void @llvm.assume(i1 [[TMP1]]) +// CHECK-NEXT: [[TMP2:%.*]] = and i32 [[ASMRESULT1]], 1 +// CHECK-NEXT: [[LOR_EXT:%.*]] = xor i32 [[TMP2]], 1 +// CHECK-NEXT: ret i32 [[LOR_EXT]] +// +int test_equal_high_transformation(int x) { + int cc; + asm ("ahi %[x],42\n" : [x] "+d"(x), "=@cc" (cc)); + return cc == 0 || cc == 2; +} +//. +// CHECK: [[META2]] = !{i64 788} +// CHECK: [[META3]] = !{i64 1670} +// CHECK: [[META4]] = !{i64 2505} +//. diff --git a/clang/test/CodeGen/target-builtin-noerror.c b/clang/test/CodeGen/target-builtin-noerror.c index 120f1a5..47d5ae5 100644 --- a/clang/test/CodeGen/target-builtin-noerror.c +++ b/clang/test/CodeGen/target-builtin-noerror.c @@ -178,6 +178,8 @@ void verifycpustrings(void) { (void)__builtin_cpu_is("lunarlake"); (void)__builtin_cpu_is("clearwaterforest"); (void)__builtin_cpu_is("pantherlake"); + (void)__builtin_cpu_is("wildcatlake"); + (void)__builtin_cpu_is("novalake"); (void)__builtin_cpu_is("haswell"); (void)__builtin_cpu_is("icelake-client"); (void)__builtin_cpu_is("icelake-server"); diff --git a/clang/test/CodeGen/target-data.c b/clang/test/CodeGen/target-data.c index eecee69..e950794 100644 --- a/clang/test/CodeGen/target-data.c +++ b/clang/test/CodeGen/target-data.c @@ -152,20 +152,20 @@ // RUN: %clang_cc1 -triple r600-unknown -o - -emit-llvm %s | \ // RUN: FileCheck %s -check-prefix=R600 -// R600: target datalayout = "e-p:32:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-v2048:2048-n32:64-S32-A5-G1" +// R600: target datalayout = "e-m:e-p:32:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-v2048:2048-n32:64-S32-A5-G1" // RUN: %clang_cc1 -triple r600-unknown -target-cpu cayman -o - -emit-llvm %s \ // RUN: | FileCheck %s -check-prefix=R600D -// R600D: target datalayout = "e-p:32:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-v2048:2048-n32:64-S32-A5-G1" +// R600D: target datalayout = "e-m:e-p:32:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-v2048:2048-n32:64-S32-A5-G1" // RUN: %clang_cc1 -triple amdgcn-unknown -target-cpu hawaii -o - -emit-llvm %s \ // RUN: | FileCheck %s -check-prefix=R600SI -// R600SI: target datalayout = "e-p:64:64-p1:64:64-p2:32:32-p3:32:32-p4:64:64-p5:32:32-p6:32:32-p7:160:256:256:32-p8:128:128:128:48-p9:192:256:256:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-v2048:2048-n32:64-S32-A5-G1-ni:7:8:9" +// R600SI: target datalayout = "e-m:e-p:64:64-p1:64:64-p2:32:32-p3:32:32-p4:64:64-p5:32:32-p6:32:32-p7:160:256:256:32-p8:128:128:128:48-p9:192:256:256:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-v2048:2048-n32:64-S32-A5-G1-ni:7:8:9" // Test default -target-cpu // RUN: %clang_cc1 -triple amdgcn-unknown -o - -emit-llvm %s \ // RUN: | FileCheck %s -check-prefix=R600SIDefault -// R600SIDefault: target datalayout = "e-p:64:64-p1:64:64-p2:32:32-p3:32:32-p4:64:64-p5:32:32-p6:32:32-p7:160:256:256:32-p8:128:128:128:48-p9:192:256:256:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-v2048:2048-n32:64-S32-A5-G1-ni:7:8:9" +// R600SIDefault: target datalayout = "e-m:e-p:64:64-p1:64:64-p2:32:32-p3:32:32-p4:64:64-p5:32:32-p6:32:32-p7:160:256:256:32-p8:128:128:128:48-p9:192:256:256:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-v2048:2048-n32:64-S32-A5-G1-ni:7:8:9" // RUN: %clang_cc1 -triple arm64-unknown -o - -emit-llvm %s | \ // RUN: FileCheck %s -check-prefix=AARCH64 diff --git a/clang/test/CodeGen/unified-lto-module-flag.ll b/clang/test/CodeGen/unified-lto-module-flag.ll new file mode 100644 index 0000000..deefe82 --- /dev/null +++ b/clang/test/CodeGen/unified-lto-module-flag.ll @@ -0,0 +1,11 @@ +; Test that we do not duplicate the UnifiedLTO module flag. +; +; RUN: %clang_cc1 -emit-llvm -flto=full -funified-lto -o - %s | FileCheck %s + +; CHECK: !llvm.module.flags = !{!0, !1, !2, !3} +!llvm.module.flags = !{!0, !1, !2, !3} + +!0 = !{i32 1, !"wchar_size", i32 2} +!1 = !{i32 7, !"frame-pointer", i32 2} +!2 = !{i32 1, !"EnableSplitLTOUnit", i32 1} +!3 = !{i32 1, !"UnifiedLTO", i32 1} diff --git a/clang/test/CodeGenCUDA/Inputs/cuda.h b/clang/test/CodeGenCUDA/Inputs/cuda.h index dc85eae..e7ad784 100644 --- a/clang/test/CodeGenCUDA/Inputs/cuda.h +++ b/clang/test/CodeGenCUDA/Inputs/cuda.h @@ -13,6 +13,8 @@ #endif #define __launch_bounds__(...) __attribute__((launch_bounds(__VA_ARGS__))) #define __grid_constant__ __attribute__((grid_constant)) +#define __cluster_dims__(...) __attribute__((cluster_dims(__VA_ARGS__))) +#define __no_cluster__ __attribute__((no_cluster)) #else #define __constant__ #define __device__ @@ -22,6 +24,8 @@ #define __managed__ #define __launch_bounds__(...) #define __grid_constant__ +#define __cluster_dims__(...) +#define __no_cluster__ #endif struct dim3 { diff --git a/clang/test/CodeGenCUDA/cluster_dims.cu b/clang/test/CodeGenCUDA/cluster_dims.cu new file mode 100644 index 0000000..00635e3 --- /dev/null +++ b/clang/test/CodeGenCUDA/cluster_dims.cu @@ -0,0 +1,38 @@ +// RUN: %clang_cc1 -triple amdgcn-amd-amdhsa -target-cpu gfx1250 -fcuda-is-device -emit-llvm -x hip -o - %s | FileCheck %s +// RUN: %clang_cc1 -triple x86_64-pc-linux-gnu -aux-triple amdgcn-amd-amdhsa -emit-llvm -x hip -o - %s | FileCheck --check-prefix=HOST %s + +#include "Inputs/cuda.h" + +const int constint = 4; + +// HOST-NOT: "amdgpu-cluster-dims" + +// CHECK: "amdgpu-cluster-dims"="2,2,2" +__global__ void __cluster_dims__(2, 2, 2) test_literal_3d() {} + +// CHECK: "amdgpu-cluster-dims"="2,2,1" +__global__ void __cluster_dims__(2, 2) test_literal_2d() {} + +// CHECK: "amdgpu-cluster-dims"="4,1,1" +__global__ void __cluster_dims__(4) test_literal_1d() {} + +// CHECK: "amdgpu-cluster-dims"="4,2,1" +__global__ void __cluster_dims__(constint, constint / 2, 1) test_constant() {} + +// CHECK: "amdgpu-cluster-dims"="0,0,0" +__global__ void __no_cluster__ test_no_cluster() {} + +// CHECK: "amdgpu-cluster-dims"="7,1,1" +template<unsigned a> +__global__ void __cluster_dims__(a) test_template_1d() {} +template __global__ void test_template_1d<7>(); + +// CHECK: "amdgpu-cluster-dims"="2,6,1" +template<unsigned a, unsigned b> +__global__ void __cluster_dims__(a, b) test_template_2d() {} +template __global__ void test_template_2d<2, 6>(); + +// CHECK: "amdgpu-cluster-dims"="1,2,3" +template<unsigned a, unsigned b, unsigned c> +__global__ void __cluster_dims__(a, b, c) test_template_3d() {} +template __global__ void test_template_3d<1, 2, 3>(); diff --git a/clang/test/CodeGenCXX/sizeof-unwind-exception.cpp b/clang/test/CodeGenCXX/sizeof-unwind-exception.cpp index 4fb977a..e40b2d7 100644 --- a/clang/test/CodeGenCXX/sizeof-unwind-exception.cpp +++ b/clang/test/CodeGenCXX/sizeof-unwind-exception.cpp @@ -3,6 +3,8 @@ // RUN: %clang_cc1 -triple x86_64-apple-darwin10 -emit-llvm -fcxx-exceptions -fexceptions %s -O2 -o - | FileCheck %s --check-prefix=ARM-DARWIN // RUN: %clang_cc1 -triple arm-unknown-gnueabi -emit-llvm -fcxx-exceptions -fexceptions %s -O2 -o - | FileCheck %s --check-prefix=ARM-EABI // RUN: %clang_cc1 -triple mipsel-unknown-unknown -emit-llvm -fcxx-exceptions -fexceptions %s -O2 -o - | FileCheck %s --check-prefix=MIPS +// RUN: %clang_cc1 -triple x86_64-windows-gnu -emit-llvm -fcxx-exceptions -fexceptions -exception-model=seh %s -O2 -o - | FileCheck %s --check-prefix=MINGW-X86-64 +// RUN: %clang_cc1 -triple thumbv7-windows-gnu -emit-llvm -fcxx-exceptions -fexceptions -exception-model=seh %s -O2 -o - | FileCheck %s --check-prefix=MINGW-ARMV7 void foo(); void test() { @@ -25,9 +27,15 @@ void test() { // ARM-EABI-NEXT: [[T1:%.*]] = getelementptr i8, ptr [[EXN]], i32 88 // MIPS: [[T0:%.*]] = tail call ptr @__cxa_begin_catch(ptr [[EXN:%.*]]) [[NUW:#[0-9]+]] // MIPS-NEXT: [[T1:%.*]] = getelementptr i8, ptr [[EXN]], i32 24 +// MINGW-X86-64: [[T0:%.*]] = tail call ptr @__cxa_begin_catch(ptr [[EXN:%.*]]) [[NUW:#[0-9]+]] +// MINGW-X86-64-NEXT:[[T1:%.*]] = getelementptr i8, ptr [[EXN]], i64 64 +// MINGW-ARMV7: [[T0:%.*]] = tail call arm_aapcs_vfpcc ptr @__cxa_begin_catch(ptr [[EXN:%.*]]) [[NUW:#[0-9]+]] +// MINGW-ARMV7-NEXT: [[T1:%.*]] = getelementptr i8, ptr [[EXN]], i32 48 // X86-64: attributes [[NUW]] = { nounwind } // X86-32: attributes [[NUW]] = { nounwind } // ARM-DARWIN: attributes [[NUW]] = { nounwind } // ARM-EABI: attributes [[NUW]] = { nounwind } // MIPS: attributes [[NUW]] = { nounwind } +// MINGW-X86-64: attributes [[NUW]] = { nounwind } +// MINGW-ARMV7: attributes [[NUW]] = { nounwind } diff --git a/clang/test/CodeGenHLSL/Operators/logical-not.hlsl b/clang/test/CodeGenHLSL/Operators/logical-not.hlsl new file mode 100644 index 0000000..0f9d067 --- /dev/null +++ b/clang/test/CodeGenHLSL/Operators/logical-not.hlsl @@ -0,0 +1,33 @@ +// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.6-library -disable-llvm-passes -emit-llvm -finclude-default-header -fnative-half-type -o - %s | FileCheck %s + +// CHECK-LABEL: case1 +// CHECK: [[ToBool:%.*]] = icmp ne <2 x i32> {{.*}}, zeroinitializer +// CHECK-NEXT: [[BoolCmp:%.*]] = icmp eq <2 x i1> [[ToBool]], zeroinitializer +// CHECK-NEXT: {{.*}} = zext <2 x i1> [[BoolCmp]] to <2 x i32> +export uint32_t2 case1(uint32_t2 b) { + return !b; +} + +// CHECK-LABEL: case2 +// CHECK: [[ToBool:%.*]] = icmp ne <3 x i32> {{.*}}, zeroinitializer +// CHECK-NEXT: [[BoolCmp:%.*]] = icmp eq <3 x i1> [[ToBool]], zeroinitializer +// CHECK-NEXT: {{.*}} = zext <3 x i1> [[BoolCmp]] to <3 x i32> +export int32_t3 case2(int32_t3 b) { + return !b; +} + +// CHECK-LABEL: case3 +// CHECK: [[ToBool:%.*]] = fcmp reassoc nnan ninf nsz arcp afn une half {{.*}}, 0xH0000 +// CHECK-NEXT: [[BoolCmp:%.*]] = xor i1 [[ToBool]], true +// CHECK-NEXT: {{.*}} = uitofp i1 [[BoolCmp]] to half +export float16_t case3(float16_t b) { + return !b; +} + +// CHECK-LABEL: case4 +// CHECK: [[ToBool:%.*]] = fcmp reassoc nnan ninf nsz arcp afn une <4 x float> {{.*}}, zeroinitializer +// CHECK-NEXT: [[BoolCmp:%.*]] = icmp eq <4 x i1> [[ToBool]], zeroinitializer +// CHECK-NEXT: {{.*}} = uitofp <4 x i1> [[BoolCmp]] to <4 x float> +export float4 case4(float4 b) { + return !b; +} diff --git a/clang/test/CodeGenHLSL/basic-target.c b/clang/test/CodeGenHLSL/basic-target.c index c700e06..b9482df 100644 --- a/clang/test/CodeGenHLSL/basic-target.c +++ b/clang/test/CodeGenHLSL/basic-target.c @@ -6,5 +6,5 @@ // RUN: %clang -cc1 -triple dxil-pc-shadermodel6.0-domain -emit-llvm -o - %s | FileCheck %s // RUN: %clang -cc1 -triple dxil-pc-shadermodel6.0-geometry -emit-llvm -o - %s | FileCheck %s -// CHECK: target datalayout = "e-m:e-p:32:32-i1:32-i8:8-i16:16-i32:32-i64:64-f16:16-f32:32-f64:64-n8:16:32:64" +// CHECK: target datalayout = "e-m:e-p:32:32-i1:32-i8:8-i16:16-i32:32-i64:64-f16:16-f32:32-f64:64-n8:16:32:64-v48:16:16-v96:32:32-v192:64:64" // CHECK: target triple = "dxilv1.0-pc-shadermodel6.0-{{[a-z]+}}" diff --git a/clang/test/CodeGenHLSL/resources/RWBuffer-elementtype.hlsl b/clang/test/CodeGenHLSL/resources/RWBuffer-elementtype.hlsl deleted file mode 100644 index f48521b..0000000 --- a/clang/test/CodeGenHLSL/resources/RWBuffer-elementtype.hlsl +++ /dev/null @@ -1,70 +0,0 @@ -// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.2-compute -finclude-default-header -fnative-half-type -emit-llvm -o - %s | FileCheck %s -check-prefixes=DXIL -// RUN: %clang_cc1 -triple spirv-pc-vulkan-compute -finclude-default-header -fnative-half-type -emit-llvm -o - %s | FileCheck %s -check-prefixes=SPIRV - -// DXIL: %"class.hlsl::RWBuffer" = type { target("dx.TypedBuffer", i16, 1, 0, 1) } -// DXIL: %"class.hlsl::RWBuffer.0" = type { target("dx.TypedBuffer", i16, 1, 0, 0) } -// DXIL: %"class.hlsl::RWBuffer.1" = type { target("dx.TypedBuffer", i32, 1, 0, 1) } -// DXIL: %"class.hlsl::RWBuffer.2" = type { target("dx.TypedBuffer", i32, 1, 0, 0) } -// DXIL: %"class.hlsl::RWBuffer.3" = type { target("dx.TypedBuffer", i64, 1, 0, 1) } -// DXIL: %"class.hlsl::RWBuffer.4" = type { target("dx.TypedBuffer", i64, 1, 0, 0) } -// DXIL: %"class.hlsl::RWBuffer.5" = type { target("dx.TypedBuffer", half, 1, 0, 0) } -// DXIL: %"class.hlsl::RWBuffer.6" = type { target("dx.TypedBuffer", float, 1, 0, 0) } -// DXIL: %"class.hlsl::RWBuffer.7" = type { target("dx.TypedBuffer", double, 1, 0, 0) } -// DXIL: %"class.hlsl::RWBuffer.8" = type { target("dx.TypedBuffer", <4 x i16>, 1, 0, 1) } -// DXIL: %"class.hlsl::RWBuffer.9" = type { target("dx.TypedBuffer", <3 x i32>, 1, 0, 0) } -// DXIL: %"class.hlsl::RWBuffer.10" = type { target("dx.TypedBuffer", <2 x half>, 1, 0, 0) } -// DXIL: %"class.hlsl::RWBuffer.11" = type { target("dx.TypedBuffer", <3 x float>, 1, 0, 0) } -// DXIL: %"class.hlsl::RWBuffer.12" = type { target("dx.TypedBuffer", <4 x i32>, 1, 0, 1) } - -// SPIRV: %"class.hlsl::RWBuffer" = type { target("spirv.SignedImage", i16, 5, 2, 0, 0, 2, 0) } -// SPIRV: %"class.hlsl::RWBuffer.0" = type { target("spirv.Image", i16, 5, 2, 0, 0, 2, 0) } -// SPIRV: %"class.hlsl::RWBuffer.1" = type { target("spirv.SignedImage", i32, 5, 2, 0, 0, 2, 24) } -// SPIRV: %"class.hlsl::RWBuffer.2" = type { target("spirv.Image", i32, 5, 2, 0, 0, 2, 33) } -// SPIRV: %"class.hlsl::RWBuffer.3" = type { target("spirv.SignedImage", i64, 5, 2, 0, 0, 2, 41) } -// SPIRV: %"class.hlsl::RWBuffer.4" = type { target("spirv.Image", i64, 5, 2, 0, 0, 2, 40) } -// SPIRV: %"class.hlsl::RWBuffer.5" = type { target("spirv.Image", half, 5, 2, 0, 0, 2, 0) } -// SPIRV: %"class.hlsl::RWBuffer.6" = type { target("spirv.Image", float, 5, 2, 0, 0, 2, 3) } -// SPIRV: %"class.hlsl::RWBuffer.7" = type { target("spirv.Image", double, 5, 2, 0, 0, 2, 0) } -// SPIRV: %"class.hlsl::RWBuffer.8" = type { target("spirv.SignedImage", i16, 5, 2, 0, 0, 2, 0) } -// SPIRV: %"class.hlsl::RWBuffer.9" = type { target("spirv.Image", i32, 5, 2, 0, 0, 2, 0) } -// SPIRV: %"class.hlsl::RWBuffer.10" = type { target("spirv.Image", half, 5, 2, 0, 0, 2, 0) } -// SPIRV: %"class.hlsl::RWBuffer.11" = type { target("spirv.Image", float, 5, 2, 0, 0, 2, 0) } -// SPIRV: %"class.hlsl::RWBuffer.12" = type { target("spirv.SignedImage", i32, 5, 2, 0, 0, 2, 21) } - -RWBuffer<int16_t> BufI16; -RWBuffer<uint16_t> BufU16; -RWBuffer<int> BufI32; -RWBuffer<uint> BufU32; -RWBuffer<int64_t> BufI64; -RWBuffer<uint64_t> BufU64; -RWBuffer<half> BufF16; -RWBuffer<float> BufF32; -RWBuffer<double> BufF64; -RWBuffer< vector<int16_t, 4> > BufI16x4; -RWBuffer< vector<uint, 3> > BufU32x3; -RWBuffer<half2> BufF16x2; -RWBuffer<float3> BufF32x3; -RWBuffer<int4> BufI32x4; -// TODO: RWBuffer<snorm half> BufSNormF16; -> 11 -// TODO: RWBuffer<unorm half> BufUNormF16; -> 12 -// TODO: RWBuffer<snorm float> BufSNormF32; -> 13 -// TODO: RWBuffer<unorm float> BufUNormF32; -> 14 -// TODO: RWBuffer<snorm double> BufSNormF64; -> 15 -// TODO: RWBuffer<unorm double> BufUNormF64; -> 16 - -[numthreads(1,1,1)] -void main(int GI : SV_GroupIndex) { - BufI16[GI] = 0; - BufU16[GI] = 0; - BufI32[GI] = 0; - BufU32[GI] = 0; - BufI64[GI] = 0; - BufU64[GI] = 0; - BufF16[GI] = 0; - BufF32[GI] = 0; - BufF64[GI] = 0; - BufI16x4[GI] = 0; - BufU32x3[GI] = 0; - BufF16x2[GI] = 0; - BufF32x3[GI] = 0; -} diff --git a/clang/test/CodeGenHLSL/resources/RWBuffer-subscript.hlsl b/clang/test/CodeGenHLSL/resources/RWBuffer-subscript.hlsl deleted file mode 100644 index 0de171c..0000000 --- a/clang/test/CodeGenHLSL/resources/RWBuffer-subscript.hlsl +++ /dev/null @@ -1,26 +0,0 @@ -// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.0-compute -emit-llvm -o - -O0 %s | FileCheck %s --check-prefixes=DXC,CHECK -// RUN: %clang_cc1 -triple spirv1.6-pc-vulkan1.3-compute -fspv-use-unknown-image-format -emit-llvm -o - -O0 %s | FileCheck %s --check-prefixes=SPIRV,CHECK - -RWBuffer<int> In; -RWBuffer<int> Out; - -[numthreads(1,1,1)] -void main(unsigned GI : SV_GroupIndex) { - // CHECK: define void @main() - - // DXC: %[[INPTR:.*]] = call noundef nonnull align 4 dereferenceable(4) ptr @llvm.dx.resource.getpointer.p0.tdx.TypedBuffer_i32_1_0_1t(target("dx.TypedBuffer", i32, 1, 0, 1) %{{.*}}, i32 %{{.*}}) - // SPIRV: %[[INPTR:.*]] = call noundef align 4 dereferenceable(4) ptr addrspace(11) @llvm.spv.resource.getpointer.p11.tspirv.SignedImage_i32_5_2_0_0_2_0t(target("spirv.SignedImage", i32, 5, 2, 0, 0, 2, 0) %{{.*}}, i32 %{{.*}}) - // CHECK: %[[LOAD:.*]] = load i32, ptr {{.*}}%[[INPTR]] - // DXC: %[[OUTPTR:.*]] = call noundef nonnull align 4 dereferenceable(4) ptr @llvm.dx.resource.getpointer.p0.tdx.TypedBuffer_i32_1_0_1t(target("dx.TypedBuffer", i32, 1, 0, 1) %{{.*}}, i32 %{{.*}}) - // SPIRV: %[[OUTPTR:.*]] = call noundef align 4 dereferenceable(4) ptr addrspace(11) @llvm.spv.resource.getpointer.p11.tspirv.SignedImage_i32_5_2_0_0_2_0t(target("spirv.SignedImage", i32, 5, 2, 0, 0, 2, 0) %{{.*}}, i32 %{{.*}}) - // CHECK: store i32 %[[LOAD]], ptr {{.*}}%[[OUTPTR]] - Out[GI] = In[GI]; - - // DXC: %[[INPTR:.*]] = call ptr @llvm.dx.resource.getpointer.p0.tdx.TypedBuffer_i32_1_0_1t(target("dx.TypedBuffer", i32, 1, 0, 1) %{{.*}}, i32 %{{.*}}) - // SPIRV: %[[INPTR:.*]] = call ptr addrspace(11) @llvm.spv.resource.getpointer.p11.tspirv.SignedImage_i32_5_2_0_0_2_0t(target("spirv.SignedImage", i32, 5, 2, 0, 0, 2, 0) %{{.*}}, i32 %{{.*}}) - // CHECK: %[[LOAD:.*]] = load i32, ptr {{.*}}%[[INPTR]] - // DXC: %[[OUTPTR:.*]] = call noundef nonnull align 4 dereferenceable(4) ptr @llvm.dx.resource.getpointer.p0.tdx.TypedBuffer_i32_1_0_1t(target("dx.TypedBuffer", i32, 1, 0, 1) %{{.*}}, i32 %{{.*}}) - // SPIRV: %[[OUTPTR:.*]] = call noundef align 4 dereferenceable(4) ptr addrspace(11) @llvm.spv.resource.getpointer.p11.tspirv.SignedImage_i32_5_2_0_0_2_0t(target("spirv.SignedImage", i32, 5, 2, 0, 0, 2, 0) %{{.*}}, i32 %{{.*}}) - // CHECK: store i32 %[[LOAD]], ptr {{.*}}%[[OUTPTR]] - Out[GI] = In.Load(GI); -} diff --git a/clang/test/CodeGenHLSL/resources/StructuredBuffers-methods-lib.hlsl b/clang/test/CodeGenHLSL/resources/StructuredBuffers-methods-lib.hlsl index 43ddd2e..2cf6a10 100644 --- a/clang/test/CodeGenHLSL/resources/StructuredBuffers-methods-lib.hlsl +++ b/clang/test/CodeGenHLSL/resources/StructuredBuffers-methods-lib.hlsl @@ -1,64 +1,112 @@ -// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.3-library -x hlsl -emit-llvm -disable-llvm-passes -o - %s | FileCheck %s --check-prefixes=CHECK,CHECK-DXIL -// RUN-DISABLED: %clang_cc1 -triple spirv-vulkan-library -x hlsl -emit-llvm -disable-llvm-passes -o - %s | FileCheck %s --check-prefixes=CHECK,CHECK-SPIRV +// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.3-library -finclude-default-header -emit-llvm -disable-llvm-passes -o - %s | llvm-cxxfilt | FileCheck %s --check-prefixes=CHECK,DXIL +// RUN-DISABLED: %clang_cc1 -triple spirv-vulkan-library -finclude-default-header -emit-llvm -disable-llvm-passes -o - %s | llvm-cxxfilt | FileCheck %s --check-prefixes=CHECK,SPV // NOTE: SPIRV codegen for resource methods is not yet implemented StructuredBuffer<float> SB1 : register(t0); RWStructuredBuffer<float> RWSB1 : register(u0); -RWStructuredBuffer<float> RWSB2 : register(u1); +RWStructuredBuffer<uint4> RWSB2 : register(u1); AppendStructuredBuffer<float> ASB : register(u2); -ConsumeStructuredBuffer<float> CSB : register(u3); +ConsumeStructuredBuffer<double> CSB : register(u3); -// CHECK: %"class.hlsl::StructuredBuffer" = type { target("dx.RawBuffer", float, 0, 0) } -// CHECK: %"class.hlsl::RWStructuredBuffer" = type { target("dx.RawBuffer", float, 1, 0), target("dx.RawBuffer", float, 1, 0) } -// CHECK: %"class.hlsl::AppendStructuredBuffer" = type { target("dx.RawBuffer", float, 1, 0), target("dx.RawBuffer", float, 1, 0) } -// CHECK: %"class.hlsl::ConsumeStructuredBuffer" = type { target("dx.RawBuffer", float, 1, 0), target("dx.RawBuffer", float, 1, 0) } +// DXIL: %"class.hlsl::StructuredBuffer" = type { target("dx.RawBuffer", float, 0, 0) } +// DXIL: %"class.hlsl::RWStructuredBuffer" = type { target("dx.RawBuffer", float, 1, 0), target("dx.RawBuffer", float, 1, 0) } +// DXIL: %"class.hlsl::RWStructuredBuffer.0" = type { target("dx.RawBuffer", <4 x i32>, 1, 0), target("dx.RawBuffer", <4 x i32>, 1, 0) } +// DXIL: %"class.hlsl::AppendStructuredBuffer" = type { target("dx.RawBuffer", float, 1, 0), target("dx.RawBuffer", float, 1, 0) } +// DXIL: %"class.hlsl::ConsumeStructuredBuffer" = type { target("dx.RawBuffer", double, 1, 0), target("dx.RawBuffer", double, 1, 0) } export int TestIncrementCounter() { return RWSB1.IncrementCounter(); } -// CHECK: define noundef i32 @_Z20TestIncrementCounterv() -// CHECK-DXIL: %[[INDEX:.*]] = call i32 @llvm.dx.resource.updatecounter.tdx.RawBuffer_f32_1_0t(target("dx.RawBuffer", float, 1, 0) %{{[0-9]+}}, i8 1) -// CHECK-DXIL: ret i32 %[[INDEX]] +// CHECK: define noundef i32 @TestIncrementCounter()() +// CHECK: call noundef i32 @hlsl::RWStructuredBuffer<float>::IncrementCounter()(ptr {{.*}} @RWSB1) +// CHECK: ret + +// CHECK: define {{.*}} noundef i32 @hlsl::RWStructuredBuffer<float>::IncrementCounter()(ptr {{.*}} %this) +// CHECK: %__counter_handle = getelementptr inbounds nuw %"class.hlsl::RWStructuredBuffer", ptr %{{.*}}, i32 0, i32 1 +// DXIL-NEXT: %[[COUNTER_HANDLE:.*]] = load target("dx.RawBuffer", float, 1, 0), ptr %__counter_handle +// DXIL-NEXT: %[[COUNTER:.*]] = call i32 @llvm.dx.resource.updatecounter.tdx.RawBuffer_f32_1_0t(target("dx.RawBuffer", float, 1, 0) %[[COUNTER_HANDLE]], i8 1) +// CHECK-NEXT: ret i32 %[[COUNTER]] + export int TestDecrementCounter() { return RWSB2.DecrementCounter(); } +// CHECK: define {{.*}} i32 @TestDecrementCounter()() +// CHECK: call noundef i32 @hlsl::RWStructuredBuffer<unsigned int vector[4]>::DecrementCounter()(ptr {{.*}} @RWSB2) +// CHECK: ret -// CHECK: define noundef i32 @_Z20TestDecrementCounterv() -// CHECK-DXIL: %[[INDEX:.*]] = call i32 @llvm.dx.resource.updatecounter.tdx.RawBuffer_f32_1_0t(target("dx.RawBuffer", float, 1, 0) %{{[0-9]+}}, i8 -1) -// CHECK-DXIL: ret i32 %[[INDEX]] +// CHECK: define {{.*}} noundef i32 @hlsl::RWStructuredBuffer<unsigned int vector[4]>::DecrementCounter()(ptr {{.*}} %this) +// CHECK: %__counter_handle = getelementptr inbounds nuw %"class.hlsl::RWStructuredBuffer.0", ptr %{{.*}}, i32 0, i32 1 +// DXIL-NEXT: %[[COUNTER_HANDLE:.*]] = load target("dx.RawBuffer", <4 x i32>, 1, 0), ptr %__counter_handle +// DXIL-NEXT: %[[COUNTER:.*]] = call i32 @llvm.dx.resource.updatecounter.tdx.RawBuffer_v4i32_1_0t(target("dx.RawBuffer", <4 x i32>, 1, 0) %[[COUNTER_HANDLE]], i8 -1) +// CHECK-NEXT: ret i32 %[[COUNTER]] export void TestAppend(float value) { ASB.Append(value); } -// CHECK: define void @_Z10TestAppendf(float noundef nofpclass(nan inf) %value) -// CHECK-DXIL: %[[VALUE:.*]] = load float, ptr %value.addr, align 4 -// CHECK-DXIL: %[[INDEX:.*]] = call i32 @llvm.dx.resource.updatecounter.tdx.RawBuffer_f32_1_0t(target("dx.RawBuffer", float, 1, 0) %{{[0-9]+}}, i8 1) -// CHECK-DXIL: %[[RESPTR:.*]] = call ptr @llvm.dx.resource.getpointer.p0.tdx.RawBuffer_f32_1_0t(target("dx.RawBuffer", float, 1, 0) %{{[0-9]+}}, i32 %[[INDEX]]) -// CHECK-DXIL: store float %[[VALUE]], ptr %[[RESPTR]], align 4 +// CHECK: define void @TestAppend(float)(float {{.*}} %value) +// CHECK: call void @hlsl::AppendStructuredBuffer<float>::Append(float)(ptr {{.*}} @ASB, float noundef nofpclass(nan inf) %0) +// CHECK: ret void + +// CHECK: define {{.*}} void @hlsl::AppendStructuredBuffer<float>::Append(float)(ptr {{.*}} %this, float noundef nofpclass(nan inf) %value) +// CHECK: %[[VALUE:.*]] = load float, ptr %value.addr +// CHECK-NEXT: %__handle = getelementptr inbounds nuw %"class.hlsl::AppendStructuredBuffer", ptr %{{.*}}, i32 0, i32 0 +// DXIL-NEXT: %[[HANDLE:.*]] = load target("dx.RawBuffer", float, 1, 0), ptr %__handle +// CHECK-NEXT: %__counter_handle = getelementptr inbounds nuw %"class.hlsl::AppendStructuredBuffer", ptr %{{.*}}, i32 0, i32 1 +// DXIL-NEXT: %[[COUNTER_HANDLE:.*]] = load target("dx.RawBuffer", float, 1, 0), ptr %__counter_handle +// DXIL-NEXT: %[[COUNTER:.*]] = call i32 @llvm.dx.resource.updatecounter.tdx.RawBuffer_f32_1_0t(target("dx.RawBuffer", float, 1, 0) %[[COUNTER_HANDLE]], i8 1) +// DXIL-NEXT: %[[PTR:.*]] = call ptr @llvm.dx.resource.getpointer.p0.tdx.RawBuffer_f32_1_0t(target("dx.RawBuffer", float, 1, 0) %[[HANDLE]], i32 %[[COUNTER]]) +// CHECK-NEXT: store float %[[VALUE]], ptr %[[PTR]] +// CHECK-NEXT: ret void -export float TestConsume() { +export double TestConsume() { return CSB.Consume(); } - -// CHECK: define noundef nofpclass(nan inf) float @_Z11TestConsumev() -// CHECK-DXIL: %[[INDEX:.*]] = call i32 @llvm.dx.resource.updatecounter.tdx.RawBuffer_f32_1_0t(target("dx.RawBuffer", float, 1, 0) %1, i8 -1) -// CHECK-DXIL: %[[RESPTR:.*]] = call ptr @llvm.dx.resource.getpointer.p0.tdx.RawBuffer_f32_1_0t(target("dx.RawBuffer", float, 1, 0) %0, i32 %[[INDEX]]) -// CHECK-DXIL: %[[VALUE:.*]] = load float, ptr %[[RESPTR]], align 4 -// CHECK-DXIL: ret float %[[VALUE]] +// CHECK: define {{.*}} double @TestConsume()() +// CHECK: call {{.*}} double @hlsl::ConsumeStructuredBuffer<double>::Consume()(ptr {{.*}} @CSB) +// CHECK: ret double + +// CHECK: define {{.*}} double @hlsl::ConsumeStructuredBuffer<double>::Consume()(ptr {{.*}} %this) +// CHECK: %__handle = getelementptr inbounds nuw %"class.hlsl::ConsumeStructuredBuffer", ptr %{{.*}}, i32 0, i32 0 +// DXIL-NEXT: %[[HANDLE:.*]] = load target("dx.RawBuffer", double, 1, 0), ptr %__handle +// CHECK-NEXT: %__counter_handle = getelementptr inbounds nuw %"class.hlsl::ConsumeStructuredBuffer", ptr %{{.*}}, i32 0, i32 1 +// DXIL-NEXT: %[[COUNTER_HANDLE:.*]] = load target("dx.RawBuffer", double, 1, 0), ptr %__counter_handle +// DXIL-NEXT: %[[COUNTER:.*]] = call i32 @llvm.dx.resource.updatecounter.tdx.RawBuffer_f64_1_0t(target("dx.RawBuffer", double, 1, 0) %[[COUNTER_HANDLE]], i8 -1) +// DXIL-NEXT: %[[PTR:.*]] = call ptr @llvm.dx.resource.getpointer.p0.tdx.RawBuffer_f64_1_0t(target("dx.RawBuffer", double, 1, 0) %[[HANDLE]], i32 %[[COUNTER]]) +// CHECK-NEXT: %[[VAL:.*]] = load double, ptr %[[PTR]], align 8 +// CHECK-NEXT: ret double %[[VAL]] export float TestLoad() { return RWSB1.Load(1) + SB1.Load(2); } -// CHECK: define noundef nofpclass(nan inf) float @_Z8TestLoadv() -// CHECK: %[[PTR1:.*]] = call ptr @llvm.dx.resource.getpointer.p0.tdx.RawBuffer_f32_1_0t(target("dx.RawBuffer", float, 1, 0) %{{[0-9]+}}, i32 %{{[0-9]+}}) -// CHECK: %[[VALUE1:.*]] = load float, ptr %[[PTR1]] -// CHECK: %[[PTR2:.*]] = call ptr @llvm.dx.resource.getpointer.p0.tdx.RawBuffer_f32_0_0t(target("dx.RawBuffer", float, 0, 0) %{{[0-9]+}}, i32 %{{[0-9]+}}) -// CHECK: %[[VALUE2:.*]] = load float, ptr %[[PTR2]] +// CHECK: define noundef nofpclass(nan inf) float @TestLoad()() +// CHECK: call {{.*}} float @hlsl::RWStructuredBuffer<float>::Load(unsigned int)(ptr {{.*}} @RWSB1, i32 noundef 1) +// CHECK: call {{.*}} float @hlsl::StructuredBuffer<float>::Load(unsigned int)(ptr {{.*}} @SB1, i32 noundef 2) +// CHECK: add +// CHECK: ret float + +// CHECK: define {{.*}} float @hlsl::RWStructuredBuffer<float>::Load(unsigned int)(ptr {{.*}} %this, i32 noundef %Index) +// CHECK: %__handle = getelementptr inbounds nuw %"class.hlsl::RWStructuredBuffer", ptr %{{.*}}, i32 0, i32 0 +// DXIL-NEXT: %[[HANDLE:.*]] = load target("dx.RawBuffer", float, 1, 0), ptr %__handle +// CHECK-NEXT: %[[INDEX:.*]] = load i32, ptr %Index.addr +// DXIL-NEXT: %[[PTR:.*]] = call ptr @llvm.dx.resource.getpointer.p0.tdx.RawBuffer_f32_1_0t(target("dx.RawBuffer", float, 1, 0) %[[HANDLE]], i32 %[[INDEX]]) +// CHECK-NEXT: %[[VAL:.*]] = load float, ptr %[[PTR]] +// CHECK-NEXT: ret float %[[VAL]] + +// CHECK: define {{.*}} float @hlsl::StructuredBuffer<float>::Load(unsigned int)(ptr {{.*}} %this, i32 noundef %Index) +// CHECK: %__handle = getelementptr inbounds nuw %"class.hlsl::StructuredBuffer", ptr %{{.*}}, i32 0, i32 0 +// DXIL-NEXT: %[[HANDLE:.*]] = load target("dx.RawBuffer", float, 0, 0), ptr %__handle +// CHECK-NEXT: %[[INDEX:.*]] = load i32, ptr %Index.addr +// DXIL-NEXT: %[[PTR:.*]] = call ptr @llvm.dx.resource.getpointer.p0.tdx.RawBuffer_f32_0_0t(target("dx.RawBuffer", float, 0, 0) %[[HANDLE]], i32 %[[INDEX]]) +// CHECK-NEXT: %[[VAL:.*]] = load float, ptr %[[PTR]] +// CHECK-NEXT: ret float %[[VAL]] -// CHECK: declare i32 @llvm.dx.resource.updatecounter.tdx.RawBuffer_f32_1_0t(target("dx.RawBuffer", float, 1, 0), i8) -// CHECK: declare ptr @llvm.dx.resource.getpointer.p0.tdx.RawBuffer_f32_1_0t(target("dx.RawBuffer", float, 1, 0), i32) -// CHECK: declare ptr @llvm.dx.resource.getpointer.p0.tdx.RawBuffer_f32_0_0t(target("dx.RawBuffer", float, 0, 0), i32) +// DXIL: declare i32 @llvm.dx.resource.updatecounter.tdx.RawBuffer_f32_1_0t(target("dx.RawBuffer", float, 1, 0), i8) +// DXIL: declare i32 @llvm.dx.resource.updatecounter.tdx.RawBuffer_v4i32_1_0t(target("dx.RawBuffer", <4 x i32>, 1, 0), i8) +// DXIL: declare ptr @llvm.dx.resource.getpointer.p0.tdx.RawBuffer_f32_1_0t(target("dx.RawBuffer", float, 1, 0), i32) +// DXIL: declare i32 @llvm.dx.resource.updatecounter.tdx.RawBuffer_f64_1_0t(target("dx.RawBuffer", double, 1, 0), i8) +// DXIL: declare ptr @llvm.dx.resource.getpointer.p0.tdx.RawBuffer_f64_1_0t(target("dx.RawBuffer", double, 1, 0), i32) +// DXIL: declare ptr @llvm.dx.resource.getpointer.p0.tdx.RawBuffer_f32_0_0t(target("dx.RawBuffer", float, 0, 0), i32) diff --git a/clang/test/CodeGenHLSL/resources/StructuredBuffers-methods-ps.hlsl b/clang/test/CodeGenHLSL/resources/StructuredBuffers-methods-ps.hlsl index 9e08a6d..47c1d0d 100644 --- a/clang/test/CodeGenHLSL/resources/StructuredBuffers-methods-ps.hlsl +++ b/clang/test/CodeGenHLSL/resources/StructuredBuffers-methods-ps.hlsl @@ -1,37 +1,71 @@ -// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.3-pixel -x hlsl -emit-llvm -disable-llvm-passes -o - %s | FileCheck %s --check-prefixes=CHECK,CHECK-DXIL -// RUN-DISABLED: %clang_cc1 -triple spirv-vulkan-pixel -x hlsl -emit-llvm -disable-llvm-passes -o - %s | FileCheck %s --check-prefixes=CHECK,CHECK-SPIRV +// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.3-pixel -finclude-default-header -emit-llvm -disable-llvm-passes -o - %s | llvm-cxxfilt | FileCheck %s --check-prefixes=CHECK,DXIL +// RUN-DISABLED: %clang_cc1 -triple spirv-vulkan-pixel -finclude-default-header -emit-llvm -disable-llvm-passes -o - %s | llvm-cxxfilt | FileCheck %s --check-prefixes=CHECK,SPV // NOTE: SPIRV codegen for resource methods is not yet implemented -RWStructuredBuffer<float> RWSB1, RWSB2; -RasterizerOrderedStructuredBuffer<float> ROSB1, ROSB2; +RasterizerOrderedStructuredBuffer<float> ROSB1; +RasterizerOrderedStructuredBuffer<int2> ROSB2; -// CHECK: %"class.hlsl::RWStructuredBuffer" = type { target("dx.RawBuffer", float, 1, 0), target("dx.RawBuffer", float, 1, 0) } +// %"class.hlsl::RasterizerOrderedStructuredBuffer" = type { target("dx.RawBuffer", float, 1, 1), target("dx.RawBuffer", float, 1, 1) } +// %"class.hlsl::RasterizerOrderedStructuredBuffer.0" = type { target("dx.RawBuffer", <2 x i32>, 1, 1), target("dx.RawBuffer", <2 x i32>, 1, 1) } + +// CHECK: @ROSB1 = internal global %"class.hlsl::RasterizerOrderedStructuredBuffer" poison +// CHECK: @ROSB2 = internal global %"class.hlsl::RasterizerOrderedStructuredBuffer.0" poison export void TestIncrementCounter() { -// CHECK: define void @_Z20TestIncrementCounterv() -// CHECK-DXIL: call i32 @llvm.dx.resource.updatecounter.tdx.RawBuffer_f32_1_0t(target("dx.RawBuffer", float, 1, 0) %{{[0-9]+}}, i8 1) -// CHECK-DXIL: call i32 @llvm.dx.resource.updatecounter.tdx.RawBuffer_f32_1_1t(target("dx.RawBuffer", float, 1, 1) %{{[0-9]+}}, i8 1) - RWSB1.IncrementCounter(); ROSB1.IncrementCounter(); } +// CHECK: define void @TestIncrementCounter()() +// CHECK: call noundef i32 @hlsl::RasterizerOrderedStructuredBuffer<float>::IncrementCounter()(ptr {{.*}} @ROSB1) +// CHECK-NEXT: ret void + +// CHECK: define {{.*}} i32 @hlsl::RasterizerOrderedStructuredBuffer<float>::IncrementCounter()(ptr {{.*}} %this) +// CHECK: %__counter_handle = getelementptr inbounds nuw %"class.hlsl::RasterizerOrderedStructuredBuffer", ptr %{{.*}}, i32 0, i32 1 +// CHECK-NEXT: %[[COUNTER_HANDLE:.*]] = load target("dx.RawBuffer", float, 1, 1), ptr %__counter_handle +// DXIL-NEXT: %[[VAL:.*]] = call i32 @llvm.dx.resource.updatecounter.tdx.RawBuffer_f32_1_1t(target("dx.RawBuffer", float, 1, 1) %[[COUNTER_HANDLE]], i8 1) +// CHECK-NEXT: ret i32 %[[VAL]] + export void TestDecrementCounter() { -// CHECK: define void @_Z20TestDecrementCounterv() -// CHECK-DXIL: call i32 @llvm.dx.resource.updatecounter.tdx.RawBuffer_f32_1_0t(target("dx.RawBuffer", float, 1, 0) %{{[0-9]+}}, i8 -1) -// CHECK-DXIL: call i32 @llvm.dx.resource.updatecounter.tdx.RawBuffer_f32_1_1t(target("dx.RawBuffer", float, 1, 1) %{{[0-9]+}}, i8 -1) - RWSB2.DecrementCounter(); ROSB2.DecrementCounter(); } +// CHECK: define void @TestDecrementCounter()() +// CHECK: call noundef i32 @hlsl::RasterizerOrderedStructuredBuffer<int vector[2]>::DecrementCounter()(ptr {{.*}} @ROSB2) +// CHECK-NEXT: ret void + +// CHECK: define {{.*}} i32 @hlsl::RasterizerOrderedStructuredBuffer<int vector[2]>::DecrementCounter()(ptr {{.*}} %this) +// CHECK: %__counter_handle = getelementptr inbounds nuw %"class.hlsl::RasterizerOrderedStructuredBuffer.0", ptr %{{.*}}, i32 0, i32 1 +// CHECK-NEXT: %[[COUNTER_HANDLE:.*]] = load target("dx.RawBuffer", <2 x i32>, 1, 1), ptr %__counter_handle +// DXIL-NEXT: %[[VAL:.*]] = call i32 @llvm.dx.resource.updatecounter.tdx.RawBuffer_v2i32_1_1t(target("dx.RawBuffer", <2 x i32>, 1, 1) %[[COUNTER_HANDLE]], i8 -1) +// CHECK-NEXT: ret i32 %[[VAL]] + export float TestLoad() { - return ROSB1.Load(10); + return ROSB1.Load(10).x + ROSB2.Load(20).x; } -// CHECK: define noundef nofpclass(nan inf) float @_Z8TestLoadv() -// CHECK: %[[PTR1:.*]] = call ptr @llvm.dx.resource.getpointer.p0.tdx.RawBuffer_f32_1_1t(target("dx.RawBuffer", float, 1, 1) %{{[0-9]+}}, i32 %{{[0-9]+}}) -// CHECK: %[[VALUE1:.*]] = load float, ptr %[[PTR1]] +// CHECK: define {{.*}} float @TestLoad()() +// CHECK: call {{.*}} float @hlsl::RasterizerOrderedStructuredBuffer<float>::Load(unsigned int)(ptr {{.*}} @ROSB1, i32 noundef 10) +// CHECK: call {{.*}} <2 x i32> @hlsl::RasterizerOrderedStructuredBuffer<int vector[2]>::Load(unsigned int)(ptr {{.*}} @ROSB2, i32 noundef 20) +// CHECK: ret + +// CHECK: define {{.*}} float @hlsl::RasterizerOrderedStructuredBuffer<float>::Load(unsigned int)(ptr {{.*}} %Index) +// CHECK: %__handle = getelementptr inbounds nuw %"class.hlsl::RasterizerOrderedStructuredBuffer", ptr {{.*}}, i32 0, i32 0 +// CHECK-NEXT: %[[HANDLE:.*]] = load target("dx.RawBuffer", float, 1, 1), ptr %__handle +// CHECK-NEXT: %[[INDEX:.*]] = load i32, ptr %Index.addr +// DXIL-NEXT: %[[BUFPTR:.*]] = call ptr @llvm.dx.resource.getpointer.p0.tdx.RawBuffer_f32_1_1t(target("dx.RawBuffer", float, 1, 1) %[[HANDLE]], i32 %[[INDEX]]) +// CHECK-NEXT: %[[VAL:.*]] = load float, ptr %[[BUFPTR]] +// CHECK-NEXT: ret float %[[VAL]] + +// CHECK: define {{.*}} <2 x i32> @hlsl::RasterizerOrderedStructuredBuffer<int vector[2]>::Load(unsigned int)(ptr {{.*}} %Index) +// CHECK: %__handle = getelementptr inbounds nuw %"class.hlsl::RasterizerOrderedStructuredBuffer.0", ptr {{.*}}, i32 0, i32 0 +// CHECK-NEXT: %[[HANDLE:.*]] = load target("dx.RawBuffer", <2 x i32>, 1, 1), ptr %__handle +// CHECK-NEXT: %[[INDEX:.*]] = load i32, ptr %Index.addr +// DXIL-NEXT: %[[BUFPTR:.*]] = call ptr @llvm.dx.resource.getpointer.p0.tdx.RawBuffer_v2i32_1_1t(target("dx.RawBuffer", <2 x i32>, 1, 1) %[[HANDLE]], i32 %[[INDEX]]) +// CHECK-NEXT: %[[VAL:.*]] = load <2 x i32>, ptr %[[BUFPTR]] +// CHECK-NEXT: ret <2 x i32> %[[VAL]] -// CHECK: declare i32 @llvm.dx.resource.updatecounter.tdx.RawBuffer_f32_1_0t(target("dx.RawBuffer", float, 1, 0), i8) -// CHECK: declare i32 @llvm.dx.resource.updatecounter.tdx.RawBuffer_f32_1_1t(target("dx.RawBuffer", float, 1, 1), i8) -// CHECK: declare ptr @llvm.dx.resource.getpointer.p0.tdx.RawBuffer_f32_1_1t(target("dx.RawBuffer", float, 1, 1), i32) +// DXIL: declare i32 @llvm.dx.resource.updatecounter.tdx.RawBuffer_f32_1_1t(target("dx.RawBuffer", float, 1, 1), i8) +// DXIL: declare i32 @llvm.dx.resource.updatecounter.tdx.RawBuffer_v2i32_1_1t(target("dx.RawBuffer", <2 x i32>, 1, 1), i8) +// DXIL: declare ptr @llvm.dx.resource.getpointer.p0.tdx.RawBuffer_f32_1_1t(target("dx.RawBuffer", float, 1, 1), i32) +// DXIL: declare ptr @llvm.dx.resource.getpointer.p0.tdx.RawBuffer_v2i32_1_1t(target("dx.RawBuffer", <2 x i32>, 1, 1), i32) diff --git a/clang/test/CodeGenHLSL/resources/RWBuffer-constructor.hlsl b/clang/test/CodeGenHLSL/resources/TypedBuffers-constructor.hlsl index ca33c42..1ec9f0f 100644 --- a/clang/test/CodeGenHLSL/resources/RWBuffer-constructor.hlsl +++ b/clang/test/CodeGenHLSL/resources/TypedBuffers-constructor.hlsl @@ -1,5 +1,5 @@ // RUN: %clang_cc1 -triple dxil-pc-shadermodel6.3-library -emit-llvm -disable-llvm-passes -o - %s | \ -// RUN: llvm-cxxfilt | FileCheck %s --check-prefixes=CHECK,CHECK-DXIL +// RUN: llvm-cxxfilt | FileCheck %s --check-prefixes=CHECK,CHECK-DXIL // FIXME: SPIR-V codegen of llvm.spv.resource.handlefrombinding and resource types is not yet implemented // RUN-DISABLED: %clang_cc1 -triple spirv-vulkan-library -x hlsl -emit-llvm -disable-llvm-passes -o - %s | \ // llvm-cxxfilt | FileCheck %s --check-prefixes=CHECK,CHECK-SPIRV @@ -14,7 +14,7 @@ RWBuffer<float> Buf1 : register(u5, space3); // Resource with implicit binding -RWBuffer<double> Buf2; +Buffer<double> Buf2; export void foo() { // Local resource declaration @@ -22,12 +22,12 @@ export void foo() { } // CHECK: %"class.hlsl::RWBuffer" = type { target("dx.TypedBuffer", float, 1, 0, 0) } -// CHECK: %"class.hlsl::RWBuffer.0" = type { target("dx.TypedBuffer", double, 1, 0, 0) } -// CHECK: %"class.hlsl::RWBuffer.1" = type { target("dx.TypedBuffer", i32, 1, 0, 1) } +// CHECK: %"class.hlsl::Buffer" = type { target("dx.TypedBuffer", double, 0, 0, 0) } +// CHECK: %"class.hlsl::RWBuffer.0" = type { target("dx.TypedBuffer", i32, 1, 0, 1) } // CHECK: @Buf1 = internal global %"class.hlsl::RWBuffer" poison, align 4 // CHECK: @[[Buf1Str:.*]] = private unnamed_addr constant [5 x i8] c"Buf1\00", align 1 -// CHECK: @Buf2 = internal global %"class.hlsl::RWBuffer.0" poison, align 4 +// CHECK: @Buf2 = internal global %"class.hlsl::Buffer" poison, align 4 // CHECK: @[[Buf2Str:.*]] = private unnamed_addr constant [5 x i8] c"Buf2\00", align 1 // Buf1 initialization part 1 - global init function that calls RWBuffer<float>::__createFromBinding @@ -50,24 +50,24 @@ export void foo() { // Buf2 initialization part 1 - global init function that RWBuffer<float>::__createFromImplicitBinding // CHECK: define internal void @__cxx_global_var_init.1() // CHECK-NEXT: entry: -// CHECK-NEXT: call void @hlsl::RWBuffer<double>::__createFromImplicitBinding(unsigned int, unsigned int, int, unsigned int, char const*) +// CHECK-NEXT: call void @hlsl::Buffer<double>::__createFromImplicitBinding(unsigned int, unsigned int, int, unsigned int, char const*) // CHECK-SAME: (ptr {{.*}} @Buf2, i32 noundef 0, i32 noundef 0, i32 noundef 1, i32 noundef 0, ptr noundef @[[Buf2Str]]) -// Buf2 initialization part 2 - body of RWBuffer<float>::__createFromImplicitBinding call -// CHECK: define linkonce_odr hidden void @hlsl::RWBuffer<double>::__createFromImplicitBinding(unsigned int, unsigned int, int, unsigned int, char const*) -// CHECK-SAME: (ptr {{.*}} sret(%"class.hlsl::RWBuffer.0") align 4 %[[RetValue2:.*]], i32 noundef %orderId, +// Buf2 initialization part 2 - body of Buffer<double>::__createFromImplicitBinding call +// CHECK: define linkonce_odr hidden void @hlsl::Buffer<double>::__createFromImplicitBinding(unsigned int, unsigned int, int, unsigned int, char const*) +// CHECK-SAME: (ptr {{.*}} sret(%"class.hlsl::Buffer") align 4 %[[RetValue2:.*]], i32 noundef %orderId, // CHECK-SAME: i32 noundef %spaceNo, i32 noundef %range, i32 noundef %index, ptr noundef %name) -// CHECK: %[[Tmp2:.*]] = alloca %"class.hlsl::RWBuffer.0", align 4 -// CHECK: %[[Handle2:.*]] = call target("dx.TypedBuffer", double, 1, 0, 0) -// CHECK-SAME: @llvm.dx.resource.handlefromimplicitbinding.tdx.TypedBuffer_f64_1_0_0t( -// CHECK: %__handle = getelementptr inbounds nuw %"class.hlsl::RWBuffer.0", ptr %[[Tmp2]], i32 0, i32 0 -// CHECK-DXIL: store target("dx.TypedBuffer", double, 1, 0, 0) %[[Handle2]], ptr %__handle, align 4 -// CHECK: call void @hlsl::RWBuffer<double>::RWBuffer(hlsl::RWBuffer<double> const&)(ptr {{.*}} %[[RetValue2]], ptr {{.*}} %[[Tmp2]]) +// CHECK: %[[Tmp2:.*]] = alloca %"class.hlsl::Buffer", align 4 +// CHECK: %[[Handle2:.*]] = call target("dx.TypedBuffer", double, 0, 0, 0) +// CHECK-SAME: @llvm.dx.resource.handlefromimplicitbinding.tdx.TypedBuffer_f64_0_0_0t( +// CHECK: %__handle = getelementptr inbounds nuw %"class.hlsl::Buffer", ptr %[[Tmp2]], i32 0, i32 0 +// CHECK-DXIL: store target("dx.TypedBuffer", double, 0, 0, 0) %[[Handle2]], ptr %__handle, align 4 +// CHECK: call void @hlsl::Buffer<double>::Buffer(hlsl::Buffer<double> const&)(ptr {{.*}} %[[RetValue2]], ptr {{.*}} %[[Tmp2]]) // Buf3 initialization part 1 - local variable declared in function foo() is initialized by RWBuffer<int> C1 default constructor // CHECK: define void @foo() // CHECK-NEXT: entry: -// CHECK-NEXT: %Buf3 = alloca %"class.hlsl::RWBuffer.1", align 4 +// CHECK-NEXT: %Buf3 = alloca %"class.hlsl::RWBuffer.0", align 4 // CHECK-NEXT: call void @hlsl::RWBuffer<int>::RWBuffer()(ptr {{.*}} %Buf3) // Buf3 initialization part 2 - body of RWBuffer<int> default C1 constructor that calls the default C2 constructor @@ -76,11 +76,11 @@ export void foo() { // Buf3 initialization part 3 - body of RWBuffer<int> default C2 constructor that initializes handle to poison // CHECK: define linkonce_odr hidden void @hlsl::RWBuffer<int>::RWBuffer()(ptr {{.*}} %this) -// CHECK: %__handle = getelementptr inbounds nuw %"class.hlsl::RWBuffer.1", ptr %{{.*}}, i32 0, i32 0 +// CHECK: %__handle = getelementptr inbounds nuw %"class.hlsl::RWBuffer.0", ptr %{{.*}}, i32 0, i32 0 // CHECK-NEXT: store target("dx.TypedBuffer", i32, 1, 0, 1) poison, ptr %__handle, align 4 // Module initialization -// CHECK: define internal void @_GLOBAL__sub_I_RWBuffer_constructor.hlsl() +// CHECK: define internal void @_GLOBAL__sub_I_TypedBuffers_constructor.hlsl() // CHECK-NEXT: entry: // CHECK-NEXT: call void @__cxx_global_var_init() // CHECK-NEXT: call void @__cxx_global_var_init.1() diff --git a/clang/test/CodeGenHLSL/resources/TypedBuffers-elementtype.hlsl b/clang/test/CodeGenHLSL/resources/TypedBuffers-elementtype.hlsl new file mode 100644 index 0000000..d3dba8a --- /dev/null +++ b/clang/test/CodeGenHLSL/resources/TypedBuffers-elementtype.hlsl @@ -0,0 +1,94 @@ +// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.2-compute -finclude-default-header -fnative-half-type \ +// RUN: -emit-llvm -o - -DRESOURCE=Buffer %s | FileCheck %s -DRESOURCE=Buffer -DRW=0 -check-prefixes=DXIL + +// RUN: %clang_cc1 -triple spirv-pc-vulkan-compute -finclude-default-header -fnative-half-type \ +// RUN: -emit-llvm -o - -DRESOURCE=Buffer %s | FileCheck %s -DRESOURCE=Buffer -DRW=1 -check-prefixes=SPV-RO + +// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.2-compute -finclude-default-header -fnative-half-type \ +// RUN: -emit-llvm -o - -DRESOURCE=RWBuffer %s | FileCheck %s -DRESOURCE=RWBuffer -DRW=1 -check-prefixes=DXIL + +// RUN: %clang_cc1 -triple spirv-pc-vulkan-compute -finclude-default-header -fnative-half-type \ +// RUN: -emit-llvm -o - -DRESOURCE=RWBuffer %s | FileCheck %s -DRESOURCE=RWBuffer --DRW=2 -check-prefixes=SPV-RW + +// DXIL: %"class.hlsl::[[RESOURCE]]" = type { target("dx.TypedBuffer", i16, [[RW]], 0, 1) } +// DXIL: %"class.hlsl::[[RESOURCE]].0" = type { target("dx.TypedBuffer", i16, [[RW]], 0, 0) } +// DXIL: %"class.hlsl::[[RESOURCE]].1" = type { target("dx.TypedBuffer", i32, [[RW]], 0, 1) } +// DXIL: %"class.hlsl::[[RESOURCE]].2" = type { target("dx.TypedBuffer", i32, [[RW]], 0, 0) } +// DXIL: %"class.hlsl::[[RESOURCE]].3" = type { target("dx.TypedBuffer", i64, [[RW]], 0, 1) } +// DXIL: %"class.hlsl::[[RESOURCE]].4" = type { target("dx.TypedBuffer", i64, [[RW]], 0, 0) } +// DXIL: %"class.hlsl::[[RESOURCE]].5" = type { target("dx.TypedBuffer", half, [[RW]], 0, 0) } +// DXIL: %"class.hlsl::[[RESOURCE]].6" = type { target("dx.TypedBuffer", float, [[RW]], 0, 0) } +// DXIL: %"class.hlsl::[[RESOURCE]].7" = type { target("dx.TypedBuffer", double, [[RW]], 0, 0) } +// DXIL: %"class.hlsl::[[RESOURCE]].8" = type { target("dx.TypedBuffer", <4 x i16>, [[RW]], 0, 1) } +// DXIL: %"class.hlsl::[[RESOURCE]].9" = type { target("dx.TypedBuffer", <3 x i32>, [[RW]], 0, 0) } +// DXIL: %"class.hlsl::[[RESOURCE]].10" = type { target("dx.TypedBuffer", <2 x half>, [[RW]], 0, 0) } +// DXIL: %"class.hlsl::[[RESOURCE]].11" = type { target("dx.TypedBuffer", <3 x float>, [[RW]], 0, 0) } +// DXIL: %"class.hlsl::[[RESOURCE]].12" = type { target("dx.TypedBuffer", <4 x i32>, [[RW]], 0, 1) } + +// SPV-RO: %"class.hlsl::[[RESOURCE]]" = type { target("spirv.SignedImage", i16, 5, 2, 0, 0, 1, 0) } +// SPV-RO: %"class.hlsl::[[RESOURCE]].0" = type { target("spirv.Image", i16, 5, 2, 0, 0, 1, 0) } +// SPV-RO: %"class.hlsl::[[RESOURCE]].1" = type { target("spirv.SignedImage", i32, 5, 2, 0, 0, 1, 0) } +// SPV-RO: %"class.hlsl::[[RESOURCE]].2" = type { target("spirv.Image", i32, 5, 2, 0, 0, 1, 0) } +// SPV-RO: %"class.hlsl::[[RESOURCE]].3" = type { target("spirv.SignedImage", i64, 5, 2, 0, 0, 1, 0) } +// SPV-RO: %"class.hlsl::[[RESOURCE]].4" = type { target("spirv.Image", i64, 5, 2, 0, 0, 1, 0) } +// SPV-RO: %"class.hlsl::[[RESOURCE]].5" = type { target("spirv.Image", half, 5, 2, 0, 0, 1, 0) } +// SPV-RO: %"class.hlsl::[[RESOURCE]].6" = type { target("spirv.Image", float, 5, 2, 0, 0, 1, 0) } +// SPV-RO: %"class.hlsl::[[RESOURCE]].7" = type { target("spirv.Image", double, 5, 2, 0, 0, 1, 0) } +// SPV-RO: %"class.hlsl::[[RESOURCE]].8" = type { target("spirv.SignedImage", i16, 5, 2, 0, 0, 1, 0) } +// SPV-RO: %"class.hlsl::[[RESOURCE]].9" = type { target("spirv.Image", i32, 5, 2, 0, 0, 1, 0) } +// SPV-RO: %"class.hlsl::[[RESOURCE]].10" = type { target("spirv.Image", half, 5, 2, 0, 0, 1, 0) } +// SPV-RO: %"class.hlsl::[[RESOURCE]].11" = type { target("spirv.Image", float, 5, 2, 0, 0, 1, 0) } +// SPV-RO: %"class.hlsl::[[RESOURCE]].12" = type { target("spirv.SignedImage", i32, 5, 2, 0, 0, 1, 0) } + +// SPV-RW: %"class.hlsl::[[RESOURCE]]" = type { target("spirv.SignedImage", i16, 5, 2, 0, 0, 2, 0) } +// SPV-RW: %"class.hlsl::[[RESOURCE]].0" = type { target("spirv.Image", i16, 5, 2, 0, 0, 2, 0) } +// SPV-RW: %"class.hlsl::[[RESOURCE]].1" = type { target("spirv.SignedImage", i32, 5, 2, 0, 0, 2, 24) } +// SPV-RW: %"class.hlsl::[[RESOURCE]].2" = type { target("spirv.Image", i32, 5, 2, 0, 0, 2, 33) } +// SPV-RW: %"class.hlsl::[[RESOURCE]].3" = type { target("spirv.SignedImage", i64, 5, 2, 0, 0, 2, 41) } +// SPV-RW: %"class.hlsl::[[RESOURCE]].4" = type { target("spirv.Image", i64, 5, 2, 0, 0, 2, 40) } +// SPV-RW: %"class.hlsl::[[RESOURCE]].5" = type { target("spirv.Image", half, 5, 2, 0, 0, 2, 0) } +// SPV-RW: %"class.hlsl::[[RESOURCE]].6" = type { target("spirv.Image", float, 5, 2, 0, 0, 2, 3) } +// SPV-RW: %"class.hlsl::[[RESOURCE]].7" = type { target("spirv.Image", double, 5, 2, 0, 0, 2, 0) } +// SPV-RW: %"class.hlsl::[[RESOURCE]].8" = type { target("spirv.SignedImage", i16, 5, 2, 0, 0, 2, 0) } +// SPV-RW: %"class.hlsl::[[RESOURCE]].9" = type { target("spirv.Image", i32, 5, 2, 0, 0, 2, 0) } +// SPV-RW: %"class.hlsl::[[RESOURCE]].10" = type { target("spirv.Image", half, 5, 2, 0, 0, 2, 0) } +// SPV-RW: %"class.hlsl::[[RESOURCE]].11" = type { target("spirv.Image", float, 5, 2, 0, 0, 2, 0) } +// SPV-RW: %"class.hlsl::[[RESOURCE]].12" = type { target("spirv.SignedImage", i32, 5, 2, 0, 0, 2, 21) } + +RESOURCE<int16_t> BufI16; +RESOURCE<uint16_t> BufU16; +RESOURCE<int> BufI32; +RESOURCE<uint> BufU32; +RESOURCE<int64_t> BufI64; +RESOURCE<uint64_t> BufU64; +RESOURCE<half> BufF16; +RESOURCE<float> BufF32; +RESOURCE<double> BufF64; +RESOURCE< vector<int16_t, 4> > BufI16x4; +RESOURCE< vector<uint, 3> > BufU32x3; +RESOURCE<half2> BufF16x2; +RESOURCE<float3> BufF32x3; +RESOURCE<int4> BufI32x4; +// TODO: RESOURCE<snorm half> BufSNormF16; -> 11 +// TODO: RESOURCE<unorm half> BufUNormF16; -> 12 +// TODO: RESOURCE<snorm float> BufSNormF32; -> 13 +// TODO: RESOURCE<unorm float> BufUNormF32; -> 14 +// TODO: RESOURCE<snorm double> BufSNormF64; -> 15 +// TODO: RESOURCE<unorm double> BufUNormF64; -> 16 + +[numthreads(1,1,1)] +void main(int GI : SV_GroupIndex) { + int16_t v1 = BufI16[GI]; + uint16_t v2 = BufU16[GI]; + int v3 = BufI32[GI]; + uint v4 = BufU32[GI]; + int64_t v5 = BufI64[GI]; + uint64_t v6 = BufU64[GI]; + half v7 = BufF16[GI]; + float v8 = BufF32[GI]; + double v9 = BufF64[GI]; + vector<int16_t,4> v10 = BufI16x4[GI]; + vector<int, 3> v11 = BufU32x3[GI]; + half2 v12 = BufF16x2[GI]; + float3 v13 = BufF32x3[GI]; +} diff --git a/clang/test/CodeGenHLSL/resources/TypedBuffers-methods.hlsl b/clang/test/CodeGenHLSL/resources/TypedBuffers-methods.hlsl new file mode 100644 index 0000000..b153bda --- /dev/null +++ b/clang/test/CodeGenHLSL/resources/TypedBuffers-methods.hlsl @@ -0,0 +1,42 @@ +// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.3-library -finclude-default-header -emit-llvm -disable-llvm-passes -o - %s | llvm-cxxfilt | FileCheck %s --check-prefixes=CHECK,DXIL +// RUN-DISABLED: %clang_cc1 -triple spirv-vulkan-library -finclude-default-header -emit-llvm -disable-llvm-passes -o - %s | llvm-cxxfilt | FileCheck %s --check-prefixes=CHECK,SPIRV + +// NOTE: SPIRV codegen for resource methods is not yet implemented + +Buffer<float> Buf : register(t0); +RWBuffer<uint4> RWBuf : register(u0); + +// DXIL: %"class.hlsl::Buffer" = type { target("dx.TypedBuffer", float, 0, 0, 0) } +// DXIL: %"class.hlsl::RWBuffer" = type { target("dx.TypedBuffer", <4 x i32>, 1, 0, 0) } + +// DXIL: @Buf = internal global %"class.hlsl::Buffer" poison +// DXIL: @RWBuf = internal global %"class.hlsl::RWBuffer" poison + +export float TestLoad() { + return Buf.Load(1) + RWBuf.Load(2).y; +} + +// CHECK: define noundef nofpclass(nan inf) float @TestLoad()() +// CHECK: call {{.*}} float @hlsl::Buffer<float>::Load(unsigned int)(ptr {{.*}} @Buf, i32 noundef 1) +// CHECK: call {{.*}} <4 x i32> @hlsl::RWBuffer<unsigned int vector[4]>::Load(unsigned int)(ptr {{.*}} @RWBuf, i32 noundef 2) +// CHECK: add +// CHECK: ret float + +// CHECK: define {{.*}} float @hlsl::Buffer<float>::Load(unsigned int)(ptr {{.*}} %this, i32 noundef %Index) +// CHECK: %__handle = getelementptr inbounds nuw %"class.hlsl::Buffer", ptr %{{.*}}, i32 0, i32 0 +// DXIL-NEXT: %[[HANDLE:.*]] = load target("dx.TypedBuffer", float, 0, 0, 0), ptr %__handle +// CHECK-NEXT: %[[INDEX:.*]] = load i32, ptr %Index.addr +// DXIL-NEXT: %[[PTR:.*]] = call ptr @llvm.dx.resource.getpointer.p0.tdx.TypedBuffer_f32_0_0_0t(target("dx.TypedBuffer", float, 0, 0, 0) %[[HANDLE]], i32 %[[INDEX]]) +// CHECK-NEXT: %[[VAL:.*]] = load float, ptr %[[PTR]] +// CHECK-NEXT: ret float %[[VAL]] + +// CHECK: define {{.*}} <4 x i32> @hlsl::RWBuffer<unsigned int vector[4]>::Load(unsigned int)(ptr {{.*}} %this, i32 noundef %Index) +// CHECK: %__handle = getelementptr inbounds nuw %"class.hlsl::RWBuffer", ptr %{{.*}}, i32 0, i32 0 +// DXIL-NEXT: %[[HANDLE:.*]] = load target("dx.TypedBuffer", <4 x i32>, 1, 0, 0), ptr %__handle +// CHECK-NEXT: %[[INDEX:.*]] = load i32, ptr %Index.addr +// DXIL-NEXT: %[[PTR:.*]] = call ptr @llvm.dx.resource.getpointer.p0.tdx.TypedBuffer_v4i32_1_0_0t(target("dx.TypedBuffer", <4 x i32>, 1, 0, 0) %[[HANDLE]], i32 %[[INDEX]]) +// CHECK-NEXT: %[[VEC:.*]] = load <4 x i32>, ptr %[[PTR]] +// CHECK-NEXT: ret <4 x i32> %[[VEC]] + +// DXIL: declare ptr @llvm.dx.resource.getpointer.p0.tdx.TypedBuffer_f32_0_0_0t(target("dx.TypedBuffer", float, 0, 0, 0), i32) +// DXIL: declare ptr @llvm.dx.resource.getpointer.p0.tdx.TypedBuffer_v4i32_1_0_0t(target("dx.TypedBuffer", <4 x i32>, 1, 0, 0), i32) diff --git a/clang/test/CodeGenHLSL/resources/TypedBuffers-subscript.hlsl b/clang/test/CodeGenHLSL/resources/TypedBuffers-subscript.hlsl new file mode 100644 index 0000000..adc35f6 --- /dev/null +++ b/clang/test/CodeGenHLSL/resources/TypedBuffers-subscript.hlsl @@ -0,0 +1,26 @@ +// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.0-compute -emit-llvm -o - -O0 %s | FileCheck %s --check-prefixes=DXIL,CHECK +// RUN: %clang_cc1 -triple spirv1.6-pc-vulkan1.3-compute -fspv-use-unknown-image-format -emit-llvm -o - -O0 %s | FileCheck %s --check-prefixes=SPIRV,CHECK + +Buffer<int> In; +RWBuffer<int> Out; + +[numthreads(1,1,1)] +void main(unsigned GI : SV_GroupIndex) { + // CHECK: define void @main() + + // DXIL: %[[INPTR:.*]] = call {{.*}} ptr @llvm.dx.resource.getpointer.p0.tdx.TypedBuffer_i32_0_0_1t(target("dx.TypedBuffer", i32, 0, 0, 1) %{{.*}}, i32 %{{.*}}) + // SPIRV: %[[INPTR:.*]] = call {{.*}} ptr addrspace(11) @llvm.spv.resource.getpointer.p11.tspirv.SignedImage_i32_5_2_0_0_1_0t(target("spirv.SignedImage", i32, 5, 2, 0, 0, 1, 0) %{{.*}}, i32 %{{.*}}) + // CHECK: %[[LOAD:.*]] = load i32, ptr {{.*}}%[[INPTR]] + // DXIL: %[[OUTPTR:.*]] = call {{.*}} ptr @llvm.dx.resource.getpointer.p0.tdx.TypedBuffer_i32_1_0_1t(target("dx.TypedBuffer", i32, 1, 0, 1) %{{.*}}, i32 %{{.*}}) + // SPIRV: %[[OUTPTR:.*]] = call {{.*}} ptr addrspace(11) @llvm.spv.resource.getpointer.p11.tspirv.SignedImage_i32_5_2_0_0_2_0t(target("spirv.SignedImage", i32, 5, 2, 0, 0, 2, 0) %{{.*}}, i32 %{{.*}}) + // CHECK: store i32 %[[LOAD]], ptr {{.*}}%[[OUTPTR]] + Out[GI] = In[GI]; + + // DXIL: %[[INPTR:.*]] = call {{.*}} ptr @llvm.dx.resource.getpointer.p0.tdx.TypedBuffer_i32_1_0_1t(target("dx.TypedBuffer", i32, 1, 0, 1) %{{.*}}, i32 %{{.*}}) + // SPIRV: %[[INPTR:.*]] = call {{.*}} ptr addrspace(11) @llvm.spv.resource.getpointer.p11.tspirv.SignedImage_i32_5_2_0_0_2_0t(target("spirv.SignedImage", i32, 5, 2, 0, 0, 2, 0) %{{.*}}, i32 %{{.*}}) + // CHECK: %[[LOAD:.*]] = load i32, ptr {{.*}}%[[INPTR]] + // DXIL: %[[OUTPTR:.*]] = call {{.*}} ptr @llvm.dx.resource.getpointer.p0.tdx.TypedBuffer_i32_1_0_1t(target("dx.TypedBuffer", i32, 1, 0, 1) %{{.*}}, i32 %{{.*}}) + // SPIRV: %[[OUTPTR:.*]] = call {{.*}} ptr addrspace(11) @llvm.spv.resource.getpointer.p11.tspirv.SignedImage_i32_5_2_0_0_2_0t(target("spirv.SignedImage", i32, 5, 2, 0, 0, 2, 0) %{{.*}}, i32 %{{.*}}) + // CHECK: store i32 %[[LOAD]], ptr {{.*}}%[[OUTPTR]] + Out[GI + 1] = Out[GI]; +} diff --git a/clang/test/CodeGenHLSL/vk-features/maximal_reconvergence.hlsl b/clang/test/CodeGenHLSL/vk-features/maximal_reconvergence.hlsl new file mode 100644 index 0000000..f23ac7c --- /dev/null +++ b/clang/test/CodeGenHLSL/vk-features/maximal_reconvergence.hlsl @@ -0,0 +1,17 @@ +// RUN: %clang_cc1 -triple spirv1.6-unknown-vulkan1.3-compute -fspv-enable-maximal-reconvergence -emit-llvm -o - -O0 %s | FileCheck %s --check-prefixes=CHECK +// RUN: %clang_cc1 -triple spirv1.6-unknown-vulkan1.3-compute -hlsl-entry test -fspv-enable-maximal-reconvergence -emit-llvm -o - -O0 %s | FileCheck %s --check-prefixes=CHECK-ENTRY + +[numthreads(1,1,1)] +void main() { +// CHECK: define void @main() [[attributeNumber:#[0-9]+]] { +} + +// CHECK: attributes [[attributeNumber]] = {{.*}} "enable-maximal-reconvergence"="true" {{.*}} + + +[numthreads(1,1,1)] +void test() { +// CHECK-ENTRY: define void @test() [[attributeNumber:#[0-9]+]] { +} + +// CHECK-ENTRY: attributes [[attributeNumber]] = {{.*}} "enable-maximal-reconvergence"="true" {{.*}} diff --git a/clang/test/CodeGenOpenCL/amdgpu-env-amdgcn.cl b/clang/test/CodeGenOpenCL/amdgpu-env-amdgcn.cl index 713ae48..72ce726 100644 --- a/clang/test/CodeGenOpenCL/amdgpu-env-amdgcn.cl +++ b/clang/test/CodeGenOpenCL/amdgpu-env-amdgcn.cl @@ -1,5 +1,5 @@ // RUN: %clang_cc1 %s -O0 -triple amdgcn -emit-llvm -o - | FileCheck %s // RUN: %clang_cc1 %s -O0 -triple amdgcn---opencl -emit-llvm -o - | FileCheck %s -// CHECK: target datalayout = "e-p:64:64-p1:64:64-p2:32:32-p3:32:32-p4:64:64-p5:32:32-p6:32:32-p7:160:256:256:32-p8:128:128:128:48-p9:192:256:256:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-v2048:2048-n32:64-S32-A5-G1-ni:7:8:9" +// CHECK: target datalayout = "e-m:e-p:64:64-p1:64:64-p2:32:32-p3:32:32-p4:64:64-p5:32:32-p6:32:32-p7:160:256:256:32-p8:128:128:128:48-p9:192:256:256:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-v2048:2048-n32:64-S32-A5-G1-ni:7:8:9" void foo(void) {} diff --git a/clang/test/DebugInfo/CXX/versioned-language.cpp b/clang/test/DebugInfo/CXX/versioned-language.cpp new file mode 100644 index 0000000..4cb2b29 --- /dev/null +++ b/clang/test/DebugInfo/CXX/versioned-language.cpp @@ -0,0 +1,23 @@ +// RUN: %clang_cc1 -emit-llvm %s -o - -debug-info-kind=limited -dwarf-version=5 -std=c++98 \ +// RUN: | FileCheck %s --implicit-check-not "sourceLanguageName" --implicit-check-not "sourceLanguageVersion" +// +// RUN: %clang_cc1 -emit-llvm %s -o - -debug-info-kind=limited -dwarf-version=6 -std=c++98 | FileCheck %s --check-prefix=CHECK-CPP98 +// RUN: %clang_cc1 -emit-llvm %s -o - -debug-info-kind=limited -dwarf-version=6 -std=c++03 | FileCheck %s --check-prefix=CHECK-CPP03 +// RUN: %clang_cc1 -emit-llvm %s -o - -debug-info-kind=limited -dwarf-version=6 -std=c++11 | FileCheck %s --check-prefix=CHECK-CPP11 +// RUN: %clang_cc1 -emit-llvm %s -o - -debug-info-kind=limited -dwarf-version=6 -std=c++14 | FileCheck %s --check-prefix=CHECK-CPP14 +// RUN: %clang_cc1 -emit-llvm %s -o - -debug-info-kind=limited -dwarf-version=6 -std=c++17 | FileCheck %s --check-prefix=CHECK-CPP17 +// RUN: %clang_cc1 -emit-llvm %s -o - -debug-info-kind=limited -dwarf-version=6 -std=c++20 | FileCheck %s --check-prefix=CHECK-CPP20 +// RUN: %clang_cc1 -emit-llvm %s -o - -debug-info-kind=limited -dwarf-version=6 -std=c++23 | FileCheck %s --check-prefix=CHECK-CPP23 +// RUN: %clang_cc1 -emit-llvm %s -o - -debug-info-kind=limited -dwarf-version=6 -std=c++2c | FileCheck %s --check-prefix=CHECK-CPP2C + +struct Foo {} globalVar; + +// CHECK-CPP98: !DICompileUnit(sourceLanguageName: DW_LNAME_C_plus_plus, sourceLanguageVersion: 199711 +// FIXME: C++03 technically has no official standard version code. From Clang's point of view C++03 and C++98 are interchangable. +// CHECK-CPP03: !DICompileUnit(sourceLanguageName: DW_LNAME_C_plus_plus, sourceLanguageVersion: 199711 +// CHECK-CPP11: !DICompileUnit(sourceLanguageName: DW_LNAME_C_plus_plus, sourceLanguageVersion: 201103 +// CHECK-CPP14: !DICompileUnit(sourceLanguageName: DW_LNAME_C_plus_plus, sourceLanguageVersion: 201402 +// CHECK-CPP17: !DICompileUnit(sourceLanguageName: DW_LNAME_C_plus_plus, sourceLanguageVersion: 201703 +// CHECK-CPP20: !DICompileUnit(sourceLanguageName: DW_LNAME_C_plus_plus, sourceLanguageVersion: 202002 +// CHECK-CPP23: !DICompileUnit(sourceLanguageName: DW_LNAME_C_plus_plus, sourceLanguageVersion: 202302 +// CHECK-CPP2C: !DICompileUnit(sourceLanguageName: DW_LNAME_C_plus_plus, sourceLanguageVersion: 202400 diff --git a/clang/test/DebugInfo/Generic/versioned-language.c b/clang/test/DebugInfo/Generic/versioned-language.c new file mode 100644 index 0000000..1faa7b4 --- /dev/null +++ b/clang/test/DebugInfo/Generic/versioned-language.c @@ -0,0 +1,17 @@ +// RUN: %clang_cc1 -emit-llvm %s -o - -debug-info-kind=limited -dwarf-version=5 -std=c99 \ +// RUN: | FileCheck %s --implicit-check-not "sourceLanguageName" --implicit-check-not "sourceLanguageVersion" +// +// RUN: %clang_cc1 -emit-llvm %s -o - -debug-info-kind=limited -dwarf-version=6 -std=c89 | FileCheck %s --check-prefix=CHECK-C89 --implicit-check-not "sourceLanguageVersion" +// RUN: %clang_cc1 -emit-llvm %s -o - -debug-info-kind=limited -dwarf-version=6 -std=c99 | FileCheck %s --check-prefix=CHECK-C99 +// RUN: %clang_cc1 -emit-llvm %s -o - -debug-info-kind=limited -dwarf-version=6 -std=c17 | FileCheck %s --check-prefix=CHECK-C17 +// RUN: %clang_cc1 -emit-llvm %s -o - -debug-info-kind=limited -dwarf-version=6 -std=c23 | FileCheck %s --check-prefix=CHECK-C23 +// RUN: %clang_cc1 -emit-llvm %s -o - -debug-info-kind=limited -dwarf-version=6 -std=c2y | FileCheck %s --check-prefix=CHECK-C2Y + +int globalVar = 10; + +// CHECK-C89: !DICompileUnit(sourceLanguageName: DW_LNAME_C, +// CHECK-C99: !DICompileUnit(sourceLanguageName: DW_LNAME_C, sourceLanguageVersion: 199901 +// CHECK-C11: !DICompileUnit(sourceLanguageName: DW_LNAME_C, sourceLanguageVersion: 201112 +// CHECK-C17: !DICompileUnit(sourceLanguageName: DW_LNAME_C, sourceLanguageVersion: 201710 +// CHECK-C23: !DICompileUnit(sourceLanguageName: DW_LNAME_C, sourceLanguageVersion: 202311 +// CHECK-C2Y: !DICompileUnit(sourceLanguageName: DW_LNAME_C, sourceLanguageVersion: 202400 diff --git a/clang/test/DebugInfo/ObjC/versioned-language.m b/clang/test/DebugInfo/ObjC/versioned-language.m new file mode 100644 index 0000000..178c47b --- /dev/null +++ b/clang/test/DebugInfo/ObjC/versioned-language.m @@ -0,0 +1,9 @@ +// RUN: %clang_cc1 -emit-llvm %s -o - -debug-info-kind=limited -dwarf-version=5 \ +// RUN: | FileCheck %s --implicit-check-not "sourceLanguageName" --implicit-check-not "sourceLanguageVersion" +// +// RUN: %clang_cc1 -emit-llvm %s -o - -debug-info-kind=limited -dwarf-version=6 \ +// RUN: | FileCheck %s --implicit-check-not "sourceLanguageVersion" --check-prefix=CHECK-OBJC + +int globalVar = 10; + +// CHECK-OBJC: !DICompileUnit(sourceLanguageName: DW_LNAME_ObjC, diff --git a/clang/test/DebugInfo/ObjCXX/versioned-language.mm b/clang/test/DebugInfo/ObjCXX/versioned-language.mm new file mode 100644 index 0000000..bfdce46 --- /dev/null +++ b/clang/test/DebugInfo/ObjCXX/versioned-language.mm @@ -0,0 +1,9 @@ +// RUN: %clang_cc1 -emit-llvm %s -o - -debug-info-kind=limited -dwarf-version=5 \ +// RUN: | FileCheck %s --implicit-check-not "sourceLanguageName" --implicit-check-not "sourceLanguageVersion" +// +// RUN: %clang_cc1 -emit-llvm %s -o - -debug-info-kind=limited -dwarf-version=6 \ +// RUN: | FileCheck %s --implicit-check-not "sourceLanguageVersion" --check-prefix=CHECK-OBJCXX + +int globalVar = 10; + +// CHECK-OBJCXX: !DICompileUnit(sourceLanguageName: DW_LNAME_ObjC_plus_plus, diff --git a/clang/test/Driver/gpu-libc-headers.c b/clang/test/Driver/gpu-libc-headers.c deleted file mode 100644 index 1802919..0000000 --- a/clang/test/Driver/gpu-libc-headers.c +++ /dev/null @@ -1,14 +0,0 @@ -// RUN: %clang -### --target=amdgcn-amd-amdhsa -mcpu=gfx90a --sysroot=%S/Inputs/basic_gpu_tree \ -// RUN: -ccc-install-dir %S/Inputs/basic_gpu_tree/bin -nogpulib %s 2>&1 | FileCheck %s --check-prefix=CHECK-HEADERS-AMDGPU -// RUN: %clang -### --target=nvptx64-nvidia-cuda -march=sm_89 --sysroot=%S/Inputs/basic_gpu_tree \ -// RUN: -ccc-install-dir %S/Inputs/basic_gpu_tree/bin -nogpulib %s 2>&1 | FileCheck %s --check-prefix=CHECK-HEADERS-NVPTX -// CHECK-HEADERS-AMDGPU: "-cc1"{{.*}}"-isysroot"{{.*}}"-internal-isystem" "{{.*}}include{{.*}}amdgcn-amd-amdhsa" -// CHECK-HEADERS-NVPTX: "-cc1"{{.*}}"-isysroot"{{.*}}"-internal-isystem" "{{.*}}include{{.*}}nvptx64-nvidia-cuda" - -// RUN: %clang -### --target=amdgcn-amd-amdhsa -mcpu=gfx1030 -nogpulib \ -// RUN: -nogpuinc %s 2>&1 | FileCheck %s --check-prefix=CHECK-HEADERS-DISABLED -// RUN: %clang -### --target=amdgcn-amd-amdhsa -mcpu=gfx1030 -nogpulib \ -// RUN: -nostdinc %s 2>&1 | FileCheck %s --check-prefix=CHECK-HEADERS-DISABLED -// RUN: %clang -### --target=amdgcn-amd-amdhsa -mcpu=gfx1030 -nogpulib \ -// RUN: -nobuiltininc %s 2>&1 | FileCheck %s --check-prefix=CHECK-HEADERS-DISABLED -// CHECK-HEADERS-DISABLED-NOT: "-cc1"{{.*}}"-internal-isystem" "{{.*}}include{{.*}}gpu-none-llvm" diff --git a/clang/test/Driver/gpu-libc.c b/clang/test/Driver/gpu-libc.c new file mode 100644 index 0000000..88f346f3 --- /dev/null +++ b/clang/test/Driver/gpu-libc.c @@ -0,0 +1,32 @@ +// RUN: %clang -### --target=amdgcn-amd-amdhsa -mcpu=gfx90a --sysroot=%S/Inputs/basic_gpu_tree \ +// RUN: -ccc-install-dir %S/Inputs/basic_gpu_tree/bin -nogpulib %s 2>&1 | FileCheck %s --check-prefix=CHECK-HEADERS-AMDGPU +// RUN: %clang -### --target=nvptx64-nvidia-cuda -march=sm_89 --sysroot=%S/Inputs/basic_gpu_tree \ +// RUN: -ccc-install-dir %S/Inputs/basic_gpu_tree/bin -nogpulib %s 2>&1 | FileCheck %s --check-prefix=CHECK-HEADERS-NVPTX +// CHECK-HEADERS-AMDGPU: "-cc1"{{.*}}"-isysroot"{{.*}}"-internal-isystem" "{{.*}}include{{.*}}amdgcn-amd-amdhsa" +// CHECK-HEADERS-NVPTX: "-cc1"{{.*}}"-isysroot"{{.*}}"-internal-isystem" "{{.*}}include{{.*}}nvptx64-nvidia-cuda" + +// RUN: %clang -### --target=amdgcn-amd-amdhsa -mcpu=gfx1030 -nogpulib \ +// RUN: -nogpuinc %s 2>&1 | FileCheck %s --check-prefix=CHECK-HEADERS-DISABLED +// RUN: %clang -### --target=amdgcn-amd-amdhsa -mcpu=gfx1030 -nogpulib \ +// RUN: -nostdinc %s 2>&1 | FileCheck %s --check-prefix=CHECK-HEADERS-DISABLED +// RUN: %clang -### --target=amdgcn-amd-amdhsa -mcpu=gfx1030 -nogpulib \ +// RUN: -nobuiltininc %s 2>&1 | FileCheck %s --check-prefix=CHECK-HEADERS-DISABLED +// CHECK-HEADERS-DISABLED-NOT: "-cc1"{{.*}}"-internal-isystem" "{{.*}}include{{.*}}gpu-none-llvm" + + +// RUN: %clang -### -fopenmp=libomp --target=x86_64-unknown-linux-gnu \ +// RUN: --offload-arch=gfx908 --rocm-path=%S/Inputs/rocm --sysroot=%S/Inputs/basic_gpu_tree \ +// RUN: -ccc-install-dir %S/Inputs/basic_gpu_tree/bin %s 2>&1 | FileCheck %s --check-prefix=OPENMP-AMDGPU +// OPENMP-AMDGPU: clang-linker-wrapper{{.*}}"--device-linker=amdgcn-amd-amdhsa=-lc" +// RUN: %clang -### -fopenmp=libomp --target=x86_64-unknown-linux-gnu -foffload-lto \ +// RUN: --offload-arch=sm_52 --cuda-path=%S/Inputs/CUDA_111/usr/local/cuda --sysroot=%S/Inputs/basic_gpu_tree \ +// RUN: -ccc-install-dir %S/Inputs/basic_gpu_tree/bin %s 2>&1 | FileCheck %s --check-prefix=OPENMP-NVPTX +// OPENMP-NVPTX: clang-linker-wrapper{{.*}}"--device-linker=nvptx64-nvidia-cuda=-lc" +// RUN: %clang -### --target=x86_64-unknown-linux-gnu --offload-arch=gfx908 \ +// RUN: --offload-new-driver --rocm-path=%S/Inputs/rocm --sysroot=%S/Inputs/basic_gpu_tree \ +// RUN: -ccc-install-dir %S/Inputs/basic_gpu_tree/bin -x hip %s 2>&1 | FileCheck %s --check-prefix=HIP +// HIP-NOT: "--device-linker=amdgcn-amd-amdhsa=-lc" +// RUN: %clang -### --target=x86_64-unknown-linux-gnu -fgpu-rdc --offload-arch=sm_52 \ +// RUN: --cuda-path=%S/Inputs/CUDA_111/usr/local/cuda --sysroot=%S/Inputs/basic_gpu_tree \ +// RUN: -ccc-install-dir %S/Inputs/basic_gpu_tree/bin -x cuda %s 2>&1 | FileCheck %s --check-prefix=CUDA +// CUDA-NOT: "--device-linker=nvptx64-nvidia-cuda=-lc" diff --git a/clang/test/Driver/x86-march.c b/clang/test/Driver/x86-march.c index 341f01c..15f8254 100644 --- a/clang/test/Driver/x86-march.c +++ b/clang/test/Driver/x86-march.c @@ -116,6 +116,14 @@ // RUN: | FileCheck %s -check-prefix=pantherlake // pantherlake: "-target-cpu" "pantherlake" // +// RUN: %clang --target=x86_64 -c -### %s -march=wildcatlake 2>&1 \ +// RUN: | FileCheck %s -check-prefix=wildcatlake +// wildcatlake: "-target-cpu" "wildcatlake" +// +// RUN: %clang --target=x86_64 -c -### %s -march=novalake 2>&1 \ +// RUN: | FileCheck %s -check-prefix=novalake +// novalake: "-target-cpu" "novalake" +// // RUN: %clang --target=x86_64 -c -### %s -march=clearwaterforest 2>&1 \ // RUN: | FileCheck %s -check-prefix=clearwaterforest // clearwaterforest: "-target-cpu" "clearwaterforest" diff --git a/clang/test/FixIt/fixit-constrained-structured-binding.cpp b/clang/test/FixIt/fixit-constrained-structured-binding.cpp index 3f21c1d..bfb004f 100644 --- a/clang/test/FixIt/fixit-constrained-structured-binding.cpp +++ b/clang/test/FixIt/fixit-constrained-structured-binding.cpp @@ -14,20 +14,20 @@ T get_T(); void use() { UnaryC auto [a, b] = get_S(); - // CHECK: error: decomposition declaration cannot be declared with constrained 'auto' + // CHECK: error: structured binding declaration cannot be declared with constrained 'auto' // CHECK: fix-it:{{.*}}:{16:3-16:10}:"" BinaryC<int> auto [c, d] = get_S(); - // CHECK: error: decomposition declaration cannot be declared with constrained 'auto' + // CHECK: error: structured binding declaration cannot be declared with constrained 'auto' // CHECK: fix-it:{{.*}}:{19:3-19:16}:"" } template<typename T> void TemplUse() { UnaryC auto [a, b] = get_T<T>(); - // CHECK: error: decomposition declaration cannot be declared with constrained 'auto' + // CHECK: error: structured binding declaration cannot be declared with constrained 'auto' // XCHECK: fix-it:{{.*}}:{26:3-26:10}:"" BinaryC<T> auto [c, d] = get_T<T>(); - // CHECK: error: decomposition declaration cannot be declared with constrained 'auto' + // CHECK: error: structured binding declaration cannot be declared with constrained 'auto' // XCHECK: fix-it:{{.*}}:{29:3-29:14}:"" } diff --git a/clang/test/Misc/pragma-attribute-supported-attributes-list.test b/clang/test/Misc/pragma-attribute-supported-attributes-list.test index 73d4cb1..ab4153a 100644 --- a/clang/test/Misc/pragma-attribute-supported-attributes-list.test +++ b/clang/test/Misc/pragma-attribute-supported-attributes-list.test @@ -35,6 +35,7 @@ // CHECK-NEXT: CFUnknownTransfer (SubjectMatchRule_function) // CHECK-NEXT: CPUDispatch (SubjectMatchRule_function) // CHECK-NEXT: CPUSpecific (SubjectMatchRule_function) +// CHECK-NEXT: CUDAClusterDims (SubjectMatchRule_objc_method, SubjectMatchRule_hasType_functionType) // CHECK-NEXT: CUDAConstant (SubjectMatchRule_variable) // CHECK-NEXT: CUDADevice (SubjectMatchRule_function, SubjectMatchRule_variable) // CHECK-NEXT: CUDADeviceBuiltinSurfaceType (SubjectMatchRule_record) @@ -43,6 +44,7 @@ // CHECK-NEXT: CUDAGridConstant (SubjectMatchRule_variable_is_parameter) // CHECK-NEXT: CUDAHost (SubjectMatchRule_function) // CHECK-NEXT: CUDALaunchBounds (SubjectMatchRule_objc_method, SubjectMatchRule_hasType_functionType) +// CHECK-NEXT: CUDANoCluster (SubjectMatchRule_objc_method, SubjectMatchRule_hasType_functionType) // CHECK-NEXT: CUDAShared (SubjectMatchRule_variable) // CHECK-NEXT: CXX11NoReturn (SubjectMatchRule_function) // CHECK-NEXT: CallableWhen (SubjectMatchRule_function_is_member) diff --git a/clang/test/Misc/target-invalid-cpu-note/x86.c b/clang/test/Misc/target-invalid-cpu-note/x86.c index f89cdc2..4a70e9b 100644 --- a/clang/test/Misc/target-invalid-cpu-note/x86.c +++ b/clang/test/Misc/target-invalid-cpu-note/x86.c @@ -63,6 +63,8 @@ // X86-SAME: {{^}}, lunarlake // X86-SAME: {{^}}, gracemont // X86-SAME: {{^}}, pantherlake +// X86-SAME: {{^}}, wildcatlake +// X86-SAME: {{^}}, novalake // X86-SAME: {{^}}, sierraforest // X86-SAME: {{^}}, grandridge // X86-SAME: {{^}}, graniterapids @@ -150,6 +152,8 @@ // X86_64-SAME: {{^}}, lunarlake // X86_64-SAME: {{^}}, gracemont // X86_64-SAME: {{^}}, pantherlake +// X86_64-SAME: {{^}}, wildcatlake +// X86_64-SAME: {{^}}, novalake // X86_64-SAME: {{^}}, sierraforest // X86_64-SAME: {{^}}, grandridge // X86_64-SAME: {{^}}, graniterapids @@ -246,6 +250,8 @@ // TUNE_X86-SAME: {{^}}, lunarlake // TUNE_X86-SAME: {{^}}, gracemont // TUNE_X86-SAME: {{^}}, pantherlake +// TUNE_X86-SAME: {{^}}, wildcatlake +// TUNE_X86-SAME: {{^}}, novalake // TUNE_X86-SAME: {{^}}, sierraforest // TUNE_X86-SAME: {{^}}, grandridge // TUNE_X86-SAME: {{^}}, graniterapids @@ -349,6 +355,8 @@ // TUNE_X86_64-SAME: {{^}}, lunarlake // TUNE_X86_64-SAME: {{^}}, gracemont // TUNE_X86_64-SAME: {{^}}, pantherlake +// TUNE_X86_64-SAME: {{^}}, wildcatlake +// TUNE_X86_64-SAME: {{^}}, novalake // TUNE_X86_64-SAME: {{^}}, sierraforest // TUNE_X86_64-SAME: {{^}}, grandridge // TUNE_X86_64-SAME: {{^}}, graniterapids diff --git a/clang/test/Modules/fmodules-validate-once-per-build-session.c b/clang/test/Modules/fmodules-validate-once-per-build-session.c index d9d79b0..2348ca1 100644 --- a/clang/test/Modules/fmodules-validate-once-per-build-session.c +++ b/clang/test/Modules/fmodules-validate-once-per-build-session.c @@ -1,119 +1,134 @@ -#include "foo.h" -#include "bar.h" - -// Clear the module cache. -// RUN: rm -rf %t -// RUN: mkdir -p %t/Inputs -// RUN: mkdir -p %t/modules-to-compare +// This tests the behavior of -fmodules-validate-once-per-build-session with +// different combinations of flags and states of the module cache. -// === -// Create a module. We will use -I or -isystem to determine whether to treat -// foo.h as a system header. -// RUN: echo 'void meow(void);' > %t/Inputs/foo.h -// RUN: echo 'void woof(void);' > %t/Inputs/bar.h -// RUN: echo 'module Foo { header "foo.h" }' > %t/Inputs/module.modulemap -// RUN: echo 'extern module Bar "bar.modulemap"' >> %t/Inputs/module.modulemap -// RUN: echo 'module Bar { header "bar.h" }' > %t/Inputs/bar.modulemap +// Note: The `sleep 1` commands sprinkled throughout this test make the strict +// comparisons of epoch mtimes work as expected. Some may be unnecessary, +// but make the intent clearer. -// === -// Compile the module. -// RUN: %clang_cc1 -cc1 -fmodules -fimplicit-module-maps -fdisable-module-hash -fmodules-cache-path=%t/modules-cache -fsyntax-only -isystem %t/Inputs -fmodules-validate-system-headers -fbuild-session-timestamp=1390000000 -fmodules-validate-once-per-build-session %s -// RUN: %clang_cc1 -cc1 -fmodules -fimplicit-module-maps -fdisable-module-hash -fmodules-cache-path=%t/modules-cache-user -fsyntax-only -I %t/Inputs -fmodules-validate-system-headers -fbuild-session-timestamp=1390000000 -fmodules-validate-once-per-build-session %s -// RUN: %clang_cc1 -cc1 -fmodules -fimplicit-module-maps -fdisable-module-hash -fmodules-cache-path=%t/modules-cache-user-no-force -fsyntax-only -I %t/Inputs -fno-modules-force-validate-user-headers -fmodules-validate-system-headers -fbuild-session-timestamp=1390000000 -fmodules-validate-once-per-build-session %s -// RUN: ls -R %t/modules-cache | grep Foo.pcm.timestamp -// RUN: ls -R %t/modules-cache | grep Bar.pcm.timestamp -// RUN: ls -R %t/modules-cache-user | grep Foo.pcm.timestamp -// RUN: ls -R %t/modules-cache-user | grep Bar.pcm.timestamp -// RUN: ls -R %t/modules-cache-user-no-force | grep Foo.pcm.timestamp -// RUN: ls -R %t/modules-cache-user-no-force | grep Bar.pcm.timestamp -// RUN: cp %t/modules-cache/Foo.pcm %t/modules-to-compare/Foo-before.pcm -// RUN: cp %t/modules-cache/Bar.pcm %t/modules-to-compare/Bar-before.pcm -// RUN: cp %t/modules-cache-user/Foo.pcm %t/modules-to-compare/Foo-before-user.pcm -// RUN: cp %t/modules-cache-user/Bar.pcm %t/modules-to-compare/Bar-before-user.pcm -// RUN: cp %t/modules-cache-user-no-force/Foo.pcm %t/modules-to-compare/Foo-before-user-no-force.pcm -// RUN: cp %t/modules-cache-user-no-force/Bar.pcm %t/modules-to-compare/Bar-before-user-no-force.pcm - -// === -// Use it, and make sure that we did not recompile it. -// RUN: %clang_cc1 -cc1 -fmodules -fimplicit-module-maps -fdisable-module-hash -fmodules-cache-path=%t/modules-cache -fsyntax-only -isystem %t/Inputs -fmodules-validate-system-headers -fbuild-session-timestamp=1390000000 -fmodules-validate-once-per-build-session %s -// RUN: %clang_cc1 -cc1 -fmodules -fimplicit-module-maps -fdisable-module-hash -fmodules-cache-path=%t/modules-cache-user -fsyntax-only -I %t/Inputs -fmodules-validate-system-headers -fbuild-session-timestamp=1390000000 -fmodules-validate-once-per-build-session %s -// RUN: %clang_cc1 -cc1 -fmodules -fimplicit-module-maps -fdisable-module-hash -fmodules-cache-path=%t/modules-cache-use-no-force -fsyntax-only -I %t/Inputs -fno-modules-force-validate-user-headers -fmodules-validate-system-headers -fbuild-session-timestamp=1390000000 -fmodules-validate-once-per-build-session %s -// RUN: ls -R %t/modules-cache | grep Foo.pcm.timestamp -// RUN: ls -R %t/modules-cache | grep Bar.pcm.timestamp -// RUN: ls -R %t/modules-cache-user | grep Foo.pcm.timestamp -// RUN: ls -R %t/modules-cache-user | grep Bar.pcm.timestamp -// RUN: ls -R %t/modules-cache-user-no-force | grep Foo.pcm.timestamp -// RUN: ls -R %t/modules-cache-user-no-force | grep Bar.pcm.timestamp -// RUN: cp %t/modules-cache/Foo.pcm %t/modules-to-compare/Foo-after.pcm -// RUN: cp %t/modules-cache/Bar.pcm %t/modules-to-compare/Bar-after.pcm -// RUN: cp %t/modules-cache-user/Foo.pcm %t/modules-to-compare/Foo-after-user.pcm -// RUN: cp %t/modules-cache-user/Bar.pcm %t/modules-to-compare/Bar-after-user.pcm -// RUN: cp %t/modules-cache-user-no-force/Foo.pcm %t/modules-to-compare/Foo-after-user-no-force.pcm -// RUN: cp %t/modules-cache-user-no-force/Bar.pcm %t/modules-to-compare/Bar-after-user-no-force.pcm +// RUN: rm -rf %t +// RUN: split-file %s %t +// RUN: echo "-fsyntax-only -fmodules -fmodules-cache-path=%/t/module-cache" > %t/ctx.rsp +// RUN: echo "-fbuild-session-file=%/t/module-cache/session.timestamp" >> %t/ctx.rsp +// RUN: echo "-fmodules-validate-once-per-build-session" >> %t/ctx.rsp +// RUN: echo "-Rmodule-build -Rmodule-validation" >> %t/ctx.rsp -// RUN: diff %t/modules-to-compare/Foo-before.pcm %t/modules-to-compare/Foo-after.pcm -// RUN: diff %t/modules-to-compare/Bar-before.pcm %t/modules-to-compare/Bar-after.pcm -// RUN: diff %t/modules-to-compare/Foo-before-user.pcm %t/modules-to-compare/Foo-after-user.pcm -// RUN: diff %t/modules-to-compare/Bar-before-user.pcm %t/modules-to-compare/Bar-after-user.pcm -// RUN: diff %t/modules-to-compare/Foo-before-user-no-force.pcm %t/modules-to-compare/Foo-after-user-no-force.pcm -// RUN: diff %t/modules-to-compare/Bar-before-user-no-force.pcm %t/modules-to-compare/Bar-after-user-no-force.pcm +//--- include/foo.h +//--- include/module.modulemap +module Foo { header "foo.h" } -// === -// Change the sources. +//--- clean.c +// Clean module cache. Modules will get compiled regardless of validation settings. +// RUN: mkdir %t/module-cache // RUN: sleep 1 -// RUN: echo 'void meow2(void);' > %t/Inputs/foo.h -// RUN: echo 'module Bar { header "bar.h" export * }' > %t/Inputs/bar.modulemap +// RUN: touch %t/module-cache/session.timestamp +// RUN: sleep 1 +// RUN: %clang @%t/ctx.rsp %t/clean.c -DCTX=1 \ +// RUN: -isystem %t/include -fmodules-validate-system-headers \ +// RUN: 2>&1 | FileCheck %t/clean.c +// RUN: %clang @%t/ctx.rsp %t/clean.c -DCTX=2 \ +// RUN: -I %t/include -fmodules-validate-system-headers \ +// RUN: 2>&1 | FileCheck %t/clean.c +// RUN: %clang @%t/ctx.rsp %t/clean.c -DCTX=3 \ +// RUN: -I %t/include -fmodules-validate-system-headers -Xclang -fno-modules-force-validate-user-headers \ +// RUN: 2>&1 | FileCheck %t/clean.c +#include "foo.h" +// CHECK: building module 'Foo' -// === -// Use the module, and make sure that we did not recompile it if foo.h or -// module.modulemap are system files or user files with force validation disabled, -// even though the sources changed. -// RUN: %clang_cc1 -cc1 -fmodules -fimplicit-module-maps -fdisable-module-hash -fmodules-cache-path=%t/modules-cache -fsyntax-only -isystem %t/Inputs -fmodules-validate-system-headers -fbuild-session-timestamp=1390000000 -fmodules-validate-once-per-build-session %s -// RUN: %clang_cc1 -cc1 -fmodules -fimplicit-module-maps -fdisable-module-hash -fmodules-cache-path=%t/modules-cache-user -fsyntax-only -I %t/Inputs -fmodules-validate-system-headers -fbuild-session-timestamp=1390000000 -fmodules-validate-once-per-build-session %s -// RUN: %clang_cc1 -cc1 -fmodules -fimplicit-module-maps -fdisable-module-hash -fmodules-cache-path=%t/modules-cache-user-no-force -fsyntax-only -I %t/Inputs -fno-modules-force-validate-user-headers -fmodules-validate-system-headers -fbuild-session-timestamp=1390000000 -fmodules-validate-once-per-build-session %s -// RUN: ls -R %t/modules-cache | grep Foo.pcm.timestamp -// RUN: ls -R %t/modules-cache | grep Bar.pcm.timestamp -// RUN: ls -R %t/modules-cache-user | grep Foo.pcm.timestamp -// RUN: ls -R %t/modules-cache-user | grep Bar.pcm.timestamp -// RUN: ls -R %t/modules-cache-user-no-force | grep Foo.pcm.timestamp -// RUN: ls -R %t/modules-cache-user-no-force | grep Bar.pcm.timestamp -// RUN: cp %t/modules-cache/Foo.pcm %t/modules-to-compare/Foo-after.pcm -// RUN: cp %t/modules-cache/Bar.pcm %t/modules-to-compare/Bar-after.pcm -// RUN: cp %t/modules-cache-user/Foo.pcm %t/modules-to-compare/Foo-after-user.pcm -// RUN: cp %t/modules-cache-user/Bar.pcm %t/modules-to-compare/Bar-after-user.pcm -// RUN: cp %t/modules-cache-user-no-force/Foo.pcm %t/modules-to-compare/Foo-after-user-no-force.pcm -// RUN: cp %t/modules-cache-user-no-force/Bar.pcm %t/modules-to-compare/Bar-after-user-no-force.pcm +//--- no-change-same-session.c +// Populated module cache in the same build session with unchanged inputs. +// Validation only happens when it's forced for user headers. No compiles. +// RUN: sleep 1 +// RUN: %clang @%t/ctx.rsp %t/no-change-same-session.c -DCTX=1 \ +// RUN: -isystem %t/include -fmodules-validate-system-headers \ +// RUN: 2>&1 | FileCheck %t/no-change-same-session.c --check-prefix=CHECK-NO-VALIDATION-OR-BUILD --allow-empty +// RUN: %clang @%t/ctx.rsp %t/no-change-same-session.c -DCTX=2 \ +// RUN: -I %t/include -fmodules-validate-system-headers \ +// RUN: 2>&1 | FileCheck %t/no-change-same-session.c --check-prefix=CHECK-VALIDATION-ONLY +// RUN: %clang @%t/ctx.rsp %t/no-change-same-session.c -DCTX=3 \ +// RUN: -I %t/include -fmodules-validate-system-headers -Xclang -fno-modules-force-validate-user-headers \ +// RUN: 2>&1 | FileCheck %t/no-change-same-session.c --check-prefix=CHECK-NO-VALIDATION-OR-BUILD --allow-empty +#include "foo.h" +// CHECK-NO-VALIDATION-OR-BUILD-NOT: validating {{[0-9]+}} input files in module 'Foo' +// CHECK-NO-VALIDATION-OR-BUILD-NOT: building module 'Foo' +// CHECK-VALIDATION-ONLY: validating {{[0-9]+}} input files in module 'Foo' +// CHECK-VALIDATION-ONLY-NOT: building module 'Foo' -// RUN: diff %t/modules-to-compare/Foo-before.pcm %t/modules-to-compare/Foo-after.pcm -// RUN: diff %t/modules-to-compare/Bar-before.pcm %t/modules-to-compare/Bar-after.pcm -// When foo.h is an user header, we will validate it by default. -// RUN: not diff %t/modules-to-compare/Foo-before-user.pcm %t/modules-to-compare/Foo-after-user.pcm -// RUN: not diff %t/modules-to-compare/Bar-before-user.pcm %t/modules-to-compare/Bar-after-user.pcm -// When foo.h is an user header, we will not validate it if force validation is turned off. -// RUN: diff %t/modules-to-compare/Foo-before-user-no-force.pcm %t/modules-to-compare/Foo-after-user-no-force.pcm -// RUN: diff %t/modules-to-compare/Bar-before-user-no-force.pcm %t/modules-to-compare/Bar-after-user-no-force.pcm +//--- change-same-session.c +// Populated module cache in the same build session with changed inputs. +// Validation only happens when it's forced for user headers and results in compilation. +// RUN: sleep 1 +// RUN: touch %t/include/foo.h +// RUN: sleep 1 +// RUN: %clang @%t/ctx.rsp %t/change-same-session.c -DCTX=1 \ +// RUN: -isystem %t/include -fmodules-validate-system-headers \ +// RUN: 2>&1 | FileCheck %t/change-same-session.c --check-prefix=CHECK-NO-VALIDATION-OR-BUILD --allow-empty +// RUN: %clang @%t/ctx.rsp %t/change-same-session.c -DCTX=2 \ +// RUN: -I %t/include -fmodules-validate-system-headers \ +// RUN: 2>&1 | FileCheck %t/change-same-session.c --check-prefix=CHECK-VALIDATION-AND-BUILD +// RUN: %clang @%t/ctx.rsp %t/change-same-session.c -DCTX=3 \ +// RUN: -I %t/include -fmodules-validate-system-headers -Xclang -fno-modules-force-validate-user-headers \ +// RUN: 2>&1 | FileCheck %t/change-same-session.c --check-prefix=CHECK-NO-VALIDATION-OR-BUILD --allow-empty +#include "foo.h" +// CHECK-NO-VALIDATION-OR-BUILD-NOT: validating {{[0-9]+}} input files in module 'Foo' +// CHECK-NO-VALIDATION-OR-BUILD-NOT: building module 'Foo' +// CHECK-VALIDATION-AND-BUILD: validating {{[0-9]+}} input files in module 'Foo' +// CHECK-VALIDATION-AND-BUILD: building module 'Foo' -// === -// Recompile the module if the today's date is before 01 January 2100. -// RUN: %clang_cc1 -cc1 -fmodules -fimplicit-module-maps -fdisable-module-hash -fmodules-cache-path=%t/modules-cache -fsyntax-only -isystem %t/Inputs -fmodules-validate-system-headers -fbuild-session-timestamp=4102441200 -fmodules-validate-once-per-build-session %s -// RUN: %clang_cc1 -cc1 -fmodules -fimplicit-module-maps -fdisable-module-hash -fmodules-cache-path=%t/modules-cache-user -fsyntax-only -I %t/Inputs -fmodules-validate-system-headers -fbuild-session-timestamp=4102441200 -fmodules-validate-once-per-build-session %s -// RUN: %clang_cc1 -cc1 -fmodules -fimplicit-module-maps -fdisable-module-hash -fmodules-cache-path=%t/modules-cache-user-no-force -fsyntax-only -I %t/Inputs -fno-modules-force-validate-user-headers -fmodules-validate-system-headers -fbuild-session-timestamp=4102441200 -fmodules-validate-once-per-build-session %s -// RUN: ls -R %t/modules-cache | grep Foo.pcm.timestamp -// RUN: ls -R %t/modules-cache | grep Bar.pcm.timestamp -// RUN: ls -R %t/modules-cache-user | grep Foo.pcm.timestamp -// RUN: ls -R %t/modules-cache-user | grep Bar.pcm.timestamp -// RUN: ls -R %t/modules-cache-user-no-force | grep Foo.pcm.timestamp -// RUN: ls -R %t/modules-cache-user-no-force | grep Bar.pcm.timestamp -// RUN: cp %t/modules-cache/Foo.pcm %t/modules-to-compare/Foo-after.pcm -// RUN: cp %t/modules-cache/Bar.pcm %t/modules-to-compare/Bar-after.pcm -// RUN: cp %t/modules-cache-user/Foo.pcm %t/modules-to-compare/Foo-after-user.pcm -// RUN: cp %t/modules-cache-user/Bar.pcm %t/modules-to-compare/Bar-after-user.pcm -// RUN: cp %t/modules-cache-user-no-force/Foo.pcm %t/modules-to-compare/Foo-after-user-no-force.pcm -// RUN: cp %t/modules-cache-user-no-force/Bar.pcm %t/modules-to-compare/Bar-after-user-no-force.pcm +//--- change-new-session.c +// Populated module cache in a new build session with changed inputs. +// All configurations validate and recompile. +// RUN: sleep 1 +// RUN: touch %t/include/foo.h +// RUN: sleep 1 +// RUN: touch %t/module-cache/session.timestamp +// RUN: sleep 1 +// RUN: %clang @%t/ctx.rsp %t/change-new-session.c -DCTX=1 \ +// RUN: -isystem %t/include -fmodules-validate-system-headers \ +// RUN: 2>&1 | FileCheck %t/change-new-session.c --check-prefixes=CHECK,CHECK-VALIDATE-ONCE +// NOTE: Forced user headers validation causes redundant validation of the just-built module. +// RUN: %clang @%t/ctx.rsp %t/change-new-session.c -DCTX=2 \ +// RUN: -I %t/include -fmodules-validate-system-headers \ +// RUN: 2>&1 | FileCheck %t/change-new-session.c --check-prefixes=CHECK,CHECK-FORCE-VALIDATE-TWICE +// RUN: %clang @%t/ctx.rsp %t/change-new-session.c -DCTX=3 \ +// RUN: -I %t/include -fmodules-validate-system-headers -Xclang -fno-modules-force-validate-user-headers \ +// RUN: 2>&1 | FileCheck %t/change-new-session.c --check-prefixes=CHECK,CHECK-VALIDATE-ONCE +#include "foo.h" +// CHECK: validating {{[0-9]+}} input files in module 'Foo' +// CHECK: building module 'Foo' +// CHECK-VALIDATE-ONCE-NOT: validating {{[0-9]+}} input files in module 'Foo' +// CHECK-FORCE-VALIDATE-TWICE: validating {{[0-9]+}} input files in module 'Foo' -// RUN: not diff %t/modules-to-compare/Foo-before.pcm %t/modules-to-compare/Foo-after.pcm -// RUN: not diff %t/modules-to-compare/Bar-before.pcm %t/modules-to-compare/Bar-after.pcm -// RUN: not diff %t/modules-to-compare/Foo-before-user.pcm %t/modules-to-compare/Foo-after-user.pcm -// RUN: not diff %t/modules-to-compare/Bar-before-user.pcm %t/modules-to-compare/Bar-after-user.pcm -// RUN: not diff %t/modules-to-compare/Foo-before-user-no-force.pcm %t/modules-to-compare/Foo-after-user-no-force.pcm -// RUN: not diff %t/modules-to-compare/Bar-before-user-no-force.pcm %t/modules-to-compare/Bar-after-user-no-force.pcm +//--- no-change-new-session-twice.c +// Populated module cache in a new build session with unchanged inputs. +// At first, all configurations validate but don't recompile. +// RUN: sleep 1 +// RUN: touch %t/module-cache/session.timestamp +// RUN: sleep 1 +// RUN: %clang @%t/ctx.rsp %t/no-change-new-session-twice.c -DCTX=1 \ +// RUN: -isystem %t/include -fmodules-validate-system-headers \ +// RUN: 2>&1 | FileCheck %t/no-change-new-session-twice.c --check-prefix=CHECK-ONCE +// RUN: %clang @%t/ctx.rsp %t/no-change-new-session-twice.c -DCTX=2 \ +// RUN: -I %t/include -fmodules-validate-system-headers \ +// RUN: 2>&1 | FileCheck %t/no-change-new-session-twice.c --check-prefix=CHECK-ONCE +// RUN: %clang @%t/ctx.rsp %t/no-change-new-session-twice.c -DCTX=3 \ +// RUN: -I %t/include -fmodules-validate-system-headers -Xclang -fno-modules-force-validate-user-headers \ +// RUN: 2>&1 | FileCheck %t/no-change-new-session-twice.c --check-prefix=CHECK-ONCE +// +// Then, only the forced user header validation performs redundant validation (but no compilation). +// All other configurations do not validate and do not compile. +// RUN: sleep 1 +// RUN: %clang @%t/ctx.rsp %t/no-change-new-session-twice.c -DCTX=1 \ +// RUN: -isystem %t/include -fmodules-validate-system-headers \ +// RUN: 2>&1 | FileCheck %t/no-change-new-session-twice.c --check-prefix=CHECK-NOT-TWICE --allow-empty +// NOTE: Forced user headers validation causes redundant validation of the just-validated module. +// RUN: %clang @%t/ctx.rsp %t/no-change-new-session-twice.c -DCTX=2 \ +// RUN: -I %t/include -fmodules-validate-system-headers \ +// RUN: 2>&1 | FileCheck %t/no-change-new-session-twice.c --check-prefix=CHECK-ONCE +// RUN: %clang @%t/ctx.rsp %t/no-change-new-session-twice.c -DCTX=3 \ +// RUN: -I %t/include -fmodules-validate-system-headers -Xclang -fno-modules-force-validate-user-headers \ +// RUN: 2>&1 | FileCheck %t/no-change-new-session-twice.c --check-prefix=CHECK-NOT-TWICE --allow-empty +#include "foo.h" +// CHECK-ONCE: validating {{[0-9]+}} input files in module 'Foo' +// CHECK-ONCE-NOT: building module 'Foo' +// CHECK-NOT-TWICE-NOT: validating {{[0-9]+}} input files in module 'Foo' +// CHECK-NOT-TWICE-NOT: building module 'Foo' diff --git a/clang/test/OpenMP/nowait_ast_print.cpp b/clang/test/OpenMP/nowait_ast_print.cpp new file mode 100644 index 0000000..df0a77c --- /dev/null +++ b/clang/test/OpenMP/nowait_ast_print.cpp @@ -0,0 +1,55 @@ +// Check no warnings/errors +// RUN: %clang_cc1 -triple x86_64-pc-linux-gnu -fopenmp -fopenmp-version=60 -fsyntax-only -verify %s +// expected-no-diagnostics + +// Check AST and unparsing +// RUN: %clang_cc1 -triple x86_64-pc-linux-gnu -fopenmp -fopenmp-version=60 -ast-dump %s | FileCheck %s --check-prefix=DUMP +// RUN: %clang_cc1 -triple x86_64-pc-linux-gnu -fopenmp -fopenmp-version=60 -ast-print %s | FileCheck %s --check-prefix=PRINT + +// Check same results after serialization round-trip +// RUN: %clang_cc1 -triple x86_64-pc-linux-gnu -fopenmp -fopenmp-version=60 -emit-pch -o %t %s +// RUN: %clang_cc1 -triple x86_64-pc-linux-gnu -fopenmp -fopenmp-version=60 -include-pch %t -ast-dump-all %s | FileCheck %s --check-prefix=DUMP +// RUN: %clang_cc1 -triple x86_64-pc-linux-gnu -fopenmp -fopenmp-version=60 -include-pch %t -ast-print %s | FileCheck %s --check-prefix=PRINT + +#ifndef HEADER +#define HEADER + +void nowait() { + int A=1; + + // DUMP: OMPTargetDirective + // DUMP-NEXT: OMPNowaitClause + // PRINT: #pragma omp target nowait + #pragma omp target nowait + { + } + + // DUMP: OMPTargetDirective + // DUMP-NEXT: OMPNowaitClause + // DUMP-NEXT: XXBoolLiteralExpr {{.*}} 'bool' false + // PRINT: #pragma omp target nowait(false) + #pragma omp target nowait(false) + { + } + + // DUMP: OMPTargetDirective + // DUMP-NEXT: OMPNowaitClause + // DUMP-NEXT: XXBoolLiteralExpr {{.*}} 'bool' true + // PRINT: #pragma omp target nowait(true) + #pragma omp target nowait(true) + { + } + + // DUMP: OMPTargetDirective + // DUMP-NEXT: OMPNowaitClause + // DUMP-NEXT: BinaryOperator {{.*}} 'bool' '>' + // DUMP-NEXT: ImplicitCastExpr {{.*}} 'int' <LValueToRValue> + // DUMP-NEXT: DeclRefExpr {{.*}} 'int' lvalue Var {{.*}} 'A' 'int' + // DUMP-NEXT: IntegerLiteral {{.*}} 'int' 5 + // PRINT: #pragma omp target nowait(A > 5) + #pragma omp target nowait(A>5) + { + } + +} +#endif diff --git a/clang/test/OpenMP/target_enter_data_nowait_messages.cpp b/clang/test/OpenMP/target_enter_data_nowait_messages.cpp index ba5eaf1..8c7d236 100644 --- a/clang/test/OpenMP/target_enter_data_nowait_messages.cpp +++ b/clang/test/OpenMP/target_enter_data_nowait_messages.cpp @@ -1,6 +1,12 @@ -// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -verify -fopenmp -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=45 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized -// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -verify -fopenmp-simd -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=45 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized int main(int argc, char **argv) { int i; @@ -13,19 +19,27 @@ int main(int argc, char **argv) { {} #pragma omp target enter nowait data map(to: i) // expected-error {{expected an OpenMP directive}} {} - #pragma omp target enter data nowait() map(to: i) // expected-warning {{extra tokens at the end of '#pragma omp target enter data' are ignored}} expected-error {{expected at least one 'map' clause for '#pragma omp target enter data'}} + #pragma omp target enter data map(to: i) nowait() // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target enter data' are ignored}} omp-60-and-later-error {{expected expression}} {} - #pragma omp target enter data map(to: i) nowait( // expected-warning {{extra tokens at the end of '#pragma omp target enter data' are ignored}} + #pragma omp target enter data map(to: i) nowait( // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target enter data' are ignored}} omp-60-and-later-error {{expected expression}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} {} - #pragma omp target enter data map(to: i) nowait (argc)) // expected-warning {{extra tokens at the end of '#pragma omp target enter data' are ignored}} + #pragma omp target enter data map(to: i) nowait (argc)) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target enter data' are ignored}} omp-60-and-later-warning {{extra tokens at the end of '#pragma omp target_enter_data' are ignored}} {} #pragma omp target enter data map(to: i) nowait device (-10u) {} - #pragma omp target enter data map(to: i) nowait (3.14) device (-10u) // expected-warning {{extra tokens at the end of '#pragma omp target enter data' are ignored}} + #pragma omp target enter data map(to: i) nowait (3.14) device (-10u) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target enter data' are ignored}} {} - #pragma omp target enter data map(to: i) nowait nowait // expected-error {{directive '#pragma omp target enter data' cannot contain more than one 'nowait' clause}} + #pragma omp target enter data map(to: i) nowait (argc>> i) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target enter data' are ignored}} {} - #pragma omp target enter data nowait map(to: i) nowait // expected-error {{directive '#pragma omp target enter data' cannot contain more than one 'nowait' clause}} + #pragma omp target enter data map(to: i) nowait (argv[1] = 2) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target enter data' are ignored}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} + {} + #pragma omp target enter data map(to: i) nowait (argc > 0 ? argv[1] : argv[2]) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target enter data' are ignored}} + {} + #pragma omp target enter data map(to: i) nowait (S1) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target enter data' are ignored}} omp-60-and-later-error {{use of undeclared identifier 'S1'}} + {} + #pragma omp target enter data map(to: i) nowait nowait // omp-52-and-earlier-error {{directive '#pragma omp target enter data' cannot contain more than one 'nowait' clause}} omp-60-and-later-error {{directive '#pragma omp target_enter_data' cannot contain more than one 'nowait' clause}} + {} + #pragma omp target enter data nowait map(to: i) nowait // omp-52-and-earlier-error {{directive '#pragma omp target enter data' cannot contain more than one 'nowait' clause}} omp-60-and-later-error {{directive '#pragma omp target_enter_data' cannot contain more than one 'nowait' clause}} {} return 0; } diff --git a/clang/test/OpenMP/target_exit_data_nowait_messages.cpp b/clang/test/OpenMP/target_exit_data_nowait_messages.cpp index 307e2c3..a8e7925 100644 --- a/clang/test/OpenMP/target_exit_data_nowait_messages.cpp +++ b/clang/test/OpenMP/target_exit_data_nowait_messages.cpp @@ -1,6 +1,10 @@ -// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -verify -fopenmp -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized -// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -verify -fopenmp-simd -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized int main(int argc, char **argv) { int i; @@ -8,12 +12,16 @@ int main(int argc, char **argv) { #pragma omp nowait target exit data map(from: i) // expected-error {{expected an OpenMP directive}} #pragma omp target nowait exit data map(from: i) // expected-warning {{extra tokens at the end of '#pragma omp target' are ignored}} #pragma omp target exit nowait data map(from: i) // expected-error {{expected an OpenMP directive}} - #pragma omp target exit data nowait() map(from: i) // expected-warning {{extra tokens at the end of '#pragma omp target exit data' are ignored}} expected-error {{expected at least one 'map' clause for '#pragma omp target exit data'}} - #pragma omp target exit data map(from: i) nowait( // expected-warning {{extra tokens at the end of '#pragma omp target exit data' are ignored}} - #pragma omp target exit data map(from: i) nowait (argc)) // expected-warning {{extra tokens at the end of '#pragma omp target exit data' are ignored}} + #pragma omp target exit data map(from: i) nowait() // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target exit data' are ignored}} omp-60-and-later-error {{expected expression}} + #pragma omp target exit data map(from: i) nowait( // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target exit data' are ignored}} omp-60-and-later-error {{expected expression}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} + #pragma omp target exit data map(from: i) nowait (argc)) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target exit data' are ignored}} omp-60-and-later-warning {{extra tokens at the end of '#pragma omp target_exit_data' are ignored}} #pragma omp target exit data map(from: i) nowait device (-10u) - #pragma omp target exit data map(from: i) nowait (3.14) device (-10u) // expected-warning {{extra tokens at the end of '#pragma omp target exit data' are ignored}} - #pragma omp target exit data map(from: i) nowait nowait // expected-error {{directive '#pragma omp target exit data' cannot contain more than one 'nowait' clause}} - #pragma omp target exit data nowait map(from: i) nowait // expected-error {{directive '#pragma omp target exit data' cannot contain more than one 'nowait' clause}} + #pragma omp target exit data map(from: i) nowait (3.14) device (-10u) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target exit data' are ignored}} + #pragma omp target exit data map(from: i) nowait (argc>> i) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target exit data' are ignored}} + #pragma omp target exit data map(from: i) nowait (argv[1] = 2) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target exit data' are ignored}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} + #pragma omp target exit data map(from: i) nowait (argc > 0 ? argv[1] : argv[2]) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target exit data' are ignored}} + #pragma omp target exit data map(from: i) nowait (S1) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target exit data' are ignored}} omp-60-and-later-error {{use of undeclared identifier 'S1'}} + #pragma omp target exit data map(from: i) nowait nowait // omp-52-and-earlier-error {{directive '#pragma omp target exit data' cannot contain more than one 'nowait' clause}} omp-60-and-later-error {{directive '#pragma omp target_exit_data' cannot contain more than one 'nowait' clause}} + #pragma omp target exit data nowait map(from: i) nowait // omp-52-and-earlier-error {{directive '#pragma omp target exit data' cannot contain more than one 'nowait' clause}} omp-60-and-later-error {{directive '#pragma omp target_exit_data' cannot contain more than one 'nowait' clause}} return 0; } diff --git a/clang/test/OpenMP/target_indirect_codegen.cpp b/clang/test/OpenMP/target_indirect_codegen.cpp index 20a36c2..ec249dd 100644 --- a/clang/test/OpenMP/target_indirect_codegen.cpp +++ b/clang/test/OpenMP/target_indirect_codegen.cpp @@ -4,6 +4,12 @@ // RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=51 -x c++ -triple amdgcn-amd-amdhsa %s -fopenmp-is-target-device -fvisibility=protected -fopenmp-host-ir-file-path %t-host.bc -emit-pch -o %t // RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=51 -x c++ -triple amdgcn-amd-amdhsa -emit-llvm %s -fopenmp-is-target-device -fvisibility=protected -fopenmp-host-ir-file-path %t-host.bc -include-pch %t -o - | FileCheck %s --check-prefix=DEVICE +// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=51 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=spirv64-intel -emit-llvm %s -o - | FileCheck %s --check-prefix=HOST +// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=51 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=spirv64-intel -emit-llvm-bc %s -o %t-spirv-host.bc +// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=51 -x c++ -triple spirv64-intel -emit-llvm %s -fopenmp-is-target-device -fvisibility=protected -fopenmp-host-ir-file-path %t-spirv-host.bc -o - | FileCheck %s --check-prefix=DEVICE +// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=51 -x c++ -triple spirv64-intel %s -fopenmp-is-target-device -fvisibility=protected -fopenmp-host-ir-file-path %t-spirv-host.bc -emit-pch -o %t +// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=51 -x c++ -triple spirv64-intel -emit-llvm %s -fopenmp-is-target-device -fvisibility=protected -fopenmp-host-ir-file-path %t-spirv-host.bc -include-pch %t -o - | FileCheck %s --check-prefix=DEVICE + // expected-no-diagnostics #ifndef HEADER #define HEADER @@ -19,10 +25,10 @@ // HOST: @[[BAR_ENTRY_NAME:.+]] = internal unnamed_addr constant [{{[0-9]+}} x i8] c"[[BAR_NAME:__omp_offloading_[0-9a-z]+_[0-9a-z]+_bar_l[0-9]+]]\00" // HOST: @.offloading.entry.[[BAR_NAME]] = weak constant %struct.__tgt_offload_entry { i64 0, i16 1, i16 1, i32 8, ptr @_ZL3barv, ptr @[[BAR_ENTRY_NAME]], i64 8, i64 0, ptr null } //. -// DEVICE: @[[FOO_NAME:__omp_offloading_[0-9a-z]+_[0-9a-z]+_foo_l[0-9]+]] = protected addrspace(1) constant ptr @_Z3foov -// DEVICE: @[[BAZ_NAME:__omp_offloading_[0-9a-z]+_[0-9a-z]+_baz_l[0-9]+]] = protected addrspace(1) constant ptr @_Z3bazv +// DEVICE: @[[FOO_NAME:__omp_offloading_[0-9a-z]+_[0-9a-z]+_foo_l[0-9]+]] = protected addrspace(1) constant {{ptr|ptr addrspace\(9\)}} @_Z3foov +// DEVICE: @[[BAZ_NAME:__omp_offloading_[0-9a-z]+_[0-9a-z]+_baz_l[0-9]+]] = protected addrspace(1) constant {{ptr|ptr addrspace\(9\)}} @_Z3bazv // DEVICE: @var = protected addrspace(1) global i8 0, align 1 -// DEVICE: @[[BAR_NAME:__omp_offloading_[0-9a-z]+_[0-9a-z]+_bar_l[0-9]+]] = protected addrspace(1) constant ptr @_ZL3barv +// DEVICE: @[[BAR_NAME:__omp_offloading_[0-9a-z]+_[0-9a-z]+_bar_l[0-9]+]] = protected addrspace(1) constant {{ptr|ptr addrspace\(9\)}} @_ZL3barv //. void foo() { } #pragma omp declare target to(foo) indirect diff --git a/clang/test/OpenMP/target_nowait_messages.cpp b/clang/test/OpenMP/target_nowait_messages.cpp index 6b8a1f4..d3690f3 100644 --- a/clang/test/OpenMP/target_nowait_messages.cpp +++ b/clang/test/OpenMP/target_nowait_messages.cpp @@ -1,18 +1,33 @@ -// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -verify -fopenmp -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=45 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized -// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -verify -fopenmp-simd -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=45 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized void foo() { } int main(int argc, char **argv) { - #pragma omp target nowait( // expected-warning {{extra tokens at the end of '#pragma omp target' are ignored}} + int i; + #pragma omp target nowait( // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target' are ignored}} omp-60-and-later-error {{expected expression}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} foo(); #pragma omp target nowait (argc)) // expected-warning {{extra tokens at the end of '#pragma omp target' are ignored}} foo(); #pragma omp target nowait device (-10u) foo(); - #pragma omp target nowait (3.14) device (-10u) // expected-warning {{extra tokens at the end of '#pragma omp target' are ignored}} + #pragma omp target nowait (3.14) device (-10u) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target' are ignored}} + foo(); + #pragma omp target nowait (argc>> i) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target' are ignored}} + foo(); + #pragma omp target nowait (argv[1] = 2) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target' are ignored}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} + foo(); + #pragma omp target nowait (argc > 0 ? argv[1] : argv[2]) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target' are ignored}} + foo(); + #pragma omp target nowait (S1) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target' are ignored}} omp-60-and-later-error {{use of undeclared identifier 'S1'}} foo(); return 0; diff --git a/clang/test/OpenMP/target_parallel_for_nowait_messages.cpp b/clang/test/OpenMP/target_parallel_for_nowait_messages.cpp index 2f88c65..6a0cdd7 100644 --- a/clang/test/OpenMP/target_parallel_for_nowait_messages.cpp +++ b/clang/test/OpenMP/target_parallel_for_nowait_messages.cpp @@ -1,19 +1,31 @@ -// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -verify -fopenmp -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized -// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -verify -fopenmp-simd -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized void foo() { } int main(int argc, char **argv) { - int i; - #pragma omp target parallel for nowait( // expected-warning {{extra tokens at the end of '#pragma omp target parallel for' are ignored}} + int i, z; + #pragma omp target parallel for nowait( // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target parallel for' are ignored}} omp-60-and-later-error {{expected expression}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} for (i = 0; i < argc; ++i) foo(); #pragma omp target parallel for nowait (argc)) // expected-warning {{extra tokens at the end of '#pragma omp target parallel for' are ignored}} for (i = 0; i < argc; ++i) foo(); #pragma omp target parallel for nowait device (-10u) for (i = 0; i < argc; ++i) foo(); - #pragma omp target parallel for nowait (3.14) device (-10u) // expected-warning {{extra tokens at the end of '#pragma omp target parallel for' are ignored}} + #pragma omp target parallel for nowait (3.14) device (-10u) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target parallel for' are ignored}} + for (i = 0; i < argc; ++i) foo(); + #pragma omp target parallel for nowait (argc>> z) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target parallel for' are ignored}} + for (i = 0; i < argc; ++i) foo(); + #pragma omp target parallel for nowait (argv[1] = 2) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target parallel for' are ignored}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} + for (i = 0; i < argc; ++i) foo(); + #pragma omp target parallel for nowait (argc > 0 ? argv[1] : argv[2]) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target parallel for' are ignored}} + for (i = 0; i < argc; ++i) foo(); + #pragma omp target parallel for nowait (S1) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target parallel for' are ignored}} omp-60-and-later-error {{use of undeclared identifier 'S1'}} for (i = 0; i < argc; ++i) foo(); return 0; diff --git a/clang/test/OpenMP/target_parallel_for_simd_nowait_messages.cpp b/clang/test/OpenMP/target_parallel_for_simd_nowait_messages.cpp index 4220bfe..4df04a0 100644 --- a/clang/test/OpenMP/target_parallel_for_simd_nowait_messages.cpp +++ b/clang/test/OpenMP/target_parallel_for_simd_nowait_messages.cpp @@ -1,19 +1,31 @@ -// RUN: %clang_cc1 -verify -fopenmp %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized -// RUN: %clang_cc1 -verify -fopenmp-simd %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized void foo() { } int main(int argc, char **argv) { - int i; - #pragma omp target parallel for simd nowait( // expected-warning {{extra tokens at the end of '#pragma omp target parallel for simd' are ignored}} + int i, z; + #pragma omp target parallel for simd nowait( // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target parallel for simd' are ignored}} omp-60-and-later-error {{expected expression}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} for (i = 0; i < argc; ++i) foo(); #pragma omp target parallel for simd nowait (argc)) // expected-warning {{extra tokens at the end of '#pragma omp target parallel for simd' are ignored}} for (i = 0; i < argc; ++i) foo(); #pragma omp target parallel for simd nowait device (-10u) for (i = 0; i < argc; ++i) foo(); - #pragma omp target parallel for simd nowait (3.14) device (-10u) // expected-warning {{extra tokens at the end of '#pragma omp target parallel for simd' are ignored}} + #pragma omp target parallel for simd nowait (3.14) device (-10u) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target parallel for simd' are ignored}} + for (i = 0; i < argc; ++i) foo(); + #pragma omp target parallel for simd nowait (argc>> z) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target parallel for simd' are ignored}} + for (i = 0; i < argc; ++i) foo(); + #pragma omp target parallel for simd nowait (argv[1] = 2) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target parallel for simd' are ignored}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} + for (i = 0; i < argc; ++i) foo(); + #pragma omp target parallel for simd nowait (argc > 0 ? argv[1] : argv[2]) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target parallel for simd' are ignored}} + for (i = 0; i < argc; ++i) foo(); + #pragma omp target parallel for simd nowait (S1) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target parallel for simd' are ignored}} omp-60-and-later-error {{use of undeclared identifier 'S1'}} for (i = 0; i < argc; ++i) foo(); return 0; diff --git a/clang/test/OpenMP/target_parallel_nowait_messages.cpp b/clang/test/OpenMP/target_parallel_nowait_messages.cpp index 3e285fc..f8f43cc 100644 --- a/clang/test/OpenMP/target_parallel_nowait_messages.cpp +++ b/clang/test/OpenMP/target_parallel_nowait_messages.cpp @@ -1,18 +1,31 @@ -// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -verify -fopenmp -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized -// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -verify -fopenmp-simd -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized void foo() { } int main(int argc, char **argv) { - #pragma omp target parallel nowait( // expected-warning {{extra tokens at the end of '#pragma omp target parallel' are ignored}} - foo(); + int z; + #pragma omp target parallel nowait( // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target parallel' are ignored}} omp-60-and-later-error {{expected expression}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} + foo(); #pragma omp target parallel nowait (argc)) // expected-warning {{extra tokens at the end of '#pragma omp target parallel' are ignored}} foo(); #pragma omp target parallel nowait device (-10u) foo(); - #pragma omp target parallel nowait (3.14) device (-10u) // expected-warning {{extra tokens at the end of '#pragma omp target parallel' are ignored}} + #pragma omp target parallel nowait (3.14) device (-10u) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target parallel' are ignored}} + foo(); + #pragma omp target parallel nowait (argc>> z) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target parallel' are ignored}} + foo(); + #pragma omp target parallel nowait (argv[1] = 2) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target parallel' are ignored}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} + foo(); + #pragma omp target parallel nowait (argc > 0 ? argv[1] : argv[2]) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target parallel' are ignored}} + foo(); + #pragma omp target parallel nowait (S1) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target parallel' are ignored}} omp-60-and-later-error {{use of undeclared identifier 'S1'}} foo(); return 0; diff --git a/clang/test/OpenMP/amdgcn_parallel_num_threads_strict_messages.cpp b/clang/test/OpenMP/target_parallel_num_threads_strict_messages.cpp index 513754b..8ceff02 100644 --- a/clang/test/OpenMP/amdgcn_parallel_num_threads_strict_messages.cpp +++ b/clang/test/OpenMP/target_parallel_num_threads_strict_messages.cpp @@ -5,6 +5,13 @@ // RUN: %clang_cc1 -DF3 -verify -fopenmp -fopenmp-version=60 -triple x86_64-unknown-unknown -fopenmp-targets=amdgcn-amd-amdhsa -emit-llvm-bc %s -o %t-ppc-host-ppc.bc // RUN: %clang_cc1 -DF3 -DTARGET -verify -fopenmp -fopenmp-version=60 -triple amdgcn-amd-amdhsa -fopenmp-targets=amdgcn-amd-amdhsa -emit-llvm %s -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-ppc-host-ppc.bc -o /dev/null +// RUN: %clang_cc1 -DF1 -verify -fopenmp -fopenmp-version=60 -triple x86_64-unknown-unknown -fopenmp-targets=spirv64-intel -emit-llvm-bc %s -o %t-spirv-ppc-host-ppc.bc +// RUN: %clang_cc1 -DF1 -DTARGET -verify -fopenmp -fopenmp-version=60 -triple spirv64-intel -fopenmp-targets=spirv64-intel -emit-llvm %s -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-spirv-ppc-host-ppc.bc -o /dev/null +// RUN: %clang_cc1 -DF2 -verify -fopenmp -fopenmp-version=60 -triple x86_64-unknown-unknown -fopenmp-targets=spirv64-intel -emit-llvm-bc %s -o %t-spirv-ppc-host-ppc.bc +// RUN: %clang_cc1 -DF2 -DTARGET -verify -fopenmp -fopenmp-version=60 -triple spirv64-intel -fopenmp-targets=spirv64-intel -emit-llvm %s -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-spirv-ppc-host-ppc.bc -o /dev/null +// RUN: %clang_cc1 -DF3 -verify -fopenmp -fopenmp-version=60 -triple x86_64-unknown-unknown -fopenmp-targets=spirv64-intel -emit-llvm-bc %s -o %t-spirv-ppc-host-ppc.bc +// RUN: %clang_cc1 -DF3 -DTARGET -verify -fopenmp -fopenmp-version=60 -triple spirv64-intel -fopenmp-targets=spirv64-intel -emit-llvm %s -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-spirv-ppc-host-ppc.bc -o /dev/null + #ifndef TARGET // expected-no-diagnostics #endif diff --git a/clang/test/OpenMP/target_simd_nowait_messages.cpp b/clang/test/OpenMP/target_simd_nowait_messages.cpp index 1aee110..4920168 100644 --- a/clang/test/OpenMP/target_simd_nowait_messages.cpp +++ b/clang/test/OpenMP/target_simd_nowait_messages.cpp @@ -1,19 +1,33 @@ -// RUN: %clang_cc1 -verify -fopenmp %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=45 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized -// RUN: %clang_cc1 -verify -fopenmp-simd %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=45 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized void foo() { } int main(int argc, char **argv) { - int i; - #pragma omp target simd nowait( // expected-warning {{extra tokens at the end of '#pragma omp target simd' are ignored}} + int i, z; + #pragma omp target simd nowait( // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target simd' are ignored}} omp-60-and-later-error {{expected expression}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} for (i = 0; i < argc; ++i) foo(); #pragma omp target simd nowait (argc)) // expected-warning {{extra tokens at the end of '#pragma omp target simd' are ignored}} for (i = 0; i < argc; ++i) foo(); #pragma omp target simd nowait device (-10u) for (i = 0; i < argc; ++i) foo(); - #pragma omp target simd nowait (3.14) device (-10u) // expected-warning {{extra tokens at the end of '#pragma omp target simd' are ignored}} + #pragma omp target simd nowait (3.14) device (-10u) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target simd' are ignored}} + for (i = 0; i < argc; ++i) foo(); + #pragma omp target simd nowait (argc>> z) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target simd' are ignored}} + for (i = 0; i < argc; ++i) foo(); + #pragma omp target simd nowait (argv[1] = 2) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target simd' are ignored}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} + for (i = 0; i < argc; ++i) foo(); + #pragma omp target simd nowait (argc > 0 ? argv[1] : argv[2]) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target simd' are ignored}} + for (i = 0; i < argc; ++i) foo(); + #pragma omp target simd nowait (S1) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target simd' are ignored}} omp-60-and-later-error {{use of undeclared identifier 'S1'}} for (i = 0; i < argc; ++i) foo(); return 0; diff --git a/clang/test/OpenMP/target_teams_distribute_nowait_messages.cpp b/clang/test/OpenMP/target_teams_distribute_nowait_messages.cpp index 75bab80..6517631 100644 --- a/clang/test/OpenMP/target_teams_distribute_nowait_messages.cpp +++ b/clang/test/OpenMP/target_teams_distribute_nowait_messages.cpp @@ -1,19 +1,33 @@ -// RUN: %clang_cc1 -verify -fopenmp -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=45 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized -// RUN: %clang_cc1 -verify -fopenmp-simd -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=45 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized void foo() { } int main(int argc, char **argv) { - int i; -#pragma omp target teams distribute nowait( // expected-warning {{extra tokens at the end of '#pragma omp target teams distribute' are ignored}} + int i, z; +#pragma omp target teams distribute nowait( // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams distribute' are ignored}} omp-60-and-later-error {{expected expression}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} for (i = 0; i < argc; ++i) foo(); #pragma omp target teams distribute nowait (argc)) // expected-warning {{extra tokens at the end of '#pragma omp target teams distribute' are ignored}} for (i = 0; i < argc; ++i) foo(); #pragma omp target teams distribute nowait device (-10u) for (i = 0; i < argc; ++i) foo(); -#pragma omp target teams distribute nowait (3.14) device (-10u) // expected-warning {{extra tokens at the end of '#pragma omp target teams distribute' are ignored}} +#pragma omp target teams distribute nowait (3.14) device (-10u) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams distribute' are ignored}} + for (i = 0; i < argc; ++i) foo(); +#pragma omp target teams distribute nowait (argc>> z) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams distribute' are ignored}} + for (i = 0; i < argc; ++i) foo(); +#pragma omp target teams distribute nowait (argv[1] = 2) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams distribute' are ignored}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} + for (i = 0; i < argc; ++i) foo(); +#pragma omp target teams distribute nowait (argc > 0 ? argv[1] : argv[2]) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams distribute' are ignored}} + for (i = 0; i < argc; ++i) foo(); +#pragma omp target teams distribute nowait (S1) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams distribute' are ignored}} omp-60-and-later-error {{use of undeclared identifier 'S1'}} for (i = 0; i < argc; ++i) foo(); return 0; diff --git a/clang/test/OpenMP/target_teams_distribute_parallel_for_nowait_messages.cpp b/clang/test/OpenMP/target_teams_distribute_parallel_for_nowait_messages.cpp index f6b6061..0e72618 100644 --- a/clang/test/OpenMP/target_teams_distribute_parallel_for_nowait_messages.cpp +++ b/clang/test/OpenMP/target_teams_distribute_parallel_for_nowait_messages.cpp @@ -1,19 +1,33 @@ -// RUN: %clang_cc1 -fsyntax-only -verify -fopenmp %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=45 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized -// RUN: %clang_cc1 -fsyntax-only -verify -fopenmp-simd %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=45 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized void foo() { } int main(int argc, char **argv) { - int i; -#pragma omp target teams distribute parallel for nowait( // expected-warning {{extra tokens at the end of '#pragma omp target teams distribute parallel for' are ignored}} + int i, z; +#pragma omp target teams distribute parallel for nowait( // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams distribute parallel for' are ignored}} omp-60-and-later-error {{expected expression}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} for (i = 0; i < argc; ++i) foo(); #pragma omp target teams distribute parallel for nowait (argc)) // expected-warning {{extra tokens at the end of '#pragma omp target teams distribute parallel for' are ignored}} for (i = 0; i < argc; ++i) foo(); #pragma omp target teams distribute parallel for nowait device (-10u) for (i = 0; i < argc; ++i) foo(); -#pragma omp target teams distribute parallel for nowait (3.14) device (-10u) // expected-warning {{extra tokens at the end of '#pragma omp target teams distribute parallel for' are ignored}} +#pragma omp target teams distribute parallel for nowait (3.14) device (-10u) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams distribute parallel for' are ignored}} + for (i = 0; i < argc; ++i) foo(); +#pragma omp target teams distribute parallel for nowait (argc>> z) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams distribute parallel for' are ignored}} + for (i = 0; i < argc; ++i) foo(); +#pragma omp target teams distribute parallel for nowait (argv[1] = 2) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams distribute parallel for' are ignored}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} + for (i = 0; i < argc; ++i) foo(); +#pragma omp target teams distribute parallel for nowait (argc > 0 ? argv[1] : argv[2]) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams distribute parallel for' are ignored}} + for (i = 0; i < argc; ++i) foo(); +#pragma omp target teams distribute parallel for nowait (S1) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams distribute parallel for' are ignored}} omp-60-and-later-error {{use of undeclared identifier 'S1'}} for (i = 0; i < argc; ++i) foo(); return 0; diff --git a/clang/test/OpenMP/target_teams_distribute_parallel_for_simd_nowait_messages.cpp b/clang/test/OpenMP/target_teams_distribute_parallel_for_simd_nowait_messages.cpp index 0f13d35..f1f6c42 100644 --- a/clang/test/OpenMP/target_teams_distribute_parallel_for_simd_nowait_messages.cpp +++ b/clang/test/OpenMP/target_teams_distribute_parallel_for_simd_nowait_messages.cpp @@ -1,19 +1,33 @@ -// RUN: %clang_cc1 -fsyntax-only -verify -fopenmp %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=45 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized -// RUN: %clang_cc1 -fsyntax-only -verify -fopenmp-simd %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=45 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized void foo() { } int main(int argc, char **argv) { - int i; -#pragma omp target teams distribute parallel for simd nowait( // expected-warning {{extra tokens at the end of '#pragma omp target teams distribute parallel for simd' are ignored}} + int i, z; +#pragma omp target teams distribute parallel for simd nowait( // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams distribute parallel for simd' are ignored}} omp-60-and-later-error {{expected expression}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} for (i = 0; i < argc; ++i) foo(); #pragma omp target teams distribute parallel for simd nowait (argc)) // expected-warning {{extra tokens at the end of '#pragma omp target teams distribute parallel for simd' are ignored}} for (i = 0; i < argc; ++i) foo(); #pragma omp target teams distribute parallel for simd nowait device (-10u) for (i = 0; i < argc; ++i) foo(); -#pragma omp target teams distribute parallel for simd nowait (3.14) device (-10u) // expected-warning {{extra tokens at the end of '#pragma omp target teams distribute parallel for simd' are ignored}} +#pragma omp target teams distribute parallel for simd nowait (3.14) device (-10u) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams distribute parallel for simd' are ignored}} + for (i = 0; i < argc; ++i) foo(); +#pragma omp target teams distribute parallel for simd nowait (argc>> z) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams distribute parallel for simd' are ignored}} + for (i = 0; i < argc; ++i) foo(); +#pragma omp target teams distribute parallel for simd nowait (argv[1] = 2) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams distribute parallel for simd' are ignored}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} + for (i = 0; i < argc; ++i) foo(); +#pragma omp target teams distribute parallel for simd nowait (argc > 0 ? argv[1] : argv[2]) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams distribute parallel for simd' are ignored}} + for (i = 0; i < argc; ++i) foo(); +#pragma omp target teams distribute parallel for simd nowait (S1) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams distribute parallel for simd' are ignored}} omp-60-and-later-error {{use of undeclared identifier 'S1'}} for (i = 0; i < argc; ++i) foo(); return 0; diff --git a/clang/test/OpenMP/target_teams_distribute_simd_nowait_messages.cpp b/clang/test/OpenMP/target_teams_distribute_simd_nowait_messages.cpp index 1a9d0b5..7f60427 100644 --- a/clang/test/OpenMP/target_teams_distribute_simd_nowait_messages.cpp +++ b/clang/test/OpenMP/target_teams_distribute_simd_nowait_messages.cpp @@ -1,19 +1,33 @@ -// RUN: %clang_cc1 -fsyntax-only -verify -fopenmp %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=45 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized -// RUN: %clang_cc1 -fsyntax-only -verify -fopenmp-simd %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=45 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized void foo() { } int main(int argc, char **argv) { - int i; -#pragma omp target teams distribute simd nowait( // expected-warning {{extra tokens at the end of '#pragma omp target teams distribute simd' are ignored}} + int i, z; +#pragma omp target teams distribute simd nowait( // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams distribute simd' are ignored}} omp-60-and-later-error {{expected expression}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} for (i = 0; i < argc; ++i) foo(); #pragma omp target teams distribute simd nowait (argc)) // expected-warning {{extra tokens at the end of '#pragma omp target teams distribute simd' are ignored}} for (i = 0; i < argc; ++i) foo(); #pragma omp target teams distribute simd nowait device (-10u) for (i = 0; i < argc; ++i) foo(); -#pragma omp target teams distribute simd nowait (3.14) device (-10u) // expected-warning {{extra tokens at the end of '#pragma omp target teams distribute simd' are ignored}} +#pragma omp target teams distribute simd nowait (3.14) device (-10u) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams distribute simd' are ignored}} + for (i = 0; i < argc; ++i) foo(); +#pragma omp target teams distribute simd nowait (argc>> z) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams distribute simd' are ignored}} + for (i = 0; i < argc; ++i) foo(); +#pragma omp target teams distribute simd nowait (argv[1] = 2) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams distribute simd' are ignored}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} + for (i = 0; i < argc; ++i) foo(); +#pragma omp target teams distribute simd nowait (argc > 0 ? argv[1] : argv[2]) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams distribute simd' are ignored}} + for (i = 0; i < argc; ++i) foo(); +#pragma omp target teams distribute simd nowait (S1) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams distribute simd' are ignored}} omp-60-and-later-error {{use of undeclared identifier 'S1'}} for (i = 0; i < argc; ++i) foo(); return 0; diff --git a/clang/test/OpenMP/target_teams_nowait_messages.cpp b/clang/test/OpenMP/target_teams_nowait_messages.cpp index bed2f97..1298d3e 100644 --- a/clang/test/OpenMP/target_teams_nowait_messages.cpp +++ b/clang/test/OpenMP/target_teams_nowait_messages.cpp @@ -1,18 +1,31 @@ -// RUN: %clang_cc1 -verify -fopenmp -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=45 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized -// RUN: %clang_cc1 -verify -fopenmp-simd -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=45 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized void foo() { } int main(int argc, char **argv) { -#pragma omp target teams nowait( // expected-warning {{extra tokens at the end of '#pragma omp target teams' are ignored}} - foo(); -#pragma omp target teams nowait (argc)) // expected-warning {{extra tokens at the end of '#pragma omp target teams' are ignored}} + int z; +#pragma omp target teams nowait( // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams' are ignored}} omp-60-and-later-error {{expected expression}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} foo(); #pragma omp target teams nowait device (-10u) foo(); -#pragma omp target teams nowait (3.14) device (-10u) // expected-warning {{extra tokens at the end of '#pragma omp target teams' are ignored}} +#pragma omp target teams nowait (3.14) device (-10u) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams' are ignored}} + foo(); +#pragma omp target teams nowait (argc>> z) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams' are ignored}} + foo(); +#pragma omp target teams nowait (argv[1] = 2) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams' are ignored}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} + foo(); +#pragma omp target teams nowait (argc > 0 ? argv[1] : argv[2]) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams' are ignored}} + foo(); +#pragma omp target teams nowait (S1) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target teams' are ignored}} omp-60-and-later-error {{use of undeclared identifier 'S1'}} foo(); return 0; diff --git a/clang/test/OpenMP/target_update_nowait_messages.cpp b/clang/test/OpenMP/target_update_nowait_messages.cpp index fc0314a..1eb5764 100644 --- a/clang/test/OpenMP/target_update_nowait_messages.cpp +++ b/clang/test/OpenMP/target_update_nowait_messages.cpp @@ -1,19 +1,29 @@ -// RUN: %clang_cc1 -verify -fopenmp -ferror-limit 100 %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=45 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized -// RUN: %clang_cc1 -verify -fopenmp-simd -ferror-limit 100 %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=45 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=51 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=52 -verify=expected,omp-52-and-earlier -ferror-limit 100 -o - %s -Wuninitialized +// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -fopenmp-simd -fopenmp-version=60 -verify=expected,omp-60-and-later -ferror-limit 100 -o - %s -Wuninitialized int main(int argc, char **argv) { - int i; + int i, z; #pragma omp nowait target update to(i) // expected-error {{expected an OpenMP directive}} #pragma omp target nowait update to(i) // expected-error {{unexpected OpenMP clause 'update' in directive '#pragma omp target'}} expected-error {{unexpected OpenMP clause 'to' in directive '#pragma omp target'}} {} - #pragma omp target update nowait() to(i) // expected-warning {{extra tokens at the end of '#pragma omp target update' are ignored}} expected-error {{expected at least one 'to' clause or 'from' clause specified to '#pragma omp target update'}} - #pragma omp target update to(i) nowait( // expected-warning {{extra tokens at the end of '#pragma omp target update' are ignored}} - #pragma omp target update to(i) nowait (argc)) // expected-warning {{extra tokens at the end of '#pragma omp target update' are ignored}} + #pragma omp target update to(i) nowait() // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target update' are ignored}} omp-60-and-later-error {{expected expression}} + #pragma omp target update to(i) nowait( // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target update' are ignored}} omp-60-and-later-error {{expected expression}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} + #pragma omp target update to(i) nowait (argc)) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target update' are ignored}} omp-60-and-later-warning {{extra tokens at the end of '#pragma omp target_update' are ignored}} #pragma omp target update to(i) nowait device (-10u) - #pragma omp target update to(i) nowait (3.14) device (-10u) // expected-warning {{extra tokens at the end of '#pragma omp target update' are ignored}} - #pragma omp target update to(i) nowait nowait // expected-error {{directive '#pragma omp target update' cannot contain more than one 'nowait' clause}} - #pragma omp target update nowait to(i) nowait // expected-error {{directive '#pragma omp target update' cannot contain more than one 'nowait' clause}} + #pragma omp target update to(i) nowait (3.14) device (-10u) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target update' are ignored}} + #pragma omp target update to(i) nowait nowait // omp-52-and-earlier-error {{directive '#pragma omp target update' cannot contain more than one 'nowait' clause}} omp-60-and-later-error {{directive '#pragma omp target_update' cannot contain more than one 'nowait' clause}} + #pragma omp target update nowait to(i) nowait // omp-52-and-earlier-error {{directive '#pragma omp target update' cannot contain more than one 'nowait' clause}} omp-60-and-later-error {{directive '#pragma omp target_update' cannot contain more than one 'nowait' clause}} + #pragma omp target update to(i) nowait (argc>> z) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target update' are ignored}} + #pragma omp target update to(i) nowait (argv[1] = 2) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target update' are ignored}} omp-60-and-later-error {{expected ')'}} omp-60-and-later-note {{to match this '('}} + #pragma omp target update to(i) nowait (argc > 0 ? argv[1] : argv[2]) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target update' are ignored}} + #pragma omp target update to(i) nowait (S1) // omp-52-and-earlier-warning {{extra tokens at the end of '#pragma omp target update' are ignored}} omp-60-and-later-error {{use of undeclared identifier 'S1'}} return 0; } diff --git a/clang/test/PCH/cxx1z-decomposition.cpp b/clang/test/PCH/cxx1z-decomposition.cpp index 914ce80..340d5ea 100644 --- a/clang/test/PCH/cxx1z-decomposition.cpp +++ b/clang/test/PCH/cxx1z-decomposition.cpp @@ -22,7 +22,7 @@ constexpr int foo(Q &&q) { return a * 10 + b; } -auto [noinit]; // expected-error{{decomposition declaration '[noinit]' requires an initializer}} +auto [noinit]; // expected-error{{structured binding declaration '[noinit]' requires an initializer}} #else @@ -31,7 +31,7 @@ int k = decomp(arr); static_assert(foo({1, 2}) == 12); -// expected-error@15 {{cannot decompose non-class, non-array type 'const int'}} +// expected-error@15 {{cannot bind non-class, non-array type 'const int'}} int z = decomp(10); // expected-note {{instantiation of}} #endif diff --git a/clang/test/Parser/DelayedTemplateParsing.cpp b/clang/test/Parser/DelayedTemplateParsing.cpp index bcd286a..072c7ce 100644 --- a/clang/test/Parser/DelayedTemplateParsing.cpp +++ b/clang/test/Parser/DelayedTemplateParsing.cpp @@ -43,10 +43,10 @@ void undeclared() } -template <class T> void foo5() {} //expected-note {{previous definition is here}} +template <class T> void foo5() {} //expected-note {{previous definition is here}} template <class T> void foo5() {} // expected-error {{redefinition of 'foo5'}} - + namespace PR11931 { @@ -195,3 +195,12 @@ template <typename> struct PR38460_2 { } }; template struct PR38460_2<int>; + +namespace LateParsedAttrs { + template <class> + void f(int a) __attribute__((__diagnose_if__(a > 0, "foo", "error"))) {} + // expected-note@-1 {{from 'diagnose_if' attribute on 'f<int>'}} + void g() { + f<int>(1); // expected-error {{foo}} + } +} // namespace LateParsedAttrs diff --git a/clang/test/Parser/c2x-auto.c b/clang/test/Parser/c2x-auto.c index b878a5b..7f80b07 100644 --- a/clang/test/Parser/c2x-auto.c +++ b/clang/test/Parser/c2x-auto.c @@ -130,3 +130,30 @@ void atomic(void) { void attributes(void) { auto ident [[clang::annotate("this works")]] = 12; // c17-error {{type specifier missing, defaults to 'int'; ISO C99 and later do not support implicit int}} } + +/** GH163090 */ +constexpr auto int a1 = 0; // c23-error {{illegal storage class on file-scoped variable}} \ + c23-error {{cannot combine with previous 'auto' declaration specifier}} \ + c17-error {{illegal storage class on file-scoped variable}} \ + c17-error {{unknown type name 'constexpr'}} + +constexpr int auto a2 = 0; // c23-error {{cannot combine with previous 'int' declaration specifier}} \ + c17-error {{illegal storage class on file-scoped variable}} \ + c17-error {{unknown type name 'constexpr'}} + +auto int b1 = 0; // c23-error {{illegal storage class on file-scoped variable}} \ + c17-error {{illegal storage class on file-scoped variable}} + +int auto b2 = 0; // c23-error {{cannot combine with previous 'int' declaration specifier}} \ + c17-error {{illegal storage class on file-scoped variable}} + +void f() { + constexpr auto int c1 = 0; // c23-error {{cannot combine with previous 'auto' declaration specifier}} \ + c17-error {{use of undeclared identifier 'constexpr'}} + + constexpr int auto c2 = 0; // c23-error {{cannot combine with previous 'int' declaration specifier}} \ + c17-error {{use of undeclared identifier 'constexpr'}} + + auto int d1 = 0; + int auto d2 = 0; // c23-error {{cannot combine with previous 'int' declaration specifier}} +} diff --git a/clang/test/Parser/cxx1z-class-template-argument-deduction.cpp b/clang/test/Parser/cxx1z-class-template-argument-deduction.cpp index 9d27f83..ece00a0 100644 --- a/clang/test/Parser/cxx1z-class-template-argument-deduction.cpp +++ b/clang/test/Parser/cxx1z-class-template-argument-deduction.cpp @@ -158,7 +158,7 @@ namespace decl { A arr[3] = 0; // expected-error {{cannot form array of deduced class template specialization type}} A F::*pm = 0; // expected-error {{cannot form pointer to deduced class template specialization type}} A (*fp)() = 0; // expected-error {{cannot form function returning deduced class template specialization type}} - A [x, y] = 0; // expected-error {{cannot be declared with type 'A'}} expected-error {{type 'A<int>' decomposes into 0 elements, but 2 names were provided}} + A [x, y] = 0; // expected-error {{cannot be declared with type 'A'}} expected-error {{type 'A<int>' binds to 0 elements, but 2 names were provided}} } namespace typename_specifier { @@ -185,7 +185,7 @@ namespace typename_specifier { typename ::A arr[3] = 0; // expected-error {{cannot form array of deduced class template specialization type}} typename ::A F::*pm = 0; // expected-error {{cannot form pointer to deduced class template specialization type}} typename ::A (*fp)() = 0; // expected-error {{cannot form function returning deduced class template specialization type}} - typename ::A [x, y] = 0; // expected-error {{cannot be declared with type 'typename ::A'}} expected-error {{type 'typename ::A<int>' (aka 'A<int>') decomposes into 0}} + typename ::A [x, y] = 0; // expected-error {{cannot be declared with type 'typename ::A'}} expected-error {{type 'typename ::A<int>' (aka 'A<int>') binds to 0}} struct X { template<typename T> struct A { A(T); }; }; // expected-note 8{{declared here}} @@ -208,7 +208,7 @@ namespace typename_specifier { {typename T::A arr[3] = 0;} // expected-error {{refers to class template member}} {typename T::A F::*pm = 0;} // expected-error {{refers to class template member}} {typename T::A (*fp)() = 0;} // expected-error {{refers to class template member}} - {typename T::A [x, y] = 0;} // expected-error {{cannot be declared with type 'typename T::A'}} expected-error {{type 'typename typename_specifier::X::A<int>' (aka 'typename_specifier::X::A<int>') decomposes into 0}} + {typename T::A [x, y] = 0;} // expected-error {{cannot be declared with type 'typename T::A'}} expected-error {{type 'typename typename_specifier::X::A<int>' (aka 'typename_specifier::X::A<int>') binds to 0}} } template void f<X>(); // expected-note {{instantiation of}} diff --git a/clang/test/Parser/cxx1z-decomposition.cpp b/clang/test/Parser/cxx1z-decomposition.cpp index 274e24e..21c9419 100644 --- a/clang/test/Parser/cxx1z-decomposition.cpp +++ b/clang/test/Parser/cxx1z-decomposition.cpp @@ -5,7 +5,7 @@ struct S { int a, b, c; }; // expected-note 2 {{'S::a' declared here}} -// A simple-declaration can be a decompsition declaration. +// A simple-declaration can be a structured binding declaration. namespace SimpleDecl { auto [a_x, b_x, c_x] = S(); @@ -19,7 +19,7 @@ namespace SimpleDecl { } } -// A for-range-declaration can be a decomposition declaration. +// A for-range-declaration can be a structured binding declaration. namespace ForRangeDecl { extern S arr[10]; void h() { @@ -100,12 +100,12 @@ namespace BadSpecifiers { inline auto &[k] = n; // expected-error {{cannot be declared 'inline'}} const int K = 5; - auto ([c]) = s; // expected-error {{decomposition declaration cannot be declared with parentheses}} + auto ([c]) = s; // expected-error {{structured binding declaration cannot be declared with parentheses}} void g() { // defining-type-specifiers other than cv-qualifiers and 'auto' S [a] = s; // expected-error {{cannot be declared with type 'S'}} decltype(auto) [b] = s; // expected-error {{cannot be declared with type 'decltype(auto)'}} - auto ([c2]) = s; // cxx17-error {{decomposition declaration cannot be declared with parenthese}} \ + auto ([c2]) = s; // cxx17-error {{structured binding declaration cannot be declared with parenthese}} \ // post2b-error {{use of undeclared identifier 'c2'}} \ // post2b-error {{expected body of lambda expression}} \ @@ -114,7 +114,7 @@ namespace BadSpecifiers { auto [e][1] = s; // expected-error {{expected ';'}} expected-error {{requires an initializer}} // FIXME: This should fire the 'misplaced array declarator' diagnostic. - int [K] arr = {0}; // expected-error {{expected ';'}} expected-error {{cannot be declared with type 'int'}} expected-error {{decomposition declaration '[K]' requires an initializer}} + int [K] arr = {0}; // expected-error {{expected ';'}} expected-error {{cannot be declared with type 'int'}} expected-error {{structured binding declaration '[K]' requires an initializer}} int [5] arr = {0}; // expected-error {{place the brackets after the name}} auto *[f] = s; // expected-error {{cannot be declared with type 'auto *'}} expected-error {{incompatible initializer}} @@ -145,14 +145,14 @@ namespace MultiDeclarator { namespace Template { int n[3]; // Structured binding template is not allowed. - template<typename T> auto [a, b, c] = n; // expected-error {{decomposition declaration cannot be a template}} + template<typename T> auto [a, b, c] = n; // expected-error {{structured binding declaration cannot be a template}} } namespace Init { void f() { int arr[1]; struct S { int n; }; - auto &[bad1]; // expected-error {{decomposition declaration '[bad1]' requires an initializer}} + auto &[bad1]; // expected-error {{structured binding declaration '[bad1]' requires an initializer}} const auto &[bad2](S{}, S{}); // expected-error {{initializer for variable '[bad2]' with type 'const auto &' contains multiple expressions}} const auto &[bad3](); // expected-error {{expected expression}} auto &[good1] = arr; diff --git a/clang/test/Parser/cxx2c-binding-pack.cpp b/clang/test/Parser/cxx2c-binding-pack.cpp index 0daaad3..40d843e 100644 --- a/clang/test/Parser/cxx2c-binding-pack.cpp +++ b/clang/test/Parser/cxx2c-binding-pack.cpp @@ -12,7 +12,7 @@ void decompose_array() { auto [...] = arr; // #2 // expected-error@#2{{expected identifier}} // expected-error@#2{{{no names were provided}}} - // expected-warning@#2{{{does not allow a decomposition group to be empty}}} + // expected-warning@#2{{{does not allow a structured binding group to be empty}}} auto [a, ..., b] = arr; // #3 // expected-error@#3{{expected identifier}} // expected-error@#3{{{only 1 name was provided}}} diff --git a/clang/test/Preprocessor/embed___has_embed_parsing_errors.c b/clang/test/Preprocessor/embed___has_embed_parsing_errors.c index 9c512a48..8ab53f6 100644 --- a/clang/test/Preprocessor/embed___has_embed_parsing_errors.c +++ b/clang/test/Preprocessor/embed___has_embed_parsing_errors.c @@ -250,3 +250,35 @@ #if __has_embed("") // expected-error {{empty filename}} #endif + +// expected-error@+3 {{missing ')' after '__has_embed'}} \ + expected-error@+3 {{expected value in expression}} \ + expected-note@+3 {{to match this '('}} +#if __has_embed (__FILE__ foo limit(1) +#endif + +//--- test3.c +// expected-error@+3 {{missing ')' after '__has_embed'}} \ + expected-error@+3 {{expected value in expression}} \ + expected-note@+3 {{to match this '('}} +#if __has_embed (__FILE__ foo +#endif + +// expected-error@+3 {{missing ')' after '__has_embed'}} \ + expected-error@+3 {{expected value in expression}} \ + expected-note@+3 {{to match this '('}} +#if __has_embed ("a" foo() +#endif + +// expected-error@+3 {{missing ')' after '__has_embed'}} \ + expected-error@+3 {{expected value in expression}} \ + expected-note@+3 {{to match this '('}} +#if __has_embed ("a" bar() foo +#endif + +// expected-error@+3 {{missing ')' after '__has_embed'}} \ + expected-error@+3 {{expected value in expression}} \ + expected-note@+3 {{to match this '('}} +#if __has_embed (__FILE__ limit(1) foo +int a = __has_embed (__FILE__); +#endif diff --git a/clang/test/Preprocessor/embed_empty_file.c b/clang/test/Preprocessor/embed_empty_file.c new file mode 100644 index 0000000..5ad807e --- /dev/null +++ b/clang/test/Preprocessor/embed_empty_file.c @@ -0,0 +1,4 @@ +// RUN: %clang_cc1 -std=c23 %s -E -verify + +#embed <> // expected-error {{empty filename}} +#embed "" // expected-error {{empty filename}} diff --git a/clang/test/Preprocessor/init-mips.c b/clang/test/Preprocessor/init-mips.c index 125872a..c829eeb 100644 --- a/clang/test/Preprocessor/init-mips.c +++ b/clang/test/Preprocessor/init-mips.c @@ -1649,6 +1649,28 @@ // MIPS-ARCH-OCTEONP:#define __OCTEON__ 1 // MIPS-ARCH-OCTEONP:#define __mips_isa_rev 2 +// RUN: %clang_cc1 -E -dM -ffreestanding -triple=mips64-none-none \ +// RUN: -target-cpu i6400 < /dev/null \ +// RUN: | FileCheck -match-full-lines -check-prefix MIPS-ARCH-I6400 %s +// +// MIPS-ARCH-I6400:#define _MIPS_ARCH "i6400" +// MIPS-ARCH-I6400:#define _MIPS_ARCH_I6400 1 +// MIPS-ARCH-I6400:#define _MIPS_ISA _MIPS_ISA_MIPS64 +// MIPS-ARCH-I6400:#define __mips_abs2008 1 +// MIPS-ARCH-I6400:#define __mips_isa_rev 6 +// MIPS-ARCH-I6400:#define __mips_nan2008 1 + +// RUN: %clang_cc1 -E -dM -ffreestanding -triple=mips64-none-none \ +// RUN: -target-cpu i6500 < /dev/null \ +// RUN: | FileCheck -match-full-lines -check-prefix MIPS-ARCH-I6500 %s +// +// MIPS-ARCH-I6500:#define _MIPS_ARCH "i6500" +// MIPS-ARCH-I6500:#define _MIPS_ARCH_I6500 1 +// MIPS-ARCH-I6500:#define _MIPS_ISA _MIPS_ISA_MIPS64 +// MIPS-ARCH-I6500:#define __mips_abs2008 1 +// MIPS-ARCH-I6500:#define __mips_isa_rev 6 +// MIPS-ARCH-I6500:#define __mips_nan2008 1 + // Check MIPS float ABI macros // // RUN: %clang_cc1 -E -dM -ffreestanding \ diff --git a/clang/test/Preprocessor/predefined-arch-macros.c b/clang/test/Preprocessor/predefined-arch-macros.c index ecddf13..dac3649 100644 --- a/clang/test/Preprocessor/predefined-arch-macros.c +++ b/clang/test/Preprocessor/predefined-arch-macros.c @@ -2525,10 +2525,16 @@ // RUN: | FileCheck -match-full-lines %s -check-prefixes=CHECK_ARL_M32,CHECK_ARLS_M32,CHECK_KL_M32 // RUN: %clang -march=pantherlake -m32 -E -dM %s -o - 2>&1 \ // RUN: -target i386-unknown-linux \ -// RUN: | FileCheck -match-full-lines %s -check-prefixes=CHECK_ARL_M32,CHECK_ARLS_M32,CHECK_PTL_M32,CHECK_NKL_M32 +// RUN: | FileCheck -match-full-lines %s -check-prefixes=CHECK_ARL_M32,CHECK_ARLS_M32,CHECK_NKL_M32 +// RUN: %clang -march=wildcatlake -m32 -E -dM %s -o - 2>&1 \ +// RUN: -target i386-unknown-linux \ +// RUN: | FileCheck -match-full-lines %s -check-prefixes=CHECK_ARL_M32,CHECK_ARLS_M32,CHECK_NKL_M32 +// RUN: %clang -march=novalake -m32 -E -dM %s -o - 2>&1 \ +// RUN: -target i386-unknown-linux \ +// RUN: | FileCheck -match-full-lines %s -check-prefixes=CHECK_ARL_M32,CHECK_ARLS_M32,CHECK_NVL_M32,CHECK_NKL_M32 // RUN: %clang -march=clearwaterforest -m32 -E -dM %s -o - 2>&1 \ // RUN: -target i386-unknown-linux \ -// RUN: | FileCheck -match-full-lines %s -check-prefixes=CHECK_SRF_M32,CHECK_ARLS_M32,CHECK_PTL_M32,CHECK_CWF_M32,CHECK_NKL_M32 +// RUN: | FileCheck -match-full-lines %s -check-prefixes=CHECK_SRF_M32,CHECK_ARLS_M32,CHECK_NVL_M32,CHECK_UMSR_M32,CHECK_NKL_M32 // CHECK_ARL_M32: #define __ADX__ 1 // CHECK_ARL_M32: #define __AES__ 1 // CHECK_ARL_M32: #define __AVX2__ 1 @@ -2568,7 +2574,7 @@ // CHECK_ARL_M32: #define __POPCNT__ 1 // CHECK_ARL_M32-NOT: #define __PREFETCHI__ 1 // CHECK_ARLS_M32-NOT: #define __PREFETCHI__ 1 -// CHECK_PTL_M32: #define __PREFETCHI__ 1 +// CHECK_NVL_M32: #define __PREFETCHI__ 1 // CHECK_ARL_M32: #define __PRFCHW__ 1 // CHECK_ARL_M32: #define __PTWRITE__ 1 // CHECK_ARL_M32-NOT: #define __RAOINT__ 1 @@ -2595,8 +2601,7 @@ // CHECK_ARL_M32: #define __UINTR__ 1 // CHECK_ARL_M32-NOT: #define __USERMSR__ 1 // CHECK_ARLS_M32-NOT: #define __USERMSR__ 1 -// CHECK_PTL_M32-NOT: #define __USERMSR__ 1 -// CHECK_CWF_M32: #define __USERMSR__ 1 +// CHECK_UMSR_M32: #define __USERMSR__ 1 // CHECK_ARL_M32: #define __VAES__ 1 // CHECK_ARL_M32: #define __VPCLMULQDQ__ 1 // CHECK_ARL_M32: #define __WAITPKG__ 1 @@ -2630,10 +2635,16 @@ // RUN: | FileCheck -match-full-lines %s -check-prefixes=CHECK_ARL_M64,CHECK_ARLS_M64,CHECK_KL_M64 // RUN: %clang -march=pantherlake -m64 -E -dM %s -o - 2>&1 \ // RUN: -target i386-unknown-linux \ -// RUN: | FileCheck -match-full-lines %s -check-prefixes=CHECK_ARL_M64,CHECK_ARLS_M64,CHECK_PTL_M64,CHECK_NKL_M64 +// RUN: | FileCheck -match-full-lines %s -check-prefixes=CHECK_ARL_M64,CHECK_ARLS_M64,CHECK_NKL_M64 +// RUN: %clang -march=wildcatlake -m64 -E -dM %s -o - 2>&1 \ +// RUN: -target i386-unknown-linux \ +// RUN: | FileCheck -match-full-lines %s -check-prefixes=CHECK_ARL_M64,CHECK_ARLS_M64,CHECK_NKL_M64 +// RUN: %clang -march=novalake -m64 -E -dM %s -o - 2>&1 \ +// RUN: -target i386-unknown-linux \ +// RUN: | FileCheck -match-full-lines %s -check-prefixes=CHECK_ARL_M64,CHECK_ARLS_M64,CHECK_NVL_M64,CHECK_NKL_M64 // RUN: %clang -march=clearwaterforest -m64 -E -dM %s -o - 2>&1 \ // RUN: -target i386-unknown-linux \ -// RUN: | FileCheck -match-full-lines %s -check-prefixes=CHECK_ARL_M64,CHECK_SRF_M64,CHECK_ARLS_M64,CHECK_PTL_M64,CHECK_CWF_M64,CHECK_NKL_M64 +// RUN: | FileCheck -match-full-lines %s -check-prefixes=CHECK_ARL_M64,CHECK_SRF_M64,CHECK_ARLS_M64,CHECK_NVL_M64,CHECK_UMSR_M64,CHECK_NKL_M64 // CHECK_ARL_M64: #define __ADX__ 1 // CHECK_ARL_M64: #define __AES__ 1 // CHECK_ARL_M64: #define __AVX2__ 1 @@ -2673,7 +2684,7 @@ // CHECK_ARL_M64: #define __POPCNT__ 1 // CHECK_ARL_M64-NOT: #define __PREFETCHI__ 1 // CHECK_ARLS_M64-NOT: #define __PREFETCHI__ 1 -// CHECK_PTL_M64: #define __PREFETCHI__ 1 +// CHECK_NVL_M64: #define __PREFETCHI__ 1 // CHECK_ARL_M64: #define __PRFCHW__ 1 // CHECK_ARL_M64: #define __PTWRITE__ 1 // CHECK_ARL_M64-NOT: #define __RAOINT__ 1 @@ -2701,8 +2712,7 @@ // CHECK_ARL_M64: #define __UINTR__ 1 // CHECK_ARL_M64-NOT: #define __USERMSR__ 1 // CHECK_ARLS_M64-NOT: #define __USERMSR__ 1 -// CHECK_PTL_M64-NOT: #define __USERMSR__ 1 -// CHECK_CWF_M64: #define __USERMSR__ 1 +// CHECK_UMSR_M64: #define __USERMSR__ 1 // CHECK_ARL_M64: #define __VAES__ 1 // CHECK_ARL_M64: #define __VPCLMULQDQ__ 1 // CHECK_ARL_M64: #define __WAITPKG__ 1 diff --git a/clang/test/Preprocessor/systemz_asm_flag_output.c b/clang/test/Preprocessor/systemz_asm_flag_output.c new file mode 100644 index 0000000..b627499 --- /dev/null +++ b/clang/test/Preprocessor/systemz_asm_flag_output.c @@ -0,0 +1,4 @@ +// RUN: %clang -target systemz-unknown-unknown -x c -E -dM -o - %s | FileCheck -match-full-lines %s +// RUN: %clang -target s390x-unknown-unknown -x c -E -dM -o - %s | FileCheck -match-full-lines %s + +// CHECK: #define __GCC_ASM_FLAG_OUTPUTS__ 1 diff --git a/clang/test/Sema/atomic-ops-fp-minmax.c b/clang/test/Sema/atomic-ops-fp-minmax.c new file mode 100644 index 0000000..970af9f --- /dev/null +++ b/clang/test/Sema/atomic-ops-fp-minmax.c @@ -0,0 +1,16 @@ +// RUN: %clang_cc1 -verify -ffreestanding -triple=aarch64-linux-gnu %s +// REQUIRES: aarch64-registered-target + +#include <stdatomic.h> + +void memory_checks(_Float16 *p16, __bf16 *pbf, float *pf, double *pd) { + (void)__atomic_fetch_min(p16, (_Float16)1.0f, memory_order_relaxed); + (void)__atomic_fetch_max(pbf, (__bf16)2.0f, memory_order_acquire); + (void)__atomic_fetch_min(pf, 3.0f, memory_order_release); + (void)__atomic_fetch_max(pd, 4.0, memory_order_seq_cst); +} + +void nullPointerWarning(void) { + (void)__atomic_fetch_min((volatile float*)0, 42.0, memory_order_relaxed); // expected-warning {{null passed to a callee that requires a non-null argument}} + (void)__atomic_fetch_max((float*)0, 42.0, memory_order_relaxed); // expected-warning {{null passed to a callee that requires a non-null argument}} +} diff --git a/clang/test/Sema/attr-cpuspecific-cpus.c b/clang/test/Sema/attr-cpuspecific-cpus.c index 48543ac..2360d86 100644 --- a/clang/test/Sema/attr-cpuspecific-cpus.c +++ b/clang/test/Sema/attr-cpuspecific-cpus.c @@ -87,3 +87,5 @@ ATTR(cpu_specific(lunarlake)) void CPU37(void){} ATTR(cpu_specific(gracemont)) void CPU38(void){} ATTR(cpu_specific(pantherlake)) void CPU39(void){} ATTR(cpu_specific(clearwaterforest)) void CPU40(void){} +ATTR(cpu_specific(wildcatlake)) void CPU41(void){} +ATTR(cpu_specific(novalake)) void CPU42(void){} diff --git a/clang/test/SemaCUDA/Inputs/cuda.h b/clang/test/SemaCUDA/Inputs/cuda.h index 10db947..2bf45e03 100644 --- a/clang/test/SemaCUDA/Inputs/cuda.h +++ b/clang/test/SemaCUDA/Inputs/cuda.h @@ -13,6 +13,8 @@ #define __managed__ __attribute__((managed)) #define __grid_constant__ __attribute__((grid_constant)) #define __launch_bounds__(...) __attribute__((launch_bounds(__VA_ARGS__))) +#define __cluster_dims__(...) __attribute__((cluster_dims(__VA_ARGS__))) +#define __no_cluster__ __attribute__((no_cluster)) struct dim3 { unsigned x, y, z; diff --git a/clang/test/SemaCUDA/cluster_dims.cu b/clang/test/SemaCUDA/cluster_dims.cu new file mode 100644 index 0000000..dcb8737 --- /dev/null +++ b/clang/test/SemaCUDA/cluster_dims.cu @@ -0,0 +1,64 @@ +// RUN: %clang_cc1 -triple amdgcn-amd-amdhsa -fcuda-is-device -ast-print -x hip -verify=NS,all %s +// RUN: %clang_cc1 -triple nvptx-nvidia-cuda -fcuda-is-device -ast-print -x hip -verify=NS,all %s +// RUN: %clang_cc1 -triple amdgcn-amd-amdhsa -target-cpu gfx1250 -fcuda-is-device -ast-print -x hip -verify=amd,common,all %s | FileCheck -check-prefixes=CHECK %s +// RUN: %clang_cc1 -triple nvptx-nvidia-cuda -target-cpu sm_90 -fcuda-is-device -ast-print -x hip -verify=cuda,common,all %s | FileCheck -check-prefixes=CHECK %s +// RUN: %clang_cc1 -triple x86_64-pc-linux-gnu -aux-triple amdgcn-amd-amdhsa -ast-print -x hip -verify=amd,common,all %s | FileCheck -check-prefixes=CHECK %s + +#include "Inputs/cuda.h" + +const int constint = 4; + +// CHECK: __attribute__((global)) __attribute__((cluster_dims(2, 2, 2))) void test_literal_3d() +__global__ void __cluster_dims__(2, 2, 2) test_literal_3d() {} //NS-error {{'cluster_dims' is not supported for this GPU architecture}} + +// CHECK: __attribute__((global)) __attribute__((cluster_dims(2, 2))) void test_literal_2d() +__global__ void __cluster_dims__(2, 2) test_literal_2d() {} //NS-error {{'cluster_dims' is not supported for this GPU architecture}} + +// CHECK: __attribute__((global)) __attribute__((cluster_dims(4))) void test_literal_1d() +__global__ void __cluster_dims__(4) test_literal_1d() {} //NS-error {{'cluster_dims' is not supported for this GPU architecture}} + +// CHECK: __attribute__((global)) __attribute__((cluster_dims(constint, constint / 4, 1))) void test_constant() +__global__ void __cluster_dims__(constint, constint / 4, 1) test_constant() {} //NS-error {{'cluster_dims' is not supported for this GPU architecture}} + +// CHECK: template <int x, int y, int z> void test_template() __attribute__((cluster_dims(x, y, z))) +template <int x, int y, int z> void test_template(void) __cluster_dims__(x, y, z){} //NS-error {{'cluster_dims' is not supported for this GPU architecture}} + +// CHECK: template <int x, int y, int z> void test_template_expr() __attribute__((cluster_dims(x + constint, y, z))) +template <int x, int y, int z> void test_template_expr(void) __cluster_dims__(x + constint, y, z) {} //NS-error {{'cluster_dims' is not supported for this GPU architecture}} + +//NS-error@+1 {{'cluster_dims' is not supported for this GPU architecture}} +__global__ void __cluster_dims__(32, 2, 4) test_too_large_dim_0() {} // common-error {{integer constant expression evaluates to value 32 that cannot be represented in a 4-bit unsigned integer type}} + +// cuda-error@+2 {{cluster does not support more than 8 thread blocks; 64 provided}} +// amd-error@+1 {{cluster does not support more than 16 thread blocks; 64 provided}} +__global__ void __cluster_dims__(4, 4, 4) test_too_large_dim_1() {} // NS-error {{'cluster_dims' is not supported for this GPU architecture}} + +// cuda-error@+3 {{cluster does not support more than 8 thread blocks; 64 provided}} +// amd-error@+2 {{cluster does not support more than 16 thread blocks; 64 provided}} +template<unsigned a, unsigned b, unsigned c> +__global__ void __cluster_dims__(a, b, c) test_too_large_dim_template() {} // NS-error {{'cluster_dims' is not supported for this GPU architecture}} +template __global__ void test_too_large_dim_template<4, 4, 4>(); // common-note {{in instantiation of function template specialization 'test_too_large_dim_template<4U, 4U, 4U>' requested here}} + +int none_const_int = 4; + +//NS-error@+1 {{'cluster_dims' is not supported for this GPU architecture}} +__global__ void __cluster_dims__(none_const_int, 2, 4) test_non_constant_0() {} // common-error {{'cluster_dims' attribute requires parameter 0 to be an integer constant}} + +//NS-error@+1 {{'cluster_dims' is not supported for this GPU architecture}} +__global__ void __cluster_dims__(8, none_const_int / 2, 4) test_non_constant_1() {} // common-error {{'cluster_dims' attribute requires parameter 1 to be an integer constant}} + +//NS-error@+1 {{'cluster_dims' is not supported for this GPU architecture}} +__global__ void __cluster_dims__(8, 2, none_const_int / 4) test_non_constant_2() {} // common-error {{'cluster_dims' attribute requires parameter 2 to be an integer constant}} + +//NS-error@+1 {{'no_cluster' is not supported for this GPU architecture}} +__global__ void __no_cluster__ test_no_cluster() {} + +//NS-error@+2 {{'no_cluster' is not supported for this GPU architecture}} +//NS-error@+1 {{'cluster_dims' is not supported for this GPU architecture}} +__global__ void __no_cluster__ __cluster_dims__(2,2,2) test_have_both() {} // common-error {{'cluster_dims' and 'no_cluster' attributes are not compatible}} common-note {{conflicting attribute is here}} + +template <int... args> +__cluster_dims__(args) void test_template_variadic_args(void) {} // all-error {{expression contains unexpanded parameter pack 'args'}} + +template <int... args> +__cluster_dims__(1, args) void test_template_variadic_args_2(void) {} // all-error {{expression contains unexpanded parameter pack 'args'}} diff --git a/clang/test/SemaCXX/builtin-structured-binding-size.cpp b/clang/test/SemaCXX/builtin-structured-binding-size.cpp index bcd13a6..a8f6705 100644 --- a/clang/test/SemaCXX/builtin-structured-binding-size.cpp +++ b/clang/test/SemaCXX/builtin-structured-binding-size.cpp @@ -40,8 +40,8 @@ static_assert(__builtin_structured_binding_size(S5) == 2); // expected-error@-1 {{static assertion failed due to requirement '__builtin_structured_binding_size(S5) == 2'}} \ // expected-note@-1 {{expression evaluates to '1 == 2'}} static_assert(__builtin_structured_binding_size(S6) == 2); -// expected-error@-1 {{cannot decompose class type 'S6' because it has an anonymous union member}} \ -// expected-error@-1 {{type 'S6' cannot be decomposed}} \ +// expected-error@-1 {{cannot bind class type 'S6' because it has an anonymous union member}} \ +// expected-error@-1 {{type 'S6' cannot be bound}} \ // expected-error@-1 {{static assertion expression is not an integral constant expression}} \ // expected-note@#note-anon-union {{declared here}} static_assert(__builtin_structured_binding_size(S7) == 1); @@ -49,15 +49,15 @@ static_assert(__builtin_structured_binding_size(S7) == 1); static_assert(__builtin_structured_binding_size(SD) == 1); static_assert(__builtin_structured_binding_size(SE1) == 1); -// expected-error@-1 {{cannot decompose class type 'SE1': both it and its base class 'S1' have non-static data members}} \ -// expected-error@-1 {{type 'SE1' cannot be decomposed}} \ +// expected-error@-1 {{cannot bind class type 'SE1': both it and its base class 'S1' have non-static data members}} \ +// expected-error@-1 {{type 'SE1' cannot be bound}} \ // expected-error@-1 {{static assertion expression is not an integral constant expression}} static_assert(__builtin_structured_binding_size(U1) == 0); -// expected-error@-1 {{type 'U1' cannot be decomposed}} \ +// expected-error@-1 {{type 'U1' cannot be bound}} \ // expected-error@-1 {{static assertion expression is not an integral constant expression}} static_assert(__builtin_structured_binding_size(U2) == 0); -// expected-error@-1 {{type 'U2' cannot be decomposed}} \ +// expected-error@-1 {{type 'U2' cannot be bound}} \ // expected-error@-1 {{static assertion expression is not an integral constant expression}} @@ -75,7 +75,7 @@ static_assert(__builtin_structured_binding_size(decltype(__builtin_complex(0., 0 int VLASize; // expected-note {{declared here}} static_assert(__builtin_structured_binding_size(int[VLASize]) == 42); -// expected-error@-1 {{type 'int[VLASize]' cannot be decomposed}} \ +// expected-error@-1 {{type 'int[VLASize]' cannot be bound}} \ // expected-warning@-1 {{variable length arrays in C++ are a Clang extension}} \ // expected-note@-1 {{read of non-const variable 'VLASize' is not allowed in a constant expression}} \ // expected-error@-1 {{static assertion expression is not an integral constant expression}} @@ -84,10 +84,10 @@ static_assert(__builtin_structured_binding_size(int[VLASize]) == 42); struct Incomplete; // expected-note {{forward declaration of 'Incomplete'}} static_assert(__builtin_structured_binding_size(Incomplete) == 1); // expected-error@-1 {{incomplete type 'Incomplete' where a complete type is required}} \ -// expected-error@-1 {{type 'Incomplete' cannot be decomposed}} \ +// expected-error@-1 {{type 'Incomplete' cannot be bound}} \ // expected-error@-1 {{static assertion expression is not an integral constant expression}} static_assert(__builtin_structured_binding_size(Incomplete[]) == 1); -// expected-error@-1 {{type 'Incomplete[]' cannot be decomposed}} \ +// expected-error@-1 {{type 'Incomplete[]' cannot be bound}} \ // expected-error@-1 {{static assertion expression is not an integral constant expression}} static_assert(__builtin_structured_binding_size(Incomplete[0]) == 0); static_assert(__builtin_structured_binding_size(Incomplete[1]) == 1); @@ -97,12 +97,12 @@ static_assert(__builtin_structured_binding_size(Incomplete[42]) == 42); static_assert(__builtin_structured_binding_size(P1) == 0); // expected-error@-1 {{static assertion failed due to requirement '__builtin_structured_binding_size(P1) == 0'}} \ // expected-note@-1 {{expression evaluates to '1 == 0'}} \ -// expected-error@-1 {{cannot decompose private member 'a' of 'P1}} \ +// expected-error@-1 {{cannot bind private member 'a' of 'P1}} \ // expected-note@#note-private {{implicitly declared private here}} void func(int array[14], int x = __builtin_structured_binding_size(decltype(array))); -//expected-error@-1 {{type 'decltype(array)' (aka 'int *') cannot be decomposed}} +//expected-error@-1 {{type 'decltype(array)' (aka 'int *') cannot be bound}} struct SM { static int array[14]; @@ -115,7 +115,7 @@ struct T { }; T<int> t1; -// expected-error@#tpl-1 {{type 'int' cannot be decomposed}} \ +// expected-error@#tpl-1 {{type 'int' cannot be bound}} \ // expected-error@#tpl-1 {{non-type template argument is not a constant expression}} \ // expected-note@-1 {{in instantiation of default argument for 'T<int>' required here}} \ // expected-note@-1 {{while checking a default template argument used here}} \ @@ -183,8 +183,8 @@ static_assert(!is_destructurable<T0&>); static_assert(__builtin_structured_binding_size(T1) == 1); static_assert(__builtin_structured_binding_size(T42) == 42); static_assert(__builtin_structured_binding_size(TSizeError) == 42); -// expected-error@-1 {{cannot decompose this type; 'std::tuple_size<TSizeError>::value' is not a valid integral constant expression}} \ -// expected-error@-1 {{type 'TSizeError' cannot be decomposed}} \ +// expected-error@-1 {{cannot bind this type; 'std::tuple_size<TSizeError>::value' is not a valid integral constant expression}} \ +// expected-error@-1 {{type 'TSizeError' cannot be bound}} \ // expected-error@-1 {{static assertion expression is not an integral constant expression}} static_assert(!is_destructurable<TSizeError>); } @@ -195,7 +195,7 @@ struct S { int y; static_assert(__builtin_structured_binding_size(S) == 2); //expected-error@-1 {{incomplete type 'S' where a complete type is required}} \ - // expected-error@-1 {{type 'S' cannot be decomposed}} \ + // expected-error@-1 {{type 'S' cannot be bound}} \ // expected-error@-1 {{static assertion expression is not an integral constant expression}} \ // expected-note@-4 {{definition of 'S' is not complete until the closing '}'}} }; @@ -228,20 +228,20 @@ static_assert(__is_same_as(tag_of_t<S1>, int)); static_assert(__is_same_as(tag_of_t<int>, int)); // error // expected-error@-1 {{constraints not satisfied for alias template 'tag_of_t' [with T = int]}} -// expected-note@#tag-of-constr {{because substituted constraint expression is ill-formed: type 'int' cannot be decomposed}} +// expected-note@#tag-of-constr {{because substituted constraint expression is ill-formed: type 'int' cannot be bound}} struct MinusOne; template <> struct ::std::tuple_size<MinusOne> { static constexpr int value = -1; }; int minus_one = __builtin_structured_binding_size(MinusOne); -// expected-error@-1 {{cannot decompose this type; 'std::tuple_size<MinusOne>::value' is not a valid size: -1}} -// expected-error@-2 {{type 'MinusOne' cannot be decomposed}} +// expected-error@-1 {{cannot bind this type; 'std::tuple_size<MinusOne>::value' is not a valid size: -1}} +// expected-error@-2 {{type 'MinusOne' cannot be bound}} struct UintMax; template <> struct ::std::tuple_size<UintMax> { static constexpr unsigned value = -1; }; int uint_max = __builtin_structured_binding_size(UintMax); -// expected-error@-1 {{cannot decompose this type; 'std::tuple_size<UintMax>::value' is not a valid size: 4294967295}} -// expected-error@-2 {{type 'UintMax' cannot be decomposed}} +// expected-error@-1 {{cannot bind this type; 'std::tuple_size<UintMax>::value' is not a valid size: 4294967295}} +// expected-error@-2 {{type 'UintMax' cannot be bound}} diff --git a/clang/test/SemaCXX/cxx17-compat.cpp b/clang/test/SemaCXX/cxx17-compat.cpp index 99e41d8..1c9060d 100644 --- a/clang/test/SemaCXX/cxx17-compat.cpp +++ b/clang/test/SemaCXX/cxx17-compat.cpp @@ -76,18 +76,18 @@ struct ConstexprVirtual { struct C { int x, y, z; }; static auto [cx, cy, cz] = C(); #if __cplusplus <= 201703L - // expected-warning@-2 {{decomposition declaration declared 'static' is a C++20 extension}} + // expected-warning@-2 {{structured binding declaration declared 'static' is a C++20 extension}} #else - // expected-warning@-4 {{decomposition declaration declared 'static' is incompatible with C++ standards before C++20}} + // expected-warning@-4 {{structured binding declaration declared 'static' is incompatible with C++ standards before C++20}} #endif void f() { static thread_local auto [cx, cy, cz] = C(); #if __cplusplus <= 201703L - // expected-warning@-2 {{decomposition declaration declared 'static' is a C++20 extension}} - // expected-warning@-3 {{decomposition declaration declared 'thread_local' is a C++20 extension}} + // expected-warning@-2 {{structured binding declaration declared 'static' is a C++20 extension}} + // expected-warning@-3 {{structured binding declaration declared 'thread_local' is a C++20 extension}} #else - // expected-warning@-5 {{decomposition declaration declared 'static' is incompatible with C++ standards before C++20}} - // expected-warning@-6 {{decomposition declaration declared 'thread_local' is incompatible with C++ standards before C++20}} + // expected-warning@-5 {{structured binding declaration declared 'static' is incompatible with C++ standards before C++20}} + // expected-warning@-6 {{structured binding declaration declared 'thread_local' is incompatible with C++ standards before C++20}} #endif } diff --git a/clang/test/SemaCXX/cxx1z-decomposition.cpp b/clang/test/SemaCXX/cxx1z-decomposition.cpp index 6ee1249..158a3a6 100644 --- a/clang/test/SemaCXX/cxx1z-decomposition.cpp +++ b/clang/test/SemaCXX/cxx1z-decomposition.cpp @@ -3,22 +3,22 @@ // RUN: %clang_cc1 -std=c++20 -Wpre-c++20-compat -fexperimental-new-constant-interpreter -verify=expected %s void use_from_own_init() { - auto [a] = a; // expected-error {{binding 'a' cannot appear in the initializer of its own decomposition declaration}} + auto [a] = a; // expected-error {{binding 'a' cannot appear in the initializer of its own structured binding declaration}} } void num_elems() { struct A0 {} a0; int a1[1], a2[2]; - auto [] = a0; // expected-warning {{does not allow a decomposition group to be empty}} - auto [v1] = a0; // expected-error {{type 'struct A0' decomposes into 0 elements, but 1 name was provided}} - auto [] = a1; // expected-error {{type 'int[1]' decomposes into 1 element, but no names were provided}} expected-warning {{empty}} + auto [] = a0; // expected-warning {{does not allow a structured binding group to be empty}} + auto [v1] = a0; // expected-error {{type 'struct A0' binds to 0 elements, but 1 name was provided}} + auto [] = a1; // expected-error {{type 'int[1]' binds to 1 element, but no names were provided}} expected-warning {{empty}} auto [v2] = a1; - auto [v3, v4] = a1; // expected-error {{type 'int[1]' decomposes into 1 element, but 2 names were provided}} - auto [] = a2; // expected-error {{type 'int[2]' decomposes into 2 elements, but no names were provided}} expected-warning {{empty}} - auto [v5] = a2; // expected-error {{type 'int[2]' decomposes into 2 elements, but only 1 name was provided}} + auto [v3, v4] = a1; // expected-error {{type 'int[1]' binds to 1 element, but 2 names were provided}} + auto [] = a2; // expected-error {{type 'int[2]' binds to 2 elements, but no names were provided}} expected-warning {{empty}} + auto [v5] = a2; // expected-error {{type 'int[2]' binds to 2 elements, but only 1 name was provided}} auto [v6, v7] = a2; - auto [v8, v9, v10] = a2; // expected-error {{type 'int[2]' decomposes into 2 elements, but 3 names were provided}} + auto [v8, v9, v10] = a2; // expected-error {{type 'int[2]' binds to 2 elements, but 3 names were provided}} } // As a Clang extension, _Complex can be decomposed. @@ -105,7 +105,7 @@ void enclosing() { void bitfield() { struct { int a : 3, : 4, b : 5; } a; auto &[x, y] = a; - auto &[p, q, r] = a; // expected-error-re {{type 'struct (unnamed struct at {{.*}})' decomposes into 2 elements, but 3 names were provided}} + auto &[p, q, r] = a; // expected-error-re {{type 'struct (unnamed struct at {{.*}})' binds to 2 elements, but 3 names were provided}} } void for_range() { @@ -115,7 +115,7 @@ void for_range() { } int y[5]; - for (auto[c] : y) { // expected-error {{cannot decompose non-class, non-array type 'int'}} + for (auto[c] : y) { // expected-error {{cannot bind non-class, non-array type 'int'}} c++; } } @@ -157,16 +157,16 @@ int f2() { namespace lambdas { void f() { int n; - auto [a] = // expected-error {{cannot decompose lambda closure type}} + auto [a] = // expected-error {{cannot bind lambda closure type}} [n] {}; // expected-note {{lambda expression}} } - auto [] = []{}; // expected-warning {{ISO C++17 does not allow a decomposition group to be empty}} + auto [] = []{}; // expected-warning {{ISO C++17 does not allow a structured binding group to be empty}} int g() { int n = 0; auto a = [=](auto &self) { // expected-note {{lambda expression}} - auto &[capture] = self; // expected-error {{cannot decompose lambda closure type}} + auto &[capture] = self; // expected-error {{cannot bind lambda closure type}} ++capture; return n; }; @@ -188,14 +188,14 @@ namespace lambdas { struct A : decltype(x) { int n; }; - auto &&[r] = A{x, 0}; // expected-error-re {{cannot decompose class type 'A': both it and its base class 'decltype(x)' (aka '(lambda {{.*}})') have non-static data members}} + auto &&[r] = A{x, 0}; // expected-error-re {{cannot bind class type 'A': both it and its base class 'decltype(x)' (aka '(lambda {{.*}})') have non-static data members}} return r; } void j() { auto x = [] {}; struct A : decltype(x) {}; - auto &&[] = A{x}; // expected-warning {{ISO C++17 does not allow a decomposition group to be empty}} + auto &&[] = A{x}; // expected-warning {{ISO C++17 does not allow a structured binding group to be empty}} } } diff --git a/clang/test/SemaCXX/cxx2b-consteval-propagate.cpp b/clang/test/SemaCXX/cxx2b-consteval-propagate.cpp index 6cf0e02..331fe83 100644 --- a/clang/test/SemaCXX/cxx2b-consteval-propagate.cpp +++ b/clang/test/SemaCXX/cxx2b-consteval-propagate.cpp @@ -626,3 +626,20 @@ void fn() { } } + + +namespace GH109096 { +consteval void undefined(); +template <typename T> +struct scope_exit { + T t; + constexpr ~scope_exit() { t(); } + // expected-error@-1 {{call to immediate function 'GH109096::(anonymous class)::operator()' is not a constant expression}} \ + // expected-note@-1 {{implicit use of 'this' pointer is only allowed within the evaluation}} +}; + +scope_exit guard( // expected-note {{in instantiation of member function}} + []() { undefined(); } +); + +} diff --git a/clang/test/SemaCXX/cxx2c-binding-pack-nontemplate.cpp b/clang/test/SemaCXX/cxx2c-binding-pack-nontemplate.cpp index 638a2d8..0dfb52b 100644 --- a/clang/test/SemaCXX/cxx2c-binding-pack-nontemplate.cpp +++ b/clang/test/SemaCXX/cxx2c-binding-pack-nontemplate.cpp @@ -10,7 +10,7 @@ void decompose_array() { auto [x, ...rest, y] = arr; // cxx26-warning@+4 {{structured binding packs are incompatible with C++ standards before C++2c}} - // cxx23-error@+3 {{decomposition declaration cannot be declared 'constexpr'}} + // cxx23-error@+3 {{structured binding declaration cannot be declared 'constexpr'}} // cxx23-warning@+2 {{structured binding packs are a C++2c extension}} // nontemplate-error@+1 {{pack declaration outside of template}} constexpr auto [x_c, ...rest_c, y_c] = arr; diff --git a/clang/test/SemaCXX/cxx2c-binding-pack.cpp b/clang/test/SemaCXX/cxx2c-binding-pack.cpp index a8c1386..0b0eb88 100644 --- a/clang/test/SemaCXX/cxx2c-binding-pack.cpp +++ b/clang/test/SemaCXX/cxx2c-binding-pack.cpp @@ -82,7 +82,7 @@ void decompose_array() { static_assert(sizeof...(b) == 0); auto [...c] = arr1; static_assert(sizeof...(c) == 1); - auto [a1, ...b1, c1] = arr1; // expected-error{{decomposes into 1 element, but 3 names were provided}} + auto [a1, ...b1, c1] = arr1; // expected-error{{binds to 1 element, but 3 names were provided}} } // Test case by Younan Zhang. @@ -160,7 +160,7 @@ void now_i_know_my() { static_assert(sizeof...(e) == 2); auto [h, i, j, ...k] = C(); // OK, the pack k is empty static_assert(sizeof...(e) == 0); - auto [l, m, n, o, ...p] = C(); // expected-error{{{decomposes into 3 elements, but 5 names were provided}}} + auto [l, m, n, o, ...p] = C(); // expected-error{{{binds to 3 elements, but 5 names were provided}}} } } // namespace @@ -225,7 +225,7 @@ namespace GH125165 { template <typename = void> auto f(auto t) { const auto& [...pack] = t; - // expected-error@-1 {{cannot decompose non-class, non-array type 'char const'}} + // expected-error@-1 {{cannot bind non-class, non-array type 'char const'}} (pack, ...); }; diff --git a/clang/test/SemaCXX/cxx2c-template-template-param.cpp b/clang/test/SemaCXX/cxx2c-template-template-param.cpp index 4ad3fd9..704df31 100644 --- a/clang/test/SemaCXX/cxx2c-template-template-param.cpp +++ b/clang/test/SemaCXX/cxx2c-template-template-param.cpp @@ -350,3 +350,87 @@ template <A<concept missing<int>> T> // expected-error {{expected expression}} \ // expected-error {{expected unqualified-id}} auto f(); } + +namespace concept_arg_normalization { + +template <typename T, + template <typename...> concept C1> +concept one = (C1<T>); // #concept-arg-one + +template <typename T> +concept A = true; // #concept-arg-A + +template <typename T> +concept BetterA = A<T> && true; + +template <typename T> +concept B = true; // #concept-arg-B + +template <typename T> +concept False = false; // #concept-arg-False + +template <typename T> +requires one<T, A> +void f1(T){} // #concept-arg-f1-1 + +template <typename T> +requires one<T, B> +void f1(T){} // #concept-arg-f1-2 + +template <typename T> +requires one<T, A> +void f2(T){} + +template <typename T> +requires one<T, BetterA> +void f2(T){} + + +template <template <typename> concept CT> +requires one<int, A> +void f3(){} // #concept-arg-f3-1 + +template <template <typename> concept CT> +requires one<int, CT> +void f3(){} // #concept-arg-f3-2 + +template <typename T> +requires one<T, False> void f4(T){} // #concept-arg-f4 + + +void test() { + f1(0); + // expected-error@-1 {{call to 'f1' is ambiguous}} + // expected-note@#concept-arg-f1-1{{candidate function [with T = int]}} + // expected-note@#concept-arg-f1-2{{candidate function [with T = int]}} + // expected-note@#concept-arg-A {{similar constraint expressions not considered equivalent}} + // expected-note@#concept-arg-B {{similar constraint expression here}} + f2(0); + + f3<BetterA>(); + // expected-error@-1 {{call to 'f3' is ambiguous}} + // expected-note@#concept-arg-f3-1 {{candidate function [with CT = concept_arg_normalization::BetterA]}} + // expected-note@#concept-arg-f3-2 {{candidate function [with CT = concept_arg_normalization::BetterA]}} + +static_assert(one<int, A>); +static_assert(one<int, False>); +// expected-error@-1 {{static assertion failed}} \ +// expected-note@-1 {{because 'one<int, False>' evaluated to false}} +// expected-note@#concept-arg-one {{because 'int' does not satisfy 'False'}} +// expected-note@#concept-arg-False {{because 'false' evaluated to false}} + +f4(0); +// expected-error@-1 {{no matching function for call to 'f4'}} +// expected-note@#concept-arg-f4 {{candidate template ignored: constraints not satisfied [with T = int]}} +// expected-note@#concept-arg-f4 {{because 'one<int, False>'}} +// expected-note@#concept-arg-one {{because 'int' does not satisfy 'False'}} +// expected-note@#concept-arg-False {{because 'false' evaluated to false}} + +} + +template <typename T, template <typename...> concept C1> +concept TestBinary = T::a || C1<T>; +static_assert(TestBinary<int, A>); + + +} diff --git a/clang/test/SemaCXX/matrix-type.cpp b/clang/test/SemaCXX/matrix-type.cpp index 186d3b6..0f9bff8 100644 --- a/clang/test/SemaCXX/matrix-type.cpp +++ b/clang/test/SemaCXX/matrix-type.cpp @@ -14,6 +14,7 @@ void matrix_var_dimensions(int Rows, unsigned Columns, char C) { using matrix7_t = int __attribute__((matrix_type(1, 0))); // expected-error{{zero matrix size}} using matrix7_t = int __attribute__((matrix_type(char, 0))); // expected-error{{expected '(' for function-style cast or type construction}} using matrix8_t = int __attribute__((matrix_type(1048576, 1))); // expected-error{{matrix row size too large}} + using matrix8_t = int __attribute__((matrix_type(1048576, 1048576))); // expected-error{{matrix row and column size too large}} } struct S1 {}; diff --git a/clang/test/SemaCXX/sizeless-1.cpp b/clang/test/SemaCXX/sizeless-1.cpp index 688bbf0..ef3eec9 100644 --- a/clang/test/SemaCXX/sizeless-1.cpp +++ b/clang/test/SemaCXX/sizeless-1.cpp @@ -532,7 +532,7 @@ void cxx_only(int sel) { auto auto_int8 = local_int8; auto auto_int16 = local_int16; #if __cplusplus >= 201703L - auto [auto_int8_a] = local_int8; // expected-error {{cannot decompose non-class, non-array type 'svint8_t' (aka '__SVInt8_t')}} + auto [auto_int8_a] = local_int8; // expected-error {{cannot bind non-class, non-array type 'svint8_t' (aka '__SVInt8_t')}} #endif #endif diff --git a/clang/test/SemaCXX/type-traits.cpp b/clang/test/SemaCXX/type-traits.cpp index 901d510..9ef44d03 100644 --- a/clang/test/SemaCXX/type-traits.cpp +++ b/clang/test/SemaCXX/type-traits.cpp @@ -2066,7 +2066,28 @@ public: UserProvidedConstructor(const UserProvidedConstructor&) = delete; UserProvidedConstructor& operator=(const UserProvidedConstructor&) = delete; }; +struct Ctr { + Ctr(); +}; +struct Ctr2 { + Ctr2(); +private: + NoEligibleTrivialContructor inner; +}; + +struct NonCopyable{ + NonCopyable() = default; + NonCopyable(const NonCopyable&) = delete; +}; + +class C { + NonCopyable nc; +}; +static_assert(__builtin_is_implicit_lifetime(Ctr)); +static_assert(!__builtin_is_implicit_lifetime(Ctr2)); +static_assert(__builtin_is_implicit_lifetime(C)); +static_assert(!__builtin_is_implicit_lifetime(NoEligibleTrivialContructor)); static_assert(__builtin_is_implicit_lifetime(NonAggregate)); static_assert(!__builtin_is_implicit_lifetime(DataMemberInitializer)); static_assert(!__builtin_is_implicit_lifetime(UserProvidedConstructor)); @@ -2076,9 +2097,27 @@ template <typename T> class Tpl { Tpl() requires false = default ; }; -static_assert(!__builtin_is_implicit_lifetime(Tpl<int>)); +static_assert(__builtin_is_implicit_lifetime(Tpl<int>)); + +template <typename> +class MultipleDefaults { + MultipleDefaults() {}; + MultipleDefaults() requires true = default; +}; +static_assert(__builtin_is_implicit_lifetime(MultipleDefaults<int>)); +template <typename> +class MultipleDefaults2 { + MultipleDefaults2() requires true {}; + MultipleDefaults2() = default; +}; + +static_assert(__builtin_is_implicit_lifetime(MultipleDefaults2<int>)); + #endif + + + } void is_signed() diff --git a/clang/test/SemaCXX/warn-shadow-in-lambdas.cpp b/clang/test/SemaCXX/warn-shadow-in-lambdas.cpp index 2388c5f..0042ef0 100644 --- a/clang/test/SemaCXX/warn-shadow-in-lambdas.cpp +++ b/clang/test/SemaCXX/warn-shadow-in-lambdas.cpp @@ -259,11 +259,11 @@ struct S { int foo() { #ifdef AVOID - auto [a] = S{0}; // cxx14-warning {{decomposition declarations are a C++17 extension}} + auto [a] = S{0}; // cxx14-warning {{structured binding declarations are a C++17 extension}} [a = a] () { // No warning with basic -Wshadow due to uncaptured-local classification }(); #else - auto [a] = S{0}; // cxx14-warning {{decomposition declarations are a C++17 extension}} expected-note {{previous declaration is here}} + auto [a] = S{0}; // cxx14-warning {{structured binding declarations are a C++17 extension}} expected-note {{previous declaration is here}} [a = a] () { // expected-warning {{declaration shadows a structured binding}} }(); #endif diff --git a/clang/test/SemaHLSL/BuiltIns/matrix-basic_types-errors.hlsl b/clang/test/SemaHLSL/BuiltIns/matrix-basic_types-errors.hlsl index 6f6b01b..5ad1d6a 100644 --- a/clang/test/SemaHLSL/BuiltIns/matrix-basic_types-errors.hlsl +++ b/clang/test/SemaHLSL/BuiltIns/matrix-basic_types-errors.hlsl @@ -10,3 +10,26 @@ uint16_t4x4 mat2; matrix<int, 5, 5> mat3; // expected-error@-1 {{constraints not satisfied for alias template 'matrix' [with element = int, rows_count = 5, cols_count = 5]}} // expected-note@* {{because '5 <= 4' (5 <= 4) evaluated to false}} + +using float8x4 = __attribute__((matrix_type(8,4))) float; +// expected-error@-1 {{matrix row size too large}} + +using float4x8 = __attribute__((matrix_type(4,8))) float; +// expected-error@-1 {{matrix column size too large}} + +using float8x8 = __attribute__((matrix_type(8,8))) float; +// expected-error@-1 {{matrix row and column size too large}} + +using floatNeg1x4 = __attribute__((matrix_type(-1,4))) float; +// expected-error@-1 {{matrix row size too large}} +using float4xNeg1 = __attribute__((matrix_type(4,-1))) float; +// expected-error@-1 {{matrix column size too large}} +using floatNeg1xNeg1 = __attribute__((matrix_type(-1,-1))) float; +// expected-error@-1 {{matrix row and column size too large}} + +using float0x4 = __attribute__((matrix_type(0,4))) float; +// expected-error@-1 {{zero matrix size}} +using float4x0 = __attribute__((matrix_type(4,0))) float; +// expected-error@-1 {{zero matrix size}} +using float0x0 = __attribute__((matrix_type(0,0))) float; +// expected-error@-1 {{zero matrix size}} diff --git a/clang/test/SemaHLSL/Language/AggregateSplatCasts.hlsl b/clang/test/SemaHLSL/Language/AggregateSplatCasts.hlsl index e5a851c..623fca0 100644 --- a/clang/test/SemaHLSL/Language/AggregateSplatCasts.hlsl +++ b/clang/test/SemaHLSL/Language/AggregateSplatCasts.hlsl @@ -3,8 +3,9 @@ // splat from vec1 to vec // CHECK-LABEL: call1 // CHECK: CStyleCastExpr {{.*}} 'int3':'vector<int, 3>' <HLSLAggregateSplatCast> -// CHECK-NEXT: ImplicitCastExpr {{.*}} 'int' lvalue <FloatingToIntegral> part_of_explicit_cast -// CHECK-NEXT: ImplicitCastExpr {{.*}} 'float' lvalue <HLSLVectorTruncation> part_of_explicit_cast +// CHECK-NEXT: ImplicitCastExpr {{.*}} 'int' <FloatingToIntegral> part_of_explicit_cast +// CHECK-NEXT: ImplicitCastExpr {{.*}} 'float' <HLSLVectorTruncation> part_of_explicit_cast +// CHECK-NEXT: ImplicitCastExpr {{.*}} 'float1':'vector<float, 1>' <LValueToRValue> part_of_explicit_cast // CHECK-NEXT: DeclRefExpr {{.*}} 'float1':'vector<float, 1>' lvalue Var {{.*}} 'A' 'float1':'vector<float, 1>' export void call1() { float1 A = {1.0}; @@ -21,7 +22,7 @@ struct S { // splat from scalar to aggregate // CHECK-LABEL: call2 // CHECK: CStyleCastExpr {{.*}} 'S' <HLSLAggregateSplatCast> -// CHECK-NEXt: IntegerLiteral {{.*}} 'int' 5 +// CHECK-NEXT: IntegerLiteral {{.*}} 'int' 5 export void call2() { S s = (S)5; } diff --git a/clang/test/SemaHLSL/Language/ElementwiseCasts.hlsl b/clang/test/SemaHLSL/Language/ElementwiseCasts.hlsl index 563d3f0..8481cfc 100644 --- a/clang/test/SemaHLSL/Language/ElementwiseCasts.hlsl +++ b/clang/test/SemaHLSL/Language/ElementwiseCasts.hlsl @@ -21,3 +21,27 @@ export void call2() { float B[1] = {1.0}; B = (float[1])A; } + +struct S { + int A; + float F; +}; + +// cast from a struct +// CHECK-LABEL: call3 +// CHECK: CStyleCastExpr {{.*}} 'int[2]' <HLSLElementwiseCast> +// CHECK-NEXT: DeclRefExpr {{.*}} 'S' lvalue Var {{.*}} 'SS' 'S' +export void call3() { + S SS = {1,1.0}; + int A[2] = (int[2])SS; +} + +// cast from a vector +// CHECK-LABEL: call4 +// CHECK: CStyleCastExpr {{.*}} 'float[3]' <HLSLElementwiseCast> +// CHECK-NEXT: ImplicitCastExpr {{.*}} 'int3':'vector<int, 3>' <LValueToRValue> part_of_explicit_cast +// CHECK-NEXT: DeclRefExpr {{.*}} 'int3':'vector<int, 3>' lvalue Var {{.*}} 'A' 'int3' +export void call4() { + int3 A = {1,2,3}; + float B[3] = (float[3])A; +} diff --git a/clang/test/SemaHLSL/Language/TemplateOutArg.hlsl b/clang/test/SemaHLSL/Language/TemplateOutArg.hlsl index 2d6252c..3365dbe 100644 --- a/clang/test/SemaHLSL/Language/TemplateOutArg.hlsl +++ b/clang/test/SemaHLSL/Language/TemplateOutArg.hlsl @@ -195,6 +195,81 @@ T buzz(int X, T Y) { return X + Y; } +// Case 4: Verify that the parameter modifier attributes are instantiated +// for both templated and non-templated arguments, and that the non-templated +// out argument type is not modified by the template instantiation. + +// CHECK-LABEL: FunctionTemplateDecl {{.*}} fizz_two + +// Check the pattern decl. +// CHECK: FunctionDecl {{.*}} fizz_two 'void (inout T, out int)' +// CHECK-NEXT: ParmVarDecl {{.*}} referenced V 'T' +// CHECK-NEXT: HLSLParamModifierAttr {{.*}} inout +// CHECK-NEXT: ParmVarDecl {{.*}} referenced I 'int &__restrict' +// CHECK-NEXT: HLSLParamModifierAttr {{.*}} out + +// Check the 3 instantiations (int, float, & double). + +// CHECK-LABEL: FunctionDecl {{.*}} used fizz_two 'void (inout int, out int)' implicit_instantiation +// CHECK: ParmVarDecl {{.*}} used V 'int &__restrict' +// CHECK-NEXT: HLSLParamModifierAttr {{.*}} inout +// CHECK: ParmVarDecl {{.*}} used I 'int &__restrict' +// CHECK-NEXT: HLSLParamModifierAttr {{.*}} out + +// CHECK-LABEL: FunctionDecl {{.*}} used fizz_two 'void (inout float, out int)' implicit_instantiation +// CHECK: ParmVarDecl {{.*}} used V 'float &__restrict' +// CHECK-NEXT: HLSLParamModifierAttr {{.*}} inout +// CHECK: ParmVarDecl {{.*}} used I 'int &__restrict' +// CHECK-NEXT: HLSLParamModifierAttr {{.*}} out + +// CHECK-LABEL: FunctionDecl {{.*}} used fizz_two 'void (inout double, out int)' implicit_instantiation +// CHECK: ParmVarDecl {{.*}} used V 'double &__restrict' +// CHECK-NEXT: HLSLParamModifierAttr {{.*}} inout +// CHECK: ParmVarDecl {{.*}} used I 'int &__restrict' +// CHECK-NEXT: HLSLParamModifierAttr {{.*}} out +template <typename T> +void fizz_two(inout T V, out int I) { + V += 2; + I = V; +} + +// Case 5: Verify that `in` parameter modifier attributes are instantiated +// for both templated and non-templated arguments and argument types are not +// modified + +// CHECK-LABEL: FunctionTemplateDecl {{.*}} buzz_two + +// Check the pattern decl. +// CHECK: FunctionDecl {{.*}} buzz_two 'int (T, int)' +// CHECK-NEXT: ParmVarDecl {{.*}} referenced A 'T' +// CHECK-NEXT: HLSLParamModifierAttr {{.*}} in +// CHECK-NEXT: ParmVarDecl {{.*}} referenced B 'int' +// CHECK-NEXT: HLSLParamModifierAttr {{.*}} in + +// Check the 3 instantiations (int, float, & double). + +// CHECK-LABEL: FunctionDecl {{.*}} used buzz_two 'int (int, int)' implicit_instantiation +// CHECK: ParmVarDecl {{.*}} used A 'int' +// CHECK-NEXT: HLSLParamModifierAttr {{.*}} in +// CHECK: ParmVarDecl {{.*}} used B 'int' +// CHECK-NEXT: HLSLParamModifierAttr {{.*}} in + +// CHECK-LABEL: FunctionDecl {{.*}} used buzz_two 'int (float, int)' implicit_instantiation +// CHECK: ParmVarDecl {{.*}} used A 'float' +// CHECK-NEXT: HLSLParamModifierAttr {{.*}} in +// CHECK: ParmVarDecl {{.*}} used B 'int' +// CHECK-NEXT: HLSLParamModifierAttr {{.*}} in + +// CHECK-LABEL: FunctionDecl {{.*}} used buzz_two 'int (double, int)' implicit_instantiation +// CHECK: ParmVarDecl {{.*}} used A 'double' +// CHECK-NEXT: HLSLParamModifierAttr {{.*}} in +// CHECK: ParmVarDecl {{.*}} used B 'int' +// CHECK-NEXT: HLSLParamModifierAttr {{.*}} in +template <typename T> +int buzz_two(in T A, in int B) { + return A + B; +} + export void caller() { int X = 2; float Y = 3.3; @@ -211,4 +286,12 @@ export void caller() { X = buzz(X, X); Y = buzz(X, Y); Z = buzz(X, Z); + + fizz_two(X, X); + fizz_two(Y, X); + fizz_two(Z, X); + + X = buzz_two(X, X); + X = buzz_two(Y, X); + X = buzz_two(Z, X); } diff --git a/clang/test/SemaHLSL/Operators/logical-not.hlsl b/clang/test/SemaHLSL/Operators/logical-not.hlsl new file mode 100644 index 0000000..d06ca39 --- /dev/null +++ b/clang/test/SemaHLSL/Operators/logical-not.hlsl @@ -0,0 +1,53 @@ +// RUN: %clang_cc1 -finclude-default-header -triple dxil-pc-shadermodel6.6-library %s -fnative-half-type -ast-dump -ast-dump-filter=case | FileCheck %s + +// CHECK-LABEL: FunctionDecl {{.*}} used case1 'uint32_t2 (uint32_t2)' +// CHECK-NEXT: ParmVarDecl {{.*}} used b 'uint32_t2':'vector<uint32_t, 2>' +// CHECK-NEXT: CompoundStmt +// CHECK-NEXT: ReturnStmt +// CHECK-NEXT: ImplicitCastExpr {{.*}} 'vector<uint32_t, 2>' <IntegralCast> +// CHECK-NEXT: UnaryOperator {{.*}} 'vector<bool, 2>' prefix '!' cannot overflow +// CHECK-NEXT: ImplicitCastExpr {{.*}} 'vector<bool, 2>' <IntegralToBoolean> +// CHECK-NEXT: ImplicitCastExpr {{.*}} 'uint32_t2':'vector<uint32_t, 2>' <LValueToRValue> +// CHECK-NEXT: DeclRefExpr {{.*}} 'uint32_t2':'vector<uint32_t, 2>' lvalue ParmVar {{.*}} 'b' 'uint32_t2':'vector<uint32_t, 2>' +export uint32_t2 case1(uint32_t2 b) { + return !b; +} + +// CHECK-LABEL: FunctionDecl {{.*}} used case2 'int32_t3 (int32_t3)' +// CHECK-NEXT: ParmVarDecl {{.*}} used b 'int32_t3':'vector<int32_t, 3>' +// CHECK-NEXT: CompoundStmt +// CHECK-NEXT: ReturnStmt +// CHECK-NEXT: ImplicitCastExpr {{.*}} 'vector<int32_t, 3>' <IntegralCast> +// CHECK-NEXT: UnaryOperator {{.*}} 'vector<bool, 3>' prefix '!' cannot overflow +// CHECK-NEXT: ImplicitCastExpr {{.*}} 'vector<bool, 3>' <IntegralToBoolean> +// CHECK-NEXT: ImplicitCastExpr {{.*}} 'int32_t3':'vector<int32_t, 3>' <LValueToRValue> +// CHECK-NEXT: DeclRefExpr {{.*}} 'int32_t3':'vector<int32_t, 3>' lvalue ParmVar {{.*}} 'b' 'int32_t3':'vector<int32_t, 3>' +export int32_t3 case2(int32_t3 b) { + return !b; +} + +// CHECK-LABEL: FunctionDecl {{.*}} used case3 'float16_t (float16_t)' +// CHECK-NEXT: ParmVarDecl {{.*}} used b 'float16_t':'half' +// CHECK-NEXT: CompoundStmt +// CHECK-NEXT: ReturnStmt +// CHECK-NEXT: ImplicitCastExpr {{.*}} 'float16_t':'half' <IntegralToFloating> +// CHECK-NEXT: UnaryOperator {{.*}} 'bool' prefix '!' cannot overflow +// CHECK-NEXT: ImplicitCastExpr {{.*}} 'bool' <FloatingToBoolean> +// CHECK-NEXT: ImplicitCastExpr {{.*}} 'float16_t':'half' <LValueToRValue> +// CHECK-NEXT: DeclRefExpr {{.*}} 'float16_t':'half' lvalue ParmVar {{.*}} 'b' 'float16_t':'half' +export float16_t case3(float16_t b) { + return !b; +} + +// CHECK-LABEL: FunctionDecl {{.*}} used case4 'float4 (float4)' +// CHECK-NEXT: ParmVarDecl {{.*}} used b 'float4':'vector<float, 4>' +// CHECK-NEXT: CompoundStmt +// CHECK-NEXT: ReturnStmt +// CHECK-NEXT: ImplicitCastExpr {{.*}} 'vector<float, 4>' <IntegralToFloating> +// CHECK-NEXT: UnaryOperator {{.*}} 'vector<bool, 4>' prefix '!' cannot overflow +// CHECK-NEXT: ImplicitCastExpr {{.*}} 'vector<bool, 4>' <FloatingToBoolean> +// CHECK-NEXT: ImplicitCastExpr {{.*}} 'float4':'vector<float, 4>' <LValueToRValue> +// CHECK-NEXT: DeclRefExpr {{.*}} 'float4':'vector<float, 4>' lvalue ParmVar {{.*}} 'b' 'float4':'vector<float, 4>' +export float4 case4(float4 b) { + return !b; +} diff --git a/clang/test/SemaOpenACC/combined-construct-reduction-clause.cpp b/clang/test/SemaOpenACC/combined-construct-reduction-clause.cpp index 72d7e6b..1b50336 100644 --- a/clang/test/SemaOpenACC/combined-construct-reduction-clause.cpp +++ b/clang/test/SemaOpenACC/combined-construct-reduction-clause.cpp @@ -2,7 +2,7 @@ struct CompositeOfScalars { int I; - float F; + float F; // #COS_FLOAT short J; char C; double D; @@ -30,6 +30,11 @@ void uses(unsigned Parm) { for(int i = 0; i < 5; ++i); #pragma acc serial loop reduction(&: CoS, I, F) + // expected-error@-1{{variable of type 'float' referenced in OpenACC 'reduction' clause does not have a valid operation available}} + // expected-error@-2{{invalid operands to binary expression ('float' and 'float')}} + // expected-error@-3{{variable of type 'float' referenced in OpenACC 'reduction' clause does not have a valid operation available}} + // expected-note@#COS_FLOAT{{while forming combiner for compound type 'CompositeOfScalars'}} + // expected-error@-5{{invalid operands to binary expression ('float' and 'float')}} for(int i = 0; i < 5; ++i); #pragma acc kernels loop reduction(min: CoS, Array[I], Array[0:I]) diff --git a/clang/test/SemaOpenACC/compute-construct-clause-ast.cpp b/clang/test/SemaOpenACC/compute-construct-clause-ast.cpp index babff53..c4ff12d 100644 --- a/clang/test/SemaOpenACC/compute-construct-clause-ast.cpp +++ b/clang/test/SemaOpenACC/compute-construct-clause-ast.cpp @@ -86,15 +86,6 @@ void NormalFunc(int i, float f) { // CHECK-NEXT: CXXBoolLiteralExpr // CHECK-NEXT: NullStmt -#pragma acc serial reduction(|: f) - while(true); - // CHECK-NEXT: OpenACCComputeConstruct{{.*}}serial - // CHECK-NEXT: reduction clause Operator: | - // CHECK-NEXT: DeclRefExpr{{.*}} 'float' lvalue ParmVar{{.*}} 'f' 'float' - // CHECK-NEXT: WhileStmt - // CHECK-NEXT: CXXBoolLiteralExpr - // CHECK-NEXT: NullStmt - #pragma acc parallel reduction(^: i) while(true); @@ -283,16 +274,6 @@ void TemplFunc() { // CHECK-NEXT: CXXBoolLiteralExpr // CHECK-NEXT: NullStmt -#pragma acc parallel reduction(&: T::SomeFloat) - while(true); - // CHECK-NEXT: OpenACCComputeConstruct{{.*}}parallel - // CHECK-NEXT: reduction clause Operator: & - // CHECK-NEXT: DependentScopeDeclRefExpr{{.*}} '<dependent type>' lvalue - // CHECK-NEXT: NestedNameSpecifier TypeSpec 'T' - // CHECK-NEXT: WhileStmt - // CHECK-NEXT: CXXBoolLiteralExpr - // CHECK-NEXT: NullStmt - #pragma acc serial reduction(|: i) while(true); // CHECK-NEXT: OpenACCComputeConstruct{{.*}}serial @@ -475,14 +456,6 @@ void TemplFunc() { // CHECK-NEXT: CXXBoolLiteralExpr // CHECK-NEXT: NullStmt - // CHECK-NEXT: OpenACCComputeConstruct{{.*}}parallel - // CHECK-NEXT: reduction clause Operator: & - // CHECK-NEXT: DeclRefExpr{{.*}} 'const float' lvalue Var{{.*}} 'SomeFloat' 'const float' - // CHECK-NEXT: NestedNameSpecifier TypeSpec 'InstTy' - // CHECK-NEXT: WhileStmt - // CHECK-NEXT: CXXBoolLiteralExpr - // CHECK-NEXT: NullStmt - // CHECK-NEXT: OpenACCComputeConstruct{{.*}}serial // CHECK-NEXT: reduction clause Operator: | // CHECK-NEXT: DeclRefExpr{{.*}} 'typename InstTy::IntTy':'int' lvalue Var{{.*}} 'i' 'typename InstTy::IntTy':'int' diff --git a/clang/test/SemaOpenACC/compute-construct-reduction-clause.c b/clang/test/SemaOpenACC/compute-construct-reduction-clause.c index 265c498..96c01d0 100644 --- a/clang/test/SemaOpenACC/compute-construct-reduction-clause.c +++ b/clang/test/SemaOpenACC/compute-construct-reduction-clause.c @@ -2,7 +2,7 @@ struct CompositeOfScalars { int I; - float F; + float F; // #COS_FLOAT short J; char C; double D; @@ -58,6 +58,11 @@ void uses(unsigned Parm) { // Vars in a reduction must be a scalar or a composite of scalars. #pragma acc parallel reduction(&: CoS, I, F) + // expected-error@-1{{variable of type 'float' referenced in OpenACC 'reduction' clause does not have a valid operation available}} + // expected-error@-2{{invalid operands to binary expression ('float' and 'float')}} + // expected-error@-3{{variable of type 'float' referenced in OpenACC 'reduction' clause does not have a valid operation available}} + // expected-note@#COS_FLOAT{{while forming combiner for compound type 'CompositeOfScalars'}} + // expected-error@-5{{invalid operands to binary expression ('float' and 'float')}} while (1); // expected-error@+3{{invalid type 'struct CompositeOfScalars' used in OpenACC 'reduction' variable reference; type is not a scalar value}} // expected-note@#COS_FIELD{{used as field 'COS' of composite 'CompositeHasComposite'}} @@ -69,12 +74,18 @@ void uses(unsigned Parm) { while (1); #pragma acc parallel reduction(&: CoS, Array[I], Array[0:I]) + // expected-error@-1{{variable of type 'float' referenced in OpenACC 'reduction' clause does not have a valid operation available}} + // expected-note@#COS_FLOAT{{while forming combiner for compound type 'CompositeOfScalars'}} + // expected-error@-3{{invalid operands to binary expression ('float' and 'float')}} while (1); struct CompositeHasComposite ChCArray[5]; - // expected-error@+3{{invalid type 'struct CompositeOfScalars' used in OpenACC 'reduction' variable reference; type is not a scalar value}} + // expected-error@+6{{invalid type 'struct CompositeOfScalars' used in OpenACC 'reduction' variable reference; type is not a scalar value}} // expected-note@#COS_FIELD{{used as field 'COS' of composite 'CompositeHasComposite'}} - // expected-note@+1{{OpenACC 'reduction' variable reference must be a scalar variable or a composite of scalars, or an array, sub-array, or element of scalar types}} + // expected-note@+4{{OpenACC 'reduction' variable reference must be a scalar variable or a composite of scalars, or an array, sub-array, or element of scalar types}} + // expected-error@+3{{variable of type 'float' referenced in OpenACC 'reduction' clause does not have a valid operation available}} + // expected-note@#COS_FLOAT{{while forming combiner for compound type 'CompositeOfScalars'}} + // expected-error@+1{{invalid operands to binary expression ('float' and 'float')}} #pragma acc parallel reduction(&: CoS, Array[I], ChCArray[0:I]) while (1); diff --git a/clang/test/SemaOpenACC/compute-construct-reduction-clause.cpp b/clang/test/SemaOpenACC/compute-construct-reduction-clause.cpp index edc67ce..e3a487a 100644 --- a/clang/test/SemaOpenACC/compute-construct-reduction-clause.cpp +++ b/clang/test/SemaOpenACC/compute-construct-reduction-clause.cpp @@ -2,7 +2,7 @@ struct CompositeOfScalars { int I; - float F; + float F; // #COS_FLOAT short J; char C; double D; @@ -61,6 +61,11 @@ void uses(unsigned Parm) { // Vars in a reduction must be a scalar or a composite of scalars. #pragma acc parallel reduction(&: CoS, I, F) + // expected-error@-1{{variable of type 'float' referenced in OpenACC 'reduction' clause does not have a valid operation available}} + // expected-error@-2{{invalid operands to binary expression ('float' and 'float')}} + // expected-error@-3{{variable of type 'float' referenced in OpenACC 'reduction' clause does not have a valid operation available}} + // expected-note@#COS_FLOAT{{while forming combiner for compound type 'CompositeOfScalars'}} + // expected-error@-5{{invalid operands to binary expression ('float' and 'float')}} while (1); // expected-error@+3{{invalid type 'struct CompositeOfScalars' used in OpenACC 'reduction' variable reference; type is not a scalar value}} // expected-note@#COS_FIELD{{used as field 'COS' of composite 'CompositeHasComposite'}} @@ -71,6 +76,9 @@ void uses(unsigned Parm) { while (1); #pragma acc parallel reduction(&: CoS, Array[I], Array[0:I]) + // expected-error@-1{{variable of type 'float' referenced in OpenACC 'reduction' clause does not have a valid operation available}} + // expected-note@#COS_FLOAT{{while forming combiner for compound type 'CompositeOfScalars'}} + // expected-error@-3{{invalid operands to binary expression ('float' and 'float')}} while (1); // expected-error@+1{{OpenACC variable is not a valid variable name, sub-array, array element, or composite variable member}} @@ -211,6 +219,9 @@ void TemplUses(T Parm, U CoS, V ChC) { // Vars in a reduction must be a scalar or a composite of scalars. #pragma acc parallel reduction(&: CoS, Var, Parm) + // expected-error@-1{{variable of type 'float' referenced in OpenACC 'reduction' clause does not have a valid operation available}} + // expected-note@#COS_FLOAT{{while forming combiner for compound type 'CompositeOfScalars'}} + // expected-error@-3{{invalid operands to binary expression ('float' and 'float')}} while (1); // expected-error@+3{{invalid type 'struct CompositeOfScalars' used in OpenACC 'reduction' variable reference; type is not a scalar value}} // expected-note@#COS_FIELD{{used as field 'COS' of composite 'CompositeHasComposite'}} @@ -223,6 +234,9 @@ void TemplUses(T Parm, U CoS, V ChC) { while (1); #pragma acc parallel reduction(&: CoS, Array[Var], Array[0:Var]) + // expected-error@-1{{variable of type 'float' referenced in OpenACC 'reduction' clause does not have a valid operation available}} + // expected-note@#COS_FLOAT{{while forming combiner for compound type 'CompositeOfScalars'}} + // expected-error@-3{{invalid operands to binary expression ('float' and 'float')}} while (1); // expected-error@+1{{OpenACC variable is not a valid variable name, sub-array, array element, or composite variable member}} diff --git a/clang/test/SemaOpenACC/loop-construct-reduction-ast.cpp b/clang/test/SemaOpenACC/loop-construct-reduction-ast.cpp index 4afb370..2daa56b4 100644 --- a/clang/test/SemaOpenACC/loop-construct-reduction-ast.cpp +++ b/clang/test/SemaOpenACC/loop-construct-reduction-ast.cpp @@ -102,24 +102,6 @@ void NormalFunc(int i, float f) { // CHECK-NEXT: DeclRefExpr{{.*}}'i' 'int' // CHECK-NEXT: NullStmt -#pragma acc loop reduction(|: f) - for(int i = 0; i < 5; ++i); - // CHECK-NEXT: OpenACCLoopConstruct{{.*}}<orphan> - // CHECK-NEXT: reduction clause Operator: | - // CHECK-NEXT: DeclRefExpr{{.*}} 'float' lvalue ParmVar{{.*}} 'f' 'float' - // CHECK-NEXT: ForStmt - // CHECK-NEXT: DeclStmt - // CHECK-NEXT: VarDecl{{.*}} i 'int' - // CHECK-NEXT: IntegerLiteral{{.*}} 'int' 0 - // CHECK-NEXT: <<<NULL>>> - // CHECK-NEXT: BinaryOperator{{.*}}'<' - // CHECK-NEXT: ImplicitCastExpr - // CHECK-NEXT: DeclRefExpr{{.*}}'i' 'int' - // CHECK-NEXT: IntegerLiteral{{.*}} 'int' 5 - // CHECK-NEXT: UnaryOperator{{.*}}++ - // CHECK-NEXT: DeclRefExpr{{.*}}'i' 'int' - // CHECK-NEXT: NullStmt - #pragma acc loop reduction(^: i) for(int i = 0; i < 5; ++i); @@ -267,25 +249,6 @@ void TemplFunc() { // CHECK-NEXT: DeclRefExpr{{.*}}'i' 'int' // CHECK-NEXT: NullStmt -#pragma acc loop reduction(&: T::SomeFloat) - for(int i = 0; i < 5; ++i); - // CHECK-NEXT: OpenACCLoopConstruct{{.*}}<orphan> - // CHECK-NEXT: reduction clause Operator: & - // CHECK-NEXT: DependentScopeDeclRefExpr{{.*}} '<dependent type>' lvalue - // CHECK-NEXT: NestedNameSpecifier TypeSpec 'T' - // CHECK-NEXT: ForStmt - // CHECK-NEXT: DeclStmt - // CHECK-NEXT: VarDecl{{.*}} i 'int' - // CHECK-NEXT: IntegerLiteral{{.*}} 'int' 0 - // CHECK-NEXT: <<<NULL>>> - // CHECK-NEXT: BinaryOperator{{.*}}'<' - // CHECK-NEXT: ImplicitCastExpr - // CHECK-NEXT: DeclRefExpr{{.*}}'i' 'int' - // CHECK-NEXT: IntegerLiteral{{.*}} 'int' 5 - // CHECK-NEXT: UnaryOperator{{.*}}++ - // CHECK-NEXT: DeclRefExpr{{.*}}'i' 'int' - // CHECK-NEXT: NullStmt - #pragma acc loop reduction(|: i) for(int i = 0; i < 5; ++i); // CHECK-NEXT: OpenACCLoopConstruct{{.*}}<orphan> @@ -440,23 +403,6 @@ void TemplFunc() { // CHECK-NEXT: NullStmt // // CHECK-NEXT: OpenACCLoopConstruct{{.*}}<orphan> - // CHECK-NEXT: reduction clause Operator: & - // CHECK-NEXT: DeclRefExpr{{.*}} 'const float' lvalue Var{{.*}} 'SomeFloat' 'const float' - // CHECK-NEXT: NestedNameSpecifier TypeSpec 'InstTy' - // CHECK-NEXT: ForStmt - // CHECK-NEXT: DeclStmt - // CHECK-NEXT: VarDecl{{.*}} i 'int' - // CHECK-NEXT: IntegerLiteral{{.*}} 'int' 0 - // CHECK-NEXT: <<<NULL>>> - // CHECK-NEXT: BinaryOperator{{.*}}'<' - // CHECK-NEXT: ImplicitCastExpr - // CHECK-NEXT: DeclRefExpr{{.*}}'i' 'int' - // CHECK-NEXT: IntegerLiteral{{.*}} 'int' 5 - // CHECK-NEXT: UnaryOperator{{.*}}++ - // CHECK-NEXT: DeclRefExpr{{.*}}'i' 'int' - // CHECK-NEXT: NullStmt - // - // CHECK-NEXT: OpenACCLoopConstruct{{.*}}<orphan> // CHECK-NEXT: reduction clause Operator: | // CHECK-NEXT: DeclRefExpr{{.*}} 'typename InstTy::IntTy':'int' lvalue Var{{.*}} 'i' 'typename InstTy::IntTy':'int' // CHECK-NEXT: ForStmt diff --git a/clang/test/SemaTemplate/concepts.cpp b/clang/test/SemaTemplate/concepts.cpp index 3fbe7c0..aaa20f6 100644 --- a/clang/test/SemaTemplate/concepts.cpp +++ b/clang/test/SemaTemplate/concepts.cpp @@ -1416,6 +1416,31 @@ concept IsEntitySpec = } +namespace case8 { + +template <class T> +struct type_identity { + using type = T; +}; + +template <typename Inner> +struct Cat {}; + +template <typename T> +concept CatConcept = requires { + []<class Inner>(type_identity<Cat<Inner>>) {}(type_identity<T>{}); +}; + +template <typename Dummy> +struct Feeder { + template <CatConcept Dummy2> + void feed() noexcept {} +}; + +void main() { Feeder<int>{}.feed<Cat<int>>(); } + +} + } namespace GH162125 { diff --git a/clang/test/SemaTemplate/cxx1z-decomposition.cpp b/clang/test/SemaTemplate/cxx1z-decomposition.cpp index 779c4cf..b0cc4e6 100644 --- a/clang/test/SemaTemplate/cxx1z-decomposition.cpp +++ b/clang/test/SemaTemplate/cxx1z-decomposition.cpp @@ -15,7 +15,7 @@ namespace std { template<> struct std::tuple_size<C> { enum { value = 2 }; }; template<typename T> int decomp(T &t) { - auto &[a, b] = t; // expected-error {{type 'D' decomposes into 3 elements, but only 2 names were provided}} + auto &[a, b] = t; // expected-error {{type 'D' binds to 3 elements, but only 2 names were provided}} return a + b; // expected-error {{cannot initialize return object of type 'int' with an rvalue of type 'int *'}} } diff --git a/clang/tools/driver/cc1as_main.cpp b/clang/tools/driver/cc1as_main.cpp index ef79e45..50da2f8 100644 --- a/clang/tools/driver/cc1as_main.cpp +++ b/clang/tools/driver/cc1as_main.cpp @@ -417,7 +417,8 @@ getOutputStream(StringRef Path, DiagnosticsEngine &Diags, bool Binary) { } static bool ExecuteAssemblerImpl(AssemblerInvocation &Opts, - DiagnosticsEngine &Diags) { + DiagnosticsEngine &Diags, + IntrusiveRefCntPtr<vfs::FileSystem> VFS) { // Get the target specific parser. std::string Error; const Target *TheTarget = TargetRegistry::lookupTarget(Opts.Triple, Error); @@ -440,6 +441,7 @@ static bool ExecuteAssemblerImpl(AssemblerInvocation &Opts, // Record the location of the include directories so that the lexer can find // it later. SrcMgr.setIncludeDirs(Opts.IncludePaths); + SrcMgr.setVirtualFileSystem(VFS); std::unique_ptr<MCRegisterInfo> MRI(TheTarget->createMCRegInfo(Opts.Triple)); assert(MRI && "Unable to create target register info!"); @@ -632,8 +634,9 @@ static bool ExecuteAssemblerImpl(AssemblerInvocation &Opts, } static bool ExecuteAssembler(AssemblerInvocation &Opts, - DiagnosticsEngine &Diags) { - bool Failed = ExecuteAssemblerImpl(Opts, Diags); + DiagnosticsEngine &Diags, + IntrusiveRefCntPtr<vfs::FileSystem> VFS) { + bool Failed = ExecuteAssemblerImpl(Opts, Diags, VFS); // Delete output file if there were errors. if (Failed) { @@ -714,7 +717,7 @@ int cc1as_main(ArrayRef<const char *> Argv, const char *Argv0, void *MainAddr) { } // Execute the invocation, unless there were parsing errors. - bool Failed = Diags.hasErrorOccurred() || ExecuteAssembler(Asm, Diags); + bool Failed = Diags.hasErrorOccurred() || ExecuteAssembler(Asm, Diags, VFS); // If any timers were active but haven't been destroyed yet, print their // results now. diff --git a/clang/tools/libclang/CIndex.cpp b/clang/tools/libclang/CIndex.cpp index c39f337..fc27fd2 100644 --- a/clang/tools/libclang/CIndex.cpp +++ b/clang/tools/libclang/CIndex.cpp @@ -1644,9 +1644,9 @@ bool CursorVisitor::VisitTagTypeLoc(TagTypeLoc TL) { return true; if (TL.isDefinition()) - return Visit(MakeCXCursor(TL.getOriginalDecl(), TU, RegionOfInterest)); + return Visit(MakeCXCursor(TL.getDecl(), TU, RegionOfInterest)); - return Visit(MakeCursorTypeRef(TL.getOriginalDecl(), TL.getNameLoc(), TU)); + return Visit(MakeCursorTypeRef(TL.getDecl(), TL.getNameLoc(), TU)); } bool CursorVisitor::VisitTemplateTypeParmTypeLoc(TemplateTypeParmTypeLoc TL) { @@ -1852,7 +1852,7 @@ bool CursorVisitor::VisitPackIndexingTypeLoc(PackIndexingTypeLoc TL) { } bool CursorVisitor::VisitInjectedClassNameTypeLoc(InjectedClassNameTypeLoc TL) { - return Visit(MakeCursorTypeRef(TL.getOriginalDecl(), TL.getNameLoc(), TU)); + return Visit(MakeCursorTypeRef(TL.getDecl(), TL.getNameLoc(), TU)); } bool CursorVisitor::VisitAtomicTypeLoc(AtomicTypeLoc TL) { @@ -2386,7 +2386,9 @@ void OMPClauseEnqueue::VisitOMPDetachClause(const OMPDetachClause *C) { Visitor->AddStmt(C->getEventHandler()); } -void OMPClauseEnqueue::VisitOMPNowaitClause(const OMPNowaitClause *) {} +void OMPClauseEnqueue::VisitOMPNowaitClause(const OMPNowaitClause *C) { + Visitor->AddStmt(C->getCondition()); +} void OMPClauseEnqueue::VisitOMPUntiedClause(const OMPUntiedClause *) {} diff --git a/clang/tools/libclang/CIndexCodeCompletion.cpp b/clang/tools/libclang/CIndexCodeCompletion.cpp index 6d14f28..81448b4 100644 --- a/clang/tools/libclang/CIndexCodeCompletion.cpp +++ b/clang/tools/libclang/CIndexCodeCompletion.cpp @@ -617,7 +617,7 @@ namespace { if (!baseType.isNull()) { // Get the declaration for a class/struct/union/enum type if (const TagType *Tag = baseType->getAs<TagType>()) - D = Tag->getOriginalDecl(); + D = Tag->getDecl(); // Get the @interface declaration for a (possibly-qualified) Objective-C // object pointer type, e.g., NSString* else if (const ObjCObjectPointerType *ObjPtr = @@ -629,7 +629,7 @@ namespace { // Get the class for a C++ injected-class-name else if (const InjectedClassNameType *Injected = baseType->getAs<InjectedClassNameType>()) - D = Injected->getOriginalDecl(); + D = Injected->getDecl(); } if (D != nullptr) { diff --git a/clang/tools/libclang/CXCursor.cpp b/clang/tools/libclang/CXCursor.cpp index 56f113c..0a43d73 100644 --- a/clang/tools/libclang/CXCursor.cpp +++ b/clang/tools/libclang/CXCursor.cpp @@ -1338,7 +1338,7 @@ CXCursor cxcursor::getTypeRefCursor(CXCursor cursor) { if (const TypedefType *Typedef = Ty->getAs<TypedefType>()) return MakeCursorTypeRef(Typedef->getDecl(), Loc, TU); if (const TagType *Tag = Ty->getAs<TagType>()) - return MakeCursorTypeRef(Tag->getOriginalDecl(), Loc, TU); + return MakeCursorTypeRef(Tag->getDecl(), Loc, TU); if (const TemplateTypeParmType *TemplP = Ty->getAs<TemplateTypeParmType>()) return MakeCursorTypeRef(TemplP->getDecl(), Loc, TU); diff --git a/clang/tools/libclang/CXIndexDataConsumer.cpp b/clang/tools/libclang/CXIndexDataConsumer.cpp index 932201a..c97aefa 100644 --- a/clang/tools/libclang/CXIndexDataConsumer.cpp +++ b/clang/tools/libclang/CXIndexDataConsumer.cpp @@ -357,7 +357,7 @@ CXIndexDataConsumer::CXXBasesListInfo::CXXBasesListInfo(const CXXRecordDecl *D, TST = T->getAs<TemplateSpecializationType>()) { BaseD = TST->getTemplateName().getAsTemplateDecl(); } else if (const RecordType *RT = T->getAs<RecordType>()) { - BaseD = RT->getOriginalDecl(); + BaseD = RT->getDecl(); } if (BaseD) diff --git a/clang/tools/libclang/CXType.cpp b/clang/tools/libclang/CXType.cpp index d21ac7c..3feb563 100644 --- a/clang/tools/libclang/CXType.cpp +++ b/clang/tools/libclang/CXType.cpp @@ -546,11 +546,11 @@ try_again: break; case Type::Record: case Type::Enum: - D = cast<TagType>(TP)->getOriginalDecl(); + D = cast<TagType>(TP)->getDecl(); break; case Type::TemplateSpecialization: if (const RecordType *Record = TP->getAs<RecordType>()) - D = Record->getOriginalDecl(); + D = Record->getDecl(); else D = cast<TemplateSpecializationType>(TP)->getTemplateName() .getAsTemplateDecl(); @@ -564,7 +564,7 @@ try_again: break; case Type::InjectedClassName: - D = cast<InjectedClassNameType>(TP)->getOriginalDecl(); + D = cast<InjectedClassNameType>(TP)->getDecl(); break; // FIXME: Template type parameters! @@ -1037,7 +1037,7 @@ static long long visitRecordForValidation(const RecordDecl *RD) { return CXTypeLayoutError_Dependent; // recurse if (const RecordType *ChildType = I->getType()->getAs<RecordType>()) { - if (const RecordDecl *Child = ChildType->getOriginalDecl()) { + if (const RecordDecl *Child = ChildType->getDecl()) { long long ret = visitRecordForValidation(Child); if (ret < 0) return ret; diff --git a/clang/unittests/AST/ASTContextParentMapTest.cpp b/clang/unittests/AST/ASTContextParentMapTest.cpp index 4a8aa48..545eaf3 100644 --- a/clang/unittests/AST/ASTContextParentMapTest.cpp +++ b/clang/unittests/AST/ASTContextParentMapTest.cpp @@ -132,8 +132,7 @@ TEST(GetParents, FriendTypeLoc) { TypeLoc FrALoc = FrA.getFriendType()->getTypeLoc(); TypeLoc FrBLoc = FrB.getFriendType()->getTypeLoc(); bool FrAOwnsTag = FrALoc.getTypePtr()->getAs<TagType>()->isTagOwned(); - TagDecl *FrATagDecl = - FrALoc.getTypePtr()->getAs<TagType>()->getOriginalDecl(); + TagDecl *FrATagDecl = FrALoc.getTypePtr()->getAs<TagType>()->getDecl(); bool FrBOwnsTag = FrBLoc.getTypePtr()->getAs<TagType>()->isTagOwned(); EXPECT_THAT(Ctx.getParents(A), ElementsAre(DynTypedNode::create(TU))); diff --git a/clang/unittests/AST/ASTImporterTest.cpp b/clang/unittests/AST/ASTImporterTest.cpp index e7160bc..4c7ea5e 100644 --- a/clang/unittests/AST/ASTImporterTest.cpp +++ b/clang/unittests/AST/ASTImporterTest.cpp @@ -29,7 +29,7 @@ using internal::Matcher; static const RecordDecl *getRecordDeclOfFriend(FriendDecl *FD) { QualType Ty = FD->getFriendType()->getType().getCanonicalType(); - return cast<RecordType>(Ty)->getOriginalDecl(); + return cast<RecordType>(Ty)->getDecl(); } struct ImportExpr : TestImportBase {}; @@ -4614,7 +4614,7 @@ TEST_P(ImportFriendClasses, ImportOfClassDefinitionAndFwdFriendShouldBeLinked) { auto *Friend = FirstDeclMatcher<FriendDecl>().match(FromTU0, friendDecl()); QualType FT = Friend->getFriendType()->getType(); FT = FromTU0->getASTContext().getCanonicalType(FT); - auto *Fwd = cast<TagType>(FT)->getOriginalDecl(); + auto *Fwd = cast<TagType>(FT)->getDecl(); auto *ImportedFwd = Import(Fwd, Lang_CXX03); Decl *FromTU1 = getTuDecl( R"( diff --git a/clang/unittests/AST/StructuralEquivalenceTest.cpp b/clang/unittests/AST/StructuralEquivalenceTest.cpp index bee288d..24e20c7 100644 --- a/clang/unittests/AST/StructuralEquivalenceTest.cpp +++ b/clang/unittests/AST/StructuralEquivalenceTest.cpp @@ -719,13 +719,11 @@ TEST_F(StructuralEquivalenceRecordTest, AnonymousRecordsShouldBeInequivalent) { auto *A = FirstDeclMatcher<IndirectFieldDecl>().match( TU, indirectFieldDecl(hasName("a"))); auto *FA = cast<FieldDecl>(A->chain().front()); - RecordDecl *RA = - cast<RecordType>(FA->getType().getTypePtr())->getOriginalDecl(); + RecordDecl *RA = cast<RecordType>(FA->getType().getTypePtr())->getDecl(); auto *B = FirstDeclMatcher<IndirectFieldDecl>().match( TU, indirectFieldDecl(hasName("b"))); auto *FB = cast<FieldDecl>(B->chain().front()); - RecordDecl *RB = - cast<RecordType>(FB->getType().getTypePtr())->getOriginalDecl(); + RecordDecl *RB = cast<RecordType>(FB->getType().getTypePtr())->getDecl(); ASSERT_NE(RA, RB); EXPECT_TRUE(testStructuralMatch(RA, RA)); @@ -754,15 +752,13 @@ TEST_F(StructuralEquivalenceRecordTest, auto *A = FirstDeclMatcher<IndirectFieldDecl>().match( TU, indirectFieldDecl(hasName("a"))); auto *FA = cast<FieldDecl>(A->chain().front()); - RecordDecl *RA = - cast<RecordType>(FA->getType().getTypePtr())->getOriginalDecl(); + RecordDecl *RA = cast<RecordType>(FA->getType().getTypePtr())->getDecl(); auto *TU1 = get<1>(t); auto *A1 = FirstDeclMatcher<IndirectFieldDecl>().match( TU1, indirectFieldDecl(hasName("a"))); auto *FA1 = cast<FieldDecl>(A1->chain().front()); - RecordDecl *RA1 = - cast<RecordType>(FA1->getType().getTypePtr())->getOriginalDecl(); + RecordDecl *RA1 = cast<RecordType>(FA1->getType().getTypePtr())->getDecl(); RecordDecl *X = FirstDeclMatcher<RecordDecl>().match(TU, recordDecl(hasName("X"))); diff --git a/clang/unittests/Analysis/FlowSensitive/CMakeLists.txt b/clang/unittests/Analysis/FlowSensitive/CMakeLists.txt index 3bd4a6e..3508238 100644 --- a/clang/unittests/Analysis/FlowSensitive/CMakeLists.txt +++ b/clang/unittests/Analysis/FlowSensitive/CMakeLists.txt @@ -12,6 +12,7 @@ add_clang_unittest(ClangAnalysisFlowSensitiveTests LoggerTest.cpp MapLatticeTest.cpp MatchSwitchTest.cpp + MockHeaders.cpp MultiVarConstantPropagationTest.cpp RecordOpsTest.cpp SignAnalysisTest.cpp diff --git a/clang/unittests/Analysis/FlowSensitive/MockHeaders.cpp b/clang/unittests/Analysis/FlowSensitive/MockHeaders.cpp new file mode 100644 index 0000000..d3dee58 --- /dev/null +++ b/clang/unittests/Analysis/FlowSensitive/MockHeaders.cpp @@ -0,0 +1,2257 @@ +//===--- MockHeaders.cpp - Mock headers for dataflow analyses -*- C++ ---*-===// +// +// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. +// See https://llvm.org/LICENSE.txt for license information. +// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception +// +//===----------------------------------------------------------------------===// +// +// This file defines mock headers for testing of dataflow analyses. +// +//===----------------------------------------------------------------------===// + +#include "MockHeaders.h" + +namespace clang { +namespace dataflow { +namespace test { +static constexpr char CStdDefHeader[] = R"( +#ifndef CSTDDEF_H +#define CSTDDEF_H + +namespace std { + +typedef decltype(sizeof(char)) size_t; + +using nullptr_t = decltype(nullptr); + +} // namespace std + +typedef decltype(sizeof(char)) size_t; +typedef decltype(sizeof(char*)) ptrdiff_t; + +#endif // CSTDDEF_H +)"; + +static constexpr char StdTypeTraitsHeader[] = R"( +#ifndef STD_TYPE_TRAITS_H +#define STD_TYPE_TRAITS_H + +#include "cstddef.h" + +namespace std { + +template <typename T, T V> +struct integral_constant { + static constexpr T value = V; +}; + +using true_type = integral_constant<bool, true>; +using false_type = integral_constant<bool, false>; + +template< class T > struct remove_reference {typedef T type;}; +template< class T > struct remove_reference<T&> {typedef T type;}; +template< class T > struct remove_reference<T&&> {typedef T type;}; + +template <class T> + using remove_reference_t = typename remove_reference<T>::type; + +template <class T> +struct remove_extent { + typedef T type; +}; + +template <class T> +struct remove_extent<T[]> { + typedef T type; +}; + +template <class T, size_t N> +struct remove_extent<T[N]> { + typedef T type; +}; + +template <class T> +struct is_array : false_type {}; + +template <class T> +struct is_array<T[]> : true_type {}; + +template <class T, size_t N> +struct is_array<T[N]> : true_type {}; + +template <class> +struct is_function : false_type {}; + +template <class Ret, class... Args> +struct is_function<Ret(Args...)> : true_type {}; + +namespace detail { + +template <class T> +struct type_identity { + using type = T; +}; // or use type_identity (since C++20) + +template <class T> +auto try_add_pointer(int) -> type_identity<typename remove_reference<T>::type*>; +template <class T> +auto try_add_pointer(...) -> type_identity<T>; + +} // namespace detail + +template <class T> +struct add_pointer : decltype(detail::try_add_pointer<T>(0)) {}; + +template <bool B, class T, class F> +struct conditional { + typedef T type; +}; + +template <class T, class F> +struct conditional<false, T, F> { + typedef F type; +}; + +template <class T> +struct remove_cv { + typedef T type; +}; +template <class T> +struct remove_cv<const T> { + typedef T type; +}; +template <class T> +struct remove_cv<volatile T> { + typedef T type; +}; +template <class T> +struct remove_cv<const volatile T> { + typedef T type; +}; + +template <class T> +using remove_cv_t = typename remove_cv<T>::type; + +template <class T> +struct decay { + private: + typedef typename remove_reference<T>::type U; + + public: + typedef typename conditional< + is_array<U>::value, typename remove_extent<U>::type*, + typename conditional<is_function<U>::value, typename add_pointer<U>::type, + typename remove_cv<U>::type>::type>::type type; +}; + +template <bool B, class T = void> +struct enable_if {}; + +template <class T> +struct enable_if<true, T> { + typedef T type; +}; + +template <bool B, class T = void> +using enable_if_t = typename enable_if<B, T>::type; + +template <class T, class U> +struct is_same : false_type {}; + +template <class T> +struct is_same<T, T> : true_type {}; + +template <class T> +struct is_void : is_same<void, typename remove_cv<T>::type> {}; + +namespace detail { + +template <class T> +auto try_add_lvalue_reference(int) -> type_identity<T&>; +template <class T> +auto try_add_lvalue_reference(...) -> type_identity<T>; + +template <class T> +auto try_add_rvalue_reference(int) -> type_identity<T&&>; +template <class T> +auto try_add_rvalue_reference(...) -> type_identity<T>; + +} // namespace detail + +template <class T> +struct add_lvalue_reference : decltype(detail::try_add_lvalue_reference<T>(0)) { +}; + +template <class T> +struct add_rvalue_reference : decltype(detail::try_add_rvalue_reference<T>(0)) { +}; + +template <class T> +typename add_rvalue_reference<T>::type declval() noexcept; + +namespace detail { + +template <class T> +auto test_returnable(int) + -> decltype(void(static_cast<T (*)()>(nullptr)), true_type{}); +template <class> +auto test_returnable(...) -> false_type; + +template <class From, class To> +auto test_implicitly_convertible(int) + -> decltype(void(declval<void (&)(To)>()(declval<From>())), true_type{}); +template <class, class> +auto test_implicitly_convertible(...) -> false_type; + +} // namespace detail + +template <class From, class To> +struct is_convertible + : integral_constant<bool, + (decltype(detail::test_returnable<To>(0))::value && + decltype(detail::test_implicitly_convertible<From, To>( + 0))::value) || + (is_void<From>::value && is_void<To>::value)> {}; + +template <class From, class To> +inline constexpr bool is_convertible_v = is_convertible<From, To>::value; + +template <class...> +using void_t = void; + +template <class, class T, class... Args> +struct is_constructible_ : false_type {}; + +template <class T, class... Args> +struct is_constructible_<void_t<decltype(T(declval<Args>()...))>, T, Args...> + : true_type {}; + +template <class T, class... Args> +using is_constructible = is_constructible_<void_t<>, T, Args...>; + +template <class T, class... Args> +inline constexpr bool is_constructible_v = is_constructible<T, Args...>::value; + +template <class _Tp> +struct __uncvref { + typedef typename remove_cv<typename remove_reference<_Tp>::type>::type type; +}; + +template <class _Tp> +using __uncvref_t = typename __uncvref<_Tp>::type; + +template <bool _Val> +using _BoolConstant = integral_constant<bool, _Val>; + +template <class _Tp, class _Up> +using _IsSame = _BoolConstant<__is_same(_Tp, _Up)>; + +template <class _Tp, class _Up> +using _IsNotSame = _BoolConstant<!__is_same(_Tp, _Up)>; + +template <bool> +struct _MetaBase; +template <> +struct _MetaBase<true> { + template <class _Tp, class _Up> + using _SelectImpl = _Tp; + template <template <class...> class _FirstFn, template <class...> class, + class... _Args> + using _SelectApplyImpl = _FirstFn<_Args...>; + template <class _First, class...> + using _FirstImpl = _First; + template <class, class _Second, class...> + using _SecondImpl = _Second; + template <class _Result, class _First, class... _Rest> + using _OrImpl = + typename _MetaBase<_First::value != true && sizeof...(_Rest) != 0>:: + template _OrImpl<_First, _Rest...>; +}; + +template <> +struct _MetaBase<false> { + template <class _Tp, class _Up> + using _SelectImpl = _Up; + template <template <class...> class, template <class...> class _SecondFn, + class... _Args> + using _SelectApplyImpl = _SecondFn<_Args...>; + template <class _Result, class...> + using _OrImpl = _Result; +}; + +template <bool _Cond, class _IfRes, class _ElseRes> +using _If = typename _MetaBase<_Cond>::template _SelectImpl<_IfRes, _ElseRes>; + +template <class... _Rest> +using _Or = typename _MetaBase<sizeof...(_Rest) != + 0>::template _OrImpl<false_type, _Rest...>; + +template <bool _Bp, class _Tp = void> +using __enable_if_t = typename enable_if<_Bp, _Tp>::type; + +template <class...> +using __expand_to_true = true_type; +template <class... _Pred> +__expand_to_true<__enable_if_t<_Pred::value>...> __and_helper(int); +template <class...> +false_type __and_helper(...); +template <class... _Pred> +using _And = decltype(__and_helper<_Pred...>(0)); + +template <class _Pred> +struct _Not : _BoolConstant<!_Pred::value> {}; + +struct __check_tuple_constructor_fail { + static constexpr bool __enable_explicit_default() { return false; } + static constexpr bool __enable_implicit_default() { return false; } + template <class...> + static constexpr bool __enable_explicit() { + return false; + } + template <class...> + static constexpr bool __enable_implicit() { + return false; + } +}; + +template <typename, typename _Tp> +struct __select_2nd { + typedef _Tp type; +}; +template <class _Tp, class _Arg> +typename __select_2nd<decltype((declval<_Tp>() = declval<_Arg>())), + true_type>::type +__is_assignable_test(int); +template <class, class> +false_type __is_assignable_test(...); +template <class _Tp, class _Arg, + bool = is_void<_Tp>::value || is_void<_Arg>::value> +struct __is_assignable_imp + : public decltype((__is_assignable_test<_Tp, _Arg>(0))) {}; +template <class _Tp, class _Arg> +struct __is_assignable_imp<_Tp, _Arg, true> : public false_type {}; +template <class _Tp, class _Arg> +struct is_assignable : public __is_assignable_imp<_Tp, _Arg> {}; + +template <class _Tp> +struct __libcpp_is_integral : public false_type {}; +template <> +struct __libcpp_is_integral<bool> : public true_type {}; +template <> +struct __libcpp_is_integral<char> : public true_type {}; +template <> +struct __libcpp_is_integral<signed char> : public true_type {}; +template <> +struct __libcpp_is_integral<unsigned char> : public true_type {}; +template <> +struct __libcpp_is_integral<wchar_t> : public true_type {}; +template <> +struct __libcpp_is_integral<short> : public true_type {}; // NOLINT +template <> +struct __libcpp_is_integral<unsigned short> : public true_type {}; // NOLINT +template <> +struct __libcpp_is_integral<int> : public true_type {}; +template <> +struct __libcpp_is_integral<unsigned int> : public true_type {}; +template <> +struct __libcpp_is_integral<long> : public true_type {}; // NOLINT +template <> +struct __libcpp_is_integral<unsigned long> : public true_type {}; // NOLINT +template <> +struct __libcpp_is_integral<long long> : public true_type {}; // NOLINT +template <> // NOLINTNEXTLINE +struct __libcpp_is_integral<unsigned long long> : public true_type {}; +template <class _Tp> +struct is_integral + : public __libcpp_is_integral<typename remove_cv<_Tp>::type> {}; + +template <class _Tp> +struct __libcpp_is_floating_point : public false_type {}; +template <> +struct __libcpp_is_floating_point<float> : public true_type {}; +template <> +struct __libcpp_is_floating_point<double> : public true_type {}; +template <> +struct __libcpp_is_floating_point<long double> : public true_type {}; +template <class _Tp> +struct is_floating_point + : public __libcpp_is_floating_point<typename remove_cv<_Tp>::type> {}; + +template <class _Tp> +struct is_arithmetic + : public integral_constant<bool, is_integral<_Tp>::value || + is_floating_point<_Tp>::value> {}; + +template <class _Tp> +struct __libcpp_is_pointer : public false_type {}; +template <class _Tp> +struct __libcpp_is_pointer<_Tp*> : public true_type {}; +template <class _Tp> +struct is_pointer : public __libcpp_is_pointer<typename remove_cv<_Tp>::type> { +}; + +template <class _Tp> +struct __libcpp_is_member_pointer : public false_type {}; +template <class _Tp, class _Up> +struct __libcpp_is_member_pointer<_Tp _Up::*> : public true_type {}; +template <class _Tp> +struct is_member_pointer + : public __libcpp_is_member_pointer<typename remove_cv<_Tp>::type> {}; + +template <class _Tp> +struct __libcpp_union : public false_type {}; +template <class _Tp> +struct is_union : public __libcpp_union<typename remove_cv<_Tp>::type> {}; + +template <class T> +struct is_reference : false_type {}; +template <class T> +struct is_reference<T&> : true_type {}; +template <class T> +struct is_reference<T&&> : true_type {}; + +template <class T> +inline constexpr bool is_reference_v = is_reference<T>::value; + +struct __two { + char __lx[2]; +}; + +namespace __is_class_imp { +template <class _Tp> +char __test(int _Tp::*); +template <class _Tp> +__two __test(...); +} // namespace __is_class_imp +template <class _Tp> +struct is_class + : public integral_constant<bool, + sizeof(__is_class_imp::__test<_Tp>(0)) == 1 && + !is_union<_Tp>::value> {}; + +template <class _Tp> +struct __is_nullptr_t_impl : public false_type {}; +template <> +struct __is_nullptr_t_impl<nullptr_t> : public true_type {}; +template <class _Tp> +struct __is_nullptr_t + : public __is_nullptr_t_impl<typename remove_cv<_Tp>::type> {}; +template <class _Tp> +struct is_null_pointer + : public __is_nullptr_t_impl<typename remove_cv<_Tp>::type> {}; + +template <class _Tp> +struct is_enum + : public integral_constant< + bool, !is_void<_Tp>::value && !is_integral<_Tp>::value && + !is_floating_point<_Tp>::value && !is_array<_Tp>::value && + !is_pointer<_Tp>::value && !is_reference<_Tp>::value && + !is_member_pointer<_Tp>::value && !is_union<_Tp>::value && + !is_class<_Tp>::value && !is_function<_Tp>::value> {}; + +template <class _Tp> +struct is_scalar + : public integral_constant< + bool, is_arithmetic<_Tp>::value || is_member_pointer<_Tp>::value || + is_pointer<_Tp>::value || __is_nullptr_t<_Tp>::value || + is_enum<_Tp>::value> {}; +template <> +struct is_scalar<nullptr_t> : public true_type {}; + +} // namespace std + +#endif // STD_TYPE_TRAITS_H +)"; + +static constexpr char AbslTypeTraitsHeader[] = R"( +#ifndef ABSL_TYPE_TRAITS_H +#define ABSL_TYPE_TRAITS_H + +#include "std_type_traits.h" + +namespace absl { + +template <typename... Ts> +struct conjunction : std::true_type {}; + +template <typename T, typename... Ts> +struct conjunction<T, Ts...> + : std::conditional<T::value, conjunction<Ts...>, T>::type {}; + +template <typename T> +struct conjunction<T> : T {}; + +template <typename... Ts> +struct disjunction : std::false_type {}; + +template <typename T, typename... Ts> +struct disjunction<T, Ts...> + : std::conditional<T::value, T, disjunction<Ts...>>::type {}; + +template <typename T> +struct disjunction<T> : T {}; + +template <typename T> +struct negation : std::integral_constant<bool, !T::value> {}; + +template <bool B, typename T = void> +using enable_if_t = typename std::enable_if<B, T>::type; + + +template <bool B, typename T, typename F> +using conditional_t = typename std::conditional<B, T, F>::type; + +template <typename T> +using remove_cv_t = typename std::remove_cv<T>::type; + +template <typename T> +using remove_reference_t = typename std::remove_reference<T>::type; + +template <typename T> +using decay_t = typename std::decay<T>::type; + +struct in_place_t {}; + +constexpr in_place_t in_place; +} // namespace absl + +#endif // ABSL_TYPE_TRAITS_H +)"; + +static constexpr char StdStringHeader[] = R"( +#ifndef STRING_H +#define STRING_H + +namespace std { + +struct string { + string(const char*); + ~string(); + const char *c_str() const; + bool empty(); +}; + +struct string_view { + string_view(const char*); + ~string_view(); + bool empty(); +}; + +bool operator!=(const string &LHS, const char *RHS); + +} // namespace std + +#endif // STRING_H +)"; + +static constexpr char StdUtilityHeader[] = R"( +#ifndef UTILITY_H +#define UTILITY_H + +#include "std_type_traits.h" + +namespace std { + +template <typename T> +constexpr remove_reference_t<T>&& move(T&& x); + +template <typename T> +void swap(T& a, T& b) noexcept; + +} // namespace std + +#endif // UTILITY_H +)"; + +static constexpr char StdInitializerListHeader[] = R"( +#ifndef INITIALIZER_LIST_H +#define INITIALIZER_LIST_H + +namespace std { + +template <typename T> +class initializer_list { + public: + const T *a, *b; + initializer_list() noexcept; +}; + +} // namespace std + +#endif // INITIALIZER_LIST_H +)"; + +static constexpr char StdOptionalHeader[] = R"( +#include "std_initializer_list.h" +#include "std_type_traits.h" +#include "std_utility.h" + +namespace std { + +struct in_place_t {}; +constexpr in_place_t in_place; + +struct nullopt_t { + constexpr explicit nullopt_t() {} +}; +constexpr nullopt_t nullopt; + +template <class _Tp> +struct __optional_destruct_base { + constexpr void reset() noexcept; +}; + +template <class _Tp> +struct __optional_storage_base : __optional_destruct_base<_Tp> { + constexpr bool has_value() const noexcept; +}; + +template <typename _Tp> +class optional : private __optional_storage_base<_Tp> { + using __base = __optional_storage_base<_Tp>; + + public: + using value_type = _Tp; + + private: + struct _CheckOptionalArgsConstructor { + template <class _Up> + static constexpr bool __enable_implicit() { + return is_constructible_v<_Tp, _Up&&> && is_convertible_v<_Up&&, _Tp>; + } + + template <class _Up> + static constexpr bool __enable_explicit() { + return is_constructible_v<_Tp, _Up&&> && !is_convertible_v<_Up&&, _Tp>; + } + }; + template <class _Up> + using _CheckOptionalArgsCtor = + _If<_IsNotSame<__uncvref_t<_Up>, in_place_t>::value && + _IsNotSame<__uncvref_t<_Up>, optional>::value, + _CheckOptionalArgsConstructor, __check_tuple_constructor_fail>; + template <class _QualUp> + struct _CheckOptionalLikeConstructor { + template <class _Up, class _Opt = optional<_Up>> + using __check_constructible_from_opt = + _Or<is_constructible<_Tp, _Opt&>, is_constructible<_Tp, _Opt const&>, + is_constructible<_Tp, _Opt&&>, is_constructible<_Tp, _Opt const&&>, + is_convertible<_Opt&, _Tp>, is_convertible<_Opt const&, _Tp>, + is_convertible<_Opt&&, _Tp>, is_convertible<_Opt const&&, _Tp>>; + template <class _Up, class _QUp = _QualUp> + static constexpr bool __enable_implicit() { + return is_convertible<_QUp, _Tp>::value && + !__check_constructible_from_opt<_Up>::value; + } + template <class _Up, class _QUp = _QualUp> + static constexpr bool __enable_explicit() { + return !is_convertible<_QUp, _Tp>::value && + !__check_constructible_from_opt<_Up>::value; + } + }; + + template <class _Up, class _QualUp> + using _CheckOptionalLikeCtor = + _If<_And<_IsNotSame<_Up, _Tp>, is_constructible<_Tp, _QualUp>>::value, + _CheckOptionalLikeConstructor<_QualUp>, + __check_tuple_constructor_fail>; + + + template <class _Up, class _QualUp> + using _CheckOptionalLikeAssign = _If< + _And< + _IsNotSame<_Up, _Tp>, + is_constructible<_Tp, _QualUp>, + is_assignable<_Tp&, _QualUp> + >::value, + _CheckOptionalLikeConstructor<_QualUp>, + __check_tuple_constructor_fail + >; + + public: + constexpr optional() noexcept {} + constexpr optional(const optional&) = default; + constexpr optional(optional&&) = default; + constexpr optional(nullopt_t) noexcept {} + + template < + class _InPlaceT, class... _Args, + class = enable_if_t<_And<_IsSame<_InPlaceT, in_place_t>, + is_constructible<value_type, _Args...>>::value>> + constexpr explicit optional(_InPlaceT, _Args&&... __args); + + template <class _Up, class... _Args, + class = enable_if_t<is_constructible_v< + value_type, initializer_list<_Up>&, _Args...>>> + constexpr explicit optional(in_place_t, initializer_list<_Up> __il, + _Args&&... __args); + + template < + class _Up = value_type, + enable_if_t<_CheckOptionalArgsCtor<_Up>::template __enable_implicit<_Up>(), + int> = 0> + constexpr optional(_Up&& __v); + + template < + class _Up, + enable_if_t<_CheckOptionalArgsCtor<_Up>::template __enable_explicit<_Up>(), + int> = 0> + constexpr explicit optional(_Up&& __v); + + template <class _Up, enable_if_t<_CheckOptionalLikeCtor<_Up, _Up const&>:: + template __enable_implicit<_Up>(), + int> = 0> + constexpr optional(const optional<_Up>& __v); + + template <class _Up, enable_if_t<_CheckOptionalLikeCtor<_Up, _Up const&>:: + template __enable_explicit<_Up>(), + int> = 0> + constexpr explicit optional(const optional<_Up>& __v); + + template <class _Up, enable_if_t<_CheckOptionalLikeCtor<_Up, _Up&&>:: + template __enable_implicit<_Up>(), + int> = 0> + constexpr optional(optional<_Up>&& __v); + + template <class _Up, enable_if_t<_CheckOptionalLikeCtor<_Up, _Up&&>:: + template __enable_explicit<_Up>(), + int> = 0> + constexpr explicit optional(optional<_Up>&& __v); + + constexpr optional& operator=(nullopt_t) noexcept; + + optional& operator=(const optional&); + + optional& operator=(optional&&); + + template <class _Up = value_type, + class = enable_if_t<_And<_IsNotSame<__uncvref_t<_Up>, optional>, + _Or<_IsNotSame<__uncvref_t<_Up>, value_type>, + _Not<is_scalar<value_type>>>, + is_constructible<value_type, _Up>, + is_assignable<value_type&, _Up>>::value>> + constexpr optional& operator=(_Up&& __v); + + template <class _Up, enable_if_t<_CheckOptionalLikeAssign<_Up, _Up const&>:: + template __enable_assign<_Up>(), + int> = 0> + constexpr optional& operator=(const optional<_Up>& __v); + + template <class _Up, enable_if_t<_CheckOptionalLikeCtor<_Up, _Up&&>:: + template __enable_assign<_Up>(), + int> = 0> + constexpr optional& operator=(optional<_Up>&& __v); + + const _Tp& operator*() const&; + _Tp& operator*() &; + const _Tp&& operator*() const&&; + _Tp&& operator*() &&; + + const _Tp* operator->() const; + _Tp* operator->(); + + const _Tp& value() const&; + _Tp& value() &; + const _Tp&& value() const&&; + _Tp&& value() &&; + + template <typename U> + constexpr _Tp value_or(U&& v) const&; + template <typename U> + _Tp value_or(U&& v) &&; + + template <typename... Args> + _Tp& emplace(Args&&... args); + + template <typename U, typename... Args> + _Tp& emplace(std::initializer_list<U> ilist, Args&&... args); + + using __base::reset; + + constexpr explicit operator bool() const noexcept; + using __base::has_value; + + constexpr void swap(optional& __opt) noexcept; +}; + +template <typename T> +constexpr optional<typename std::decay<T>::type> make_optional(T&& v); + +template <typename T, typename... Args> +constexpr optional<T> make_optional(Args&&... args); + +template <typename T, typename U, typename... Args> +constexpr optional<T> make_optional(std::initializer_list<U> il, + Args&&... args); + +template <typename T, typename U> +constexpr bool operator==(const optional<T> &lhs, const optional<U> &rhs); +template <typename T, typename U> +constexpr bool operator!=(const optional<T> &lhs, const optional<U> &rhs); + +template <typename T> +constexpr bool operator==(const optional<T> &opt, nullopt_t); + +// C++20 and later do not define the following overloads because they are +// provided by rewritten candidates instead. +#if __cplusplus < 202002L +template <typename T> +constexpr bool operator==(nullopt_t, const optional<T> &opt); +template <typename T> +constexpr bool operator!=(const optional<T> &opt, nullopt_t); +template <typename T> +constexpr bool operator!=(nullopt_t, const optional<T> &opt); +#endif // __cplusplus < 202002L + +template <typename T, typename U> +constexpr bool operator==(const optional<T> &opt, const U &value); +template <typename T, typename U> +constexpr bool operator==(const T &value, const optional<U> &opt); +template <typename T, typename U> +constexpr bool operator!=(const optional<T> &opt, const U &value); +template <typename T, typename U> +constexpr bool operator!=(const T &value, const optional<U> &opt); + +} // namespace std +)"; + +static constexpr char AbslOptionalHeader[] = R"( +#include "absl_type_traits.h" +#include "std_initializer_list.h" +#include "std_type_traits.h" +#include "std_utility.h" + +namespace absl { + +struct nullopt_t { + constexpr explicit nullopt_t() {} +}; +constexpr nullopt_t nullopt; + +template <typename T> +class optional; + +namespace optional_internal { + +template <typename T, typename U> +struct is_constructible_convertible_from_optional + : std::integral_constant< + bool, std::is_constructible<T, optional<U>&>::value || + std::is_constructible<T, optional<U>&&>::value || + std::is_constructible<T, const optional<U>&>::value || + std::is_constructible<T, const optional<U>&&>::value || + std::is_convertible<optional<U>&, T>::value || + std::is_convertible<optional<U>&&, T>::value || + std::is_convertible<const optional<U>&, T>::value || + std::is_convertible<const optional<U>&&, T>::value> {}; + +template <typename T, typename U> +struct is_constructible_convertible_assignable_from_optional + : std::integral_constant< + bool, is_constructible_convertible_from_optional<T, U>::value || + std::is_assignable<T&, optional<U>&>::value || + std::is_assignable<T&, optional<U>&&>::value || + std::is_assignable<T&, const optional<U>&>::value || + std::is_assignable<T&, const optional<U>&&>::value> {}; + +} // namespace optional_internal + +template <typename T> +class optional { + public: + constexpr optional() noexcept; + + constexpr optional(nullopt_t) noexcept; + + optional(const optional&) = default; + + optional(optional&&) = default; + + template <typename InPlaceT, typename... Args, + absl::enable_if_t<absl::conjunction< + std::is_same<InPlaceT, in_place_t>, + std::is_constructible<T, Args&&...>>::value>* = nullptr> + constexpr explicit optional(InPlaceT, Args&&... args); + + template <typename U, typename... Args, + typename = typename std::enable_if<std::is_constructible< + T, std::initializer_list<U>&, Args&&...>::value>::type> + constexpr explicit optional(in_place_t, std::initializer_list<U> il, + Args&&... args); + + template < + typename U = T, + typename std::enable_if< + absl::conjunction<absl::negation<std::is_same< + in_place_t, typename std::decay<U>::type>>, + absl::negation<std::is_same< + optional<T>, typename std::decay<U>::type>>, + std::is_convertible<U&&, T>, + std::is_constructible<T, U&&>>::value, + bool>::type = false> + constexpr optional(U&& v); + + template < + typename U = T, + typename std::enable_if< + absl::conjunction<absl::negation<std::is_same< + in_place_t, typename std::decay<U>::type>>, + absl::negation<std::is_same< + optional<T>, typename std::decay<U>::type>>, + absl::negation<std::is_convertible<U&&, T>>, + std::is_constructible<T, U&&>>::value, + bool>::type = false> + explicit constexpr optional(U&& v); + + template <typename U, + typename std::enable_if< + absl::conjunction< + absl::negation<std::is_same<T, U>>, + std::is_constructible<T, const U&>, + absl::negation< + optional_internal:: + is_constructible_convertible_from_optional<T, U>>, + std::is_convertible<const U&, T>>::value, + bool>::type = false> + optional(const optional<U>& rhs); + + template <typename U, + typename std::enable_if< + absl::conjunction< + absl::negation<std::is_same<T, U>>, + std::is_constructible<T, const U&>, + absl::negation< + optional_internal:: + is_constructible_convertible_from_optional<T, U>>, + absl::negation<std::is_convertible<const U&, T>>>::value, + bool>::type = false> + explicit optional(const optional<U>& rhs); + + template < + typename U, + typename std::enable_if< + absl::conjunction< + absl::negation<std::is_same<T, U>>, std::is_constructible<T, U&&>, + absl::negation< + optional_internal::is_constructible_convertible_from_optional< + T, U>>, + std::is_convertible<U&&, T>>::value, + bool>::type = false> + optional(optional<U>&& rhs); + + template < + typename U, + typename std::enable_if< + absl::conjunction< + absl::negation<std::is_same<T, U>>, std::is_constructible<T, U&&>, + absl::negation< + optional_internal::is_constructible_convertible_from_optional< + T, U>>, + absl::negation<std::is_convertible<U&&, T>>>::value, + bool>::type = false> + explicit optional(optional<U>&& rhs); + + optional& operator=(nullopt_t) noexcept; + + optional& operator=(const optional& src); + + optional& operator=(optional&& src); + + template < + typename U = T, + typename = typename std::enable_if<absl::conjunction< + absl::negation< + std::is_same<optional<T>, typename std::decay<U>::type>>, + absl::negation< + absl::conjunction<std::is_scalar<T>, + std::is_same<T, typename std::decay<U>::type>>>, + std::is_constructible<T, U>, std::is_assignable<T&, U>>::value>::type> + optional& operator=(U&& v); + + template < + typename U, + typename = typename std::enable_if<absl::conjunction< + absl::negation<std::is_same<T, U>>, + std::is_constructible<T, const U&>, std::is_assignable<T&, const U&>, + absl::negation< + optional_internal:: + is_constructible_convertible_assignable_from_optional< + T, U>>>::value>::type> + optional& operator=(const optional<U>& rhs); + + template <typename U, + typename = typename std::enable_if<absl::conjunction< + absl::negation<std::is_same<T, U>>, std::is_constructible<T, U>, + std::is_assignable<T&, U>, + absl::negation< + optional_internal:: + is_constructible_convertible_assignable_from_optional< + T, U>>>::value>::type> + optional& operator=(optional<U>&& rhs); + + const T& operator*() const&; + T& operator*() &; + const T&& operator*() const&&; + T&& operator*() &&; + + const T* operator->() const; + T* operator->(); + + const T& value() const&; + T& value() &; + const T&& value() const&&; + T&& value() &&; + + template <typename U> + constexpr T value_or(U&& v) const&; + template <typename U> + T value_or(U&& v) &&; + + template <typename... Args> + T& emplace(Args&&... args); + + template <typename U, typename... Args> + T& emplace(std::initializer_list<U> ilist, Args&&... args); + + void reset() noexcept; + + constexpr explicit operator bool() const noexcept; + constexpr bool has_value() const noexcept; + + void swap(optional& rhs) noexcept; +}; + +template <typename T> +constexpr optional<typename std::decay<T>::type> make_optional(T&& v); + +template <typename T, typename... Args> +constexpr optional<T> make_optional(Args&&... args); + +template <typename T, typename U, typename... Args> +constexpr optional<T> make_optional(std::initializer_list<U> il, + Args&&... args); + +template <typename T, typename U> +constexpr bool operator==(const optional<T> &lhs, const optional<U> &rhs); +template <typename T, typename U> +constexpr bool operator!=(const optional<T> &lhs, const optional<U> &rhs); + +template <typename T> +constexpr bool operator==(const optional<T> &opt, nullopt_t); +template <typename T> +constexpr bool operator==(nullopt_t, const optional<T> &opt); +template <typename T> +constexpr bool operator!=(const optional<T> &opt, nullopt_t); +template <typename T> +constexpr bool operator!=(nullopt_t, const optional<T> &opt); + +template <typename T, typename U> +constexpr bool operator==(const optional<T> &opt, const U &value); +template <typename T, typename U> +constexpr bool operator==(const T &value, const optional<U> &opt); +template <typename T, typename U> +constexpr bool operator!=(const optional<T> &opt, const U &value); +template <typename T, typename U> +constexpr bool operator!=(const T &value, const optional<U> &opt); + +} // namespace absl +)"; + +static constexpr char BaseOptionalHeader[] = R"( +#include "std_initializer_list.h" +#include "std_type_traits.h" +#include "std_utility.h" + +namespace base { + +struct in_place_t {}; +constexpr in_place_t in_place; + +struct nullopt_t { + constexpr explicit nullopt_t() {} +}; +constexpr nullopt_t nullopt; + +template <typename T> +class Optional; + +namespace internal { + +template <typename T> +using RemoveCvRefT = std::remove_cv_t<std::remove_reference_t<T>>; + +template <typename T, typename U> +struct IsConvertibleFromOptional + : std::integral_constant< + bool, std::is_constructible<T, Optional<U>&>::value || + std::is_constructible<T, const Optional<U>&>::value || + std::is_constructible<T, Optional<U>&&>::value || + std::is_constructible<T, const Optional<U>&&>::value || + std::is_convertible<Optional<U>&, T>::value || + std::is_convertible<const Optional<U>&, T>::value || + std::is_convertible<Optional<U>&&, T>::value || + std::is_convertible<const Optional<U>&&, T>::value> {}; + +template <typename T, typename U> +struct IsAssignableFromOptional + : std::integral_constant< + bool, IsConvertibleFromOptional<T, U>::value || + std::is_assignable<T&, Optional<U>&>::value || + std::is_assignable<T&, const Optional<U>&>::value || + std::is_assignable<T&, Optional<U>&&>::value || + std::is_assignable<T&, const Optional<U>&&>::value> {}; + +} // namespace internal + +template <typename T> +class Optional { + public: + using value_type = T; + + constexpr Optional() = default; + constexpr Optional(const Optional& other) noexcept = default; + constexpr Optional(Optional&& other) noexcept = default; + + constexpr Optional(nullopt_t); + + template <typename U, + typename std::enable_if< + std::is_constructible<T, const U&>::value && + !internal::IsConvertibleFromOptional<T, U>::value && + std::is_convertible<const U&, T>::value, + bool>::type = false> + Optional(const Optional<U>& other) noexcept; + + template <typename U, + typename std::enable_if< + std::is_constructible<T, const U&>::value && + !internal::IsConvertibleFromOptional<T, U>::value && + !std::is_convertible<const U&, T>::value, + bool>::type = false> + explicit Optional(const Optional<U>& other) noexcept; + + template <typename U, + typename std::enable_if< + std::is_constructible<T, U&&>::value && + !internal::IsConvertibleFromOptional<T, U>::value && + std::is_convertible<U&&, T>::value, + bool>::type = false> + Optional(Optional<U>&& other) noexcept; + + template <typename U, + typename std::enable_if< + std::is_constructible<T, U&&>::value && + !internal::IsConvertibleFromOptional<T, U>::value && + !std::is_convertible<U&&, T>::value, + bool>::type = false> + explicit Optional(Optional<U>&& other) noexcept; + + template <class... Args> + constexpr explicit Optional(in_place_t, Args&&... args); + + template <class U, class... Args, + class = typename std::enable_if<std::is_constructible< + value_type, std::initializer_list<U>&, Args...>::value>::type> + constexpr explicit Optional(in_place_t, std::initializer_list<U> il, + Args&&... args); + + template < + typename U = value_type, + typename std::enable_if< + std::is_constructible<T, U&&>::value && + !std::is_same<internal::RemoveCvRefT<U>, in_place_t>::value && + !std::is_same<internal::RemoveCvRefT<U>, Optional<T>>::value && + std::is_convertible<U&&, T>::value, + bool>::type = false> + constexpr Optional(U&& value); + + template < + typename U = value_type, + typename std::enable_if< + std::is_constructible<T, U&&>::value && + !std::is_same<internal::RemoveCvRefT<U>, in_place_t>::value && + !std::is_same<internal::RemoveCvRefT<U>, Optional<T>>::value && + !std::is_convertible<U&&, T>::value, + bool>::type = false> + constexpr explicit Optional(U&& value); + + Optional& operator=(const Optional& other) noexcept; + + Optional& operator=(Optional&& other) noexcept; + + Optional& operator=(nullopt_t); + + template <typename U> + typename std::enable_if< + !std::is_same<internal::RemoveCvRefT<U>, Optional<T>>::value && + std::is_constructible<T, U>::value && + std::is_assignable<T&, U>::value && + (!std::is_scalar<T>::value || + !std::is_same<typename std::decay<U>::type, T>::value), + Optional&>::type + operator=(U&& value) noexcept; + + template <typename U> + typename std::enable_if<!internal::IsAssignableFromOptional<T, U>::value && + std::is_constructible<T, const U&>::value && + std::is_assignable<T&, const U&>::value, + Optional&>::type + operator=(const Optional<U>& other) noexcept; + + template <typename U> + typename std::enable_if<!internal::IsAssignableFromOptional<T, U>::value && + std::is_constructible<T, U>::value && + std::is_assignable<T&, U>::value, + Optional&>::type + operator=(Optional<U>&& other) noexcept; + + const T& operator*() const&; + T& operator*() &; + const T&& operator*() const&&; + T&& operator*() &&; + + const T* operator->() const; + T* operator->(); + + const T& value() const&; + T& value() &; + const T&& value() const&&; + T&& value() &&; + + template <typename U> + constexpr T value_or(U&& v) const&; + template <typename U> + T value_or(U&& v) &&; + + template <typename... Args> + T& emplace(Args&&... args); + + template <typename U, typename... Args> + T& emplace(std::initializer_list<U> ilist, Args&&... args); + + void reset() noexcept; + + constexpr explicit operator bool() const noexcept; + constexpr bool has_value() const noexcept; + + void swap(Optional& other); +}; + +template <typename T> +constexpr Optional<typename std::decay<T>::type> make_optional(T&& v); + +template <typename T, typename... Args> +constexpr Optional<T> make_optional(Args&&... args); + +template <typename T, typename U, typename... Args> +constexpr Optional<T> make_optional(std::initializer_list<U> il, + Args&&... args); + +template <typename T, typename U> +constexpr bool operator==(const Optional<T> &lhs, const Optional<U> &rhs); +template <typename T, typename U> +constexpr bool operator!=(const Optional<T> &lhs, const Optional<U> &rhs); + +template <typename T> +constexpr bool operator==(const Optional<T> &opt, nullopt_t); +template <typename T> +constexpr bool operator==(nullopt_t, const Optional<T> &opt); +template <typename T> +constexpr bool operator!=(const Optional<T> &opt, nullopt_t); +template <typename T> +constexpr bool operator!=(nullopt_t, const Optional<T> &opt); + +template <typename T, typename U> +constexpr bool operator==(const Optional<T> &opt, const U &value); +template <typename T, typename U> +constexpr bool operator==(const T &value, const Optional<U> &opt); +template <typename T, typename U> +constexpr bool operator!=(const Optional<T> &opt, const U &value); +template <typename T, typename U> +constexpr bool operator!=(const T &value, const Optional<U> &opt); + +} // namespace base +)"; + +constexpr const char StatusDefsHeader[] = + R"cc( +#ifndef STATUS_H_ +#define STATUS_H_ + +#include "absl_type_traits.h" +#include "std_initializer_list.h" +#include "std_string.h" +#include "std_type_traits.h" +#include "std_utility.h" + +namespace absl { +struct SourceLocation { + static constexpr SourceLocation current(); + static constexpr SourceLocation + DoNotInvokeDirectlyNoSeriouslyDont(int line, const char *file_name); +}; +} // namespace absl +namespace absl { +enum class StatusCode : int { + kOk, + kCancelled, + kUnknown, + kInvalidArgument, + kDeadlineExceeded, + kNotFound, + kAlreadyExists, + kPermissionDenied, + kResourceExhausted, + kFailedPrecondition, + kAborted, + kOutOfRange, + kUnimplemented, + kInternal, + kUnavailable, + kDataLoss, + kUnauthenticated, +}; +} // namespace absl + +namespace absl { +enum class StatusToStringMode : int { + kWithNoExtraData = 0, + kWithPayload = 1 << 0, + kWithSourceLocation = 1 << 1, + kWithEverything = ~kWithNoExtraData, + kDefault = kWithPayload, +}; +class Status { +public: + Status(); + template <typename Enum> Status(Enum code, std::string_view msg); + Status(absl::StatusCode code, std::string_view msg, + absl::SourceLocation loc = SourceLocation::current()); + Status(const Status &base_status, absl::SourceLocation loc); + Status(Status &&base_status, absl::SourceLocation loc); + ~Status() {} + + Status(const Status &); + Status &operator=(const Status &x); + + Status(Status &&) noexcept; + Status &operator=(Status &&); + + friend bool operator==(const Status &, const Status &); + friend bool operator!=(const Status &, const Status &); + + bool ok() const { return true; } + void CheckSuccess() const; + void IgnoreError() const; + int error_code() const; + absl::Status ToCanonical() const; + std::string + ToString(StatusToStringMode m = StatusToStringMode::kDefault) const; + void Update(const Status &new_status); + void Update(Status &&new_status); +}; + +bool operator==(const Status &lhs, const Status &rhs); +bool operator!=(const Status &lhs, const Status &rhs); + +Status OkStatus(); +Status InvalidArgumentError(char *); + +#endif // STATUS_H +)cc"; + +constexpr const char StatusOrDefsHeader[] = R"cc( +#ifndef STATUSOR_H_ +#define STATUSOR_H_ +#include "absl_type_traits.h" +#include "status_defs.h" +#include "std_initializer_list.h" +#include "std_type_traits.h" +#include "std_utility.h" + +template <typename T> struct StatusOr; + +namespace internal_statusor { + +template <typename T, typename U, typename = void> +struct HasConversionOperatorToStatusOr : std::false_type {}; + +template <typename T, typename U> +void test(char (*)[sizeof(std::declval<U>().operator absl::StatusOr<T>())]); + +template <typename T, typename U> +struct HasConversionOperatorToStatusOr<T, U, decltype(test<T, U>(0))> + : std::true_type {}; + +template <typename T, typename U> +using IsConstructibleOrConvertibleFromStatusOr = + absl::disjunction<std::is_constructible<T, StatusOr<U> &>, + std::is_constructible<T, const StatusOr<U> &>, + std::is_constructible<T, StatusOr<U> &&>, + std::is_constructible<T, const StatusOr<U> &&>, + std::is_convertible<StatusOr<U> &, T>, + std::is_convertible<const StatusOr<U> &, T>, + std::is_convertible<StatusOr<U> &&, T>, + std::is_convertible<const StatusOr<U> &&, T>>; + +template <typename T, typename U> +using IsConstructibleOrConvertibleOrAssignableFromStatusOr = + absl::disjunction<IsConstructibleOrConvertibleFromStatusOr<T, U>, + std::is_assignable<T &, StatusOr<U> &>, + std::is_assignable<T &, const StatusOr<U> &>, + std::is_assignable<T &, StatusOr<U> &&>, + std::is_assignable<T &, const StatusOr<U> &&>>; + +template <typename T, typename U> +struct IsDirectInitializationAmbiguous + : public absl::conditional_t< + std::is_same<absl::remove_cv_t<absl::remove_reference_t<U>>, + U>::value, + std::false_type, + IsDirectInitializationAmbiguous< + T, absl::remove_cv_t<absl::remove_reference_t<U>>>> {}; + +template <typename T, typename V> +struct IsDirectInitializationAmbiguous<T, absl::StatusOr<V>> + : public IsConstructibleOrConvertibleFromStatusOr<T, V> {}; + +template <typename T, typename U> +using IsDirectInitializationValid = absl::disjunction< + // Short circuits if T is basically U. + std::is_same<T, absl::remove_cv_t<absl::remove_reference_t<U>>>, + absl::negation<absl::disjunction< + std::is_same<absl::StatusOr<T>, + absl::remove_cv_t<absl::remove_reference_t<U>>>, + std::is_same<absl::Status, + absl::remove_cv_t<absl::remove_reference_t<U>>>, + std::is_same<absl::in_place_t, + absl::remove_cv_t<absl::remove_reference_t<U>>>, + IsDirectInitializationAmbiguous<T, U>>>>; + +template <typename T, typename U> +struct IsForwardingAssignmentAmbiguous + : public absl::conditional_t< + std::is_same<absl::remove_cv_t<absl::remove_reference_t<U>>, + U>::value, + std::false_type, + IsForwardingAssignmentAmbiguous< + T, absl::remove_cv_t<absl::remove_reference_t<U>>>> {}; + +template <typename T, typename U> +struct IsForwardingAssignmentAmbiguous<T, absl::StatusOr<U>> + : public IsConstructibleOrConvertibleOrAssignableFromStatusOr<T, U> {}; + +template <typename T, typename U> +using IsForwardingAssignmentValid = absl::disjunction< + // Short circuits if T is basically U. + std::is_same<T, absl::remove_cv_t<absl::remove_reference_t<U>>>, + absl::negation<absl::disjunction< + std::is_same<absl::StatusOr<T>, + absl::remove_cv_t<absl::remove_reference_t<U>>>, + std::is_same<absl::Status, + absl::remove_cv_t<absl::remove_reference_t<U>>>, + std::is_same<absl::in_place_t, + absl::remove_cv_t<absl::remove_reference_t<U>>>, + IsForwardingAssignmentAmbiguous<T, U>>>>; + +template <typename T, typename U> +using IsForwardingAssignmentValid = absl::disjunction< + // Short circuits if T is basically U. + std::is_same<T, absl::remove_cv_t<absl::remove_reference_t<U>>>, + absl::negation<absl::disjunction< + std::is_same<absl::StatusOr<T>, + absl::remove_cv_t<absl::remove_reference_t<U>>>, + std::is_same<absl::Status, + absl::remove_cv_t<absl::remove_reference_t<U>>>, + std::is_same<absl::in_place_t, + absl::remove_cv_t<absl::remove_reference_t<U>>>, + IsForwardingAssignmentAmbiguous<T, U>>>>; + +template <typename T> struct OperatorBase { + const T &value() const &; + T &value() &; + const T &&value() const &&; + T &&value() &&; + + const T &operator*() const &; + T &operator*() &; + const T &&operator*() const &&; + T &&operator*() &&; + + // To test that analyses are okay if there is a use of operator* + // within this base class. + const T *operator->() const { return __builtin_addressof(**this); } + T *operator->() { return __builtin_addressof(**this); } +}; + +} // namespace internal_statusor + +template <typename T> +struct StatusOr : private internal_statusor::OperatorBase<T> { + explicit StatusOr(); + + StatusOr(const StatusOr &) = default; + StatusOr &operator=(const StatusOr &) = default; + + StatusOr(StatusOr &&) = default; + StatusOr &operator=(StatusOr &&) = default; + + template < + typename U, + absl::enable_if_t< + absl::conjunction< + absl::negation<std::is_same<T, U>>, + std::is_constructible<T, const U &>, + std::is_convertible<const U &, T>, + absl::negation< + internal_statusor::IsConstructibleOrConvertibleFromStatusOr< + T, U>>>::value, + int> = 0> + StatusOr(const StatusOr<U> &); + + template < + typename U, + absl::enable_if_t< + absl::conjunction< + absl::negation<std::is_same<T, U>>, + std::is_constructible<T, const U &>, + absl::negation<std::is_convertible<const U &, T>>, + absl::negation< + internal_statusor::IsConstructibleOrConvertibleFromStatusOr< + T, U>>>::value, + int> = 0> + explicit StatusOr(const StatusOr<U> &); + + template < + typename U, + absl::enable_if_t< + absl::conjunction< + absl::negation<std::is_same<T, U>>, + std::is_constructible<T, U &&>, std::is_convertible<U &&, T>, + absl::negation< + internal_statusor::IsConstructibleOrConvertibleFromStatusOr< + T, U>>>::value, + int> = 0> + StatusOr(StatusOr<U> &&); + + template < + typename U, + absl::enable_if_t< + absl::conjunction< + absl::negation<std::is_same<T, U>>, + std::is_constructible<T, U &&>, + absl::negation<std::is_convertible<U &&, T>>, + absl::negation< + internal_statusor::IsConstructibleOrConvertibleFromStatusOr< + T, U>>>::value, + int> = 0> + explicit StatusOr(StatusOr<U> &&); + + template < + typename U, + absl::enable_if_t< + absl::conjunction< + absl::negation<std::is_same<T, U>>, + std::is_constructible<T, const U &>, + std::is_assignable<T, const U &>, + absl::negation< + internal_statusor:: + IsConstructibleOrConvertibleOrAssignableFromStatusOr< + T, U>>>::value, + int> = 0> + StatusOr &operator=(const StatusOr<U> &); + + template < + typename U, + absl::enable_if_t< + absl::conjunction< + absl::negation<std::is_same<T, U>>, + std::is_constructible<T, U &&>, std::is_assignable<T, U &&>, + absl::negation< + internal_statusor:: + IsConstructibleOrConvertibleOrAssignableFromStatusOr< + T, U>>>::value, + int> = 0> + StatusOr &operator=(StatusOr<U> &&); + + template < + typename U = absl::Status, + absl::enable_if_t< + absl::conjunction< + std::is_convertible<U &&, absl::Status>, + std::is_constructible<absl::Status, U &&>, + absl::negation<std::is_same<absl::decay_t<U>, absl::StatusOr<T>>>, + absl::negation<std::is_same<absl::decay_t<U>, T>>, + absl::negation<std::is_same<absl::decay_t<U>, absl::in_place_t>>, + absl::negation<internal_statusor::HasConversionOperatorToStatusOr< + T, U &&>>>::value, + int> = 0> + StatusOr(U &&); + + template < + typename U = absl::Status, + absl::enable_if_t< + absl::conjunction< + absl::negation<std::is_convertible<U &&, absl::Status>>, + std::is_constructible<absl::Status, U &&>, + absl::negation<std::is_same<absl::decay_t<U>, absl::StatusOr<T>>>, + absl::negation<std::is_same<absl::decay_t<U>, T>>, + absl::negation<std::is_same<absl::decay_t<U>, absl::in_place_t>>, + absl::negation<internal_statusor::HasConversionOperatorToStatusOr< + T, U &&>>>::value, + int> = 0> + explicit StatusOr(U &&); + + template < + typename U = absl::Status, + absl::enable_if_t< + absl::conjunction< + std::is_convertible<U &&, absl::Status>, + std::is_constructible<absl::Status, U &&>, + absl::negation<std::is_same<absl::decay_t<U>, absl::StatusOr<T>>>, + absl::negation<std::is_same<absl::decay_t<U>, T>>, + absl::negation<std::is_same<absl::decay_t<U>, absl::in_place_t>>, + absl::negation<internal_statusor::HasConversionOperatorToStatusOr< + T, U &&>>>::value, + int> = 0> + StatusOr &operator=(U &&); + + template < + typename U = T, + typename = typename std::enable_if<absl::conjunction< + std::is_constructible<T, U &&>, std::is_assignable<T &, U &&>, + absl::disjunction< + std::is_same<absl::remove_cv_t<absl::remove_reference_t<U>>, T>, + absl::conjunction< + absl::negation<std::is_convertible<U &&, absl::Status>>, + absl::negation< + internal_statusor::HasConversionOperatorToStatusOr< + T, U &&>>>>, + internal_statusor::IsForwardingAssignmentValid<T, U &&>>::value>:: + type> + StatusOr &operator=(U &&); + + template <typename... Args> explicit StatusOr(absl::in_place_t, Args &&...); + + template <typename U, typename... Args> + explicit StatusOr(absl::in_place_t, std::initializer_list<U>, Args &&...); + + template < + typename U = T, + absl::enable_if_t< + absl::conjunction< + internal_statusor::IsDirectInitializationValid<T, U &&>, + std::is_constructible<T, U &&>, std::is_convertible<U &&, T>, + absl::disjunction< + std::is_same<absl::remove_cv_t<absl::remove_reference_t<U>>, + T>, + absl::conjunction< + absl::negation<std::is_convertible<U &&, absl::Status>>, + absl::negation< + internal_statusor::HasConversionOperatorToStatusOr< + T, U &&>>>>>::value, + int> = 0> + StatusOr(U &&); + + template < + typename U = T, + absl::enable_if_t< + absl::conjunction< + internal_statusor::IsDirectInitializationValid<T, U &&>, + absl::disjunction< + std::is_same<absl::remove_cv_t<absl::remove_reference_t<U>>, + T>, + absl::conjunction< + absl::negation<std::is_constructible<absl::Status, U &&>>, + absl::negation< + internal_statusor::HasConversionOperatorToStatusOr< + T, U &&>>>>, + std::is_constructible<T, U &&>, + absl::negation<std::is_convertible<U &&, T>>>::value, + int> = 0> + explicit StatusOr(U &&); + + bool ok() const; + + const Status &status() const & { return status_; } + Status status() &&; + + using StatusOr::OperatorBase::value; + + const T &ValueOrDie() const &; + T &ValueOrDie() &; + const T &&ValueOrDie() const &&; + T &&ValueOrDie() &&; + + using StatusOr::OperatorBase::operator*; + using StatusOr::OperatorBase::operator->; + + template <typename U> T value_or(U &&default_value) const &; + template <typename U> T value_or(U &&default_value) &&; + + template <typename... Args> T &emplace(Args &&...args); + + template < + typename U, typename... Args, + absl::enable_if_t<std::is_constructible<T, std::initializer_list<U> &, + Args &&...>::value, + int> = 0> + T &emplace(std::initializer_list<U> ilist, Args &&...args); + +private: + absl::Status status_; +}; + +template <typename T> +bool operator==(const StatusOr<T> &lhs, const StatusOr<T> &rhs); + +template <typename T> +bool operator!=(const StatusOr<T> &lhs, const StatusOr<T> &rhs); + +} // namespace absl + +#endif // STATUSOR_H_ +)cc"; + +static constexpr char StdVectorHeader[] = R"cc( +#ifndef STD_VECTOR_H +#define STD_VECTOR_H +namespace std { +template <class T> struct allocator { + typedef size_t size_type; + typedef ptrdiff_t difference_type; + typedef T *pointer; + typedef const T *const_pointer; + typedef T value_type; + + T *allocate(size_t n); +}; + +template <class Alloc> struct allocator_traits { + typedef Alloc allocator_type; + typedef typename allocator_type::value_type value_type; + typedef typename allocator_type::pointer pointer; + typedef typename allocator_type::const_pointer const_pointer; + typedef typename allocator_type::difference_type difference_type; + typedef typename allocator_type::size_type size_type; +}; + +template <typename T, class Allocator = allocator<T>> class vector { +public: + using value_type = T; + using size_type = typename allocator_traits<Allocator>::size_type; + + // Constructors. + vector() {} + vector(size_type, const Allocator & = Allocator()) {} + vector(initializer_list<T> initializer_list, + const Allocator & = Allocator()) {} + vector(const vector &vector) {} + ~vector(); + + // Modifiers. + void push_back(const T &value); + void push_back(T &&value); + template <typename... Args> T &emplace_back(Args &&...args); + + // Iterators + class InputIterator { + public: + InputIterator(const InputIterator &); + ~InputIterator(); + InputIterator &operator=(const InputIterator &); + InputIterator &operator++(); + T &operator*() const; + bool operator!=(const InputIterator &) const; + bool operator==(const InputIterator &) const; + }; + typedef InputIterator iterator; + typedef const InputIterator const_iterator; + iterator begin() noexcept; + const_iterator begin() const noexcept; + const_iterator cbegin() const noexcept; + iterator end() noexcept; + const_iterator end() const noexcept; + const_iterator cend() const noexcept; + T *data() noexcept; + const T *data() const noexcept; + T &operator[](int n); + const T &operator[](int n) const; + T &at(int n); + const T &at(int n) const; + size_t size() const; +}; +} // namespace std +#endif // STD_VECTOR_H +)cc"; + +static constexpr char StdPairHeader[] = R"cc( +#ifndef STD_PAIR_H +#define STD_PAIR_H +namespace std { +template <class T1, class T2> struct pair { + T1 first; + T2 second; + + typedef T1 first_type; + typedef T2 second_type; + + constexpr pair(); + + template <class U1, class U2> pair(pair<U1, U2> &&p); + + template <class U1, class U2> pair(U1 &&x, U2 &&y); +}; + +template <class T1, class T2> pair<T1, T2> make_pair(T1 &&t1, T2 &&t2); +} // namespace std +#endif // STD_PAIR_H +)cc"; + +constexpr const char AbslLogHeader[] = R"cc( +#ifndef ABSL_LOG_H +#define ABSL_LOG_H + +#include "std_pair.h" + +namespace absl { + +#define ABSL_PREDICT_FALSE(x) (__builtin_expect(false || (x), false)) +#define ABSL_PREDICT_TRUE(x) (__builtin_expect(false || (x), true)) + +namespace log_internal { +class LogMessage { +public: + LogMessage(); + LogMessage &stream(); + LogMessage &InternalStream(); + LogMessage &WithVerbosity(int verboselevel); + template <typename T> LogMessage &operator<<(const T &); +}; +class LogMessageFatal : public LogMessage { +public: + LogMessageFatal(); + ~LogMessageFatal() __attribute__((noreturn)); +}; +class LogMessageQuietlyFatal : public LogMessage { +public: + LogMessageQuietlyFatal(); + ~LogMessageQuietlyFatal() __attribute__((noreturn)); +}; +class Voidify final { +public: + // This has to be an operator with a precedence lower than << but higher + // than + // ?: + template <typename T> void operator&&(const T &) const && {} +}; +} // namespace log_internal +} // namespace absl + +#ifndef NULL +#define NULL __null +#endif +extern "C" void abort() {} +#define ABSL_LOG_INTERNAL_LOG_INFO ::absl::log_internal::LogMessage() +#define ABSL_LOG_INTERNAL_LOG_WARNING ::absl::log_internal::LogMessage() +#define ABSL_LOG_INTERNAL_LOG_ERROR ::absl::log_internal::LogMessage() +#define ABSL_LOG_INTERNAL_LOG_FATAL ::absl::log_internal::LogMessageFatal() +#define ABSL_LOG_INTERNAL_LOG_QFATAL \ + ::absl::log_internal::LogMessageQuietlyFatal() +#define LOG(severity) ABSL_LOG_INTERNAL_LOG_##severity.InternalStream() + +#define PREDICT_FALSE(x) (__builtin_expect(x, 0)) +#define ABSL_LOG_INTERNAL_STRIP_STRING_LITERAL(lit) lit + +#define ABSL_LOG_INTERNAL_STATELESS_CONDITION(condition) \ + switch (0) \ + case 0: \ + default: \ + !(condition) ? (void)0 : ::absl::log_internal::Voidify() && + +#define ABSL_LOG_INTERNAL_CONDITION_INFO(type, condition) \ + ABSL_LOG_INTERNAL_##type##_CONDITION(condition) + +#define ABSL_LOG_INTERNAL_CONDITION_FATAL(type, condition) \ + ABSL_LOG_INTERNAL_##type##_CONDITION(condition) + +#define ABSL_LOG_INTERNAL_CONDITION_QFATAL(type, condition) \ + ABSL_LOG_INTERNAL_##type##_CONDITION(condition) + +#define ABSL_CHECK_IMPL(condition, condition_text) \ + ABSL_LOG_INTERNAL_CONDITION_FATAL(STATELESS, \ + ABSL_PREDICT_FALSE(!(condition))) \ + ABSL_LOG_INTERNAL_CHECK(condition_text).InternalStream() + +#define ABSL_QCHECK_IMPL(condition, condition_text) \ + ABSL_LOG_INTERNAL_CONDITION_QFATAL(STATELESS, \ + ABSL_PREDICT_FALSE(!(condition))) \ + ABSL_LOG_INTERNAL_QCHECK(condition_text).InternalStream() + +#define CHECK(condition) ABSL_CHECK_IMPL((condition), #condition) +#define DCHECK(condition) CHECK(condition) +#define QCHECK(condition) ABSL_QCHECK_IMPL((condition), #condition) + +#define ABSL_LOG_INTERNAL_MAX_LOG_VERBOSITY_CHECK(x) + +namespace absl { + +template <typename T> class StatusOr; +class Status; + +namespace status_internal { +std::string *MakeCheckFailString(const absl::Status *status, + const char *prefix); +} // namespace status_internal + +namespace log_internal { +template <class T> const T &GetReferenceableValue(const T &t); +char GetReferenceableValue(char t); +unsigned char GetReferenceableValue(unsigned char t); +signed char GetReferenceableValue(signed char t); +short GetReferenceableValue(short t); +unsigned short GetReferenceableValue(unsigned short t); +int GetReferenceableValue(int t); +unsigned int GetReferenceableValue(unsigned int t); +long GetReferenceableValue(long t); +unsigned long GetReferenceableValue(unsigned long t); +long long GetReferenceableValue(long long t); +unsigned long long GetReferenceableValue(unsigned long long t); +const absl::Status *AsStatus(const absl::Status &s); +template <typename T> const absl::Status *AsStatus(const absl::StatusOr<T> &s); +} // namespace log_internal +} // namespace absl +// TODO(tkd): this still doesn't allow operator<<, unlike the real CHECK_ +// macros. +#define ABSL_LOG_INTERNAL_CHECK_OP(name, op, val1, val2) \ + while (char *_result = ::absl::log_internal::name##Impl( \ + ::absl::log_internal::GetReferenceableValue(val1), \ + ::absl::log_internal::GetReferenceableValue(val2), \ + #val1 " " #op " " #val2)) \ + (void)0 +#define ABSL_LOG_INTERNAL_QCHECK_OP(name, op, val1, val2) \ + while (char *_result = ::absl::log_internal::name##Impl( \ + ::absl::log_internal::GetReferenceableValue(val1), \ + ::absl::log_internal::GetReferenceableValue(val2), \ + #val1 " " #op " " #val2)) \ + (void)0 +namespace absl { +namespace log_internal { +template <class T1, class T2> +char *Check_NEImpl(const T1 &v1, const T2 &v2, const char *names); +template <class T1, class T2> +char *Check_EQImpl(const T1 &v1, const T2 &v2, const char *names); +template <class T1, class T2> +char *Check_LTImpl(const T1 &v1, const T2 &v2, const char *names); + +#define CHECK_EQ(a, b) ABSL_LOG_INTERNAL_CHECK_OP(Check_EQ, ==, a, b) +#define CHECK_NE(a, b) ABSL_LOG_INTERNAL_CHECK_OP(Check_NE, !=, a, b) +#define CHECK_LT(a, b) ABSL_LOG_INTERNAL_CHECK_OP(Check_EQ, <, a, b) + +#define QCHECK_EQ(a, b) ABSL_LOG_INTERNAL_QCHECK_OP(Check_EQ, ==, a, b) +#define QCHECK_NE(a, b) ABSL_LOG_INTERNAL_QCHECK_OP(Check_NE, !=, a, b) +} // namespace log_internal +} // namespace absl + +#define CHECK_NOTNULL(x) CHECK((x) != nullptr) + +#define ABSL_LOG_INTERNAL_CHECK(failure_message) \ + ::absl::log_internal::LogMessageFatal() +#define ABSL_LOG_INTERNAL_QCHECK(failure_message) \ + ::absl::log_internal::LogMessageQuietlyFatal() +#define ABSL_LOG_INTERNAL_CHECK_OK(val) \ + for (::std::pair<const ::absl::Status *, ::std::string *> \ + absl_log_internal_check_ok_goo; \ + absl_log_internal_check_ok_goo.first = \ + ::absl::log_internal::AsStatus(val), \ + absl_log_internal_check_ok_goo.second = \ + ABSL_PREDICT_TRUE(absl_log_internal_check_ok_goo.first->ok()) \ + ? nullptr \ + : ::absl::status_internal::MakeCheckFailString( \ + absl_log_internal_check_ok_goo.first, \ + ABSL_LOG_INTERNAL_STRIP_STRING_LITERAL(#val " is OK")), \ + !ABSL_PREDICT_TRUE(absl_log_internal_check_ok_goo.first->ok());) \ + ABSL_LOG_INTERNAL_CHECK(*absl_log_internal_check_ok_goo.second) \ + .InternalStream() +#define ABSL_LOG_INTERNAL_QCHECK_OK(val) \ + for (::std::pair<const ::absl::Status *, ::std::string *> \ + absl_log_internal_check_ok_goo; \ + absl_log_internal_check_ok_goo.first = \ + ::absl::log_internal::AsStatus(val), \ + absl_log_internal_check_ok_goo.second = \ + ABSL_PREDICT_TRUE(absl_log_internal_check_ok_goo.first->ok()) \ + ? nullptr \ + : ::absl::status_internal::MakeCheckFailString( \ + absl_log_internal_check_ok_goo.first, \ + ABSL_LOG_INTERNAL_STRIP_STRING_LITERAL(#val " is OK")), \ + !ABSL_PREDICT_TRUE(absl_log_internal_check_ok_goo.first->ok());) \ + ABSL_LOG_INTERNAL_QCHECK(*absl_log_internal_check_ok_goo.second) \ + .InternalStream() + +#define CHECK_OK(val) ABSL_LOG_INTERNAL_CHECK_OK(val) +#define DCHECK_OK(val) ABSL_LOG_INTERNAL_CHECK_OK(val) +#define QCHECK_OK(val) ABSL_LOG_INTERNAL_QCHECK_OK(val) + +#endif // ABSL_LOG_H +)cc"; + +constexpr const char TestingDefsHeader[] = R"cc( +#pragma clang system_header + +#ifndef TESTING_DEFS_H +#define TESTING_DEFS_H + +#include "absl_type_traits.h" +#include "std_initializer_list.h" +#include "std_string.h" +#include "std_type_traits.h" +#include "std_utility.h" + +namespace testing { +struct AssertionResult { + template <typename T> + explicit AssertionResult(const T &res, bool enable_if = true) {} + ~AssertionResult(); + operator bool() const; + template <typename T> AssertionResult &operator<<(const T &value); + const char *failure_message() const; +}; + +class TestPartResult { +public: + enum Type { kSuccess, kNonFatalFailure, kFatalFailure, kSkip }; +}; + +class Test { +public: + virtual ~Test() = default; + +protected: + virtual void SetUp() {} +}; + +class Message { +public: + template <typename T> Message &operator<<(const T &val); +}; + +namespace internal { +class AssertHelper { +public: + AssertHelper(TestPartResult::Type type, const char *file, int line, + const char *message); + void operator=(const Message &message) const; +}; + +class EqHelper { +public: + template <typename T1, typename T2> + static AssertionResult Compare(const char *lhx, const char *rhx, + const T1 &lhs, const T2 &rhs); +}; + +#define GTEST_IMPL_CMP_HELPER_(op_name) \ + template <typename T1, typename T2> \ + AssertionResult CmpHelper##op_name(const char *expr1, const char *expr2, \ + const T1 &val1, const T2 &val2); + +GTEST_IMPL_CMP_HELPER_(NE) +GTEST_IMPL_CMP_HELPER_(LE) +GTEST_IMPL_CMP_HELPER_(LT) +GTEST_IMPL_CMP_HELPER_(GE) +GTEST_IMPL_CMP_HELPER_(GT) + +#undef GTEST_IMPL_CMP_HELPER_ + +std::string GetBoolAssertionFailureMessage( + const AssertionResult &assertion_result, const char *expression_text, + const char *actual_predicate_value, const char *expected_predicate_value); + +template <typename M> class PredicateFormatterFromMatcher { +public: + template <typename T> + AssertionResult operator()(const char *value_text, const T &x) const; +}; + +template <typename M> +inline PredicateFormatterFromMatcher<M> +MakePredicateFormatterFromMatcher(M matcher) { + return PredicateFormatterFromMatcher<M>(); +} +} // namespace internal + +namespace status { +namespace internal_status { +class IsOkMatcher {}; + +class StatusIsMatcher {}; + +class CanonicalStatusIsMatcher {}; + +template <typename M> class IsOkAndHoldsMatcher {}; + +} // namespace internal_status + +internal_status::IsOkMatcher IsOk(); + +template <typename StatusCodeMatcher> +internal_status::StatusIsMatcher StatusIs(StatusCodeMatcher &&code_matcher); + +template <typename StatusCodeMatcher> +internal_status::CanonicalStatusIsMatcher +CanonicalStatusIs(StatusCodeMatcher &&code_matcher); + +template <typename InnerMatcher> +internal_status::IsOkAndHoldsMatcher<InnerMatcher> IsOkAndHolds(InnerMatcher m); +} // namespace status + +class IsTrueMatcher {}; +IsTrueMatcher IsTrue(); + +class IsFalseMatcher {}; +IsFalseMatcher IsFalse(); + +} // namespace testing + +namespace absl_testing { +namespace status_internal { +class IsOkMatcher {}; +template <typename M> class IsOkAndHoldsMatcher {}; +class StatusIsMatcher {}; +class CanonicalStatusIsMatcher {}; +} // namespace status_internal +status_internal::IsOkMatcher IsOk(); +template <typename InnerMatcher> +status_internal::IsOkAndHoldsMatcher<InnerMatcher> IsOkAndHolds(InnerMatcher m); +template <typename StatusCodeMatcher> +status_internal::StatusIsMatcher StatusIs(StatusCodeMatcher &&code_matcher); + +template <typename StatusCodeMatcher> +status_internal::CanonicalStatusIsMatcher +CanonicalStatusIs(StatusCodeMatcher &&code_matcher); +} // namespace absl_testing + +using testing::AssertionResult; +#define EXPECT_TRUE(x) \ + switch (0) \ + case 0: \ + default: \ + if (const AssertionResult gtest_ar_ = AssertionResult(x)) { \ + } else /* NOLINT */ \ + ::testing::Message() +#define EXPECT_FALSE(x) EXPECT_TRUE(!(x)) + +#define GTEST_AMBIGUOUS_ELSE_BLOCKER_ \ + switch (0) \ + case 0: \ + default: + +#define GTEST_ASSERT_(expression, on_failure) \ + GTEST_AMBIGUOUS_ELSE_BLOCKER_ \ + if (const ::testing::AssertionResult gtest_ar = (expression)) \ + ; \ + else \ + on_failure(gtest_ar.failure_message()) +#define GTEST_PRED_FORMAT1_(pred_format, v1, on_failure) \ + GTEST_ASSERT_(pred_format(#v1, v1), on_failure) +#define GTEST_PRED_FORMAT2_(pred_format, v1, v2, on_failure) \ + GTEST_ASSERT_(pred_format(#v1, #v2, v1, v2), on_failure) +#define GTEST_MESSAGE_AT_(file, line, message, result_type) \ + ::testing::internal::AssertHelper(result_type, file, line, message) = \ + ::testing::Message() +#define GTEST_MESSAGE_(message, result_type) \ + GTEST_MESSAGE_AT_(__FILE__, __LINE__, message, result_type) +#define GTEST_FATAL_FAILURE_(message) \ + return GTEST_MESSAGE_(message, ::testing::TestPartResult::kFatalFailure) +#define GTEST_NONFATAL_FAILURE_(message) \ + GTEST_MESSAGE_(message, ::testing::TestPartResult::kNonFatalFailure) + +#define ASSERT_PRED_FORMAT1(pred_format, v1) \ + GTEST_PRED_FORMAT1_(pred_format, v1, GTEST_FATAL_FAILURE_) +#define ASSERT_PRED_FORMAT2(pred_format, v1, v2) \ + GTEST_PRED_FORMAT2_(pred_format, v1, v2, GTEST_FATAL_FAILURE_) + +#define ASSERT_THAT(value, matcher) \ + ASSERT_PRED_FORMAT1( \ + ::testing::internal::MakePredicateFormatterFromMatcher(matcher), value) +#define ASSERT_OK(x) ASSERT_THAT(x, ::testing::status::IsOk()) + +#define EXPECT_PRED_FORMAT1(pred_format, v1) \ + GTEST_PRED_FORMAT1_(pred_format, v1, GTEST_NONFATAL_FAILURE_) +#define EXPECT_PRED_FORMAT2(pred_format, v1, v2) \ + GTEST_PRED_FORMAT2_(pred_format, v1, v2, GTEST_NONFATAL_FAILURE_) +#define EXPECT_THAT(value, matcher) \ + EXPECT_PRED_FORMAT1( \ + ::testing::internal::MakePredicateFormatterFromMatcher(matcher), value) +#define EXPECT_OK(expression) EXPECT_THAT(expression, ::testing::status::IsOk()) + +#define GTEST_TEST_BOOLEAN_(expression, text, actual, expected, fail) \ + GTEST_AMBIGUOUS_ELSE_BLOCKER_ \ + if (const ::testing::AssertionResult gtest_ar_ = \ + ::testing::AssertionResult(expression)) \ + ; \ + else \ + fail(::testing::internal::GetBoolAssertionFailureMessage( \ + gtest_ar_, text, #actual, #expected) \ + .c_str()) +#define GTEST_ASSERT_TRUE(condition) \ + GTEST_TEST_BOOLEAN_(condition, #condition, false, true, GTEST_FATAL_FAILURE_) +#define GTEST_ASSERT_FALSE(condition) \ + GTEST_TEST_BOOLEAN_(!(condition), #condition, true, false, \ + GTEST_FATAL_FAILURE_) +#define ASSERT_TRUE(condition) GTEST_ASSERT_TRUE(condition) +#define ASSERT_FALSE(condition) GTEST_ASSERT_FALSE(condition) + +#define EXPECT_EQ(x, y) \ + EXPECT_PRED_FORMAT2(::testing::internal::EqHelper::Compare, x, y) +#define EXPECT_NE(x, y) \ + EXPECT_PRED_FORMAT2(::testing::internal::CmpHelperNE, x, y) +#define EXPECT_LT(x, y) \ + EXPECT_PRED_FORMAT2(::testing::internal::CmpHelperLT, x, y) +#define EXPECT_GT(x, y) \ + EXPECT_PRED_FORMAT2(::testing::internal::CmpHelperGT, x, y) +#define EXPECT_LE(x, y) \ + EXPECT_PRED_FORMAT2(::testing::internal::CmpHelperLE, x, y) +#define EXPECT_GE(x, y) \ + EXPECT_PRED_FORMAT2(::testing::internal::CmpHelperGE, x, y) + +#define ASSERT_EQ(x, y) \ + ASSERT_PRED_FORMAT2(testing::internal::EqHelper::Compare, x, y) +#define ASSERT_NE(x, y) \ + ASSERT_PRED_FORMAT2(testing::internal::CmpHelperNE, x, y) +#define ASSERT_LT(x, y) \ + ASSERT_PRED_FORMAT2(testing::internal::CmpHelperLT, x, y) +#define ASSERT_GT(x, y) \ + ASSERT_PRED_FORMAT2(testing::internal::CmpHelperGT, x, y) +#define ASSERT_LE(x, y) \ + ASSERT_PRED_FORMAT2(testing::internal::CmpHelperLE, x, y) +#define ASSERT_GE(x, y) \ + ASSERT_PRED_FORMAT2(testing::internal::CmpHelperGE, x, y) + +#endif // TESTING_DEFS_H +)cc"; + +std::vector<std::pair<std::string, std::string>> getMockHeaders() { + std::vector<std::pair<std::string, std::string>> Headers; + Headers.emplace_back("cstddef.h", CStdDefHeader); + Headers.emplace_back("std_initializer_list.h", StdInitializerListHeader); + Headers.emplace_back("std_string.h", StdStringHeader); + Headers.emplace_back("std_type_traits.h", StdTypeTraitsHeader); + Headers.emplace_back("std_utility.h", StdUtilityHeader); + Headers.emplace_back("std_optional.h", StdOptionalHeader); + Headers.emplace_back("absl_type_traits.h", AbslTypeTraitsHeader); + Headers.emplace_back("absl_optional.h", AbslOptionalHeader); + Headers.emplace_back("base_optional.h", BaseOptionalHeader); + Headers.emplace_back("std_vector.h", StdVectorHeader); + Headers.emplace_back("std_pair.h", StdPairHeader); + Headers.emplace_back("status_defs.h", StatusDefsHeader); + Headers.emplace_back("statusor_defs.h", StatusOrDefsHeader); + Headers.emplace_back("absl_log.h", AbslLogHeader); + Headers.emplace_back("testing_defs.h", TestingDefsHeader); + return Headers; +} + +} // namespace test +} // namespace dataflow +} // namespace clang
\ No newline at end of file diff --git a/clang/unittests/Analysis/FlowSensitive/MockHeaders.h b/clang/unittests/Analysis/FlowSensitive/MockHeaders.h new file mode 100644 index 0000000..c0b544f --- /dev/null +++ b/clang/unittests/Analysis/FlowSensitive/MockHeaders.h @@ -0,0 +1,30 @@ +//===--- MockHeaders.h - Mock headers for dataflow analyses -*- C++ -----*-===// +// +// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. +// See https://llvm.org/LICENSE.txt for license information. +// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception +// +//===----------------------------------------------------------------------===// +// +// This file defines mock headers for testing of dataflow analyses. +// +//===----------------------------------------------------------------------===// + +#ifndef LLVM_CLANG_ANALYSIS_FLOW_SENSITIVE_MOCK_HEADERS_H_ +#define LLVM_CLANG_ANALYSIS_FLOW_SENSITIVE_MOCK_HEADERS_H_ + +#include <string> +#include <utility> +#include <vector> + +namespace clang { +namespace dataflow { +namespace test { + +std::vector<std::pair<std::string, std::string>> getMockHeaders(); + +} // namespace test +} // namespace dataflow +} // namespace clang + +#endif // LLVM_CLANG_ANALYSIS_FLOW_SENSITIVE_MOCK_HEADERS_H_ diff --git a/clang/unittests/Analysis/FlowSensitive/UncheckedOptionalAccessModelTest.cpp b/clang/unittests/Analysis/FlowSensitive/UncheckedOptionalAccessModelTest.cpp index 1dd0783..ba509e8 100644 --- a/clang/unittests/Analysis/FlowSensitive/UncheckedOptionalAccessModelTest.cpp +++ b/clang/unittests/Analysis/FlowSensitive/UncheckedOptionalAccessModelTest.cpp @@ -8,6 +8,7 @@ // FIXME: Move this to clang/unittests/Analysis/FlowSensitive/Models. #include "clang/Analysis/FlowSensitive/Models/UncheckedOptionalAccessModel.h" +#include "MockHeaders.h" #include "TestingSupport.h" #include "clang/AST/ASTContext.h" #include "clang/ASTMatchers/ASTMatchers.h" @@ -31,1232 +32,6 @@ using namespace test; using ::testing::ContainerEq; -// FIXME: Move header definitions in separate file(s). -static constexpr char CSDtdDefHeader[] = R"( -#ifndef CSTDDEF_H -#define CSTDDEF_H - -namespace std { - -typedef decltype(sizeof(char)) size_t; - -using nullptr_t = decltype(nullptr); - -} // namespace std - -#endif // CSTDDEF_H -)"; - -static constexpr char StdTypeTraitsHeader[] = R"( -#ifndef STD_TYPE_TRAITS_H -#define STD_TYPE_TRAITS_H - -#include "cstddef.h" - -namespace std { - -template <typename T, T V> -struct integral_constant { - static constexpr T value = V; -}; - -using true_type = integral_constant<bool, true>; -using false_type = integral_constant<bool, false>; - -template< class T > struct remove_reference {typedef T type;}; -template< class T > struct remove_reference<T&> {typedef T type;}; -template< class T > struct remove_reference<T&&> {typedef T type;}; - -template <class T> - using remove_reference_t = typename remove_reference<T>::type; - -template <class T> -struct remove_extent { - typedef T type; -}; - -template <class T> -struct remove_extent<T[]> { - typedef T type; -}; - -template <class T, size_t N> -struct remove_extent<T[N]> { - typedef T type; -}; - -template <class T> -struct is_array : false_type {}; - -template <class T> -struct is_array<T[]> : true_type {}; - -template <class T, size_t N> -struct is_array<T[N]> : true_type {}; - -template <class> -struct is_function : false_type {}; - -template <class Ret, class... Args> -struct is_function<Ret(Args...)> : true_type {}; - -namespace detail { - -template <class T> -struct type_identity { - using type = T; -}; // or use type_identity (since C++20) - -template <class T> -auto try_add_pointer(int) -> type_identity<typename remove_reference<T>::type*>; -template <class T> -auto try_add_pointer(...) -> type_identity<T>; - -} // namespace detail - -template <class T> -struct add_pointer : decltype(detail::try_add_pointer<T>(0)) {}; - -template <bool B, class T, class F> -struct conditional { - typedef T type; -}; - -template <class T, class F> -struct conditional<false, T, F> { - typedef F type; -}; - -template <class T> -struct remove_cv { - typedef T type; -}; -template <class T> -struct remove_cv<const T> { - typedef T type; -}; -template <class T> -struct remove_cv<volatile T> { - typedef T type; -}; -template <class T> -struct remove_cv<const volatile T> { - typedef T type; -}; - -template <class T> -using remove_cv_t = typename remove_cv<T>::type; - -template <class T> -struct decay { - private: - typedef typename remove_reference<T>::type U; - - public: - typedef typename conditional< - is_array<U>::value, typename remove_extent<U>::type*, - typename conditional<is_function<U>::value, typename add_pointer<U>::type, - typename remove_cv<U>::type>::type>::type type; -}; - -template <bool B, class T = void> -struct enable_if {}; - -template <class T> -struct enable_if<true, T> { - typedef T type; -}; - -template <bool B, class T = void> -using enable_if_t = typename enable_if<B, T>::type; - -template <class T, class U> -struct is_same : false_type {}; - -template <class T> -struct is_same<T, T> : true_type {}; - -template <class T> -struct is_void : is_same<void, typename remove_cv<T>::type> {}; - -namespace detail { - -template <class T> -auto try_add_lvalue_reference(int) -> type_identity<T&>; -template <class T> -auto try_add_lvalue_reference(...) -> type_identity<T>; - -template <class T> -auto try_add_rvalue_reference(int) -> type_identity<T&&>; -template <class T> -auto try_add_rvalue_reference(...) -> type_identity<T>; - -} // namespace detail - -template <class T> -struct add_lvalue_reference : decltype(detail::try_add_lvalue_reference<T>(0)) { -}; - -template <class T> -struct add_rvalue_reference : decltype(detail::try_add_rvalue_reference<T>(0)) { -}; - -template <class T> -typename add_rvalue_reference<T>::type declval() noexcept; - -namespace detail { - -template <class T> -auto test_returnable(int) - -> decltype(void(static_cast<T (*)()>(nullptr)), true_type{}); -template <class> -auto test_returnable(...) -> false_type; - -template <class From, class To> -auto test_implicitly_convertible(int) - -> decltype(void(declval<void (&)(To)>()(declval<From>())), true_type{}); -template <class, class> -auto test_implicitly_convertible(...) -> false_type; - -} // namespace detail - -template <class From, class To> -struct is_convertible - : integral_constant<bool, - (decltype(detail::test_returnable<To>(0))::value && - decltype(detail::test_implicitly_convertible<From, To>( - 0))::value) || - (is_void<From>::value && is_void<To>::value)> {}; - -template <class From, class To> -inline constexpr bool is_convertible_v = is_convertible<From, To>::value; - -template <class...> -using void_t = void; - -template <class, class T, class... Args> -struct is_constructible_ : false_type {}; - -template <class T, class... Args> -struct is_constructible_<void_t<decltype(T(declval<Args>()...))>, T, Args...> - : true_type {}; - -template <class T, class... Args> -using is_constructible = is_constructible_<void_t<>, T, Args...>; - -template <class T, class... Args> -inline constexpr bool is_constructible_v = is_constructible<T, Args...>::value; - -template <class _Tp> -struct __uncvref { - typedef typename remove_cv<typename remove_reference<_Tp>::type>::type type; -}; - -template <class _Tp> -using __uncvref_t = typename __uncvref<_Tp>::type; - -template <bool _Val> -using _BoolConstant = integral_constant<bool, _Val>; - -template <class _Tp, class _Up> -using _IsSame = _BoolConstant<__is_same(_Tp, _Up)>; - -template <class _Tp, class _Up> -using _IsNotSame = _BoolConstant<!__is_same(_Tp, _Up)>; - -template <bool> -struct _MetaBase; -template <> -struct _MetaBase<true> { - template <class _Tp, class _Up> - using _SelectImpl = _Tp; - template <template <class...> class _FirstFn, template <class...> class, - class... _Args> - using _SelectApplyImpl = _FirstFn<_Args...>; - template <class _First, class...> - using _FirstImpl = _First; - template <class, class _Second, class...> - using _SecondImpl = _Second; - template <class _Result, class _First, class... _Rest> - using _OrImpl = - typename _MetaBase<_First::value != true && sizeof...(_Rest) != 0>:: - template _OrImpl<_First, _Rest...>; -}; - -template <> -struct _MetaBase<false> { - template <class _Tp, class _Up> - using _SelectImpl = _Up; - template <template <class...> class, template <class...> class _SecondFn, - class... _Args> - using _SelectApplyImpl = _SecondFn<_Args...>; - template <class _Result, class...> - using _OrImpl = _Result; -}; - -template <bool _Cond, class _IfRes, class _ElseRes> -using _If = typename _MetaBase<_Cond>::template _SelectImpl<_IfRes, _ElseRes>; - -template <class... _Rest> -using _Or = typename _MetaBase<sizeof...(_Rest) != - 0>::template _OrImpl<false_type, _Rest...>; - -template <bool _Bp, class _Tp = void> -using __enable_if_t = typename enable_if<_Bp, _Tp>::type; - -template <class...> -using __expand_to_true = true_type; -template <class... _Pred> -__expand_to_true<__enable_if_t<_Pred::value>...> __and_helper(int); -template <class...> -false_type __and_helper(...); -template <class... _Pred> -using _And = decltype(__and_helper<_Pred...>(0)); - -template <class _Pred> -struct _Not : _BoolConstant<!_Pred::value> {}; - -struct __check_tuple_constructor_fail { - static constexpr bool __enable_explicit_default() { return false; } - static constexpr bool __enable_implicit_default() { return false; } - template <class...> - static constexpr bool __enable_explicit() { - return false; - } - template <class...> - static constexpr bool __enable_implicit() { - return false; - } -}; - -template <typename, typename _Tp> -struct __select_2nd { - typedef _Tp type; -}; -template <class _Tp, class _Arg> -typename __select_2nd<decltype((declval<_Tp>() = declval<_Arg>())), - true_type>::type -__is_assignable_test(int); -template <class, class> -false_type __is_assignable_test(...); -template <class _Tp, class _Arg, - bool = is_void<_Tp>::value || is_void<_Arg>::value> -struct __is_assignable_imp - : public decltype((__is_assignable_test<_Tp, _Arg>(0))) {}; -template <class _Tp, class _Arg> -struct __is_assignable_imp<_Tp, _Arg, true> : public false_type {}; -template <class _Tp, class _Arg> -struct is_assignable : public __is_assignable_imp<_Tp, _Arg> {}; - -template <class _Tp> -struct __libcpp_is_integral : public false_type {}; -template <> -struct __libcpp_is_integral<bool> : public true_type {}; -template <> -struct __libcpp_is_integral<char> : public true_type {}; -template <> -struct __libcpp_is_integral<signed char> : public true_type {}; -template <> -struct __libcpp_is_integral<unsigned char> : public true_type {}; -template <> -struct __libcpp_is_integral<wchar_t> : public true_type {}; -template <> -struct __libcpp_is_integral<short> : public true_type {}; // NOLINT -template <> -struct __libcpp_is_integral<unsigned short> : public true_type {}; // NOLINT -template <> -struct __libcpp_is_integral<int> : public true_type {}; -template <> -struct __libcpp_is_integral<unsigned int> : public true_type {}; -template <> -struct __libcpp_is_integral<long> : public true_type {}; // NOLINT -template <> -struct __libcpp_is_integral<unsigned long> : public true_type {}; // NOLINT -template <> -struct __libcpp_is_integral<long long> : public true_type {}; // NOLINT -template <> // NOLINTNEXTLINE -struct __libcpp_is_integral<unsigned long long> : public true_type {}; -template <class _Tp> -struct is_integral - : public __libcpp_is_integral<typename remove_cv<_Tp>::type> {}; - -template <class _Tp> -struct __libcpp_is_floating_point : public false_type {}; -template <> -struct __libcpp_is_floating_point<float> : public true_type {}; -template <> -struct __libcpp_is_floating_point<double> : public true_type {}; -template <> -struct __libcpp_is_floating_point<long double> : public true_type {}; -template <class _Tp> -struct is_floating_point - : public __libcpp_is_floating_point<typename remove_cv<_Tp>::type> {}; - -template <class _Tp> -struct is_arithmetic - : public integral_constant<bool, is_integral<_Tp>::value || - is_floating_point<_Tp>::value> {}; - -template <class _Tp> -struct __libcpp_is_pointer : public false_type {}; -template <class _Tp> -struct __libcpp_is_pointer<_Tp*> : public true_type {}; -template <class _Tp> -struct is_pointer : public __libcpp_is_pointer<typename remove_cv<_Tp>::type> { -}; - -template <class _Tp> -struct __libcpp_is_member_pointer : public false_type {}; -template <class _Tp, class _Up> -struct __libcpp_is_member_pointer<_Tp _Up::*> : public true_type {}; -template <class _Tp> -struct is_member_pointer - : public __libcpp_is_member_pointer<typename remove_cv<_Tp>::type> {}; - -template <class _Tp> -struct __libcpp_union : public false_type {}; -template <class _Tp> -struct is_union : public __libcpp_union<typename remove_cv<_Tp>::type> {}; - -template <class T> -struct is_reference : false_type {}; -template <class T> -struct is_reference<T&> : true_type {}; -template <class T> -struct is_reference<T&&> : true_type {}; - -template <class T> -inline constexpr bool is_reference_v = is_reference<T>::value; - -struct __two { - char __lx[2]; -}; - -namespace __is_class_imp { -template <class _Tp> -char __test(int _Tp::*); -template <class _Tp> -__two __test(...); -} // namespace __is_class_imp -template <class _Tp> -struct is_class - : public integral_constant<bool, - sizeof(__is_class_imp::__test<_Tp>(0)) == 1 && - !is_union<_Tp>::value> {}; - -template <class _Tp> -struct __is_nullptr_t_impl : public false_type {}; -template <> -struct __is_nullptr_t_impl<nullptr_t> : public true_type {}; -template <class _Tp> -struct __is_nullptr_t - : public __is_nullptr_t_impl<typename remove_cv<_Tp>::type> {}; -template <class _Tp> -struct is_null_pointer - : public __is_nullptr_t_impl<typename remove_cv<_Tp>::type> {}; - -template <class _Tp> -struct is_enum - : public integral_constant< - bool, !is_void<_Tp>::value && !is_integral<_Tp>::value && - !is_floating_point<_Tp>::value && !is_array<_Tp>::value && - !is_pointer<_Tp>::value && !is_reference<_Tp>::value && - !is_member_pointer<_Tp>::value && !is_union<_Tp>::value && - !is_class<_Tp>::value && !is_function<_Tp>::value> {}; - -template <class _Tp> -struct is_scalar - : public integral_constant< - bool, is_arithmetic<_Tp>::value || is_member_pointer<_Tp>::value || - is_pointer<_Tp>::value || __is_nullptr_t<_Tp>::value || - is_enum<_Tp>::value> {}; -template <> -struct is_scalar<nullptr_t> : public true_type {}; - -} // namespace std - -#endif // STD_TYPE_TRAITS_H -)"; - -static constexpr char AbslTypeTraitsHeader[] = R"( -#ifndef ABSL_TYPE_TRAITS_H -#define ABSL_TYPE_TRAITS_H - -#include "std_type_traits.h" - -namespace absl { - -template <typename... Ts> -struct conjunction : std::true_type {}; - -template <typename T, typename... Ts> -struct conjunction<T, Ts...> - : std::conditional<T::value, conjunction<Ts...>, T>::type {}; - -template <typename T> -struct conjunction<T> : T {}; - -template <typename T> -struct negation : std::integral_constant<bool, !T::value> {}; - -template <bool B, typename T = void> -using enable_if_t = typename std::enable_if<B, T>::type; - -} // namespace absl - -#endif // ABSL_TYPE_TRAITS_H -)"; - -static constexpr char StdStringHeader[] = R"( -#ifndef STRING_H -#define STRING_H - -namespace std { - -struct string { - string(const char*); - ~string(); - bool empty(); -}; -bool operator!=(const string &LHS, const char *RHS); - -} // namespace std - -#endif // STRING_H -)"; - -static constexpr char StdUtilityHeader[] = R"( -#ifndef UTILITY_H -#define UTILITY_H - -#include "std_type_traits.h" - -namespace std { - -template <typename T> -constexpr remove_reference_t<T>&& move(T&& x); - -template <typename T> -void swap(T& a, T& b) noexcept; - -} // namespace std - -#endif // UTILITY_H -)"; - -static constexpr char StdInitializerListHeader[] = R"( -#ifndef INITIALIZER_LIST_H -#define INITIALIZER_LIST_H - -namespace std { - -template <typename T> -class initializer_list { - public: - const T *a, *b; - initializer_list() noexcept; -}; - -} // namespace std - -#endif // INITIALIZER_LIST_H -)"; - -static constexpr char StdOptionalHeader[] = R"( -#include "std_initializer_list.h" -#include "std_type_traits.h" -#include "std_utility.h" - -namespace std { - -struct in_place_t {}; -constexpr in_place_t in_place; - -struct nullopt_t { - constexpr explicit nullopt_t() {} -}; -constexpr nullopt_t nullopt; - -template <class _Tp> -struct __optional_destruct_base { - constexpr void reset() noexcept; -}; - -template <class _Tp> -struct __optional_storage_base : __optional_destruct_base<_Tp> { - constexpr bool has_value() const noexcept; -}; - -template <typename _Tp> -class optional : private __optional_storage_base<_Tp> { - using __base = __optional_storage_base<_Tp>; - - public: - using value_type = _Tp; - - private: - struct _CheckOptionalArgsConstructor { - template <class _Up> - static constexpr bool __enable_implicit() { - return is_constructible_v<_Tp, _Up&&> && is_convertible_v<_Up&&, _Tp>; - } - - template <class _Up> - static constexpr bool __enable_explicit() { - return is_constructible_v<_Tp, _Up&&> && !is_convertible_v<_Up&&, _Tp>; - } - }; - template <class _Up> - using _CheckOptionalArgsCtor = - _If<_IsNotSame<__uncvref_t<_Up>, in_place_t>::value && - _IsNotSame<__uncvref_t<_Up>, optional>::value, - _CheckOptionalArgsConstructor, __check_tuple_constructor_fail>; - template <class _QualUp> - struct _CheckOptionalLikeConstructor { - template <class _Up, class _Opt = optional<_Up>> - using __check_constructible_from_opt = - _Or<is_constructible<_Tp, _Opt&>, is_constructible<_Tp, _Opt const&>, - is_constructible<_Tp, _Opt&&>, is_constructible<_Tp, _Opt const&&>, - is_convertible<_Opt&, _Tp>, is_convertible<_Opt const&, _Tp>, - is_convertible<_Opt&&, _Tp>, is_convertible<_Opt const&&, _Tp>>; - template <class _Up, class _QUp = _QualUp> - static constexpr bool __enable_implicit() { - return is_convertible<_QUp, _Tp>::value && - !__check_constructible_from_opt<_Up>::value; - } - template <class _Up, class _QUp = _QualUp> - static constexpr bool __enable_explicit() { - return !is_convertible<_QUp, _Tp>::value && - !__check_constructible_from_opt<_Up>::value; - } - }; - - template <class _Up, class _QualUp> - using _CheckOptionalLikeCtor = - _If<_And<_IsNotSame<_Up, _Tp>, is_constructible<_Tp, _QualUp>>::value, - _CheckOptionalLikeConstructor<_QualUp>, - __check_tuple_constructor_fail>; - - - template <class _Up, class _QualUp> - using _CheckOptionalLikeAssign = _If< - _And< - _IsNotSame<_Up, _Tp>, - is_constructible<_Tp, _QualUp>, - is_assignable<_Tp&, _QualUp> - >::value, - _CheckOptionalLikeConstructor<_QualUp>, - __check_tuple_constructor_fail - >; - - public: - constexpr optional() noexcept {} - constexpr optional(const optional&) = default; - constexpr optional(optional&&) = default; - constexpr optional(nullopt_t) noexcept {} - - template < - class _InPlaceT, class... _Args, - class = enable_if_t<_And<_IsSame<_InPlaceT, in_place_t>, - is_constructible<value_type, _Args...>>::value>> - constexpr explicit optional(_InPlaceT, _Args&&... __args); - - template <class _Up, class... _Args, - class = enable_if_t<is_constructible_v< - value_type, initializer_list<_Up>&, _Args...>>> - constexpr explicit optional(in_place_t, initializer_list<_Up> __il, - _Args&&... __args); - - template < - class _Up = value_type, - enable_if_t<_CheckOptionalArgsCtor<_Up>::template __enable_implicit<_Up>(), - int> = 0> - constexpr optional(_Up&& __v); - - template < - class _Up, - enable_if_t<_CheckOptionalArgsCtor<_Up>::template __enable_explicit<_Up>(), - int> = 0> - constexpr explicit optional(_Up&& __v); - - template <class _Up, enable_if_t<_CheckOptionalLikeCtor<_Up, _Up const&>:: - template __enable_implicit<_Up>(), - int> = 0> - constexpr optional(const optional<_Up>& __v); - - template <class _Up, enable_if_t<_CheckOptionalLikeCtor<_Up, _Up const&>:: - template __enable_explicit<_Up>(), - int> = 0> - constexpr explicit optional(const optional<_Up>& __v); - - template <class _Up, enable_if_t<_CheckOptionalLikeCtor<_Up, _Up&&>:: - template __enable_implicit<_Up>(), - int> = 0> - constexpr optional(optional<_Up>&& __v); - - template <class _Up, enable_if_t<_CheckOptionalLikeCtor<_Up, _Up&&>:: - template __enable_explicit<_Up>(), - int> = 0> - constexpr explicit optional(optional<_Up>&& __v); - - constexpr optional& operator=(nullopt_t) noexcept; - - optional& operator=(const optional&); - - optional& operator=(optional&&); - - template <class _Up = value_type, - class = enable_if_t<_And<_IsNotSame<__uncvref_t<_Up>, optional>, - _Or<_IsNotSame<__uncvref_t<_Up>, value_type>, - _Not<is_scalar<value_type>>>, - is_constructible<value_type, _Up>, - is_assignable<value_type&, _Up>>::value>> - constexpr optional& operator=(_Up&& __v); - - template <class _Up, enable_if_t<_CheckOptionalLikeAssign<_Up, _Up const&>:: - template __enable_assign<_Up>(), - int> = 0> - constexpr optional& operator=(const optional<_Up>& __v); - - template <class _Up, enable_if_t<_CheckOptionalLikeCtor<_Up, _Up&&>:: - template __enable_assign<_Up>(), - int> = 0> - constexpr optional& operator=(optional<_Up>&& __v); - - const _Tp& operator*() const&; - _Tp& operator*() &; - const _Tp&& operator*() const&&; - _Tp&& operator*() &&; - - const _Tp* operator->() const; - _Tp* operator->(); - - const _Tp& value() const&; - _Tp& value() &; - const _Tp&& value() const&&; - _Tp&& value() &&; - - template <typename U> - constexpr _Tp value_or(U&& v) const&; - template <typename U> - _Tp value_or(U&& v) &&; - - template <typename... Args> - _Tp& emplace(Args&&... args); - - template <typename U, typename... Args> - _Tp& emplace(std::initializer_list<U> ilist, Args&&... args); - - using __base::reset; - - constexpr explicit operator bool() const noexcept; - using __base::has_value; - - constexpr void swap(optional& __opt) noexcept; -}; - -template <typename T> -constexpr optional<typename std::decay<T>::type> make_optional(T&& v); - -template <typename T, typename... Args> -constexpr optional<T> make_optional(Args&&... args); - -template <typename T, typename U, typename... Args> -constexpr optional<T> make_optional(std::initializer_list<U> il, - Args&&... args); - -template <typename T, typename U> -constexpr bool operator==(const optional<T> &lhs, const optional<U> &rhs); -template <typename T, typename U> -constexpr bool operator!=(const optional<T> &lhs, const optional<U> &rhs); - -template <typename T> -constexpr bool operator==(const optional<T> &opt, nullopt_t); - -// C++20 and later do not define the following overloads because they are -// provided by rewritten candidates instead. -#if __cplusplus < 202002L -template <typename T> -constexpr bool operator==(nullopt_t, const optional<T> &opt); -template <typename T> -constexpr bool operator!=(const optional<T> &opt, nullopt_t); -template <typename T> -constexpr bool operator!=(nullopt_t, const optional<T> &opt); -#endif // __cplusplus < 202002L - -template <typename T, typename U> -constexpr bool operator==(const optional<T> &opt, const U &value); -template <typename T, typename U> -constexpr bool operator==(const T &value, const optional<U> &opt); -template <typename T, typename U> -constexpr bool operator!=(const optional<T> &opt, const U &value); -template <typename T, typename U> -constexpr bool operator!=(const T &value, const optional<U> &opt); - -} // namespace std -)"; - -static constexpr char AbslOptionalHeader[] = R"( -#include "absl_type_traits.h" -#include "std_initializer_list.h" -#include "std_type_traits.h" -#include "std_utility.h" - -namespace absl { - -struct nullopt_t { - constexpr explicit nullopt_t() {} -}; -constexpr nullopt_t nullopt; - -struct in_place_t {}; -constexpr in_place_t in_place; - -template <typename T> -class optional; - -namespace optional_internal { - -template <typename T, typename U> -struct is_constructible_convertible_from_optional - : std::integral_constant< - bool, std::is_constructible<T, optional<U>&>::value || - std::is_constructible<T, optional<U>&&>::value || - std::is_constructible<T, const optional<U>&>::value || - std::is_constructible<T, const optional<U>&&>::value || - std::is_convertible<optional<U>&, T>::value || - std::is_convertible<optional<U>&&, T>::value || - std::is_convertible<const optional<U>&, T>::value || - std::is_convertible<const optional<U>&&, T>::value> {}; - -template <typename T, typename U> -struct is_constructible_convertible_assignable_from_optional - : std::integral_constant< - bool, is_constructible_convertible_from_optional<T, U>::value || - std::is_assignable<T&, optional<U>&>::value || - std::is_assignable<T&, optional<U>&&>::value || - std::is_assignable<T&, const optional<U>&>::value || - std::is_assignable<T&, const optional<U>&&>::value> {}; - -} // namespace optional_internal - -template <typename T> -class optional { - public: - constexpr optional() noexcept; - - constexpr optional(nullopt_t) noexcept; - - optional(const optional&) = default; - - optional(optional&&) = default; - - template <typename InPlaceT, typename... Args, - absl::enable_if_t<absl::conjunction< - std::is_same<InPlaceT, in_place_t>, - std::is_constructible<T, Args&&...>>::value>* = nullptr> - constexpr explicit optional(InPlaceT, Args&&... args); - - template <typename U, typename... Args, - typename = typename std::enable_if<std::is_constructible< - T, std::initializer_list<U>&, Args&&...>::value>::type> - constexpr explicit optional(in_place_t, std::initializer_list<U> il, - Args&&... args); - - template < - typename U = T, - typename std::enable_if< - absl::conjunction<absl::negation<std::is_same< - in_place_t, typename std::decay<U>::type>>, - absl::negation<std::is_same< - optional<T>, typename std::decay<U>::type>>, - std::is_convertible<U&&, T>, - std::is_constructible<T, U&&>>::value, - bool>::type = false> - constexpr optional(U&& v); - - template < - typename U = T, - typename std::enable_if< - absl::conjunction<absl::negation<std::is_same< - in_place_t, typename std::decay<U>::type>>, - absl::negation<std::is_same< - optional<T>, typename std::decay<U>::type>>, - absl::negation<std::is_convertible<U&&, T>>, - std::is_constructible<T, U&&>>::value, - bool>::type = false> - explicit constexpr optional(U&& v); - - template <typename U, - typename std::enable_if< - absl::conjunction< - absl::negation<std::is_same<T, U>>, - std::is_constructible<T, const U&>, - absl::negation< - optional_internal:: - is_constructible_convertible_from_optional<T, U>>, - std::is_convertible<const U&, T>>::value, - bool>::type = false> - optional(const optional<U>& rhs); - - template <typename U, - typename std::enable_if< - absl::conjunction< - absl::negation<std::is_same<T, U>>, - std::is_constructible<T, const U&>, - absl::negation< - optional_internal:: - is_constructible_convertible_from_optional<T, U>>, - absl::negation<std::is_convertible<const U&, T>>>::value, - bool>::type = false> - explicit optional(const optional<U>& rhs); - - template < - typename U, - typename std::enable_if< - absl::conjunction< - absl::negation<std::is_same<T, U>>, std::is_constructible<T, U&&>, - absl::negation< - optional_internal::is_constructible_convertible_from_optional< - T, U>>, - std::is_convertible<U&&, T>>::value, - bool>::type = false> - optional(optional<U>&& rhs); - - template < - typename U, - typename std::enable_if< - absl::conjunction< - absl::negation<std::is_same<T, U>>, std::is_constructible<T, U&&>, - absl::negation< - optional_internal::is_constructible_convertible_from_optional< - T, U>>, - absl::negation<std::is_convertible<U&&, T>>>::value, - bool>::type = false> - explicit optional(optional<U>&& rhs); - - optional& operator=(nullopt_t) noexcept; - - optional& operator=(const optional& src); - - optional& operator=(optional&& src); - - template < - typename U = T, - typename = typename std::enable_if<absl::conjunction< - absl::negation< - std::is_same<optional<T>, typename std::decay<U>::type>>, - absl::negation< - absl::conjunction<std::is_scalar<T>, - std::is_same<T, typename std::decay<U>::type>>>, - std::is_constructible<T, U>, std::is_assignable<T&, U>>::value>::type> - optional& operator=(U&& v); - - template < - typename U, - typename = typename std::enable_if<absl::conjunction< - absl::negation<std::is_same<T, U>>, - std::is_constructible<T, const U&>, std::is_assignable<T&, const U&>, - absl::negation< - optional_internal:: - is_constructible_convertible_assignable_from_optional< - T, U>>>::value>::type> - optional& operator=(const optional<U>& rhs); - - template <typename U, - typename = typename std::enable_if<absl::conjunction< - absl::negation<std::is_same<T, U>>, std::is_constructible<T, U>, - std::is_assignable<T&, U>, - absl::negation< - optional_internal:: - is_constructible_convertible_assignable_from_optional< - T, U>>>::value>::type> - optional& operator=(optional<U>&& rhs); - - const T& operator*() const&; - T& operator*() &; - const T&& operator*() const&&; - T&& operator*() &&; - - const T* operator->() const; - T* operator->(); - - const T& value() const&; - T& value() &; - const T&& value() const&&; - T&& value() &&; - - template <typename U> - constexpr T value_or(U&& v) const&; - template <typename U> - T value_or(U&& v) &&; - - template <typename... Args> - T& emplace(Args&&... args); - - template <typename U, typename... Args> - T& emplace(std::initializer_list<U> ilist, Args&&... args); - - void reset() noexcept; - - constexpr explicit operator bool() const noexcept; - constexpr bool has_value() const noexcept; - - void swap(optional& rhs) noexcept; -}; - -template <typename T> -constexpr optional<typename std::decay<T>::type> make_optional(T&& v); - -template <typename T, typename... Args> -constexpr optional<T> make_optional(Args&&... args); - -template <typename T, typename U, typename... Args> -constexpr optional<T> make_optional(std::initializer_list<U> il, - Args&&... args); - -template <typename T, typename U> -constexpr bool operator==(const optional<T> &lhs, const optional<U> &rhs); -template <typename T, typename U> -constexpr bool operator!=(const optional<T> &lhs, const optional<U> &rhs); - -template <typename T> -constexpr bool operator==(const optional<T> &opt, nullopt_t); -template <typename T> -constexpr bool operator==(nullopt_t, const optional<T> &opt); -template <typename T> -constexpr bool operator!=(const optional<T> &opt, nullopt_t); -template <typename T> -constexpr bool operator!=(nullopt_t, const optional<T> &opt); - -template <typename T, typename U> -constexpr bool operator==(const optional<T> &opt, const U &value); -template <typename T, typename U> -constexpr bool operator==(const T &value, const optional<U> &opt); -template <typename T, typename U> -constexpr bool operator!=(const optional<T> &opt, const U &value); -template <typename T, typename U> -constexpr bool operator!=(const T &value, const optional<U> &opt); - -} // namespace absl -)"; - -static constexpr char BaseOptionalHeader[] = R"( -#include "std_initializer_list.h" -#include "std_type_traits.h" -#include "std_utility.h" - -namespace base { - -struct in_place_t {}; -constexpr in_place_t in_place; - -struct nullopt_t { - constexpr explicit nullopt_t() {} -}; -constexpr nullopt_t nullopt; - -template <typename T> -class Optional; - -namespace internal { - -template <typename T> -using RemoveCvRefT = std::remove_cv_t<std::remove_reference_t<T>>; - -template <typename T, typename U> -struct IsConvertibleFromOptional - : std::integral_constant< - bool, std::is_constructible<T, Optional<U>&>::value || - std::is_constructible<T, const Optional<U>&>::value || - std::is_constructible<T, Optional<U>&&>::value || - std::is_constructible<T, const Optional<U>&&>::value || - std::is_convertible<Optional<U>&, T>::value || - std::is_convertible<const Optional<U>&, T>::value || - std::is_convertible<Optional<U>&&, T>::value || - std::is_convertible<const Optional<U>&&, T>::value> {}; - -template <typename T, typename U> -struct IsAssignableFromOptional - : std::integral_constant< - bool, IsConvertibleFromOptional<T, U>::value || - std::is_assignable<T&, Optional<U>&>::value || - std::is_assignable<T&, const Optional<U>&>::value || - std::is_assignable<T&, Optional<U>&&>::value || - std::is_assignable<T&, const Optional<U>&&>::value> {}; - -} // namespace internal - -template <typename T> -class Optional { - public: - using value_type = T; - - constexpr Optional() = default; - constexpr Optional(const Optional& other) noexcept = default; - constexpr Optional(Optional&& other) noexcept = default; - - constexpr Optional(nullopt_t); - - template <typename U, - typename std::enable_if< - std::is_constructible<T, const U&>::value && - !internal::IsConvertibleFromOptional<T, U>::value && - std::is_convertible<const U&, T>::value, - bool>::type = false> - Optional(const Optional<U>& other) noexcept; - - template <typename U, - typename std::enable_if< - std::is_constructible<T, const U&>::value && - !internal::IsConvertibleFromOptional<T, U>::value && - !std::is_convertible<const U&, T>::value, - bool>::type = false> - explicit Optional(const Optional<U>& other) noexcept; - - template <typename U, - typename std::enable_if< - std::is_constructible<T, U&&>::value && - !internal::IsConvertibleFromOptional<T, U>::value && - std::is_convertible<U&&, T>::value, - bool>::type = false> - Optional(Optional<U>&& other) noexcept; - - template <typename U, - typename std::enable_if< - std::is_constructible<T, U&&>::value && - !internal::IsConvertibleFromOptional<T, U>::value && - !std::is_convertible<U&&, T>::value, - bool>::type = false> - explicit Optional(Optional<U>&& other) noexcept; - - template <class... Args> - constexpr explicit Optional(in_place_t, Args&&... args); - - template <class U, class... Args, - class = typename std::enable_if<std::is_constructible< - value_type, std::initializer_list<U>&, Args...>::value>::type> - constexpr explicit Optional(in_place_t, std::initializer_list<U> il, - Args&&... args); - - template < - typename U = value_type, - typename std::enable_if< - std::is_constructible<T, U&&>::value && - !std::is_same<internal::RemoveCvRefT<U>, in_place_t>::value && - !std::is_same<internal::RemoveCvRefT<U>, Optional<T>>::value && - std::is_convertible<U&&, T>::value, - bool>::type = false> - constexpr Optional(U&& value); - - template < - typename U = value_type, - typename std::enable_if< - std::is_constructible<T, U&&>::value && - !std::is_same<internal::RemoveCvRefT<U>, in_place_t>::value && - !std::is_same<internal::RemoveCvRefT<U>, Optional<T>>::value && - !std::is_convertible<U&&, T>::value, - bool>::type = false> - constexpr explicit Optional(U&& value); - - Optional& operator=(const Optional& other) noexcept; - - Optional& operator=(Optional&& other) noexcept; - - Optional& operator=(nullopt_t); - - template <typename U> - typename std::enable_if< - !std::is_same<internal::RemoveCvRefT<U>, Optional<T>>::value && - std::is_constructible<T, U>::value && - std::is_assignable<T&, U>::value && - (!std::is_scalar<T>::value || - !std::is_same<typename std::decay<U>::type, T>::value), - Optional&>::type - operator=(U&& value) noexcept; - - template <typename U> - typename std::enable_if<!internal::IsAssignableFromOptional<T, U>::value && - std::is_constructible<T, const U&>::value && - std::is_assignable<T&, const U&>::value, - Optional&>::type - operator=(const Optional<U>& other) noexcept; - - template <typename U> - typename std::enable_if<!internal::IsAssignableFromOptional<T, U>::value && - std::is_constructible<T, U>::value && - std::is_assignable<T&, U>::value, - Optional&>::type - operator=(Optional<U>&& other) noexcept; - - const T& operator*() const&; - T& operator*() &; - const T&& operator*() const&&; - T&& operator*() &&; - - const T* operator->() const; - T* operator->(); - - const T& value() const&; - T& value() &; - const T&& value() const&&; - T&& value() &&; - - template <typename U> - constexpr T value_or(U&& v) const&; - template <typename U> - T value_or(U&& v) &&; - - template <typename... Args> - T& emplace(Args&&... args); - - template <typename U, typename... Args> - T& emplace(std::initializer_list<U> ilist, Args&&... args); - - void reset() noexcept; - - constexpr explicit operator bool() const noexcept; - constexpr bool has_value() const noexcept; - - void swap(Optional& other); -}; - -template <typename T> -constexpr Optional<typename std::decay<T>::type> make_optional(T&& v); - -template <typename T, typename... Args> -constexpr Optional<T> make_optional(Args&&... args); - -template <typename T, typename U, typename... Args> -constexpr Optional<T> make_optional(std::initializer_list<U> il, - Args&&... args); - -template <typename T, typename U> -constexpr bool operator==(const Optional<T> &lhs, const Optional<U> &rhs); -template <typename T, typename U> -constexpr bool operator!=(const Optional<T> &lhs, const Optional<U> &rhs); - -template <typename T> -constexpr bool operator==(const Optional<T> &opt, nullopt_t); -template <typename T> -constexpr bool operator==(nullopt_t, const Optional<T> &opt); -template <typename T> -constexpr bool operator!=(const Optional<T> &opt, nullopt_t); -template <typename T> -constexpr bool operator!=(nullopt_t, const Optional<T> &opt); - -template <typename T, typename U> -constexpr bool operator==(const Optional<T> &opt, const U &value); -template <typename T, typename U> -constexpr bool operator==(const T &value, const Optional<U> &opt); -template <typename T, typename U> -constexpr bool operator!=(const Optional<T> &opt, const U &value); -template <typename T, typename U> -constexpr bool operator!=(const T &value, const Optional<U> &opt); - -} // namespace base -)"; - /// Replaces all occurrences of `Pattern` in `S` with `Replacement`. static void ReplaceAllOccurrences(std::string &S, const std::string &Pattern, const std::string &Replacement) { @@ -1323,16 +98,7 @@ protected: ReplaceAllOccurrences(SourceCode, "$ns", GetParam().NamespaceName); ReplaceAllOccurrences(SourceCode, "$optional", GetParam().TypeName); - std::vector<std::pair<std::string, std::string>> Headers; - Headers.emplace_back("cstddef.h", CSDtdDefHeader); - Headers.emplace_back("std_initializer_list.h", StdInitializerListHeader); - Headers.emplace_back("std_string.h", StdStringHeader); - Headers.emplace_back("std_type_traits.h", StdTypeTraitsHeader); - Headers.emplace_back("std_utility.h", StdUtilityHeader); - Headers.emplace_back("std_optional.h", StdOptionalHeader); - Headers.emplace_back("absl_type_traits.h", AbslTypeTraitsHeader); - Headers.emplace_back("absl_optional.h", AbslOptionalHeader); - Headers.emplace_back("base_optional.h", BaseOptionalHeader); + auto Headers = getMockHeaders(); Headers.emplace_back("unchecked_optional_access_test.h", R"( #include "absl_optional.h" #include "base_optional.h" diff --git a/clang/unittests/Basic/CMakeLists.txt b/clang/unittests/Basic/CMakeLists.txt index 8c8baa5..f20c8db 100644 --- a/clang/unittests/Basic/CMakeLists.txt +++ b/clang/unittests/Basic/CMakeLists.txt @@ -6,6 +6,7 @@ add_distinct_clang_unittest(BasicTests DiagnosticTest.cpp FileEntryTest.cpp FileManagerTest.cpp + LangOptionsTest.cpp LineOffsetMappingTest.cpp OffloadArchTest.cpp SanitizersTest.cpp diff --git a/clang/unittests/Basic/LangOptionsTest.cpp b/clang/unittests/Basic/LangOptionsTest.cpp new file mode 100644 index 0000000..0d7d5ec --- /dev/null +++ b/clang/unittests/Basic/LangOptionsTest.cpp @@ -0,0 +1,56 @@ +//===- unittests/Basic/LangOptionsTest.cpp --------------------------------===// +// +// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. +// See https://llvm.org/LICENSE.txt for license information. +// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception +// +//===----------------------------------------------------------------------===// + +#include "clang/Basic/LangOptions.h" +#include "gtest/gtest.h" + +using namespace llvm; +using namespace clang; + +namespace { +TEST(LangOptsTest, CStdLang) { + LangOptions opts; + EXPECT_FALSE(opts.getCLangStd()); + opts.GNUMode = 0; + opts.Digraphs = 1; + EXPECT_EQ(opts.getCLangStd(), 199409); + opts.C99 = 1; + EXPECT_EQ(opts.getCLangStd(), 199901); + opts.C11 = 1; + EXPECT_EQ(opts.getCLangStd(), 201112); + opts.C17 = 1; + EXPECT_EQ(opts.getCLangStd(), 201710); + opts.C23 = 1; + EXPECT_EQ(opts.getCLangStd(), 202311); + opts.C2y = 1; + EXPECT_EQ(opts.getCLangStd(), 202400); + + EXPECT_FALSE(opts.getCPlusPlusLangStd()); +} + +TEST(LangOptsTest, CppStdLang) { + LangOptions opts; + EXPECT_FALSE(opts.getCPlusPlusLangStd()); + opts.CPlusPlus = 1; + EXPECT_EQ(opts.getCPlusPlusLangStd(), 199711); + opts.CPlusPlus11 = 1; + EXPECT_EQ(opts.getCPlusPlusLangStd(), 201103); + opts.CPlusPlus14 = 1; + EXPECT_EQ(opts.getCPlusPlusLangStd(), 201402); + opts.CPlusPlus17 = 1; + EXPECT_EQ(opts.getCPlusPlusLangStd(), 201703); + opts.CPlusPlus20 = 1; + EXPECT_EQ(opts.getCPlusPlusLangStd(), 202002); + opts.CPlusPlus23 = 1; + EXPECT_EQ(opts.getCPlusPlusLangStd(), 202302); + opts.CPlusPlus26 = 1; + EXPECT_EQ(opts.getCPlusPlusLangStd(), 202400); + + EXPECT_FALSE(opts.getCLangStd()); +} +} // namespace diff --git a/clang/unittests/Format/ConfigParseTest.cpp b/clang/unittests/Format/ConfigParseTest.cpp index 6111e86..52f02c3 100644 --- a/clang/unittests/Format/ConfigParseTest.cpp +++ b/clang/unittests/Format/ConfigParseTest.cpp @@ -1264,6 +1264,13 @@ TEST(ConfigParseTest, ParsesConfigurationWithLanguages) { IndentWidth, 56u); } +TEST(ConfigParseTest, AllowCommentOnlyConfigFile) { + FormatStyle Style = {}; + Style.Language = FormatStyle::LK_Cpp; + EXPECT_EQ(parseConfiguration("#Language: C", &Style), ParseError::Success); + EXPECT_EQ(Style.Language, FormatStyle::LK_Cpp); +} + TEST(ConfigParseTest, AllowCppForC) { FormatStyle Style = {}; Style.Language = FormatStyle::LK_C; diff --git a/clang/unittests/Format/FormatTest.cpp b/clang/unittests/Format/FormatTest.cpp index 450c34f..b9ad930 100644 --- a/clang/unittests/Format/FormatTest.cpp +++ b/clang/unittests/Format/FormatTest.cpp @@ -19539,6 +19539,15 @@ TEST_F(FormatTest, AlignConsecutiveAssignments) { "int j = 2;", Alignment); + verifyFormat("int abcdefghijk = 111;\n" + "auto lambda = [] {\n" + " int c = call(1, //\n" + " 2, //\n" + " 3, //\n" + " 4);\n" + "};", + Alignment); + verifyFormat("template <typename T, typename T_0 = very_long_type_name_0,\n" " typename B = very_long_type_name_1,\n" " typename T_2 = very_long_type_name_2>\n" @@ -19577,6 +19586,12 @@ TEST_F(FormatTest, AlignConsecutiveAssignments) { Alignment); verifyFormat("auto aaaaaaaaaaaaaaaaaaaaa = {};\n" "auto b = g([] {\n" + " return \"Hello \"\n" + " \"World\";\n" + "});", + Alignment); + verifyFormat("auto aaaaaaaaaaaaaaaaaaaaa = {};\n" + "auto b = g([] {\n" " f();\n" " return;\n" "});", @@ -19599,12 +19614,11 @@ TEST_F(FormatTest, AlignConsecutiveAssignments) { " ccc ? aaaaa : bbbbb,\n" " dddddddddddddddddddddddddd);", Alignment); - // FIXME: https://llvm.org/PR53497 - // verifyFormat("auto aaaaaaaaaaaa = f();\n" - // "auto b = f(aaaaaaaaaaaaaaaaaaaaaaaaa,\n" - // " ccc ? aaaaa : bbbbb,\n" - // " dddddddddddddddddddddddddd);", - // Alignment); + verifyFormat("auto aaaaaaaaaaaa = f();\n" + "auto b = f(aaaaaaaaaaaaaaaaaaaaaaaaa,\n" + " ccc ? aaaaa : bbbbb,\n" + " dddddddddddddddddddddddddd);", + Alignment); // Confirm proper handling of AlignConsecutiveAssignments with // BinPackArguments. @@ -20192,6 +20206,11 @@ TEST_F(FormatTest, AlignConsecutiveDeclarations) { " i = 3 //\n" "};", Alignment); + // When assignments are nested, each level should be aligned. + verifyFormat("float i2 = 0;\n" + "auto v = type{i2 = 1, //\n" + " i = 3};", + Alignment); Alignment.AlignConsecutiveAssignments.Enabled = false; verifyFormat( @@ -20681,6 +20700,21 @@ TEST_F(FormatTest, AlignWithLineBreaks) { "}", Style); + verifyFormat("void foo() {\n" + " int myVar = 5;\n" + " double x = 3.14;\n" + " auto str = (\"Hello \"\n" + " \"World\");\n" + " auto s = (\"Hello \"\n" + " \"Again\");\n" + "}", + Style); + + verifyFormat("A B = {\"Hello \"\n" + " \"World\"};\n" + "BYTE payload = 2;", + Style); + // clang-format off verifyFormat("void foo() {\n" " const int capacityBefore = Entries.capacity();\n" @@ -20763,6 +20797,28 @@ TEST_F(FormatTest, AlignWithInitializerPeriods) { "}", Style); + // The lines inside the braces are supposed to be indented by + // BracedInitializerIndentWidth from the start of the line. They should not + // move with the opening brace. + verifyFormat("void foo2(void) {\n" + " BYTE p[1] = 1;\n" + " A B = {\n" + " .one_foooooooooooooooo = 2,\n" + " .two_fooooooooooooo = 3,\n" + " .three_fooooooooooooo = 4,\n" + " };\n" + " BYTE payload = 2;\n" + "}", + Style); + + verifyFormat("auto aaaaaaaaaaaaaaaaaaaaa = {};\n" + "auto b = g([] {\n" + " x = {.one_foooooooooooooooo = 2, //\n" + " .two_fooooooooooooo = 3, //\n" + " .three_fooooooooooooo = 4};\n" + "});", + Style); + Style.AlignConsecutiveAssignments.Enabled = false; Style.AlignConsecutiveDeclarations.Enabled = true; verifyFormat("void foo3(void) {\n" diff --git a/clang/unittests/Format/TokenAnnotatorTest.cpp b/clang/unittests/Format/TokenAnnotatorTest.cpp index c21b118..1152466 100644 --- a/clang/unittests/Format/TokenAnnotatorTest.cpp +++ b/clang/unittests/Format/TokenAnnotatorTest.cpp @@ -1490,6 +1490,11 @@ TEST_F(TokenAnnotatorTest, UnderstandsRequiresExpressions) { EXPECT_TOKEN(Tokens[4], tok::l_paren, TT_RequiresExpressionLParen); EXPECT_TOKEN(Tokens[8], tok::l_brace, TT_RequiresExpressionLBrace); + Tokens = annotate("bool foo{requires { 0; }};"); + ASSERT_EQ(Tokens.size(), 11u) << Tokens; + EXPECT_TOKEN(Tokens[3], tok::kw_requires, TT_RequiresExpression); + EXPECT_TOKEN(Tokens[4], tok::l_brace, TT_RequiresExpressionLBrace); + Tokens = annotate("if (requires(int i) { i + 5; }) return;"); ASSERT_EQ(Tokens.size(), 17u) << Tokens; EXPECT_TOKEN(Tokens[2], tok::kw_requires, TT_RequiresExpression); diff --git a/clang/unittests/Frontend/CompilerInvocationTest.cpp b/clang/unittests/Frontend/CompilerInvocationTest.cpp index 75390aa..1332422 100644 --- a/clang/unittests/Frontend/CompilerInvocationTest.cpp +++ b/clang/unittests/Frontend/CompilerInvocationTest.cpp @@ -732,6 +732,26 @@ TEST_F(CommandLineTest, ConditionalParsingIfTrueFlagPresent) { ASSERT_THAT(GeneratedArgs, Contains(StrEq("-sycl-std=2017"))); } +TEST_F(CommandLineTest, ConditionalParsingIfHLSLFlagPresent) { + const char *Args[] = {"-xhlsl"}; + + CompilerInvocation::CreateFromArgs(Invocation, Args, *Diags); + + ASSERT_EQ(Invocation.getLangOpts().MaxMatrixDimension, 4u); + + Invocation.generateCC1CommandLine(GeneratedArgs, *this); +} + +TEST_F(CommandLineTest, ConditionalParsingIfHLSLFlagNotPresent) { + const char *Args[] = {""}; + + CompilerInvocation::CreateFromArgs(Invocation, Args, *Diags); + + ASSERT_EQ(Invocation.getLangOpts().MaxMatrixDimension, 1048575u); + + Invocation.generateCC1CommandLine(GeneratedArgs, *this); +} + // Wide integer option. TEST_F(CommandLineTest, WideIntegerHighValue) { diff --git a/clang/unittests/StaticAnalyzer/CMakeLists.txt b/clang/unittests/StaticAnalyzer/CMakeLists.txt index 9e10c4a..caf686e 100644 --- a/clang/unittests/StaticAnalyzer/CMakeLists.txt +++ b/clang/unittests/StaticAnalyzer/CMakeLists.txt @@ -20,6 +20,7 @@ add_clang_unittest(StaticAnalysisTests SValSimplifyerTest.cpp SValTest.cpp TestReturnValueUnderConstruction.cpp + UnsignedStatDemo.cpp Z3CrosscheckOracleTest.cpp CLANG_LIBS clangBasic diff --git a/clang/unittests/StaticAnalyzer/RangeSetTest.cpp b/clang/unittests/StaticAnalyzer/RangeSetTest.cpp index 9e36aab..a8c7626 100644 --- a/clang/unittests/StaticAnalyzer/RangeSetTest.cpp +++ b/clang/unittests/StaticAnalyzer/RangeSetTest.cpp @@ -27,21 +27,21 @@ template <class RangeOrSet> static std::string toString(const RangeOrSet &Obj) { Obj.dump(SS); return ObjRepresentation; } -LLVM_ATTRIBUTE_UNUSED static std::string toString(const llvm::APSInt &Point) { +[[maybe_unused]] static std::string toString(const llvm::APSInt &Point) { return toString(Point, 10); } // We need it here for better fail diagnostics from gtest. -LLVM_ATTRIBUTE_UNUSED static std::ostream &operator<<(std::ostream &OS, - const RangeSet &Set) { +[[maybe_unused]] static std::ostream &operator<<(std::ostream &OS, + const RangeSet &Set) { return OS << toString(Set); } // We need it here for better fail diagnostics from gtest. -LLVM_ATTRIBUTE_UNUSED static std::ostream &operator<<(std::ostream &OS, - const Range &R) { +[[maybe_unused]] static std::ostream &operator<<(std::ostream &OS, + const Range &R) { return OS << toString(R); } -LLVM_ATTRIBUTE_UNUSED static std::ostream &operator<<(std::ostream &OS, - APSIntType Ty) { +[[maybe_unused]] static std::ostream &operator<<(std::ostream &OS, + APSIntType Ty) { return OS << (Ty.isUnsigned() ? "u" : "s") << Ty.getBitWidth(); } diff --git a/clang/unittests/StaticAnalyzer/SValTest.cpp b/clang/unittests/StaticAnalyzer/SValTest.cpp index 1f4a18b..f96456a 100644 --- a/clang/unittests/StaticAnalyzer/SValTest.cpp +++ b/clang/unittests/StaticAnalyzer/SValTest.cpp @@ -34,13 +34,12 @@ namespace clang { // getType() tests include whole bunch of type comparisons, // so when something is wrong, it's good to have gtest telling us // what are those types. -LLVM_ATTRIBUTE_UNUSED std::ostream &operator<<(std::ostream &OS, - const QualType &T) { +[[maybe_unused]] std::ostream &operator<<(std::ostream &OS, const QualType &T) { return OS << T.getAsString(); } -LLVM_ATTRIBUTE_UNUSED std::ostream &operator<<(std::ostream &OS, - const CanQualType &T) { +[[maybe_unused]] std::ostream &operator<<(std::ostream &OS, + const CanQualType &T) { return OS << QualType{T}; } @@ -302,13 +301,13 @@ void foo(int x) { ASSERT_FALSE(B.getType(Context).isNull()); const auto *BRecordType = dyn_cast<RecordType>(B.getType(Context)); ASSERT_NE(BRecordType, nullptr); - EXPECT_EQ("TestStruct", BRecordType->getOriginalDecl()->getName()); + EXPECT_EQ("TestStruct", BRecordType->getDecl()->getName()); SVal C = getByName("c"); ASSERT_FALSE(C.getType(Context).isNull()); const auto *CRecordType = dyn_cast<RecordType>(C.getType(Context)); ASSERT_NE(CRecordType, nullptr); - EXPECT_EQ("TestUnion", CRecordType->getOriginalDecl()->getName()); + EXPECT_EQ("TestUnion", CRecordType->getDecl()->getName()); auto D = getByName("d").getAs<nonloc::CompoundVal>(); ASSERT_TRUE(D.has_value()); @@ -322,7 +321,7 @@ void foo(int x) { ASSERT_FALSE(LDT.isNull()); const auto *DRecordType = dyn_cast<RecordType>(LDT); ASSERT_NE(DRecordType, nullptr); - EXPECT_EQ("TestStruct", DRecordType->getOriginalDecl()->getName()); + EXPECT_EQ("TestStruct", DRecordType->getDecl()->getName()); } SVAL_TEST(GetStringType, R"( @@ -351,7 +350,7 @@ void TestClass::foo() { ASSERT_NE(APtrTy, nullptr); const auto *ARecordType = dyn_cast<RecordType>(APtrTy->getPointeeType()); ASSERT_NE(ARecordType, nullptr); - EXPECT_EQ("TestClass", ARecordType->getOriginalDecl()->getName()); + EXPECT_EQ("TestClass", ARecordType->getDecl()->getName()); } SVAL_TEST(GetFunctionPtrType, R"( diff --git a/clang/unittests/StaticAnalyzer/UnsignedStatDemo.cpp b/clang/unittests/StaticAnalyzer/UnsignedStatDemo.cpp new file mode 100644 index 0000000..2d1323b --- /dev/null +++ b/clang/unittests/StaticAnalyzer/UnsignedStatDemo.cpp @@ -0,0 +1,150 @@ +//=== UnsignedStatDemo.cpp --------------------------------------*- C++ -*-===// +// +// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. +// See https://llvm.org/LICENSE.txt for license information. +// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception +// +//===----------------------------------------------------------------------===// +// +// This checker demonstrates the use of UnsignedEPStat for per-entry-point +// statistics. It conditionally sets a statistic based on the entry point name. +// +//===----------------------------------------------------------------------===// + +#include "CheckerRegistration.h" +#include "clang/StaticAnalyzer/Core/Checker.h" +#include "clang/StaticAnalyzer/Core/CheckerManager.h" +#include "clang/StaticAnalyzer/Core/PathSensitive/CheckerContext.h" +#include "clang/StaticAnalyzer/Core/PathSensitive/EntryPointStats.h" +#include "llvm/ADT/STLExtras.h" +#include "llvm/ADT/ScopeExit.h" +#include "llvm/ADT/StringMap.h" +#include "llvm/Support/MemoryBuffer.h" +#include "gtest/gtest.h" +#include <optional> + +using namespace clang; +using namespace ento; + +static UnsignedEPStat DemoStat("DemoStat"); + +namespace { +class UnsignedStatTesterChecker : public Checker<check::BeginFunction> { +public: + void checkBeginFunction(CheckerContext &C) const { + StringRef Name; + if (const Decl *D = C.getLocationContext()->getDecl()) + if (const FunctionDecl *F = D->getAsFunction()) + Name = F->getName(); + + // Conditionally set the statistic based on the function name (leaving it + // undefined for all other functions) + if (Name == "func_one") + DemoStat.set(1); + else if (Name == "func_two") + DemoStat.set(2); + else + ; // For any other function (e.g., "func_none") don't set the statistic + } +}; + +void addUnsignedStatTesterChecker(AnalysisASTConsumer &AnalysisConsumer, + AnalyzerOptions &AnOpts) { + AnOpts.CheckersAndPackages = {{"test.DemoStatChecker", true}}; + AnalysisConsumer.AddCheckerRegistrationFn([](CheckerRegistry &Registry) { + Registry.addChecker<UnsignedStatTesterChecker>( + "test.DemoStatChecker", "DescriptionOfDemoStatChecker"); + }); +} + +// Find the index of a column in the CSV header. +// Returns std::nullopt if the column is not found. +static std::optional<unsigned> +findColumnIndex(llvm::ArrayRef<llvm::StringRef> Header, + llvm::StringRef ColumnName) { + auto Iter = llvm::find(Header, ColumnName); + if (Iter != Header.end()) + return std::distance(Header.begin(), Iter); + return std::nullopt; +} + +// Parse CSV content and extract a mapping from one column to another. +// KeyColumn is used as the map key (e.g., "DebugName"). +// ValueColumn is used as the map value (e.g., "DemoStat"). +// Returns a map from key column values to value column values. +static llvm::StringMap<std::string> +parseCSVColumnMapping(llvm::StringRef CSVContent, llvm::StringRef KeyColumn, + llvm::StringRef ValueColumn) { + llvm::StringMap<std::string> Result; + + // Parse CSV: first line is header, subsequent lines are data + llvm::SmallVector<llvm::StringRef, 8> Lines; + CSVContent.split(Lines, '\n', -1, false); + if (Lines.size() < 2) // Need at least header + one data row + return Result; + + // Parse header to find column indices + llvm::SmallVector<llvm::StringRef, 32> Header; + Lines[0].split(Header, ','); + std::optional<unsigned> KeyIdx = findColumnIndex(Header, KeyColumn); + std::optional<unsigned> ValueIdx = findColumnIndex(Header, ValueColumn); + + if (!KeyIdx || !ValueIdx) + return Result; + + // Parse data rows and extract mappings + for (auto Line : llvm::drop_begin(Lines)) { + llvm::SmallVector<llvm::StringRef, 32> Row; + Line.split(Row, ','); + if (Row.size() <= std::max(*KeyIdx, *ValueIdx)) + continue; + + llvm::StringRef KeyVal = Row[*KeyIdx].trim().trim('"'); + llvm::StringRef ValueVal = Row[*ValueIdx].trim().trim('"'); + + if (!KeyVal.empty()) + Result[KeyVal] = ValueVal.str(); + } + + return Result; +} + +TEST(UnsignedStat, ExplicitlySetUnsignedStatistic) { + llvm::SmallString<128> TempMetricsCsvPath; + std::error_code EC = + llvm::sys::fs::createTemporaryFile("ep_stats", "csv", TempMetricsCsvPath); + ASSERT_FALSE(EC); + std::vector<std::string> Args = { + "-Xclang", "-analyzer-config", "-Xclang", + std::string("dump-entry-point-stats-to-csv=") + + TempMetricsCsvPath.str().str()}; + // Clean up on exit + auto Cleanup = llvm::make_scope_exit( + [&]() { llvm::sys::fs::remove(TempMetricsCsvPath); }); + EXPECT_TRUE(runCheckerOnCodeWithArgs<addUnsignedStatTesterChecker>( + R"cpp( + void func_one() {} + void func_two() {} + void func_none() {} + )cpp", + Args)); + + auto BufferOrError = llvm::MemoryBuffer::getFile(TempMetricsCsvPath); + ASSERT_TRUE(BufferOrError); + llvm::StringRef CSVContent = BufferOrError.get()->getBuffer(); + + // Parse the CSV and extract function statistics + llvm::StringMap<std::string> FunctionStats = + parseCSVColumnMapping(CSVContent, "DebugName", "DemoStat"); + + // Verify the expected values + ASSERT_TRUE(FunctionStats.count("func_one()")); + EXPECT_EQ(FunctionStats["func_one()"], "1"); + + ASSERT_TRUE(FunctionStats.count("func_two()")); + EXPECT_EQ(FunctionStats["func_two()"], "2"); + + ASSERT_TRUE(FunctionStats.count("func_none()")); + EXPECT_EQ(FunctionStats["func_none()"], ""); // Not set, should be empty +} +} // namespace diff --git a/clang/unittests/Tooling/LookupTest.cpp b/clang/unittests/Tooling/LookupTest.cpp index 4c49ebe..ed6f5d4 100644 --- a/clang/unittests/Tooling/LookupTest.cpp +++ b/clang/unittests/Tooling/LookupTest.cpp @@ -200,9 +200,9 @@ TEST(LookupTest, replaceNestedClassName) { Visitor.OnRecordTypeLoc = [&](RecordTypeLoc Type) { // Filter Types by name since there are other `RecordTypeLoc` in the test // file. - if (Type.getOriginalDecl()->getQualifiedNameAsString() == "a::b::Foo") { - EXPECT_EQ("x::Bar", replaceTypeLoc(Type.getOriginalDecl(), - Type.getBeginLoc(), "::a::x::Bar")); + if (Type.getDecl()->getQualifiedNameAsString() == "a::b::Foo") { + EXPECT_EQ("x::Bar", replaceTypeLoc(Type.getDecl(), Type.getBeginLoc(), + "::a::x::Bar")); } }; Visitor.runOver("namespace a { namespace b {\n" @@ -227,9 +227,9 @@ TEST(LookupTest, replaceNestedClassName) { // `x::y::Foo` in c.cc [1], it should not make "Foo" at [0] ambiguous because // it's not visible at [0]. Visitor.OnRecordTypeLoc = [&](RecordTypeLoc Type) { - if (Type.getOriginalDecl()->getQualifiedNameAsString() == "x::y::Old") { - EXPECT_EQ("Foo", replaceTypeLoc(Type.getOriginalDecl(), - Type.getBeginLoc(), "::x::Foo")); + if (Type.getDecl()->getQualifiedNameAsString() == "x::y::Old") { + EXPECT_EQ("Foo", + replaceTypeLoc(Type.getDecl(), Type.getBeginLoc(), "::x::Foo")); } }; Visitor.runOver(R"( diff --git a/clang/unittests/Tooling/RecursiveASTVisitorTests/MemberPointerTypeLoc.cpp b/clang/unittests/Tooling/RecursiveASTVisitorTests/MemberPointerTypeLoc.cpp index 88cebb7..587a00d 100644 --- a/clang/unittests/Tooling/RecursiveASTVisitorTests/MemberPointerTypeLoc.cpp +++ b/clang/unittests/Tooling/RecursiveASTVisitorTests/MemberPointerTypeLoc.cpp @@ -24,7 +24,7 @@ public: bool VisitRecordTypeLoc(RecordTypeLoc RTL) override { if (!RTL) return true; - Match(RTL.getOriginalDecl()->getName(), RTL.getNameLoc()); + Match(RTL.getDecl()->getName(), RTL.getNameLoc()); return true; } }; diff --git a/clang/unittests/Tooling/RecursiveASTVisitorTests/NestedNameSpecifiers.cpp b/clang/unittests/Tooling/RecursiveASTVisitorTests/NestedNameSpecifiers.cpp index 4181cd2..120b14b 100644 --- a/clang/unittests/Tooling/RecursiveASTVisitorTests/NestedNameSpecifiers.cpp +++ b/clang/unittests/Tooling/RecursiveASTVisitorTests/NestedNameSpecifiers.cpp @@ -18,7 +18,7 @@ public: bool VisitRecordTypeLoc(RecordTypeLoc RTL) override { if (!RTL) return true; - Match(RTL.getOriginalDecl()->getName(), RTL.getNameLoc()); + Match(RTL.getDecl()->getName(), RTL.getNameLoc()); return true; } diff --git a/clang/utils/TableGen/MveEmitter.cpp b/clang/utils/TableGen/MveEmitter.cpp index a003b5e..f55a5f5 100644 --- a/clang/utils/TableGen/MveEmitter.cpp +++ b/clang/utils/TableGen/MveEmitter.cpp @@ -1684,7 +1684,8 @@ void EmitterBase::EmitBuiltinCG(raw_ostream &OS) { OS << " case ARM::BI__builtin_arm_" << OI.Int->builtinExtension() << "_" << OI.Name << ":\n"; for (size_t i = 0, e = MG.ParamTypes.size(); i < e; ++i) - OS << " Param" << utostr(i) << " = " << OI.ParamValues[i] << ";\n"; + OS << " Param" << utostr(i) << " = static_cast<" + << MG.ParamTypes[i] << ">(" << OI.ParamValues[i] << ");\n"; OS << " break;\n"; } OS << " }\n"; |