diff options
Diffstat (limited to 'llvm')
329 files changed, 3303 insertions, 1287 deletions
diff --git a/llvm/docs/CommandGuide/llvm-config.rst b/llvm/docs/CommandGuide/llvm-config.rst index 63658d0..1c5c9c7 100644 --- a/llvm/docs/CommandGuide/llvm-config.rst +++ b/llvm/docs/CommandGuide/llvm-config.rst @@ -126,6 +126,11 @@ OPTIONS Print the installation prefix for LLVM. +**--quote-paths** + + Quote and escape paths when needed, most notably when a quote, space, backslash + or dollar sign characters are present in the path. + **--shared-mode** Print how the provided components can be collectively linked (`shared` or `static`). diff --git a/llvm/docs/LangRef.rst b/llvm/docs/LangRef.rst index 54c7d0f..3c089b5 100644 --- a/llvm/docs/LangRef.rst +++ b/llvm/docs/LangRef.rst @@ -668,7 +668,7 @@ representation is not just an integer address are called "non-integral". Non-integral pointers have at least one of the following three properties: * the pointer representation contains non-address bits -* the pointer representation is unstable (may changed at any time in a +* the pointer representation is unstable (may change at any time in a target-specific way) * the pointer representation has external state @@ -757,7 +757,7 @@ The following restrictions apply to IR level optimization passes: The ``inttoptr`` instruction does not recreate the external state and therefore it is target dependent whether it can be used to create a dereferenceable -pointer. In general passes should assume that the result of such an inttoptr +pointer. In general passes should assume that the result of such an ``inttoptr`` is not dereferenceable. For example, on CHERI targets an ``inttoptr`` will yield a capability with the external state (the validity tag bit) set to zero, which will cause any dereference to trap. @@ -784,7 +784,7 @@ be performed as loads and stores of the correct type since stores of other types may not propagate the external data. Therefore it is not legal to convert an existing load/store (or a ``llvm.memcpy`` / ``llvm.memmove`` intrinsic) of pointer types with external -state to a load/store of an integer type with same bitwidth, as that may drop +state to a load/store of an integer type with the same bitwidth, as that may drop the external state. @@ -806,7 +806,7 @@ Global variables can optionally specify a :ref:`linkage type <linkage>`. Either global variable definitions or declarations may have an explicit section to be placed in and may have an optional explicit alignment specified. If there is a mismatch between the explicit or inferred section information for the -variable declaration and its definition the resulting behavior is undefined. +variable declaration and its definition, the resulting behavior is undefined. A variable may be defined as a global ``constant``, which indicates that the contents of the variable will **never** be modified (enabling better @@ -1334,7 +1334,7 @@ Currently, only the following parameter attributes are defined: The byval type argument indicates the in-memory value type. The byval attribute also supports specifying an alignment with the - align attribute. It indicates the alignment of the stack slot to + ``align`` attribute. It indicates the alignment of the stack slot to form and the known alignment of the pointer specified to the call site. If the alignment is not specified, then the code generator makes a target-specific assumption. @@ -1355,7 +1355,7 @@ Currently, only the following parameter attributes are defined: This is not a valid attribute for return values. - The alignment for an ``byref`` parameter can be explicitly + The alignment for a ``byref`` parameter can be explicitly specified by combining it with the ``align`` attribute, similar to ``byval``. If the alignment is not specified, then the code generator makes a target-specific assumption. @@ -1382,7 +1382,7 @@ Currently, only the following parameter attributes are defined: The preallocated attribute requires a type argument. The preallocated attribute also supports specifying an alignment with the - align attribute. It indicates the alignment of the stack slot to + ``align`` attribute. It indicates the alignment of the stack slot to form and the known alignment of the pointer specified to the call site. If the alignment is not specified, then the code generator makes a target-specific assumption. @@ -1550,7 +1550,7 @@ Currently, only the following parameter attributes are defined: ``nonnull`` This indicates that the parameter or return pointer is not null. This - attribute may only be applied to pointer typed parameters. This is not + attribute may only be applied to pointer-typed parameters. This is not checked or enforced by LLVM; if the parameter or return pointer is null, :ref:`poison value <poisonvalues>` is returned or passed instead. The ``nonnull`` attribute should be combined with the ``noundef`` attribute @@ -1558,7 +1558,7 @@ Currently, only the following parameter attributes are defined: ``dereferenceable(<n>)`` This indicates that the parameter or return pointer is dereferenceable. This - attribute may only be applied to pointer typed parameters. A pointer that + attribute may only be applied to pointer-typed parameters. A pointer that is dereferenceable can be loaded from speculatively without a risk of trapping. The number of bytes known to be dereferenceable must be provided in parentheses. It is legal for the number of bytes to be less than the @@ -1584,7 +1584,7 @@ Currently, only the following parameter attributes are defined: implies that a pointer is at least one of ``dereferenceable(<n>)`` or ``null`` (i.e., it may be both ``null`` and ``dereferenceable(<n>)``). This attribute may only be applied to - pointer typed parameters. + pointer-typed parameters. ``swiftself`` This indicates that the parameter is the self/context parameter. This is not @@ -1601,7 +1601,7 @@ Currently, only the following parameter attributes are defined: ``swifterror`` This attribute is motivated to model and optimize Swift error handling. It - can be applied to a parameter with pointer to pointer type or a + can be applied to a parameter with pointer-to-pointer type or a pointer-sized alloca. At the call site, the actual argument that corresponds to a ``swifterror`` parameter has to come from a ``swifterror`` alloca or the ``swifterror`` parameter of the caller. A ``swifterror`` value (either diff --git a/llvm/docs/ProgrammersManual.rst b/llvm/docs/ProgrammersManual.rst index 7b7a1ce..270a635 100644 --- a/llvm/docs/ProgrammersManual.rst +++ b/llvm/docs/ProgrammersManual.rst @@ -113,7 +113,7 @@ rarely have to include this file directly). ``isa<>``: The ``isa<>`` operator works exactly like the Java "``instanceof``" operator. - It returns true or false depending on whether a reference or pointer points to + It returns ``true`` or ``false`` depending on whether a reference or pointer points to an instance of the specified class. This can be very useful for constraint checking of various sorts (example below). @@ -167,7 +167,7 @@ rarely have to include this file directly). ``isa_and_present<>``: The ``isa_and_present<>`` operator works just like the ``isa<>`` operator, except that it allows for a null pointer as an argument (which it then - returns false). This can sometimes be useful, allowing you to combine several + returns ``false``). This can sometimes be useful, allowing you to combine several null checks into one. ``cast_if_present<>``: @@ -402,7 +402,7 @@ doxygen documentation or by looking at the unit test suite. Error handling -------------- -Proper error handling helps us identify bugs in our code, and helps end-users +Proper error handling helps us identify bugs in our code, and helps end users understand errors in their tool usage. Errors fall into two broad categories: *programmatic* and *recoverable*, with different strategies for handling and reporting. @@ -449,10 +449,10 @@ violations even in builds that do not enable assertions: Recoverable Errors ^^^^^^^^^^^^^^^^^^ -Recoverable errors represent an error in the program's environment, for example +Recoverable errors represent an error in the program's environment, for example, a resource failure (a missing file, a dropped network connection, etc.), or malformed input. These errors should be detected and communicated to a level of -the program where they can be handled appropriately. Handling the error may be +the program that can handle them appropriately. Handling the error may be as simple as reporting the issue to the user, or it may involve attempts at recovery. @@ -668,7 +668,7 @@ Since the list of handlers passed to ``handleErrors`` may not cover every error type that can occur, the ``handleErrors`` function also returns an Error value that must be checked or propagated. If the error value that is passed to ``handleErrors`` does not match any of the handlers it will be returned from -handleErrors. Idiomatic use of ``handleErrors`` thus looks like: +``handleErrors``. Idiomatic use of ``handleErrors`` thus looks like: .. code-block:: c++ @@ -683,18 +683,18 @@ handleErrors. Idiomatic use of ``handleErrors`` thus looks like: })) return Err; -In cases where you truly know that the handler list is exhaustive the +In cases where you truly know that the handler list is exhaustive, the ``handleAllErrors`` function can be used instead. This is identical to ``handleErrors`` except that it will terminate the program if an unhandled error is passed in, and can therefore return void. The ``handleAllErrors`` function should generally be avoided: the introduction of a new error type elsewhere in the program can easily turn a formerly exhaustive list of errors into a non-exhaustive list, risking unexpected program termination. Where -possible, use handleErrors and propagate unknown errors up the stack instead. +possible, use ``handleErrors`` and propagate unknown errors up the stack instead. For tool code, where errors can be handled by printing an error message then exiting with an error code, the :ref:`ExitOnError <err_exitonerr>` utility -may be a better choice than handleErrors, as it simplifies control flow when +may be a better choice than ``handleErrors``, as it simplifies control flow when calling fallible functions. In situations where it is known that a particular call to a fallible function @@ -706,9 +706,9 @@ simplifying control flow. StringError """"""""""" -Many kinds of errors have no recovery strategy, the only action that can be +Many kinds of errors have no recovery strategy; the only action that can be taken is to report them to the user so that the user can attempt to fix the -environment. In this case representing the error as a string makes perfect +environment. In this case, representing the error as a string makes perfect sense. LLVM provides the ``StringError`` class for this purpose. It takes two arguments: A string error message, and an equivalent ``std::error_code`` for interoperability. It also provides a ``createStringError`` function to simplify @@ -721,7 +721,7 @@ common usage of this class: createStringError(errc::executable_format_error, "Bad executable"); If you're certain that the error you're building will never need to be converted -to a ``std::error_code`` you can use the ``inconvertibleErrorCode()`` function: +to a ``std::error_code``, you can use the ``inconvertibleErrorCode()`` function: .. code-block:: c++ @@ -791,18 +791,18 @@ actually recognises three different forms of handler signature: Error(std::unique_ptr<UserDefinedError> E); Any error returned from a handler will be returned from the ``handleErrors`` -function so that it can be handled itself, or propagated up the stack. +function so that it can be handled itself or propagated up the stack. .. _err_exitonerr: Using ExitOnError to simplify tool code """"""""""""""""""""""""""""""""""""""" -Library code should never call ``exit`` for a recoverable error, however in tool +Library code should never call ``exit`` for a recoverable error; however, in tool code (especially command line tools) this can be a reasonable approach. Calling ``exit`` upon encountering an error dramatically simplifies control flow as the error no longer needs to be propagated up the stack. This allows code to be -written in straight-line style, as long as each fallible call is wrapped in a +written in a straight-line style, as long as each fallible call is wrapped in a check and call to exit. The ``ExitOnError`` class supports this pattern by providing call operators that inspect ``Error`` values, stripping the error away in the success case and logging to ``stderr`` then exiting in the failure case. @@ -827,7 +827,7 @@ turning them into non-failing calls: } On failure, the error's log message will be written to ``stderr``, optionally -preceded by a string "banner" that can be set by calling the setBanner method. A +preceded by a string "banner" that can be set by calling the ``setBanner`` method. A mapping can also be supplied from ``Error`` values to exit codes using the ``setExitCodeMapper`` method: @@ -854,8 +854,8 @@ Some functions may only fail for a subset of their inputs, so calls using known safe inputs can be assumed to succeed. The cantFail functions encapsulate this by wrapping an assertion that their -argument is a success value and, in the case of Expected<T>, unwrapping the -T value: +argument is a success value and, in the case of ``Expected<T>``, unwrapping the +``T`` value: .. code-block:: c++ @@ -868,16 +868,16 @@ T value: ... } -Like the ExitOnError utility, cantFail simplifies control flow. Their treatment +Like the ExitOnError utility, ``cantFail`` simplifies control flow. Their treatment of error cases is very different, however: Where ExitOnError is guaranteed to -terminate the program on an error input, cantFail simply asserts that the result +terminate the program on an error input, ``cantFail`` simply asserts that the result is success. In debug builds this will result in an assertion failure if an error -is encountered. In release builds, the behavior of cantFail for failure values is -undefined. As such, care must be taken in the use of cantFail: clients must be -certain that a cantFail wrapped call really can not fail with the given +is encountered. In release builds, the behavior of ``cantFail`` for failure values is +undefined. As such, care must be taken in the use of ``cantFail``: clients must be +certain that a ``cantFail`` wrapped call really can not fail with the given arguments. -Use of the cantFail functions should be rare in library code, but they are +Use of the ``cantFail`` functions should be rare in library code, but they are likely to be of more use in tool and unit-test code where inputs and/or mocked-up classes or functions may be known to be safe. @@ -979,7 +979,7 @@ completing the walk over the archive they could use the ``joinErrors`` utility: } The ``joinErrors`` routine builds a special error type called ``ErrorList``, -which holds a list of user defined errors. The ``handleErrors`` routine +which holds a list of user-defined errors. The ``handleErrors`` routine recognizes this type and will attempt to handle each of the contained errors in order. If all contained errors can be handled, ``handleErrors`` will return ``Error::success()``; otherwise, ``handleErrors`` will concatenate the remaining @@ -1146,7 +1146,7 @@ be passed by value. The ``LDBG`` and ``LLVM_DEBUG()`` macros and ``-debug`` option -------------------------------------------------------------- -Often when working on your pass you will put a bunch of debugging printouts and +Often, when working on your pass, you will put a bunch of debugging printouts and other code into your pass. After you get it working, you want to remove it, but you may need it again in the future (to work out new bugs that you run across). @@ -1183,7 +1183,7 @@ The debug output can be enabled by passing the ``-debug`` command line argument. $ opt < a.bc > /dev/null -mypass -debug [my-pass MyPass.cpp:123 2] I am here! -While `LDBG()` is useful to add debug output to your code, there are cases +While ``LDBG()`` is useful to add debug output to your code, there are cases where you may need to guard a block of code with a debug check. The ``llvm/Support/Debug.h`` (`doxygen <https://llvm.org/doxygen/Debug_8h_source.html>`__) file provides a macro named @@ -1220,7 +1220,7 @@ with ``-debug``. Fine grained debug info with ``DEBUG_TYPE`` and the ``-debug-only`` option ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ -Sometimes you may find yourself in a situation where enabling ``-debug`` just +Sometimes, you may find yourself in a situation where enabling ``-debug`` just turns on **too much** information (such as when working on the code generator). If you want to enable debug information with more fine-grained control, you can control the debug type and level with associate with each logging statement @@ -1389,7 +1389,7 @@ maintainable and useful. Adding debug counters to aid in debugging your code --------------------------------------------------- -Sometimes, when writing new passes, or trying to track down bugs, it +Sometimes, when writing new passes or trying to track down bugs, it is useful to be able to control whether certain things in your pass happen or not. For example, there are times the minimization tooling can only easily give you large testcases. You would like to narrow @@ -1729,17 +1729,17 @@ page and one extra indirection when accessing elements with their positional index. In order to minimise the memory footprint of this container, it's important to -balance the ``PageSize`` so that it's not too small (otherwise the overhead of the -pointer per page might become too high) and not too big (otherwise the memory +balance the ``PageSize`` so that it's not too small (otherwise, the overhead of the +pointer per page might become too high) and not too big (otherwise, the memory is wasted if the page is not fully used). Moreover, while retaining the order of the elements based on their insertion index, like a vector, iterating over the elements via ``begin()`` and ``end()`` -is not provided in the API, due to the fact accessing the elements in order +is not provided in the API, due to the fact that accessing the elements in order would allocate all the iterated pages, defeating memory savings and the purpose of the ``PagedVector``. -Finally a ``materialized_begin()`` and ``materialized_end`` iterators are +Finally, ``materialized_begin()`` and ``materialized_end`` iterators are provided to access the elements associated to the accessed pages, which could speed up operations that need to iterate over initialized elements in a non-ordered manner. @@ -1782,9 +1782,9 @@ loop. ^^^^^^^ ``std::deque`` is, in some senses, a generalized version of ``std::vector``. -Like ``std::vector``, it provides constant time random access and other similar +Like ``std::vector``, it provides constant-time random access and other similar properties, but it also provides efficient access to the front of the list. It -does not guarantee continuity of elements within memory. +does not guarantee the continuity of elements within memory. In exchange for this extra flexibility, ``std::deque`` has significantly higher constant factor costs than ``std::vector``. If possible, use ``std::vector`` or @@ -1843,7 +1843,7 @@ Related classes of interest are explained in the following subsections: llvm/ADT/PackedVector.h ^^^^^^^^^^^^^^^^^^^^^^^ -Useful for storing a vector of values using only a few number of bits for each +Useful for storing a vector of values using only a few bits for each value. Apart from the standard operations of a vector-like container, it can also perform an 'or' set operation. @@ -1901,13 +1901,13 @@ non-empty ``ilist``\ s. The only sensible solution to this problem is to allocate a so-called *sentinel* along with the intrusive list, which serves as the ``end`` iterator, providing -the back-link to the last element. However conforming to the C++ convention it +the back-link to the last element. However, conforming to the C++ convention it is illegal to ``operator++`` beyond the sentinel and it also must not be dereferenced. These constraints allow for some implementation freedom to the ``ilist`` how to allocate and store the sentinel. The corresponding policy is dictated by -``ilist_traits<T>``. By default a ``T`` gets heap-allocated whenever the need +``ilist_traits<T>``. By default, a ``T`` gets heap-allocated whenever the need for a sentinel arises. While the default policy is sufficient in most cases, it may break down when @@ -1941,7 +1941,7 @@ String-like containers There are a variety of ways to pass around and use strings in C and C++, and LLVM adds a few new options to choose from. Pick the first option on this list -that will do what you need, they are ordered according to their relative cost. +that will do what you need; they are ordered according to their relative cost. Note that it is generally preferred to *not* pass strings around as ``const char*``'s. These have a number of problems, including the fact that they @@ -1973,12 +1973,12 @@ either because they are C string literals, ``std::string``, a C array, or a ``StringRef`` has a few major limitations which make more powerful string containers useful: -#. You cannot directly convert a ``StringRef`` to a 'const char*' because there is +#. You cannot directly convert a ``StringRef`` to a ``const char*`` because there is no way to add a trailing nul (unlike the ``.c_str()`` method on various stronger classes). #. ``StringRef`` doesn't own or keep alive the underlying string bytes. - As such it can easily lead to dangling pointers, and is not suitable for + As such, it can easily lead to dangling pointers, and is not suitable for embedding in datastructures in most cases (instead, use an ``std::string`` or something like that). diff --git a/llvm/docs/ReleaseNotes.md b/llvm/docs/ReleaseNotes.md index 49158fb..bfe6827 100644 --- a/llvm/docs/ReleaseNotes.md +++ b/llvm/docs/ReleaseNotes.md @@ -180,6 +180,10 @@ Changes to the LLVM tools * Some code paths for supporting Python 2.7 in `llvm-lit` have been removed. * Support for `%T` in lit has been removed. +* `llvm-config` gained a new flag `--quote-paths` which quotes and escapes paths + emitted on stdout, to account for spaces or other special characters in path. + (`#97305 <https://github.com/llvm/llvm-project/pull/97305>`_). + Changes to LLDB --------------------------------- diff --git a/llvm/examples/Kaleidoscope/Chapter9/toy.cpp b/llvm/examples/Kaleidoscope/Chapter9/toy.cpp index 51457a3..14081fb 100644 --- a/llvm/examples/Kaleidoscope/Chapter9/toy.cpp +++ b/llvm/examples/Kaleidoscope/Chapter9/toy.cpp @@ -203,7 +203,7 @@ class ExprAST { public: ExprAST(SourceLocation Loc = CurLoc) : Loc(Loc) {} - virtual ~ExprAST() {} + virtual ~ExprAST() = default; virtual Value *codegen() = 0; int getLine() const { return Loc.Line; } int getCol() const { return Loc.Col; } diff --git a/llvm/examples/OptSubcommand/llvm-hello-sub.cpp b/llvm/examples/OptSubcommand/llvm-hello-sub.cpp index 8071f56..8c0363f 100644 --- a/llvm/examples/OptSubcommand/llvm-hello-sub.cpp +++ b/llvm/examples/OptSubcommand/llvm-hello-sub.cpp @@ -46,7 +46,7 @@ public: HelloSubOptTable() : GenericOptTable(OptionStrTable, OptionPrefixesTable, InfoTable, /*IgnoreCase=*/false, OptionSubCommands, - OptionSubCommandIDsTable) {} + OptionSubCommandIDsTable) = default; }; } // namespace diff --git a/llvm/include/llvm/ADT/AddressRanges.h b/llvm/include/llvm/ADT/AddressRanges.h index 79ba5d5..6ea097d 100644 --- a/llvm/include/llvm/ADT/AddressRanges.h +++ b/llvm/include/llvm/ADT/AddressRanges.h @@ -21,7 +21,7 @@ namespace llvm { /// a start and an end address: [Start, End). class AddressRange { public: - AddressRange() {} + AddressRange() = default; AddressRange(uint64_t S, uint64_t E) : Start(S), End(E) { assert(Start <= End); } diff --git a/llvm/include/llvm/ADT/FloatingPointMode.h b/llvm/include/llvm/ADT/FloatingPointMode.h index 0314b4c..a9702c6 100644 --- a/llvm/include/llvm/ADT/FloatingPointMode.h +++ b/llvm/include/llvm/ADT/FloatingPointMode.h @@ -191,7 +191,7 @@ inline DenormalMode::DenormalModeKind parseDenormalFPAttributeComponent(StringRef Str) { // Assume ieee on unspecified attribute. return StringSwitch<DenormalMode::DenormalModeKind>(Str) - .Cases("", "ieee", DenormalMode::IEEE) + .Cases({"", "ieee"}, DenormalMode::IEEE) .Case("preserve-sign", DenormalMode::PreserveSign) .Case("positive-zero", DenormalMode::PositiveZero) .Case("dynamic", DenormalMode::Dynamic) diff --git a/llvm/include/llvm/ADT/StringMap.h b/llvm/include/llvm/ADT/StringMap.h index 01cbf2d3..7901365 100644 --- a/llvm/include/llvm/ADT/StringMap.h +++ b/llvm/include/llvm/ADT/StringMap.h @@ -302,7 +302,7 @@ public: if (FindInRHS == RHS.end()) return false; - if constexpr (!std::is_same_v<ValueTy, std::nullopt_t>) { + if constexpr (!std::is_same_v<ValueTy, EmptyStringSetTag>) { if (!(KeyValue.getValue() == FindInRHS->getValue())) return false; } diff --git a/llvm/include/llvm/ADT/StringMapEntry.h b/llvm/include/llvm/ADT/StringMapEntry.h index 21be5ec..b0a3c8c 100644 --- a/llvm/include/llvm/ADT/StringMapEntry.h +++ b/llvm/include/llvm/ADT/StringMapEntry.h @@ -21,6 +21,9 @@ namespace llvm { +/// The "value type" of StringSet represented as an empty struct. +struct EmptyStringSetTag {}; + /// StringMapEntryBase - Shared base class of StringMapEntry instances. class StringMapEntryBase { size_t keyLength; @@ -85,14 +88,13 @@ public: }; template <> -class StringMapEntryStorage<std::nullopt_t> : public StringMapEntryBase { +class StringMapEntryStorage<EmptyStringSetTag> : public StringMapEntryBase { public: - explicit StringMapEntryStorage(size_t keyLength, - std::nullopt_t = std::nullopt) + explicit StringMapEntryStorage(size_t keyLength, EmptyStringSetTag = {}) : StringMapEntryBase(keyLength) {} StringMapEntryStorage(StringMapEntryStorage &entry) = delete; - std::nullopt_t getValue() const { return std::nullopt; } + EmptyStringSetTag getValue() const { return {}; } }; /// StringMapEntry - This is used to represent one value that is inserted into diff --git a/llvm/include/llvm/ADT/StringSet.h b/llvm/include/llvm/ADT/StringSet.h index c8be3f2..dc154af 100644 --- a/llvm/include/llvm/ADT/StringSet.h +++ b/llvm/include/llvm/ADT/StringSet.h @@ -22,8 +22,8 @@ namespace llvm { /// StringSet - A wrapper for StringMap that provides set-like functionality. template <class AllocatorTy = MallocAllocator> -class StringSet : public StringMap<std::nullopt_t, AllocatorTy> { - using Base = StringMap<std::nullopt_t, AllocatorTy>; +class StringSet : public StringMap<EmptyStringSetTag, AllocatorTy> { + using Base = StringMap<EmptyStringSetTag, AllocatorTy>; public: StringSet() = default; diff --git a/llvm/include/llvm/ADT/StringSwitch.h b/llvm/include/llvm/ADT/StringSwitch.h index 98685de..8c8d31b 100644 --- a/llvm/include/llvm/ADT/StringSwitch.h +++ b/llvm/include/llvm/ADT/StringSwitch.h @@ -89,6 +89,7 @@ public: return CasesImpl(CaseStrings, Value); } + [[deprecated("Pass cases in std::initializer_list instead")]] StringSwitch &Cases(StringLiteral S0, StringLiteral S1, T Value) { return CasesImpl({S0, S1}, Value); } @@ -173,6 +174,7 @@ public: return CasesLowerImpl(CaseStrings, Value); } + [[deprecated("Pass cases in std::initializer_list instead")]] StringSwitch &CasesLower(StringLiteral S0, StringLiteral S1, T Value) { return CasesLowerImpl({S0, S1}, Value); } diff --git a/llvm/include/llvm/Analysis/AliasAnalysis.h b/llvm/include/llvm/Analysis/AliasAnalysis.h index 1681079..878b7e7 100644 --- a/llvm/include/llvm/Analysis/AliasAnalysis.h +++ b/llvm/include/llvm/Analysis/AliasAnalysis.h @@ -861,7 +861,7 @@ protected: // Provide all the copy and move constructors so that derived types aren't // constrained. - AAResultBase(const AAResultBase &Arg) {} + AAResultBase(const AAResultBase &Arg) = default; AAResultBase(AAResultBase &&Arg) {} public: diff --git a/llvm/include/llvm/Analysis/ConstantFolding.h b/llvm/include/llvm/Analysis/ConstantFolding.h index 5f91f97..ea22ed4 100644 --- a/llvm/include/llvm/Analysis/ConstantFolding.h +++ b/llvm/include/llvm/Analysis/ConstantFolding.h @@ -119,12 +119,6 @@ ConstantFoldFPInstOperands(unsigned Opcode, Constant *LHS, Constant *RHS, LLVM_ABI Constant *FlushFPConstant(Constant *Operand, const Instruction *I, bool IsOutput); -/// Attempt to constant fold a select instruction with the specified -/// operands. The constant result is returned if successful; if not, null is -/// returned. -LLVM_ABI Constant *ConstantFoldSelectInstruction(Constant *Cond, Constant *V1, - Constant *V2); - /// Attempt to constant fold a cast with the specified operand. If it /// fails, it returns a constant expression of the specified operand. LLVM_ABI Constant *ConstantFoldCastOperand(unsigned Opcode, Constant *C, @@ -135,40 +129,6 @@ LLVM_ABI Constant *ConstantFoldCastOperand(unsigned Opcode, Constant *C, LLVM_ABI Constant *ConstantFoldIntegerCast(Constant *C, Type *DestTy, bool IsSigned, const DataLayout &DL); -/// ConstantFoldInsertValueInstruction - Attempt to constant fold an insertvalue -/// instruction with the specified operands and indices. The constant result is -/// returned if successful; if not, null is returned. -LLVM_ABI Constant *ConstantFoldInsertValueInstruction(Constant *Agg, - Constant *Val, - ArrayRef<unsigned> Idxs); - -/// Attempt to constant fold an extractvalue instruction with the -/// specified operands and indices. The constant result is returned if -/// successful; if not, null is returned. -LLVM_ABI Constant *ConstantFoldExtractValueInstruction(Constant *Agg, - ArrayRef<unsigned> Idxs); - -/// Attempt to constant fold an insertelement instruction with the -/// specified operands and indices. The constant result is returned if -/// successful; if not, null is returned. -LLVM_ABI Constant *ConstantFoldInsertElementInstruction(Constant *Val, - Constant *Elt, - Constant *Idx); - -/// Attempt to constant fold an extractelement instruction with the -/// specified operands and indices. The constant result is returned if -/// successful; if not, null is returned. -LLVM_ABI Constant *ConstantFoldExtractElementInstruction(Constant *Val, - Constant *Idx); - -/// Attempt to constant fold a shufflevector instruction with the -/// specified operands and mask. See class ShuffleVectorInst for a description -/// of the mask representation. The constant result is returned if successful; -/// if not, null is returned. -LLVM_ABI Constant *ConstantFoldShuffleVectorInstruction(Constant *V1, - Constant *V2, - ArrayRef<int> Mask); - /// Extract value of C at the given Offset reinterpreted as Ty. If bits past /// the end of C are accessed, they are assumed to be poison. LLVM_ABI Constant *ConstantFoldLoadFromConst(Constant *C, Type *Ty, diff --git a/llvm/include/llvm/Analysis/ConstraintSystem.h b/llvm/include/llvm/Analysis/ConstraintSystem.h index 307ad50..1d9ac49 100644 --- a/llvm/include/llvm/Analysis/ConstraintSystem.h +++ b/llvm/include/llvm/Analysis/ConstraintSystem.h @@ -64,7 +64,7 @@ class ConstraintSystem { SmallVector<std::string> getVarNamesList() const; public: - ConstraintSystem() {} + ConstraintSystem() = default; ConstraintSystem(ArrayRef<Value *> FunctionArgs) { NumVariables += FunctionArgs.size(); for (auto *Arg : FunctionArgs) { diff --git a/llvm/include/llvm/Analysis/DDG.h b/llvm/include/llvm/Analysis/DDG.h index 1c53291..120bb46 100644 --- a/llvm/include/llvm/Analysis/DDG.h +++ b/llvm/include/llvm/Analysis/DDG.h @@ -60,11 +60,7 @@ public: DDGNode(DDGNode &&N) : DDGNodeBase(std::move(N)), Kind(N.Kind) {} virtual ~DDGNode() = 0; - DDGNode &operator=(const DDGNode &N) { - DGNode::operator=(N); - Kind = N.Kind; - return *this; - } + DDGNode &operator=(const DDGNode &N) = default; DDGNode &operator=(DDGNode &&N) { DGNode::operator=(std::move(N)); diff --git a/llvm/include/llvm/Analysis/DOTGraphTraitsPass.h b/llvm/include/llvm/Analysis/DOTGraphTraitsPass.h index ba5ee1d..19a202f 100644 --- a/llvm/include/llvm/Analysis/DOTGraphTraitsPass.h +++ b/llvm/include/llvm/Analysis/DOTGraphTraitsPass.h @@ -80,7 +80,7 @@ protected: /// virtual destructor needed. Making this dtor protected stops accidental /// invocation when the derived class destructor should have been called. /// Those derived classes sould be marked final to avoid the warning. - ~DOTGraphTraitsViewer() {} + ~DOTGraphTraitsViewer() = default; private: StringRef Name; @@ -161,7 +161,7 @@ protected: /// virtual destructor needed. Making this dtor protected stops accidental /// invocation when the derived class destructor should have been called. /// Those derived classes sould be marked final to avoid the warning. - ~DOTGraphTraitsPrinter() {} + ~DOTGraphTraitsPrinter() = default; private: StringRef Name; diff --git a/llvm/include/llvm/Analysis/IR2Vec.h b/llvm/include/llvm/Analysis/IR2Vec.h index 71055dd16..7a68773 100644 --- a/llvm/include/llvm/Analysis/IR2Vec.h +++ b/llvm/include/llvm/Analysis/IR2Vec.h @@ -72,7 +72,7 @@ enum class IR2VecKind { Symbolic, FlowAware }; namespace ir2vec { -extern llvm::cl::OptionCategory IR2VecCategory; +LLVM_ABI extern llvm::cl::OptionCategory IR2VecCategory; LLVM_ABI extern cl::opt<float> OpcWeight; LLVM_ABI extern cl::opt<float> TypeWeight; LLVM_ABI extern cl::opt<float> ArgWeight; @@ -110,8 +110,8 @@ public: return Data[Itr]; } - using iterator = typename std::vector<double>::iterator; - using const_iterator = typename std::vector<double>::const_iterator; + using iterator = std::vector<double>::iterator; + using const_iterator = std::vector<double>::const_iterator; iterator begin() { return Data.begin(); } iterator end() { return Data.end(); } diff --git a/llvm/include/llvm/Analysis/LoopIterator.h b/llvm/include/llvm/Analysis/LoopIterator.h index 523d2a2..1ac8e68 100644 --- a/llvm/include/llvm/Analysis/LoopIterator.h +++ b/llvm/include/llvm/Analysis/LoopIterator.h @@ -45,12 +45,12 @@ struct LoopBodyTraits { class WrappedSuccIterator : public iterator_adaptor_base< WrappedSuccIterator, succ_iterator, - typename std::iterator_traits<succ_iterator>::iterator_category, - NodeRef, std::ptrdiff_t, NodeRef *, NodeRef> { + std::iterator_traits<succ_iterator>::iterator_category, NodeRef, + std::ptrdiff_t, NodeRef *, NodeRef> { using BaseT = iterator_adaptor_base< WrappedSuccIterator, succ_iterator, - typename std::iterator_traits<succ_iterator>::iterator_category, - NodeRef, std::ptrdiff_t, NodeRef *, NodeRef>; + std::iterator_traits<succ_iterator>::iterator_category, NodeRef, + std::ptrdiff_t, NodeRef *, NodeRef>; const Loop *L; diff --git a/llvm/include/llvm/Analysis/MemorySSA.h b/llvm/include/llvm/Analysis/MemorySSA.h index cbb942f..07d39ab 100644 --- a/llvm/include/llvm/Analysis/MemorySSA.h +++ b/llvm/include/llvm/Analysis/MemorySSA.h @@ -1247,7 +1247,7 @@ public: return DefIterator == Other.DefIterator; } - typename std::iterator_traits<BaseT>::reference operator*() const { + std::iterator_traits<BaseT>::reference operator*() const { assert(DefIterator != OriginalAccess->defs_end() && "Tried to access past the end of our iterator"); return CurrentPair; diff --git a/llvm/include/llvm/Analysis/TargetFolder.h b/llvm/include/llvm/Analysis/TargetFolder.h index d27455c..cbce482 100644 --- a/llvm/include/llvm/Analysis/TargetFolder.h +++ b/llvm/include/llvm/Analysis/TargetFolder.h @@ -20,6 +20,7 @@ #include "llvm/ADT/ArrayRef.h" #include "llvm/Analysis/ConstantFolding.h" +#include "llvm/IR/ConstantFold.h" #include "llvm/IR/Constants.h" #include "llvm/IR/IRBuilderFolder.h" #include "llvm/IR/Operator.h" diff --git a/llvm/include/llvm/Analysis/TargetTransformInfo.h b/llvm/include/llvm/Analysis/TargetTransformInfo.h index 7b7dc1b..0f17312 100644 --- a/llvm/include/llvm/Analysis/TargetTransformInfo.h +++ b/llvm/include/llvm/Analysis/TargetTransformInfo.h @@ -1764,7 +1764,7 @@ public: /// \param Types List of types to check. LLVM_ABI bool areTypesABICompatible(const Function *Caller, const Function *Callee, - const ArrayRef<Type *> &Types) const; + ArrayRef<Type *> Types) const; /// The type of load/store indexing. enum MemIndexedMode { diff --git a/llvm/include/llvm/Analysis/TargetTransformInfoImpl.h b/llvm/include/llvm/Analysis/TargetTransformInfoImpl.h index 4cd607c..aacb88d 100644 --- a/llvm/include/llvm/Analysis/TargetTransformInfoImpl.h +++ b/llvm/include/llvm/Analysis/TargetTransformInfoImpl.h @@ -1028,7 +1028,7 @@ public: virtual bool areTypesABICompatible(const Function *Caller, const Function *Callee, - const ArrayRef<Type *> &Types) const { + ArrayRef<Type *> Types) const { return (Caller->getFnAttribute("target-cpu") == Callee->getFnAttribute("target-cpu")) && (Caller->getFnAttribute("target-features") == diff --git a/llvm/include/llvm/CodeGen/BasicBlockSectionsProfileReader.h b/llvm/include/llvm/CodeGen/BasicBlockSectionsProfileReader.h index 8237530..7b1a5f5 100644 --- a/llvm/include/llvm/CodeGen/BasicBlockSectionsProfileReader.h +++ b/llvm/include/llvm/CodeGen/BasicBlockSectionsProfileReader.h @@ -66,7 +66,7 @@ public: BasicBlockSectionsProfileReader(const MemoryBuffer *Buf) : MBuf(Buf), LineIt(*Buf, /*SkipBlanks=*/true, /*CommentMarker=*/'#'){}; - BasicBlockSectionsProfileReader(){}; + BasicBlockSectionsProfileReader() = default; // Returns true if basic block sections profile exist for function \p // FuncName. diff --git a/llvm/include/llvm/CodeGen/DIE.h b/llvm/include/llvm/CodeGen/DIE.h index 32f4651..92265fd 100644 --- a/llvm/include/llvm/CodeGen/DIE.h +++ b/llvm/include/llvm/CodeGen/DIE.h @@ -653,7 +653,7 @@ public: public: const_iterator() = default; // Placate MSVC by explicitly scoping 'iterator'. - const_iterator(typename IntrusiveBackList<T>::iterator X) : N(X.N) {} + const_iterator(IntrusiveBackList<T>::iterator X) : N(X.N) {} explicit const_iterator(const T *N) : N(N) {} const_iterator &operator++() { diff --git a/llvm/include/llvm/CodeGen/GlobalISel/CombinerHelper.h b/llvm/include/llvm/CodeGen/GlobalISel/CombinerHelper.h index 36cb90b..96cb7cd 100644 --- a/llvm/include/llvm/CodeGen/GlobalISel/CombinerHelper.h +++ b/llvm/include/llvm/CodeGen/GlobalISel/CombinerHelper.h @@ -293,7 +293,7 @@ public: SmallVectorImpl<Register> &Ops) const; /// Replace \p MI with a concat_vectors with \p Ops. void applyCombineShuffleVector(MachineInstr &MI, - const ArrayRef<Register> Ops) const; + ArrayRef<Register> Ops) const; /// Optimize memcpy intrinsics et al, e.g. constant len calls. /// /p MaxLen if non-zero specifies the max length of a mem libcall to inline. diff --git a/llvm/include/llvm/CodeGen/GlobalMergeFunctions.h b/llvm/include/llvm/CodeGen/GlobalMergeFunctions.h index caea5b6..54ea68a 100644 --- a/llvm/include/llvm/CodeGen/GlobalMergeFunctions.h +++ b/llvm/include/llvm/CodeGen/GlobalMergeFunctions.h @@ -58,7 +58,7 @@ public: /// The suffix used to identify the merged function that parameterizes /// the constant values. Note that the original function, without this suffix, /// becomes a thunk supplying contexts to the merged function via parameters. - static constexpr const char MergingInstanceSuffix[] = ".Tgm"; + static constexpr char MergingInstanceSuffix[] = ".Tgm"; GlobalMergeFunc(const ModuleSummaryIndex *Index) : Index(Index) {}; diff --git a/llvm/include/llvm/CodeGen/MIR2Vec.h b/llvm/include/llvm/CodeGen/MIR2Vec.h index 44f009c..18b1290 100644 --- a/llvm/include/llvm/CodeGen/MIR2Vec.h +++ b/llvm/include/llvm/CodeGen/MIR2Vec.h @@ -73,7 +73,7 @@ namespace mir2vec { class MIREmbedder; class SymbolicMIREmbedder; -extern llvm::cl::OptionCategory MIR2VecCategory; +LLVM_ABI extern llvm::cl::OptionCategory MIR2VecCategory; extern cl::opt<float> OpcWeight, CommonOperandWeight, RegOperandWeight; using Embedding = ir2vec::Embedding; @@ -154,14 +154,14 @@ class MIRVocabulary { void buildRegisterOperandMapping(); /// Get canonical index for a machine opcode - unsigned getCanonicalOpcodeIndex(unsigned Opcode) const; + LLVM_ABI unsigned getCanonicalOpcodeIndex(unsigned Opcode) const; /// Get index for a common (non-register) machine operand unsigned getCommonOperandIndex(MachineOperand::MachineOperandType OperandType) const; /// Get index for a register machine operand - unsigned getRegisterOperandIndex(Register Reg) const; + LLVM_ABI unsigned getRegisterOperandIndex(Register Reg) const; // Accessors for operand types const Embedding & @@ -192,7 +192,7 @@ class MIRVocabulary { /// Get entity ID (flat index) for a common operand type /// This is used for triplet generation - unsigned getEntityIDForCommonOperand( + LLVM_ABI unsigned getEntityIDForCommonOperand( MachineOperand::MachineOperandType OperandType) const { return Layout.CommonOperandBase + getCommonOperandIndex(OperandType); } @@ -221,7 +221,7 @@ public: bool IsPhysical = true) const; /// Get the string key for a vocabulary entry at the given position - std::string getStringKey(unsigned Pos) const; + LLVM_ABI std::string getStringKey(unsigned Pos) const; unsigned getDimension() const { return Storage.getDimension(); } @@ -268,7 +268,7 @@ public: const TargetRegisterInfo &TRI, const MachineRegisterInfo &MRI); /// Create a dummy vocabulary for testing purposes. - static Expected<MIRVocabulary> + LLVM_ABI static Expected<MIRVocabulary> createDummyVocabForTest(const TargetInstrInfo &TII, const TargetRegisterInfo &TRI, const MachineRegisterInfo &MRI, unsigned Dim = 1); @@ -302,10 +302,10 @@ protected: RegOperandWeight(mir2vec::RegOperandWeight) {} /// Function to compute embeddings. - Embedding computeEmbeddings() const; + LLVM_ABI Embedding computeEmbeddings() const; /// Function to compute the embedding for a given machine basic block. - Embedding computeEmbeddings(const MachineBasicBlock &MBB) const; + LLVM_ABI Embedding computeEmbeddings(const MachineBasicBlock &MBB) const; /// Function to compute the embedding for a given machine instruction. /// Specific to the kind of embeddings being computed. @@ -316,9 +316,9 @@ public: /// Factory method to create an Embedder object of the specified kind /// Returns nullptr if the requested kind is not supported. - static std::unique_ptr<MIREmbedder> create(MIR2VecKind Mode, - const MachineFunction &MF, - const MIRVocabulary &Vocab); + LLVM_ABI static std::unique_ptr<MIREmbedder> + create(MIR2VecKind Mode, const MachineFunction &MF, + const MIRVocabulary &Vocab); /// Computes and returns the embedding for a given machine instruction MI in /// the machine function MF. @@ -369,7 +369,7 @@ class MIR2VecVocabProvider { public: MIR2VecVocabProvider(const MachineModuleInfo &MMI) : MMI(MMI) {} - Expected<mir2vec::MIRVocabulary> getVocabulary(const Module &M); + LLVM_ABI Expected<mir2vec::MIRVocabulary> getVocabulary(const Module &M); private: Error readVocabulary(VocabMap &OpcVocab, VocabMap &CommonOperandVocab, @@ -454,7 +454,7 @@ public: }; /// Create a machine pass that prints MIR2Vec embeddings -MachineFunctionPass *createMIR2VecPrinterLegacyPass(raw_ostream &OS); +LLVM_ABI MachineFunctionPass *createMIR2VecPrinterLegacyPass(raw_ostream &OS); } // namespace llvm diff --git a/llvm/include/llvm/CodeGen/MachineScheduler.h b/llvm/include/llvm/CodeGen/MachineScheduler.h index 5a2aee2..6c5c27c 100644 --- a/llvm/include/llvm/CodeGen/MachineScheduler.h +++ b/llvm/include/llvm/CodeGen/MachineScheduler.h @@ -829,7 +829,7 @@ private: public: // constructor for empty set - explicit ResourceSegments(){}; + explicit ResourceSegments() = default; bool empty() const { return _Intervals.empty(); } explicit ResourceSegments(const std::list<IntervalTy> &Intervals) : _Intervals(Intervals) { diff --git a/llvm/include/llvm/CodeGen/RDFRegisters.h b/llvm/include/llvm/CodeGen/RDFRegisters.h index 82027ca..3b7454e 100644 --- a/llvm/include/llvm/CodeGen/RDFRegisters.h +++ b/llvm/include/llvm/CodeGen/RDFRegisters.h @@ -294,7 +294,7 @@ struct RegisterAggr { ref_iterator ref_begin() const { return ref_iterator(*this, false); } ref_iterator ref_end() const { return ref_iterator(*this, true); } - using unit_iterator = typename BitVector::const_set_bits_iterator; + using unit_iterator = BitVector::const_set_bits_iterator; unit_iterator unit_begin() const { return Units.set_bits_begin(); } unit_iterator unit_end() const { return Units.set_bits_end(); } diff --git a/llvm/include/llvm/CodeGen/RegAllocRegistry.h b/llvm/include/llvm/CodeGen/RegAllocRegistry.h index cd81e08..db62640 100644 --- a/llvm/include/llvm/CodeGen/RegAllocRegistry.h +++ b/llvm/include/llvm/CodeGen/RegAllocRegistry.h @@ -67,7 +67,7 @@ public: /// RegisterRegAlloc's global Registry tracks allocator registration. template <class T> MachinePassRegistry<typename RegisterRegAllocBase<T>::FunctionPassCtor> -RegisterRegAllocBase<T>::Registry; + RegisterRegAllocBase<T>::Registry; } // end namespace llvm diff --git a/llvm/include/llvm/CodeGen/SDPatternMatch.h b/llvm/include/llvm/CodeGen/SDPatternMatch.h index 9a6bf5f..511cb56 100644 --- a/llvm/include/llvm/CodeGen/SDPatternMatch.h +++ b/llvm/include/llvm/CodeGen/SDPatternMatch.h @@ -1311,7 +1311,7 @@ template <typename... PatternTs> struct ReassociatableOpc_match { } [[nodiscard]] inline bool - reassociatableMatchHelper(const ArrayRef<SmallBitVector> Matches, + reassociatableMatchHelper(ArrayRef<SmallBitVector> Matches, SmallBitVector &Used, size_t Curr = 0) { if (Curr == Matches.size()) return true; diff --git a/llvm/include/llvm/CodeGen/WindowScheduler.h b/llvm/include/llvm/CodeGen/WindowScheduler.h index 476d5ad..97776de 100644 --- a/llvm/include/llvm/CodeGen/WindowScheduler.h +++ b/llvm/include/llvm/CodeGen/WindowScheduler.h @@ -105,7 +105,7 @@ protected: public: WindowScheduler(MachineSchedContext *C, MachineLoop &ML); - virtual ~WindowScheduler() {} + virtual ~WindowScheduler() = default; bool run(); diff --git a/llvm/include/llvm/CodeGenTypes/LowLevelType.h b/llvm/include/llvm/CodeGenTypes/LowLevelType.h index 4c1fe13..472a3f3 100644 --- a/llvm/include/llvm/CodeGenTypes/LowLevelType.h +++ b/llvm/include/llvm/CodeGenTypes/LowLevelType.h @@ -340,18 +340,18 @@ private: /// valid encodings, SizeInBits/SizeOfElement must be larger than 0. /// * Non-pointer scalar (isPointer == 0 && isVector == 0): /// SizeInBits: 32; - static const constexpr BitFieldInfo ScalarSizeFieldInfo{32, 29}; + static constexpr BitFieldInfo ScalarSizeFieldInfo{32, 29}; /// * Pointer (isPointer == 1 && isVector == 0): /// SizeInBits: 16; /// AddressSpace: 24; - static const constexpr BitFieldInfo PointerSizeFieldInfo{16, 45}; - static const constexpr BitFieldInfo PointerAddressSpaceFieldInfo{24, 21}; + static constexpr BitFieldInfo PointerSizeFieldInfo{16, 45}; + static constexpr BitFieldInfo PointerAddressSpaceFieldInfo{24, 21}; /// * Vector-of-non-pointer (isPointer == 0 && isVector == 1): /// NumElements: 16; /// SizeOfElement: 32; /// Scalable: 1; - static const constexpr BitFieldInfo VectorElementsFieldInfo{16, 5}; - static const constexpr BitFieldInfo VectorScalableFieldInfo{1, 0}; + static constexpr BitFieldInfo VectorElementsFieldInfo{16, 5}; + static constexpr BitFieldInfo VectorScalableFieldInfo{1, 0}; /// * Vector-of-pointer (isPointer == 1 && isVector == 1): /// NumElements: 16; /// SizeOfElement: 16; diff --git a/llvm/include/llvm/DWARFLinker/StringPool.h b/llvm/include/llvm/DWARFLinker/StringPool.h index d0f4e21..7838e3b 100644 --- a/llvm/include/llvm/DWARFLinker/StringPool.h +++ b/llvm/include/llvm/DWARFLinker/StringPool.h @@ -20,7 +20,7 @@ namespace dwarf_linker { /// StringEntry keeps data of the string: the length, external offset /// and a string body which is placed right after StringEntry. -using StringEntry = StringMapEntry<std::nullopt_t>; +using StringEntry = StringMapEntry<EmptyStringSetTag>; class StringPoolEntryInfo { public: diff --git a/llvm/include/llvm/DebugInfo/DWARF/DWARFUnit.h b/llvm/include/llvm/DebugInfo/DWARF/DWARFUnit.h index be78647..b7d6e72 100644 --- a/llvm/include/llvm/DebugInfo/DWARF/DWARFUnit.h +++ b/llvm/include/llvm/DebugInfo/DWARF/DWARFUnit.h @@ -136,8 +136,8 @@ class DWARFUnitVector final : public SmallVector<std::unique_ptr<DWARFUnit>, 1> public: using UnitVector = SmallVectorImpl<std::unique_ptr<DWARFUnit>>; - using iterator = typename UnitVector::iterator; - using iterator_range = llvm::iterator_range<typename UnitVector::iterator>; + using iterator = UnitVector::iterator; + using iterator_range = llvm::iterator_range<UnitVector::iterator>; using compile_unit_range = decltype(make_filter_range(std::declval<iterator_range>(), isCompileUnit)); diff --git a/llvm/include/llvm/DebugInfo/PDB/Native/DbiModuleList.h b/llvm/include/llvm/DebugInfo/PDB/Native/DbiModuleList.h index 8992fae..bbed56b 100644 --- a/llvm/include/llvm/DebugInfo/PDB/Native/DbiModuleList.h +++ b/llvm/include/llvm/DebugInfo/PDB/Native/DbiModuleList.h @@ -32,7 +32,7 @@ struct FileInfoSubstreamHeader; class DbiModuleSourceFilesIterator : public iterator_facade_base<DbiModuleSourceFilesIterator, std::random_access_iterator_tag, StringRef> { - using BaseType = typename DbiModuleSourceFilesIterator::iterator_facade_base; + using BaseType = DbiModuleSourceFilesIterator::iterator_facade_base; public: LLVM_ABI DbiModuleSourceFilesIterator(const DbiModuleList &Modules, diff --git a/llvm/include/llvm/Demangle/Utility.h b/llvm/include/llvm/Demangle/Utility.h index 002a1f5..6e6203d 100644 --- a/llvm/include/llvm/Demangle/Utility.h +++ b/llvm/include/llvm/Demangle/Utility.h @@ -81,7 +81,7 @@ public: OutputBuffer(const OutputBuffer &) = delete; OutputBuffer &operator=(const OutputBuffer &) = delete; - virtual ~OutputBuffer() {} + virtual ~OutputBuffer() = default; operator std::string_view() const { return std::string_view(Buffer, CurrentPosition); diff --git a/llvm/include/llvm/ExecutionEngine/JITLink/aarch32.h b/llvm/include/llvm/ExecutionEngine/JITLink/aarch32.h index 98170f6..9479c10 100644 --- a/llvm/include/llvm/ExecutionEngine/JITLink/aarch32.h +++ b/llvm/include/llvm/ExecutionEngine/JITLink/aarch32.h @@ -175,7 +175,7 @@ struct HalfWords { /// FixupInfo base class is required for dynamic lookups. struct FixupInfoBase { LLVM_ABI static const FixupInfoBase *getDynFixupInfo(Edge::Kind K); - virtual ~FixupInfoBase() {} + virtual ~FixupInfoBase() = default; }; /// FixupInfo checks for Arm edge kinds work on 32-bit words diff --git a/llvm/include/llvm/ExecutionEngine/Orc/MachOBuilder.h b/llvm/include/llvm/ExecutionEngine/Orc/MachOBuilder.h index dd41025..1296e24 100644 --- a/llvm/include/llvm/ExecutionEngine/Orc/MachOBuilder.h +++ b/llvm/include/llvm/ExecutionEngine/Orc/MachOBuilder.h @@ -36,7 +36,7 @@ size_t writeMachOStruct(MutableArrayRef<char> Buf, size_t Offset, MachOStruct S, /// Base type for MachOBuilder load command wrappers. struct MachOBuilderLoadCommandBase { - virtual ~MachOBuilderLoadCommandBase() {} + virtual ~MachOBuilderLoadCommandBase() = default; virtual size_t size() const = 0; virtual size_t write(MutableArrayRef<char> Buf, size_t Offset, bool SwapStruct) = 0; diff --git a/llvm/include/llvm/ExecutionEngine/Orc/TargetProcess/ExecutorSharedMemoryMapperService.h b/llvm/include/llvm/ExecutionEngine/Orc/TargetProcess/ExecutorSharedMemoryMapperService.h index 2c385de..8f87650 100644 --- a/llvm/include/llvm/ExecutionEngine/Orc/TargetProcess/ExecutorSharedMemoryMapperService.h +++ b/llvm/include/llvm/ExecutionEngine/Orc/TargetProcess/ExecutorSharedMemoryMapperService.h @@ -29,7 +29,7 @@ namespace rt_bootstrap { class LLVM_ABI ExecutorSharedMemoryMapperService final : public ExecutorBootstrapService { public: - ~ExecutorSharedMemoryMapperService() override {}; + ~ExecutorSharedMemoryMapperService() override = default; Expected<std::pair<ExecutorAddr, std::string>> reserve(uint64_t Size); Expected<ExecutorAddr> initialize(ExecutorAddr Reservation, diff --git a/llvm/include/llvm/ExecutionEngine/Orc/TargetProcess/LibraryResolver.h b/llvm/include/llvm/ExecutionEngine/Orc/TargetProcess/LibraryResolver.h index 9182995..7cc78d4 100644 --- a/llvm/include/llvm/ExecutionEngine/Orc/TargetProcess/LibraryResolver.h +++ b/llvm/include/llvm/ExecutionEngine/Orc/TargetProcess/LibraryResolver.h @@ -100,7 +100,7 @@ public: class FilteredView { public: using Map = StringMap<std::shared_ptr<LibraryInfo>>; - using Iterator = typename Map::const_iterator; + using Iterator = Map::const_iterator; class FilterIterator { public: FilterIterator(Iterator it_, Iterator end_, LibState S, PathType K) diff --git a/llvm/include/llvm/Frontend/OpenMP/OMPIRBuilder.h b/llvm/include/llvm/Frontend/OpenMP/OMPIRBuilder.h index 5331cb5..b3d7ab4 100644 --- a/llvm/include/llvm/Frontend/OpenMP/OMPIRBuilder.h +++ b/llvm/include/llvm/Frontend/OpenMP/OMPIRBuilder.h @@ -2383,7 +2383,7 @@ public: /// runtime library for debugging Value *MapNamesArray = nullptr; - explicit TargetDataRTArgs() {} + explicit TargetDataRTArgs() = default; explicit TargetDataRTArgs(Value *BasePointersArray, Value *PointersArray, Value *SizesArray, Value *MapTypesArray, Value *MapTypesArrayEnd, Value *MappersArray, @@ -2451,7 +2451,7 @@ public: bool HasNoWait = false; // Constructors for TargetKernelArgs. - TargetKernelArgs() {} + TargetKernelArgs() = default; TargetKernelArgs(unsigned NumTargetItems, TargetDataRTArgs RTArgs, Value *NumIterations, ArrayRef<Value *> NumTeams, ArrayRef<Value *> NumThreads, Value *DynCGGroupMem, @@ -2494,7 +2494,7 @@ public: /// Whether the `target ... data` directive has a `nowait` clause. bool HasNoWait = false; - explicit TargetDataInfo() {} + explicit TargetDataInfo() = default; explicit TargetDataInfo(bool RequiresDevicePointerInfo, bool SeparateBeginEndCalls) : RequiresDevicePointerInfo(RequiresDevicePointerInfo), diff --git a/llvm/include/llvm/IR/ConstantFold.h b/llvm/include/llvm/IR/ConstantFold.h index f9f2b35..4056f1f 100644 --- a/llvm/include/llvm/IR/ConstantFold.h +++ b/llvm/include/llvm/IR/ConstantFold.h @@ -26,42 +26,66 @@ #include <optional> namespace llvm { - template <typename T> class ArrayRef; - class Value; - class Constant; - class Type; +template <typename T> class ArrayRef; +class Value; +class Constant; +class Type; - // Constant fold various types of instruction... - LLVM_ABI Constant * - ConstantFoldCastInstruction(unsigned opcode, ///< The opcode of the cast - Constant *V, ///< The source constant - Type *DestTy ///< The destination type - ); - LLVM_ABI Constant *ConstantFoldSelectInstruction(Constant *Cond, Constant *V1, - Constant *V2); - LLVM_ABI Constant *ConstantFoldExtractElementInstruction(Constant *Val, - Constant *Idx); - LLVM_ABI Constant *ConstantFoldInsertElementInstruction(Constant *Val, - Constant *Elt, - Constant *Idx); - LLVM_ABI Constant *ConstantFoldShuffleVectorInstruction(Constant *V1, - Constant *V2, - ArrayRef<int> Mask); - LLVM_ABI Constant * - ConstantFoldExtractValueInstruction(Constant *Agg, ArrayRef<unsigned> Idxs); - LLVM_ABI Constant * - ConstantFoldInsertValueInstruction(Constant *Agg, Constant *Val, - ArrayRef<unsigned> Idxs); - LLVM_ABI Constant *ConstantFoldUnaryInstruction(unsigned Opcode, Constant *V); - LLVM_ABI Constant *ConstantFoldBinaryInstruction(unsigned Opcode, - Constant *V1, Constant *V2); - LLVM_ABI Constant * - ConstantFoldCompareInstruction(CmpInst::Predicate Predicate, Constant *C1, - Constant *C2); - LLVM_ABI Constant * - ConstantFoldGetElementPtr(Type *Ty, Constant *C, - std::optional<ConstantRange> InRange, - ArrayRef<Value *> Idxs); -} // End llvm namespace +// Constant fold various types of instruction... +LLVM_ABI Constant * +ConstantFoldCastInstruction(unsigned opcode, ///< The opcode of the cast + Constant *V, ///< The source constant + Type *DestTy ///< The destination type +); + +/// Attempt to constant fold a select instruction with the specified +/// operands. The constant result is returned if successful; if not, null is +/// returned. +LLVM_ABI Constant *ConstantFoldSelectInstruction(Constant *Cond, Constant *V1, + Constant *V2); + +/// Attempt to constant fold an extractelement instruction with the +/// specified operands and indices. The constant result is returned if +/// successful; if not, null is returned. +LLVM_ABI Constant *ConstantFoldExtractElementInstruction(Constant *Val, + Constant *Idx); + +/// Attempt to constant fold an insertelement instruction with the +/// specified operands and indices. The constant result is returned if +/// successful; if not, null is returned. +LLVM_ABI Constant *ConstantFoldInsertElementInstruction(Constant *Val, + Constant *Elt, + Constant *Idx); + +/// Attempt to constant fold a shufflevector instruction with the +/// specified operands and mask. See class ShuffleVectorInst for a description +/// of the mask representation. The constant result is returned if successful; +/// if not, null is returned. +LLVM_ABI Constant *ConstantFoldShuffleVectorInstruction(Constant *V1, + Constant *V2, + ArrayRef<int> Mask); + +/// Attempt to constant fold an extractvalue instruction with the +/// specified operands and indices. The constant result is returned if +/// successful; if not, null is returned. +LLVM_ABI Constant *ConstantFoldExtractValueInstruction(Constant *Agg, + ArrayRef<unsigned> Idxs); + +/// Attempt to constant fold an insertvalue instruction with the specified +/// operands and indices. The constant result is returned if successful; if +/// not, null is returned. +LLVM_ABI Constant *ConstantFoldInsertValueInstruction(Constant *Agg, + Constant *Val, + ArrayRef<unsigned> Idxs); +LLVM_ABI Constant *ConstantFoldUnaryInstruction(unsigned Opcode, Constant *V); +LLVM_ABI Constant *ConstantFoldBinaryInstruction(unsigned Opcode, Constant *V1, + Constant *V2); +LLVM_ABI Constant *ConstantFoldCompareInstruction(CmpInst::Predicate Predicate, + Constant *C1, Constant *C2); +LLVM_ABI Constant * +ConstantFoldGetElementPtr(Type *Ty, Constant *C, + std::optional<ConstantRange> InRange, + ArrayRef<Value *> Idxs); +} // namespace llvm #endif diff --git a/llvm/include/llvm/IR/DataLayout.h b/llvm/include/llvm/IR/DataLayout.h index 56fc749..5445820 100644 --- a/llvm/include/llvm/IR/DataLayout.h +++ b/llvm/include/llvm/IR/DataLayout.h @@ -590,7 +590,7 @@ public: /// /// This is the amount that alloca reserves for this type. For example, /// returns 12 or 16 for x86_fp80, depending on alignment. - TypeSize getTypeAllocSize(Type *Ty) const; + LLVM_ABI TypeSize getTypeAllocSize(Type *Ty) const; /// Returns the offset in bits between successive objects of the /// specified type, including alignment padding; always a multiple of 8. diff --git a/llvm/include/llvm/IR/DebugProgramInstruction.h b/llvm/include/llvm/IR/DebugProgramInstruction.h index 457c60e3b..66f44fe 100644 --- a/llvm/include/llvm/IR/DebugProgramInstruction.h +++ b/llvm/include/llvm/IR/DebugProgramInstruction.h @@ -589,7 +589,7 @@ filterDbgVars(iterator_range<simple_ilist<DbgRecord>::iterator> R) { /// date. class DbgMarker { public: - DbgMarker() {} + DbgMarker() = default; /// Link back to the Instruction that owns this marker. Can be null during /// operations that move a marker from one instruction to another. Instruction *MarkedInstr = nullptr; diff --git a/llvm/include/llvm/IR/DroppedVariableStats.h b/llvm/include/llvm/IR/DroppedVariableStats.h index 42e86dd..8a1dbd6a 100644 --- a/llvm/include/llvm/IR/DroppedVariableStats.h +++ b/llvm/include/llvm/IR/DroppedVariableStats.h @@ -42,7 +42,7 @@ class DroppedVariableStats { public: LLVM_ABI DroppedVariableStats(bool DroppedVarStatsEnabled); - virtual ~DroppedVariableStats() {} + virtual ~DroppedVariableStats() = default; // We intend this to be unique per-compilation, thus no copies. DroppedVariableStats(const DroppedVariableStats &) = delete; diff --git a/llvm/include/llvm/IR/IntrinsicsHexagonDep.td b/llvm/include/llvm/IR/IntrinsicsHexagonDep.td index fe95377..dde4132 100644 --- a/llvm/include/llvm/IR/IntrinsicsHexagonDep.td +++ b/llvm/include/llvm/IR/IntrinsicsHexagonDep.td @@ -6835,6 +6835,180 @@ Hexagon_v64i32_v32i32v32i32_Intrinsic<"HEXAGON_V6_vsub_hf_f8_128B">; // V81 HVX Instructions. +def int_hexagon_V6_vabs_qf16_hf : +Hexagon_v16i32_v16i32_Intrinsic<"HEXAGON_V6_vabs_qf16_hf">; + +def int_hexagon_V6_vabs_qf16_hf_128B : +Hexagon_v32i32_v32i32_Intrinsic<"HEXAGON_V6_vabs_qf16_hf_128B">; + +def int_hexagon_V6_vabs_qf16_qf16 : +Hexagon_v16i32_v16i32_Intrinsic<"HEXAGON_V6_vabs_qf16_qf16">; + +def int_hexagon_V6_vabs_qf16_qf16_128B : +Hexagon_v32i32_v32i32_Intrinsic<"HEXAGON_V6_vabs_qf16_qf16_128B">; + +def int_hexagon_V6_vabs_qf32_qf32 : +Hexagon_v16i32_v16i32_Intrinsic<"HEXAGON_V6_vabs_qf32_qf32">; + +def int_hexagon_V6_vabs_qf32_qf32_128B : +Hexagon_v32i32_v32i32_Intrinsic<"HEXAGON_V6_vabs_qf32_qf32_128B">; + +def int_hexagon_V6_vabs_qf32_sf : +Hexagon_v16i32_v16i32_Intrinsic<"HEXAGON_V6_vabs_qf32_sf">; + +def int_hexagon_V6_vabs_qf32_sf_128B : +Hexagon_v32i32_v32i32_Intrinsic<"HEXAGON_V6_vabs_qf32_sf_128B">; + +def int_hexagon_V6_valign4 : +Hexagon_v16i32_v16i32v16i32i32_Intrinsic<"HEXAGON_V6_valign4">; + +def int_hexagon_V6_valign4_128B : +Hexagon_v32i32_v32i32v32i32i32_Intrinsic<"HEXAGON_V6_valign4_128B">; + +def int_hexagon_V6_vconv_bf_qf32 : +Hexagon_v16i32_v32i32_Intrinsic<"HEXAGON_V6_vconv_bf_qf32">; + +def int_hexagon_V6_vconv_bf_qf32_128B : +Hexagon_v32i32_v64i32_Intrinsic<"HEXAGON_V6_vconv_bf_qf32_128B">; + +def int_hexagon_V6_vconv_f8_qf16 : +Hexagon_v16i32_v16i32_Intrinsic<"HEXAGON_V6_vconv_f8_qf16">; + +def int_hexagon_V6_vconv_f8_qf16_128B : +Hexagon_v32i32_v32i32_Intrinsic<"HEXAGON_V6_vconv_f8_qf16_128B">; + +def int_hexagon_V6_vconv_h_hf_rnd : +Hexagon_v16i32_v16i32_Intrinsic<"HEXAGON_V6_vconv_h_hf_rnd">; + +def int_hexagon_V6_vconv_h_hf_rnd_128B : +Hexagon_v32i32_v32i32_Intrinsic<"HEXAGON_V6_vconv_h_hf_rnd_128B">; + +def int_hexagon_V6_vconv_qf16_f8 : +Hexagon_v32i32_v16i32_Intrinsic<"HEXAGON_V6_vconv_qf16_f8">; + +def int_hexagon_V6_vconv_qf16_f8_128B : +Hexagon_v64i32_v32i32_Intrinsic<"HEXAGON_V6_vconv_qf16_f8_128B">; + +def int_hexagon_V6_vconv_qf16_hf : +Hexagon_v16i32_v16i32_Intrinsic<"HEXAGON_V6_vconv_qf16_hf">; + +def int_hexagon_V6_vconv_qf16_hf_128B : +Hexagon_v32i32_v32i32_Intrinsic<"HEXAGON_V6_vconv_qf16_hf_128B">; + +def int_hexagon_V6_vconv_qf16_qf16 : +Hexagon_v16i32_v16i32_Intrinsic<"HEXAGON_V6_vconv_qf16_qf16">; + +def int_hexagon_V6_vconv_qf16_qf16_128B : +Hexagon_v32i32_v32i32_Intrinsic<"HEXAGON_V6_vconv_qf16_qf16_128B">; + +def int_hexagon_V6_vconv_qf32_qf32 : +Hexagon_v16i32_v16i32_Intrinsic<"HEXAGON_V6_vconv_qf32_qf32">; + +def int_hexagon_V6_vconv_qf32_qf32_128B : +Hexagon_v32i32_v32i32_Intrinsic<"HEXAGON_V6_vconv_qf32_qf32_128B">; + +def int_hexagon_V6_vconv_qf32_sf : +Hexagon_v16i32_v16i32_Intrinsic<"HEXAGON_V6_vconv_qf32_sf">; + +def int_hexagon_V6_vconv_qf32_sf_128B : +Hexagon_v32i32_v32i32_Intrinsic<"HEXAGON_V6_vconv_qf32_sf_128B">; + +def int_hexagon_V6_veqhf : +Hexagon_v64i1_v16i32v16i32_Intrinsic<"HEXAGON_V6_veqhf">; + +def int_hexagon_V6_veqhf_128B : +Hexagon_v128i1_v32i32v32i32_Intrinsic<"HEXAGON_V6_veqhf_128B">; + +def int_hexagon_V6_veqhf_and : +Hexagon_v64i1_v64i1v16i32v16i32_Intrinsic<"HEXAGON_V6_veqhf_and">; + +def int_hexagon_V6_veqhf_and_128B : +Hexagon_v128i1_v128i1v32i32v32i32_Intrinsic<"HEXAGON_V6_veqhf_and_128B">; + +def int_hexagon_V6_veqhf_or : +Hexagon_v64i1_v64i1v16i32v16i32_Intrinsic<"HEXAGON_V6_veqhf_or">; + +def int_hexagon_V6_veqhf_or_128B : +Hexagon_v128i1_v128i1v32i32v32i32_Intrinsic<"HEXAGON_V6_veqhf_or_128B">; + +def int_hexagon_V6_veqhf_xor : +Hexagon_v64i1_v64i1v16i32v16i32_Intrinsic<"HEXAGON_V6_veqhf_xor">; + +def int_hexagon_V6_veqhf_xor_128B : +Hexagon_v128i1_v128i1v32i32v32i32_Intrinsic<"HEXAGON_V6_veqhf_xor_128B">; + +def int_hexagon_V6_veqsf : +Hexagon_v64i1_v16i32v16i32_Intrinsic<"HEXAGON_V6_veqsf">; + +def int_hexagon_V6_veqsf_128B : +Hexagon_v128i1_v32i32v32i32_Intrinsic<"HEXAGON_V6_veqsf_128B">; + +def int_hexagon_V6_veqsf_and : +Hexagon_v64i1_v64i1v16i32v16i32_Intrinsic<"HEXAGON_V6_veqsf_and">; + +def int_hexagon_V6_veqsf_and_128B : +Hexagon_v128i1_v128i1v32i32v32i32_Intrinsic<"HEXAGON_V6_veqsf_and_128B">; + +def int_hexagon_V6_veqsf_or : +Hexagon_v64i1_v64i1v16i32v16i32_Intrinsic<"HEXAGON_V6_veqsf_or">; + +def int_hexagon_V6_veqsf_or_128B : +Hexagon_v128i1_v128i1v32i32v32i32_Intrinsic<"HEXAGON_V6_veqsf_or_128B">; + +def int_hexagon_V6_veqsf_xor : +Hexagon_v64i1_v64i1v16i32v16i32_Intrinsic<"HEXAGON_V6_veqsf_xor">; + +def int_hexagon_V6_veqsf_xor_128B : +Hexagon_v128i1_v128i1v32i32v32i32_Intrinsic<"HEXAGON_V6_veqsf_xor_128B">; + +def int_hexagon_V6_vilog2_hf : +Hexagon_v16i32_v16i32_Intrinsic<"HEXAGON_V6_vilog2_hf">; + +def int_hexagon_V6_vilog2_hf_128B : +Hexagon_v32i32_v32i32_Intrinsic<"HEXAGON_V6_vilog2_hf_128B">; + +def int_hexagon_V6_vilog2_qf16 : +Hexagon_v16i32_v16i32_Intrinsic<"HEXAGON_V6_vilog2_qf16">; + +def int_hexagon_V6_vilog2_qf16_128B : +Hexagon_v32i32_v32i32_Intrinsic<"HEXAGON_V6_vilog2_qf16_128B">; + +def int_hexagon_V6_vilog2_qf32 : +Hexagon_v16i32_v16i32_Intrinsic<"HEXAGON_V6_vilog2_qf32">; + +def int_hexagon_V6_vilog2_qf32_128B : +Hexagon_v32i32_v32i32_Intrinsic<"HEXAGON_V6_vilog2_qf32_128B">; + +def int_hexagon_V6_vilog2_sf : +Hexagon_v16i32_v16i32_Intrinsic<"HEXAGON_V6_vilog2_sf">; + +def int_hexagon_V6_vilog2_sf_128B : +Hexagon_v32i32_v32i32_Intrinsic<"HEXAGON_V6_vilog2_sf_128B">; + +def int_hexagon_V6_vneg_qf16_hf : +Hexagon_v16i32_v16i32_Intrinsic<"HEXAGON_V6_vneg_qf16_hf">; + +def int_hexagon_V6_vneg_qf16_hf_128B : +Hexagon_v32i32_v32i32_Intrinsic<"HEXAGON_V6_vneg_qf16_hf_128B">; + +def int_hexagon_V6_vneg_qf16_qf16 : +Hexagon_v16i32_v16i32_Intrinsic<"HEXAGON_V6_vneg_qf16_qf16">; + +def int_hexagon_V6_vneg_qf16_qf16_128B : +Hexagon_v32i32_v32i32_Intrinsic<"HEXAGON_V6_vneg_qf16_qf16_128B">; + +def int_hexagon_V6_vneg_qf32_qf32 : +Hexagon_v16i32_v16i32_Intrinsic<"HEXAGON_V6_vneg_qf32_qf32">; + +def int_hexagon_V6_vneg_qf32_qf32_128B : +Hexagon_v32i32_v32i32_Intrinsic<"HEXAGON_V6_vneg_qf32_qf32_128B">; + +def int_hexagon_V6_vneg_qf32_sf : +Hexagon_v16i32_v16i32_Intrinsic<"HEXAGON_V6_vneg_qf32_sf">; + +def int_hexagon_V6_vneg_qf32_sf_128B : +Hexagon_v32i32_v32i32_Intrinsic<"HEXAGON_V6_vneg_qf32_sf_128B">; + def int_hexagon_V6_vsub_hf_mix : Hexagon_v16i32_v16i32v16i32_Intrinsic<"HEXAGON_V6_vsub_hf_mix">; diff --git a/llvm/include/llvm/IR/IntrinsicsNVVM.td b/llvm/include/llvm/IR/IntrinsicsNVVM.td index 719181a..2710853 100644 --- a/llvm/include/llvm/IR/IntrinsicsNVVM.td +++ b/llvm/include/llvm/IR/IntrinsicsNVVM.td @@ -1334,15 +1334,8 @@ let TargetPrefix = "nvvm" in { // let IntrProperties = [IntrNoMem] in { foreach ftz = ["", "_ftz"] in - def int_nvvm_ex2_approx # ftz # _f : NVVMBuiltin, - DefaultAttrsIntrinsic<[llvm_float_ty], [llvm_float_ty]>; - - def int_nvvm_ex2_approx_d : NVVMBuiltin, - DefaultAttrsIntrinsic<[llvm_double_ty], [llvm_double_ty]>; - def int_nvvm_ex2_approx_f16 : - DefaultAttrsIntrinsic<[llvm_half_ty], [llvm_half_ty]>; - def int_nvvm_ex2_approx_f16x2 : - DefaultAttrsIntrinsic<[llvm_v2f16_ty], [llvm_v2f16_ty]>; + def int_nvvm_ex2_approx # ftz : + DefaultAttrsIntrinsic<[llvm_anyfloat_ty], [LLVMMatchType<0>]>; foreach ftz = ["", "_ftz"] in def int_nvvm_lg2_approx # ftz # _f : NVVMBuiltin, diff --git a/llvm/include/llvm/IR/ModuleSummaryIndexYAML.h b/llvm/include/llvm/IR/ModuleSummaryIndexYAML.h index 3381e17..ccb77e7 100644 --- a/llvm/include/llvm/IR/ModuleSummaryIndexYAML.h +++ b/llvm/include/llvm/IR/ModuleSummaryIndexYAML.h @@ -79,7 +79,7 @@ struct CustomMappingTraits< } Args.push_back(Arg); } - io.mapRequired(Key.str().c_str(), V[Args]); + io.mapRequired(Key, V[Args]); } static void output( IO &io, @@ -91,7 +91,7 @@ struct CustomMappingTraits< Key += ','; Key += llvm::utostr(Arg); } - io.mapRequired(Key.c_str(), P.second); + io.mapRequired(Key, P.second); } } }; @@ -122,11 +122,11 @@ struct CustomMappingTraits<std::map<uint64_t, WholeProgramDevirtResolution>> { io.setError("key not an integer"); return; } - io.mapRequired(Key.str().c_str(), V[KeyInt]); + io.mapRequired(Key, V[KeyInt]); } static void output(IO &io, std::map<uint64_t, WholeProgramDevirtResolution> &V) { for (auto &P : V) - io.mapRequired(llvm::utostr(P.first).c_str(), P.second); + io.mapRequired(llvm::utostr(P.first), P.second); } }; @@ -215,7 +215,7 @@ namespace yaml { template <> struct CustomMappingTraits<GlobalValueSummaryMapTy> { static void inputOne(IO &io, StringRef Key, GlobalValueSummaryMapTy &V) { std::vector<GlobalValueSummaryYaml> GVSums; - io.mapRequired(Key.str().c_str(), GVSums); + io.mapRequired(Key, GVSums); uint64_t KeyInt; if (Key.getAsInteger(0, KeyInt)) { io.setError("key not an integer"); @@ -290,7 +290,7 @@ template <> struct CustomMappingTraits<GlobalValueSummaryMapTy> { } } if (!GVSums.empty()) - io.mapRequired(llvm::utostr(P.first).c_str(), GVSums); + io.mapRequired(llvm::utostr(P.first), GVSums); } } static void fixAliaseeLinks(GlobalValueSummaryMapTy &V) { @@ -313,12 +313,12 @@ template <> struct CustomMappingTraits<GlobalValueSummaryMapTy> { template <> struct CustomMappingTraits<TypeIdSummaryMapTy> { static void inputOne(IO &io, StringRef Key, TypeIdSummaryMapTy &V) { TypeIdSummary TId; - io.mapRequired(Key.str().c_str(), TId); + io.mapRequired(Key, TId); V.insert({GlobalValue::getGUIDAssumingExternalLinkage(Key), {Key, TId}}); } static void output(IO &io, TypeIdSummaryMapTy &V) { for (auto &TidIter : V) - io.mapRequired(TidIter.second.first.str().c_str(), TidIter.second.second); + io.mapRequired(TidIter.second.first, TidIter.second.second); } }; diff --git a/llvm/include/llvm/IR/TrackingMDRef.h b/llvm/include/llvm/IR/TrackingMDRef.h index d737739..7ad7225 100644 --- a/llvm/include/llvm/IR/TrackingMDRef.h +++ b/llvm/include/llvm/IR/TrackingMDRef.h @@ -111,17 +111,14 @@ public: explicit TypedTrackingMDRef(T *MD) : Ref(static_cast<Metadata *>(MD)) {} TypedTrackingMDRef(TypedTrackingMDRef &&X) : Ref(std::move(X.Ref)) {} - TypedTrackingMDRef(const TypedTrackingMDRef &X) : Ref(X.Ref) {} + TypedTrackingMDRef(const TypedTrackingMDRef &X) = default; TypedTrackingMDRef &operator=(TypedTrackingMDRef &&X) { Ref = std::move(X.Ref); return *this; } - TypedTrackingMDRef &operator=(const TypedTrackingMDRef &X) { - Ref = X.Ref; - return *this; - } + TypedTrackingMDRef &operator=(const TypedTrackingMDRef &X) = default; T *get() const { return (T *)Ref.get(); } operator T *() const { return get(); } diff --git a/llvm/include/llvm/MC/MCAssembler.h b/llvm/include/llvm/MC/MCAssembler.h index 6e1d642..bbb8bee 100644 --- a/llvm/include/llvm/MC/MCAssembler.h +++ b/llvm/include/llvm/MC/MCAssembler.h @@ -198,8 +198,8 @@ public: const_iterator end() const { return Sections.end(); } SmallVectorImpl<const MCSymbol *> &getSymbols() { return Symbols; } - iterator_range<pointee_iterator< - typename SmallVector<const MCSymbol *, 0>::const_iterator>> + iterator_range< + pointee_iterator<SmallVector<const MCSymbol *, 0>::const_iterator>> symbols() const { return make_pointee_range(Symbols); } diff --git a/llvm/include/llvm/MC/MCRegisterInfo.h b/llvm/include/llvm/MC/MCRegisterInfo.h index e6fc707..f611edd 100644 --- a/llvm/include/llvm/MC/MCRegisterInfo.h +++ b/llvm/include/llvm/MC/MCRegisterInfo.h @@ -272,7 +272,7 @@ public: friend class MCRegUnitRootIterator; friend class MCRegAliasIterator; - virtual ~MCRegisterInfo() {} + virtual ~MCRegisterInfo() = default; /// Initialize MCRegisterInfo, called by TableGen /// auto-generated routines. *DO NOT USE*. diff --git a/llvm/include/llvm/MCA/SourceMgr.h b/llvm/include/llvm/MCA/SourceMgr.h index 16a60d1..300961c 100644 --- a/llvm/include/llvm/MCA/SourceMgr.h +++ b/llvm/include/llvm/MCA/SourceMgr.h @@ -50,7 +50,7 @@ struct SourceMgr { /// Advance to the next \a SourceRef. virtual void updateNext() = 0; - virtual ~SourceMgr() {} + virtual ~SourceMgr() = default; }; /// The default implementation of \a SourceMgr. It always takes a fixed number diff --git a/llvm/include/llvm/ObjCopy/ConfigManager.h b/llvm/include/llvm/ObjCopy/ConfigManager.h index 1568799..45f847f 100644 --- a/llvm/include/llvm/ObjCopy/ConfigManager.h +++ b/llvm/include/llvm/ObjCopy/ConfigManager.h @@ -23,7 +23,7 @@ namespace llvm { namespace objcopy { struct LLVM_ABI ConfigManager : public MultiFormatConfig { - ~ConfigManager() override {} + ~ConfigManager() override = default; const CommonConfig &getCommonConfig() const override { return Common; } diff --git a/llvm/include/llvm/ObjCopy/MultiFormatConfig.h b/llvm/include/llvm/ObjCopy/MultiFormatConfig.h index bb93f64..91baf9b 100644 --- a/llvm/include/llvm/ObjCopy/MultiFormatConfig.h +++ b/llvm/include/llvm/ObjCopy/MultiFormatConfig.h @@ -24,7 +24,7 @@ struct DXContainerConfig; class MultiFormatConfig { public: - virtual ~MultiFormatConfig() {} + virtual ~MultiFormatConfig() = default; virtual const CommonConfig &getCommonConfig() const = 0; virtual Expected<const ELFConfig &> getELFConfig() const = 0; diff --git a/llvm/include/llvm/Object/ELF.h b/llvm/include/llvm/Object/ELF.h index 03d5ee2..cc1e5f9 100644 --- a/llvm/include/llvm/Object/ELF.h +++ b/llvm/include/llvm/Object/ELF.h @@ -261,6 +261,8 @@ public: ELFFile(const ELFFile &) = default; ELFFile &operator=(const ELFFile &) = default; + ELFFile(ELFFile &&) = default; + // This is a callback that can be passed to a number of functions. // It can be used to ignore non-critical errors (warnings), which is // useful for dumpers, like llvm-readobj. diff --git a/llvm/include/llvm/Object/ELFObjectFile.h b/llvm/include/llvm/Object/ELFObjectFile.h index ced1afd..ca41357 100644 --- a/llvm/include/llvm/Object/ELFObjectFile.h +++ b/llvm/include/llvm/Object/ELFObjectFile.h @@ -1218,12 +1218,12 @@ ELFObjectFile<ELFT>::ELFObjectFile(MemoryBufferRef Object, ELFFile<ELFT> EF, : ELFObjectFileBase(getELFType(ELFT::Endianness == llvm::endianness::little, ELFT::Is64Bits), Object), - EF(EF), DotDynSymSec(DotDynSymSec), DotSymtabSec(DotSymtabSec), + EF(std::move(EF)), DotDynSymSec(DotDynSymSec), DotSymtabSec(DotSymtabSec), DotSymtabShndxSec(DotSymtabShndx) {} template <class ELFT> ELFObjectFile<ELFT>::ELFObjectFile(ELFObjectFile<ELFT> &&Other) - : ELFObjectFile(Other.Data, Other.EF, Other.DotDynSymSec, + : ELFObjectFile(Other.Data, std::move(Other.EF), Other.DotDynSymSec, Other.DotSymtabSec, Other.DotSymtabShndxSec) {} template <class ELFT> diff --git a/llvm/include/llvm/Object/SFrameParser.h b/llvm/include/llvm/Object/SFrameParser.h index 3ce5d70..23298357 100644 --- a/llvm/include/llvm/Object/SFrameParser.h +++ b/llvm/include/llvm/Object/SFrameParser.h @@ -90,7 +90,7 @@ public: uint32_t Idx, uint32_t Size, uint64_t Offset) : Data(Data), FREType(FREType), Idx(Idx), Size(Size), Offset(Offset) {} - Error inc(); + LLVM_ABI Error inc(); const FrameRowEntry &operator*() const { return FRE; } friend bool operator==(const FallibleFREIterator &LHS, diff --git a/llvm/include/llvm/ObjectYAML/DXContainerYAML.h b/llvm/include/llvm/ObjectYAML/DXContainerYAML.h index b5b110d..fbfe306 100644 --- a/llvm/include/llvm/ObjectYAML/DXContainerYAML.h +++ b/llvm/include/llvm/ObjectYAML/DXContainerYAML.h @@ -115,7 +115,7 @@ struct RootParameterHeaderYaml { dxbc::ShaderVisibility Visibility; uint32_t Offset; - RootParameterHeaderYaml(){}; + RootParameterHeaderYaml() = default; RootParameterHeaderYaml(dxbc::RootParameterType T) : Type(T) {} }; @@ -123,7 +123,7 @@ struct RootParameterLocationYaml { RootParameterHeaderYaml Header; std::optional<size_t> IndexInSignature; - RootParameterLocationYaml(){}; + RootParameterLocationYaml() = default; explicit RootParameterLocationYaml(RootParameterHeaderYaml Header) : Header(Header) {} }; diff --git a/llvm/include/llvm/ProfileData/DataAccessProf.h b/llvm/include/llvm/ProfileData/DataAccessProf.h index 608306f..ea256ef 100644 --- a/llvm/include/llvm/ProfileData/DataAccessProf.h +++ b/llvm/include/llvm/ProfileData/DataAccessProf.h @@ -42,7 +42,7 @@ struct SourceLocation { : FileName(FileNameRef.str()), Line(Line) {} // Empty constructor is used in yaml conversion. - SourceLocation() {} + SourceLocation() = default; /// The filename where the data is located. std::string FileName; /// The line number in the source code. diff --git a/llvm/include/llvm/ProfileData/MemProfYAML.h b/llvm/include/llvm/ProfileData/MemProfYAML.h index d66e16d..c55f780 100644 --- a/llvm/include/llvm/ProfileData/MemProfYAML.h +++ b/llvm/include/llvm/ProfileData/MemProfYAML.h @@ -141,7 +141,7 @@ template <> struct CustomMappingTraits<memprof::PortableMemInfoBlock> { #define MIBEntryDef(NameTag, Name, Type) \ if (KeyStr == #Name) { \ uint64_t Value; \ - Io.mapRequired(KeyStr.str().c_str(), Value); \ + Io.mapRequired(KeyStr, Value); \ MIB.Name = static_cast<Type>(Value); \ MIB.Schema.set(llvm::to_underlying(memprof::Meta::Name)); \ return; \ diff --git a/llvm/include/llvm/SandboxIR/Context.h b/llvm/include/llvm/SandboxIR/Context.h index 7d8b2c8..a8966db 100644 --- a/llvm/include/llvm/SandboxIR/Context.h +++ b/llvm/include/llvm/SandboxIR/Context.h @@ -51,7 +51,7 @@ public: // Uses a 64-bit integer so we don't have to worry about the unlikely case // of overflowing a 32-bit counter. using ValTy = uint64_t; - static constexpr const ValTy InvalidVal = 0; + static constexpr ValTy InvalidVal = 0; private: // Default initialization results in an invalid ID. diff --git a/llvm/include/llvm/SandboxIR/Instruction.h b/llvm/include/llvm/SandboxIR/Instruction.h index e1c1ca0..5e369a4 100644 --- a/llvm/include/llvm/SandboxIR/Instruction.h +++ b/llvm/include/llvm/SandboxIR/Instruction.h @@ -1866,7 +1866,7 @@ class SwitchInst : public SingleLLVMInstructionImpl<llvm::SwitchInst> { friend class Context; // For accessing the constructor in create*() public: - static constexpr const unsigned DefaultPseudoIndex = + static constexpr unsigned DefaultPseudoIndex = llvm::SwitchInst::DefaultPseudoIndex; LLVM_ABI static SwitchInst *create(Value *V, BasicBlock *Dest, diff --git a/llvm/include/llvm/SandboxIR/Pass.h b/llvm/include/llvm/SandboxIR/Pass.h index 267389a..eb84f21 100644 --- a/llvm/include/llvm/SandboxIR/Pass.h +++ b/llvm/include/llvm/SandboxIR/Pass.h @@ -56,7 +56,7 @@ public: "A pass name should not contain whitespaces!"); assert(!Name.starts_with('-') && "A pass name should not start with '-'!"); } - virtual ~Pass() {} + virtual ~Pass() = default; /// \Returns the name of the pass. StringRef getName() const { return Name; } #ifndef NDEBUG diff --git a/llvm/include/llvm/SandboxIR/PassManager.h b/llvm/include/llvm/SandboxIR/PassManager.h index 93ca710..a8117aa 100644 --- a/llvm/include/llvm/SandboxIR/PassManager.h +++ b/llvm/include/llvm/SandboxIR/PassManager.h @@ -59,10 +59,10 @@ public: Passes.push_back(std::move(Pass)); } - static constexpr const char EndToken = '\0'; - static constexpr const char BeginArgsToken = '<'; - static constexpr const char EndArgsToken = '>'; - static constexpr const char PassDelimToken = ','; + static constexpr char EndToken = '\0'; + static constexpr char BeginArgsToken = '<'; + static constexpr char EndArgsToken = '>'; + static constexpr char PassDelimToken = ','; /// Parses \p Pipeline as a comma-separated sequence of pass names and sets /// the pass pipeline, using \p CreatePass to instantiate passes by name. diff --git a/llvm/include/llvm/Support/CommandLine.h b/llvm/include/llvm/Support/CommandLine.h index 5a5f00e..d737fbc 100644 --- a/llvm/include/llvm/Support/CommandLine.h +++ b/llvm/include/llvm/Support/CommandLine.h @@ -2099,7 +2099,7 @@ getRegisteredOptions(SubCommand &Sub = SubCommand::getTopLevel()); /// /// This interface is useful for defining subcommands in libraries and /// the dispatch from a single point (like in the main function). -LLVM_ABI iterator_range<typename SmallPtrSet<SubCommand *, 4>::iterator> +LLVM_ABI iterator_range<SmallPtrSet<SubCommand *, 4>::iterator> getRegisteredSubcommands(); //===----------------------------------------------------------------------===// diff --git a/llvm/include/llvm/Support/ELFAttributeParser.h b/llvm/include/llvm/Support/ELFAttributeParser.h index 97350ed..c2ad812 100644 --- a/llvm/include/llvm/Support/ELFAttributeParser.h +++ b/llvm/include/llvm/Support/ELFAttributeParser.h @@ -17,7 +17,7 @@ namespace llvm { class ELFAttributeParser { public: - virtual ~ELFAttributeParser() {} + virtual ~ELFAttributeParser() = default; virtual Error parse(ArrayRef<uint8_t> Section, llvm::endianness Endian) { return llvm::Error::success(); diff --git a/llvm/include/llvm/Support/FormatProviders.h b/llvm/include/llvm/Support/FormatProviders.h index 8eaa5e38..3377781 100644 --- a/llvm/include/llvm/Support/FormatProviders.h +++ b/llvm/include/llvm/Support/FormatProviders.h @@ -261,7 +261,7 @@ template <> struct format_provider<bool> { .Case("y", B ? "yes" : "no") .CaseLower("D", B ? "1" : "0") .Case("T", B ? "TRUE" : "FALSE") - .Cases("t", "", B ? "true" : "false") + .Cases({"t", ""}, B ? "true" : "false") .Default(B ? "1" : "0"); } }; diff --git a/llvm/include/llvm/Support/GraphWriter.h b/llvm/include/llvm/Support/GraphWriter.h index 3bef75c..43d9b0c 100644 --- a/llvm/include/llvm/Support/GraphWriter.h +++ b/llvm/include/llvm/Support/GraphWriter.h @@ -128,7 +128,7 @@ public: DTraits = DOTTraits(SN); RenderUsingHTML = DTraits.renderNodesUsingHTML(); } - virtual ~GraphWriterBase() {} + virtual ~GraphWriterBase() = default; void writeGraph(const std::string &Title = "") { // Output the header for the graph... @@ -369,7 +369,7 @@ class GraphWriter : public GraphWriterBase<GraphType, GraphWriter<GraphType>> { public: GraphWriter(raw_ostream &o, const GraphType &g, bool SN) : GraphWriterBase<GraphType, GraphWriter<GraphType>>(o, g, SN) {} - ~GraphWriter() override {} + ~GraphWriter() override = default; }; template <typename GraphType> diff --git a/llvm/include/llvm/Support/JSON.h b/llvm/include/llvm/Support/JSON.h index d8c6de4..37baa7b 100644 --- a/llvm/include/llvm/Support/JSON.h +++ b/llvm/include/llvm/Support/JSON.h @@ -154,7 +154,7 @@ public: LLVM_ABI const json::Array *getArray(StringRef K) const; LLVM_ABI json::Array *getArray(StringRef K); - friend bool operator==(const Object &LHS, const Object &RHS); + friend LLVM_ABI bool operator==(const Object &LHS, const Object &RHS); }; LLVM_ABI bool operator==(const Object &LHS, const Object &RHS); inline bool operator!=(const Object &LHS, const Object &RHS) { @@ -318,7 +318,7 @@ public: Value(std::string V) : Type(T_String) { if (LLVM_UNLIKELY(!isUTF8(V))) { assert(false && "Invalid UTF-8 in value used as JSON"); - V = fixUTF8(std::move(V)); + V = fixUTF8(V); } create<std::string>(std::move(V)); } @@ -549,10 +549,10 @@ inline const Value &Array::back() const { return V.back(); } inline Value *Array::data() { return V.data(); } inline const Value *Array::data() const { return V.data(); } -inline typename Array::iterator Array::begin() { return V.begin(); } -inline typename Array::const_iterator Array::begin() const { return V.begin(); } -inline typename Array::iterator Array::end() { return V.end(); } -inline typename Array::const_iterator Array::end() const { return V.end(); } +inline Array::iterator Array::begin() { return V.begin(); } +inline Array::const_iterator Array::begin() const { return V.begin(); } +inline Array::iterator Array::end() { return V.end(); } +inline Array::const_iterator Array::end() const { return V.end(); } inline bool Array::empty() const { return V.empty(); } inline size_t Array::size() const { return V.size(); } @@ -565,18 +565,18 @@ template <typename... Args> inline void Array::emplace_back(Args &&...A) { V.emplace_back(std::forward<Args>(A)...); } inline void Array::pop_back() { V.pop_back(); } -inline typename Array::iterator Array::insert(const_iterator P, const Value &E) { +inline Array::iterator Array::insert(const_iterator P, const Value &E) { return V.insert(P, E); } -inline typename Array::iterator Array::insert(const_iterator P, Value &&E) { +inline Array::iterator Array::insert(const_iterator P, Value &&E) { return V.insert(P, std::move(E)); } template <typename It> -inline typename Array::iterator Array::insert(const_iterator P, It A, It Z) { +inline Array::iterator Array::insert(const_iterator P, It A, It Z) { return V.insert(P, A, Z); } template <typename... Args> -inline typename Array::iterator Array::emplace(const_iterator P, Args &&...A) { +inline Array::iterator Array::emplace(const_iterator P, Args &&...A) { return V.emplace(P, std::forward<Args>(A)...); } inline bool operator==(const Array &L, const Array &R) { return L.V == R.V; } @@ -591,7 +591,7 @@ public: ObjectKey(std::string S) : Owned(new std::string(std::move(S))) { if (LLVM_UNLIKELY(!isUTF8(*Owned))) { assert(false && "Invalid UTF-8 in value used as JSON"); - *Owned = fixUTF8(std::move(*Owned)); + *Owned = fixUTF8(*Owned); } Data = *Owned; } diff --git a/llvm/include/llvm/Support/SourceMgr.h b/llvm/include/llvm/Support/SourceMgr.h index 8320006..43f7e27 100644 --- a/llvm/include/llvm/Support/SourceMgr.h +++ b/llvm/include/llvm/Support/SourceMgr.h @@ -103,7 +103,7 @@ private: public: /// Create new source manager without support for include files. - SourceMgr(); + LLVM_ABI SourceMgr(); /// Create new source manager with the capability of finding include files /// via the provided file system. explicit SourceMgr(IntrusiveRefCntPtr<vfs::FileSystem> FS); @@ -111,10 +111,10 @@ public: SourceMgr &operator=(const SourceMgr &) = delete; SourceMgr(SourceMgr &&); SourceMgr &operator=(SourceMgr &&); - ~SourceMgr(); + LLVM_ABI ~SourceMgr(); IntrusiveRefCntPtr<vfs::FileSystem> getVirtualFileSystem() const; - void setVirtualFileSystem(IntrusiveRefCntPtr<vfs::FileSystem> FS); + LLVM_ABI void setVirtualFileSystem(IntrusiveRefCntPtr<vfs::FileSystem> FS); /// Return the include directories of this source manager. ArrayRef<std::string> getIncludeDirs() const { return IncludeDirectories; } diff --git a/llvm/include/llvm/Support/VirtualFileSystem.h b/llvm/include/llvm/Support/VirtualFileSystem.h index c8911a0..dbd5a5c 100644 --- a/llvm/include/llvm/Support/VirtualFileSystem.h +++ b/llvm/include/llvm/Support/VirtualFileSystem.h @@ -1116,8 +1116,9 @@ protected: /// Collect all pairs of <virtual path, real path> entries from the /// \p VFS. This is used by the module dependency collector to forward /// the entries into the reproducer output VFS YAML file. -void collectVFSEntries(RedirectingFileSystem &VFS, - SmallVectorImpl<YAMLVFSEntry> &CollectedEntries); +LLVM_ABI void +collectVFSEntries(RedirectingFileSystem &VFS, + SmallVectorImpl<YAMLVFSEntry> &CollectedEntries); class YAMLVFSWriter { std::vector<YAMLVFSEntry> Mappings; diff --git a/llvm/include/llvm/Support/VirtualOutputBackend.h b/llvm/include/llvm/Support/VirtualOutputBackend.h index 85caa021..78ed4b9b 100644 --- a/llvm/include/llvm/Support/VirtualOutputBackend.h +++ b/llvm/include/llvm/Support/VirtualOutputBackend.h @@ -32,7 +32,7 @@ namespace llvm::vfs { /// If virtual functions are added here, also add them to \a /// ProxyOutputBackend. class OutputBackend : public RefCountedBase<OutputBackend> { - virtual void anchor(); + LLVM_ABI virtual void anchor(); public: /// Get a backend that points to the same destination as this one but that @@ -47,7 +47,7 @@ public: /// have been customized). /// /// Thread-safe. - Expected<OutputFile> + LLVM_ABI Expected<OutputFile> createFile(const Twine &Path, std::optional<OutputConfig> Config = std::nullopt); diff --git a/llvm/include/llvm/Support/VirtualOutputBackends.h b/llvm/include/llvm/Support/VirtualOutputBackends.h index 219bc30..13a9611 100644 --- a/llvm/include/llvm/Support/VirtualOutputBackends.h +++ b/llvm/include/llvm/Support/VirtualOutputBackends.h @@ -77,14 +77,14 @@ private: /// An output backend that creates files on disk, wrapping APIs in sys::fs. class OnDiskOutputBackend : public OutputBackend { - void anchor() override; + LLVM_ABI void anchor() override; protected: IntrusiveRefCntPtr<OutputBackend> cloneImpl() const override { return clone(); } - Expected<std::unique_ptr<OutputFileImpl>> + LLVM_ABI Expected<std::unique_ptr<OutputFileImpl>> createFileImpl(StringRef Path, std::optional<OutputConfig> Config) override; public: diff --git a/llvm/include/llvm/Support/VirtualOutputError.h b/llvm/include/llvm/Support/VirtualOutputError.h index 2293ff9..44590a1 100644 --- a/llvm/include/llvm/Support/VirtualOutputError.h +++ b/llvm/include/llvm/Support/VirtualOutputError.h @@ -43,7 +43,7 @@ public: void log(raw_ostream &OS) const override; // Used by ErrorInfo::classID. - static char ID; + LLVM_ABI static char ID; OutputError(const Twine &OutputPath, std::error_code EC) : ErrorInfo<OutputError, ECError>(EC), OutputPath(OutputPath.str()) { @@ -99,7 +99,7 @@ public: void log(raw_ostream &OS) const override; // Used by ErrorInfo::classID. - static char ID; + LLVM_ABI static char ID; TempFileOutputError(const Twine &TempPath, const Twine &OutputPath, std::error_code EC) diff --git a/llvm/include/llvm/Support/VirtualOutputFile.h b/llvm/include/llvm/Support/VirtualOutputFile.h index dd50437..d53701c 100644 --- a/llvm/include/llvm/Support/VirtualOutputFile.h +++ b/llvm/include/llvm/Support/VirtualOutputFile.h @@ -80,13 +80,13 @@ public: /// /// If there's an open proxy from \a createProxy(), calls \a discard() to /// clean up temporaries followed by \a report_fatal_error(). - Error keep(); + LLVM_ABI Error keep(); /// Discard an output, cleaning up any temporary state. Errors if clean-up /// fails. /// /// If it has already been closed, calls \a report_fatal_error(). - Error discard(); + LLVM_ABI Error discard(); /// Discard the output when destroying it if it's still open, sending the /// result to \a Handler. @@ -98,7 +98,7 @@ public: /// producer. Errors if there's already a proxy. The proxy must be deleted /// before calling \a keep(). The proxy will crash if it's written to after /// calling \a discard(). - Expected<std::unique_ptr<raw_pwrite_stream>> createProxy(); + LLVM_ABI Expected<std::unique_ptr<raw_pwrite_stream>> createProxy(); bool hasOpenProxy() const { return OpenProxy; } @@ -132,7 +132,7 @@ public: private: /// Destroy \a Impl. Reports fatal error if the file is open and there's no /// handler from \a discardOnDestroy(). - void destroy(); + LLVM_ABI void destroy(); OutputFile &moveFrom(OutputFile &O) { Path = std::move(O.Path); Impl = std::move(O.Impl); diff --git a/llvm/include/llvm/Support/YAMLTraits.h b/llvm/include/llvm/Support/YAMLTraits.h index 3d36f41..b53b28d 100644 --- a/llvm/include/llvm/Support/YAMLTraits.h +++ b/llvm/include/llvm/Support/YAMLTraits.h @@ -1921,12 +1921,12 @@ template <typename T> struct StdMapStringCustomMappingTraitsImpl { using map_type = std::map<std::string, T>; static void inputOne(IO &io, StringRef key, map_type &v) { - io.mapRequired(key.str().c_str(), v[std::string(key)]); + io.mapRequired(key, v[std::string(key)]); } static void output(IO &io, map_type &v) { for (auto &p : v) - io.mapRequired(p.first.c_str(), p.second); + io.mapRequired(p.first, p.second); } }; diff --git a/llvm/include/llvm/Transforms/Coroutines/CoroAnnotationElide.h b/llvm/include/llvm/Transforms/Coroutines/CoroAnnotationElide.h index 352c9e14..2061098 100644 --- a/llvm/include/llvm/Transforms/Coroutines/CoroAnnotationElide.h +++ b/llvm/include/llvm/Transforms/Coroutines/CoroAnnotationElide.h @@ -24,7 +24,7 @@ namespace llvm { struct CoroAnnotationElidePass : PassInfoMixin<CoroAnnotationElidePass> { - CoroAnnotationElidePass() {} + CoroAnnotationElidePass() = default; PreservedAnalyses run(LazyCallGraph::SCC &C, CGSCCAnalysisManager &AM, LazyCallGraph &CG, CGSCCUpdateResult &UR); diff --git a/llvm/include/llvm/Transforms/IPO/FatLTOCleanup.h b/llvm/include/llvm/Transforms/IPO/FatLTOCleanup.h index 17eab85..6fc1b262 100644 --- a/llvm/include/llvm/Transforms/IPO/FatLTOCleanup.h +++ b/llvm/include/llvm/Transforms/IPO/FatLTOCleanup.h @@ -26,7 +26,7 @@ class ModuleSummaryIndex; class FatLtoCleanup : public PassInfoMixin<FatLtoCleanup> { public: - FatLtoCleanup() {} + FatLtoCleanup() = default; PreservedAnalyses run(Module &M, ModuleAnalysisManager &AM); static bool isRequired() { return true; } }; diff --git a/llvm/include/llvm/Transforms/IPO/InferFunctionAttrs.h b/llvm/include/llvm/Transforms/IPO/InferFunctionAttrs.h index 8addf49..272b960 100644 --- a/llvm/include/llvm/Transforms/IPO/InferFunctionAttrs.h +++ b/llvm/include/llvm/Transforms/IPO/InferFunctionAttrs.h @@ -23,7 +23,7 @@ class Module; /// A pass which infers function attributes from the names and signatures of /// function declarations in a module. struct InferFunctionAttrsPass : PassInfoMixin<InferFunctionAttrsPass> { - PreservedAnalyses run(Module &M, ModuleAnalysisManager &AM); + LLVM_ABI PreservedAnalyses run(Module &M, ModuleAnalysisManager &AM); }; } diff --git a/llvm/include/llvm/Transforms/Instrumentation/SanitizerCoverage.h b/llvm/include/llvm/Transforms/Instrumentation/SanitizerCoverage.h index a8a09fb..346e7f0 100644 --- a/llvm/include/llvm/Transforms/Instrumentation/SanitizerCoverage.h +++ b/llvm/include/llvm/Transforms/Instrumentation/SanitizerCoverage.h @@ -33,7 +33,7 @@ class FileSystem; /// appends globals to llvm.compiler.used. class SanitizerCoveragePass : public PassInfoMixin<SanitizerCoveragePass> { public: - explicit SanitizerCoveragePass( + LLVM_ABI explicit SanitizerCoveragePass( SanitizerCoverageOptions Options = SanitizerCoverageOptions(), IntrusiveRefCntPtr<vfs::FileSystem> VFS = nullptr, const std::vector<std::string> &AllowlistFiles = {}, diff --git a/llvm/include/llvm/Transforms/Vectorize/SandboxVectorizer/Legality.h b/llvm/include/llvm/Transforms/Vectorize/SandboxVectorizer/Legality.h index 96a2348..3d76cda 100644 --- a/llvm/include/llvm/Transforms/Vectorize/SandboxVectorizer/Legality.h +++ b/llvm/include/llvm/Transforms/Vectorize/SandboxVectorizer/Legality.h @@ -167,7 +167,7 @@ protected: LegalityResult &operator=(const LegalityResult &) = delete; public: - virtual ~LegalityResult() {} + virtual ~LegalityResult() = default; LegalityResultID getSubclassID() const { return ID; } #ifndef NDEBUG virtual void print(raw_ostream &OS) const { diff --git a/llvm/include/llvm/Transforms/Vectorize/SandboxVectorizer/SeedCollector.h b/llvm/include/llvm/Transforms/Vectorize/SandboxVectorizer/SeedCollector.h index b289520..821382b 100644 --- a/llvm/include/llvm/Transforms/Vectorize/SandboxVectorizer/SeedCollector.h +++ b/llvm/include/llvm/Transforms/Vectorize/SandboxVectorizer/SeedCollector.h @@ -36,7 +36,7 @@ public: /// No need to allow copies. SeedBundle(const SeedBundle &) = delete; SeedBundle &operator=(const SeedBundle &) = delete; - virtual ~SeedBundle() {} + virtual ~SeedBundle() = default; using iterator = SmallVector<Instruction *>::iterator; using const_iterator = SmallVector<Instruction *>::const_iterator; diff --git a/llvm/lib/Analysis/AliasAnalysis.cpp b/llvm/lib/Analysis/AliasAnalysis.cpp index f2dc25f..26a5602 100644 --- a/llvm/lib/Analysis/AliasAnalysis.cpp +++ b/llvm/lib/Analysis/AliasAnalysis.cpp @@ -75,7 +75,7 @@ AAResults::AAResults(const TargetLibraryInfo &TLI) : TLI(TLI) {} AAResults::AAResults(AAResults &&Arg) : TLI(Arg.TLI), AAs(std::move(Arg.AAs)), AADeps(std::move(Arg.AADeps)) {} -AAResults::~AAResults() {} +AAResults::~AAResults() = default; bool AAResults::invalidate(Function &F, const PreservedAnalyses &PA, FunctionAnalysisManager::Invalidator &Inv) { diff --git a/llvm/lib/Analysis/ScalarEvolution.cpp b/llvm/lib/Analysis/ScalarEvolution.cpp index c9baeda..a31f17b 100644 --- a/llvm/lib/Analysis/ScalarEvolution.cpp +++ b/llvm/lib/Analysis/ScalarEvolution.cpp @@ -2424,10 +2424,10 @@ ScalarEvolution::getStrengthenedNoWrapFlagsFromBinOp( // We're trying to construct a SCEV of type `Type' with `Ops' as operands and // `OldFlags' as can't-wrap behavior. Infer a more aggressive set of // can't-overflow flags for the operation if possible. -static SCEV::NoWrapFlags -StrengthenNoWrapFlags(ScalarEvolution *SE, SCEVTypes Type, - const ArrayRef<const SCEV *> Ops, - SCEV::NoWrapFlags Flags) { +static SCEV::NoWrapFlags StrengthenNoWrapFlags(ScalarEvolution *SE, + SCEVTypes Type, + ArrayRef<const SCEV *> Ops, + SCEV::NoWrapFlags Flags) { using namespace std::placeholders; using OBO = OverflowingBinaryOperator; @@ -2540,7 +2540,7 @@ const SCEV *ScalarEvolution::getAddExpr(SmallVectorImpl<const SCEV *> &Ops, unsigned Idx = isa<SCEVConstant>(Ops[0]) ? 1 : 0; // Delay expensive flag strengthening until necessary. - auto ComputeFlags = [this, OrigFlags](const ArrayRef<const SCEV *> Ops) { + auto ComputeFlags = [this, OrigFlags](ArrayRef<const SCEV *> Ops) { return StrengthenNoWrapFlags(this, scAddExpr, Ops, OrigFlags); }; @@ -3125,7 +3125,7 @@ const SCEV *ScalarEvolution::getMulExpr(SmallVectorImpl<const SCEV *> &Ops, return Folded; // Delay expensive flag strengthening until necessary. - auto ComputeFlags = [this, OrigFlags](const ArrayRef<const SCEV *> Ops) { + auto ComputeFlags = [this, OrigFlags](ArrayRef<const SCEV *> Ops) { return StrengthenNoWrapFlags(this, scMulExpr, Ops, OrigFlags); }; @@ -15510,6 +15510,78 @@ static const SCEV *getNextSCEVDivisibleByDivisor(const SCEV *Expr, return SE.getConstant(*ExprVal + DivisorVal - Rem); } +static bool collectDivisibilityInformation( + ICmpInst::Predicate Predicate, const SCEV *LHS, const SCEV *RHS, + DenseMap<const SCEV *, const SCEV *> &DivInfo, + DenseMap<const SCEV *, APInt> &Multiples, ScalarEvolution &SE) { + // If we have LHS == 0, check if LHS is computing a property of some unknown + // SCEV %v which we can rewrite %v to express explicitly. + if (Predicate != CmpInst::ICMP_EQ || !match(RHS, m_scev_Zero())) + return false; + // If LHS is A % B, i.e. A % B == 0, rewrite A to (A /u B) * B to + // explicitly express that. + const SCEVUnknown *URemLHS = nullptr; + const SCEV *URemRHS = nullptr; + if (!match(LHS, m_scev_URem(m_SCEVUnknown(URemLHS), m_SCEV(URemRHS), SE))) + return false; + + const SCEV *Multiple = + SE.getMulExpr(SE.getUDivExpr(URemLHS, URemRHS), URemRHS); + DivInfo[URemLHS] = Multiple; + if (auto *C = dyn_cast<SCEVConstant>(URemRHS)) + Multiples[URemLHS] = C->getAPInt(); + return true; +} + +// Check if the condition is a divisibility guard (A % B == 0). +static bool isDivisibilityGuard(const SCEV *LHS, const SCEV *RHS, + ScalarEvolution &SE) { + const SCEV *X, *Y; + return match(LHS, m_scev_URem(m_SCEV(X), m_SCEV(Y), SE)) && RHS->isZero(); +} + +// Apply divisibility by \p Divisor on MinMaxExpr with constant values, +// recursively. This is done by aligning up/down the constant value to the +// Divisor. +static const SCEV *applyDivisibilityOnMinMaxExpr(const SCEV *MinMaxExpr, + APInt Divisor, + ScalarEvolution &SE) { + // Return true if \p Expr is a MinMax SCEV expression with a non-negative + // constant operand. If so, return in \p SCTy the SCEV type and in \p RHS + // the non-constant operand and in \p LHS the constant operand. + auto IsMinMaxSCEVWithNonNegativeConstant = + [&](const SCEV *Expr, SCEVTypes &SCTy, const SCEV *&LHS, + const SCEV *&RHS) { + if (auto *MinMax = dyn_cast<SCEVMinMaxExpr>(Expr)) { + if (MinMax->getNumOperands() != 2) + return false; + if (auto *C = dyn_cast<SCEVConstant>(MinMax->getOperand(0))) { + if (C->getAPInt().isNegative()) + return false; + SCTy = MinMax->getSCEVType(); + LHS = MinMax->getOperand(0); + RHS = MinMax->getOperand(1); + return true; + } + } + return false; + }; + + const SCEV *MinMaxLHS = nullptr, *MinMaxRHS = nullptr; + SCEVTypes SCTy; + if (!IsMinMaxSCEVWithNonNegativeConstant(MinMaxExpr, SCTy, MinMaxLHS, + MinMaxRHS)) + return MinMaxExpr; + auto IsMin = isa<SCEVSMinExpr>(MinMaxExpr) || isa<SCEVUMinExpr>(MinMaxExpr); + assert(SE.isKnownNonNegative(MinMaxLHS) && "Expected non-negative operand!"); + auto *DivisibleExpr = + IsMin ? getPreviousSCEVDivisibleByDivisor(MinMaxLHS, Divisor, SE) + : getNextSCEVDivisibleByDivisor(MinMaxLHS, Divisor, SE); + SmallVector<const SCEV *> Ops = { + applyDivisibilityOnMinMaxExpr(MinMaxRHS, Divisor, SE), DivisibleExpr}; + return SE.getMinMaxExpr(SCTy, Ops); +} + void ScalarEvolution::LoopGuards::collectFromBlock( ScalarEvolution &SE, ScalarEvolution::LoopGuards &Guards, const BasicBlock *Block, const BasicBlock *Pred, @@ -15520,19 +15592,13 @@ void ScalarEvolution::LoopGuards::collectFromBlock( SmallVector<const SCEV *> ExprsToRewrite; auto CollectCondition = [&](ICmpInst::Predicate Predicate, const SCEV *LHS, const SCEV *RHS, - DenseMap<const SCEV *, const SCEV *> - &RewriteMap) { + DenseMap<const SCEV *, const SCEV *> &RewriteMap, + const LoopGuards &DivGuards) { // WARNING: It is generally unsound to apply any wrap flags to the proposed // replacement SCEV which isn't directly implied by the structure of that // SCEV. In particular, using contextual facts to imply flags is *NOT* // legal. See the scoping rules for flags in the header to understand why. - // If LHS is a constant, apply information to the other expression. - if (isa<SCEVConstant>(LHS)) { - std::swap(LHS, RHS); - Predicate = CmpInst::getSwappedPredicate(Predicate); - } - // Check for a condition of the form (-C1 + X < C2). InstCombine will // create this form when combining two checks of the form (X u< C2 + C1) and // (X >=u C1). @@ -15565,67 +15631,6 @@ void ScalarEvolution::LoopGuards::collectFromBlock( if (MatchRangeCheckIdiom()) return; - // Return true if \p Expr is a MinMax SCEV expression with a non-negative - // constant operand. If so, return in \p SCTy the SCEV type and in \p RHS - // the non-constant operand and in \p LHS the constant operand. - auto IsMinMaxSCEVWithNonNegativeConstant = - [&](const SCEV *Expr, SCEVTypes &SCTy, const SCEV *&LHS, - const SCEV *&RHS) { - const APInt *C; - SCTy = Expr->getSCEVType(); - return match(Expr, m_scev_MinMax(m_SCEV(LHS), m_SCEV(RHS))) && - match(LHS, m_scev_APInt(C)) && C->isNonNegative(); - }; - - // Apply divisibilty by \p Divisor on MinMaxExpr with constant values, - // recursively. This is done by aligning up/down the constant value to the - // Divisor. - std::function<const SCEV *(const SCEV *, const SCEV *)> - ApplyDivisibiltyOnMinMaxExpr = [&](const SCEV *MinMaxExpr, - const SCEV *Divisor) { - auto *ConstDivisor = dyn_cast<SCEVConstant>(Divisor); - if (!ConstDivisor) - return MinMaxExpr; - const APInt &DivisorVal = ConstDivisor->getAPInt(); - - const SCEV *MinMaxLHS = nullptr, *MinMaxRHS = nullptr; - SCEVTypes SCTy; - if (!IsMinMaxSCEVWithNonNegativeConstant(MinMaxExpr, SCTy, MinMaxLHS, - MinMaxRHS)) - return MinMaxExpr; - auto IsMin = - isa<SCEVSMinExpr>(MinMaxExpr) || isa<SCEVUMinExpr>(MinMaxExpr); - assert(SE.isKnownNonNegative(MinMaxLHS) && - "Expected non-negative operand!"); - auto *DivisibleExpr = - IsMin - ? getPreviousSCEVDivisibleByDivisor(MinMaxLHS, DivisorVal, SE) - : getNextSCEVDivisibleByDivisor(MinMaxLHS, DivisorVal, SE); - SmallVector<const SCEV *> Ops = { - ApplyDivisibiltyOnMinMaxExpr(MinMaxRHS, Divisor), DivisibleExpr}; - return SE.getMinMaxExpr(SCTy, Ops); - }; - - // If we have LHS == 0, check if LHS is computing a property of some unknown - // SCEV %v which we can rewrite %v to express explicitly. - if (Predicate == CmpInst::ICMP_EQ && match(RHS, m_scev_Zero())) { - // If LHS is A % B, i.e. A % B == 0, rewrite A to (A /u B) * B to - // explicitly express that. - const SCEVUnknown *URemLHS = nullptr; - const SCEV *URemRHS = nullptr; - if (match(LHS, - m_scev_URem(m_SCEVUnknown(URemLHS), m_SCEV(URemRHS), SE))) { - auto I = RewriteMap.find(URemLHS); - const SCEV *RewrittenLHS = I != RewriteMap.end() ? I->second : URemLHS; - RewrittenLHS = ApplyDivisibiltyOnMinMaxExpr(RewrittenLHS, URemRHS); - const auto *Multiple = - SE.getMulExpr(SE.getUDivExpr(RewrittenLHS, URemRHS), URemRHS); - RewriteMap[URemLHS] = Multiple; - ExprsToRewrite.push_back(URemLHS); - return; - } - } - // Do not apply information for constants or if RHS contains an AddRec. if (isa<SCEVConstant>(LHS) || SE.containsAddRecurrence(RHS)) return; @@ -15655,7 +15660,9 @@ void ScalarEvolution::LoopGuards::collectFromBlock( }; const SCEV *RewrittenLHS = GetMaybeRewritten(LHS); - const APInt &DividesBy = SE.getConstantMultiple(RewrittenLHS); + // Apply divisibility information when computing the constant multiple. + const APInt &DividesBy = + SE.getConstantMultiple(DivGuards.rewrite(RewrittenLHS)); // Collect rewrites for LHS and its transitive operands based on the // condition. @@ -15840,8 +15847,11 @@ void ScalarEvolution::LoopGuards::collectFromBlock( // Now apply the information from the collected conditions to // Guards.RewriteMap. Conditions are processed in reverse order, so the - // earliest conditions is processed first. This ensures the SCEVs with the + // earliest conditions is processed first, except guards with divisibility + // information, which are moved to the back. This ensures the SCEVs with the // shortest dependency chains are constructed first. + SmallVector<std::tuple<CmpInst::Predicate, const SCEV *, const SCEV *>> + GuardsToProcess; for (auto [Term, EnterIfTrue] : reverse(Terms)) { SmallVector<Value *, 8> Worklist; SmallPtrSet<Value *, 8> Visited; @@ -15856,7 +15866,14 @@ void ScalarEvolution::LoopGuards::collectFromBlock( EnterIfTrue ? Cmp->getPredicate() : Cmp->getInversePredicate(); const auto *LHS = SE.getSCEV(Cmp->getOperand(0)); const auto *RHS = SE.getSCEV(Cmp->getOperand(1)); - CollectCondition(Predicate, LHS, RHS, Guards.RewriteMap); + // If LHS is a constant, apply information to the other expression. + // TODO: If LHS is not a constant, check if using CompareSCEVComplexity + // can improve results. + if (isa<SCEVConstant>(LHS)) { + std::swap(LHS, RHS); + Predicate = CmpInst::getSwappedPredicate(Predicate); + } + GuardsToProcess.emplace_back(Predicate, LHS, RHS); continue; } @@ -15869,6 +15886,31 @@ void ScalarEvolution::LoopGuards::collectFromBlock( } } + // Process divisibility guards in reverse order to populate DivGuards early. + DenseMap<const SCEV *, APInt> Multiples; + LoopGuards DivGuards(SE); + for (const auto &[Predicate, LHS, RHS] : GuardsToProcess) { + if (!isDivisibilityGuard(LHS, RHS, SE)) + continue; + collectDivisibilityInformation(Predicate, LHS, RHS, DivGuards.RewriteMap, + Multiples, SE); + } + + for (const auto &[Predicate, LHS, RHS] : GuardsToProcess) + CollectCondition(Predicate, LHS, RHS, Guards.RewriteMap, DivGuards); + + // Apply divisibility information last. This ensures it is applied to the + // outermost expression after other rewrites for the given value. + for (const auto &[K, Divisor] : Multiples) { + const SCEV *DivisorSCEV = SE.getConstant(Divisor); + Guards.RewriteMap[K] = + SE.getMulExpr(SE.getUDivExpr(applyDivisibilityOnMinMaxExpr( + Guards.rewrite(K), Divisor, SE), + DivisorSCEV), + DivisorSCEV); + ExprsToRewrite.push_back(K); + } + // Let the rewriter preserve NUW/NSW flags if the unsigned/signed ranges of // the replacement expressions are contained in the ranges of the replaced // expressions. diff --git a/llvm/lib/Analysis/TargetTransformInfo.cpp b/llvm/lib/Analysis/TargetTransformInfo.cpp index c47a1c1..0426ac7 100644 --- a/llvm/lib/Analysis/TargetTransformInfo.cpp +++ b/llvm/lib/Analysis/TargetTransformInfo.cpp @@ -1353,9 +1353,9 @@ TargetTransformInfo::getInlineCallPenalty(const Function *F, return TTIImpl->getInlineCallPenalty(F, Call, DefaultCallPenalty); } -bool TargetTransformInfo::areTypesABICompatible( - const Function *Caller, const Function *Callee, - const ArrayRef<Type *> &Types) const { +bool TargetTransformInfo::areTypesABICompatible(const Function *Caller, + const Function *Callee, + ArrayRef<Type *> Types) const { return TTIImpl->areTypesABICompatible(Caller, Callee, Types); } diff --git a/llvm/lib/BinaryFormat/Dwarf.cpp b/llvm/lib/BinaryFormat/Dwarf.cpp index 55fa2df..a6c7e6a 100644 --- a/llvm/lib/BinaryFormat/Dwarf.cpp +++ b/llvm/lib/BinaryFormat/Dwarf.cpp @@ -1076,10 +1076,3 @@ StringRef (*const llvm::dwarf::EnumTraits<LineNumberOps>::StringFn)(unsigned) = LNStandardString; StringRef (*const llvm::dwarf::EnumTraits<Index>::StringFn)(unsigned) = IndexString; - -constexpr char llvm::dwarf::EnumTraits<Attribute>::Type[]; -constexpr char llvm::dwarf::EnumTraits<Form>::Type[]; -constexpr char llvm::dwarf::EnumTraits<Index>::Type[]; -constexpr char llvm::dwarf::EnumTraits<Tag>::Type[]; -constexpr char llvm::dwarf::EnumTraits<LineNumberOps>::Type[]; -constexpr char llvm::dwarf::EnumTraits<LocationAtom>::Type[]; diff --git a/llvm/lib/BinaryFormat/MsgPackDocumentYAML.cpp b/llvm/lib/BinaryFormat/MsgPackDocumentYAML.cpp index 3de3dcc..80b421d 100644 --- a/llvm/lib/BinaryFormat/MsgPackDocumentYAML.cpp +++ b/llvm/lib/BinaryFormat/MsgPackDocumentYAML.cpp @@ -209,12 +209,12 @@ template <> struct CustomMappingTraits<MapDocNode> { static void inputOne(IO &IO, StringRef Key, MapDocNode &M) { ScalarDocNode KeyObj = M.getDocument()->getNode(); KeyObj.fromString(Key, ""); - IO.mapRequired(Key.str().c_str(), M.getMap()[KeyObj]); + IO.mapRequired(Key, M.getMap()[KeyObj]); } static void output(IO &IO, MapDocNode &M) { for (auto I : M.getMap()) { - IO.mapRequired(I.first.toString().c_str(), I.second); + IO.mapRequired(I.first.toString(), I.second); } } }; diff --git a/llvm/lib/CGData/OutlinedHashTreeRecord.cpp b/llvm/lib/CGData/OutlinedHashTreeRecord.cpp index cc76063..2b6e2f0 100644 --- a/llvm/lib/CGData/OutlinedHashTreeRecord.cpp +++ b/llvm/lib/CGData/OutlinedHashTreeRecord.cpp @@ -37,7 +37,7 @@ template <> struct MappingTraits<HashNodeStable> { template <> struct CustomMappingTraits<IdHashNodeStableMapTy> { static void inputOne(IO &io, StringRef Key, IdHashNodeStableMapTy &V) { HashNodeStable NodeStable; - io.mapRequired(Key.str().c_str(), NodeStable); + io.mapRequired(Key, NodeStable); unsigned Id; if (Key.getAsInteger(0, Id)) { io.setError("Id not an integer"); @@ -48,7 +48,7 @@ template <> struct CustomMappingTraits<IdHashNodeStableMapTy> { static void output(IO &io, IdHashNodeStableMapTy &V) { for (auto Iter = V.begin(); Iter != V.end(); ++Iter) - io.mapRequired(utostr(Iter->first).c_str(), Iter->second); + io.mapRequired(utostr(Iter->first), Iter->second); } }; diff --git a/llvm/lib/CodeGen/AsmPrinter/AccelTable.cpp b/llvm/lib/CodeGen/AsmPrinter/AccelTable.cpp index e5c85d5..1ea30d8 100644 --- a/llvm/lib/CodeGen/AsmPrinter/AccelTable.cpp +++ b/llvm/lib/CodeGen/AsmPrinter/AccelTable.cpp @@ -745,11 +745,6 @@ void AppleAccelTableStaticTypeData::emit(AsmPrinter *Asm) const { Asm->emitInt32(QualifiedNameHash); } -constexpr AppleAccelTableData::Atom AppleAccelTableTypeData::Atoms[]; -constexpr AppleAccelTableData::Atom AppleAccelTableOffsetData::Atoms[]; -constexpr AppleAccelTableData::Atom AppleAccelTableStaticOffsetData::Atoms[]; -constexpr AppleAccelTableData::Atom AppleAccelTableStaticTypeData::Atoms[]; - #ifndef NDEBUG void AppleAccelTableWriter::Header::print(raw_ostream &OS) const { OS << "Magic: " << format("0x%x", Magic) << "\n" diff --git a/llvm/lib/CodeGen/AsmPrinter/DbgEntityHistoryCalculator.cpp b/llvm/lib/CodeGen/AsmPrinter/DbgEntityHistoryCalculator.cpp index 171fb83..98cdada 100644 --- a/llvm/lib/CodeGen/AsmPrinter/DbgEntityHistoryCalculator.cpp +++ b/llvm/lib/CodeGen/AsmPrinter/DbgEntityHistoryCalculator.cpp @@ -112,8 +112,7 @@ void DbgValueHistoryMap::Entry::endEntry(EntryIndex Index) { /// to the first intersecting scope range if one exists. static std::optional<ArrayRef<InsnRange>::iterator> intersects(const MachineInstr *StartMI, const MachineInstr *EndMI, - const ArrayRef<InsnRange> &Ranges, - const InstructionOrdering &Ordering) { + ArrayRef<InsnRange> Ranges, const InstructionOrdering &Ordering) { for (auto RangesI = Ranges.begin(), RangesE = Ranges.end(); RangesI != RangesE; ++RangesI) { if (EndMI && Ordering.isBefore(EndMI, RangesI->first)) diff --git a/llvm/lib/CodeGen/CodeGenPrepare.cpp b/llvm/lib/CodeGen/CodeGenPrepare.cpp index 8ea1326..0309e22 100644 --- a/llvm/lib/CodeGen/CodeGenPrepare.cpp +++ b/llvm/lib/CodeGen/CodeGenPrepare.cpp @@ -368,7 +368,7 @@ class CodeGenPrepare { std::unique_ptr<DominatorTree> DT; public: - CodeGenPrepare(){}; + CodeGenPrepare() = default; CodeGenPrepare(const TargetMachine *TM) : TM(TM){}; /// If encounter huge function, we need to limit the build time. bool IsHugeFunc = false; diff --git a/llvm/lib/CodeGen/GlobalISel/CombinerHelper.cpp b/llvm/lib/CodeGen/GlobalISel/CombinerHelper.cpp index 9ace7d6..ec4d13f 100644 --- a/llvm/lib/CodeGen/GlobalISel/CombinerHelper.cpp +++ b/llvm/lib/CodeGen/GlobalISel/CombinerHelper.cpp @@ -589,8 +589,8 @@ bool CombinerHelper::matchCombineShuffleVector( return true; } -void CombinerHelper::applyCombineShuffleVector( - MachineInstr &MI, const ArrayRef<Register> Ops) const { +void CombinerHelper::applyCombineShuffleVector(MachineInstr &MI, + ArrayRef<Register> Ops) const { Register DstReg = MI.getOperand(0).getReg(); Builder.setInsertPt(*MI.getParent(), MI); Register NewDstReg = MRI.cloneVirtualRegister(DstReg); diff --git a/llvm/lib/CodeGen/MachineOperand.cpp b/llvm/lib/CodeGen/MachineOperand.cpp index bb9c76f..8c6d219 100644 --- a/llvm/lib/CodeGen/MachineOperand.cpp +++ b/llvm/lib/CodeGen/MachineOperand.cpp @@ -363,8 +363,9 @@ bool MachineOperand::isIdenticalTo(const MachineOperand &Other) const { case MachineOperand::MO_RegisterMask: case MachineOperand::MO_RegisterLiveOut: { // Shallow compare of the two RegMasks - const uint32_t *RegMask = getRegMask(); - const uint32_t *OtherRegMask = Other.getRegMask(); + const uint32_t *RegMask = isRegMask() ? getRegMask() : getRegLiveOut(); + const uint32_t *OtherRegMask = + isRegMask() ? Other.getRegMask() : Other.getRegLiveOut(); if (RegMask == OtherRegMask) return true; @@ -434,7 +435,8 @@ hash_code llvm::hash_value(const MachineOperand &MO) { if (const MachineFunction *MF = getMFIfAvailable(MO)) { const TargetRegisterInfo *TRI = MF->getSubtarget().getRegisterInfo(); unsigned RegMaskSize = MachineOperand::getRegMaskSize(TRI->getNumRegs()); - const uint32_t *RegMask = MO.getRegMask(); + const uint32_t *RegMask = + MO.isRegMask() ? MO.getRegMask() : MO.getRegLiveOut(); std::vector<stable_hash> RegMaskHashes(RegMask, RegMask + RegMaskSize); return hash_combine(MO.getType(), MO.getTargetFlags(), stable_hash_combine(RegMaskHashes)); diff --git a/llvm/lib/CodeGen/MachineScheduler.cpp b/llvm/lib/CodeGen/MachineScheduler.cpp index 3ed1045..f18c051 100644 --- a/llvm/lib/CodeGen/MachineScheduler.cpp +++ b/llvm/lib/CodeGen/MachineScheduler.cpp @@ -334,7 +334,7 @@ public: LiveIntervals &LIS; }; - MachineSchedulerImpl() {} + MachineSchedulerImpl() = default; // Migration only void setLegacyPass(MachineFunctionPass *P) { this->P = P; } void setMFAM(MachineFunctionAnalysisManager *MFAM) { this->MFAM = MFAM; } @@ -358,7 +358,7 @@ public: MachineLoopInfo &MLI; AAResults &AA; }; - PostMachineSchedulerImpl() {} + PostMachineSchedulerImpl() = default; // Migration only void setLegacyPass(MachineFunctionPass *P) { this->P = P; } void setMFAM(MachineFunctionAnalysisManager *MFAM) { this->MFAM = MFAM; } diff --git a/llvm/lib/CodeGen/MachineStableHash.cpp b/llvm/lib/CodeGen/MachineStableHash.cpp index 9d56696..6da708d 100644 --- a/llvm/lib/CodeGen/MachineStableHash.cpp +++ b/llvm/lib/CodeGen/MachineStableHash.cpp @@ -136,7 +136,8 @@ stable_hash llvm::stableHashValue(const MachineOperand &MO) { const TargetRegisterInfo *TRI = MF->getSubtarget().getRegisterInfo(); unsigned RegMaskSize = MachineOperand::getRegMaskSize(TRI->getNumRegs()); - const uint32_t *RegMask = MO.getRegMask(); + const uint32_t *RegMask = + MO.isRegMask() ? MO.getRegMask() : MO.getRegLiveOut(); std::vector<llvm::stable_hash> RegMaskHashes(RegMask, RegMask + RegMaskSize); return stable_hash_combine(MO.getType(), MO.getTargetFlags(), diff --git a/llvm/lib/CodeGen/RegAllocFast.cpp b/llvm/lib/CodeGen/RegAllocFast.cpp index 697b779..ec6ffd4 100644 --- a/llvm/lib/CodeGen/RegAllocFast.cpp +++ b/llvm/lib/CodeGen/RegAllocFast.cpp @@ -206,7 +206,7 @@ private: bool Error = false; ///< Could not allocate. explicit LiveReg(Register VirtReg) : VirtReg(VirtReg) {} - explicit LiveReg() {} + explicit LiveReg() = default; unsigned getSparseSetIndex() const { return VirtReg.virtRegIndex(); } }; diff --git a/llvm/lib/CodeGen/RegisterCoalescer.cpp b/llvm/lib/CodeGen/RegisterCoalescer.cpp index e17a214b..38f6deb 100644 --- a/llvm/lib/CodeGen/RegisterCoalescer.cpp +++ b/llvm/lib/CodeGen/RegisterCoalescer.cpp @@ -378,7 +378,7 @@ class RegisterCoalescer : private LiveRangeEdit::Delegate { public: // For legacy pass only. - RegisterCoalescer() {} + RegisterCoalescer() = default; RegisterCoalescer &operator=(RegisterCoalescer &&Other) = default; RegisterCoalescer(LiveIntervals *LIS, SlotIndexes *SI, diff --git a/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp b/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp index bdd6bf0..46c4bb8 100644 --- a/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp +++ b/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp @@ -9374,7 +9374,7 @@ static unsigned bigEndianByteAt(unsigned BW, unsigned i) { // Check if the bytes offsets we are looking at match with either big or // little endian value loaded. Return true for big endian, false for little // endian, and std::nullopt if match failed. -static std::optional<bool> isBigEndian(const ArrayRef<int64_t> ByteOffsets, +static std::optional<bool> isBigEndian(ArrayRef<int64_t> ByteOffsets, int64_t FirstOffset) { // The endian can be decided only when it is 2 bytes at least. unsigned Width = ByteOffsets.size(); diff --git a/llvm/lib/CodeGenTypes/LowLevelType.cpp b/llvm/lib/CodeGenTypes/LowLevelType.cpp index 4785f26..92b7fad 100644 --- a/llvm/lib/CodeGenTypes/LowLevelType.cpp +++ b/llvm/lib/CodeGenTypes/LowLevelType.cpp @@ -54,9 +54,3 @@ LLVM_DUMP_METHOD void LLT::dump() const { dbgs() << '\n'; } #endif - -const constexpr LLT::BitFieldInfo LLT::ScalarSizeFieldInfo; -const constexpr LLT::BitFieldInfo LLT::PointerSizeFieldInfo; -const constexpr LLT::BitFieldInfo LLT::PointerAddressSpaceFieldInfo; -const constexpr LLT::BitFieldInfo LLT::VectorElementsFieldInfo; -const constexpr LLT::BitFieldInfo LLT::VectorScalableFieldInfo; diff --git a/llvm/lib/DWARFLinker/Parallel/DWARFLinkerUnit.h b/llvm/lib/DWARFLinker/Parallel/DWARFLinkerUnit.h index 84757ae..970abdc 100644 --- a/llvm/lib/DWARFLinker/Parallel/DWARFLinkerUnit.h +++ b/llvm/lib/DWARFLinker/Parallel/DWARFLinkerUnit.h @@ -28,7 +28,7 @@ using MacroOffset2UnitMapTy = DenseMap<uint64_t, DwarfUnit *>; /// Base class for all Dwarf units(Compile unit/Type table unit). class DwarfUnit : public OutputSections { public: - virtual ~DwarfUnit() {} + virtual ~DwarfUnit() = default; DwarfUnit(LinkingGlobalData &GlobalData, unsigned ID, StringRef ClangModuleName) : OutputSections(GlobalData), ID(ID), ClangModuleName(ClangModuleName), diff --git a/llvm/lib/DWARFLinker/Parallel/StringEntryToDwarfStringPoolEntryMap.h b/llvm/lib/DWARFLinker/Parallel/StringEntryToDwarfStringPoolEntryMap.h index f67536e..8ccb4a5 100644 --- a/llvm/lib/DWARFLinker/Parallel/StringEntryToDwarfStringPoolEntryMap.h +++ b/llvm/lib/DWARFLinker/Parallel/StringEntryToDwarfStringPoolEntryMap.h @@ -22,7 +22,7 @@ class StringEntryToDwarfStringPoolEntryMap { public: StringEntryToDwarfStringPoolEntryMap(LinkingGlobalData &GlobalData) : GlobalData(GlobalData) {} - ~StringEntryToDwarfStringPoolEntryMap() {} + ~StringEntryToDwarfStringPoolEntryMap() = default; /// Create DwarfStringPoolEntry for specified StringEntry if necessary. /// Initialize DwarfStringPoolEntry with initial values. diff --git a/llvm/lib/DebugInfo/CodeView/LazyRandomTypeCollection.cpp b/llvm/lib/DebugInfo/CodeView/LazyRandomTypeCollection.cpp index 6c23ba8..23ab534 100644 --- a/llvm/lib/DebugInfo/CodeView/LazyRandomTypeCollection.cpp +++ b/llvm/lib/DebugInfo/CodeView/LazyRandomTypeCollection.cpp @@ -102,7 +102,8 @@ std::optional<CVType> LazyRandomTypeCollection::tryGetType(TypeIndex Index) { return std::nullopt; } - assert(contains(Index)); + if (!contains(Index)) + return std::nullopt; return Records[Index.toArrayIndex()].Type; } diff --git a/llvm/lib/ExecutionEngine/Orc/MemoryMapper.cpp b/llvm/lib/ExecutionEngine/Orc/MemoryMapper.cpp index 7e606c6a..4e7db82 100644 --- a/llvm/lib/ExecutionEngine/Orc/MemoryMapper.cpp +++ b/llvm/lib/ExecutionEngine/Orc/MemoryMapper.cpp @@ -27,7 +27,7 @@ namespace llvm { namespace orc { -MemoryMapper::~MemoryMapper() {} +MemoryMapper::~MemoryMapper() = default; InProcessMemoryMapper::InProcessMemoryMapper(size_t PageSize) : PageSize(PageSize) {} diff --git a/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp b/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp index 0e5926f..fff9a81 100644 --- a/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp +++ b/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp @@ -528,7 +528,7 @@ void OpenMPIRBuilder::getKernelArgsVector(TargetKernelArgs &KernelArgs, Value *Version = Builder.getInt32(OMP_KERNEL_ARG_VERSION); Value *PointerNum = Builder.getInt32(KernelArgs.NumTargetItems); auto Int32Ty = Type::getInt32Ty(Builder.getContext()); - constexpr const size_t MaxDim = 3; + constexpr size_t MaxDim = 3; Value *ZeroArray = Constant::getNullValue(ArrayType::get(Int32Ty, MaxDim)); Value *Flags = Builder.getInt64(KernelArgs.HasNoWait); diff --git a/llvm/lib/IR/AutoUpgrade.cpp b/llvm/lib/IR/AutoUpgrade.cpp index b838e36..58b7ddd 100644 --- a/llvm/lib/IR/AutoUpgrade.cpp +++ b/llvm/lib/IR/AutoUpgrade.cpp @@ -730,7 +730,7 @@ static bool upgradeArmOrAarch64IntrinsicFunction(bool IsArm, Function *F, // (arm|aarch64).neon.bfdot.*'. Intrinsic::ID ID = StringSwitch<Intrinsic::ID>(Name) - .Cases("v2f32.v8i8", "v4f32.v16i8", + .Cases({"v2f32.v8i8", "v4f32.v16i8"}, IsArm ? (Intrinsic::ID)Intrinsic::arm_neon_bfdot : (Intrinsic::ID)Intrinsic::aarch64_neon_bfdot) .Default(Intrinsic::not_intrinsic); @@ -1456,7 +1456,7 @@ static bool upgradeIntrinsicFunction1(Function *F, Function *&NewFn, if (F->arg_size() == 1) { Intrinsic::ID IID = StringSwitch<Intrinsic::ID>(Name) - .Cases("brev32", "brev64", Intrinsic::bitreverse) + .Cases({"brev32", "brev64"}, Intrinsic::bitreverse) .Case("clz.i", Intrinsic::ctlz) .Case("popc.i", Intrinsic::ctpop) .Default(Intrinsic::not_intrinsic); @@ -1504,6 +1504,10 @@ static bool upgradeIntrinsicFunction1(Function *F, Function *&NewFn, else if (Name.consume_front("fabs.")) // nvvm.fabs.{f,ftz.f,d} Expand = Name == "f" || Name == "ftz.f" || Name == "d"; + else if (Name.consume_front("ex2.approx.")) + // nvvm.ex2.approx.{f,ftz.f,d,f16x2} + Expand = + Name == "f" || Name == "ftz.f" || Name == "d" || Name == "f16x2"; else if (Name.consume_front("max.") || Name.consume_front("min.")) // nvvm.{min,max}.{i,ii,ui,ull} Expand = Name == "s" || Name == "i" || Name == "ll" || Name == "us" || @@ -2550,6 +2554,11 @@ static Value *upgradeNVVMIntrinsicCall(StringRef Name, CallBase *CI, Intrinsic::ID IID = (Name == "fabs.ftz.f") ? Intrinsic::nvvm_fabs_ftz : Intrinsic::nvvm_fabs; Rep = Builder.CreateUnaryIntrinsic(IID, CI->getArgOperand(0)); + } else if (Name.consume_front("ex2.approx.")) { + // nvvm.ex2.approx.{f,ftz.f,d,f16x2} + Intrinsic::ID IID = Name.starts_with("ftz") ? Intrinsic::nvvm_ex2_approx_ftz + : Intrinsic::nvvm_ex2_approx; + Rep = Builder.CreateUnaryIntrinsic(IID, CI->getArgOperand(0)); } else if (Name.starts_with("atomic.load.add.f32.p") || Name.starts_with("atomic.load.add.f64.p")) { Value *Ptr = CI->getArgOperand(0); diff --git a/llvm/lib/IR/ConstantsContext.h b/llvm/lib/IR/ConstantsContext.h index 51fb40b..e3e8d89 100644 --- a/llvm/lib/IR/ConstantsContext.h +++ b/llvm/lib/IR/ConstantsContext.h @@ -535,7 +535,7 @@ struct ConstantPtrAuthKeyType { unsigned getHash() const { return hash_combine_range(Operands); } - using TypeClass = typename ConstantInfo<ConstantPtrAuth>::TypeClass; + using TypeClass = ConstantInfo<ConstantPtrAuth>::TypeClass; ConstantPtrAuth *create(TypeClass *Ty) const { return new ConstantPtrAuth(Operands[0], cast<ConstantInt>(Operands[1]), diff --git a/llvm/lib/IR/ModuleSummaryIndex.cpp b/llvm/lib/IR/ModuleSummaryIndex.cpp index 62fd62c..3394754 100644 --- a/llvm/lib/IR/ModuleSummaryIndex.cpp +++ b/llvm/lib/IR/ModuleSummaryIndex.cpp @@ -34,8 +34,6 @@ static cl::opt<bool> ImportConstantsWithRefs( "import-constants-with-refs", cl::init(true), cl::Hidden, cl::desc("Import constant global variables with references")); -constexpr uint32_t FunctionSummary::ParamAccess::RangeWidth; - FunctionSummary FunctionSummary::ExternalNode = FunctionSummary::makeDummyFunctionSummary( SmallVector<FunctionSummary::EdgeTy, 0>()); @@ -88,8 +86,6 @@ std::pair<unsigned, unsigned> FunctionSummary::specialRefCounts() const { return {RORefCnt, WORefCnt}; } -constexpr uint64_t ModuleSummaryIndex::BitcodeSummaryVersion; - uint64_t ModuleSummaryIndex::getFlags() const { uint64_t Flags = 0; // Flags & 0x4 is reserved. DO NOT REUSE. diff --git a/llvm/lib/MC/GOFFObjectWriter.cpp b/llvm/lib/MC/GOFFObjectWriter.cpp index 71bd397..a3eaaa7 100644 --- a/llvm/lib/MC/GOFFObjectWriter.cpp +++ b/llvm/lib/MC/GOFFObjectWriter.cpp @@ -520,7 +520,7 @@ GOFFObjectWriter::GOFFObjectWriter( std::unique_ptr<MCGOFFObjectTargetWriter> MOTW, raw_pwrite_stream &OS) : TargetObjectWriter(std::move(MOTW)), OS(OS) {} -GOFFObjectWriter::~GOFFObjectWriter() {} +GOFFObjectWriter::~GOFFObjectWriter() = default; uint64_t GOFFObjectWriter::writeObject() { uint64_t Size = GOFFWriter(OS, *Asm).writeObject(); diff --git a/llvm/lib/MC/MCDXContainerWriter.cpp b/llvm/lib/MC/MCDXContainerWriter.cpp index 5eda039..ebed411 100644 --- a/llvm/lib/MC/MCDXContainerWriter.cpp +++ b/llvm/lib/MC/MCDXContainerWriter.cpp @@ -16,7 +16,7 @@ using namespace llvm; -MCDXContainerTargetWriter::~MCDXContainerTargetWriter() {} +MCDXContainerTargetWriter::~MCDXContainerTargetWriter() = default; uint64_t DXContainerObjectWriter::writeObject() { auto &Asm = *this->Asm; diff --git a/llvm/lib/MC/MCGOFFStreamer.cpp b/llvm/lib/MC/MCGOFFStreamer.cpp index 8b228db..ad6397b 100644 --- a/llvm/lib/MC/MCGOFFStreamer.cpp +++ b/llvm/lib/MC/MCGOFFStreamer.cpp @@ -20,7 +20,7 @@ using namespace llvm; -MCGOFFStreamer::~MCGOFFStreamer() {} +MCGOFFStreamer::~MCGOFFStreamer() = default; GOFFObjectWriter &MCGOFFStreamer::getWriter() { return static_cast<GOFFObjectWriter &>(getAssembler().getWriter()); diff --git a/llvm/lib/MC/MCParser/ELFAsmParser.cpp b/llvm/lib/MC/MCParser/ELFAsmParser.cpp index 1a3752f..911d92c 100644 --- a/llvm/lib/MC/MCParser/ELFAsmParser.cpp +++ b/llvm/lib/MC/MCParser/ELFAsmParser.cpp @@ -695,15 +695,15 @@ bool ELFAsmParser::parseDirectivePrevious(StringRef DirName, SMLoc) { static MCSymbolAttr MCAttrForString(StringRef Type) { return StringSwitch<MCSymbolAttr>(Type) - .Cases("STT_FUNC", "function", MCSA_ELF_TypeFunction) - .Cases("STT_OBJECT", "object", MCSA_ELF_TypeObject) - .Cases("STT_TLS", "tls_object", MCSA_ELF_TypeTLS) - .Cases("STT_COMMON", "common", MCSA_ELF_TypeCommon) - .Cases("STT_NOTYPE", "notype", MCSA_ELF_TypeNoType) - .Cases("STT_GNU_IFUNC", "gnu_indirect_function", - MCSA_ELF_TypeIndFunction) - .Case("gnu_unique_object", MCSA_ELF_TypeGnuUniqueObject) - .Default(MCSA_Invalid); + .Cases({"STT_FUNC", "function"}, MCSA_ELF_TypeFunction) + .Cases({"STT_OBJECT", "object"}, MCSA_ELF_TypeObject) + .Cases({"STT_TLS", "tls_object"}, MCSA_ELF_TypeTLS) + .Cases({"STT_COMMON", "common"}, MCSA_ELF_TypeCommon) + .Cases({"STT_NOTYPE", "notype"}, MCSA_ELF_TypeNoType) + .Cases({"STT_GNU_IFUNC", "gnu_indirect_function"}, + MCSA_ELF_TypeIndFunction) + .Case("gnu_unique_object", MCSA_ELF_TypeGnuUniqueObject) + .Default(MCSA_Invalid); } /// parseDirectiveELFType diff --git a/llvm/lib/MC/MCParser/MasmParser.cpp b/llvm/lib/MC/MCParser/MasmParser.cpp index 3462954..3a85770 100644 --- a/llvm/lib/MC/MCParser/MasmParser.cpp +++ b/llvm/lib/MC/MCParser/MasmParser.cpp @@ -5323,10 +5323,10 @@ void MasmParser::initializeDirectiveKindMap() { bool MasmParser::isMacroLikeDirective() { if (getLexer().is(AsmToken::Identifier)) { bool IsMacroLike = StringSwitch<bool>(getTok().getIdentifier()) - .CasesLower("repeat", "rept", true) + .CasesLower({"repeat", "rept"}, true) .CaseLower("while", true) - .CasesLower("for", "irp", true) - .CasesLower("forc", "irpc", true) + .CasesLower({"for", "irp"}, true) + .CasesLower({"forc", "irpc"}, true) .Default(false); if (IsMacroLike) return true; diff --git a/llvm/lib/ObjCopy/COFF/COFFWriter.h b/llvm/lib/ObjCopy/COFF/COFFWriter.h index 66d7f01..3ee0e06 100644 --- a/llvm/lib/ObjCopy/COFF/COFFWriter.h +++ b/llvm/lib/ObjCopy/COFF/COFFWriter.h @@ -50,7 +50,7 @@ class COFFWriter { Expected<uint32_t> virtualAddressToFileAddress(uint32_t RVA); public: - virtual ~COFFWriter() {} + virtual ~COFFWriter() = default; Error write(); COFFWriter(Object &Obj, raw_ostream &Out) diff --git a/llvm/lib/ObjCopy/ELF/ELFObject.h b/llvm/lib/ObjCopy/ELF/ELFObject.h index 4f6473f..2783ef27 100644 --- a/llvm/lib/ObjCopy/ELF/ELFObject.h +++ b/llvm/lib/ObjCopy/ELF/ELFObject.h @@ -134,7 +134,7 @@ private: using Elf_Sym = typename ELFT::Sym; public: - ~ELFSectionWriter() override {} + ~ELFSectionWriter() override = default; Error visit(const SymbolTableSection &Sec) override; Error visit(const RelocationSection &Sec) override; Error visit(const GnuDebugLinkSection &Sec) override; @@ -180,7 +180,7 @@ public: class BinarySectionWriter : public SectionWriter { public: - ~BinarySectionWriter() override {} + ~BinarySectionWriter() override = default; Error visit(const SymbolTableSection &Sec) override; Error visit(const RelocationSection &Sec) override; @@ -346,7 +346,7 @@ private: size_t totalSize() const; public: - ~ELFWriter() override {} + ~ELFWriter() override = default; bool WriteSectionHeaders; // For --only-keep-debug, select an alternative section/segment layout @@ -367,7 +367,7 @@ private: uint64_t TotalSize = 0; public: - ~BinaryWriter() override {} + ~BinaryWriter() override = default; Error finalize() override; Error write() override; BinaryWriter(Object &Obj, raw_ostream &Out, const CommonConfig &Config) @@ -784,7 +784,7 @@ private: SymbolTableSection *Symbols = nullptr; public: - ~SectionIndexSection() override {} + ~SectionIndexSection() override = default; void addIndex(uint32_t Index) { assert(Size > 0); Indexes.push_back(Index); diff --git a/llvm/lib/ObjCopy/MachO/MachOReader.h b/llvm/lib/ObjCopy/MachO/MachOReader.h index e315e6fd..940ba4c 100644 --- a/llvm/lib/ObjCopy/MachO/MachOReader.h +++ b/llvm/lib/ObjCopy/MachO/MachOReader.h @@ -23,7 +23,7 @@ namespace macho { // raw binaries and regular MachO object files. class Reader { public: - virtual ~Reader(){}; + virtual ~Reader() = default; virtual Expected<std::unique_ptr<Object>> create() const = 0; }; diff --git a/llvm/lib/ObjCopy/XCOFF/XCOFFWriter.h b/llvm/lib/ObjCopy/XCOFF/XCOFFWriter.h index 8620548..47639ad 100644 --- a/llvm/lib/ObjCopy/XCOFF/XCOFFWriter.h +++ b/llvm/lib/ObjCopy/XCOFF/XCOFFWriter.h @@ -20,7 +20,7 @@ namespace xcoff { class XCOFFWriter { public: - virtual ~XCOFFWriter() {} + virtual ~XCOFFWriter() = default; XCOFFWriter(Object &Obj, raw_ostream &Out) : Obj(Obj), Out(Out) {} Error write(); diff --git a/llvm/lib/Object/WindowsMachineFlag.cpp b/llvm/lib/Object/WindowsMachineFlag.cpp index caf357e8..14c14f6 100644 --- a/llvm/lib/Object/WindowsMachineFlag.cpp +++ b/llvm/lib/Object/WindowsMachineFlag.cpp @@ -23,8 +23,8 @@ using namespace llvm; COFF::MachineTypes llvm::getMachineType(StringRef S) { // Flags must be a superset of Microsoft lib.exe /machine flags. return StringSwitch<COFF::MachineTypes>(S.lower()) - .Cases("x64", "amd64", COFF::IMAGE_FILE_MACHINE_AMD64) - .Cases("x86", "i386", COFF::IMAGE_FILE_MACHINE_I386) + .Cases({"x64", "amd64"}, COFF::IMAGE_FILE_MACHINE_AMD64) + .Cases({"x86", "i386"}, COFF::IMAGE_FILE_MACHINE_I386) .Case("arm", COFF::IMAGE_FILE_MACHINE_ARMNT) .Case("arm64", COFF::IMAGE_FILE_MACHINE_ARM64) .Case("arm64ec", COFF::IMAGE_FILE_MACHINE_ARM64EC) diff --git a/llvm/lib/ObjectYAML/GOFFYAML.cpp b/llvm/lib/ObjectYAML/GOFFYAML.cpp index 60bc1f7..ecd7fb6 100644 --- a/llvm/lib/ObjectYAML/GOFFYAML.cpp +++ b/llvm/lib/ObjectYAML/GOFFYAML.cpp @@ -15,7 +15,7 @@ namespace llvm { namespace GOFFYAML { -Object::Object() {} +Object::Object() = default; } // namespace GOFFYAML diff --git a/llvm/lib/Passes/StandardInstrumentations.cpp b/llvm/lib/Passes/StandardInstrumentations.cpp index 7290a86..6b7e980 100644 --- a/llvm/lib/Passes/StandardInstrumentations.cpp +++ b/llvm/lib/Passes/StandardInstrumentations.cpp @@ -537,7 +537,7 @@ void IRChangedPrinter::handleAfter(StringRef PassID, std::string &Name, Out << "*** IR Dump After " << PassID << " on " << Name << " ***\n" << After; } -IRChangedTester::~IRChangedTester() {} +IRChangedTester::~IRChangedTester() = default; void IRChangedTester::registerCallbacks(PassInstrumentationCallbacks &PIC) { if (TestChanged != "") @@ -1566,7 +1566,7 @@ void InLineChangePrinter::registerCallbacks(PassInstrumentationCallbacks &PIC) { TextChangeReporter<IRDataT<EmptyData>>::registerRequiredCallbacks(PIC); } -TimeProfilingPassesHandler::TimeProfilingPassesHandler() {} +TimeProfilingPassesHandler::TimeProfilingPassesHandler() = default; void TimeProfilingPassesHandler::registerCallbacks( PassInstrumentationCallbacks &PIC) { diff --git a/llvm/lib/Remarks/RemarkFormat.cpp b/llvm/lib/Remarks/RemarkFormat.cpp index 1c52e35..f9fd4af 100644 --- a/llvm/lib/Remarks/RemarkFormat.cpp +++ b/llvm/lib/Remarks/RemarkFormat.cpp @@ -19,7 +19,7 @@ using namespace llvm::remarks; Expected<Format> llvm::remarks::parseFormat(StringRef FormatStr) { auto Result = StringSwitch<Format>(FormatStr) - .Cases("", "yaml", Format::YAML) + .Cases({"", "yaml"}, Format::YAML) .Case("bitstream", Format::Bitstream) .Default(Format::Unknown); diff --git a/llvm/lib/SandboxIR/Context.cpp b/llvm/lib/SandboxIR/Context.cpp index fb6ff62..6f5d072 100644 --- a/llvm/lib/SandboxIR/Context.cpp +++ b/llvm/lib/SandboxIR/Context.cpp @@ -637,7 +637,7 @@ Context::Context(LLVMContext &LLVMCtx) : LLVMCtx(LLVMCtx), IRTracker(*this), LLVMIRBuilder(LLVMCtx, ConstantFolder()) {} -Context::~Context() {} +Context::~Context() = default; void Context::clear() { // TODO: Ideally we should clear only function-scope objects, and keep global diff --git a/llvm/lib/Support/AArch64BuildAttributes.cpp b/llvm/lib/Support/AArch64BuildAttributes.cpp index 4a6b2fd..be4d1f1 100644 --- a/llvm/lib/Support/AArch64BuildAttributes.cpp +++ b/llvm/lib/Support/AArch64BuildAttributes.cpp @@ -67,8 +67,8 @@ StringRef AArch64BuildAttributes::getTypeStr(unsigned Type) { } SubsectionType AArch64BuildAttributes::getTypeID(StringRef Type) { return StringSwitch<SubsectionType>(Type) - .Cases("uleb128", "ULEB128", ULEB128) - .Cases("ntbs", "NTBS", NTBS) + .Cases({"uleb128", "ULEB128"}, ULEB128) + .Cases({"ntbs", "NTBS"}, NTBS) .Default(TYPE_NOT_FOUND); } StringRef AArch64BuildAttributes::getSubsectionTypeUnknownError() { diff --git a/llvm/lib/Support/APFloat.cpp b/llvm/lib/Support/APFloat.cpp index e21cf8e..e2645fa 100644 --- a/llvm/lib/Support/APFloat.cpp +++ b/llvm/lib/Support/APFloat.cpp @@ -269,12 +269,6 @@ bool APFloatBase::isRepresentableBy(const fltSemantics &A, A.precision <= B.precision; } -constexpr RoundingMode APFloatBase::rmNearestTiesToEven; -constexpr RoundingMode APFloatBase::rmTowardPositive; -constexpr RoundingMode APFloatBase::rmTowardNegative; -constexpr RoundingMode APFloatBase::rmTowardZero; -constexpr RoundingMode APFloatBase::rmNearestTiesToAway; - /* A tight upper bound on number of parts required to hold the value pow(5, power) is diff --git a/llvm/lib/Support/BranchProbability.cpp b/llvm/lib/Support/BranchProbability.cpp index ea42f34..143e58a 100644 --- a/llvm/lib/Support/BranchProbability.cpp +++ b/llvm/lib/Support/BranchProbability.cpp @@ -20,8 +20,6 @@ using namespace llvm; -constexpr uint32_t BranchProbability::D; - raw_ostream &BranchProbability::print(raw_ostream &OS) const { if (isUnknown()) return OS << "?%"; diff --git a/llvm/lib/Support/CommandLine.cpp b/llvm/lib/Support/CommandLine.cpp index 9491ec0..de5bd79 100644 --- a/llvm/lib/Support/CommandLine.cpp +++ b/llvm/lib/Support/CommandLine.cpp @@ -382,7 +382,7 @@ public: RegisteredSubCommands.erase(sub); } - iterator_range<typename SmallPtrSet<SubCommand *, 4>::iterator> + iterator_range<SmallPtrSet<SubCommand *, 4>::iterator> getRegisteredSubcommands() { return make_range(RegisteredSubCommands.begin(), RegisteredSubCommands.end()); @@ -2830,7 +2830,7 @@ StringMap<Option *> &cl::getRegisteredOptions(SubCommand &Sub) { return Sub.OptionsMap; } -iterator_range<typename SmallPtrSet<SubCommand *, 4>::iterator> +iterator_range<SmallPtrSet<SubCommand *, 4>::iterator> cl::getRegisteredSubcommands() { return GlobalParser->getRegisteredSubcommands(); } diff --git a/llvm/lib/Support/StringRef.cpp b/llvm/lib/Support/StringRef.cpp index b6a2f8a..2e8fba8 100644 --- a/llvm/lib/Support/StringRef.cpp +++ b/llvm/lib/Support/StringRef.cpp @@ -17,11 +17,6 @@ using namespace llvm; -// MSVC emits references to this into the translation units which reference it. -#ifndef _MSC_VER -constexpr size_t StringRef::npos; -#endif - // strncasecmp() is not available on non-POSIX systems, so define an // alternative function here. static int ascii_strncasecmp(StringRef LHS, StringRef RHS) { diff --git a/llvm/lib/Support/UnicodeNameToCodepoint.cpp b/llvm/lib/Support/UnicodeNameToCodepoint.cpp index 6f8e091..8f0d24e 100644 --- a/llvm/lib/Support/UnicodeNameToCodepoint.cpp +++ b/llvm/lib/Support/UnicodeNameToCodepoint.cpp @@ -251,10 +251,10 @@ constexpr const char *const HangulSyllables[][3] = { // Unicode 15.0 // 3.12 Conjoining Jamo Behavior Common constants -constexpr const char32_t SBase = 0xAC00; -constexpr const uint32_t LCount = 19; -constexpr const uint32_t VCount = 21; -constexpr const uint32_t TCount = 28; +constexpr char32_t SBase = 0xAC00; +constexpr uint32_t LCount = 19; +constexpr uint32_t VCount = 21; +constexpr uint32_t TCount = 28; static std::size_t findSyllable(StringRef Name, bool Strict, char &PreviousInName, int &Pos, int Column) { diff --git a/llvm/lib/Support/Windows/Signals.inc b/llvm/lib/Support/Windows/Signals.inc index 648d6a5..da68994 100644 --- a/llvm/lib/Support/Windows/Signals.inc +++ b/llvm/lib/Support/Windows/Signals.inc @@ -421,8 +421,13 @@ bool sys::RemoveFileOnSignal(StringRef Filename, std::string *ErrMsg) { return true; } - if (FilesToRemove == NULL) + if (FilesToRemove == NULL) { FilesToRemove = new std::vector<std::string>; + std::atexit([]() { + delete FilesToRemove; + FilesToRemove = NULL; + }); + } FilesToRemove->push_back(std::string(Filename)); diff --git a/llvm/lib/Support/raw_ostream.cpp b/llvm/lib/Support/raw_ostream.cpp index 07b9989..d6f27fb 100644 --- a/llvm/lib/Support/raw_ostream.cpp +++ b/llvm/lib/Support/raw_ostream.cpp @@ -61,17 +61,6 @@ using namespace llvm; -constexpr raw_ostream::Colors raw_ostream::BLACK; -constexpr raw_ostream::Colors raw_ostream::RED; -constexpr raw_ostream::Colors raw_ostream::GREEN; -constexpr raw_ostream::Colors raw_ostream::YELLOW; -constexpr raw_ostream::Colors raw_ostream::BLUE; -constexpr raw_ostream::Colors raw_ostream::MAGENTA; -constexpr raw_ostream::Colors raw_ostream::CYAN; -constexpr raw_ostream::Colors raw_ostream::WHITE; -constexpr raw_ostream::Colors raw_ostream::SAVEDCOLOR; -constexpr raw_ostream::Colors raw_ostream::RESET; - raw_ostream::~raw_ostream() { // raw_ostream's subclasses should take care to flush the buffer // in their destructors. diff --git a/llvm/lib/Support/raw_socket_stream.cpp b/llvm/lib/Support/raw_socket_stream.cpp index 3b510d3..f716317 100644 --- a/llvm/lib/Support/raw_socket_stream.cpp +++ b/llvm/lib/Support/raw_socket_stream.cpp @@ -332,7 +332,7 @@ ListeningSocket::~ListeningSocket() { raw_socket_stream::raw_socket_stream(int SocketFD) : raw_fd_stream(SocketFD, true) {} -raw_socket_stream::~raw_socket_stream() {} +raw_socket_stream::~raw_socket_stream() = default; Expected<std::unique_ptr<raw_socket_stream>> raw_socket_stream::createConnectedUnix(StringRef SocketPath) { diff --git a/llvm/lib/TableGen/TGLexer.cpp b/llvm/lib/TableGen/TGLexer.cpp index 30eae6e..e8e6469 100644 --- a/llvm/lib/TableGen/TGLexer.cpp +++ b/llvm/lib/TableGen/TGLexer.cpp @@ -682,8 +682,10 @@ tgtok::TokKind TGLexer::LexExclaim() { .Case("instances", tgtok::XInstances) .Case("substr", tgtok::XSubstr) .Case("find", tgtok::XFind) - .Cases("setdagop", "setop", tgtok::XSetDagOp) // !setop is deprecated. - .Cases("getdagop", "getop", tgtok::XGetDagOp) // !getop is deprecated. + .Cases({"setdagop", "setop"}, + tgtok::XSetDagOp) // !setop is deprecated. + .Cases({"getdagop", "getop"}, + tgtok::XGetDagOp) // !getop is deprecated. .Case("setdagopname", tgtok::XSetDagOpName) .Case("getdagopname", tgtok::XGetDagOpName) .Case("getdagarg", tgtok::XGetDagArg) diff --git a/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp b/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp index 47c1ac4..e8352be 100644 --- a/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp +++ b/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp @@ -308,9 +308,9 @@ bool AArch64TTIImpl::areInlineCompatible(const Function *Caller, return (EffectiveCallerBits & EffectiveCalleeBits) == EffectiveCalleeBits; } -bool AArch64TTIImpl::areTypesABICompatible( - const Function *Caller, const Function *Callee, - const ArrayRef<Type *> &Types) const { +bool AArch64TTIImpl::areTypesABICompatible(const Function *Caller, + const Function *Callee, + ArrayRef<Type *> Types) const { if (!BaseT::areTypesABICompatible(Caller, Callee, Types)) return false; @@ -2227,7 +2227,7 @@ static std::optional<Instruction *> instCombineSVEPTest(InstCombiner &IC, return std::nullopt; } -template <Intrinsic::ID MulOpc, typename Intrinsic::ID FuseOpc> +template <Intrinsic::ID MulOpc, Intrinsic::ID FuseOpc> static std::optional<Instruction *> instCombineSVEVectorFuseMulAddSub(InstCombiner &IC, IntrinsicInst &II, bool MergeIntoAddendOp) { diff --git a/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.h b/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.h index fe2e849..b39546a 100644 --- a/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.h +++ b/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.h @@ -84,7 +84,7 @@ public: const Function *Callee) const override; bool areTypesABICompatible(const Function *Caller, const Function *Callee, - const ArrayRef<Type *> &Types) const override; + ArrayRef<Type *> Types) const override; unsigned getInlineCallPenalty(const Function *F, const CallBase &Call, unsigned DefaultCallPenalty) const override; diff --git a/llvm/lib/Target/AMDGPU/AMDGPU.h b/llvm/lib/Target/AMDGPU/AMDGPU.h index cd8b249..67042b7 100644 --- a/llvm/lib/Target/AMDGPU/AMDGPU.h +++ b/llvm/lib/Target/AMDGPU/AMDGPU.h @@ -69,7 +69,7 @@ FunctionPass *createAMDGPUPreloadKernArgPrologLegacyPass(); ModulePass *createAMDGPUPreloadKernelArgumentsLegacyPass(const TargetMachine *); struct AMDGPUSimplifyLibCallsPass : PassInfoMixin<AMDGPUSimplifyLibCallsPass> { - AMDGPUSimplifyLibCallsPass() {} + AMDGPUSimplifyLibCallsPass() = default; PreservedAnalyses run(Function &F, FunctionAnalysisManager &AM); }; @@ -371,13 +371,13 @@ public: class AMDGPUAnnotateUniformValuesPass : public PassInfoMixin<AMDGPUAnnotateUniformValuesPass> { public: - AMDGPUAnnotateUniformValuesPass() {} + AMDGPUAnnotateUniformValuesPass() = default; PreservedAnalyses run(Function &F, FunctionAnalysisManager &AM); }; class SIModeRegisterPass : public PassInfoMixin<SIModeRegisterPass> { public: - SIModeRegisterPass() {} + SIModeRegisterPass() = default; PreservedAnalyses run(MachineFunction &F, MachineFunctionAnalysisManager &AM); }; diff --git a/llvm/lib/Target/AMDGPU/AMDGPUArgumentUsageInfo.h b/llvm/lib/Target/AMDGPU/AMDGPUArgumentUsageInfo.h index 1064e57..dad94b8 100644 --- a/llvm/lib/Target/AMDGPU/AMDGPUArgumentUsageInfo.h +++ b/llvm/lib/Target/AMDGPU/AMDGPUArgumentUsageInfo.h @@ -96,7 +96,7 @@ inline raw_ostream &operator<<(raw_ostream &OS, const ArgDescriptor &Arg) { } struct KernArgPreloadDescriptor : public ArgDescriptor { - KernArgPreloadDescriptor() {} + KernArgPreloadDescriptor() = default; SmallVector<MCRegister> Regs; }; diff --git a/llvm/lib/Target/AMDGPU/AMDGPUAttributor.cpp b/llvm/lib/Target/AMDGPU/AMDGPUAttributor.cpp index 9907c88f..8669978 100644 --- a/llvm/lib/Target/AMDGPU/AMDGPUAttributor.cpp +++ b/llvm/lib/Target/AMDGPU/AMDGPUAttributor.cpp @@ -1555,7 +1555,7 @@ private: AMDGPU::ClusterDimsAttr Attr; - static constexpr const char AttrName[] = "amdgpu-cluster-dims"; + static constexpr char AttrName[] = "amdgpu-cluster-dims"; }; AAAMDGPUClusterDims & diff --git a/llvm/lib/Target/AMDGPU/AMDGPUPerfHintAnalysis.h b/llvm/lib/Target/AMDGPU/AMDGPUPerfHintAnalysis.h index cf2ab825..a3be0f5 100644 --- a/llvm/lib/Target/AMDGPU/AMDGPUPerfHintAnalysis.h +++ b/llvm/lib/Target/AMDGPU/AMDGPUPerfHintAnalysis.h @@ -48,7 +48,7 @@ private: FuncInfoMap FIM; public: - AMDGPUPerfHintAnalysis() {} + AMDGPUPerfHintAnalysis() = default; // OldPM bool runOnSCC(const GCNTargetMachine &TM, CallGraphSCC &SCC); diff --git a/llvm/lib/Target/AMDGPU/AMDGPURegBankLegalizeRules.cpp b/llvm/lib/Target/AMDGPU/AMDGPURegBankLegalizeRules.cpp index 103cdec..dd474ac 100644 --- a/llvm/lib/Target/AMDGPU/AMDGPURegBankLegalizeRules.cpp +++ b/llvm/lib/Target/AMDGPU/AMDGPURegBankLegalizeRules.cpp @@ -202,7 +202,7 @@ bool PredicateMapping::match(const MachineInstr &MI, return true; } -SetOfRulesForOpcode::SetOfRulesForOpcode() {} +SetOfRulesForOpcode::SetOfRulesForOpcode() = default; SetOfRulesForOpcode::SetOfRulesForOpcode(FastRulesTypes FastTypes) : FastTypes(FastTypes) {} diff --git a/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp b/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp index b28c50e..b87b54f 100644 --- a/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp +++ b/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp @@ -816,7 +816,7 @@ parseAMDGPUAtomicOptimizerStrategy(StringRef Params) { Params.consume_front("strategy="); auto Result = StringSwitch<std::optional<ScanOptions>>(Params) .Case("dpp", ScanOptions::DPP) - .Cases("iterative", "", ScanOptions::Iterative) + .Cases({"iterative", ""}, ScanOptions::Iterative) .Case("none", ScanOptions::None) .Default(std::nullopt); if (Result) diff --git a/llvm/lib/Target/AMDGPU/AMDGPUWaitSGPRHazards.cpp b/llvm/lib/Target/AMDGPU/AMDGPUWaitSGPRHazards.cpp index 61c5dcd..ded2f5a 100644 --- a/llvm/lib/Target/AMDGPU/AMDGPUWaitSGPRHazards.cpp +++ b/llvm/lib/Target/AMDGPU/AMDGPUWaitSGPRHazards.cpp @@ -54,7 +54,7 @@ public: bool CullSGPRHazardsAtMemWait; unsigned CullSGPRHazardsMemWaitThreshold; - AMDGPUWaitSGPRHazards() {} + AMDGPUWaitSGPRHazards() = default; // Return the numeric ID 0-127 for a given SGPR. static std::optional<unsigned> sgprNumber(Register Reg, diff --git a/llvm/lib/Target/AMDGPU/GCNSchedStrategy.h b/llvm/lib/Target/AMDGPU/GCNSchedStrategy.h index 975781f..f357981 100644 --- a/llvm/lib/Target/AMDGPU/GCNSchedStrategy.h +++ b/llvm/lib/Target/AMDGPU/GCNSchedStrategy.h @@ -183,7 +183,7 @@ class ScheduleMetrics { unsigned BubbleCycles; public: - ScheduleMetrics() {} + ScheduleMetrics() = default; ScheduleMetrics(unsigned L, unsigned BC) : ScheduleLength(L), BubbleCycles(BC) {} unsigned getLength() const { return ScheduleLength; } @@ -217,7 +217,7 @@ class RegionPressureMap { bool IsLiveOut; public: - RegionPressureMap() {} + RegionPressureMap() = default; RegionPressureMap(GCNScheduleDAGMILive *GCNDAG, bool LiveOut) : DAG(GCNDAG), IsLiveOut(LiveOut) {} // Build the Instr->LiveReg and RegionIdx->Instr maps diff --git a/llvm/lib/Target/AMDGPU/SIInsertWaitcnts.cpp b/llvm/lib/Target/AMDGPU/SIInsertWaitcnts.cpp index 6dcbced..b7fa899 100644 --- a/llvm/lib/Target/AMDGPU/SIInsertWaitcnts.cpp +++ b/llvm/lib/Target/AMDGPU/SIInsertWaitcnts.cpp @@ -1288,18 +1288,38 @@ void WaitcntBrackets::applyWaitcnt(InstCounterType T, unsigned Count) { } void WaitcntBrackets::applyXcnt(const AMDGPU::Waitcnt &Wait) { + // On entry to a block with multiple predescessors, there may + // be pending SMEM and VMEM events active at the same time. + // In such cases, only clear one active event at a time. + auto applyPendingXcntGroup = [this](unsigned E) { + unsigned LowerBound = getScoreLB(X_CNT); + applyWaitcnt(X_CNT, 0); + PendingEvents |= (1 << E); + setScoreLB(X_CNT, LowerBound); + }; + // Wait on XCNT is redundant if we are already waiting for a load to complete. // SMEM can return out of order, so only omit XCNT wait if we are waiting till // zero. - if (Wait.KmCnt == 0 && hasPendingEvent(SMEM_GROUP)) - return applyWaitcnt(X_CNT, 0); + if (Wait.KmCnt == 0 && hasPendingEvent(SMEM_GROUP)) { + if (hasPendingEvent(VMEM_GROUP)) + applyPendingXcntGroup(VMEM_GROUP); + else + applyWaitcnt(X_CNT, 0); + return; + } // If we have pending store we cannot optimize XCnt because we do not wait for // stores. VMEM loads retun in order, so if we only have loads XCnt is // decremented to the same number as LOADCnt. if (Wait.LoadCnt != ~0u && hasPendingEvent(VMEM_GROUP) && - !hasPendingEvent(STORE_CNT)) - return applyWaitcnt(X_CNT, std::min(Wait.XCnt, Wait.LoadCnt)); + !hasPendingEvent(STORE_CNT)) { + if (hasPendingEvent(SMEM_GROUP)) + applyPendingXcntGroup(SMEM_GROUP); + else + applyWaitcnt(X_CNT, std::min(Wait.XCnt, Wait.LoadCnt)); + return; + } applyWaitcnt(X_CNT, Wait.XCnt); } diff --git a/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp b/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp index d9f76c9..45f5919 100644 --- a/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp +++ b/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp @@ -6153,7 +6153,7 @@ bool SIInstrInfo::isLegalRegOperand(const MachineInstr &MI, unsigned OpIdx, // information. if (AMDGPU::isPackedFP32Inst(MI.getOpcode()) && AMDGPU::isGFX12Plus(ST) && MO.isReg() && RI.isSGPRReg(MRI, MO.getReg())) { - constexpr const AMDGPU::OpName OpNames[] = { + constexpr AMDGPU::OpName OpNames[] = { AMDGPU::OpName::src0, AMDGPU::OpName::src1, AMDGPU::OpName::src2}; for (auto [I, OpName] : enumerate(OpNames)) { @@ -6215,8 +6215,8 @@ bool SIInstrInfo::isLegalVSrcOperand(const MachineRegisterInfo &MRI, bool SIInstrInfo::isLegalGFX12PlusPackedMathFP32Operand( const MachineRegisterInfo &MRI, const MachineInstr &MI, unsigned SrcN, const MachineOperand *MO) const { - constexpr const unsigned NumOps = 3; - constexpr const AMDGPU::OpName OpNames[NumOps * 2] = { + constexpr unsigned NumOps = 3; + constexpr AMDGPU::OpName OpNames[NumOps * 2] = { AMDGPU::OpName::src0, AMDGPU::OpName::src1, AMDGPU::OpName::src2, AMDGPU::OpName::src0_modifiers, AMDGPU::OpName::src1_modifiers, AMDGPU::OpName::src2_modifiers}; diff --git a/llvm/lib/Target/CSKY/CSKYISelLowering.cpp b/llvm/lib/Target/CSKY/CSKYISelLowering.cpp index ab4ee55..08f196b 100644 --- a/llvm/lib/Target/CSKY/CSKYISelLowering.cpp +++ b/llvm/lib/Target/CSKY/CSKYISelLowering.cpp @@ -884,13 +884,13 @@ CSKYTargetLowering::getRegForInlineAsmConstraint(const TargetRegisterInfo *TRI, .Case("{t4}", CSKY::R20) .Case("{t5}", CSKY::R21) .Case("{t6}", CSKY::R22) - .Cases("{t7}", "{fp}", CSKY::R23) - .Cases("{t8}", "{top}", CSKY::R24) - .Cases("{t9}", "{bsp}", CSKY::R25) + .Cases({"{t7}", "{fp}"}, CSKY::R23) + .Cases({"{t8}", "{top}"}, CSKY::R24) + .Cases({"{t9}", "{bsp}"}, CSKY::R25) .Case("{r26}", CSKY::R26) .Case("{r27}", CSKY::R27) .Cases({"{gb}", "{rgb}", "{rdb}"}, CSKY::R28) - .Cases("{tb}", "{rtb}", CSKY::R29) + .Cases({"{tb}", "{rtb}"}, CSKY::R29) .Case("{svbr}", CSKY::R30) .Case("{tls}", CSKY::R31) .Default(CSKY::NoRegister); @@ -907,38 +907,38 @@ CSKYTargetLowering::getRegForInlineAsmConstraint(const TargetRegisterInfo *TRI, // use the ABI names in register constraint lists. if (Subtarget.useHardFloat()) { unsigned FReg = StringSwitch<unsigned>(Constraint.lower()) - .Cases("{fr0}", "{vr0}", CSKY::F0_32) - .Cases("{fr1}", "{vr1}", CSKY::F1_32) - .Cases("{fr2}", "{vr2}", CSKY::F2_32) - .Cases("{fr3}", "{vr3}", CSKY::F3_32) - .Cases("{fr4}", "{vr4}", CSKY::F4_32) - .Cases("{fr5}", "{vr5}", CSKY::F5_32) - .Cases("{fr6}", "{vr6}", CSKY::F6_32) - .Cases("{fr7}", "{vr7}", CSKY::F7_32) - .Cases("{fr8}", "{vr8}", CSKY::F8_32) - .Cases("{fr9}", "{vr9}", CSKY::F9_32) - .Cases("{fr10}", "{vr10}", CSKY::F10_32) - .Cases("{fr11}", "{vr11}", CSKY::F11_32) - .Cases("{fr12}", "{vr12}", CSKY::F12_32) - .Cases("{fr13}", "{vr13}", CSKY::F13_32) - .Cases("{fr14}", "{vr14}", CSKY::F14_32) - .Cases("{fr15}", "{vr15}", CSKY::F15_32) - .Cases("{fr16}", "{vr16}", CSKY::F16_32) - .Cases("{fr17}", "{vr17}", CSKY::F17_32) - .Cases("{fr18}", "{vr18}", CSKY::F18_32) - .Cases("{fr19}", "{vr19}", CSKY::F19_32) - .Cases("{fr20}", "{vr20}", CSKY::F20_32) - .Cases("{fr21}", "{vr21}", CSKY::F21_32) - .Cases("{fr22}", "{vr22}", CSKY::F22_32) - .Cases("{fr23}", "{vr23}", CSKY::F23_32) - .Cases("{fr24}", "{vr24}", CSKY::F24_32) - .Cases("{fr25}", "{vr25}", CSKY::F25_32) - .Cases("{fr26}", "{vr26}", CSKY::F26_32) - .Cases("{fr27}", "{vr27}", CSKY::F27_32) - .Cases("{fr28}", "{vr28}", CSKY::F28_32) - .Cases("{fr29}", "{vr29}", CSKY::F29_32) - .Cases("{fr30}", "{vr30}", CSKY::F30_32) - .Cases("{fr31}", "{vr31}", CSKY::F31_32) + .Cases({"{fr0}", "{vr0}"}, CSKY::F0_32) + .Cases({"{fr1}", "{vr1}"}, CSKY::F1_32) + .Cases({"{fr2}", "{vr2}"}, CSKY::F2_32) + .Cases({"{fr3}", "{vr3}"}, CSKY::F3_32) + .Cases({"{fr4}", "{vr4}"}, CSKY::F4_32) + .Cases({"{fr5}", "{vr5}"}, CSKY::F5_32) + .Cases({"{fr6}", "{vr6}"}, CSKY::F6_32) + .Cases({"{fr7}", "{vr7}"}, CSKY::F7_32) + .Cases({"{fr8}", "{vr8}"}, CSKY::F8_32) + .Cases({"{fr9}", "{vr9}"}, CSKY::F9_32) + .Cases({"{fr10}", "{vr10}"}, CSKY::F10_32) + .Cases({"{fr11}", "{vr11}"}, CSKY::F11_32) + .Cases({"{fr12}", "{vr12}"}, CSKY::F12_32) + .Cases({"{fr13}", "{vr13}"}, CSKY::F13_32) + .Cases({"{fr14}", "{vr14}"}, CSKY::F14_32) + .Cases({"{fr15}", "{vr15}"}, CSKY::F15_32) + .Cases({"{fr16}", "{vr16}"}, CSKY::F16_32) + .Cases({"{fr17}", "{vr17}"}, CSKY::F17_32) + .Cases({"{fr18}", "{vr18}"}, CSKY::F18_32) + .Cases({"{fr19}", "{vr19}"}, CSKY::F19_32) + .Cases({"{fr20}", "{vr20}"}, CSKY::F20_32) + .Cases({"{fr21}", "{vr21}"}, CSKY::F21_32) + .Cases({"{fr22}", "{vr22}"}, CSKY::F22_32) + .Cases({"{fr23}", "{vr23}"}, CSKY::F23_32) + .Cases({"{fr24}", "{vr24}"}, CSKY::F24_32) + .Cases({"{fr25}", "{vr25}"}, CSKY::F25_32) + .Cases({"{fr26}", "{vr26}"}, CSKY::F26_32) + .Cases({"{fr27}", "{vr27}"}, CSKY::F27_32) + .Cases({"{fr28}", "{vr28}"}, CSKY::F28_32) + .Cases({"{fr29}", "{vr29}"}, CSKY::F29_32) + .Cases({"{fr30}", "{vr30}"}, CSKY::F30_32) + .Cases({"{fr31}", "{vr31}"}, CSKY::F31_32) .Default(CSKY::NoRegister); if (FReg != CSKY::NoRegister) { assert(CSKY::F0_32 <= FReg && FReg <= CSKY::F31_32 && "Unknown fp-reg"); diff --git a/llvm/lib/Target/Hexagon/HexagonDepIICHVX.td b/llvm/lib/Target/Hexagon/HexagonDepIICHVX.td index f4e36fa7..e661c94 100644 --- a/llvm/lib/Target/Hexagon/HexagonDepIICHVX.td +++ b/llvm/lib/Target/Hexagon/HexagonDepIICHVX.td @@ -26,6 +26,7 @@ def tc_20a4bbec : InstrItinClass; def tc_227864f7 : InstrItinClass; def tc_257f6f7c : InstrItinClass; def tc_26a377fe : InstrItinClass; +def tc_2a698a03 : InstrItinClass; def tc_2b4c548e : InstrItinClass; def tc_2c745bb8 : InstrItinClass; def tc_2d4051cd : InstrItinClass; @@ -52,6 +53,7 @@ def tc_561aaa58 : InstrItinClass; def tc_56c4f9fe : InstrItinClass; def tc_56e64202 : InstrItinClass; def tc_58d21193 : InstrItinClass; +def tc_57a4709c : InstrItinClass; def tc_5bf8afbb : InstrItinClass; def tc_5cdf8c84 : InstrItinClass; def tc_61bf7c03 : InstrItinClass; @@ -220,6 +222,11 @@ class DepHVXItinV55 { InstrStage<1, [CVI_ALL_NOMEM]>], [9, 3, 5, 2], [HVX_FWD, Hex_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_2a698a03, /*SLOT0123,VSorVP*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_SHIFT, CVI_XLANE]>], [9, 5], + [HVX_FWD, HVX_FWD]>, + InstrItinData <tc_2b4c548e, /*SLOT23,VX_DV*/ [InstrStage<1, [SLOT2, SLOT3], 0>, InstrStage<1, [CVI_MPY01]>], [9, 5, 5, 2], @@ -356,6 +363,11 @@ class DepHVXItinV55 { InstrStage<1, [CVI_MPY01, CVI_XLSHF]>], [7, 1, 2, 7, 7], [HVX_FWD, Hex_FWD, Hex_FWD, HVX_FWD, HVX_FWD]>, + InstrItinData <tc_57a4709c, /*SLOT0123,VA*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_MPY0, CVI_MPY1, CVI_SHIFT, CVI_XLANE]>], [9, 7, 7, 2], + [HVX_FWD, HVX_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_5bf8afbb, /*SLOT0123,VP*/ [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, InstrStage<1, [CVI_XLANE]>], [9, 2], @@ -812,6 +824,11 @@ class DepHVXItinV60 { InstrStage<1, [CVI_ALL_NOMEM]>], [9, 3, 5, 2], [HVX_FWD, Hex_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_2a698a03, /*SLOT0123,VSorVP*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_SHIFT, CVI_XLANE]>], [9, 5], + [HVX_FWD, HVX_FWD]>, + InstrItinData <tc_2b4c548e, /*SLOT23,VX_DV*/ [InstrStage<1, [SLOT2, SLOT3], 0>, InstrStage<1, [CVI_MPY01]>], [9, 5, 5, 2], @@ -948,6 +965,11 @@ class DepHVXItinV60 { InstrStage<1, [CVI_MPY01, CVI_XLSHF]>], [7, 1, 2, 7, 7], [HVX_FWD, Hex_FWD, Hex_FWD, HVX_FWD, HVX_FWD]>, + InstrItinData <tc_57a4709c, /*SLOT0123,VA*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_MPY0, CVI_MPY1, CVI_SHIFT, CVI_XLANE]>], [9, 7, 7, 2], + [HVX_FWD, HVX_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_5bf8afbb, /*SLOT0123,VP*/ [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, InstrStage<1, [CVI_XLANE]>], [9, 2], @@ -1404,6 +1426,11 @@ class DepHVXItinV62 { InstrStage<1, [CVI_ALL_NOMEM]>], [9, 3, 5, 2], [HVX_FWD, Hex_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_2a698a03, /*SLOT0123,VSorVP*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_SHIFT, CVI_XLANE]>], [9, 5], + [HVX_FWD, HVX_FWD]>, + InstrItinData <tc_2b4c548e, /*SLOT23,VX_DV*/ [InstrStage<1, [SLOT2, SLOT3], 0>, InstrStage<1, [CVI_MPY01]>], [9, 5, 5, 2], @@ -1540,6 +1567,11 @@ class DepHVXItinV62 { InstrStage<1, [CVI_MPY01, CVI_XLSHF]>], [7, 1, 2, 7, 7], [HVX_FWD, Hex_FWD, Hex_FWD, HVX_FWD, HVX_FWD]>, + InstrItinData <tc_57a4709c, /*SLOT0123,VA*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_MPY0, CVI_MPY1, CVI_SHIFT, CVI_XLANE]>], [9, 7, 7, 2], + [HVX_FWD, HVX_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_5bf8afbb, /*SLOT0123,VP*/ [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, InstrStage<1, [CVI_XLANE]>], [9, 2], @@ -1996,6 +2028,11 @@ class DepHVXItinV65 { InstrStage<1, [CVI_ALL_NOMEM]>], [9, 3, 5, 2], [HVX_FWD, Hex_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_2a698a03, /*SLOT0123,VSorVP*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_SHIFT, CVI_XLANE]>], [9, 5], + [HVX_FWD, HVX_FWD]>, + InstrItinData <tc_2b4c548e, /*SLOT23,VX_DV*/ [InstrStage<1, [SLOT2, SLOT3], 0>, InstrStage<1, [CVI_MPY01]>], [9, 5, 5, 2], @@ -2132,6 +2169,11 @@ class DepHVXItinV65 { InstrStage<1, [CVI_MPY01, CVI_XLSHF]>], [7, 1, 2, 7, 7], [HVX_FWD, Hex_FWD, Hex_FWD, HVX_FWD, HVX_FWD]>, + InstrItinData <tc_57a4709c, /*SLOT0123,VA*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_MPY0, CVI_MPY1, CVI_SHIFT, CVI_XLANE]>], [9, 7, 7, 2], + [HVX_FWD, HVX_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_5bf8afbb, /*SLOT0123,VP*/ [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, InstrStage<1, [CVI_XLANE]>], [9, 2], @@ -2588,6 +2630,11 @@ class DepHVXItinV66 { InstrStage<1, [CVI_ALL_NOMEM]>], [9, 3, 5, 2], [HVX_FWD, Hex_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_2a698a03, /*SLOT0123,VSorVP*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_SHIFT, CVI_XLANE]>], [9, 5], + [HVX_FWD, HVX_FWD]>, + InstrItinData <tc_2b4c548e, /*SLOT23,VX_DV*/ [InstrStage<1, [SLOT2, SLOT3], 0>, InstrStage<1, [CVI_MPY01]>], [9, 5, 5, 2], @@ -2724,6 +2771,11 @@ class DepHVXItinV66 { InstrStage<1, [CVI_MPY01, CVI_XLSHF]>], [7, 1, 2, 7, 7], [HVX_FWD, Hex_FWD, Hex_FWD, HVX_FWD, HVX_FWD]>, + InstrItinData <tc_57a4709c, /*SLOT0123,VA*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_MPY0, CVI_MPY1, CVI_SHIFT, CVI_XLANE]>], [9, 7, 7, 2], + [HVX_FWD, HVX_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_5bf8afbb, /*SLOT0123,VP*/ [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, InstrStage<1, [CVI_XLANE]>], [9, 2], @@ -3180,6 +3232,11 @@ class DepHVXItinV67 { InstrStage<1, [CVI_ALL_NOMEM]>], [9, 3, 5, 2], [HVX_FWD, Hex_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_2a698a03, /*SLOT0123,VSorVP*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_SHIFT, CVI_XLANE]>], [9, 5], + [HVX_FWD, HVX_FWD]>, + InstrItinData <tc_2b4c548e, /*SLOT23,VX_DV*/ [InstrStage<1, [SLOT2, SLOT3], 0>, InstrStage<1, [CVI_MPY01]>], [9, 5, 5, 2], @@ -3316,6 +3373,11 @@ class DepHVXItinV67 { InstrStage<1, [CVI_MPY01, CVI_XLSHF]>], [7, 1, 2, 7, 7], [HVX_FWD, Hex_FWD, Hex_FWD, HVX_FWD, HVX_FWD]>, + InstrItinData <tc_57a4709c, /*SLOT0123,VA*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_MPY0, CVI_MPY1, CVI_SHIFT, CVI_XLANE]>], [9, 7, 7, 2], + [HVX_FWD, HVX_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_5bf8afbb, /*SLOT0123,VP*/ [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, InstrStage<1, [CVI_XLANE]>], [9, 2], @@ -3772,6 +3834,11 @@ class DepHVXItinV68 { InstrStage<1, [CVI_ALL_NOMEM]>], [9, 3, 5, 2], [HVX_FWD, Hex_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_2a698a03, /*SLOT0123,VSorVP*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_SHIFT, CVI_XLANE]>], [9, 5], + [HVX_FWD, HVX_FWD]>, + InstrItinData <tc_2b4c548e, /*SLOT23,VX_DV*/ [InstrStage<1, [SLOT2, SLOT3], 0>, InstrStage<1, [CVI_MPY01]>], [9, 5, 5, 2], @@ -3908,6 +3975,11 @@ class DepHVXItinV68 { InstrStage<1, [CVI_MPY01, CVI_XLSHF]>], [7, 1, 2, 7, 7], [HVX_FWD, Hex_FWD, Hex_FWD, HVX_FWD, HVX_FWD]>, + InstrItinData <tc_57a4709c, /*SLOT0123,VA*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_MPY0, CVI_MPY1, CVI_SHIFT, CVI_XLANE]>], [9, 7, 7, 2], + [HVX_FWD, HVX_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_5bf8afbb, /*SLOT0123,VP*/ [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, InstrStage<1, [CVI_XLANE]>], [9, 2], @@ -4364,6 +4436,11 @@ class DepHVXItinV69 { InstrStage<1, [CVI_ALL_NOMEM]>], [9, 3, 5, 2], [HVX_FWD, Hex_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_2a698a03, /*SLOT0123,VSorVP*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_SHIFT, CVI_XLANE]>], [9, 5], + [HVX_FWD, HVX_FWD]>, + InstrItinData <tc_2b4c548e, /*SLOT23,VX_DV*/ [InstrStage<1, [SLOT2, SLOT3], 0>, InstrStage<1, [CVI_MPY01]>], [9, 5, 5, 2], @@ -4500,6 +4577,11 @@ class DepHVXItinV69 { InstrStage<1, [CVI_MPY01, CVI_XLSHF]>], [7, 1, 2, 7, 7], [HVX_FWD, Hex_FWD, Hex_FWD, HVX_FWD, HVX_FWD]>, + InstrItinData <tc_57a4709c, /*SLOT0123,VA*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_MPY0, CVI_MPY1, CVI_SHIFT, CVI_XLANE]>], [9, 7, 7, 2], + [HVX_FWD, HVX_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_5bf8afbb, /*SLOT0123,VP*/ [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, InstrStage<1, [CVI_XLANE]>], [9, 2], @@ -4956,6 +5038,11 @@ class DepHVXItinV71 { InstrStage<1, [CVI_ALL_NOMEM]>], [9, 3, 5, 2], [HVX_FWD, Hex_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_2a698a03, /*SLOT0123,VSorVP*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_SHIFT, CVI_XLANE]>], [9, 5], + [HVX_FWD, HVX_FWD]>, + InstrItinData <tc_2b4c548e, /*SLOT23,VX_DV*/ [InstrStage<1, [SLOT2, SLOT3], 0>, InstrStage<1, [CVI_MPY01]>], [9, 5, 5, 2], @@ -5092,6 +5179,11 @@ class DepHVXItinV71 { InstrStage<1, [CVI_MPY01, CVI_XLSHF]>], [7, 1, 2, 7, 7], [HVX_FWD, Hex_FWD, Hex_FWD, HVX_FWD, HVX_FWD]>, + InstrItinData <tc_57a4709c, /*SLOT0123,VA*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_MPY0, CVI_MPY1, CVI_SHIFT, CVI_XLANE]>], [9, 7, 7, 2], + [HVX_FWD, HVX_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_5bf8afbb, /*SLOT0123,VP*/ [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, InstrStage<1, [CVI_XLANE]>], [9, 2], @@ -5548,6 +5640,11 @@ class DepHVXItinV73 { InstrStage<1, [CVI_ALL_NOMEM]>], [9, 3, 5, 2], [HVX_FWD, Hex_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_2a698a03, /*SLOT0123,VSorVP*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_SHIFT, CVI_XLANE]>], [9, 5], + [HVX_FWD, HVX_FWD]>, + InstrItinData <tc_2b4c548e, /*SLOT23,VX_DV*/ [InstrStage<1, [SLOT2, SLOT3], 0>, InstrStage<1, [CVI_MPY01]>], [9, 5, 5, 2], @@ -5684,6 +5781,11 @@ class DepHVXItinV73 { InstrStage<1, [CVI_MPY01, CVI_XLSHF]>], [7, 1, 2, 7, 7], [HVX_FWD, Hex_FWD, Hex_FWD, HVX_FWD, HVX_FWD]>, + InstrItinData <tc_57a4709c, /*SLOT0123,VA*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_MPY0, CVI_MPY1, CVI_SHIFT, CVI_XLANE]>], [9, 7, 7, 2], + [HVX_FWD, HVX_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_5bf8afbb, /*SLOT0123,VP*/ [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, InstrStage<1, [CVI_XLANE]>], [9, 2], @@ -6140,6 +6242,11 @@ class DepHVXItinV75 { InstrStage<1, [CVI_ALL_NOMEM]>], [9, 3, 5, 2], [HVX_FWD, Hex_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_2a698a03, /*SLOT0123,VSorVP*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_SHIFT, CVI_XLANE]>], [9, 5], + [HVX_FWD, HVX_FWD]>, + InstrItinData <tc_2b4c548e, /*SLOT23,VX_DV*/ [InstrStage<1, [SLOT2, SLOT3], 0>, InstrStage<1, [CVI_MPY01]>], [9, 5, 5, 2], @@ -6276,6 +6383,11 @@ class DepHVXItinV75 { InstrStage<1, [CVI_MPY01, CVI_XLSHF]>], [7, 1, 2, 7, 7], [HVX_FWD, Hex_FWD, Hex_FWD, HVX_FWD, HVX_FWD]>, + InstrItinData <tc_57a4709c, /*SLOT0123,VA*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_MPY0, CVI_MPY1, CVI_SHIFT, CVI_XLANE]>], [9, 7, 7, 2], + [HVX_FWD, HVX_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_5bf8afbb, /*SLOT0123,VP*/ [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, InstrStage<1, [CVI_XLANE]>], [9, 2], @@ -6732,6 +6844,11 @@ class DepHVXItinV79 { InstrStage<1, [CVI_ALL_NOMEM]>], [9, 3, 5, 2], [HVX_FWD, Hex_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_2a698a03, /*SLOT0123,VSorVP*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_SHIFT, CVI_XLANE]>], [9, 5], + [HVX_FWD, HVX_FWD]>, + InstrItinData <tc_2b4c548e, /*SLOT23,VX_DV*/ [InstrStage<1, [SLOT2, SLOT3], 0>, InstrStage<1, [CVI_MPY01]>], [9, 5, 5, 2], @@ -6868,6 +6985,11 @@ class DepHVXItinV79 { InstrStage<1, [CVI_MPY01, CVI_XLSHF]>], [7, 1, 2, 7, 7], [HVX_FWD, Hex_FWD, Hex_FWD, HVX_FWD, HVX_FWD]>, + InstrItinData <tc_57a4709c, /*SLOT0123,VA*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_MPY0, CVI_MPY1, CVI_SHIFT, CVI_XLANE]>], [9, 7, 7, 2], + [HVX_FWD, HVX_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_5bf8afbb, /*SLOT0123,VP*/ [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, InstrStage<1, [CVI_XLANE]>], [9, 2], @@ -7324,6 +7446,11 @@ class DepHVXItinV81 { InstrStage<1, [CVI_ALL_NOMEM]>], [9, 3, 5, 2], [HVX_FWD, Hex_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_2a698a03, /*SLOT0123,VSorVP*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_SHIFT, CVI_XLANE]>], [9, 5], + [HVX_FWD, HVX_FWD]>, + InstrItinData <tc_2b4c548e, /*SLOT23,VX_DV*/ [InstrStage<1, [SLOT2, SLOT3], 0>, InstrStage<1, [CVI_MPY01]>], [9, 5, 5, 2], @@ -7460,6 +7587,11 @@ class DepHVXItinV81 { InstrStage<1, [CVI_MPY01, CVI_XLSHF]>], [7, 1, 2, 7, 7], [HVX_FWD, Hex_FWD, Hex_FWD, HVX_FWD, HVX_FWD]>, + InstrItinData <tc_57a4709c, /*SLOT0123,VA*/ + [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, + InstrStage<1, [CVI_MPY0, CVI_MPY1, CVI_SHIFT, CVI_XLANE]>], [9, 7, 7, 2], + [HVX_FWD, HVX_FWD, HVX_FWD, Hex_FWD]>, + InstrItinData <tc_5bf8afbb, /*SLOT0123,VP*/ [InstrStage<1, [SLOT0, SLOT1, SLOT2, SLOT3], 0>, InstrStage<1, [CVI_XLANE]>], [9, 2], diff --git a/llvm/lib/Target/Hexagon/HexagonDepInstrInfo.td b/llvm/lib/Target/Hexagon/HexagonDepInstrInfo.td index f8f1c2a..b188134 100644 --- a/llvm/lib/Target/Hexagon/HexagonDepInstrInfo.td +++ b/llvm/lib/Target/Hexagon/HexagonDepInstrInfo.td @@ -29939,6 +29939,58 @@ let opNewValue = 0; let isCVI = 1; let DecoderNamespace = "EXT_mmvec"; } +def V6_vabs_qf16_hf : HInst< +(outs HvxVR:$Vd32), +(ins HvxVR:$Vu32), +"$Vd32.qf16 = vabs($Vu32.hf)", +tc_2a698a03, TypeCVI_VS>, Enc_e7581c, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-5} = 0b110; +let Inst{13-13} = 0b1; +let Inst{31-16} = 0b0001111000001110; +let hasNewValue = 1; +let opNewValue = 0; +let isCVI = 1; +let DecoderNamespace = "EXT_mmvec"; +} +def V6_vabs_qf16_qf16 : HInst< +(outs HvxVR:$Vd32), +(ins HvxVR:$Vu32), +"$Vd32.qf16 = vabs($Vu32.qf16)", +tc_2a698a03, TypeCVI_VS>, Enc_e7581c, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-5} = 0b111; +let Inst{13-13} = 0b1; +let Inst{31-16} = 0b0001111000001110; +let hasNewValue = 1; +let opNewValue = 0; +let isCVI = 1; +let DecoderNamespace = "EXT_mmvec"; +} +def V6_vabs_qf32_qf32 : HInst< +(outs HvxVR:$Vd32), +(ins HvxVR:$Vu32), +"$Vd32.qf32 = vabs($Vu32.qf32)", +tc_2a698a03, TypeCVI_VS>, Enc_e7581c, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-5} = 0b101; +let Inst{13-13} = 0b1; +let Inst{31-16} = 0b0001111000001110; +let hasNewValue = 1; +let opNewValue = 0; +let isCVI = 1; +let DecoderNamespace = "EXT_mmvec"; +} +def V6_vabs_qf32_sf : HInst< +(outs HvxVR:$Vd32), +(ins HvxVR:$Vu32), +"$Vd32.qf32 = vabs($Vu32.sf)", +tc_2a698a03, TypeCVI_VS>, Enc_e7581c, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-5} = 0b100; +let Inst{13-13} = 0b1; +let Inst{31-16} = 0b0001111000001110; +let hasNewValue = 1; +let opNewValue = 0; +let isCVI = 1; +let DecoderNamespace = "EXT_mmvec"; +} def V6_vabs_sf : HInst< (outs HvxVR:$Vd32), (ins HvxVR:$Vu32), @@ -31302,6 +31354,21 @@ let isPseudo = 1; let isCodeGenOnly = 1; let DecoderNamespace = "EXT_mmvec"; } +def V6_valign4 : HInst< +(outs HvxVR:$Vd32), +(ins HvxVR:$Vu32, HvxVR:$Vv32, IntRegsLow8:$Rt8), +"$Vd32 = valign4($Vu32,$Vv32,$Rt8)", +tc_57a4709c, TypeCVI_VA>, Enc_a30110, Requires<[UseHVXV81]> { +let Inst{7-5} = 0b101; +let Inst{13-13} = 0b0; +let Inst{31-24} = 0b00011000; +let hasNewValue = 1; +let opNewValue = 0; +let isCVI = 1; +let isHVXALU = 1; +let isHVXALU2SRC = 1; +let DecoderNamespace = "EXT_mmvec"; +} def V6_valignb : HInst< (outs HvxVR:$Vd32), (ins HvxVR:$Vu32, HvxVR:$Vv32, IntRegsLow8:$Rt8), @@ -32583,6 +32650,32 @@ let isCVI = 1; let hasHvxTmp = 1; let DecoderNamespace = "EXT_mmvec"; } +def V6_vconv_bf_qf32 : HInst< +(outs HvxVR:$Vd32), +(ins HvxWR:$Vuu32), +"$Vd32.bf = $Vuu32.qf32", +tc_2a698a03, TypeCVI_VS>, Enc_a33d04, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-5} = 0b111; +let Inst{13-13} = 0b1; +let Inst{31-16} = 0b0001111000000110; +let hasNewValue = 1; +let opNewValue = 0; +let isCVI = 1; +let DecoderNamespace = "EXT_mmvec"; +} +def V6_vconv_f8_qf16 : HInst< +(outs HvxVR:$Vd32), +(ins HvxVR:$Vu32), +"$Vd32.f8 = $Vu32.qf16", +tc_2a698a03, TypeCVI_VS>, Enc_e7581c, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-5} = 0b111; +let Inst{13-13} = 0b1; +let Inst{31-16} = 0b0001111000001100; +let hasNewValue = 1; +let opNewValue = 0; +let isCVI = 1; +let DecoderNamespace = "EXT_mmvec"; +} def V6_vconv_h_hf : HInst< (outs HvxVR:$Vd32), (ins HvxVR:$Vu32), @@ -32596,6 +32689,19 @@ let opNewValue = 0; let isCVI = 1; let DecoderNamespace = "EXT_mmvec"; } +def V6_vconv_h_hf_rnd : HInst< +(outs HvxVR:$Vd32), +(ins HvxVR:$Vu32), +"$Vd32.h = $Vu32.hf:rnd", +tc_2a698a03, TypeCVI_VS>, Enc_e7581c, Requires<[UseHVXV81]> { +let Inst{7-5} = 0b110; +let Inst{13-13} = 0b1; +let Inst{31-16} = 0b0001111000000110; +let hasNewValue = 1; +let opNewValue = 0; +let isCVI = 1; +let DecoderNamespace = "EXT_mmvec"; +} def V6_vconv_hf_h : HInst< (outs HvxVR:$Vd32), (ins HvxVR:$Vu32), @@ -32635,6 +32741,71 @@ let opNewValue = 0; let isCVI = 1; let DecoderNamespace = "EXT_mmvec"; } +def V6_vconv_qf16_f8 : HInst< +(outs HvxWR:$Vdd32), +(ins HvxVR:$Vu32), +"$Vdd32.qf16 = $Vu32.f8", +tc_04da405a, TypeCVI_VP_VS>, Enc_dd766a, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-5} = 0b101; +let Inst{13-13} = 0b1; +let Inst{31-16} = 0b0001111000001100; +let hasNewValue = 1; +let opNewValue = 0; +let isCVI = 1; +let DecoderNamespace = "EXT_mmvec"; +} +def V6_vconv_qf16_hf : HInst< +(outs HvxVR:$Vd32), +(ins HvxVR:$Vu32), +"$Vd32.qf16 = $Vu32.hf", +tc_2a698a03, TypeCVI_VS>, Enc_e7581c, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-5} = 0b100; +let Inst{13-13} = 0b1; +let Inst{31-16} = 0b0001111000001100; +let hasNewValue = 1; +let opNewValue = 0; +let isCVI = 1; +let DecoderNamespace = "EXT_mmvec"; +} +def V6_vconv_qf16_qf16 : HInst< +(outs HvxVR:$Vd32), +(ins HvxVR:$Vu32), +"$Vd32.qf16 = $Vu32.qf16", +tc_2a698a03, TypeCVI_VS>, Enc_e7581c, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-5} = 0b110; +let Inst{13-13} = 0b1; +let Inst{31-16} = 0b0001111000001100; +let hasNewValue = 1; +let opNewValue = 0; +let isCVI = 1; +let DecoderNamespace = "EXT_mmvec"; +} +def V6_vconv_qf32_qf32 : HInst< +(outs HvxVR:$Vd32), +(ins HvxVR:$Vu32), +"$Vd32.qf32 = $Vu32.qf32", +tc_2a698a03, TypeCVI_VS>, Enc_e7581c, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-5} = 0b111; +let Inst{13-13} = 0b1; +let Inst{31-16} = 0b0001111000001101; +let hasNewValue = 1; +let opNewValue = 0; +let isCVI = 1; +let DecoderNamespace = "EXT_mmvec"; +} +def V6_vconv_qf32_sf : HInst< +(outs HvxVR:$Vd32), +(ins HvxVR:$Vu32), +"$Vd32.qf32 = $Vu32.sf", +tc_2a698a03, TypeCVI_VS>, Enc_e7581c, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-5} = 0b110; +let Inst{13-13} = 0b1; +let Inst{31-16} = 0b0001111000001101; +let hasNewValue = 1; +let opNewValue = 0; +let isCVI = 1; +let DecoderNamespace = "EXT_mmvec"; +} def V6_vconv_sf_qf32 : HInst< (outs HvxVR:$Vd32), (ins HvxVR:$Vu32), @@ -33720,6 +33891,122 @@ let isHVXALU2SRC = 1; let DecoderNamespace = "EXT_mmvec"; let Constraints = "$Qx4 = $Qx4in"; } +def V6_veqhf : HInst< +(outs HvxQR:$Qd4), +(ins HvxVR:$Vu32, HvxVR:$Vv32), +"$Qd4 = vcmp.eq($Vu32.hf,$Vv32.hf)", +tc_56c4f9fe, TypeCVI_VA>, Enc_95441f, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-2} = 0b000111; +let Inst{13-13} = 0b0; +let Inst{31-21} = 0b00011111100; +let hasNewValue = 1; +let opNewValue = 0; +let isCVI = 1; +let isHVXALU = 1; +let isHVXALU2SRC = 1; +let DecoderNamespace = "EXT_mmvec"; +} +def V6_veqhf_and : HInst< +(outs HvxQR:$Qx4), +(ins HvxQR:$Qx4in, HvxVR:$Vu32, HvxVR:$Vv32), +"$Qx4 &= vcmp.eq($Vu32.hf,$Vv32.hf)", +tc_257f6f7c, TypeCVI_VA>, Enc_eaa9f8, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-2} = 0b000111; +let Inst{13-13} = 0b1; +let Inst{31-21} = 0b00011100100; +let isCVI = 1; +let isHVXALU = 1; +let isHVXALU2SRC = 1; +let DecoderNamespace = "EXT_mmvec"; +let Constraints = "$Qx4 = $Qx4in"; +} +def V6_veqhf_or : HInst< +(outs HvxQR:$Qx4), +(ins HvxQR:$Qx4in, HvxVR:$Vu32, HvxVR:$Vv32), +"$Qx4 |= vcmp.eq($Vu32.hf,$Vv32.hf)", +tc_257f6f7c, TypeCVI_VA>, Enc_eaa9f8, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-2} = 0b010111; +let Inst{13-13} = 0b1; +let Inst{31-21} = 0b00011100100; +let isAccumulator = 1; +let isCVI = 1; +let isHVXALU = 1; +let isHVXALU2SRC = 1; +let DecoderNamespace = "EXT_mmvec"; +let Constraints = "$Qx4 = $Qx4in"; +} +def V6_veqhf_xor : HInst< +(outs HvxQR:$Qx4), +(ins HvxQR:$Qx4in, HvxVR:$Vu32, HvxVR:$Vv32), +"$Qx4 ^= vcmp.eq($Vu32.hf,$Vv32.hf)", +tc_257f6f7c, TypeCVI_VA>, Enc_eaa9f8, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-2} = 0b100111; +let Inst{13-13} = 0b1; +let Inst{31-21} = 0b00011100100; +let isCVI = 1; +let isHVXALU = 1; +let isHVXALU2SRC = 1; +let DecoderNamespace = "EXT_mmvec"; +let Constraints = "$Qx4 = $Qx4in"; +} +def V6_veqsf : HInst< +(outs HvxQR:$Qd4), +(ins HvxVR:$Vu32, HvxVR:$Vv32), +"$Qd4 = vcmp.eq($Vu32.sf,$Vv32.sf)", +tc_56c4f9fe, TypeCVI_VA>, Enc_95441f, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-2} = 0b000011; +let Inst{13-13} = 0b0; +let Inst{31-21} = 0b00011111100; +let hasNewValue = 1; +let opNewValue = 0; +let isCVI = 1; +let isHVXALU = 1; +let isHVXALU2SRC = 1; +let DecoderNamespace = "EXT_mmvec"; +} +def V6_veqsf_and : HInst< +(outs HvxQR:$Qx4), +(ins HvxQR:$Qx4in, HvxVR:$Vu32, HvxVR:$Vv32), +"$Qx4 &= vcmp.eq($Vu32.sf,$Vv32.sf)", +tc_257f6f7c, TypeCVI_VA>, Enc_eaa9f8, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-2} = 0b000011; +let Inst{13-13} = 0b1; +let Inst{31-21} = 0b00011100100; +let isCVI = 1; +let isHVXALU = 1; +let isHVXALU2SRC = 1; +let DecoderNamespace = "EXT_mmvec"; +let Constraints = "$Qx4 = $Qx4in"; +} +def V6_veqsf_or : HInst< +(outs HvxQR:$Qx4), +(ins HvxQR:$Qx4in, HvxVR:$Vu32, HvxVR:$Vv32), +"$Qx4 |= vcmp.eq($Vu32.sf,$Vv32.sf)", +tc_257f6f7c, TypeCVI_VA>, Enc_eaa9f8, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-2} = 0b010011; +let Inst{13-13} = 0b1; +let Inst{31-21} = 0b00011100100; +let isAccumulator = 1; +let isCVI = 1; +let isHVXALU = 1; +let isHVXALU2SRC = 1; +let DecoderNamespace = "EXT_mmvec"; +let Constraints = "$Qx4 = $Qx4in"; +} +def V6_veqsf_xor : HInst< +(outs HvxQR:$Qx4), +(ins HvxQR:$Qx4in, HvxVR:$Vu32, HvxVR:$Vv32), +"$Qx4 ^= vcmp.eq($Vu32.sf,$Vv32.sf)", +tc_257f6f7c, TypeCVI_VA>, Enc_eaa9f8, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-2} = 0b100011; +let Inst{13-13} = 0b1; +let Inst{31-21} = 0b00011100100; +let isCVI = 1; +let isHVXALU = 1; +let isHVXALU2SRC = 1; +let DecoderNamespace = "EXT_mmvec"; +let Constraints = "$Qx4 = $Qx4in"; +} def V6_veqw : HInst< (outs HvxQR:$Qd4), (ins HvxVR:$Vu32, HvxVR:$Vv32), @@ -34538,6 +34825,58 @@ let Inst{31-24} = 0b00011110; let isCVI = 1; let DecoderNamespace = "EXT_mmvec"; } +def V6_vilog2_hf : HInst< +(outs HvxVR:$Vd32), +(ins HvxVR:$Vu32), +"$Vd32.w = vilog2($Vu32.hf)", +tc_2a698a03, TypeCVI_VS>, Enc_e7581c, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-5} = 0b011; +let Inst{13-13} = 0b1; +let Inst{31-16} = 0b0001111000001100; +let hasNewValue = 1; +let opNewValue = 0; +let isCVI = 1; +let DecoderNamespace = "EXT_mmvec"; +} +def V6_vilog2_qf16 : HInst< +(outs HvxVR:$Vd32), +(ins HvxVR:$Vu32), +"$Vd32.w = vilog2($Vu32.qf16)", +tc_2a698a03, TypeCVI_VS>, Enc_e7581c, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-5} = 0b001; +let Inst{13-13} = 0b1; +let Inst{31-16} = 0b0001111000001100; +let hasNewValue = 1; +let opNewValue = 0; +let isCVI = 1; +let DecoderNamespace = "EXT_mmvec"; +} +def V6_vilog2_qf32 : HInst< +(outs HvxVR:$Vd32), +(ins HvxVR:$Vu32), +"$Vd32.w = vilog2($Vu32.qf32)", +tc_2a698a03, TypeCVI_VS>, Enc_e7581c, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-5} = 0b000; +let Inst{13-13} = 0b1; +let Inst{31-16} = 0b0001111000001100; +let hasNewValue = 1; +let opNewValue = 0; +let isCVI = 1; +let DecoderNamespace = "EXT_mmvec"; +} +def V6_vilog2_sf : HInst< +(outs HvxVR:$Vd32), +(ins HvxVR:$Vu32), +"$Vd32.w = vilog2($Vu32.sf)", +tc_2a698a03, TypeCVI_VS>, Enc_e7581c, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-5} = 0b010; +let Inst{13-13} = 0b1; +let Inst{31-16} = 0b0001111000001100; +let hasNewValue = 1; +let opNewValue = 0; +let isCVI = 1; +let DecoderNamespace = "EXT_mmvec"; +} def V6_vinsertwr : HInst< (outs HvxVR:$Vx32), (ins HvxVR:$Vx32in, IntRegs:$Rt32), @@ -37170,6 +37509,58 @@ let isCVI = 1; let isHVXALU = 1; let DecoderNamespace = "EXT_mmvec"; } +def V6_vneg_qf16_hf : HInst< +(outs HvxVR:$Vd32), +(ins HvxVR:$Vu32), +"$Vd32.qf16 = vneg($Vu32.hf)", +tc_2a698a03, TypeCVI_VS>, Enc_e7581c, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-5} = 0b010; +let Inst{13-13} = 0b1; +let Inst{31-16} = 0b0001111000001110; +let hasNewValue = 1; +let opNewValue = 0; +let isCVI = 1; +let DecoderNamespace = "EXT_mmvec"; +} +def V6_vneg_qf16_qf16 : HInst< +(outs HvxVR:$Vd32), +(ins HvxVR:$Vu32), +"$Vd32.qf16 = vneg($Vu32.qf16)", +tc_2a698a03, TypeCVI_VS>, Enc_e7581c, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-5} = 0b011; +let Inst{13-13} = 0b1; +let Inst{31-16} = 0b0001111000001110; +let hasNewValue = 1; +let opNewValue = 0; +let isCVI = 1; +let DecoderNamespace = "EXT_mmvec"; +} +def V6_vneg_qf32_qf32 : HInst< +(outs HvxVR:$Vd32), +(ins HvxVR:$Vu32), +"$Vd32.qf32 = vneg($Vu32.qf32)", +tc_2a698a03, TypeCVI_VS>, Enc_e7581c, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-5} = 0b001; +let Inst{13-13} = 0b1; +let Inst{31-16} = 0b0001111000001110; +let hasNewValue = 1; +let opNewValue = 0; +let isCVI = 1; +let DecoderNamespace = "EXT_mmvec"; +} +def V6_vneg_qf32_sf : HInst< +(outs HvxVR:$Vd32), +(ins HvxVR:$Vu32), +"$Vd32.qf32 = vneg($Vu32.sf)", +tc_2a698a03, TypeCVI_VS>, Enc_e7581c, Requires<[UseHVXV81,UseHVXQFloat]> { +let Inst{7-5} = 0b000; +let Inst{13-13} = 0b1; +let Inst{31-16} = 0b0001111000001110; +let hasNewValue = 1; +let opNewValue = 0; +let isCVI = 1; +let DecoderNamespace = "EXT_mmvec"; +} def V6_vnormamth : HInst< (outs HvxVR:$Vd32), (ins HvxVR:$Vu32), diff --git a/llvm/lib/Target/Hexagon/HexagonDepMapAsm2Intrin.td b/llvm/lib/Target/Hexagon/HexagonDepMapAsm2Intrin.td index 23f4b3a..c11483b 100644 --- a/llvm/lib/Target/Hexagon/HexagonDepMapAsm2Intrin.td +++ b/llvm/lib/Target/Hexagon/HexagonDepMapAsm2Intrin.td @@ -3830,6 +3830,122 @@ def: Pat<(int_hexagon_V6_vsub_hf_f8_128B HvxVR:$src1, HvxVR:$src2), // V81 HVX Instructions. +def: Pat<(int_hexagon_V6_vabs_qf16_hf HvxVR:$src1), + (V6_vabs_qf16_hf HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vabs_qf16_hf_128B HvxVR:$src1), + (V6_vabs_qf16_hf HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vabs_qf16_qf16 HvxVR:$src1), + (V6_vabs_qf16_qf16 HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vabs_qf16_qf16_128B HvxVR:$src1), + (V6_vabs_qf16_qf16 HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vabs_qf32_qf32 HvxVR:$src1), + (V6_vabs_qf32_qf32 HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vabs_qf32_qf32_128B HvxVR:$src1), + (V6_vabs_qf32_qf32 HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vabs_qf32_sf HvxVR:$src1), + (V6_vabs_qf32_sf HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vabs_qf32_sf_128B HvxVR:$src1), + (V6_vabs_qf32_sf HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_valign4 HvxVR:$src1, HvxVR:$src2, IntRegsLow8:$src3), + (V6_valign4 HvxVR:$src1, HvxVR:$src2, IntRegsLow8:$src3)>, Requires<[UseHVXV81, UseHVX64B]>; +def: Pat<(int_hexagon_V6_valign4_128B HvxVR:$src1, HvxVR:$src2, IntRegsLow8:$src3), + (V6_valign4 HvxVR:$src1, HvxVR:$src2, IntRegsLow8:$src3)>, Requires<[UseHVXV81, UseHVX128B]>; +def: Pat<(int_hexagon_V6_vconv_bf_qf32 HvxWR:$src1), + (V6_vconv_bf_qf32 HvxWR:$src1)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vconv_bf_qf32_128B HvxWR:$src1), + (V6_vconv_bf_qf32 HvxWR:$src1)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vconv_f8_qf16 HvxVR:$src1), + (V6_vconv_f8_qf16 HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vconv_f8_qf16_128B HvxVR:$src1), + (V6_vconv_f8_qf16 HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vconv_h_hf_rnd HvxVR:$src1), + (V6_vconv_h_hf_rnd HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX64B]>; +def: Pat<(int_hexagon_V6_vconv_h_hf_rnd_128B HvxVR:$src1), + (V6_vconv_h_hf_rnd HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX128B]>; +def: Pat<(int_hexagon_V6_vconv_qf16_f8 HvxVR:$src1), + (V6_vconv_qf16_f8 HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vconv_qf16_f8_128B HvxVR:$src1), + (V6_vconv_qf16_f8 HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vconv_qf16_hf HvxVR:$src1), + (V6_vconv_qf16_hf HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vconv_qf16_hf_128B HvxVR:$src1), + (V6_vconv_qf16_hf HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vconv_qf16_qf16 HvxVR:$src1), + (V6_vconv_qf16_qf16 HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vconv_qf16_qf16_128B HvxVR:$src1), + (V6_vconv_qf16_qf16 HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vconv_qf32_qf32 HvxVR:$src1), + (V6_vconv_qf32_qf32 HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vconv_qf32_qf32_128B HvxVR:$src1), + (V6_vconv_qf32_qf32 HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vconv_qf32_sf HvxVR:$src1), + (V6_vconv_qf32_sf HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vconv_qf32_sf_128B HvxVR:$src1), + (V6_vconv_qf32_sf HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_veqhf HvxVR:$src1, HvxVR:$src2), + (V6_veqhf HvxVR:$src1, HvxVR:$src2)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_veqhf_128B HvxVR:$src1, HvxVR:$src2), + (V6_veqhf HvxVR:$src1, HvxVR:$src2)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_veqhf_and HvxQR:$src1, HvxVR:$src2, HvxVR:$src3), + (V6_veqhf_and HvxQR:$src1, HvxVR:$src2, HvxVR:$src3)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_veqhf_and_128B HvxQR:$src1, HvxVR:$src2, HvxVR:$src3), + (V6_veqhf_and HvxQR:$src1, HvxVR:$src2, HvxVR:$src3)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_veqhf_or HvxQR:$src1, HvxVR:$src2, HvxVR:$src3), + (V6_veqhf_or HvxQR:$src1, HvxVR:$src2, HvxVR:$src3)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_veqhf_or_128B HvxQR:$src1, HvxVR:$src2, HvxVR:$src3), + (V6_veqhf_or HvxQR:$src1, HvxVR:$src2, HvxVR:$src3)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_veqhf_xor HvxQR:$src1, HvxVR:$src2, HvxVR:$src3), + (V6_veqhf_xor HvxQR:$src1, HvxVR:$src2, HvxVR:$src3)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_veqhf_xor_128B HvxQR:$src1, HvxVR:$src2, HvxVR:$src3), + (V6_veqhf_xor HvxQR:$src1, HvxVR:$src2, HvxVR:$src3)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_veqsf HvxVR:$src1, HvxVR:$src2), + (V6_veqsf HvxVR:$src1, HvxVR:$src2)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_veqsf_128B HvxVR:$src1, HvxVR:$src2), + (V6_veqsf HvxVR:$src1, HvxVR:$src2)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_veqsf_and HvxQR:$src1, HvxVR:$src2, HvxVR:$src3), + (V6_veqsf_and HvxQR:$src1, HvxVR:$src2, HvxVR:$src3)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_veqsf_and_128B HvxQR:$src1, HvxVR:$src2, HvxVR:$src3), + (V6_veqsf_and HvxQR:$src1, HvxVR:$src2, HvxVR:$src3)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_veqsf_or HvxQR:$src1, HvxVR:$src2, HvxVR:$src3), + (V6_veqsf_or HvxQR:$src1, HvxVR:$src2, HvxVR:$src3)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_veqsf_or_128B HvxQR:$src1, HvxVR:$src2, HvxVR:$src3), + (V6_veqsf_or HvxQR:$src1, HvxVR:$src2, HvxVR:$src3)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_veqsf_xor HvxQR:$src1, HvxVR:$src2, HvxVR:$src3), + (V6_veqsf_xor HvxQR:$src1, HvxVR:$src2, HvxVR:$src3)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_veqsf_xor_128B HvxQR:$src1, HvxVR:$src2, HvxVR:$src3), + (V6_veqsf_xor HvxQR:$src1, HvxVR:$src2, HvxVR:$src3)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vilog2_hf HvxVR:$src1), + (V6_vilog2_hf HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vilog2_hf_128B HvxVR:$src1), + (V6_vilog2_hf HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vilog2_qf16 HvxVR:$src1), + (V6_vilog2_qf16 HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vilog2_qf16_128B HvxVR:$src1), + (V6_vilog2_qf16 HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vilog2_qf32 HvxVR:$src1), + (V6_vilog2_qf32 HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vilog2_qf32_128B HvxVR:$src1), + (V6_vilog2_qf32 HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vilog2_sf HvxVR:$src1), + (V6_vilog2_sf HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vilog2_sf_128B HvxVR:$src1), + (V6_vilog2_sf HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vneg_qf16_hf HvxVR:$src1), + (V6_vneg_qf16_hf HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vneg_qf16_hf_128B HvxVR:$src1), + (V6_vneg_qf16_hf HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vneg_qf16_qf16 HvxVR:$src1), + (V6_vneg_qf16_qf16 HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vneg_qf16_qf16_128B HvxVR:$src1), + (V6_vneg_qf16_qf16 HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vneg_qf32_qf32 HvxVR:$src1), + (V6_vneg_qf32_qf32 HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vneg_qf32_qf32_128B HvxVR:$src1), + (V6_vneg_qf32_qf32 HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vneg_qf32_sf HvxVR:$src1), + (V6_vneg_qf32_sf HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; +def: Pat<(int_hexagon_V6_vneg_qf32_sf_128B HvxVR:$src1), + (V6_vneg_qf32_sf HvxVR:$src1)>, Requires<[UseHVXV81, UseHVX128B, UseHVXQFloat]>; def: Pat<(int_hexagon_V6_vsub_hf_mix HvxVR:$src1, HvxVR:$src2), (V6_vsub_hf_mix HvxVR:$src1, HvxVR:$src2)>, Requires<[UseHVXV81, UseHVX64B, UseHVXQFloat]>; def: Pat<(int_hexagon_V6_vsub_hf_mix_128B HvxVR:$src1, HvxVR:$src2), diff --git a/llvm/lib/Target/Hexagon/HexagonISelDAGToDAG.cpp b/llvm/lib/Target/Hexagon/HexagonISelDAGToDAG.cpp index 7ee280d..eadf020 100644 --- a/llvm/lib/Target/Hexagon/HexagonISelDAGToDAG.cpp +++ b/llvm/lib/Target/Hexagon/HexagonISelDAGToDAG.cpp @@ -1815,7 +1815,7 @@ struct WeightedLeaf { int Weight; int InsertionOrder; - WeightedLeaf() {} + WeightedLeaf() = default; WeightedLeaf(SDValue Value, int Weight, int InsertionOrder) : Value(Value), Weight(Weight), InsertionOrder(InsertionOrder) { diff --git a/llvm/lib/Target/LoongArch/LoongArchISelLowering.cpp b/llvm/lib/Target/LoongArch/LoongArchISelLowering.cpp index 904aabed..fe700e1 100644 --- a/llvm/lib/Target/LoongArch/LoongArchISelLowering.cpp +++ b/llvm/lib/Target/LoongArch/LoongArchISelLowering.cpp @@ -375,6 +375,8 @@ LoongArchTargetLowering::LoongArchTargetLowering(const TargetMachine &TM, setOperationAction(ISD::FFLOOR, VT, Legal); setOperationAction(ISD::FTRUNC, VT, Legal); setOperationAction(ISD::FROUNDEVEN, VT, Legal); + setOperationAction(ISD::FMINNUM, VT, Legal); + setOperationAction(ISD::FMAXNUM, VT, Legal); } setOperationAction(ISD::CTPOP, GRLenVT, Legal); setOperationAction(ISD::FCEIL, {MVT::f32, MVT::f64}, Legal); @@ -461,6 +463,8 @@ LoongArchTargetLowering::LoongArchTargetLowering(const TargetMachine &TM, setOperationAction(ISD::FFLOOR, VT, Legal); setOperationAction(ISD::FTRUNC, VT, Legal); setOperationAction(ISD::FROUNDEVEN, VT, Legal); + setOperationAction(ISD::FMINNUM, VT, Legal); + setOperationAction(ISD::FMAXNUM, VT, Legal); } } diff --git a/llvm/lib/Target/LoongArch/LoongArchLASXInstrInfo.td b/llvm/lib/Target/LoongArch/LoongArchLASXInstrInfo.td index 610ba05..b502b056 100644 --- a/llvm/lib/Target/LoongArch/LoongArchLASXInstrInfo.td +++ b/llvm/lib/Target/LoongArch/LoongArchLASXInstrInfo.td @@ -1558,6 +1558,10 @@ defm : PatXrXrF<fmul, "XVFMUL">; // XVFDIV_{S/D} defm : PatXrXrF<fdiv, "XVFDIV">; +// XVFMAX_{S/D}, XVFMIN_{S/D} +defm : PatXrXrF<fmaxnum, "XVFMAX">; +defm : PatXrXrF<fminnum, "XVFMIN">; + // XVFMADD_{S/D} def : Pat<(fma v8f32:$xj, v8f32:$xk, v8f32:$xa), (XVFMADD_S v8f32:$xj, v8f32:$xk, v8f32:$xa)>; diff --git a/llvm/lib/Target/LoongArch/LoongArchLSXInstrInfo.td b/llvm/lib/Target/LoongArch/LoongArchLSXInstrInfo.td index 6470842..6b74a4b 100644 --- a/llvm/lib/Target/LoongArch/LoongArchLSXInstrInfo.td +++ b/llvm/lib/Target/LoongArch/LoongArchLSXInstrInfo.td @@ -1760,6 +1760,10 @@ defm : PatVrVrF<fmul, "VFMUL">; // VFDIV_{S/D} defm : PatVrVrF<fdiv, "VFDIV">; +// VFMAX_{S/D}, VFMIN_{S/D} +defm : PatVrVrF<fmaxnum, "VFMAX">; +defm : PatVrVrF<fminnum, "VFMIN">; + // VFMADD_{S/D} def : Pat<(fma v4f32:$vj, v4f32:$vk, v4f32:$va), (VFMADD_S v4f32:$vj, v4f32:$vk, v4f32:$va)>; diff --git a/llvm/lib/Target/LoongArch/MCTargetDesc/LoongArchELFObjectWriter.cpp b/llvm/lib/Target/LoongArch/MCTargetDesc/LoongArchELFObjectWriter.cpp index 7d54565..6d69af5 100644 --- a/llvm/lib/Target/LoongArch/MCTargetDesc/LoongArchELFObjectWriter.cpp +++ b/llvm/lib/Target/LoongArch/MCTargetDesc/LoongArchELFObjectWriter.cpp @@ -39,7 +39,7 @@ LoongArchELFObjectWriter::LoongArchELFObjectWriter(uint8_t OSABI, bool Is64Bit) : MCELFObjectTargetWriter(Is64Bit, OSABI, ELF::EM_LOONGARCH, /*HasRelocationAddend=*/true) {} -LoongArchELFObjectWriter::~LoongArchELFObjectWriter() {} +LoongArchELFObjectWriter::~LoongArchELFObjectWriter() = default; unsigned LoongArchELFObjectWriter::getRelocType(const MCFixup &Fixup, const MCValue &Target, diff --git a/llvm/lib/Target/LoongArch/MCTargetDesc/LoongArchMCCodeEmitter.cpp b/llvm/lib/Target/LoongArch/MCTargetDesc/LoongArchMCCodeEmitter.cpp index f0e2bc4..08fa51d 100644 --- a/llvm/lib/Target/LoongArch/MCTargetDesc/LoongArchMCCodeEmitter.cpp +++ b/llvm/lib/Target/LoongArch/MCTargetDesc/LoongArchMCCodeEmitter.cpp @@ -38,7 +38,7 @@ public: LoongArchMCCodeEmitter(MCContext &ctx, MCInstrInfo const &MCII) : Ctx(ctx), MCII(MCII) {} - ~LoongArchMCCodeEmitter() override {} + ~LoongArchMCCodeEmitter() override = default; void encodeInstruction(const MCInst &MI, SmallVectorImpl<char> &CB, SmallVectorImpl<MCFixup> &Fixups, diff --git a/llvm/lib/Target/M68k/AsmParser/M68kAsmParser.cpp b/llvm/lib/Target/M68k/AsmParser/M68kAsmParser.cpp index e37f3a66..fb5cd5c2 100644 --- a/llvm/lib/Target/M68k/AsmParser/M68kAsmParser.cpp +++ b/llvm/lib/Target/M68k/AsmParser/M68kAsmParser.cpp @@ -690,9 +690,9 @@ bool M68kAsmParser::parseRegisterName(MCRegister &RegNo, SMLoc Loc, } else { // Floating point control register. RegNo = StringSwitch<unsigned>(RegisterNameLower) - .Cases("fpc", "fpcr", M68k::FPC) - .Cases("fps", "fpsr", M68k::FPS) - .Cases("fpi", "fpiar", M68k::FPIAR) + .Cases({"fpc", "fpcr"}, M68k::FPC) + .Cases({"fps", "fpsr"}, M68k::FPS) + .Cases({"fpi", "fpiar"}, M68k::FPIAR) .Default(M68k::NoRegister); assert(RegNo != M68k::NoRegister && "Unrecognized FP control register name"); diff --git a/llvm/lib/Target/Mips/AsmParser/MipsAsmParser.cpp b/llvm/lib/Target/Mips/AsmParser/MipsAsmParser.cpp index 97379d7..f588e56 100644 --- a/llvm/lib/Target/Mips/AsmParser/MipsAsmParser.cpp +++ b/llvm/lib/Target/Mips/AsmParser/MipsAsmParser.cpp @@ -6176,7 +6176,7 @@ int MipsAsmParser::matchCPURegisterName(StringRef Name) { CC = StringSwitch<unsigned>(Name) .Case("zero", 0) - .Cases("at", "AT", 1) + .Cases({"at", "AT"}, 1) .Case("a0", 4) .Case("a1", 5) .Case("a2", 6) diff --git a/llvm/lib/Target/NVPTX/NVPTXAliasAnalysis.h b/llvm/lib/Target/NVPTX/NVPTXAliasAnalysis.h index caef8fe7..b832b82 100644 --- a/llvm/lib/Target/NVPTX/NVPTXAliasAnalysis.h +++ b/llvm/lib/Target/NVPTX/NVPTXAliasAnalysis.h @@ -20,7 +20,7 @@ class MemoryLocation; class NVPTXAAResult : public AAResultBase { public: - NVPTXAAResult() {} + NVPTXAAResult() = default; NVPTXAAResult(NVPTXAAResult &&Arg) : AAResultBase(std::move(Arg)) {} /// Handle invalidation events from the new pass manager. diff --git a/llvm/lib/Target/NVPTX/NVPTXIntrinsics.td b/llvm/lib/Target/NVPTX/NVPTXIntrinsics.td index e8758aa..50827bd 100644 --- a/llvm/lib/Target/NVPTX/NVPTXIntrinsics.td +++ b/llvm/lib/Target/NVPTX/NVPTXIntrinsics.td @@ -1562,12 +1562,17 @@ def : Pat<(int_nvvm_saturate_d f64:$a), (CVT_f64_f64 $a, CvtSAT)>; // Exp2 Log2 // -def : Pat<(int_nvvm_ex2_approx_ftz_f f32:$a), (EX2_APPROX_f32 $a, FTZ)>; -def : Pat<(int_nvvm_ex2_approx_f f32:$a), (EX2_APPROX_f32 $a, NoFTZ)>; +def : Pat<(f32 (int_nvvm_ex2_approx_ftz f32:$a)), (EX2_APPROX_f32 $a, FTZ)>; +def : Pat<(f32 (int_nvvm_ex2_approx f32:$a)), (EX2_APPROX_f32 $a, NoFTZ)>; let Predicates = [hasPTX<70>, hasSM<75>] in { - def : Pat<(int_nvvm_ex2_approx_f16 f16:$a), (EX2_APPROX_f16 $a)>; - def : Pat<(int_nvvm_ex2_approx_f16x2 v2f16:$a), (EX2_APPROX_f16x2 $a)>; + def : Pat<(f16 (int_nvvm_ex2_approx f16:$a)), (EX2_APPROX_f16 $a)>; + def : Pat<(v2f16 (int_nvvm_ex2_approx v2f16:$a)), (EX2_APPROX_f16x2 $a)>; +} + +let Predicates = [hasPTX<78>, hasSM<90>] in { + def : Pat<(bf16 (int_nvvm_ex2_approx_ftz bf16:$a)), (EX2_APPROX_bf16 $a)>; + def : Pat<(v2bf16 (int_nvvm_ex2_approx_ftz v2bf16:$a)), (EX2_APPROX_bf16x2 $a)>; } def LG2_APPROX_f32 : diff --git a/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.cpp b/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.cpp index 729c077..64593e6 100644 --- a/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.cpp +++ b/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.cpp @@ -318,7 +318,7 @@ static Instruction *convertNvvmIntrinsicToLlvm(InstCombiner &IC, // answer. These include: // // - nvvm_cos_approx_{f,ftz_f} - // - nvvm_ex2_approx_{d,f,ftz_f} + // - nvvm_ex2_approx(_ftz) // - nvvm_lg2_approx_{d,f,ftz_f} // - nvvm_sin_approx_{f,ftz_f} // - nvvm_sqrt_approx_{f,ftz_f} diff --git a/llvm/lib/Target/PowerPC/PPCAsmPrinter.cpp b/llvm/lib/Target/PowerPC/PPCAsmPrinter.cpp index bcb3f50..780e124 100644 --- a/llvm/lib/Target/PowerPC/PPCAsmPrinter.cpp +++ b/llvm/lib/Target/PowerPC/PPCAsmPrinter.cpp @@ -2702,7 +2702,7 @@ static bool isSpecialLLVMGlobalArrayToSkip(const GlobalVariable *GV) { static bool isSpecialLLVMGlobalArrayForStaticInit(const GlobalVariable *GV) { return StringSwitch<bool>(GV->getName()) - .Cases("llvm.global_ctors", "llvm.global_dtors", true) + .Cases({"llvm.global_ctors", "llvm.global_dtors"}, true) .Default(false); } diff --git a/llvm/lib/Target/PowerPC/PPCTargetTransformInfo.cpp b/llvm/lib/Target/PowerPC/PPCTargetTransformInfo.cpp index 2fba090..b04e887 100644 --- a/llvm/lib/Target/PowerPC/PPCTargetTransformInfo.cpp +++ b/llvm/lib/Target/PowerPC/PPCTargetTransformInfo.cpp @@ -912,7 +912,7 @@ bool PPCTTIImpl::areInlineCompatible(const Function *Caller, bool PPCTTIImpl::areTypesABICompatible(const Function *Caller, const Function *Callee, - const ArrayRef<Type *> &Types) const { + ArrayRef<Type *> Types) const { // We need to ensure that argument promotion does not // attempt to promote pointers to MMA types (__vector_pair diff --git a/llvm/lib/Target/PowerPC/PPCTargetTransformInfo.h b/llvm/lib/Target/PowerPC/PPCTargetTransformInfo.h index 475472a..8d7f255 100644 --- a/llvm/lib/Target/PowerPC/PPCTargetTransformInfo.h +++ b/llvm/lib/Target/PowerPC/PPCTargetTransformInfo.h @@ -147,7 +147,7 @@ public: bool areInlineCompatible(const Function *Caller, const Function *Callee) const override; bool areTypesABICompatible(const Function *Caller, const Function *Callee, - const ArrayRef<Type *> &Types) const override; + ArrayRef<Type *> Types) const override; bool supportsTailCallFor(const CallBase *CB) const override; private: diff --git a/llvm/lib/Target/RISCV/RISCVISelLowering.cpp b/llvm/lib/Target/RISCV/RISCVISelLowering.cpp index c6a8b84..e0cf739 100644 --- a/llvm/lib/Target/RISCV/RISCVISelLowering.cpp +++ b/llvm/lib/Target/RISCV/RISCVISelLowering.cpp @@ -23946,7 +23946,7 @@ RISCVTargetLowering::getRegForInlineAsmConstraint(const TargetRegisterInfo *TRI, .Case("{t0}", RISCV::X5) .Case("{t1}", RISCV::X6) .Case("{t2}", RISCV::X7) - .Cases("{s0}", "{fp}", RISCV::X8) + .Cases({"{s0}", "{fp}"}, RISCV::X8) .Case("{s1}", RISCV::X9) .Case("{a0}", RISCV::X10) .Case("{a1}", RISCV::X11) @@ -23983,38 +23983,38 @@ RISCVTargetLowering::getRegForInlineAsmConstraint(const TargetRegisterInfo *TRI, // use the ABI names in register constraint lists. if (Subtarget.hasStdExtF()) { unsigned FReg = StringSwitch<unsigned>(Constraint.lower()) - .Cases("{f0}", "{ft0}", RISCV::F0_F) - .Cases("{f1}", "{ft1}", RISCV::F1_F) - .Cases("{f2}", "{ft2}", RISCV::F2_F) - .Cases("{f3}", "{ft3}", RISCV::F3_F) - .Cases("{f4}", "{ft4}", RISCV::F4_F) - .Cases("{f5}", "{ft5}", RISCV::F5_F) - .Cases("{f6}", "{ft6}", RISCV::F6_F) - .Cases("{f7}", "{ft7}", RISCV::F7_F) - .Cases("{f8}", "{fs0}", RISCV::F8_F) - .Cases("{f9}", "{fs1}", RISCV::F9_F) - .Cases("{f10}", "{fa0}", RISCV::F10_F) - .Cases("{f11}", "{fa1}", RISCV::F11_F) - .Cases("{f12}", "{fa2}", RISCV::F12_F) - .Cases("{f13}", "{fa3}", RISCV::F13_F) - .Cases("{f14}", "{fa4}", RISCV::F14_F) - .Cases("{f15}", "{fa5}", RISCV::F15_F) - .Cases("{f16}", "{fa6}", RISCV::F16_F) - .Cases("{f17}", "{fa7}", RISCV::F17_F) - .Cases("{f18}", "{fs2}", RISCV::F18_F) - .Cases("{f19}", "{fs3}", RISCV::F19_F) - .Cases("{f20}", "{fs4}", RISCV::F20_F) - .Cases("{f21}", "{fs5}", RISCV::F21_F) - .Cases("{f22}", "{fs6}", RISCV::F22_F) - .Cases("{f23}", "{fs7}", RISCV::F23_F) - .Cases("{f24}", "{fs8}", RISCV::F24_F) - .Cases("{f25}", "{fs9}", RISCV::F25_F) - .Cases("{f26}", "{fs10}", RISCV::F26_F) - .Cases("{f27}", "{fs11}", RISCV::F27_F) - .Cases("{f28}", "{ft8}", RISCV::F28_F) - .Cases("{f29}", "{ft9}", RISCV::F29_F) - .Cases("{f30}", "{ft10}", RISCV::F30_F) - .Cases("{f31}", "{ft11}", RISCV::F31_F) + .Cases({"{f0}", "{ft0}"}, RISCV::F0_F) + .Cases({"{f1}", "{ft1}"}, RISCV::F1_F) + .Cases({"{f2}", "{ft2}"}, RISCV::F2_F) + .Cases({"{f3}", "{ft3}"}, RISCV::F3_F) + .Cases({"{f4}", "{ft4}"}, RISCV::F4_F) + .Cases({"{f5}", "{ft5}"}, RISCV::F5_F) + .Cases({"{f6}", "{ft6}"}, RISCV::F6_F) + .Cases({"{f7}", "{ft7}"}, RISCV::F7_F) + .Cases({"{f8}", "{fs0}"}, RISCV::F8_F) + .Cases({"{f9}", "{fs1}"}, RISCV::F9_F) + .Cases({"{f10}", "{fa0}"}, RISCV::F10_F) + .Cases({"{f11}", "{fa1}"}, RISCV::F11_F) + .Cases({"{f12}", "{fa2}"}, RISCV::F12_F) + .Cases({"{f13}", "{fa3}"}, RISCV::F13_F) + .Cases({"{f14}", "{fa4}"}, RISCV::F14_F) + .Cases({"{f15}", "{fa5}"}, RISCV::F15_F) + .Cases({"{f16}", "{fa6}"}, RISCV::F16_F) + .Cases({"{f17}", "{fa7}"}, RISCV::F17_F) + .Cases({"{f18}", "{fs2}"}, RISCV::F18_F) + .Cases({"{f19}", "{fs3}"}, RISCV::F19_F) + .Cases({"{f20}", "{fs4}"}, RISCV::F20_F) + .Cases({"{f21}", "{fs5}"}, RISCV::F21_F) + .Cases({"{f22}", "{fs6}"}, RISCV::F22_F) + .Cases({"{f23}", "{fs7}"}, RISCV::F23_F) + .Cases({"{f24}", "{fs8}"}, RISCV::F24_F) + .Cases({"{f25}", "{fs9}"}, RISCV::F25_F) + .Cases({"{f26}", "{fs10}"}, RISCV::F26_F) + .Cases({"{f27}", "{fs11}"}, RISCV::F27_F) + .Cases({"{f28}", "{ft8}"}, RISCV::F28_F) + .Cases({"{f29}", "{ft9}"}, RISCV::F29_F) + .Cases({"{f30}", "{ft10}"}, RISCV::F30_F) + .Cases({"{f31}", "{ft11}"}, RISCV::F31_F) .Default(RISCV::NoRegister); if (FReg != RISCV::NoRegister) { assert(RISCV::F0_F <= FReg && FReg <= RISCV::F31_F && "Unknown fp-reg"); diff --git a/llvm/lib/Target/RISCV/RISCVInsertWriteVXRM.cpp b/llvm/lib/Target/RISCV/RISCVInsertWriteVXRM.cpp index a1c8e23..c58a5c0 100644 --- a/llvm/lib/Target/RISCV/RISCVInsertWriteVXRM.cpp +++ b/llvm/lib/Target/RISCV/RISCVInsertWriteVXRM.cpp @@ -48,7 +48,7 @@ class VXRMInfo { } State = Uninitialized; public: - VXRMInfo() {} + VXRMInfo() = default; static VXRMInfo getUnknown() { VXRMInfo Info; diff --git a/llvm/lib/Target/SPIRV/MCTargetDesc/SPIRVTargetStreamer.cpp b/llvm/lib/Target/SPIRV/MCTargetDesc/SPIRVTargetStreamer.cpp index 0a318e0..ed6d355 100644 --- a/llvm/lib/Target/SPIRV/MCTargetDesc/SPIRVTargetStreamer.cpp +++ b/llvm/lib/Target/SPIRV/MCTargetDesc/SPIRVTargetStreamer.cpp @@ -15,4 +15,4 @@ using namespace llvm; SPIRVTargetStreamer::SPIRVTargetStreamer(MCStreamer &S) : MCTargetStreamer(S) {} -SPIRVTargetStreamer::~SPIRVTargetStreamer() {} +SPIRVTargetStreamer::~SPIRVTargetStreamer() = default; diff --git a/llvm/lib/Target/SPIRV/SPIRVModuleAnalysis.h b/llvm/lib/Target/SPIRV/SPIRVModuleAnalysis.h index 2d19f6de..44b6c66 100644 --- a/llvm/lib/Target/SPIRV/SPIRVModuleAnalysis.h +++ b/llvm/lib/Target/SPIRV/SPIRVModuleAnalysis.h @@ -81,7 +81,7 @@ private: void initAvailableCapabilitiesForVulkan(const SPIRVSubtarget &ST); public: - RequirementHandler() {} + RequirementHandler() = default; void clear() { MinimalCaps.clear(); AllCaps.clear(); diff --git a/llvm/lib/Target/SPIRV/SPIRVTargetMachine.cpp b/llvm/lib/Target/SPIRV/SPIRVTargetMachine.cpp index 7dd0b95..5ba0356 100644 --- a/llvm/lib/Target/SPIRV/SPIRVTargetMachine.cpp +++ b/llvm/lib/Target/SPIRV/SPIRVTargetMachine.cpp @@ -69,7 +69,7 @@ static Reloc::Model getEffectiveRelocModel(std::optional<Reloc::Model> RM) { } // Pin SPIRVTargetObjectFile's vtables to this file. -SPIRVTargetObjectFile::~SPIRVTargetObjectFile() {} +SPIRVTargetObjectFile::~SPIRVTargetObjectFile() = default; SPIRVTargetMachine::SPIRVTargetMachine(const Target &T, const Triple &TT, StringRef CPU, StringRef FS, diff --git a/llvm/lib/Target/SystemZ/SystemZTargetObjectFile.h b/llvm/lib/Target/SystemZ/SystemZTargetObjectFile.h index 9d0adbb..87ec256 100644 --- a/llvm/lib/Target/SystemZ/SystemZTargetObjectFile.h +++ b/llvm/lib/Target/SystemZ/SystemZTargetObjectFile.h @@ -16,7 +16,7 @@ namespace llvm { /// This implementation is used for SystemZ ELF targets. class SystemZELFTargetObjectFile : public TargetLoweringObjectFileELF { public: - SystemZELFTargetObjectFile() {} + SystemZELFTargetObjectFile() = default; /// Describe a TLS variable address within debug info. const MCExpr *getDebugThreadLocalSymbol(const MCSymbol *Sym) const override; diff --git a/llvm/lib/Target/WebAssembly/WebAssemblyExceptionInfo.h b/llvm/lib/Target/WebAssembly/WebAssemblyExceptionInfo.h index 7845cdf..1bfc61f 100644 --- a/llvm/lib/Target/WebAssembly/WebAssemblyExceptionInfo.h +++ b/llvm/lib/Target/WebAssembly/WebAssemblyExceptionInfo.h @@ -76,7 +76,7 @@ public: BlockSet.insert(MBB); } ArrayRef<MachineBasicBlock *> getBlocks() const { return Blocks; } - using block_iterator = typename ArrayRef<MachineBasicBlock *>::const_iterator; + using block_iterator = ArrayRef<MachineBasicBlock *>::const_iterator; block_iterator block_begin() const { return getBlocks().begin(); } block_iterator block_end() const { return getBlocks().end(); } inline iterator_range<block_iterator> blocks() const { @@ -96,7 +96,7 @@ public: void addSubException(std::unique_ptr<WebAssemblyException> E) { SubExceptions.push_back(std::move(E)); } - using iterator = typename decltype(SubExceptions)::const_iterator; + using iterator = decltype(SubExceptions)::const_iterator; iterator begin() const { return SubExceptions.begin(); } iterator end() const { return SubExceptions.end(); } diff --git a/llvm/lib/Target/WebAssembly/WebAssemblyMachineFunctionInfo.h b/llvm/lib/Target/WebAssembly/WebAssemblyMachineFunctionInfo.h index ff4d6469..ee575e3 100644 --- a/llvm/lib/Target/WebAssembly/WebAssemblyMachineFunctionInfo.h +++ b/llvm/lib/Target/WebAssembly/WebAssemblyMachineFunctionInfo.h @@ -207,8 +207,7 @@ template <> struct MappingTraits<WebAssemblyFunctionInfo> { template <> struct CustomMappingTraits<BBNumberMap> { static void inputOne(IO &YamlIO, StringRef Key, BBNumberMap &SrcToUnwindDest) { - YamlIO.mapRequired(Key.str().c_str(), - SrcToUnwindDest[std::atoi(Key.str().c_str())]); + YamlIO.mapRequired(Key, SrcToUnwindDest[std::atoi(Key.str().c_str())]); } static void output(IO &YamlIO, BBNumberMap &SrcToUnwindDest) { diff --git a/llvm/lib/Target/WebAssembly/WebAssemblySortRegion.h b/llvm/lib/Target/WebAssembly/WebAssemblySortRegion.h index e92bf17..96b8a4e 100644 --- a/llvm/lib/Target/WebAssembly/WebAssemblySortRegion.h +++ b/llvm/lib/Target/WebAssembly/WebAssemblySortRegion.h @@ -35,7 +35,7 @@ public: virtual MachineBasicBlock *getHeader() const = 0; virtual bool contains(const MachineBasicBlock *MBB) const = 0; virtual unsigned getNumBlocks() const = 0; - using block_iterator = typename ArrayRef<MachineBasicBlock *>::const_iterator; + using block_iterator = ArrayRef<MachineBasicBlock *>::const_iterator; virtual iterator_range<block_iterator> blocks() const = 0; virtual bool isLoop() const = 0; }; diff --git a/llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp b/llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp index b7ea672..bac3692 100644 --- a/llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp +++ b/llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp @@ -2470,10 +2470,10 @@ bool X86AsmParser::ParseIntelOffsetOperator(const MCExpr *&Val, StringRef &ID, // Report back its kind, or IOK_INVALID if does not evaluated as a known one unsigned X86AsmParser::IdentifyIntelInlineAsmOperator(StringRef Name) { return StringSwitch<unsigned>(Name) - .Cases("TYPE","type",IOK_TYPE) - .Cases("SIZE","size",IOK_SIZE) - .Cases("LENGTH","length",IOK_LENGTH) - .Default(IOK_INVALID); + .Cases({"TYPE", "type"}, IOK_TYPE) + .Cases({"SIZE", "size"}, IOK_SIZE) + .Cases({"LENGTH", "length"}, IOK_LENGTH) + .Default(IOK_INVALID); } /// Parse the 'LENGTH', 'TYPE' and 'SIZE' operators. The LENGTH operator @@ -2516,8 +2516,8 @@ unsigned X86AsmParser::ParseIntelInlineAsmOperator(unsigned OpKind) { unsigned X86AsmParser::IdentifyMasmOperator(StringRef Name) { return StringSwitch<unsigned>(Name.lower()) .Case("type", MOK_TYPE) - .Cases("size", "sizeof", MOK_SIZEOF) - .Cases("length", "lengthof", MOK_LENGTHOF) + .Cases({"size", "sizeof"}, MOK_SIZEOF) + .Cases({"length", "lengthof"}, MOK_LENGTHOF) .Default(MOK_INVALID); } @@ -2581,21 +2581,21 @@ bool X86AsmParser::ParseMasmOperator(unsigned OpKind, int64_t &Val) { bool X86AsmParser::ParseIntelMemoryOperandSize(unsigned &Size, StringRef *SizeStr) { Size = StringSwitch<unsigned>(getTok().getString()) - .Cases("BYTE", "byte", 8) - .Cases("WORD", "word", 16) - .Cases("DWORD", "dword", 32) - .Cases("FLOAT", "float", 32) - .Cases("LONG", "long", 32) - .Cases("FWORD", "fword", 48) - .Cases("DOUBLE", "double", 64) - .Cases("QWORD", "qword", 64) - .Cases("MMWORD","mmword", 64) - .Cases("XWORD", "xword", 80) - .Cases("TBYTE", "tbyte", 80) - .Cases("XMMWORD", "xmmword", 128) - .Cases("YMMWORD", "ymmword", 256) - .Cases("ZMMWORD", "zmmword", 512) - .Default(0); + .Cases({"BYTE", "byte"}, 8) + .Cases({"WORD", "word"}, 16) + .Cases({"DWORD", "dword"}, 32) + .Cases({"FLOAT", "float"}, 32) + .Cases({"LONG", "long"}, 32) + .Cases({"FWORD", "fword"}, 48) + .Cases({"DOUBLE", "double"}, 64) + .Cases({"QWORD", "qword"}, 64) + .Cases({"MMWORD", "mmword"}, 64) + .Cases({"XWORD", "xword"}, 80) + .Cases({"TBYTE", "tbyte"}, 80) + .Cases({"XMMWORD", "xmmword"}, 128) + .Cases({"YMMWORD", "ymmword"}, 256) + .Cases({"ZMMWORD", "zmmword"}, 512) + .Default(0); if (Size) { if (SizeStr) *SizeStr = getTok().getString(); @@ -2886,22 +2886,22 @@ bool X86AsmParser::parseATTOperand(OperandVector &Operands) { // otherwise the EFLAGS Condition Code enumerator. X86::CondCode X86AsmParser::ParseConditionCode(StringRef CC) { return StringSwitch<X86::CondCode>(CC) - .Case("o", X86::COND_O) // Overflow - .Case("no", X86::COND_NO) // No Overflow - .Cases("b", "nae", X86::COND_B) // Below/Neither Above nor Equal - .Cases("ae", "nb", X86::COND_AE) // Above or Equal/Not Below - .Cases("e", "z", X86::COND_E) // Equal/Zero - .Cases("ne", "nz", X86::COND_NE) // Not Equal/Not Zero - .Cases("be", "na", X86::COND_BE) // Below or Equal/Not Above - .Cases("a", "nbe", X86::COND_A) // Above/Neither Below nor Equal - .Case("s", X86::COND_S) // Sign - .Case("ns", X86::COND_NS) // No Sign - .Cases("p", "pe", X86::COND_P) // Parity/Parity Even - .Cases("np", "po", X86::COND_NP) // No Parity/Parity Odd - .Cases("l", "nge", X86::COND_L) // Less/Neither Greater nor Equal - .Cases("ge", "nl", X86::COND_GE) // Greater or Equal/Not Less - .Cases("le", "ng", X86::COND_LE) // Less or Equal/Not Greater - .Cases("g", "nle", X86::COND_G) // Greater/Neither Less nor Equal + .Case("o", X86::COND_O) // Overflow + .Case("no", X86::COND_NO) // No Overflow + .Cases({"b", "nae"}, X86::COND_B) // Below/Neither Above nor Equal + .Cases({"ae", "nb"}, X86::COND_AE) // Above or Equal/Not Below + .Cases({"e", "z"}, X86::COND_E) // Equal/Zero + .Cases({"ne", "nz"}, X86::COND_NE) // Not Equal/Not Zero + .Cases({"be", "na"}, X86::COND_BE) // Below or Equal/Not Above + .Cases({"a", "nbe"}, X86::COND_A) // Above/Neither Below nor Equal + .Case("s", X86::COND_S) // Sign + .Case("ns", X86::COND_NS) // No Sign + .Cases({"p", "pe"}, X86::COND_P) // Parity/Parity Even + .Cases({"np", "po"}, X86::COND_NP) // No Parity/Parity Odd + .Cases({"l", "nge"}, X86::COND_L) // Less/Neither Greater nor Equal + .Cases({"ge", "nl"}, X86::COND_GE) // Greater or Equal/Not Less + .Cases({"le", "ng"}, X86::COND_LE) // Less or Equal/Not Greater + .Cases({"g", "nle"}, X86::COND_G) // Greater/Neither Less nor Equal .Default(X86::COND_INVALID); } diff --git a/llvm/lib/Target/X86/X86ISelLowering.cpp b/llvm/lib/Target/X86/X86ISelLowering.cpp index c5fb553..133406b 100644 --- a/llvm/lib/Target/X86/X86ISelLowering.cpp +++ b/llvm/lib/Target/X86/X86ISelLowering.cpp @@ -22861,6 +22861,13 @@ static SDValue combineVectorSizedSetCCEquality(EVT VT, SDValue X, SDValue Y, if (!OpVT.isScalarInteger() || OpSize < 128) return SDValue(); + // Don't do this if we're not supposed to use the FPU. + bool NoImplicitFloatOps = + DAG.getMachineFunction().getFunction().hasFnAttribute( + Attribute::NoImplicitFloat); + if (Subtarget.useSoftFloat() || NoImplicitFloatOps) + return SDValue(); + // Ignore a comparison with zero because that gets special treatment in // EmitTest(). But make an exception for the special case of a pair of // logically-combined vector-sized operands compared to zero. This pattern may @@ -22883,13 +22890,9 @@ static SDValue combineVectorSizedSetCCEquality(EVT VT, SDValue X, SDValue Y, // Use XOR (plus OR) and PTEST after SSE4.1 for 128/256-bit operands. // Use PCMPNEQ (plus OR) and KORTEST for 512-bit operands. // Otherwise use PCMPEQ (plus AND) and mask testing. - bool NoImplicitFloatOps = - DAG.getMachineFunction().getFunction().hasFnAttribute( - Attribute::NoImplicitFloat); - if (!Subtarget.useSoftFloat() && !NoImplicitFloatOps && - ((OpSize == 128 && Subtarget.hasSSE2()) || - (OpSize == 256 && Subtarget.hasAVX()) || - (OpSize == 512 && Subtarget.useAVX512Regs()))) { + if ((OpSize == 128 && Subtarget.hasSSE2()) || + (OpSize == 256 && Subtarget.hasAVX()) || + (OpSize == 512 && Subtarget.useAVX512Regs())) { bool HasPT = Subtarget.hasSSE41(); // PTEST and MOVMSK are slow on Knights Landing and Knights Mill and widened diff --git a/llvm/lib/Target/X86/X86LoadValueInjectionLoadHardening.cpp b/llvm/lib/Target/X86/X86LoadValueInjectionLoadHardening.cpp index 090060e..3b96e70 100644 --- a/llvm/lib/Target/X86/X86LoadValueInjectionLoadHardening.cpp +++ b/llvm/lib/Target/X86/X86LoadValueInjectionLoadHardening.cpp @@ -115,9 +115,9 @@ struct MachineGadgetGraph : ImmutableGraph<MachineInstr *, int> { static constexpr MachineInstr *const ArgNodeSentinel = nullptr; using GraphT = ImmutableGraph<MachineInstr *, int>; - using Node = typename GraphT::Node; - using Edge = typename GraphT::Edge; - using size_type = typename GraphT::size_type; + using Node = GraphT::Node; + using Edge = GraphT::Edge; + using size_type = GraphT::size_type; MachineGadgetGraph(std::unique_ptr<Node[]> Nodes, std::unique_ptr<Edge[]> Edges, size_type NodesSize, size_type EdgesSize, int NumFences = 0, int NumGadgets = 0) @@ -191,10 +191,10 @@ template <> struct DOTGraphTraits<MachineGadgetGraph *> : DefaultDOTGraphTraits { using GraphType = MachineGadgetGraph; using Traits = llvm::GraphTraits<GraphType *>; - using NodeRef = typename Traits::NodeRef; - using EdgeRef = typename Traits::EdgeRef; - using ChildIteratorType = typename Traits::ChildIteratorType; - using ChildEdgeIteratorType = typename Traits::ChildEdgeIteratorType; + using NodeRef = Traits::NodeRef; + using EdgeRef = Traits::EdgeRef; + using ChildIteratorType = Traits::ChildIteratorType; + using ChildEdgeIteratorType = Traits::ChildEdgeIteratorType; DOTGraphTraits(bool IsSimple = false) : DefaultDOTGraphTraits(IsSimple) {} @@ -227,9 +227,6 @@ struct DOTGraphTraits<MachineGadgetGraph *> : DefaultDOTGraphTraits { } // end namespace llvm -constexpr MachineInstr *MachineGadgetGraph::ArgNodeSentinel; -constexpr int MachineGadgetGraph::GadgetEdgeSentinel; - char X86LoadValueInjectionLoadHardeningPass::ID = 0; void X86LoadValueInjectionLoadHardeningPass::getAnalysisUsage( @@ -335,7 +332,7 @@ X86LoadValueInjectionLoadHardeningPass::getGadgetGraph( L.computePhiInfo(); GraphBuilder Builder; - using GraphIter = typename GraphBuilder::BuilderNodeRef; + using GraphIter = GraphBuilder::BuilderNodeRef; DenseMap<MachineInstr *, GraphIter> NodeMap; int FenceCount = 0, GadgetCount = 0; auto MaybeAddNode = [&NodeMap, &Builder](MachineInstr *MI) { diff --git a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp index 3d8d0a23..0b1430e 100644 --- a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp +++ b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp @@ -6562,7 +6562,7 @@ bool X86TTIImpl::areInlineCompatible(const Function *Caller, bool X86TTIImpl::areTypesABICompatible(const Function *Caller, const Function *Callee, - const ArrayRef<Type *> &Types) const { + ArrayRef<Type *> Types) const { if (!BaseT::areTypesABICompatible(Caller, Callee, Types)) return false; diff --git a/llvm/lib/Target/X86/X86TargetTransformInfo.h b/llvm/lib/Target/X86/X86TargetTransformInfo.h index 133b366..de5e1c2 100644 --- a/llvm/lib/Target/X86/X86TargetTransformInfo.h +++ b/llvm/lib/Target/X86/X86TargetTransformInfo.h @@ -296,7 +296,7 @@ public: bool areInlineCompatible(const Function *Caller, const Function *Callee) const override; bool areTypesABICompatible(const Function *Caller, const Function *Callee, - const ArrayRef<Type *> &Type) const override; + ArrayRef<Type *> Type) const override; uint64_t getMaxMemIntrinsicInlineSizeThreshold() const override { return ST->getMaxInlineSizeThreshold(); diff --git a/llvm/lib/TargetParser/PPCTargetParser.cpp b/llvm/lib/TargetParser/PPCTargetParser.cpp index d510445..f74d670 100644 --- a/llvm/lib/TargetParser/PPCTargetParser.cpp +++ b/llvm/lib/TargetParser/PPCTargetParser.cpp @@ -48,9 +48,9 @@ StringRef normalizeCPUName(StringRef CPUName) { // accepting it. Clang has always ignored it and passed the // generic CPU ID to the back end. return StringSwitch<StringRef>(CPUName) - .Cases("common", "405", "generic") - .Cases("ppc440", "440fp", "440") - .Cases("630", "power3", "pwr3") + .Cases({"common", "405"}, "generic") + .Cases({"ppc440", "440fp"}, "440") + .Cases({"630", "power3"}, "pwr3") .Case("G3", "g3") .Case("G4", "g4") .Case("G4+", "g4+") @@ -69,7 +69,7 @@ StringRef normalizeCPUName(StringRef CPUName) { .Case("power9", "pwr9") .Case("power10", "pwr10") .Case("power11", "pwr11") - .Cases("powerpc", "powerpc32", "ppc") + .Cases({"powerpc", "powerpc32"}, "ppc") .Case("powerpc64", "ppc64") .Case("powerpc64le", "ppc64le") .Default(CPUName); diff --git a/llvm/lib/TextAPI/BinaryReader/DylibReader.cpp b/llvm/lib/TextAPI/BinaryReader/DylibReader.cpp index cda07e8..f55bc9c 100644 --- a/llvm/lib/TextAPI/BinaryReader/DylibReader.cpp +++ b/llvm/lib/TextAPI/BinaryReader/DylibReader.cpp @@ -32,7 +32,7 @@ using namespace llvm::MachO; using namespace llvm::MachO::DylibReader; using TripleVec = std::vector<Triple>; -static typename TripleVec::iterator emplace(TripleVec &Container, Triple &&T) { +static TripleVec::iterator emplace(TripleVec &Container, Triple &&T) { auto I = partition_point(Container, [=](const Triple &CT) { return std::forward_as_tuple(CT.getArch(), CT.getOS(), CT.getEnvironment()) < diff --git a/llvm/lib/TextAPI/RecordVisitor.cpp b/llvm/lib/TextAPI/RecordVisitor.cpp index d333b33..24971a7 100644 --- a/llvm/lib/TextAPI/RecordVisitor.cpp +++ b/llvm/lib/TextAPI/RecordVisitor.cpp @@ -15,7 +15,7 @@ using namespace llvm; using namespace llvm::MachO; -RecordVisitor::~RecordVisitor() {} +RecordVisitor::~RecordVisitor() = default; void RecordVisitor::visitObjCInterface(const ObjCInterfaceRecord &) {} void RecordVisitor::visitObjCCategory(const ObjCCategoryRecord &) {} diff --git a/llvm/lib/Transforms/Coroutines/CoroCloner.h b/llvm/lib/Transforms/Coroutines/CoroCloner.h index e05fe28..1e549f1 100644 --- a/llvm/lib/Transforms/Coroutines/CoroCloner.h +++ b/llvm/lib/Transforms/Coroutines/CoroCloner.h @@ -77,7 +77,7 @@ public: : OrigF(OrigF), Suffix(Suffix), Shape(Shape), FKind(FKind), Builder(OrigF.getContext()), TTI(TTI) {} - virtual ~BaseCloner() {} + virtual ~BaseCloner() = default; /// Create a clone for a continuation lowering. static Function *createClone(Function &OrigF, const Twine &Suffix, diff --git a/llvm/lib/Transforms/IPO/AttributorAttributes.cpp b/llvm/lib/Transforms/IPO/AttributorAttributes.cpp index 5048561..5ed47ae 100644 --- a/llvm/lib/Transforms/IPO/AttributorAttributes.cpp +++ b/llvm/lib/Transforms/IPO/AttributorAttributes.cpp @@ -3619,7 +3619,7 @@ struct AAIntraFnReachabilityFunction final return true; RQITy StackRQI(A, From, To, ExclusionSet, false); - typename RQITy::Reachable Result; + RQITy::Reachable Result; if (!NonConstThis->checkQueryCache(A, StackRQI, Result)) return NonConstThis->isReachableImpl(A, StackRQI, /*IsTemporaryRQI=*/true); @@ -10701,7 +10701,7 @@ struct AAInterFnReachabilityFunction auto *NonConstThis = const_cast<AAInterFnReachabilityFunction *>(this); RQITy StackRQI(A, From, To, ExclusionSet, false); - typename RQITy::Reachable Result; + RQITy::Reachable Result; if (!NonConstThis->checkQueryCache(A, StackRQI, Result)) return NonConstThis->isReachableImpl(A, StackRQI, /*IsTemporaryRQI=*/true); diff --git a/llvm/lib/Transforms/IPO/MemProfContextDisambiguation.cpp b/llvm/lib/Transforms/IPO/MemProfContextDisambiguation.cpp index 894d83f..d35ae47 100644 --- a/llvm/lib/Transforms/IPO/MemProfContextDisambiguation.cpp +++ b/llvm/lib/Transforms/IPO/MemProfContextDisambiguation.cpp @@ -1034,11 +1034,11 @@ private: } // namespace template <> -struct llvm::DenseMapInfo<typename CallsiteContextGraph< +struct llvm::DenseMapInfo<CallsiteContextGraph< ModuleCallsiteContextGraph, Function, Instruction *>::CallInfo> : public DenseMapInfo<std::pair<Instruction *, unsigned>> {}; template <> -struct llvm::DenseMapInfo<typename CallsiteContextGraph< +struct llvm::DenseMapInfo<CallsiteContextGraph< IndexCallsiteContextGraph, FunctionSummary, IndexCall>::CallInfo> : public DenseMapInfo<std::pair<IndexCall, unsigned>> {}; template <> diff --git a/llvm/lib/Transforms/IPO/OpenMPOpt.cpp b/llvm/lib/Transforms/IPO/OpenMPOpt.cpp index d7eb745..2a87a0f 100644 --- a/llvm/lib/Transforms/IPO/OpenMPOpt.cpp +++ b/llvm/lib/Transforms/IPO/OpenMPOpt.cpp @@ -208,7 +208,7 @@ namespace KernelInfo { // }; #define KERNEL_ENVIRONMENT_IDX(MEMBER, IDX) \ - constexpr const unsigned MEMBER##Idx = IDX; + constexpr unsigned MEMBER##Idx = IDX; KERNEL_ENVIRONMENT_IDX(Configuration, 0) KERNEL_ENVIRONMENT_IDX(Ident, 1) @@ -216,7 +216,7 @@ KERNEL_ENVIRONMENT_IDX(Ident, 1) #undef KERNEL_ENVIRONMENT_IDX #define KERNEL_ENVIRONMENT_CONFIGURATION_IDX(MEMBER, IDX) \ - constexpr const unsigned MEMBER##Idx = IDX; + constexpr unsigned MEMBER##Idx = IDX; KERNEL_ENVIRONMENT_CONFIGURATION_IDX(UseGenericStateMachine, 0) KERNEL_ENVIRONMENT_CONFIGURATION_IDX(MayUseNestedParallelism, 1) @@ -258,7 +258,7 @@ KERNEL_ENVIRONMENT_CONFIGURATION_GETTER(MaxTeams) GlobalVariable * getKernelEnvironementGVFromKernelInitCB(CallBase *KernelInitCB) { - constexpr const int InitKernelEnvironmentArgNo = 0; + constexpr int InitKernelEnvironmentArgNo = 0; return cast<GlobalVariable>( KernelInitCB->getArgOperand(InitKernelEnvironmentArgNo) ->stripPointerCasts()); diff --git a/llvm/lib/Transforms/InstCombine/InstCombineAndOrXor.cpp b/llvm/lib/Transforms/InstCombine/InstCombineAndOrXor.cpp index 3ddf182..cbaff29 100644 --- a/llvm/lib/Transforms/InstCombine/InstCombineAndOrXor.cpp +++ b/llvm/lib/Transforms/InstCombine/InstCombineAndOrXor.cpp @@ -3997,6 +3997,27 @@ static Value *foldOrUnsignedUMulOverflowICmp(BinaryOperator &I, return nullptr; } +/// Fold select(X >s 0, 0, -X) | smax(X, 0) --> abs(X) +/// select(X <s 0, -X, 0) | smax(X, 0) --> abs(X) +static Value *FoldOrOfSelectSmaxToAbs(BinaryOperator &I, + InstCombiner::BuilderTy &Builder) { + Value *X; + Value *Sel; + if (match(&I, + m_c_Or(m_Value(Sel), m_OneUse(m_SMax(m_Value(X), m_ZeroInt()))))) { + auto NegX = m_Neg(m_Specific(X)); + if (match(Sel, m_Select(m_SpecificICmp(ICmpInst::ICMP_SGT, m_Specific(X), + m_ZeroInt()), + m_ZeroInt(), NegX)) || + match(Sel, m_Select(m_SpecificICmp(ICmpInst::ICMP_SLT, m_Specific(X), + m_ZeroInt()), + NegX, m_ZeroInt()))) + return Builder.CreateBinaryIntrinsic(Intrinsic::abs, X, + Builder.getFalse()); + } + return nullptr; +} + // FIXME: We use commutative matchers (m_c_*) for some, but not all, matches // here. We should standardize that construct where it is needed or choose some // other way to ensure that commutated variants of patterns are not missed. @@ -4545,6 +4566,9 @@ Instruction *InstCombinerImpl::visitOr(BinaryOperator &I) { if (Value *V = SimplifyAddWithRemainder(I)) return replaceInstUsesWith(I, V); + if (Value *Res = FoldOrOfSelectSmaxToAbs(I, Builder)) + return replaceInstUsesWith(I, Res); + return nullptr; } diff --git a/llvm/lib/Transforms/Instrumentation/NumericalStabilitySanitizer.cpp b/llvm/lib/Transforms/Instrumentation/NumericalStabilitySanitizer.cpp index 80e77e09..a2fad02 100644 --- a/llvm/lib/Transforms/Instrumentation/NumericalStabilitySanitizer.cpp +++ b/llvm/lib/Transforms/Instrumentation/NumericalStabilitySanitizer.cpp @@ -161,7 +161,7 @@ template <char NsanTypeId> class ShadowTypeConfigImpl : public ShadowTypeConfig { public: char getNsanTypeId() const override { return NsanTypeId; } - static constexpr const char kNsanTypeId = NsanTypeId; + static constexpr char kNsanTypeId = NsanTypeId; }; // `double` (`d`) shadow type. diff --git a/llvm/lib/Transforms/Scalar/GVNSink.cpp b/llvm/lib/Transforms/Scalar/GVNSink.cpp index a06f832..d564e32 100644 --- a/llvm/lib/Transforms/Scalar/GVNSink.cpp +++ b/llvm/lib/Transforms/Scalar/GVNSink.cpp @@ -514,7 +514,7 @@ public: class GVNSink { public: - GVNSink() {} + GVNSink() = default; bool run(Function &F) { LLVM_DEBUG(dbgs() << "GVNSink: running on function @" << F.getName() diff --git a/llvm/lib/Transforms/Scalar/LowerMatrixIntrinsics.cpp b/llvm/lib/Transforms/Scalar/LowerMatrixIntrinsics.cpp index 3487e81..7e70ba2 100644 --- a/llvm/lib/Transforms/Scalar/LowerMatrixIntrinsics.cpp +++ b/llvm/lib/Transforms/Scalar/LowerMatrixIntrinsics.cpp @@ -245,11 +245,14 @@ raw_ostream &operator<<(raw_ostream &OS, ShapeInfo SI) { } // namespace -static bool isUniformShape(Value *V) { +static bool isShapePreserving(Value *V) { Instruction *I = dyn_cast<Instruction>(V); if (!I) return true; + if (isa<SelectInst>(I)) + return true; + if (I->isBinaryOp()) return true; @@ -300,6 +303,16 @@ static bool isUniformShape(Value *V) { } } +/// Return an iterator over the operands of \p I that should share shape +/// information with \p I. +static iterator_range<Use *> getShapedOperandsForInst(Instruction *I) { + assert(isShapePreserving(I) && + "Can't retrieve shaped operands for an instruction that does not " + "preserve shape information"); + auto Ops = I->operands(); + return isa<SelectInst>(I) ? drop_begin(Ops) : Ops; +} + /// Return the ShapeInfo for the result of \p I, it it can be determined. static std::optional<ShapeInfo> computeShapeInfoForInst(Instruction *I, @@ -329,9 +342,8 @@ computeShapeInfoForInst(Instruction *I, return OpShape->second; } - if (isUniformShape(I) || isa<SelectInst>(I)) { - auto Ops = I->operands(); - auto ShapedOps = isa<SelectInst>(I) ? drop_begin(Ops) : Ops; + if (isShapePreserving(I)) { + auto ShapedOps = getShapedOperandsForInst(I); // Find the first operand that has a known shape and use that. for (auto &Op : ShapedOps) { auto OpShape = ShapeMap.find(Op.get()); @@ -710,10 +722,9 @@ public: case Intrinsic::matrix_column_major_store: return true; default: - return isUniformShape(II); + break; } - return isUniformShape(V) || isa<StoreInst>(V) || isa<LoadInst>(V) || - isa<SelectInst>(V); + return isShapePreserving(V) || isa<StoreInst>(V) || isa<LoadInst>(V); } /// Propagate the shape information of instructions to their users. @@ -800,9 +811,8 @@ public: } else if (isa<StoreInst>(V)) { // Nothing to do. We forward-propagated to this so we would just // backward propagate to an instruction with an already known shape. - } else if (isUniformShape(V) || isa<SelectInst>(V)) { - auto Ops = cast<Instruction>(V)->operands(); - auto ShapedOps = isa<SelectInst>(V) ? drop_begin(Ops) : Ops; + } else if (isShapePreserving(V)) { + auto ShapedOps = getShapedOperandsForInst(cast<Instruction>(V)); // Propagate to all operands. ShapeInfo Shape = ShapeMap[V]; for (Use &U : ShapedOps) { diff --git a/llvm/lib/Transforms/Scalar/StructurizeCFG.cpp b/llvm/lib/Transforms/Scalar/StructurizeCFG.cpp index 0f3978f..5f6f66a 100644 --- a/llvm/lib/Transforms/Scalar/StructurizeCFG.cpp +++ b/llvm/lib/Transforms/Scalar/StructurizeCFG.cpp @@ -143,8 +143,8 @@ struct SubGraphTraits { class WrappedSuccIterator : public iterator_adaptor_base< WrappedSuccIterator, BaseSuccIterator, - typename std::iterator_traits<BaseSuccIterator>::iterator_category, - NodeRef, std::ptrdiff_t, NodeRef *, NodeRef> { + std::iterator_traits<BaseSuccIterator>::iterator_category, NodeRef, + std::ptrdiff_t, NodeRef *, NodeRef> { SmallDenseSet<RegionNode *> *Nodes; public: diff --git a/llvm/lib/Transforms/Vectorize/LoopVectorizationPlanner.h b/llvm/lib/Transforms/Vectorize/LoopVectorizationPlanner.h index 5298728..04b0562 100644 --- a/llvm/lib/Transforms/Vectorize/LoopVectorizationPlanner.h +++ b/llvm/lib/Transforms/Vectorize/LoopVectorizationPlanner.h @@ -167,7 +167,7 @@ public: DebugLoc DL = DebugLoc::getUnknown(), const Twine &Name = "") { return tryInsertInstruction( - new VPInstruction(Opcode, Operands, Flags, DL, Name)); + new VPInstruction(Opcode, Operands, Flags, {}, DL, Name)); } VPInstruction *createNaryOp(unsigned Opcode, ArrayRef<VPValue *> Operands, @@ -184,7 +184,7 @@ public: DebugLoc DL = DebugLoc::getUnknown(), const Twine &Name = "") { return tryInsertInstruction( - new VPInstruction(Opcode, Operands, WrapFlags, DL, Name)); + new VPInstruction(Opcode, Operands, WrapFlags, {}, DL, Name)); } VPInstruction *createNot(VPValue *Operand, @@ -205,7 +205,7 @@ public: return tryInsertInstruction(new VPInstruction( Instruction::BinaryOps::Or, {LHS, RHS}, - VPRecipeWithIRFlags::DisjointFlagsTy(false), DL, Name)); + VPRecipeWithIRFlags::DisjointFlagsTy(false), {}, DL, Name)); } VPInstruction *createLogicalAnd(VPValue *LHS, VPValue *RHS, @@ -221,7 +221,7 @@ public: std::optional<FastMathFlags> FMFs = std::nullopt) { auto *Select = FMFs ? new VPInstruction(Instruction::Select, {Cond, TrueVal, FalseVal}, - *FMFs, DL, Name) + *FMFs, {}, DL, Name) : new VPInstruction(Instruction::Select, {Cond, TrueVal, FalseVal}, DL, Name); return tryInsertInstruction(Select); @@ -235,7 +235,7 @@ public: assert(Pred >= CmpInst::FIRST_ICMP_PREDICATE && Pred <= CmpInst::LAST_ICMP_PREDICATE && "invalid predicate"); return tryInsertInstruction( - new VPInstruction(Instruction::ICmp, {A, B}, Pred, DL, Name)); + new VPInstruction(Instruction::ICmp, {A, B}, Pred, {}, DL, Name)); } /// Create a new FCmp VPInstruction with predicate \p Pred and operands \p A @@ -246,7 +246,7 @@ public: assert(Pred >= CmpInst::FIRST_FCMP_PREDICATE && Pred <= CmpInst::LAST_FCMP_PREDICATE && "invalid predicate"); return tryInsertInstruction( - new VPInstruction(Instruction::FCmp, {A, B}, Pred, DL, Name)); + new VPInstruction(Instruction::FCmp, {A, B}, Pred, {}, DL, Name)); } VPInstruction *createPtrAdd(VPValue *Ptr, VPValue *Offset, @@ -254,7 +254,7 @@ public: const Twine &Name = "") { return tryInsertInstruction( new VPInstruction(VPInstruction::PtrAdd, {Ptr, Offset}, - GEPNoWrapFlags::none(), DL, Name)); + GEPNoWrapFlags::none(), {}, DL, Name)); } VPInstruction *createNoWrapPtrAdd(VPValue *Ptr, VPValue *Offset, @@ -262,7 +262,7 @@ public: DebugLoc DL = DebugLoc::getUnknown(), const Twine &Name = "") { return tryInsertInstruction(new VPInstruction( - VPInstruction::PtrAdd, {Ptr, Offset}, GEPFlags, DL, Name)); + VPInstruction::PtrAdd, {Ptr, Offset}, GEPFlags, {}, DL, Name)); } VPInstruction *createWidePtrAdd(VPValue *Ptr, VPValue *Offset, @@ -270,7 +270,7 @@ public: const Twine &Name = "") { return tryInsertInstruction( new VPInstruction(VPInstruction::WidePtrAdd, {Ptr, Offset}, - GEPNoWrapFlags::none(), DL, Name)); + GEPNoWrapFlags::none(), {}, DL, Name)); } VPPhi *createScalarPhi(ArrayRef<VPValue *> IncomingValues, DebugLoc DL, @@ -303,9 +303,11 @@ public: } VPInstruction *createScalarCast(Instruction::CastOps Opcode, VPValue *Op, - Type *ResultTy, DebugLoc DL) { + Type *ResultTy, DebugLoc DL, + const VPIRFlags &Flags = {}, + const VPIRMetadata &Metadata = {}) { return tryInsertInstruction( - new VPInstructionWithType(Opcode, Op, ResultTy, {}, DL)); + new VPInstructionWithType(Opcode, Op, ResultTy, DL, Flags, Metadata)); } VPValue *createScalarZExtOrTrunc(VPValue *Op, Type *ResultTy, Type *SrcTy, diff --git a/llvm/lib/Transforms/Vectorize/SandboxVectorizer/DependencyGraph.cpp b/llvm/lib/Transforms/Vectorize/SandboxVectorizer/DependencyGraph.cpp index 9c869dd..d354933 100644 --- a/llvm/lib/Transforms/Vectorize/SandboxVectorizer/DependencyGraph.cpp +++ b/llvm/lib/Transforms/Vectorize/SandboxVectorizer/DependencyGraph.cpp @@ -92,7 +92,7 @@ void MemDGNode::print(raw_ostream &OS, bool PrintDeps) const { DGNode::print(OS, false); if (PrintDeps) { // Print memory preds. - static constexpr const unsigned Indent = 4; + static constexpr unsigned Indent = 4; for (auto *Pred : MemPreds) OS.indent(Indent) << "<-" << *Pred->getInstruction() << "\n"; } diff --git a/llvm/lib/Transforms/Vectorize/SandboxVectorizer/Passes/BottomUpVec.cpp b/llvm/lib/Transforms/Vectorize/SandboxVectorizer/Passes/BottomUpVec.cpp index 86dbd21..5534da9 100644 --- a/llvm/lib/Transforms/Vectorize/SandboxVectorizer/Passes/BottomUpVec.cpp +++ b/llvm/lib/Transforms/Vectorize/SandboxVectorizer/Passes/BottomUpVec.cpp @@ -25,14 +25,14 @@ static cl::opt<bool> "emit new instructions (*very* expensive).")); #endif // NDEBUG -static constexpr const unsigned long StopAtDisabled = +static constexpr unsigned long StopAtDisabled = std::numeric_limits<unsigned long>::max(); static cl::opt<unsigned long> StopAt("sbvec-stop-at", cl::init(StopAtDisabled), cl::Hidden, cl::desc("Vectorize if the invocation count is < than this. 0 " "disables vectorization.")); -static constexpr const unsigned long StopBundleDisabled = +static constexpr unsigned long StopBundleDisabled = std::numeric_limits<unsigned long>::max(); static cl::opt<unsigned long> StopBundle("sbvec-stop-bndl", cl::init(StopBundleDisabled), cl::Hidden, diff --git a/llvm/lib/Transforms/Vectorize/SandboxVectorizer/SandboxVectorizer.cpp b/llvm/lib/Transforms/Vectorize/SandboxVectorizer/SandboxVectorizer.cpp index ed2f80b..2de6921 100644 --- a/llvm/lib/Transforms/Vectorize/SandboxVectorizer/SandboxVectorizer.cpp +++ b/llvm/lib/Transforms/Vectorize/SandboxVectorizer/SandboxVectorizer.cpp @@ -43,7 +43,7 @@ cl::opt<std::string> AllowFiles( "sbvec-allow-files", cl::init(".*"), cl::Hidden, cl::desc("Run the vectorizer only on file paths that match any in the " "list of comma-separated regex's.")); -static constexpr const char AllowFilesDelim = ','; +static constexpr char AllowFilesDelim = ','; SandboxVectorizerPass::SandboxVectorizerPass() : FPM("fpm") { if (UserDefinedPassPipeline == DefaultPipelineMagicStr) { diff --git a/llvm/lib/Transforms/Vectorize/VPlan.h b/llvm/lib/Transforms/Vectorize/VPlan.h index 08c9c15..aba6d35 100644 --- a/llvm/lib/Transforms/Vectorize/VPlan.h +++ b/llvm/lib/Transforms/Vectorize/VPlan.h @@ -939,7 +939,7 @@ class VPIRMetadata { SmallVector<std::pair<unsigned, MDNode *>> Metadata; public: - VPIRMetadata() {} + VPIRMetadata() = default; /// Adds metatadata that can be preserved from the original instruction /// \p I. @@ -950,12 +950,9 @@ public: VPIRMetadata(Instruction &I, LoopVersioning *LVer); /// Copy constructor for cloning. - VPIRMetadata(const VPIRMetadata &Other) : Metadata(Other.Metadata) {} + VPIRMetadata(const VPIRMetadata &Other) = default; - VPIRMetadata &operator=(const VPIRMetadata &Other) { - Metadata = Other.Metadata; - return *this; - } + VPIRMetadata &operator=(const VPIRMetadata &Other) = default; /// Add all metadata to \p I. void applyMetadata(Instruction &I) const; @@ -1107,14 +1104,15 @@ public: VPIRMetadata(), Opcode(Opcode), Name(Name.str()) {} VPInstruction(unsigned Opcode, ArrayRef<VPValue *> Operands, - const VPIRFlags &Flags, DebugLoc DL = DebugLoc::getUnknown(), - const Twine &Name = ""); + const VPIRFlags &Flags, const VPIRMetadata &MD = {}, + DebugLoc DL = DebugLoc::getUnknown(), const Twine &Name = ""); VP_CLASSOF_IMPL(VPDef::VPInstructionSC) VPInstruction *clone() override { SmallVector<VPValue *, 2> Operands(operands()); - auto *New = new VPInstruction(Opcode, Operands, *this, getDebugLoc(), Name); + auto *New = + new VPInstruction(Opcode, Operands, *this, *this, getDebugLoc(), Name); if (getUnderlyingValue()) New->setUnderlyingValue(getUnderlyingInstr()); return New; @@ -1196,7 +1194,14 @@ public: VPInstructionWithType(unsigned Opcode, ArrayRef<VPValue *> Operands, Type *ResultTy, const VPIRFlags &Flags, DebugLoc DL, const Twine &Name = "") - : VPInstruction(Opcode, Operands, Flags, DL, Name), ResultTy(ResultTy) {} + : VPInstruction(Opcode, Operands, Flags, {}, DL, Name), + ResultTy(ResultTy) {} + + VPInstructionWithType(unsigned Opcode, ArrayRef<VPValue *> Operands, + Type *ResultTy, DebugLoc DL, const VPIRFlags &Flags, + const VPIRMetadata &Metadata, const Twine &Name = "") + : VPInstruction(Opcode, Operands, Flags, Metadata, DL, Name), + ResultTy(ResultTy) {} static inline bool classof(const VPRecipeBase *R) { // VPInstructionWithType are VPInstructions with specific opcodes requiring @@ -3977,7 +3982,7 @@ class VPIRBasicBlock : public VPBasicBlock { IRBB(IRBB) {} public: - ~VPIRBasicBlock() override {} + ~VPIRBasicBlock() override = default; static inline bool classof(const VPBlockBase *V) { return V->getVPBlockID() == VPBlockBase::VPIRBasicBlockSC; @@ -4029,7 +4034,7 @@ class LLVM_ABI_FOR_TEST VPRegionBlock : public VPBlockBase { IsReplicator(IsReplicator) {} public: - ~VPRegionBlock() override {} + ~VPRegionBlock() override = default; /// Method to support type inquiry through isa, cast, and dyn_cast. static inline bool classof(const VPBlockBase *V) { diff --git a/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp b/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp index f9c15a3..1ee405a 100644 --- a/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp +++ b/llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp @@ -162,8 +162,12 @@ bool VPRecipeBase::mayHaveSideEffects() const { case VPPredInstPHISC: case VPVectorEndPointerSC: return false; - case VPInstructionSC: - return mayWriteToMemory(); + case VPInstructionSC: { + auto *VPI = cast<VPInstruction>(this); + return mayWriteToMemory() || + VPI->getOpcode() == VPInstruction::BranchOnCount || + VPI->getOpcode() == VPInstruction::BranchOnCond; + } case VPWidenCallSC: { Function *Fn = cast<VPWidenCallRecipe>(this)->getCalledScalarFunction(); return mayWriteToMemory() || !Fn->doesNotThrow() || !Fn->willReturn(); @@ -490,10 +494,10 @@ template class VPUnrollPartAccessor<3>; } VPInstruction::VPInstruction(unsigned Opcode, ArrayRef<VPValue *> Operands, - const VPIRFlags &Flags, DebugLoc DL, - const Twine &Name) + const VPIRFlags &Flags, const VPIRMetadata &MD, + DebugLoc DL, const Twine &Name) : VPRecipeWithIRFlags(VPDef::VPInstructionSC, Operands, Flags, DL), - VPIRMetadata(), Opcode(Opcode), Name(Name.str()) { + VPIRMetadata(MD), Opcode(Opcode), Name(Name.str()) { assert(flagsValidForOpcode(getOpcode()) && "Set flags not supported for the provided opcode"); assert((getNumOperandsForOpcode(Opcode) == -1u || @@ -1241,6 +1245,8 @@ bool VPInstruction::opcodeMayReadOrWriteFromMemory() const { case Instruction::Select: case Instruction::PHI: case VPInstruction::AnyOf: + case VPInstruction::BranchOnCond: + case VPInstruction::BranchOnCount: case VPInstruction::Broadcast: case VPInstruction::BuildStructVector: case VPInstruction::BuildVector: diff --git a/llvm/lib/Transforms/Vectorize/VPlanSLP.h b/llvm/lib/Transforms/Vectorize/VPlanSLP.h index 77ff36c..44972c68 100644 --- a/llvm/lib/Transforms/Vectorize/VPlanSLP.h +++ b/llvm/lib/Transforms/Vectorize/VPlanSLP.h @@ -89,8 +89,7 @@ class VPlanSlp { /// Width of the widest combined bundle in bits. unsigned WidestBundleBits = 0; - using MultiNodeOpTy = - typename std::pair<VPInstruction *, SmallVector<VPValue *, 4>>; + using MultiNodeOpTy = std::pair<VPInstruction *, SmallVector<VPValue *, 4>>; // Input operand bundles for the current multi node. Each multi node operand // bundle contains values not matching the multi node's opcode. They will diff --git a/llvm/lib/Transforms/Vectorize/VPlanTransforms.cpp b/llvm/lib/Transforms/Vectorize/VPlanTransforms.cpp index 6a8231b..b455368 100644 --- a/llvm/lib/Transforms/Vectorize/VPlanTransforms.cpp +++ b/llvm/lib/Transforms/Vectorize/VPlanTransforms.cpp @@ -151,7 +151,27 @@ static bool cannotHoistOrSinkRecipe(const VPRecipeBase &R) { static bool sinkScalarOperands(VPlan &Plan) { auto Iter = vp_depth_first_deep(Plan.getEntry()); + bool ScalarVFOnly = Plan.hasScalarVFOnly(); bool Changed = false; + + auto IsValidSinkCandidate = [ScalarVFOnly](VPBasicBlock *SinkTo, + VPSingleDefRecipe *Candidate) { + // We only know how to duplicate VPReplicateRecipes and + // VPScalarIVStepsRecipes for now. + if (!isa<VPReplicateRecipe, VPScalarIVStepsRecipe>(Candidate)) + return false; + + if (Candidate->getParent() == SinkTo || Candidate->mayHaveSideEffects() || + Candidate->mayReadOrWriteMemory()) + return false; + + if (auto *RepR = dyn_cast<VPReplicateRecipe>(Candidate)) + if (!ScalarVFOnly && RepR->isSingleScalar()) + return false; + + return true; + }; + // First, collect the operands of all recipes in replicate blocks as seeds for // sinking. SetVector<std::pair<VPBasicBlock *, VPSingleDefRecipe *>> WorkList; @@ -159,51 +179,37 @@ static bool sinkScalarOperands(VPlan &Plan) { VPBasicBlock *EntryVPBB = VPR->getEntryBasicBlock(); if (!VPR->isReplicator() || EntryVPBB->getSuccessors().size() != 2) continue; - VPBasicBlock *VPBB = dyn_cast<VPBasicBlock>(EntryVPBB->getSuccessors()[0]); - if (!VPBB || VPBB->getSingleSuccessor() != VPR->getExitingBasicBlock()) + VPBasicBlock *VPBB = cast<VPBasicBlock>(EntryVPBB->getSuccessors().front()); + if (VPBB->getSingleSuccessor() != VPR->getExitingBasicBlock()) continue; for (auto &Recipe : *VPBB) { - for (VPValue *Op : Recipe.operands()) + for (VPValue *Op : Recipe.operands()) { if (auto *Def = dyn_cast_or_null<VPSingleDefRecipe>(Op->getDefiningRecipe())) - WorkList.insert({VPBB, Def}); + if (IsValidSinkCandidate(VPBB, Def)) + WorkList.insert({VPBB, Def}); + } } } - bool ScalarVFOnly = Plan.hasScalarVFOnly(); // Try to sink each replicate or scalar IV steps recipe in the worklist. for (unsigned I = 0; I != WorkList.size(); ++I) { VPBasicBlock *SinkTo; VPSingleDefRecipe *SinkCandidate; std::tie(SinkTo, SinkCandidate) = WorkList[I]; - if (SinkCandidate->getParent() == SinkTo || - SinkCandidate->mayHaveSideEffects() || - SinkCandidate->mayReadOrWriteMemory()) - continue; - if (auto *RepR = dyn_cast<VPReplicateRecipe>(SinkCandidate)) { - if (!ScalarVFOnly && RepR->isSingleScalar()) - continue; - } else if (!isa<VPScalarIVStepsRecipe>(SinkCandidate)) - continue; - bool NeedsDuplicating = false; // All recipe users of the sink candidate must be in the same block SinkTo - // or all users outside of SinkTo must be uniform-after-vectorization ( - // i.e., only first lane is used) . In the latter case, we need to duplicate - // SinkCandidate. - auto CanSinkWithUser = [SinkTo, &NeedsDuplicating, - SinkCandidate](VPUser *U) { - auto *UI = cast<VPRecipeBase>(U); - if (UI->getParent() == SinkTo) - return true; - NeedsDuplicating = UI->onlyFirstLaneUsed(SinkCandidate); - // We only know how to duplicate VPReplicateRecipes and - // VPScalarIVStepsRecipes for now. - return NeedsDuplicating && - isa<VPReplicateRecipe, VPScalarIVStepsRecipe>(SinkCandidate); - }; - if (!all_of(SinkCandidate->users(), CanSinkWithUser)) + // or all users outside of SinkTo must have only their first lane used. In + // the latter case, we need to duplicate SinkCandidate. + auto UsersOutsideSinkTo = + make_filter_range(SinkCandidate->users(), [SinkTo](VPUser *U) { + return cast<VPRecipeBase>(U)->getParent() != SinkTo; + }); + if (any_of(UsersOutsideSinkTo, [SinkCandidate](VPUser *U) { + return !U->onlyFirstLaneUsed(SinkCandidate); + })) continue; + bool NeedsDuplicating = !UsersOutsideSinkTo.empty(); if (NeedsDuplicating) { if (ScalarVFOnly) @@ -230,7 +236,8 @@ static bool sinkScalarOperands(VPlan &Plan) { for (VPValue *Op : SinkCandidate->operands()) if (auto *Def = dyn_cast_or_null<VPSingleDefRecipe>(Op->getDefiningRecipe())) - WorkList.insert({SinkTo, Def}); + if (IsValidSinkCandidate(SinkTo, Def)) + WorkList.insert({SinkTo, Def}); Changed = true; } return Changed; @@ -1281,6 +1288,12 @@ static void simplifyRecipe(VPRecipeBase &R, VPTypeAnalysis &TypeInfo) { return; } + if (match(Def, m_BuildVector()) && all_equal(R.operands())) { + Def->replaceAllUsesWith( + Builder.createNaryOp(VPInstruction::Broadcast, Def->getOperand(0))); + return; + } + if (auto *Phi = dyn_cast<VPPhi>(Def)) { if (Phi->getNumOperands() == 1) Phi->replaceAllUsesWith(Phi->getOperand(0)); @@ -4168,7 +4181,7 @@ void VPlanTransforms::narrowInterleaveGroups(VPlan &Plan, ElementCount VF, unsigned VFMinVal = VF.getKnownMinValue(); SmallVector<VPInterleaveRecipe *> StoreGroups; for (auto &R : *VectorLoop->getEntryBasicBlock()) { - if (isa<VPCanonicalIVPHIRecipe>(&R) || match(&R, m_BranchOnCount())) + if (isa<VPCanonicalIVPHIRecipe>(&R)) continue; if (isa<VPDerivedIVRecipe, VPScalarIVStepsRecipe>(&R) && diff --git a/llvm/lib/Transforms/Vectorize/VPlanUtils.cpp b/llvm/lib/Transforms/Vectorize/VPlanUtils.cpp index 8c23e78..c6380d3 100644 --- a/llvm/lib/Transforms/Vectorize/VPlanUtils.cpp +++ b/llvm/lib/Transforms/Vectorize/VPlanUtils.cpp @@ -32,22 +32,17 @@ bool vputils::onlyScalarValuesUsed(const VPValue *Def) { } VPValue *vputils::getOrCreateVPValueForSCEVExpr(VPlan &Plan, const SCEV *Expr) { - VPValue *Expanded = nullptr; if (auto *E = dyn_cast<SCEVConstant>(Expr)) - Expanded = Plan.getOrAddLiveIn(E->getValue()); - else { - auto *U = dyn_cast<SCEVUnknown>(Expr); - // Skip SCEV expansion if Expr is a SCEVUnknown wrapping a non-instruction - // value. Otherwise the value may be defined in a loop and using it directly - // will break LCSSA form. The SCEV expansion takes care of preserving LCSSA - // form. - if (U && !isa<Instruction>(U->getValue())) { - Expanded = Plan.getOrAddLiveIn(U->getValue()); - } else { - Expanded = new VPExpandSCEVRecipe(Expr); - Plan.getEntry()->appendRecipe(Expanded->getDefiningRecipe()); - } - } + return Plan.getOrAddLiveIn(E->getValue()); + // Skip SCEV expansion if Expr is a SCEVUnknown wrapping a non-instruction + // value. Otherwise the value may be defined in a loop and using it directly + // will break LCSSA form. The SCEV expansion takes care of preserving LCSSA + // form. + auto *U = dyn_cast<SCEVUnknown>(Expr); + if (U && !isa<Instruction>(U->getValue())) + return Plan.getOrAddLiveIn(U->getValue()); + auto *Expanded = new VPExpandSCEVRecipe(Expr); + Plan.getEntry()->appendRecipe(Expanded); return Expanded; } diff --git a/llvm/test/Analysis/CostModel/AArch64/masked_ldst_vls.ll b/llvm/test/Analysis/CostModel/AArch64/masked_ldst_vls.ll index fa53a18..1920fc9 100644 --- a/llvm/test/Analysis/CostModel/AArch64/masked_ldst_vls.ll +++ b/llvm/test/Analysis/CostModel/AArch64/masked_ldst_vls.ll @@ -1,17 +1,6 @@ ; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=256 | FileCheck %s -D#VBITS=256 -; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=384 | FileCheck %s -D#VBITS=256 ; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=512 | FileCheck %s -D#VBITS=512 -; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=640 | FileCheck %s -D#VBITS=512 -; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=768 | FileCheck %s -D#VBITS=512 -; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=896 | FileCheck %s -D#VBITS=512 ; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=1024 | FileCheck %s -D#VBITS=1024 -; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=1152 | FileCheck %s -D#VBITS=1024 -; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=1280 | FileCheck %s -D#VBITS=1024 -; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=1408 | FileCheck %s -D#VBITS=1024 -; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=1536 | FileCheck %s -D#VBITS=1024 -; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=1664 | FileCheck %s -D#VBITS=1024 -; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=1792 | FileCheck %s -D#VBITS=1024 -; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=1920 | FileCheck %s -D#VBITS=1024 ; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=2048 | FileCheck %s -D#VBITS=2048 target datalayout = "e-m:e-i8:8:32-i16:16:32-i64:64-i128:128-n32:64-S128" diff --git a/llvm/test/Analysis/CostModel/AArch64/sve-fixed-length.ll b/llvm/test/Analysis/CostModel/AArch64/sve-fixed-length.ll index df40a96..e128987 100644 --- a/llvm/test/Analysis/CostModel/AArch64/sve-fixed-length.ll +++ b/llvm/test/Analysis/CostModel/AArch64/sve-fixed-length.ll @@ -1,19 +1,8 @@ ; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output | FileCheck %s -D#VBITS=128 ; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=128 | FileCheck %s -D#VBITS=128 ; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=256 | FileCheck %s -D#VBITS=256 -; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=384 | FileCheck %s -D#VBITS=256 ; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=512 | FileCheck %s -D#VBITS=512 -; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=640 | FileCheck %s -D#VBITS=512 -; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=768 | FileCheck %s -D#VBITS=512 -; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=896 | FileCheck %s -D#VBITS=512 ; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=1024 | FileCheck %s -D#VBITS=1024 -; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=1152 | FileCheck %s -D#VBITS=1024 -; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=1280 | FileCheck %s -D#VBITS=1024 -; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=1408 | FileCheck %s -D#VBITS=1024 -; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=1536 | FileCheck %s -D#VBITS=1024 -; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=1664 | FileCheck %s -D#VBITS=1024 -; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=1792 | FileCheck %s -D#VBITS=1024 -; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=1920 | FileCheck %s -D#VBITS=1024 ; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -aarch64-sve-vector-bits-min=2048 | FileCheck %s -D#VBITS=2048 ; VBITS represents the useful bit size of a vector register from the code diff --git a/llvm/test/Assembler/auto_upgrade_nvvm_intrinsics.ll b/llvm/test/Assembler/auto_upgrade_nvvm_intrinsics.ll index 362586a..4fc506f 100644 --- a/llvm/test/Assembler/auto_upgrade_nvvm_intrinsics.ll +++ b/llvm/test/Assembler/auto_upgrade_nvvm_intrinsics.ll @@ -87,6 +87,11 @@ declare void @llvm.nvvm.barrier(i32, i32) declare void @llvm.nvvm.barrier.sync(i32) declare void @llvm.nvvm.barrier.sync.cnt(i32, i32) +declare float @llvm.nvvm.ex2.approx.f(float) +declare double @llvm.nvvm.ex2.approx.d(double) +declare <2 x half> @llvm.nvvm.ex2.approx.f16x2(<2 x half>) +declare float @llvm.nvvm.ex2.approx.ftz.f(float) + ; CHECK-LABEL: @simple_upgrade define void @simple_upgrade(i32 %a, i64 %b, i16 %c) { ; CHECK: call i32 @llvm.bitreverse.i32(i32 %a) @@ -355,3 +360,15 @@ define void @cta_barriers(i32 %x, i32 %y) { call void @llvm.nvvm.barrier.sync.cnt(i32 %x, i32 %y) ret void } + +define void @nvvm_ex2_approx(float %a, double %b, half %c, <2 x half> %d) { +; CHECK: call float @llvm.nvvm.ex2.approx.f32(float %a) +; CHECK: call double @llvm.nvvm.ex2.approx.f64(double %b) +; CHECK: call <2 x half> @llvm.nvvm.ex2.approx.v2f16(<2 x half> %d) +; CHECK: call float @llvm.nvvm.ex2.approx.ftz.f32(float %a) + %r1 = call float @llvm.nvvm.ex2.approx.f(float %a) + %r2 = call double @llvm.nvvm.ex2.approx.d(double %b) + %r3 = call <2 x half> @llvm.nvvm.ex2.approx.f16x2(<2 x half> %d) + %r4 = call float @llvm.nvvm.ex2.approx.ftz.f(float %a) + ret void +} diff --git a/llvm/test/CodeGen/AArch64/sve-fixed-length-loads-stores.ll b/llvm/test/CodeGen/AArch64/sve-fixed-length-loads-stores.ll index becddae..b2ed8de 100644 --- a/llvm/test/CodeGen/AArch64/sve-fixed-length-loads-stores.ll +++ b/llvm/test/CodeGen/AArch64/sve-fixed-length-loads-stores.ll @@ -1,19 +1,8 @@ ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py ; RUN: llc -aarch64-sve-vector-bits-min=128 < %s | not grep ptrue ; RUN: llc -aarch64-sve-vector-bits-min=256 < %s | FileCheck %s -check-prefixes=CHECK,VBITS_GE_256 -; RUN: llc -aarch64-sve-vector-bits-min=384 < %s | FileCheck %s -check-prefixes=CHECK,VBITS_GE_256 ; RUN: llc -aarch64-sve-vector-bits-min=512 < %s | FileCheck %s -check-prefixes=CHECK,VBITS_GE_512 -; RUN: llc -aarch64-sve-vector-bits-min=640 < %s | FileCheck %s -check-prefixes=CHECK,VBITS_GE_512 -; RUN: llc -aarch64-sve-vector-bits-min=768 < %s | FileCheck %s -check-prefixes=CHECK,VBITS_GE_512 -; RUN: llc -aarch64-sve-vector-bits-min=896 < %s | FileCheck %s -check-prefixes=CHECK,VBITS_GE_512 ; RUN: llc -aarch64-sve-vector-bits-min=1024 < %s | FileCheck %s -check-prefixes=CHECK,VBITS_GE_1024 -; RUN: llc -aarch64-sve-vector-bits-min=1152 < %s | FileCheck %s -check-prefixes=CHECK,VBITS_GE_1024 -; RUN: llc -aarch64-sve-vector-bits-min=1280 < %s | FileCheck %s -check-prefixes=CHECK,VBITS_GE_1024 -; RUN: llc -aarch64-sve-vector-bits-min=1408 < %s | FileCheck %s -check-prefixes=CHECK,VBITS_GE_1024 -; RUN: llc -aarch64-sve-vector-bits-min=1536 < %s | FileCheck %s -check-prefixes=CHECK,VBITS_GE_1024 -; RUN: llc -aarch64-sve-vector-bits-min=1664 < %s | FileCheck %s -check-prefixes=CHECK,VBITS_GE_1024 -; RUN: llc -aarch64-sve-vector-bits-min=1792 < %s | FileCheck %s -check-prefixes=CHECK,VBITS_GE_1024 -; RUN: llc -aarch64-sve-vector-bits-min=1920 < %s | FileCheck %s -check-prefixes=CHECK,VBITS_GE_1024 ; RUN: llc -aarch64-sve-vector-bits-min=2048 < %s | FileCheck %s -check-prefixes=CHECK,VBITS_GE_2048 target triple = "aarch64-unknown-linux-gnu" diff --git a/llvm/test/CodeGen/AMDGPU/wait-xcnt.mir b/llvm/test/CodeGen/AMDGPU/wait-xcnt.mir index 1b8e126..a1381ec 100644 --- a/llvm/test/CodeGen/AMDGPU/wait-xcnt.mir +++ b/llvm/test/CodeGen/AMDGPU/wait-xcnt.mir @@ -945,7 +945,6 @@ body: | $vgpr0 = V_MOV_B32_e32 0, implicit $exec ... -# FIXME: Missing S_WAIT_XCNT before overwriting vgpr0. --- name: wait_kmcnt_with_outstanding_vmem_2 tracksRegLiveness: true @@ -971,6 +970,7 @@ body: | ; GCN-NEXT: {{ $}} ; GCN-NEXT: S_WAIT_KMCNT 0 ; GCN-NEXT: $sgpr2 = S_MOV_B32 $sgpr2 + ; GCN-NEXT: S_WAIT_XCNT 0 ; GCN-NEXT: $vgpr0 = V_MOV_B32_e32 0, implicit $exec bb.0: liveins: $vgpr0_vgpr1, $sgpr0_sgpr1, $scc @@ -986,6 +986,180 @@ body: | ... --- +name: wait_kmcnt_and_wait_loadcnt +tracksRegLiveness: true +machineFunctionInfo: + isEntryFunction: true +body: | + ; GCN-LABEL: name: wait_kmcnt_and_wait_loadcnt + ; GCN: bb.0: + ; GCN-NEXT: successors: %bb.2(0x40000000), %bb.1(0x40000000) + ; GCN-NEXT: liveins: $vgpr0_vgpr1, $sgpr0_sgpr1, $scc + ; GCN-NEXT: {{ $}} + ; GCN-NEXT: $sgpr2 = S_LOAD_DWORD_IMM $sgpr0_sgpr1, 0, 0 + ; GCN-NEXT: S_CBRANCH_SCC1 %bb.2, implicit $scc + ; GCN-NEXT: {{ $}} + ; GCN-NEXT: bb.1: + ; GCN-NEXT: successors: %bb.2(0x80000000) + ; GCN-NEXT: liveins: $vgpr0_vgpr1, $sgpr2 + ; GCN-NEXT: {{ $}} + ; GCN-NEXT: $vgpr2 = GLOBAL_LOAD_DWORD $vgpr0_vgpr1, 0, 0, implicit $exec + ; GCN-NEXT: {{ $}} + ; GCN-NEXT: bb.2: + ; GCN-NEXT: liveins: $sgpr2 + ; GCN-NEXT: {{ $}} + ; GCN-NEXT: S_WAIT_KMCNT 0 + ; GCN-NEXT: $sgpr2 = S_MOV_B32 $sgpr2 + ; GCN-NEXT: S_WAIT_LOADCNT 0 + ; GCN-NEXT: $vgpr2 = V_MOV_B32_e32 0, implicit $exec + bb.0: + liveins: $vgpr0_vgpr1, $sgpr0_sgpr1, $scc + $sgpr2 = S_LOAD_DWORD_IMM $sgpr0_sgpr1, 0, 0 + S_CBRANCH_SCC1 %bb.2, implicit $scc + bb.1: + liveins: $vgpr0_vgpr1, $sgpr2 + $vgpr2 = GLOBAL_LOAD_DWORD $vgpr0_vgpr1, 0, 0, implicit $exec + bb.2: + liveins: $sgpr2 + $sgpr2 = S_MOV_B32 $sgpr2 + $vgpr2 = V_MOV_B32_e32 0, implicit $exec +... + +--- +name: implicit_handling_of_pending_vmem_group +tracksRegLiveness: true +machineFunctionInfo: + isEntryFunction: true +body: | + ; GCN-LABEL: name: implicit_handling_of_pending_vmem_group + ; GCN: bb.0: + ; GCN-NEXT: successors: %bb.2(0x40000000), %bb.1(0x40000000) + ; GCN-NEXT: liveins: $vgpr0_vgpr1, $sgpr0_sgpr1, $scc + ; GCN-NEXT: {{ $}} + ; GCN-NEXT: $sgpr2 = S_LOAD_DWORD_IMM $sgpr0_sgpr1, 0, 0 + ; GCN-NEXT: S_CBRANCH_SCC1 %bb.2, implicit $scc + ; GCN-NEXT: {{ $}} + ; GCN-NEXT: bb.1: + ; GCN-NEXT: successors: %bb.2(0x80000000) + ; GCN-NEXT: liveins: $vgpr0_vgpr1, $sgpr2 + ; GCN-NEXT: {{ $}} + ; GCN-NEXT: $vgpr2 = GLOBAL_LOAD_DWORD $vgpr0_vgpr1, 0, 0, implicit $exec + ; GCN-NEXT: {{ $}} + ; GCN-NEXT: bb.2: + ; GCN-NEXT: liveins: $sgpr0_sgpr1, $sgpr2 + ; GCN-NEXT: {{ $}} + ; GCN-NEXT: S_WAIT_KMCNT 0 + ; GCN-NEXT: $sgpr2 = S_MOV_B32 $sgpr2 + ; GCN-NEXT: $sgpr2 = S_LOAD_DWORD_IMM $sgpr0_sgpr1, 0, 0 + ; GCN-NEXT: $vgpr0 = V_MOV_B32_e32 0, implicit $exec + ; GCN-NEXT: S_WAIT_XCNT 0 + ; GCN-NEXT: $sgpr0 = S_MOV_B32 $sgpr0 + bb.0: + liveins: $vgpr0_vgpr1, $sgpr0_sgpr1, $scc + $sgpr2 = S_LOAD_DWORD_IMM $sgpr0_sgpr1, 0, 0 + S_CBRANCH_SCC1 %bb.2, implicit $scc + bb.1: + liveins: $vgpr0_vgpr1, $sgpr2 + $vgpr2 = GLOBAL_LOAD_DWORD $vgpr0_vgpr1, 0, 0, implicit $exec + bb.2: + liveins: $sgpr0_sgpr1, $sgpr2 + $sgpr2 = S_MOV_B32 $sgpr2 + $sgpr2 = S_LOAD_DWORD_IMM $sgpr0_sgpr1, 0, 0 + $vgpr0 = V_MOV_B32_e32 0, implicit $exec + $sgpr0 = S_MOV_B32 $sgpr0 +... + +--- +name: pending_vmem_event_between_block +tracksRegLiveness: true +machineFunctionInfo: + isEntryFunction: true +body: | + ; GCN-LABEL: name: pending_vmem_event_between_block + ; GCN: bb.0: + ; GCN-NEXT: successors: %bb.2(0x40000000), %bb.1(0x40000000) + ; GCN-NEXT: liveins: $vgpr0_vgpr1, $sgpr0_sgpr1, $scc + ; GCN-NEXT: {{ $}} + ; GCN-NEXT: $sgpr2 = S_LOAD_DWORD_IMM $sgpr0_sgpr1, 0, 0 + ; GCN-NEXT: S_CBRANCH_SCC1 %bb.2, implicit $scc + ; GCN-NEXT: {{ $}} + ; GCN-NEXT: bb.1: + ; GCN-NEXT: successors: %bb.2(0x80000000) + ; GCN-NEXT: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3, $sgpr2 + ; GCN-NEXT: {{ $}} + ; GCN-NEXT: $vgpr4 = GLOBAL_LOAD_DWORD $vgpr0_vgpr1, 0, 0, implicit $exec + ; GCN-NEXT: $vgpr5 = GLOBAL_LOAD_DWORD $vgpr2_vgpr3, 0, 0, implicit $exec + ; GCN-NEXT: {{ $}} + ; GCN-NEXT: bb.2: + ; GCN-NEXT: liveins: $sgpr0_sgpr1, $sgpr2, $vgpr2 + ; GCN-NEXT: {{ $}} + ; GCN-NEXT: S_WAIT_KMCNT 0 + ; GCN-NEXT: $sgpr2 = S_MOV_B32 $sgpr2 + ; GCN-NEXT: S_WAIT_XCNT 1 + ; GCN-NEXT: $vgpr1 = V_MOV_B32_e32 0, implicit $exec + ; GCN-NEXT: S_WAIT_XCNT 0 + ; GCN-NEXT: $vgpr2 = V_MOV_B32_e32 0, implicit $exec + ; GCN-NEXT: $sgpr0 = S_MOV_B32 $sgpr0 + bb.0: + liveins: $vgpr0_vgpr1, $sgpr0_sgpr1, $scc + $sgpr2 = S_LOAD_DWORD_IMM $sgpr0_sgpr1, 0, 0 + S_CBRANCH_SCC1 %bb.2, implicit $scc + bb.1: + liveins: $vgpr0_vgpr1, $vgpr2_vgpr3, $sgpr2 + $vgpr4 = GLOBAL_LOAD_DWORD $vgpr0_vgpr1, 0, 0, implicit $exec + $vgpr5 = GLOBAL_LOAD_DWORD $vgpr2_vgpr3, 0, 0, implicit $exec + bb.2: + liveins: $sgpr0_sgpr1, $sgpr2, $vgpr2 + $sgpr2 = S_MOV_B32 $sgpr2 + $vgpr1 = V_MOV_B32_e32 0, implicit $exec + $vgpr2 = V_MOV_B32_e32 0, implicit $exec + $sgpr0 = S_MOV_B32 $sgpr0 +... + +--- +name: flushing_vmem_cnt_on_block_entry +tracksRegLiveness: true +machineFunctionInfo: + isEntryFunction: true +body: | + ; GCN-LABEL: name: flushing_vmem_cnt_on_block_entry + ; GCN: bb.0: + ; GCN-NEXT: successors: %bb.2(0x40000000), %bb.1(0x40000000) + ; GCN-NEXT: liveins: $vgpr0_vgpr1, $sgpr0_sgpr1, $scc + ; GCN-NEXT: {{ $}} + ; GCN-NEXT: $sgpr2 = S_LOAD_DWORD_IMM $sgpr0_sgpr1, 0, 0 + ; GCN-NEXT: S_CBRANCH_SCC1 %bb.2, implicit $scc + ; GCN-NEXT: {{ $}} + ; GCN-NEXT: bb.1: + ; GCN-NEXT: successors: %bb.2(0x80000000) + ; GCN-NEXT: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3, $sgpr2 + ; GCN-NEXT: {{ $}} + ; GCN-NEXT: $vgpr4 = GLOBAL_LOAD_DWORD $vgpr0_vgpr1, 0, 0, implicit $exec + ; GCN-NEXT: $vgpr5 = GLOBAL_LOAD_DWORD $vgpr2_vgpr3, 0, 0, implicit $exec + ; GCN-NEXT: {{ $}} + ; GCN-NEXT: bb.2: + ; GCN-NEXT: liveins: $sgpr0_sgpr1, $sgpr2, $vgpr2 + ; GCN-NEXT: {{ $}} + ; GCN-NEXT: S_WAIT_XCNT 0 + ; GCN-NEXT: $vgpr1 = V_MOV_B32_e32 0, implicit $exec + ; GCN-NEXT: $vgpr2 = V_MOV_B32_e32 0, implicit $exec + ; GCN-NEXT: $sgpr0 = S_MOV_B32 $sgpr0 + bb.0: + liveins: $vgpr0_vgpr1, $sgpr0_sgpr1, $scc + $sgpr2 = S_LOAD_DWORD_IMM $sgpr0_sgpr1, 0, 0 + S_CBRANCH_SCC1 %bb.2, implicit $scc + bb.1: + liveins: $vgpr0_vgpr1, $vgpr2_vgpr3, $sgpr2 + $vgpr4 = GLOBAL_LOAD_DWORD $vgpr0_vgpr1, 0, 0, implicit $exec + $vgpr5 = GLOBAL_LOAD_DWORD $vgpr2_vgpr3, 0, 0, implicit $exec + bb.2: + liveins: $sgpr0_sgpr1, $sgpr2, $vgpr2 + $vgpr1 = V_MOV_B32_e32 0, implicit $exec + $vgpr2 = V_MOV_B32_e32 0, implicit $exec + $sgpr0 = S_MOV_B32 $sgpr0 +... + +--- name: wait_loadcnt_with_outstanding_smem tracksRegLiveness: true machineFunctionInfo: diff --git a/llvm/test/CodeGen/LoongArch/lasx/ctpop-ctlz.ll b/llvm/test/CodeGen/LoongArch/lasx/ctpop-ctlz.ll index ba2118f..b3155c9 100644 --- a/llvm/test/CodeGen/LoongArch/lasx/ctpop-ctlz.ll +++ b/llvm/test/CodeGen/LoongArch/lasx/ctpop-ctlz.ll @@ -106,6 +106,69 @@ define void @ctlz_v4i64(ptr %src, ptr %dst) nounwind { ret void } +define void @not_ctlz_v32i8(ptr %src, ptr %dst) nounwind { +; CHECK-LABEL: not_ctlz_v32i8: +; CHECK: # %bb.0: +; CHECK-NEXT: xvld $xr0, $a0, 0 +; CHECK-NEXT: xvxori.b $xr0, $xr0, 255 +; CHECK-NEXT: xvclz.b $xr0, $xr0 +; CHECK-NEXT: xvst $xr0, $a1, 0 +; CHECK-NEXT: ret + %v = load <32 x i8>, ptr %src + %neg = xor <32 x i8> %v, <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1> + %res = call <32 x i8> @llvm.ctlz.v32i8(<32 x i8> %neg, i1 false) + store <32 x i8> %res, ptr %dst + ret void +} + +define void @not_ctlz_v16i16(ptr %src, ptr %dst) nounwind { +; CHECK-LABEL: not_ctlz_v16i16: +; CHECK: # %bb.0: +; CHECK-NEXT: xvld $xr0, $a0, 0 +; CHECK-NEXT: xvrepli.b $xr1, -1 +; CHECK-NEXT: xvxor.v $xr0, $xr0, $xr1 +; CHECK-NEXT: xvclz.h $xr0, $xr0 +; CHECK-NEXT: xvst $xr0, $a1, 0 +; CHECK-NEXT: ret + %v = load <16 x i16>, ptr %src + %neg = xor <16 x i16> %v, <i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1> + %res = call <16 x i16> @llvm.ctlz.v16i16(<16 x i16> %neg, i1 false) + store <16 x i16> %res, ptr %dst + ret void +} + +define void @not_ctlz_v8i32(ptr %src, ptr %dst) nounwind { +; CHECK-LABEL: not_ctlz_v8i32: +; CHECK: # %bb.0: +; CHECK-NEXT: xvld $xr0, $a0, 0 +; CHECK-NEXT: xvrepli.b $xr1, -1 +; CHECK-NEXT: xvxor.v $xr0, $xr0, $xr1 +; CHECK-NEXT: xvclz.w $xr0, $xr0 +; CHECK-NEXT: xvst $xr0, $a1, 0 +; CHECK-NEXT: ret + %v = load <8 x i32>, ptr %src + %neg = xor <8 x i32> %v, <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1> + %res = call <8 x i32> @llvm.ctlz.v8i32(<8 x i32> %neg, i1 false) + store <8 x i32> %res, ptr %dst + ret void +} + +define void @not_ctlz_v4i64(ptr %src, ptr %dst) nounwind { +; CHECK-LABEL: not_ctlz_v4i64: +; CHECK: # %bb.0: +; CHECK-NEXT: xvld $xr0, $a0, 0 +; CHECK-NEXT: xvrepli.b $xr1, -1 +; CHECK-NEXT: xvxor.v $xr0, $xr0, $xr1 +; CHECK-NEXT: xvclz.d $xr0, $xr0 +; CHECK-NEXT: xvst $xr0, $a1, 0 +; CHECK-NEXT: ret + %v = load <4 x i64>, ptr %src + %neg = xor <4 x i64> %v, <i64 -1, i64 -1, i64 -1, i64 -1> + %res = call <4 x i64> @llvm.ctlz.v4i64(<4 x i64> %neg, i1 false) + store <4 x i64> %res, ptr %dst + ret void +} + declare <32 x i8> @llvm.ctpop.v32i8(<32 x i8>) declare <16 x i16> @llvm.ctpop.v16i16(<16 x i16>) declare <8 x i32> @llvm.ctpop.v8i32(<8 x i32>) diff --git a/llvm/test/CodeGen/LoongArch/lasx/fp-max-min.ll b/llvm/test/CodeGen/LoongArch/lasx/fp-max-min.ll index 48ec98c..8e08e1e 100644 --- a/llvm/test/CodeGen/LoongArch/lasx/fp-max-min.ll +++ b/llvm/test/CodeGen/LoongArch/lasx/fp-max-min.ll @@ -5,40 +5,10 @@ define void @minnum_v8f32(ptr %res, ptr %x, ptr %y) nounwind { ; CHECK-LABEL: minnum_v8f32: ; CHECK: # %bb.0: # %entry -; CHECK-NEXT: xvld $xr0, $a2, 0 -; CHECK-NEXT: xvld $xr1, $a1, 0 -; CHECK-NEXT: xvpickve.w $xr2, $xr0, 5 -; CHECK-NEXT: xvpickve.w $xr3, $xr1, 5 -; CHECK-NEXT: fmin.s $fa2, $fa3, $fa2 -; CHECK-NEXT: xvpickve.w $xr3, $xr0, 4 -; CHECK-NEXT: xvpickve.w $xr4, $xr1, 4 -; CHECK-NEXT: fmin.s $fa3, $fa4, $fa3 -; CHECK-NEXT: vextrins.w $vr3, $vr2, 16 -; CHECK-NEXT: xvpickve.w $xr2, $xr0, 6 -; CHECK-NEXT: xvpickve.w $xr4, $xr1, 6 -; CHECK-NEXT: fmin.s $fa2, $fa4, $fa2 -; CHECK-NEXT: vextrins.w $vr3, $vr2, 32 -; CHECK-NEXT: xvpickve.w $xr2, $xr0, 7 -; CHECK-NEXT: xvpickve.w $xr4, $xr1, 7 -; CHECK-NEXT: fmin.s $fa2, $fa4, $fa2 -; CHECK-NEXT: vextrins.w $vr3, $vr2, 48 -; CHECK-NEXT: xvpickve.w $xr2, $xr0, 1 -; CHECK-NEXT: xvpickve.w $xr4, $xr1, 1 -; CHECK-NEXT: fmin.s $fa2, $fa4, $fa2 -; CHECK-NEXT: xvpickve.w $xr4, $xr0, 0 -; CHECK-NEXT: xvpickve.w $xr5, $xr1, 0 -; CHECK-NEXT: fmin.s $fa4, $fa5, $fa4 -; CHECK-NEXT: vextrins.w $vr4, $vr2, 16 -; CHECK-NEXT: xvpickve.w $xr2, $xr0, 2 -; CHECK-NEXT: xvpickve.w $xr5, $xr1, 2 -; CHECK-NEXT: fmin.s $fa2, $fa5, $fa2 -; CHECK-NEXT: vextrins.w $vr4, $vr2, 32 -; CHECK-NEXT: xvpickve.w $xr0, $xr0, 3 -; CHECK-NEXT: xvpickve.w $xr1, $xr1, 3 -; CHECK-NEXT: fmin.s $fa0, $fa1, $fa0 -; CHECK-NEXT: vextrins.w $vr4, $vr0, 48 -; CHECK-NEXT: xvpermi.q $xr4, $xr3, 2 -; CHECK-NEXT: xvst $xr4, $a0, 0 +; CHECK-NEXT: xvld $xr0, $a1, 0 +; CHECK-NEXT: xvld $xr1, $a2, 0 +; CHECK-NEXT: xvfmin.s $xr0, $xr0, $xr1 +; CHECK-NEXT: xvst $xr0, $a0, 0 ; CHECK-NEXT: ret entry: %v0 = load <8 x float>, ptr %x @@ -51,23 +21,9 @@ entry: define void @minnum_v4f64(ptr %res, ptr %x, ptr %y) nounwind { ; CHECK-LABEL: minnum_v4f64: ; CHECK: # %bb.0: # %entry -; CHECK-NEXT: xvld $xr0, $a2, 0 -; CHECK-NEXT: xvld $xr1, $a1, 0 -; CHECK-NEXT: xvpickve.d $xr2, $xr0, 3 -; CHECK-NEXT: xvpickve.d $xr3, $xr1, 3 -; CHECK-NEXT: fmin.d $fa2, $fa3, $fa2 -; CHECK-NEXT: xvpickve.d $xr3, $xr0, 2 -; CHECK-NEXT: xvpickve.d $xr4, $xr1, 2 -; CHECK-NEXT: fmin.d $fa3, $fa4, $fa3 -; CHECK-NEXT: vextrins.d $vr3, $vr2, 16 -; CHECK-NEXT: xvpickve.d $xr2, $xr0, 1 -; CHECK-NEXT: xvpickve.d $xr4, $xr1, 1 -; CHECK-NEXT: fmin.d $fa2, $fa4, $fa2 -; CHECK-NEXT: xvpickve.d $xr0, $xr0, 0 -; CHECK-NEXT: xvpickve.d $xr1, $xr1, 0 -; CHECK-NEXT: fmin.d $fa0, $fa1, $fa0 -; CHECK-NEXT: vextrins.d $vr0, $vr2, 16 -; CHECK-NEXT: xvpermi.q $xr0, $xr3, 2 +; CHECK-NEXT: xvld $xr0, $a1, 0 +; CHECK-NEXT: xvld $xr1, $a2, 0 +; CHECK-NEXT: xvfmin.d $xr0, $xr0, $xr1 ; CHECK-NEXT: xvst $xr0, $a0, 0 ; CHECK-NEXT: ret entry: @@ -81,40 +37,10 @@ entry: define void @maxnum_v8f32(ptr %res, ptr %x, ptr %y) nounwind { ; CHECK-LABEL: maxnum_v8f32: ; CHECK: # %bb.0: # %entry -; CHECK-NEXT: xvld $xr0, $a2, 0 -; CHECK-NEXT: xvld $xr1, $a1, 0 -; CHECK-NEXT: xvpickve.w $xr2, $xr0, 5 -; CHECK-NEXT: xvpickve.w $xr3, $xr1, 5 -; CHECK-NEXT: fmax.s $fa2, $fa3, $fa2 -; CHECK-NEXT: xvpickve.w $xr3, $xr0, 4 -; CHECK-NEXT: xvpickve.w $xr4, $xr1, 4 -; CHECK-NEXT: fmax.s $fa3, $fa4, $fa3 -; CHECK-NEXT: vextrins.w $vr3, $vr2, 16 -; CHECK-NEXT: xvpickve.w $xr2, $xr0, 6 -; CHECK-NEXT: xvpickve.w $xr4, $xr1, 6 -; CHECK-NEXT: fmax.s $fa2, $fa4, $fa2 -; CHECK-NEXT: vextrins.w $vr3, $vr2, 32 -; CHECK-NEXT: xvpickve.w $xr2, $xr0, 7 -; CHECK-NEXT: xvpickve.w $xr4, $xr1, 7 -; CHECK-NEXT: fmax.s $fa2, $fa4, $fa2 -; CHECK-NEXT: vextrins.w $vr3, $vr2, 48 -; CHECK-NEXT: xvpickve.w $xr2, $xr0, 1 -; CHECK-NEXT: xvpickve.w $xr4, $xr1, 1 -; CHECK-NEXT: fmax.s $fa2, $fa4, $fa2 -; CHECK-NEXT: xvpickve.w $xr4, $xr0, 0 -; CHECK-NEXT: xvpickve.w $xr5, $xr1, 0 -; CHECK-NEXT: fmax.s $fa4, $fa5, $fa4 -; CHECK-NEXT: vextrins.w $vr4, $vr2, 16 -; CHECK-NEXT: xvpickve.w $xr2, $xr0, 2 -; CHECK-NEXT: xvpickve.w $xr5, $xr1, 2 -; CHECK-NEXT: fmax.s $fa2, $fa5, $fa2 -; CHECK-NEXT: vextrins.w $vr4, $vr2, 32 -; CHECK-NEXT: xvpickve.w $xr0, $xr0, 3 -; CHECK-NEXT: xvpickve.w $xr1, $xr1, 3 -; CHECK-NEXT: fmax.s $fa0, $fa1, $fa0 -; CHECK-NEXT: vextrins.w $vr4, $vr0, 48 -; CHECK-NEXT: xvpermi.q $xr4, $xr3, 2 -; CHECK-NEXT: xvst $xr4, $a0, 0 +; CHECK-NEXT: xvld $xr0, $a1, 0 +; CHECK-NEXT: xvld $xr1, $a2, 0 +; CHECK-NEXT: xvfmax.s $xr0, $xr0, $xr1 +; CHECK-NEXT: xvst $xr0, $a0, 0 ; CHECK-NEXT: ret entry: %v0 = load <8 x float>, ptr %x @@ -127,23 +53,9 @@ entry: define void @maxnum_v4f64(ptr %res, ptr %x, ptr %y) nounwind { ; CHECK-LABEL: maxnum_v4f64: ; CHECK: # %bb.0: # %entry -; CHECK-NEXT: xvld $xr0, $a2, 0 -; CHECK-NEXT: xvld $xr1, $a1, 0 -; CHECK-NEXT: xvpickve.d $xr2, $xr0, 3 -; CHECK-NEXT: xvpickve.d $xr3, $xr1, 3 -; CHECK-NEXT: fmax.d $fa2, $fa3, $fa2 -; CHECK-NEXT: xvpickve.d $xr3, $xr0, 2 -; CHECK-NEXT: xvpickve.d $xr4, $xr1, 2 -; CHECK-NEXT: fmax.d $fa3, $fa4, $fa3 -; CHECK-NEXT: vextrins.d $vr3, $vr2, 16 -; CHECK-NEXT: xvpickve.d $xr2, $xr0, 1 -; CHECK-NEXT: xvpickve.d $xr4, $xr1, 1 -; CHECK-NEXT: fmax.d $fa2, $fa4, $fa2 -; CHECK-NEXT: xvpickve.d $xr0, $xr0, 0 -; CHECK-NEXT: xvpickve.d $xr1, $xr1, 0 -; CHECK-NEXT: fmax.d $fa0, $fa1, $fa0 -; CHECK-NEXT: vextrins.d $vr0, $vr2, 16 -; CHECK-NEXT: xvpermi.q $xr0, $xr3, 2 +; CHECK-NEXT: xvld $xr0, $a1, 0 +; CHECK-NEXT: xvld $xr1, $a2, 0 +; CHECK-NEXT: xvfmax.d $xr0, $xr0, $xr1 ; CHECK-NEXT: xvst $xr0, $a0, 0 ; CHECK-NEXT: ret entry: diff --git a/llvm/test/CodeGen/LoongArch/lsx/ctpop-ctlz.ll b/llvm/test/CodeGen/LoongArch/lsx/ctpop-ctlz.ll index a9a38e8..6ac7d51 100644 --- a/llvm/test/CodeGen/LoongArch/lsx/ctpop-ctlz.ll +++ b/llvm/test/CodeGen/LoongArch/lsx/ctpop-ctlz.ll @@ -106,6 +106,69 @@ define void @ctlz_v2i64(ptr %src, ptr %dst) nounwind { ret void } +define void @not_ctlz_v16i8(ptr %src, ptr %dst) nounwind { +; CHECK-LABEL: not_ctlz_v16i8: +; CHECK: # %bb.0: +; CHECK-NEXT: vld $vr0, $a0, 0 +; CHECK-NEXT: vxori.b $vr0, $vr0, 255 +; CHECK-NEXT: vclz.b $vr0, $vr0 +; CHECK-NEXT: vst $vr0, $a1, 0 +; CHECK-NEXT: ret + %v = load <16 x i8>, ptr %src + %neg = xor <16 x i8> %v, <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1> + %res = call <16 x i8> @llvm.ctlz.v16i8(<16 x i8> %neg, i1 false) + store <16 x i8> %res, ptr %dst + ret void +} + +define void @not_ctlz_v8i16(ptr %src, ptr %dst) nounwind { +; CHECK-LABEL: not_ctlz_v8i16: +; CHECK: # %bb.0: +; CHECK-NEXT: vld $vr0, $a0, 0 +; CHECK-NEXT: vrepli.b $vr1, -1 +; CHECK-NEXT: vxor.v $vr0, $vr0, $vr1 +; CHECK-NEXT: vclz.h $vr0, $vr0 +; CHECK-NEXT: vst $vr0, $a1, 0 +; CHECK-NEXT: ret + %v = load <8 x i16>, ptr %src + %neg = xor <8 x i16> %v, <i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1> + %res = call <8 x i16> @llvm.ctlz.v8i16(<8 x i16> %neg, i1 false) + store <8 x i16> %res, ptr %dst + ret void +} + +define void @not_ctlz_v4i32(ptr %src, ptr %dst) nounwind { +; CHECK-LABEL: not_ctlz_v4i32: +; CHECK: # %bb.0: +; CHECK-NEXT: vld $vr0, $a0, 0 +; CHECK-NEXT: vrepli.b $vr1, -1 +; CHECK-NEXT: vxor.v $vr0, $vr0, $vr1 +; CHECK-NEXT: vclz.w $vr0, $vr0 +; CHECK-NEXT: vst $vr0, $a1, 0 +; CHECK-NEXT: ret + %v = load <4 x i32>, ptr %src + %neg = xor <4 x i32> %v, <i32 -1, i32 -1, i32 -1, i32 -1> + %res = call <4 x i32> @llvm.ctlz.v4i32(<4 x i32> %neg, i1 false) + store <4 x i32> %res, ptr %dst + ret void +} + +define void @not_ctlz_v2i64(ptr %src, ptr %dst) nounwind { +; CHECK-LABEL: not_ctlz_v2i64: +; CHECK: # %bb.0: +; CHECK-NEXT: vld $vr0, $a0, 0 +; CHECK-NEXT: vrepli.b $vr1, -1 +; CHECK-NEXT: vxor.v $vr0, $vr0, $vr1 +; CHECK-NEXT: vclz.d $vr0, $vr0 +; CHECK-NEXT: vst $vr0, $a1, 0 +; CHECK-NEXT: ret + %v = load <2 x i64>, ptr %src + %neg = xor <2 x i64> %v, <i64 -1, i64 -1> + %res = call <2 x i64> @llvm.ctlz.v2i64(<2 x i64> %neg, i1 false) + store <2 x i64> %res, ptr %dst + ret void +} + declare <16 x i8> @llvm.ctpop.v16i8(<16 x i8>) declare <8 x i16> @llvm.ctpop.v8i16(<8 x i16>) declare <4 x i32> @llvm.ctpop.v4i32(<4 x i32>) diff --git a/llvm/test/CodeGen/LoongArch/lsx/fp-max-min.ll b/llvm/test/CodeGen/LoongArch/lsx/fp-max-min.ll index 27ecb75..c173092 100644 --- a/llvm/test/CodeGen/LoongArch/lsx/fp-max-min.ll +++ b/llvm/test/CodeGen/LoongArch/lsx/fp-max-min.ll @@ -5,24 +5,10 @@ define void @minnum_v4f32(ptr %res, ptr %x, ptr %y) nounwind { ; CHECK-LABEL: minnum_v4f32: ; CHECK: # %bb.0: # %entry -; CHECK-NEXT: vld $vr0, $a2, 0 -; CHECK-NEXT: vld $vr1, $a1, 0 -; CHECK-NEXT: vreplvei.w $vr2, $vr0, 1 -; CHECK-NEXT: vreplvei.w $vr3, $vr1, 1 -; CHECK-NEXT: fmin.s $fa2, $fa3, $fa2 -; CHECK-NEXT: vreplvei.w $vr3, $vr0, 0 -; CHECK-NEXT: vreplvei.w $vr4, $vr1, 0 -; CHECK-NEXT: fmin.s $fa3, $fa4, $fa3 -; CHECK-NEXT: vextrins.w $vr3, $vr2, 16 -; CHECK-NEXT: vreplvei.w $vr2, $vr0, 2 -; CHECK-NEXT: vreplvei.w $vr4, $vr1, 2 -; CHECK-NEXT: fmin.s $fa2, $fa4, $fa2 -; CHECK-NEXT: vextrins.w $vr3, $vr2, 32 -; CHECK-NEXT: vreplvei.w $vr0, $vr0, 3 -; CHECK-NEXT: vreplvei.w $vr1, $vr1, 3 -; CHECK-NEXT: fmin.s $fa0, $fa1, $fa0 -; CHECK-NEXT: vextrins.w $vr3, $vr0, 48 -; CHECK-NEXT: vst $vr3, $a0, 0 +; CHECK-NEXT: vld $vr0, $a1, 0 +; CHECK-NEXT: vld $vr1, $a2, 0 +; CHECK-NEXT: vfmin.s $vr0, $vr0, $vr1 +; CHECK-NEXT: vst $vr0, $a0, 0 ; CHECK-NEXT: ret entry: %v0 = load <4 x float>, ptr %x @@ -35,15 +21,9 @@ entry: define void @minnum_v2f64(ptr %res, ptr %x, ptr %y) nounwind { ; CHECK-LABEL: minnum_v2f64: ; CHECK: # %bb.0: # %entry -; CHECK-NEXT: vld $vr0, $a2, 0 -; CHECK-NEXT: vld $vr1, $a1, 0 -; CHECK-NEXT: vreplvei.d $vr2, $vr0, 1 -; CHECK-NEXT: vreplvei.d $vr3, $vr1, 1 -; CHECK-NEXT: fmin.d $fa2, $fa3, $fa2 -; CHECK-NEXT: vreplvei.d $vr0, $vr0, 0 -; CHECK-NEXT: vreplvei.d $vr1, $vr1, 0 -; CHECK-NEXT: fmin.d $fa0, $fa1, $fa0 -; CHECK-NEXT: vextrins.d $vr0, $vr2, 16 +; CHECK-NEXT: vld $vr0, $a1, 0 +; CHECK-NEXT: vld $vr1, $a2, 0 +; CHECK-NEXT: vfmin.d $vr0, $vr0, $vr1 ; CHECK-NEXT: vst $vr0, $a0, 0 ; CHECK-NEXT: ret entry: @@ -57,24 +37,10 @@ entry: define void @maxnum_v4f32(ptr %res, ptr %x, ptr %y) nounwind { ; CHECK-LABEL: maxnum_v4f32: ; CHECK: # %bb.0: # %entry -; CHECK-NEXT: vld $vr0, $a2, 0 -; CHECK-NEXT: vld $vr1, $a1, 0 -; CHECK-NEXT: vreplvei.w $vr2, $vr0, 1 -; CHECK-NEXT: vreplvei.w $vr3, $vr1, 1 -; CHECK-NEXT: fmax.s $fa2, $fa3, $fa2 -; CHECK-NEXT: vreplvei.w $vr3, $vr0, 0 -; CHECK-NEXT: vreplvei.w $vr4, $vr1, 0 -; CHECK-NEXT: fmax.s $fa3, $fa4, $fa3 -; CHECK-NEXT: vextrins.w $vr3, $vr2, 16 -; CHECK-NEXT: vreplvei.w $vr2, $vr0, 2 -; CHECK-NEXT: vreplvei.w $vr4, $vr1, 2 -; CHECK-NEXT: fmax.s $fa2, $fa4, $fa2 -; CHECK-NEXT: vextrins.w $vr3, $vr2, 32 -; CHECK-NEXT: vreplvei.w $vr0, $vr0, 3 -; CHECK-NEXT: vreplvei.w $vr1, $vr1, 3 -; CHECK-NEXT: fmax.s $fa0, $fa1, $fa0 -; CHECK-NEXT: vextrins.w $vr3, $vr0, 48 -; CHECK-NEXT: vst $vr3, $a0, 0 +; CHECK-NEXT: vld $vr0, $a1, 0 +; CHECK-NEXT: vld $vr1, $a2, 0 +; CHECK-NEXT: vfmax.s $vr0, $vr0, $vr1 +; CHECK-NEXT: vst $vr0, $a0, 0 ; CHECK-NEXT: ret entry: %v0 = load <4 x float>, ptr %x @@ -87,15 +53,9 @@ entry: define void @maxnum_v2f64(ptr %res, ptr %x, ptr %y) nounwind { ; CHECK-LABEL: maxnum_v2f64: ; CHECK: # %bb.0: # %entry -; CHECK-NEXT: vld $vr0, $a2, 0 -; CHECK-NEXT: vld $vr1, $a1, 0 -; CHECK-NEXT: vreplvei.d $vr2, $vr0, 1 -; CHECK-NEXT: vreplvei.d $vr3, $vr1, 1 -; CHECK-NEXT: fmax.d $fa2, $fa3, $fa2 -; CHECK-NEXT: vreplvei.d $vr0, $vr0, 0 -; CHECK-NEXT: vreplvei.d $vr1, $vr1, 0 -; CHECK-NEXT: fmax.d $fa0, $fa1, $fa0 -; CHECK-NEXT: vextrins.d $vr0, $vr2, 16 +; CHECK-NEXT: vld $vr0, $a1, 0 +; CHECK-NEXT: vld $vr1, $a2, 0 +; CHECK-NEXT: vfmax.d $vr0, $vr0, $vr1 ; CHECK-NEXT: vst $vr0, $a0, 0 ; CHECK-NEXT: ret entry: diff --git a/llvm/test/CodeGen/NVPTX/f16-ex2.ll b/llvm/test/CodeGen/NVPTX/f16-ex2.ll index ee79f9d..af3fe67 100644 --- a/llvm/test/CodeGen/NVPTX/f16-ex2.ll +++ b/llvm/test/CodeGen/NVPTX/f16-ex2.ll @@ -1,12 +1,13 @@ ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5 -; RUN: llc < %s -mcpu=sm_75 -mattr=+ptx70 | FileCheck --check-prefixes=CHECK-FP16 %s -; RUN: %if ptxas-sm_75 && ptxas-isa-7.0 %{ llc < %s -mcpu=sm_75 -mattr=+ptx70 | %ptxas-verify -arch=sm_75 %} +; RUN: llc < %s -mcpu=sm_90 -mattr=+ptx78 | FileCheck --check-prefixes=CHECK-FP16 %s +; RUN: %if ptxas-sm_90 && ptxas-isa-7.8 %{ llc < %s -mcpu=sm_90 -mattr=+ptx78 | %ptxas-verify -arch=sm_90 %} target triple = "nvptx64-nvidia-cuda" declare half @llvm.nvvm.ex2.approx.f16(half) -declare <2 x half> @llvm.nvvm.ex2.approx.f16x2(<2 x half>) +declare <2 x half> @llvm.nvvm.ex2.approx.v2f16(<2 x half>) +declare bfloat @llvm.nvvm.ex2.approx.ftz.bf16(bfloat) +declare <2 x bfloat> @llvm.nvvm.ex2.approx.ftz.v2bf16(<2 x bfloat>) -; CHECK-LABEL: ex2_half define half @ex2_half(half %0) { ; CHECK-FP16-LABEL: ex2_half( ; CHECK-FP16: { @@ -21,7 +22,6 @@ define half @ex2_half(half %0) { ret half %res } -; CHECK-LABEL: ex2_2xhalf define <2 x half> @ex2_2xhalf(<2 x half> %0) { ; CHECK-FP16-LABEL: ex2_2xhalf( ; CHECK-FP16: { @@ -32,6 +32,34 @@ define <2 x half> @ex2_2xhalf(<2 x half> %0) { ; CHECK-FP16-NEXT: ex2.approx.f16x2 %r2, %r1; ; CHECK-FP16-NEXT: st.param.b32 [func_retval0], %r2; ; CHECK-FP16-NEXT: ret; - %res = call <2 x half> @llvm.nvvm.ex2.approx.f16x2(<2 x half> %0) + %res = call <2 x half> @llvm.nvvm.ex2.approx.v2f16(<2 x half> %0) ret <2 x half> %res } + +define bfloat @ex2_bfloat(bfloat %0) { +; CHECK-FP16-LABEL: ex2_bfloat( +; CHECK-FP16: { +; CHECK-FP16-NEXT: .reg .b16 %rs<3>; +; CHECK-FP16-EMPTY: +; CHECK-FP16-NEXT: // %bb.0: +; CHECK-FP16-NEXT: ld.param.b16 %rs1, [ex2_bfloat_param_0]; +; CHECK-FP16-NEXT: ex2.approx.ftz.bf16 %rs2, %rs1; +; CHECK-FP16-NEXT: st.param.b16 [func_retval0], %rs2; +; CHECK-FP16-NEXT: ret; + %res = call bfloat @llvm.nvvm.ex2.approx.ftz.bf16(bfloat %0) + ret bfloat %res +} + +define <2 x bfloat> @ex2_2xbfloat(<2 x bfloat> %0) { +; CHECK-FP16-LABEL: ex2_2xbfloat( +; CHECK-FP16: { +; CHECK-FP16-NEXT: .reg .b32 %r<3>; +; CHECK-FP16-EMPTY: +; CHECK-FP16-NEXT: // %bb.0: +; CHECK-FP16-NEXT: ld.param.b32 %r1, [ex2_2xbfloat_param_0]; +; CHECK-FP16-NEXT: ex2.approx.ftz.bf16x2 %r2, %r1; +; CHECK-FP16-NEXT: st.param.b32 [func_retval0], %r2; +; CHECK-FP16-NEXT: ret; + %res = call <2 x bfloat> @llvm.nvvm.ex2.approx.ftz.v2bf16(<2 x bfloat> %0) + ret <2 x bfloat> %res +} diff --git a/llvm/test/CodeGen/NVPTX/f32-ex2.ll b/llvm/test/CodeGen/NVPTX/f32-ex2.ll index 796d80d..97b9d35 100644 --- a/llvm/test/CodeGen/NVPTX/f32-ex2.ll +++ b/llvm/test/CodeGen/NVPTX/f32-ex2.ll @@ -3,7 +3,8 @@ ; RUN: %if ptxas-sm_50 && ptxas-isa-3.2 %{ llc < %s -mtriple=nvptx64 -mcpu=sm_50 -mattr=+ptx32 | %ptxas-verify -arch=sm_50 %} target triple = "nvptx-nvidia-cuda" -declare float @llvm.nvvm.ex2.approx.f(float) +declare float @llvm.nvvm.ex2.approx.f32(float) +declare float @llvm.nvvm.ex2.approx.ftz.f32(float) ; CHECK-LABEL: ex2_float define float @ex2_float(float %0) { @@ -16,7 +17,7 @@ define float @ex2_float(float %0) { ; CHECK-NEXT: ex2.approx.f32 %r2, %r1; ; CHECK-NEXT: st.param.b32 [func_retval0], %r2; ; CHECK-NEXT: ret; - %res = call float @llvm.nvvm.ex2.approx.f(float %0) + %res = call float @llvm.nvvm.ex2.approx.f32(float %0) ret float %res } @@ -31,6 +32,6 @@ define float @ex2_float_ftz(float %0) { ; CHECK-NEXT: ex2.approx.ftz.f32 %r2, %r1; ; CHECK-NEXT: st.param.b32 [func_retval0], %r2; ; CHECK-NEXT: ret; - %res = call float @llvm.nvvm.ex2.approx.ftz.f(float %0) + %res = call float @llvm.nvvm.ex2.approx.ftz.f32(float %0) ret float %res } diff --git a/llvm/test/CodeGen/X86/isel-llvm.sincos.ll b/llvm/test/CodeGen/X86/isel-llvm.sincos.ll index 065710f..8576f8f 100644 --- a/llvm/test/CodeGen/X86/isel-llvm.sincos.ll +++ b/llvm/test/CodeGen/X86/isel-llvm.sincos.ll @@ -3,6 +3,9 @@ ; RUN: llc < %s -mtriple=x86_64-linux-gnu -fast-isel | FileCheck %s --check-prefixes=X64,FASTISEL-X64 ; RUN: llc < %s -mtriple=i686-linux-gnu -global-isel=0 -fast-isel=0 | FileCheck %s --check-prefixes=X86,SDAG-X86 ; RUN: llc < %s -mtriple=x86_64-linux-gnu -global-isel=0 -fast-isel=0 | FileCheck %s --check-prefixes=X64,SDAG-X64 +; RUN: llc < %s -mtriple=x86_64-apple-macosx10.9.0 -mcpu=core2 | FileCheck %s --check-prefix=MACOS-SINCOS-STRET +; RUN: llc < %s -mtriple=x86_64-apple-macosx10.8.0 -mcpu=core2 | FileCheck %s --check-prefix=MACOS-NOSINCOS-STRET + ; TODO: The below RUN line will fails GISEL selection and will fallback to DAG selection due to lack of support for loads/stores in i686 mode, support is expected soon enough, for this reason the llvm/test/CodeGen/X86/GlobalISel/llvm.sincos.mir test is added for now because of the lack of support for i686 in GlobalISel. ; RUN: llc < %s -mtriple=i686-linux-gnu -global-isel=1 -global-isel-abort=2 | FileCheck %s --check-prefixes=GISEL-X86 ; RUN: llc < %s -mtriple=x86_64-linux-gnu -global-isel=1 -global-isel-abort=1 | FileCheck %s --check-prefixes=GISEL-X64 @@ -34,6 +37,29 @@ define { float, float } @test_sincos_f32(float %Val) nounwind { ; X64-NEXT: popq %rax ; X64-NEXT: retq ; +; MACOS-SINCOS-STRET-LABEL: test_sincos_f32: +; MACOS-SINCOS-STRET: ## %bb.0: +; MACOS-SINCOS-STRET-NEXT: pushq %rax +; MACOS-SINCOS-STRET-NEXT: callq ___sincosf_stret +; MACOS-SINCOS-STRET-NEXT: movshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] +; MACOS-SINCOS-STRET-NEXT: popq %rax +; MACOS-SINCOS-STRET-NEXT: retq +; +; MACOS-NOSINCOS-STRET-LABEL: test_sincos_f32: +; MACOS-NOSINCOS-STRET: ## %bb.0: +; MACOS-NOSINCOS-STRET-NEXT: pushq %rax +; MACOS-NOSINCOS-STRET-NEXT: movss %xmm0, (%rsp) ## 4-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: callq _sinf +; MACOS-NOSINCOS-STRET-NEXT: movss %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 4-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movss (%rsp), %xmm0 ## 4-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = mem[0],zero,zero,zero +; MACOS-NOSINCOS-STRET-NEXT: callq _cosf +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, %xmm1 +; MACOS-NOSINCOS-STRET-NEXT: movss {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 4-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = mem[0],zero,zero,zero +; MACOS-NOSINCOS-STRET-NEXT: popq %rax +; MACOS-NOSINCOS-STRET-NEXT: retq +; ; GISEL-X86-LABEL: test_sincos_f32: ; GISEL-X86: # %bb.0: ; GISEL-X86-NEXT: subl $28, %esp @@ -93,6 +119,28 @@ define { double, double } @test_sincos_f64(double %Val) nounwind { ; X64-NEXT: addq $24, %rsp ; X64-NEXT: retq ; +; MACOS-SINCOS-STRET-LABEL: test_sincos_f64: +; MACOS-SINCOS-STRET: ## %bb.0: +; MACOS-SINCOS-STRET-NEXT: pushq %rax +; MACOS-SINCOS-STRET-NEXT: callq ___sincos_stret +; MACOS-SINCOS-STRET-NEXT: popq %rax +; MACOS-SINCOS-STRET-NEXT: retq +; +; MACOS-NOSINCOS-STRET-LABEL: test_sincos_f64: +; MACOS-NOSINCOS-STRET: ## %bb.0: +; MACOS-NOSINCOS-STRET-NEXT: subq $24, %rsp +; MACOS-NOSINCOS-STRET-NEXT: movsd %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 8-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: callq _sin +; MACOS-NOSINCOS-STRET-NEXT: movsd %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 8-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movsd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 8-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = mem[0],zero +; MACOS-NOSINCOS-STRET-NEXT: callq _cos +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, %xmm1 +; MACOS-NOSINCOS-STRET-NEXT: movsd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 8-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = mem[0],zero +; MACOS-NOSINCOS-STRET-NEXT: addq $24, %rsp +; MACOS-NOSINCOS-STRET-NEXT: retq +; ; GISEL-X86-LABEL: test_sincos_f64: ; GISEL-X86: # %bb.0: ; GISEL-X86-NEXT: subl $44, %esp @@ -153,6 +201,40 @@ define { x86_fp80, x86_fp80 } @test_sincos_f80(x86_fp80 %Val) nounwind { ; X64-NEXT: addq $56, %rsp ; X64-NEXT: retq ; +; MACOS-SINCOS-STRET-LABEL: test_sincos_f80: +; MACOS-SINCOS-STRET: ## %bb.0: +; MACOS-SINCOS-STRET-NEXT: subq $40, %rsp +; MACOS-SINCOS-STRET-NEXT: fldt {{[0-9]+}}(%rsp) +; MACOS-SINCOS-STRET-NEXT: fld %st(0) +; MACOS-SINCOS-STRET-NEXT: fstpt {{[-0-9]+}}(%r{{[sb]}}p) ## 10-byte Folded Spill +; MACOS-SINCOS-STRET-NEXT: fstpt (%rsp) +; MACOS-SINCOS-STRET-NEXT: callq _cosl +; MACOS-SINCOS-STRET-NEXT: fstpt {{[-0-9]+}}(%r{{[sb]}}p) ## 10-byte Folded Spill +; MACOS-SINCOS-STRET-NEXT: fldt {{[-0-9]+}}(%r{{[sb]}}p) ## 10-byte Folded Reload +; MACOS-SINCOS-STRET-NEXT: fstpt (%rsp) +; MACOS-SINCOS-STRET-NEXT: callq _sinl +; MACOS-SINCOS-STRET-NEXT: fldt {{[-0-9]+}}(%r{{[sb]}}p) ## 10-byte Folded Reload +; MACOS-SINCOS-STRET-NEXT: fxch %st(1) +; MACOS-SINCOS-STRET-NEXT: addq $40, %rsp +; MACOS-SINCOS-STRET-NEXT: retq +; +; MACOS-NOSINCOS-STRET-LABEL: test_sincos_f80: +; MACOS-NOSINCOS-STRET: ## %bb.0: +; MACOS-NOSINCOS-STRET-NEXT: subq $40, %rsp +; MACOS-NOSINCOS-STRET-NEXT: fldt {{[0-9]+}}(%rsp) +; MACOS-NOSINCOS-STRET-NEXT: fld %st(0) +; MACOS-NOSINCOS-STRET-NEXT: fstpt {{[-0-9]+}}(%r{{[sb]}}p) ## 10-byte Folded Spill +; MACOS-NOSINCOS-STRET-NEXT: fstpt (%rsp) +; MACOS-NOSINCOS-STRET-NEXT: callq _cosl +; MACOS-NOSINCOS-STRET-NEXT: fstpt {{[-0-9]+}}(%r{{[sb]}}p) ## 10-byte Folded Spill +; MACOS-NOSINCOS-STRET-NEXT: fldt {{[-0-9]+}}(%r{{[sb]}}p) ## 10-byte Folded Reload +; MACOS-NOSINCOS-STRET-NEXT: fstpt (%rsp) +; MACOS-NOSINCOS-STRET-NEXT: callq _sinl +; MACOS-NOSINCOS-STRET-NEXT: fldt {{[-0-9]+}}(%r{{[sb]}}p) ## 10-byte Folded Reload +; MACOS-NOSINCOS-STRET-NEXT: fxch %st(1) +; MACOS-NOSINCOS-STRET-NEXT: addq $40, %rsp +; MACOS-NOSINCOS-STRET-NEXT: retq +; ; GISEL-X86-LABEL: test_sincos_f80: ; GISEL-X86: # %bb.0: ; GISEL-X86-NEXT: subl $60, %esp @@ -288,6 +370,57 @@ define void @can_fold_with_call_in_chain(float %x, ptr noalias %a, ptr noalias % ; SDAG-X64-NEXT: popq %r14 ; SDAG-X64-NEXT: retq ; +; MACOS-SINCOS-STRET-LABEL: can_fold_with_call_in_chain: +; MACOS-SINCOS-STRET: ## %bb.0: ## %entry +; MACOS-SINCOS-STRET-NEXT: pushq %r14 +; MACOS-SINCOS-STRET-NEXT: pushq %rbx +; MACOS-SINCOS-STRET-NEXT: subq $40, %rsp +; MACOS-SINCOS-STRET-NEXT: movq %rsi, %rbx +; MACOS-SINCOS-STRET-NEXT: movq %rdi, %r14 +; MACOS-SINCOS-STRET-NEXT: callq ___sincosf_stret +; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, (%rsp) ## 16-byte Spill +; MACOS-SINCOS-STRET-NEXT: movshdup {{.*#+}} xmm0 = xmm0[1,1,3,3] +; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-SINCOS-STRET-NEXT: movq %r14, %rdi +; MACOS-SINCOS-STRET-NEXT: movq %rbx, %rsi +; MACOS-SINCOS-STRET-NEXT: callq _foo +; MACOS-SINCOS-STRET-NEXT: movaps (%rsp), %xmm0 ## 16-byte Reload +; MACOS-SINCOS-STRET-NEXT: movss %xmm0, (%r14) +; MACOS-SINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload +; MACOS-SINCOS-STRET-NEXT: movss %xmm0, (%rbx) +; MACOS-SINCOS-STRET-NEXT: addq $40, %rsp +; MACOS-SINCOS-STRET-NEXT: popq %rbx +; MACOS-SINCOS-STRET-NEXT: popq %r14 +; MACOS-SINCOS-STRET-NEXT: retq +; +; MACOS-NOSINCOS-STRET-LABEL: can_fold_with_call_in_chain: +; MACOS-NOSINCOS-STRET: ## %bb.0: ## %entry +; MACOS-NOSINCOS-STRET-NEXT: pushq %r14 +; MACOS-NOSINCOS-STRET-NEXT: pushq %rbx +; MACOS-NOSINCOS-STRET-NEXT: pushq %rax +; MACOS-NOSINCOS-STRET-NEXT: movq %rsi, %rbx +; MACOS-NOSINCOS-STRET-NEXT: movq %rdi, %r14 +; MACOS-NOSINCOS-STRET-NEXT: movss %xmm0, (%rsp) ## 4-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: callq _sinf +; MACOS-NOSINCOS-STRET-NEXT: movss %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 4-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movss (%rsp), %xmm0 ## 4-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = mem[0],zero,zero,zero +; MACOS-NOSINCOS-STRET-NEXT: callq _cosf +; MACOS-NOSINCOS-STRET-NEXT: movss %xmm0, (%rsp) ## 4-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movq %r14, %rdi +; MACOS-NOSINCOS-STRET-NEXT: movq %rbx, %rsi +; MACOS-NOSINCOS-STRET-NEXT: callq _foo +; MACOS-NOSINCOS-STRET-NEXT: movss {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 4-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = mem[0],zero,zero,zero +; MACOS-NOSINCOS-STRET-NEXT: movss %xmm0, (%r14) +; MACOS-NOSINCOS-STRET-NEXT: movss (%rsp), %xmm0 ## 4-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = mem[0],zero,zero,zero +; MACOS-NOSINCOS-STRET-NEXT: movss %xmm0, (%rbx) +; MACOS-NOSINCOS-STRET-NEXT: addq $8, %rsp +; MACOS-NOSINCOS-STRET-NEXT: popq %rbx +; MACOS-NOSINCOS-STRET-NEXT: popq %r14 +; MACOS-NOSINCOS-STRET-NEXT: retq +; ; GISEL-X86-LABEL: can_fold_with_call_in_chain: ; GISEL-X86: # %bb.0: # %entry ; GISEL-X86-NEXT: pushl %ebx diff --git a/llvm/test/CodeGen/X86/llvm.sincos.vec.ll b/llvm/test/CodeGen/X86/llvm.sincos.vec.ll index 834dd78..9b02438 100644 --- a/llvm/test/CodeGen/X86/llvm.sincos.vec.ll +++ b/llvm/test/CodeGen/X86/llvm.sincos.vec.ll @@ -1,59 +1,213 @@ ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --no_x86_scrub_sp --version 5 -; RUN: llc < %s -mtriple=i386-unknown-linux-gnu | FileCheck %s +; RUN: llc < %s -mtriple=i386-unknown-linux-gnu | FileCheck -check-prefix=X86 %s +; RUN: llc < %s -mtriple=x86_64-unknown-linux-gnu | FileCheck -check-prefix=X64 %s +; RUN: llc < %s -mtriple=x86_64-apple-macosx10.9.0 | FileCheck --check-prefix=MACOS-SINCOS-STRET %s +; RUN: llc < %s -mtriple=x86_64-apple-macosx10.8.0 | FileCheck --check-prefix=MACOS-NOSINCOS-STRET %s define void @test_sincos_v4f32(<4 x float> %x, ptr noalias %out_sin, ptr noalias %out_cos) nounwind { -; CHECK-LABEL: test_sincos_v4f32: -; CHECK: # %bb.0: -; CHECK-NEXT: pushl %edi -; CHECK-NEXT: pushl %esi -; CHECK-NEXT: subl $52, %esp -; CHECK-NEXT: movl 84(%esp), %esi -; CHECK-NEXT: flds 76(%esp) -; CHECK-NEXT: fstps {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill -; CHECK-NEXT: flds 64(%esp) -; CHECK-NEXT: fstps {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill -; CHECK-NEXT: flds 72(%esp) -; CHECK-NEXT: fstps {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill -; CHECK-NEXT: flds 68(%esp) -; CHECK-NEXT: movl 80(%esp), %edi -; CHECK-NEXT: leal 40(%esp), %eax -; CHECK-NEXT: movl %eax, 8(%esp) -; CHECK-NEXT: leal 4(%edi), %eax -; CHECK-NEXT: movl %eax, 4(%esp) -; CHECK-NEXT: fstps (%esp) -; CHECK-NEXT: calll sincosf -; CHECK-NEXT: leal 44(%esp), %eax -; CHECK-NEXT: movl %eax, 8(%esp) -; CHECK-NEXT: leal 8(%edi), %eax -; CHECK-NEXT: movl %eax, 4(%esp) -; CHECK-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload -; CHECK-NEXT: fstps (%esp) -; CHECK-NEXT: calll sincosf -; CHECK-NEXT: leal 36(%esp), %eax -; CHECK-NEXT: movl %eax, 8(%esp) -; CHECK-NEXT: movl %edi, 4(%esp) -; CHECK-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload -; CHECK-NEXT: fstps (%esp) -; CHECK-NEXT: calll sincosf -; CHECK-NEXT: leal 48(%esp), %eax -; CHECK-NEXT: movl %eax, 8(%esp) -; CHECK-NEXT: addl $12, %edi -; CHECK-NEXT: movl %edi, 4(%esp) -; CHECK-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload -; CHECK-NEXT: fstps (%esp) -; CHECK-NEXT: calll sincosf -; CHECK-NEXT: flds 36(%esp) -; CHECK-NEXT: flds 40(%esp) -; CHECK-NEXT: flds 44(%esp) -; CHECK-NEXT: flds 48(%esp) -; CHECK-NEXT: fstps 12(%esi) -; CHECK-NEXT: fstps 8(%esi) -; CHECK-NEXT: fstps 4(%esi) -; CHECK-NEXT: fstps (%esi) -; CHECK-NEXT: addl $52, %esp -; CHECK-NEXT: popl %esi -; CHECK-NEXT: popl %edi -; CHECK-NEXT: retl +; X86-LABEL: test_sincos_v4f32: +; X86: # %bb.0: +; X86-NEXT: pushl %edi +; X86-NEXT: pushl %esi +; X86-NEXT: subl $52, %esp +; X86-NEXT: movl 84(%esp), %esi +; X86-NEXT: flds 76(%esp) +; X86-NEXT: fstps {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill +; X86-NEXT: flds 64(%esp) +; X86-NEXT: fstps {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill +; X86-NEXT: flds 72(%esp) +; X86-NEXT: fstps {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill +; X86-NEXT: flds 68(%esp) +; X86-NEXT: movl 80(%esp), %edi +; X86-NEXT: leal 40(%esp), %eax +; X86-NEXT: movl %eax, 8(%esp) +; X86-NEXT: leal 4(%edi), %eax +; X86-NEXT: movl %eax, 4(%esp) +; X86-NEXT: fstps (%esp) +; X86-NEXT: calll sincosf +; X86-NEXT: leal 44(%esp), %eax +; X86-NEXT: movl %eax, 8(%esp) +; X86-NEXT: leal 8(%edi), %eax +; X86-NEXT: movl %eax, 4(%esp) +; X86-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload +; X86-NEXT: fstps (%esp) +; X86-NEXT: calll sincosf +; X86-NEXT: leal 36(%esp), %eax +; X86-NEXT: movl %eax, 8(%esp) +; X86-NEXT: movl %edi, 4(%esp) +; X86-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload +; X86-NEXT: fstps (%esp) +; X86-NEXT: calll sincosf +; X86-NEXT: leal 48(%esp), %eax +; X86-NEXT: movl %eax, 8(%esp) +; X86-NEXT: addl $12, %edi +; X86-NEXT: movl %edi, 4(%esp) +; X86-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload +; X86-NEXT: fstps (%esp) +; X86-NEXT: calll sincosf +; X86-NEXT: flds 36(%esp) +; X86-NEXT: flds 40(%esp) +; X86-NEXT: flds 44(%esp) +; X86-NEXT: flds 48(%esp) +; X86-NEXT: fstps 12(%esi) +; X86-NEXT: fstps 8(%esi) +; X86-NEXT: fstps 4(%esi) +; X86-NEXT: fstps (%esi) +; X86-NEXT: addl $52, %esp +; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: retl +; +; X64-LABEL: test_sincos_v4f32: +; X64: # %bb.0: +; X64-NEXT: pushq %r14 +; X64-NEXT: pushq %rbx +; X64-NEXT: subq $56, %rsp +; X64-NEXT: movq %rsi, %rbx +; X64-NEXT: movq %rdi, %r14 +; X64-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill +; X64-NEXT: shufps {{.*#+}} xmm0 = xmm0[3,3,3,3] +; X64-NEXT: leaq 4(%rsp), %rdi +; X64-NEXT: movq %rsp, %rsi +; X64-NEXT: callq sincosf@PLT +; X64-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload +; X64-NEXT: movhlps {{.*#+}} xmm0 = xmm0[1,1] +; X64-NEXT: leaq 12(%rsp), %rdi +; X64-NEXT: leaq 8(%rsp), %rsi +; X64-NEXT: callq sincosf@PLT +; X64-NEXT: leaq 28(%rsp), %rdi +; X64-NEXT: leaq 24(%rsp), %rsi +; X64-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload +; X64-NEXT: callq sincosf@PLT +; X64-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload +; X64-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,1,1,1] +; X64-NEXT: leaq 20(%rsp), %rdi +; X64-NEXT: leaq 16(%rsp), %rsi +; X64-NEXT: callq sincosf@PLT +; X64-NEXT: movss {{.*#+}} xmm0 = mem[0],zero,zero,zero +; X64-NEXT: movss {{.*#+}} xmm1 = mem[0],zero,zero,zero +; X64-NEXT: unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1] +; X64-NEXT: movss {{.*#+}} xmm0 = mem[0],zero,zero,zero +; X64-NEXT: movss {{.*#+}} xmm2 = mem[0],zero,zero,zero +; X64-NEXT: unpcklps {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1] +; X64-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0] +; X64-NEXT: movss {{.*#+}} xmm1 = mem[0],zero,zero,zero +; X64-NEXT: movss {{.*#+}} xmm2 = mem[0],zero,zero,zero +; X64-NEXT: unpcklps {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1] +; X64-NEXT: movss {{.*#+}} xmm1 = mem[0],zero,zero,zero +; X64-NEXT: movss {{.*#+}} xmm3 = mem[0],zero,zero,zero +; X64-NEXT: unpcklps {{.*#+}} xmm1 = xmm1[0],xmm3[0],xmm1[1],xmm3[1] +; X64-NEXT: movlhps {{.*#+}} xmm1 = xmm1[0],xmm2[0] +; X64-NEXT: movups %xmm1, (%r14) +; X64-NEXT: movups %xmm0, (%rbx) +; X64-NEXT: addq $56, %rsp +; X64-NEXT: popq %rbx +; X64-NEXT: popq %r14 +; X64-NEXT: retq +; +; MACOS-SINCOS-STRET-LABEL: test_sincos_v4f32: +; MACOS-SINCOS-STRET: ## %bb.0: +; MACOS-SINCOS-STRET-NEXT: pushq %r14 +; MACOS-SINCOS-STRET-NEXT: pushq %rbx +; MACOS-SINCOS-STRET-NEXT: subq $104, %rsp +; MACOS-SINCOS-STRET-NEXT: movq %rsi, %rbx +; MACOS-SINCOS-STRET-NEXT: movq %rdi, %r14 +; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-SINCOS-STRET-NEXT: shufps {{.*#+}} xmm0 = xmm0[3,3,3,3] +; MACOS-SINCOS-STRET-NEXT: callq ___sincosf_stret +; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-SINCOS-STRET-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,1,1,1] +; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, (%rsp) ## 16-byte Spill +; MACOS-SINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload +; MACOS-SINCOS-STRET-NEXT: movhlps {{.*#+}} xmm0 = xmm0[1,1] +; MACOS-SINCOS-STRET-NEXT: callq ___sincosf_stret +; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-SINCOS-STRET-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,1,1,1] +; MACOS-SINCOS-STRET-NEXT: unpcklps (%rsp), %xmm0 ## 16-byte Folded Reload +; MACOS-SINCOS-STRET-NEXT: ## xmm0 = xmm0[0],mem[0],xmm0[1],mem[1] +; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, (%rsp) ## 16-byte Spill +; MACOS-SINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload +; MACOS-SINCOS-STRET-NEXT: callq ___sincosf_stret +; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, %xmm1 +; MACOS-SINCOS-STRET-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,1],xmm0[1,1] +; MACOS-SINCOS-STRET-NEXT: movaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-SINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload +; MACOS-SINCOS-STRET-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,1,1,1] +; MACOS-SINCOS-STRET-NEXT: callq ___sincosf_stret +; MACOS-SINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Reload +; MACOS-SINCOS-STRET-NEXT: unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1] +; MACOS-SINCOS-STRET-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,1,1,1] +; MACOS-SINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm2 ## 16-byte Reload +; MACOS-SINCOS-STRET-NEXT: unpcklps {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1] +; MACOS-SINCOS-STRET-NEXT: unpcklpd (%rsp), %xmm2 ## 16-byte Folded Reload +; MACOS-SINCOS-STRET-NEXT: ## xmm2 = xmm2[0],mem[0] +; MACOS-SINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload +; MACOS-SINCOS-STRET-NEXT: unpcklps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Folded Reload +; MACOS-SINCOS-STRET-NEXT: ## xmm0 = xmm0[0],mem[0],xmm0[1],mem[1] +; MACOS-SINCOS-STRET-NEXT: movlhps {{.*#+}} xmm1 = xmm1[0],xmm0[0] +; MACOS-SINCOS-STRET-NEXT: movups %xmm1, (%r14) +; MACOS-SINCOS-STRET-NEXT: movups %xmm2, (%rbx) +; MACOS-SINCOS-STRET-NEXT: addq $104, %rsp +; MACOS-SINCOS-STRET-NEXT: popq %rbx +; MACOS-SINCOS-STRET-NEXT: popq %r14 +; MACOS-SINCOS-STRET-NEXT: retq +; +; MACOS-NOSINCOS-STRET-LABEL: test_sincos_v4f32: +; MACOS-NOSINCOS-STRET: ## %bb.0: +; MACOS-NOSINCOS-STRET-NEXT: pushq %r14 +; MACOS-NOSINCOS-STRET-NEXT: pushq %rbx +; MACOS-NOSINCOS-STRET-NEXT: subq $104, %rsp +; MACOS-NOSINCOS-STRET-NEXT: movq %rsi, %rbx +; MACOS-NOSINCOS-STRET-NEXT: movq %rdi, %r14 +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, (%rsp) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: shufps {{.*#+}} xmm0 = xmm0[3,3,3,3] +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: callq _cosf +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movaps (%rsp), %xmm0 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: movhlps {{.*#+}} xmm0 = xmm0[1,1] +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: callq _cosf +; MACOS-NOSINCOS-STRET-NEXT: unpcklps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Folded Reload +; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = xmm0[0],mem[0],xmm0[1],mem[1] +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movaps (%rsp), %xmm0 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: callq _cosf +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movaps (%rsp), %xmm0 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,1,1,1] +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: callq _cosf +; MACOS-NOSINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1] +; MACOS-NOSINCOS-STRET-NEXT: unpcklpd {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Folded Reload +; MACOS-NOSINCOS-STRET-NEXT: ## xmm1 = xmm1[0],mem[0] +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: callq _sinf +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: callq _sinf +; MACOS-NOSINCOS-STRET-NEXT: unpcklps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Folded Reload +; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = xmm0[0],mem[0],xmm0[1],mem[1] +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movaps (%rsp), %xmm0 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: callq _sinf +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, (%rsp) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: callq _sinf +; MACOS-NOSINCOS-STRET-NEXT: movaps (%rsp), %xmm1 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1] +; MACOS-NOSINCOS-STRET-NEXT: unpcklpd {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Folded Reload +; MACOS-NOSINCOS-STRET-NEXT: ## xmm1 = xmm1[0],mem[0] +; MACOS-NOSINCOS-STRET-NEXT: movups %xmm1, (%r14) +; MACOS-NOSINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: movups %xmm0, (%rbx) +; MACOS-NOSINCOS-STRET-NEXT: addq $104, %rsp +; MACOS-NOSINCOS-STRET-NEXT: popq %rbx +; MACOS-NOSINCOS-STRET-NEXT: popq %r14 +; MACOS-NOSINCOS-STRET-NEXT: retq %result = call { <4 x float>, <4 x float> } @llvm.sincos.v4f32(<4 x float> %x) %result.0 = extractvalue { <4 x float>, <4 x float> } %result, 0 %result.1 = extractvalue { <4 x float>, <4 x float> } %result, 1 @@ -63,36 +217,120 @@ define void @test_sincos_v4f32(<4 x float> %x, ptr noalias %out_sin, ptr noalias } define void @test_sincos_v2f64(<2 x double> %x, ptr noalias %out_sin, ptr noalias %out_cos) nounwind { -; CHECK-LABEL: test_sincos_v2f64: -; CHECK: # %bb.0: -; CHECK-NEXT: pushl %edi -; CHECK-NEXT: pushl %esi -; CHECK-NEXT: subl $52, %esp -; CHECK-NEXT: movl 84(%esp), %esi -; CHECK-NEXT: fldl 72(%esp) -; CHECK-NEXT: fstpl {{[-0-9]+}}(%e{{[sb]}}p) # 8-byte Folded Spill -; CHECK-NEXT: fldl 64(%esp) -; CHECK-NEXT: movl 80(%esp), %edi -; CHECK-NEXT: leal 24(%esp), %eax -; CHECK-NEXT: movl %eax, 12(%esp) -; CHECK-NEXT: movl %edi, 8(%esp) -; CHECK-NEXT: fstpl (%esp) -; CHECK-NEXT: calll sincos -; CHECK-NEXT: leal 32(%esp), %eax -; CHECK-NEXT: movl %eax, 12(%esp) -; CHECK-NEXT: addl $8, %edi -; CHECK-NEXT: movl %edi, 8(%esp) -; CHECK-NEXT: fldl {{[-0-9]+}}(%e{{[sb]}}p) # 8-byte Folded Reload -; CHECK-NEXT: fstpl (%esp) -; CHECK-NEXT: calll sincos -; CHECK-NEXT: fldl 24(%esp) -; CHECK-NEXT: fldl 32(%esp) -; CHECK-NEXT: fstpl 8(%esi) -; CHECK-NEXT: fstpl (%esi) -; CHECK-NEXT: addl $52, %esp -; CHECK-NEXT: popl %esi -; CHECK-NEXT: popl %edi -; CHECK-NEXT: retl +; X86-LABEL: test_sincos_v2f64: +; X86: # %bb.0: +; X86-NEXT: pushl %edi +; X86-NEXT: pushl %esi +; X86-NEXT: subl $52, %esp +; X86-NEXT: movl 84(%esp), %esi +; X86-NEXT: fldl 72(%esp) +; X86-NEXT: fstpl {{[-0-9]+}}(%e{{[sb]}}p) # 8-byte Folded Spill +; X86-NEXT: fldl 64(%esp) +; X86-NEXT: movl 80(%esp), %edi +; X86-NEXT: leal 24(%esp), %eax +; X86-NEXT: movl %eax, 12(%esp) +; X86-NEXT: movl %edi, 8(%esp) +; X86-NEXT: fstpl (%esp) +; X86-NEXT: calll sincos +; X86-NEXT: leal 32(%esp), %eax +; X86-NEXT: movl %eax, 12(%esp) +; X86-NEXT: addl $8, %edi +; X86-NEXT: movl %edi, 8(%esp) +; X86-NEXT: fldl {{[-0-9]+}}(%e{{[sb]}}p) # 8-byte Folded Reload +; X86-NEXT: fstpl (%esp) +; X86-NEXT: calll sincos +; X86-NEXT: fldl 24(%esp) +; X86-NEXT: fldl 32(%esp) +; X86-NEXT: fstpl 8(%esi) +; X86-NEXT: fstpl (%esi) +; X86-NEXT: addl $52, %esp +; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: retl +; +; X64-LABEL: test_sincos_v2f64: +; X64: # %bb.0: +; X64-NEXT: pushq %r14 +; X64-NEXT: pushq %rbx +; X64-NEXT: subq $56, %rsp +; X64-NEXT: movq %rsi, %rbx +; X64-NEXT: movq %rdi, %r14 +; X64-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill +; X64-NEXT: leaq 24(%rsp), %rdi +; X64-NEXT: leaq 16(%rsp), %rsi +; X64-NEXT: callq sincos@PLT +; X64-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload +; X64-NEXT: movhlps {{.*#+}} xmm0 = xmm0[1,1] +; X64-NEXT: leaq 8(%rsp), %rdi +; X64-NEXT: movq %rsp, %rsi +; X64-NEXT: callq sincos@PLT +; X64-NEXT: movsd {{.*#+}} xmm0 = mem[0],zero +; X64-NEXT: movhps {{.*#+}} xmm0 = xmm0[0,1],mem[0,1] +; X64-NEXT: movsd {{.*#+}} xmm1 = mem[0],zero +; X64-NEXT: movhps {{.*#+}} xmm1 = xmm1[0,1],mem[0,1] +; X64-NEXT: movups %xmm1, (%r14) +; X64-NEXT: movups %xmm0, (%rbx) +; X64-NEXT: addq $56, %rsp +; X64-NEXT: popq %rbx +; X64-NEXT: popq %r14 +; X64-NEXT: retq +; +; MACOS-SINCOS-STRET-LABEL: test_sincos_v2f64: +; MACOS-SINCOS-STRET: ## %bb.0: +; MACOS-SINCOS-STRET-NEXT: pushq %r14 +; MACOS-SINCOS-STRET-NEXT: pushq %rbx +; MACOS-SINCOS-STRET-NEXT: subq $56, %rsp +; MACOS-SINCOS-STRET-NEXT: movq %rsi, %rbx +; MACOS-SINCOS-STRET-NEXT: movq %rdi, %r14 +; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, (%rsp) ## 16-byte Spill +; MACOS-SINCOS-STRET-NEXT: callq ___sincos_stret +; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-SINCOS-STRET-NEXT: movaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-SINCOS-STRET-NEXT: movaps (%rsp), %xmm0 ## 16-byte Reload +; MACOS-SINCOS-STRET-NEXT: movhlps {{.*#+}} xmm0 = xmm0[1,1] +; MACOS-SINCOS-STRET-NEXT: callq ___sincos_stret +; MACOS-SINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm2 ## 16-byte Reload +; MACOS-SINCOS-STRET-NEXT: movlhps {{.*#+}} xmm2 = xmm2[0],xmm1[0] +; MACOS-SINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Reload +; MACOS-SINCOS-STRET-NEXT: movlhps {{.*#+}} xmm1 = xmm1[0],xmm0[0] +; MACOS-SINCOS-STRET-NEXT: movups %xmm1, (%r14) +; MACOS-SINCOS-STRET-NEXT: movups %xmm2, (%rbx) +; MACOS-SINCOS-STRET-NEXT: addq $56, %rsp +; MACOS-SINCOS-STRET-NEXT: popq %rbx +; MACOS-SINCOS-STRET-NEXT: popq %r14 +; MACOS-SINCOS-STRET-NEXT: retq +; +; MACOS-NOSINCOS-STRET-LABEL: test_sincos_v2f64: +; MACOS-NOSINCOS-STRET: ## %bb.0: +; MACOS-NOSINCOS-STRET-NEXT: pushq %r14 +; MACOS-NOSINCOS-STRET-NEXT: pushq %rbx +; MACOS-NOSINCOS-STRET-NEXT: subq $56, %rsp +; MACOS-NOSINCOS-STRET-NEXT: movq %rsi, %rbx +; MACOS-NOSINCOS-STRET-NEXT: movq %rdi, %r14 +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, (%rsp) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: callq _cos +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movaps (%rsp), %xmm0 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: movhlps {{.*#+}} xmm0 = xmm0[1,1] +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: callq _cos +; MACOS-NOSINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: movlhps {{.*#+}} xmm1 = xmm1[0],xmm0[0] +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movaps (%rsp), %xmm0 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: callq _sin +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, (%rsp) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: callq _sin +; MACOS-NOSINCOS-STRET-NEXT: movaps (%rsp), %xmm1 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: movlhps {{.*#+}} xmm1 = xmm1[0],xmm0[0] +; MACOS-NOSINCOS-STRET-NEXT: movups %xmm1, (%r14) +; MACOS-NOSINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: movups %xmm0, (%rbx) +; MACOS-NOSINCOS-STRET-NEXT: addq $56, %rsp +; MACOS-NOSINCOS-STRET-NEXT: popq %rbx +; MACOS-NOSINCOS-STRET-NEXT: popq %r14 +; MACOS-NOSINCOS-STRET-NEXT: retq %result = call { <2 x double>, <2 x double> } @llvm.sincos.v2f64(<2 x double> %x) %result.0 = extractvalue { <2 x double>, <2 x double> } %result, 0 %result.1 = extractvalue { <2 x double>, <2 x double> } %result, 1 diff --git a/llvm/test/DebugInfo/PDB/Native/pdb-native-index-overflow.test b/llvm/test/DebugInfo/PDB/Native/pdb-native-index-overflow.test new file mode 100755 index 0000000..aa3f6dc --- /dev/null +++ b/llvm/test/DebugInfo/PDB/Native/pdb-native-index-overflow.test @@ -0,0 +1,13 @@ +; Test that the native PDB reader isn't crashed by index value bigger than +; number of types in TPI or IPI stream +; RUN: llvm-pdbutil dump %p/../Inputs/empty.pdb --type-index=20000000\ +; RUN: | FileCheck -check-prefixes=TYPES,NOT_FOUND %s +; RUN: llvm-pdbutil dump %p/../Inputs/empty.pdb --id-index=20000000\ +; RUN: | FileCheck -check-prefixes=IDS,NOT_FOUND %s + +TYPES: Types (TPI Stream) +IDS: Types (IPI Stream) +NOT_FOUND:============================================================ +NOT_FOUND: Showing 1 records. +NOT_FOUND: Type 0x1312D00 doesn't exist in TPI stream + diff --git a/llvm/test/Transforms/IndVarSimplify/loop-guard-order.ll b/llvm/test/Transforms/IndVarSimplify/loop-guard-order.ll index 14ee00d..2763860 100644 --- a/llvm/test/Transforms/IndVarSimplify/loop-guard-order.ll +++ b/llvm/test/Transforms/IndVarSimplify/loop-guard-order.ll @@ -114,7 +114,7 @@ define i32 @urem_order1(i32 %n) { ; CHECK: [[LOOP]]: ; CHECK-NEXT: [[IV:%.*]] = phi i32 [ [[IV_NEXT:%.*]], %[[LOOP]] ], [ 0, %[[LOOP_PREHEADER]] ] ; CHECK-NEXT: call void @foo() -; CHECK-NEXT: [[IV_NEXT]] = add i32 [[IV]], 3 +; CHECK-NEXT: [[IV_NEXT]] = add nuw i32 [[IV]], 3 ; CHECK-NEXT: [[EC:%.*]] = icmp eq i32 [[IV_NEXT]], [[N]] ; CHECK-NEXT: br i1 [[EC]], label %[[EXIT_LOOPEXIT:.*]], label %[[LOOP]] ; CHECK: [[EXIT_LOOPEXIT]]: @@ -205,13 +205,12 @@ define i64 @test_loop_with_div_order_1(i64 %n) { ; CHECK-NEXT: [[PARITY_CHECK:%.*]] = icmp eq i64 [[IS_ODD]], 0 ; CHECK-NEXT: br i1 [[PARITY_CHECK]], label %[[LOOP_PREHEADER:.*]], label %[[EXIT]] ; CHECK: [[LOOP_PREHEADER]]: -; CHECK-NEXT: [[UMAX:%.*]] = call i64 @llvm.umax.i64(i64 [[UPPER_BOUND]], i64 1) ; CHECK-NEXT: br label %[[LOOP:.*]] ; CHECK: [[LOOP]]: ; CHECK-NEXT: [[IV:%.*]] = phi i64 [ [[IV_NEXT:%.*]], %[[LOOP]] ], [ 0, %[[LOOP_PREHEADER]] ] ; CHECK-NEXT: [[DUMMY:%.*]] = load volatile i64, ptr null, align 8 ; CHECK-NEXT: [[IV_NEXT]] = add nuw nsw i64 [[IV]], 1 -; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i64 [[IV_NEXT]], [[UMAX]] +; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i64 [[IV_NEXT]], [[UPPER_BOUND]] ; CHECK-NEXT: br i1 [[EXITCOND]], label %[[LOOP]], label %[[EXIT_LOOPEXIT:.*]] ; CHECK: [[EXIT_LOOPEXIT]]: ; CHECK-NEXT: br label %[[EXIT]] diff --git a/llvm/test/Transforms/InstCombine/or.ll b/llvm/test/Transforms/InstCombine/or.ll index 6b090e9..f61a197 100644 --- a/llvm/test/Transforms/InstCombine/or.ll +++ b/llvm/test/Transforms/InstCombine/or.ll @@ -2113,3 +2113,98 @@ define <4 x i32> @or_zext_nneg_minus_constant_splat(<4 x i8> %a) { %or = or <4 x i32> %zext, splat (i32 -9) ret <4 x i32> %or } + +define i8 @or_positive_minus_non_positive_to_abs(i8 %a){ +; CHECK-LABEL: @or_positive_minus_non_positive_to_abs( +; CHECK-NEXT: [[TMP2:%.*]] = call i8 @llvm.abs.i8(i8 [[A:%.*]], i1 false) +; CHECK-NEXT: ret i8 [[TMP2]] +; + %b = icmp sgt i8 %a, 0 + %mask = sext i1 %b to i8 + %neg = sub i8 0, %a + %mask_inv = xor i8 %mask, -1 + %c = and i8 %neg, %mask_inv + %d = and i8 %a, %mask + %or = or i8 %c, %d + ret i8 %or +} + +; TODO: Fold to smax https://alive2.llvm.org/ce/z/wDiDh2 +define i8 @or_select_smax_neg_to_abs(i8 %a){ +; CHECK-LABEL: @or_select_smax_neg_to_abs( +; CHECK-NEXT: [[SGT0:%.*]] = icmp sgt i8 [[A:%.*]], 0 +; CHECK-NEXT: [[NEG:%.*]] = sub nsw i8 0, [[A]] +; CHECK-NEXT: [[OR:%.*]] = select i1 [[SGT0]], i8 0, i8 [[NEG]] +; CHECK-NEXT: ret i8 [[OR]] +; + %sgt0 = icmp sgt i8 %a, 0 + %neg = sub nsw i8 0, %a + %sel = select i1 %sgt0, i8 0, i8 %neg + ret i8 %sel +} + +; TODO: Fold to abs https://alive2.llvm.org/ce/z/DybfHG +define i8 @or_select_smax_smax_to_abs(i8 %a){ +; CHECK-LABEL: @or_select_smax_smax_to_abs( +; CHECK-NEXT: [[NEG:%.*]] = sub nsw i8 0, [[A:%.*]] +; CHECK-NEXT: [[SEL:%.*]] = call i8 @llvm.smax.i8(i8 [[NEG]], i8 0) +; CHECK-NEXT: [[MAX:%.*]] = call i8 @llvm.smax.i8(i8 [[A]], i8 0) +; CHECK-NEXT: [[OR:%.*]] = or i8 [[SEL]], [[MAX]] +; CHECK-NEXT: ret i8 [[OR]] +; + %neg = sub nsw i8 0, %a + %sel = call i8 @llvm.smax.i8(i8 %neg, i8 0) + %max = call i8 @llvm.smax.i8(i8 %a, i8 0) + %or = or i8 %sel, %max + ret i8 %or +} + +declare i8 @llvm.abs.i8(i8, i1) +declare <2 x i8> @llvm.abs.v2i8(<2 x i8>, i1) + +define <2 x i8> @or_sgt_select_smax_to_abs(<2 x i8> %a){ +; CHECK-LABEL: @or_sgt_select_smax_to_abs( +; CHECK-NEXT: [[OR:%.*]] = call <2 x i8> @llvm.abs.v2i8(<2 x i8> [[A:%.*]], i1 false) +; CHECK-NEXT: ret <2 x i8> [[OR]] +; + %sgt0 = icmp sgt <2 x i8> %a, zeroinitializer + %neg = sub <2 x i8> zeroinitializer, %a + %sel = select <2 x i1> %sgt0, <2 x i8> zeroinitializer, <2 x i8> %neg + %max = call <2 x i8> @llvm.smax.v2i8(<2 x i8> %a, <2 x i8> zeroinitializer) + %or = or <2 x i8> %sel, %max + ret <2 x i8> %or +} + +define <2 x i8> @or_slt_select_smax_to_abs(<2 x i8> %a){ +; CHECK-LABEL: @or_slt_select_smax_to_abs( +; CHECK-NEXT: [[OR:%.*]] = call <2 x i8> @llvm.abs.v2i8(<2 x i8> [[A:%.*]], i1 false) +; CHECK-NEXT: ret <2 x i8> [[OR]] +; + %slt0 = icmp slt <2 x i8> %a, zeroinitializer + %neg = sub <2 x i8> zeroinitializer, %a + %sel = select <2 x i1> %slt0, <2 x i8> %neg, <2 x i8> zeroinitializer + %max = call <2 x i8> @llvm.smax.v2i8(<2 x i8> %a, <2 x i8> zeroinitializer) + %or = or <2 x i8> %sel, %max + ret <2 x i8> %or +} + +; negative test - %d has multiple uses. %or is not folded to abs. + +define <2 x i8> @or_select_smax_multi_uses(<2 x i8> %a){ +; CHECK-LABEL: @or_select_smax_multi_uses( +; CHECK-NEXT: [[B:%.*]] = icmp sgt <2 x i8> [[A:%.*]], zeroinitializer +; CHECK-NEXT: [[NEG:%.*]] = sub <2 x i8> zeroinitializer, [[A]] +; CHECK-NEXT: [[C:%.*]] = select <2 x i1> [[B]], <2 x i8> zeroinitializer, <2 x i8> [[NEG]] +; CHECK-NEXT: [[D:%.*]] = call <2 x i8> @llvm.smax.v2i8(<2 x i8> [[A]], <2 x i8> zeroinitializer) +; CHECK-NEXT: [[OR1:%.*]] = or <2 x i8> [[C]], [[D]] +; CHECK-NEXT: [[OR:%.*]] = add <2 x i8> [[OR1]], [[D]] +; CHECK-NEXT: ret <2 x i8> [[OR]] +; + %sgt0 = icmp sgt <2 x i8> %a, zeroinitializer + %neg = sub <2 x i8> zeroinitializer, %a + %sel = select <2 x i1> %sgt0, <2 x i8> zeroinitializer, <2 x i8> %neg + %max = call <2 x i8> @llvm.smax.v2i8(<2 x i8> %a, <2 x i8> zeroinitializer) + %or = or <2 x i8> %sel, %max + %add = add <2 x i8> %or, %max + ret <2 x i8> %add +} diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/force-target-instruction-cost.ll b/llvm/test/Transforms/LoopVectorize/AArch64/force-target-instruction-cost.ll index bfee39ea..068f82c 100644 --- a/llvm/test/Transforms/LoopVectorize/AArch64/force-target-instruction-cost.ll +++ b/llvm/test/Transforms/LoopVectorize/AArch64/force-target-instruction-cost.ll @@ -365,8 +365,8 @@ define void @invalid_legacy_cost(i64 %N, ptr %x) #0 { ; CHECK: [[VECTOR_BODY]]: ; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ] ; CHECK-NEXT: [[TMP6:%.*]] = alloca i8, i64 0, align 16 -; CHECK-NEXT: [[TMP7:%.*]] = insertelement <2 x ptr> poison, ptr [[TMP6]], i32 0 -; CHECK-NEXT: [[TMP8:%.*]] = insertelement <2 x ptr> [[TMP7]], ptr [[TMP6]], i32 1 +; CHECK-NEXT: [[BROADCAST_SPLATINSERT:%.*]] = insertelement <2 x ptr> poison, ptr [[TMP6]], i64 0 +; CHECK-NEXT: [[TMP8:%.*]] = shufflevector <2 x ptr> [[BROADCAST_SPLATINSERT]], <2 x ptr> poison, <2 x i32> zeroinitializer ; CHECK-NEXT: [[TMP9:%.*]] = getelementptr ptr, ptr [[X]], i64 [[INDEX]] ; CHECK-NEXT: store <2 x ptr> [[TMP8]], ptr [[TMP9]], align 8 ; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 2 diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/licm-calls.ll b/llvm/test/Transforms/LoopVectorize/AArch64/licm-calls.ll index ea01489..0a9494e 100644 --- a/llvm/test/Transforms/LoopVectorize/AArch64/licm-calls.ll +++ b/llvm/test/Transforms/LoopVectorize/AArch64/licm-calls.ll @@ -10,8 +10,8 @@ define void @licm_replicate_call(double %x, ptr %dst) { ; CHECK-NEXT: br label %[[VECTOR_PH:.*]] ; CHECK: [[VECTOR_PH]]: ; CHECK-NEXT: [[TMP1:%.*]] = tail call double @llvm.pow.f64(double [[X]], double 3.000000e+00) -; CHECK-NEXT: [[TMP2:%.*]] = insertelement <2 x double> poison, double [[TMP1]], i32 0 -; CHECK-NEXT: [[TMP3:%.*]] = insertelement <2 x double> [[TMP2]], double [[TMP1]], i32 1 +; CHECK-NEXT: [[BROADCAST_SPLATINSERT:%.*]] = insertelement <2 x double> poison, double [[TMP1]], i64 0 +; CHECK-NEXT: [[TMP3:%.*]] = shufflevector <2 x double> [[BROADCAST_SPLATINSERT]], <2 x double> poison, <2 x i32> zeroinitializer ; CHECK-NEXT: br label %[[VECTOR_BODY:.*]] ; CHECK: [[VECTOR_BODY]]: ; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ] diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/masked-call-scalarize.ll b/llvm/test/Transforms/LoopVectorize/AArch64/masked-call-scalarize.ll index 157b787..3558957 100644 --- a/llvm/test/Transforms/LoopVectorize/AArch64/masked-call-scalarize.ll +++ b/llvm/test/Transforms/LoopVectorize/AArch64/masked-call-scalarize.ll @@ -64,9 +64,9 @@ define void @test_widen_exp_v2(ptr noalias %p2, ptr noalias %p, i64 %n) #5 { ; TFCOMMON-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[ENTRY:%.*]] ], [ [[INDEX_NEXT:%.*]], [[PRED_STORE_CONTINUE6:%.*]] ] ; TFCOMMON-NEXT: [[ACTIVE_LANE_MASK:%.*]] = phi <2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY]], [[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], [[PRED_STORE_CONTINUE6]] ] ; TFCOMMON-NEXT: [[LD:%.*]] = load double, ptr [[P2:%.*]], align 8 -; TFCOMMON-NEXT: [[TMP5:%.*]] = tail call double @llvm.exp.f64(double [[LD]]) #[[ATTR3:[0-9]+]] -; TFCOMMON-NEXT: [[TMP7:%.*]] = insertelement <2 x double> poison, double [[TMP5]], i32 0 -; TFCOMMON-NEXT: [[TMP8:%.*]] = insertelement <2 x double> [[TMP7]], double [[TMP5]], i32 1 +; TFCOMMON-NEXT: [[TMP5:%.*]] = tail call double @llvm.exp.f64(double [[LD]]) #[[ATTR2:[0-9]+]] +; TFCOMMON-NEXT: [[BROADCAST_SPLATINSERT:%.*]] = insertelement <2 x double> poison, double [[TMP5]], i64 0 +; TFCOMMON-NEXT: [[TMP8:%.*]] = shufflevector <2 x double> [[BROADCAST_SPLATINSERT]], <2 x double> poison, <2 x i32> zeroinitializer ; TFCOMMON-NEXT: [[TMP9:%.*]] = fcmp ogt <2 x double> [[TMP8]], zeroinitializer ; TFCOMMON-NEXT: [[PREDPHI:%.*]] = select <2 x i1> [[TMP9]], <2 x double> zeroinitializer, <2 x double> splat (double 1.000000e+00) ; TFCOMMON-NEXT: [[TMP16:%.*]] = extractelement <2 x i1> [[ACTIVE_LANE_MASK]], i32 0 @@ -79,7 +79,7 @@ define void @test_widen_exp_v2(ptr noalias %p2, ptr noalias %p, i64 %n) #5 { ; TFCOMMON-NEXT: [[TMP14:%.*]] = extractelement <2 x i1> [[ACTIVE_LANE_MASK]], i32 1 ; TFCOMMON-NEXT: br i1 [[TMP14]], label [[PRED_STORE_IF1:%.*]], label [[PRED_STORE_CONTINUE6]] ; TFCOMMON: pred.store.if1: -; TFCOMMON-NEXT: [[TMP19:%.*]] = extractelement <2 x double> [[PREDPHI]], i32 1 +; TFCOMMON-NEXT: [[TMP19:%.*]] = extractelement <2 x double> [[PREDPHI]], i32 0 ; TFCOMMON-NEXT: store double [[TMP19]], ptr [[P]], align 8 ; TFCOMMON-NEXT: br label [[PRED_STORE_CONTINUE6]] ; TFCOMMON: pred.store.continue2: @@ -105,9 +105,9 @@ define void @test_widen_exp_v2(ptr noalias %p2, ptr noalias %p, i64 %n) #5 { ; TFA_INTERLEAVE-NEXT: [[ACTIVE_LANE_MASK:%.*]] = phi <2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY]], [[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], [[PRED_STORE_CONTINUE9]] ] ; TFA_INTERLEAVE-NEXT: [[ACTIVE_LANE_MASK2:%.*]] = phi <2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY1]], [[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT10:%.*]], [[PRED_STORE_CONTINUE9]] ] ; TFA_INTERLEAVE-NEXT: [[TMP4:%.*]] = load double, ptr [[P2:%.*]], align 8 -; TFA_INTERLEAVE-NEXT: [[TMP9:%.*]] = tail call double @llvm.exp.f64(double [[TMP4]]) #[[ATTR3:[0-9]+]] -; TFA_INTERLEAVE-NEXT: [[TMP11:%.*]] = insertelement <2 x double> poison, double [[TMP9]], i32 0 -; TFA_INTERLEAVE-NEXT: [[TMP12:%.*]] = insertelement <2 x double> [[TMP11]], double [[TMP9]], i32 1 +; TFA_INTERLEAVE-NEXT: [[TMP5:%.*]] = tail call double @llvm.exp.f64(double [[TMP4]]) #[[ATTR2:[0-9]+]] +; TFA_INTERLEAVE-NEXT: [[BROADCAST_SPLATINSERT:%.*]] = insertelement <2 x double> poison, double [[TMP5]], i64 0 +; TFA_INTERLEAVE-NEXT: [[TMP12:%.*]] = shufflevector <2 x double> [[BROADCAST_SPLATINSERT]], <2 x double> poison, <2 x i32> zeroinitializer ; TFA_INTERLEAVE-NEXT: [[TMP14:%.*]] = fcmp ogt <2 x double> [[TMP12]], zeroinitializer ; TFA_INTERLEAVE-NEXT: [[PREDPHI3:%.*]] = select <2 x i1> [[TMP14]], <2 x double> zeroinitializer, <2 x double> splat (double 1.000000e+00) ; TFA_INTERLEAVE-NEXT: [[TMP19:%.*]] = extractelement <2 x i1> [[ACTIVE_LANE_MASK]], i32 0 @@ -120,7 +120,7 @@ define void @test_widen_exp_v2(ptr noalias %p2, ptr noalias %p, i64 %n) #5 { ; TFA_INTERLEAVE-NEXT: [[TMP29:%.*]] = extractelement <2 x i1> [[ACTIVE_LANE_MASK]], i32 1 ; TFA_INTERLEAVE-NEXT: br i1 [[TMP29]], label [[PRED_STORE_IF4:%.*]], label [[PRED_STORE_CONTINUE5:%.*]] ; TFA_INTERLEAVE: pred.store.if3: -; TFA_INTERLEAVE-NEXT: [[TMP22:%.*]] = extractelement <2 x double> [[PREDPHI3]], i32 1 +; TFA_INTERLEAVE-NEXT: [[TMP22:%.*]] = extractelement <2 x double> [[PREDPHI3]], i32 0 ; TFA_INTERLEAVE-NEXT: store double [[TMP22]], ptr [[P]], align 8 ; TFA_INTERLEAVE-NEXT: br label [[PRED_STORE_CONTINUE5]] ; TFA_INTERLEAVE: pred.store.continue4: @@ -134,7 +134,7 @@ define void @test_widen_exp_v2(ptr noalias %p2, ptr noalias %p, i64 %n) #5 { ; TFA_INTERLEAVE-NEXT: [[TMP25:%.*]] = extractelement <2 x i1> [[ACTIVE_LANE_MASK2]], i32 1 ; TFA_INTERLEAVE-NEXT: br i1 [[TMP25]], label [[PRED_STORE_IF8:%.*]], label [[PRED_STORE_CONTINUE9]] ; TFA_INTERLEAVE: pred.store.if7: -; TFA_INTERLEAVE-NEXT: [[TMP34:%.*]] = extractelement <2 x double> [[PREDPHI3]], i32 1 +; TFA_INTERLEAVE-NEXT: [[TMP34:%.*]] = extractelement <2 x double> [[PREDPHI3]], i32 0 ; TFA_INTERLEAVE-NEXT: store double [[TMP34]], ptr [[P]], align 8 ; TFA_INTERLEAVE-NEXT: br label [[PRED_STORE_CONTINUE9]] ; TFA_INTERLEAVE: pred.store.continue8: diff --git a/llvm/test/Transforms/LoopVectorize/X86/replicate-recipe-with-only-first-lane-used.ll b/llvm/test/Transforms/LoopVectorize/X86/replicate-recipe-with-only-first-lane-used.ll index 03087bb..4590dfc 100644 --- a/llvm/test/Transforms/LoopVectorize/X86/replicate-recipe-with-only-first-lane-used.ll +++ b/llvm/test/Transforms/LoopVectorize/X86/replicate-recipe-with-only-first-lane-used.ll @@ -199,10 +199,8 @@ define float @uniform_load_replicating_select(ptr %A, ptr %B, i64 %1) { ; CHECK-NEXT: [[TMP5:%.*]] = add i64 [[INDEX]], 7 ; CHECK-NEXT: [[TMP6:%.*]] = load float, ptr [[A]], align 4 ; CHECK-NEXT: [[TMP10:%.*]] = fcmp ogt float [[TMP6]], 0.000000e+00 -; CHECK-NEXT: [[TMP8:%.*]] = insertelement <4 x i1> poison, i1 [[TMP10]], i32 0 -; CHECK-NEXT: [[TMP9:%.*]] = insertelement <4 x i1> [[TMP8]], i1 [[TMP10]], i32 1 -; CHECK-NEXT: [[TMP13:%.*]] = insertelement <4 x i1> [[TMP9]], i1 [[TMP10]], i32 2 -; CHECK-NEXT: [[TMP14:%.*]] = insertelement <4 x i1> [[TMP13]], i1 [[TMP10]], i32 3 +; CHECK-NEXT: [[BROADCAST_SPLATINSERT:%.*]] = insertelement <4 x i1> poison, i1 [[TMP10]], i64 0 +; CHECK-NEXT: [[TMP14:%.*]] = shufflevector <4 x i1> [[BROADCAST_SPLATINSERT]], <4 x i1> poison, <4 x i32> zeroinitializer ; CHECK-NEXT: [[TMP15:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[TMP2]] ; CHECK-NEXT: [[TMP16:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[TMP3]] ; CHECK-NEXT: [[TMP17:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[TMP4]] diff --git a/llvm/test/Transforms/LoopVectorize/hoist-and-sink-mem-ops-with-invariant-pointers.ll b/llvm/test/Transforms/LoopVectorize/hoist-and-sink-mem-ops-with-invariant-pointers.ll new file mode 100644 index 0000000..8615401 --- /dev/null +++ b/llvm/test/Transforms/LoopVectorize/hoist-and-sink-mem-ops-with-invariant-pointers.ll @@ -0,0 +1,247 @@ +; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-globals none --version 6 +; RUN: opt -passes=loop-vectorize -force-vector-width=4 -force-vector-interleave=1 -S %s | FileCheck %s + +define void @hoist_invariant_load_noalias_due_to_memchecks(ptr %dst, ptr %invariant_ptr, i32 %n) { +; CHECK-LABEL: define void @hoist_invariant_load_noalias_due_to_memchecks( +; CHECK-SAME: ptr [[DST:%.*]], ptr [[INVARIANT_PTR:%.*]], i32 [[N:%.*]]) { +; CHECK-NEXT: [[ENTRY:.*]]: +; CHECK-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i32 [[N]], 4 +; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK]], label %[[SCALAR_PH:.*]], label %[[VECTOR_MEMCHECK:.*]] +; CHECK: [[VECTOR_MEMCHECK]]: +; CHECK-NEXT: [[TMP0:%.*]] = add i32 [[N]], -1 +; CHECK-NEXT: [[TMP1:%.*]] = zext i32 [[TMP0]] to i64 +; CHECK-NEXT: [[TMP2:%.*]] = shl nuw nsw i64 [[TMP1]], 2 +; CHECK-NEXT: [[TMP3:%.*]] = add nuw nsw i64 [[TMP2]], 4 +; CHECK-NEXT: [[SCEVGEP:%.*]] = getelementptr i8, ptr [[DST]], i64 [[TMP3]] +; CHECK-NEXT: [[SCEVGEP1:%.*]] = getelementptr i8, ptr [[INVARIANT_PTR]], i64 4 +; CHECK-NEXT: [[BOUND0:%.*]] = icmp ult ptr [[DST]], [[SCEVGEP1]] +; CHECK-NEXT: [[BOUND1:%.*]] = icmp ult ptr [[INVARIANT_PTR]], [[SCEVGEP]] +; CHECK-NEXT: [[FOUND_CONFLICT:%.*]] = and i1 [[BOUND0]], [[BOUND1]] +; CHECK-NEXT: br i1 [[FOUND_CONFLICT]], label %[[SCALAR_PH]], label %[[VECTOR_PH:.*]] +; CHECK: [[VECTOR_PH]]: +; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i32 [[N]], 4 +; CHECK-NEXT: [[N_VEC:%.*]] = sub i32 [[N]], [[N_MOD_VF]] +; CHECK-NEXT: br label %[[VECTOR_BODY:.*]] +; CHECK: [[VECTOR_BODY]]: +; CHECK-NEXT: [[INDEX:%.*]] = phi i32 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ] +; CHECK-NEXT: [[TMP4:%.*]] = load i32, ptr [[INVARIANT_PTR]], align 4, !alias.scope [[META0:![0-9]+]] +; CHECK-NEXT: [[BROADCAST_SPLATINSERT:%.*]] = insertelement <4 x i32> poison, i32 [[TMP4]], i64 0 +; CHECK-NEXT: [[BROADCAST_SPLAT:%.*]] = shufflevector <4 x i32> [[BROADCAST_SPLATINSERT]], <4 x i32> poison, <4 x i32> zeroinitializer +; CHECK-NEXT: [[TMP5:%.*]] = getelementptr inbounds i32, ptr [[DST]], i32 [[INDEX]] +; CHECK-NEXT: store <4 x i32> [[BROADCAST_SPLAT]], ptr [[TMP5]], align 4, !alias.scope [[META3:![0-9]+]], !noalias [[META0]] +; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i32 [[INDEX]], 4 +; CHECK-NEXT: [[TMP6:%.*]] = icmp eq i32 [[INDEX_NEXT]], [[N_VEC]] +; CHECK-NEXT: br i1 [[TMP6]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP5:![0-9]+]] +; CHECK: [[MIDDLE_BLOCK]]: +; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i32 [[N]], [[N_VEC]] +; CHECK-NEXT: br i1 [[CMP_N]], label %[[EXIT:.*]], label %[[SCALAR_PH]] +; CHECK: [[SCALAR_PH]]: +; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i32 [ [[N_VEC]], %[[MIDDLE_BLOCK]] ], [ 0, %[[ENTRY]] ], [ 0, %[[VECTOR_MEMCHECK]] ] +; CHECK-NEXT: br label %[[LOOP:.*]] +; CHECK: [[LOOP]]: +; CHECK-NEXT: [[IV:%.*]] = phi i32 [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], %[[LOOP]] ] +; CHECK-NEXT: [[INV_VAL:%.*]] = load i32, ptr [[INVARIANT_PTR]], align 4 +; CHECK-NEXT: [[GEP:%.*]] = getelementptr inbounds i32, ptr [[DST]], i32 [[IV]] +; CHECK-NEXT: store i32 [[INV_VAL]], ptr [[GEP]], align 4 +; CHECK-NEXT: [[IV_NEXT]] = add nuw nsw i32 [[IV]], 1 +; CHECK-NEXT: [[EC:%.*]] = icmp eq i32 [[IV_NEXT]], [[N]] +; CHECK-NEXT: br i1 [[EC]], label %[[EXIT]], label %[[LOOP]], !llvm.loop [[LOOP8:![0-9]+]] +; CHECK: [[EXIT]]: +; CHECK-NEXT: ret void +; +entry: + br label %loop + +loop: + %iv = phi i32 [ 0, %entry ], [ %iv.next, %loop ] + %inv_val = load i32, ptr %invariant_ptr, align 4 + %gep = getelementptr inbounds i32, ptr %dst, i32 %iv + store i32 %inv_val, ptr %gep, align 4 + %iv.next = add nuw nsw i32 %iv, 1 + %ec = icmp eq i32 %iv.next, %n + br i1 %ec, label %exit, label %loop + +exit: + ret void +} + +; Test that loads with non-invariant addresses are not hoisted. +define void @dont_hoist_variant_address(ptr %dst, ptr %src, i32 %n) { +; CHECK-LABEL: define void @dont_hoist_variant_address( +; CHECK-SAME: ptr [[DST:%.*]], ptr [[SRC:%.*]], i32 [[N:%.*]]) { +; CHECK-NEXT: [[ENTRY:.*]]: +; CHECK-NEXT: [[SRC2:%.*]] = ptrtoint ptr [[SRC]] to i64 +; CHECK-NEXT: [[A1:%.*]] = ptrtoint ptr [[DST]] to i64 +; CHECK-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i32 [[N]], 4 +; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK]], label %[[SCALAR_PH:.*]], label %[[VECTOR_MEMCHECK:.*]] +; CHECK: [[VECTOR_MEMCHECK]]: +; CHECK-NEXT: [[TMP0:%.*]] = sub i64 [[A1]], [[SRC2]] +; CHECK-NEXT: [[DIFF_CHECK:%.*]] = icmp ult i64 [[TMP0]], 16 +; CHECK-NEXT: br i1 [[DIFF_CHECK]], label %[[SCALAR_PH]], label %[[VECTOR_PH:.*]] +; CHECK: [[VECTOR_PH]]: +; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i32 [[N]], 4 +; CHECK-NEXT: [[N_VEC:%.*]] = sub i32 [[N]], [[N_MOD_VF]] +; CHECK-NEXT: br label %[[VECTOR_BODY:.*]] +; CHECK: [[VECTOR_BODY]]: +; CHECK-NEXT: [[INDEX:%.*]] = phi i32 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ] +; CHECK-NEXT: [[TMP1:%.*]] = getelementptr inbounds i32, ptr [[SRC]], i32 [[INDEX]] +; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <4 x i32>, ptr [[TMP1]], align 4 +; CHECK-NEXT: [[TMP2:%.*]] = getelementptr inbounds i32, ptr [[DST]], i32 [[INDEX]] +; CHECK-NEXT: store <4 x i32> [[WIDE_LOAD]], ptr [[TMP2]], align 4 +; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i32 [[INDEX]], 4 +; CHECK-NEXT: [[TMP3:%.*]] = icmp eq i32 [[INDEX_NEXT]], [[N_VEC]] +; CHECK-NEXT: br i1 [[TMP3]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP9:![0-9]+]] +; CHECK: [[MIDDLE_BLOCK]]: +; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i32 [[N]], [[N_VEC]] +; CHECK-NEXT: br i1 [[CMP_N]], label %[[EXIT:.*]], label %[[SCALAR_PH]] +; CHECK: [[SCALAR_PH]]: +; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i32 [ [[N_VEC]], %[[MIDDLE_BLOCK]] ], [ 0, %[[ENTRY]] ], [ 0, %[[VECTOR_MEMCHECK]] ] +; CHECK-NEXT: br label %[[LOOP:.*]] +; CHECK: [[LOOP]]: +; CHECK-NEXT: [[IV:%.*]] = phi i32 [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], %[[LOOP]] ] +; CHECK-NEXT: [[GEP_SRC:%.*]] = getelementptr inbounds i32, ptr [[SRC]], i32 [[IV]] +; CHECK-NEXT: [[VAL:%.*]] = load i32, ptr [[GEP_SRC]], align 4 +; CHECK-NEXT: [[GEP_DST:%.*]] = getelementptr inbounds i32, ptr [[DST]], i32 [[IV]] +; CHECK-NEXT: store i32 [[VAL]], ptr [[GEP_DST]], align 4 +; CHECK-NEXT: [[IV_NEXT]] = add nuw nsw i32 [[IV]], 1 +; CHECK-NEXT: [[EC:%.*]] = icmp eq i32 [[IV_NEXT]], [[N]] +; CHECK-NEXT: br i1 [[EC]], label %[[EXIT]], label %[[LOOP]], !llvm.loop [[LOOP10:![0-9]+]] +; CHECK: [[EXIT]]: +; CHECK-NEXT: ret void +; +entry: + br label %loop + +loop: + %iv = phi i32 [ 0, %entry ], [ %iv.next, %loop ] + %gep.src = getelementptr inbounds i32, ptr %src, i32 %iv + %val = load i32, ptr %gep.src, align 4 + %gep.dst = getelementptr inbounds i32, ptr %dst, i32 %iv + store i32 %val, ptr %gep.dst, align 4 + %iv.next = add nuw nsw i32 %iv, 1 + %ec = icmp eq i32 %iv.next, %n + br i1 %ec, label %exit, label %loop + +exit: + ret void +} + +; Test that predicated loads are not hoisted. +define void @dont_hoist_predicated_load(ptr %dst, ptr %invariant_ptr, ptr %cond_ptr, i32 %n) { +; CHECK-LABEL: define void @dont_hoist_predicated_load( +; CHECK-SAME: ptr [[DST:%.*]], ptr [[INVARIANT_PTR:%.*]], ptr [[COND_PTR:%.*]], i32 [[N:%.*]]) { +; CHECK-NEXT: [[ENTRY:.*]]: +; CHECK-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i32 [[N]], 4 +; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK]], label %[[SCALAR_PH:.*]], label %[[VECTOR_MEMCHECK:.*]] +; CHECK: [[VECTOR_MEMCHECK]]: +; CHECK-NEXT: [[TMP5:%.*]] = add i32 [[N]], -1 +; CHECK-NEXT: [[TMP20:%.*]] = zext i32 [[TMP5]] to i64 +; CHECK-NEXT: [[TMP22:%.*]] = shl nuw nsw i64 [[TMP20]], 2 +; CHECK-NEXT: [[TMP3:%.*]] = add nuw nsw i64 [[TMP22]], 4 +; CHECK-NEXT: [[SCEVGEP:%.*]] = getelementptr i8, ptr [[DST]], i64 [[TMP3]] +; CHECK-NEXT: [[SCEVGEP1:%.*]] = getelementptr i8, ptr [[COND_PTR]], i64 [[TMP3]] +; CHECK-NEXT: [[SCEVGEP2:%.*]] = getelementptr i8, ptr [[INVARIANT_PTR]], i64 4 +; CHECK-NEXT: [[BOUND0:%.*]] = icmp ult ptr [[DST]], [[SCEVGEP1]] +; CHECK-NEXT: [[BOUND1:%.*]] = icmp ult ptr [[COND_PTR]], [[SCEVGEP]] +; CHECK-NEXT: [[FOUND_CONFLICT:%.*]] = and i1 [[BOUND0]], [[BOUND1]] +; CHECK-NEXT: [[BOUND03:%.*]] = icmp ult ptr [[DST]], [[SCEVGEP2]] +; CHECK-NEXT: [[BOUND14:%.*]] = icmp ult ptr [[INVARIANT_PTR]], [[SCEVGEP]] +; CHECK-NEXT: [[FOUND_CONFLICT5:%.*]] = and i1 [[BOUND03]], [[BOUND14]] +; CHECK-NEXT: [[CONFLICT_RDX:%.*]] = or i1 [[FOUND_CONFLICT]], [[FOUND_CONFLICT5]] +; CHECK-NEXT: br i1 [[CONFLICT_RDX]], label %[[SCALAR_PH]], label %[[VECTOR_PH:.*]] +; CHECK: [[VECTOR_PH]]: +; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i32 [[N]], 4 +; CHECK-NEXT: [[N_VEC:%.*]] = sub i32 [[N]], [[N_MOD_VF]] +; CHECK-NEXT: br label %[[VECTOR_BODY:.*]] +; CHECK: [[VECTOR_BODY]]: +; CHECK-NEXT: [[INDEX:%.*]] = phi i32 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[PRED_STORE_CONTINUE11:.*]] ] +; CHECK-NEXT: [[TMP0:%.*]] = getelementptr inbounds i32, ptr [[COND_PTR]], i32 [[INDEX]] +; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <4 x i32>, ptr [[TMP0]], align 4, !alias.scope [[META11:![0-9]+]] +; CHECK-NEXT: [[TMP1:%.*]] = icmp sgt <4 x i32> [[WIDE_LOAD]], zeroinitializer +; CHECK-NEXT: [[TMP2:%.*]] = extractelement <4 x i1> [[TMP1]], i32 0 +; CHECK-NEXT: br i1 [[TMP2]], label %[[PRED_STORE_IF:.*]], label %[[PRED_STORE_CONTINUE:.*]] +; CHECK: [[PRED_STORE_IF]]: +; CHECK-NEXT: [[TMP7:%.*]] = load i32, ptr [[INVARIANT_PTR]], align 4, !alias.scope [[META14:![0-9]+]] +; CHECK-NEXT: [[TMP4:%.*]] = add i32 [[INDEX]], 0 +; CHECK-NEXT: [[TMP9:%.*]] = getelementptr inbounds i32, ptr [[DST]], i32 [[TMP4]] +; CHECK-NEXT: store i32 [[TMP7]], ptr [[TMP9]], align 4, !alias.scope [[META16:![0-9]+]], !noalias [[META18:![0-9]+]] +; CHECK-NEXT: br label %[[PRED_STORE_CONTINUE]] +; CHECK: [[PRED_STORE_CONTINUE]]: +; CHECK-NEXT: [[TMP6:%.*]] = extractelement <4 x i1> [[TMP1]], i32 1 +; CHECK-NEXT: br i1 [[TMP6]], label %[[PRED_STORE_IF6:.*]], label %[[PRED_STORE_CONTINUE7:.*]] +; CHECK: [[PRED_STORE_IF6]]: +; CHECK-NEXT: [[TMP11:%.*]] = load i32, ptr [[INVARIANT_PTR]], align 4, !alias.scope [[META14]] +; CHECK-NEXT: [[TMP8:%.*]] = add i32 [[INDEX]], 1 +; CHECK-NEXT: [[TMP13:%.*]] = getelementptr inbounds i32, ptr [[DST]], i32 [[TMP8]] +; CHECK-NEXT: store i32 [[TMP11]], ptr [[TMP13]], align 4, !alias.scope [[META16]], !noalias [[META18]] +; CHECK-NEXT: br label %[[PRED_STORE_CONTINUE7]] +; CHECK: [[PRED_STORE_CONTINUE7]]: +; CHECK-NEXT: [[TMP10:%.*]] = extractelement <4 x i1> [[TMP1]], i32 2 +; CHECK-NEXT: br i1 [[TMP10]], label %[[PRED_STORE_IF8:.*]], label %[[PRED_STORE_CONTINUE9:.*]] +; CHECK: [[PRED_STORE_IF8]]: +; CHECK-NEXT: [[TMP15:%.*]] = load i32, ptr [[INVARIANT_PTR]], align 4, !alias.scope [[META14]] +; CHECK-NEXT: [[TMP12:%.*]] = add i32 [[INDEX]], 2 +; CHECK-NEXT: [[TMP17:%.*]] = getelementptr inbounds i32, ptr [[DST]], i32 [[TMP12]] +; CHECK-NEXT: store i32 [[TMP15]], ptr [[TMP17]], align 4, !alias.scope [[META16]], !noalias [[META18]] +; CHECK-NEXT: br label %[[PRED_STORE_CONTINUE9]] +; CHECK: [[PRED_STORE_CONTINUE9]]: +; CHECK-NEXT: [[TMP14:%.*]] = extractelement <4 x i1> [[TMP1]], i32 3 +; CHECK-NEXT: br i1 [[TMP14]], label %[[PRED_STORE_IF10:.*]], label %[[PRED_STORE_CONTINUE11]] +; CHECK: [[PRED_STORE_IF10]]: +; CHECK-NEXT: [[TMP19:%.*]] = load i32, ptr [[INVARIANT_PTR]], align 4, !alias.scope [[META14]] +; CHECK-NEXT: [[TMP16:%.*]] = add i32 [[INDEX]], 3 +; CHECK-NEXT: [[TMP21:%.*]] = getelementptr inbounds i32, ptr [[DST]], i32 [[TMP16]] +; CHECK-NEXT: store i32 [[TMP19]], ptr [[TMP21]], align 4, !alias.scope [[META16]], !noalias [[META18]] +; CHECK-NEXT: br label %[[PRED_STORE_CONTINUE11]] +; CHECK: [[PRED_STORE_CONTINUE11]]: +; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i32 [[INDEX]], 4 +; CHECK-NEXT: [[TMP18:%.*]] = icmp eq i32 [[INDEX_NEXT]], [[N_VEC]] +; CHECK-NEXT: br i1 [[TMP18]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP19:![0-9]+]] +; CHECK: [[MIDDLE_BLOCK]]: +; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i32 [[N]], [[N_VEC]] +; CHECK-NEXT: br i1 [[CMP_N]], label %[[EXIT:.*]], label %[[SCALAR_PH]] +; CHECK: [[SCALAR_PH]]: +; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i32 [ [[N_VEC]], %[[MIDDLE_BLOCK]] ], [ 0, %[[ENTRY]] ], [ 0, %[[VECTOR_MEMCHECK]] ] +; CHECK-NEXT: br label %[[LOOP:.*]] +; CHECK: [[LOOP]]: +; CHECK-NEXT: [[IV:%.*]] = phi i32 [ [[BC_RESUME_VAL]], %[[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], %[[LOOP_LATCH:.*]] ] +; CHECK-NEXT: [[GEP_COND:%.*]] = getelementptr inbounds i32, ptr [[COND_PTR]], i32 [[IV]] +; CHECK-NEXT: [[COND:%.*]] = load i32, ptr [[GEP_COND]], align 4 +; CHECK-NEXT: [[CMP:%.*]] = icmp sgt i32 [[COND]], 0 +; CHECK-NEXT: br i1 [[CMP]], label %[[IF_THEN:.*]], label %[[LOOP_LATCH]] +; CHECK: [[IF_THEN]]: +; CHECK-NEXT: [[INV_VAL:%.*]] = load i32, ptr [[INVARIANT_PTR]], align 4 +; CHECK-NEXT: [[GEP:%.*]] = getelementptr inbounds i32, ptr [[DST]], i32 [[IV]] +; CHECK-NEXT: store i32 [[INV_VAL]], ptr [[GEP]], align 4 +; CHECK-NEXT: br label %[[LOOP_LATCH]] +; CHECK: [[LOOP_LATCH]]: +; CHECK-NEXT: [[IV_NEXT]] = add nuw nsw i32 [[IV]], 1 +; CHECK-NEXT: [[EC:%.*]] = icmp eq i32 [[IV_NEXT]], [[N]] +; CHECK-NEXT: br i1 [[EC]], label %[[EXIT]], label %[[LOOP]], !llvm.loop [[LOOP20:![0-9]+]] +; CHECK: [[EXIT]]: +; CHECK-NEXT: ret void +; +entry: + br label %loop + +loop: + %iv = phi i32 [ 0, %entry ], [ %iv.next, %loop.latch ] + %gep.cond = getelementptr inbounds i32, ptr %cond_ptr, i32 %iv + %cond = load i32, ptr %gep.cond, align 4 + %cmp = icmp sgt i32 %cond, 0 + br i1 %cmp, label %if.then, label %loop.latch + +if.then: + %inv_val = load i32, ptr %invariant_ptr, align 4 + %gep = getelementptr inbounds i32, ptr %dst, i32 %iv + store i32 %inv_val, ptr %gep, align 4 + br label %loop.latch + +loop.latch: + %iv.next = add nuw nsw i32 %iv, 1 + %ec = icmp eq i32 %iv.next, %n + br i1 %ec, label %exit, label %loop + +exit: + ret void +} diff --git a/llvm/test/Transforms/LoopVectorize/vplan-printing-metadata.ll b/llvm/test/Transforms/LoopVectorize/vplan-printing-metadata.ll new file mode 100644 index 0000000..857b913 --- /dev/null +++ b/llvm/test/Transforms/LoopVectorize/vplan-printing-metadata.ll @@ -0,0 +1,100 @@ +; REQUIRES: asserts + +; RUN: opt -passes=loop-vectorize -debug-only=loop-vectorize -force-vector-interleave=1 -force-vector-width=4 -disable-output %s 2>&1 | FileCheck %s + +define void @test_widen_metadata(ptr noalias %A, ptr noalias %B, i32 %n) { +; CHECK-LABEL: Checking a loop in 'test_widen_metadata' +; CHECK: VPlan 'Initial VPlan for VF={4},UF>=1' { +; CHECK: <x1> vector loop: { +; CHECK: vector.body: +; CHECK: WIDEN ir<%lv> = load vp<{{.*}}> +; CHECK: WIDEN-CAST ir<%conv> = sitofp ir<%lv> to float +; CHECK: WIDEN ir<%mul> = fmul ir<%conv>, ir<2.000000e+00> +; CHECK: WIDEN-CAST ir<%conv.back> = fptosi ir<%mul> to i32 +; CHECK: WIDEN store vp<{{.*}}>, ir<%conv.back> +; +entry: + br label %loop + +loop: + %i = phi i32 [ 0, %entry ], [ %i.next, %loop ] + %gep.A = getelementptr inbounds i32, ptr %A, i32 %i + %lv = load i32, ptr %gep.A, align 4, !tbaa !0, !range !6 + %conv = sitofp i32 %lv to float, !fpmath !5 + %mul = fmul float %conv, 2.0, !fpmath !5 + %conv.back = fptosi float %mul to i32 + %gep.B = getelementptr inbounds i32, ptr %B, i32 %i + store i32 %conv.back, ptr %gep.B, align 4, !tbaa !0 + %i.next = add i32 %i, 1 + %cond = icmp eq i32 %i.next, %n + br i1 %cond, label %exit, label %loop + +exit: + ret void +} + +declare float @llvm.sqrt.f32(float) + +define void @test_intrinsic_with_metadata(ptr noalias %A, ptr noalias %B, i32 %n) { +; CHECK-LABEL: Checking a loop in 'test_intrinsic_with_metadata' +; CHECK: VPlan 'Initial VPlan for VF={4},UF>=1' { +; CHECK: <x1> vector loop: { +; CHECK: vector.body: +; CHECK: WIDEN ir<%lv> = load vp<{{.*}}> +; CHECK: WIDEN-INTRINSIC ir<%sqrt> = call llvm.sqrt(ir<%lv>) +; CHECK: WIDEN store vp<{{.*}}>, ir<%sqrt> +; +entry: + br label %loop + +loop: + %i = phi i32 [ 0, %entry ], [ %i.next, %loop ] + %gep.A = getelementptr inbounds float, ptr %A, i32 %i + %lv = load float, ptr %gep.A, align 4, !tbaa !0 + %sqrt = call float @llvm.sqrt.f32(float %lv), !fpmath !5 + %gep.B = getelementptr inbounds float, ptr %B, i32 %i + store float %sqrt, ptr %gep.B, align 4, !tbaa !0 + %i.next = add i32 %i, 1 + %cond = icmp eq i32 %i.next, %n + br i1 %cond, label %exit, label %loop + +exit: + ret void +} + +define void @test_widen_with_multiple_metadata(ptr noalias %A, ptr noalias %B, i32 %n) { +; CHECK-LABEL: Checking a loop in 'test_widen_with_multiple_metadata' +; CHECK: VPlan 'Initial VPlan for VF={4},UF>=1' { +; CHECK: <x1> vector loop: { +; CHECK: vector.body: +; CHECK: WIDEN ir<%lv> = load vp<{{.*}}> +; CHECK: WIDEN-CAST ir<%conv> = sitofp ir<%lv> to float +; CHECK: WIDEN ir<%mul> = fmul ir<%conv>, ir<2.000000e+00> +; CHECK: WIDEN-CAST ir<%conv.back> = fptosi ir<%mul> to i32 +; CHECK: WIDEN store vp<{{.*}}>, ir<%conv.back> +; +entry: + br label %loop + +loop: + %i = phi i32 [ 0, %entry ], [ %i.next, %loop ] + %gep.A = getelementptr inbounds i32, ptr %A, i32 %i + %lv = load i32, ptr %gep.A, align 4, !tbaa !0, !range !6 + %conv = sitofp i32 %lv to float + %mul = fmul float %conv, 2.0 + %conv.back = fptosi float %mul to i32 + %gep.B = getelementptr inbounds i32, ptr %B, i32 %i + store i32 %conv.back, ptr %gep.B, align 4, !tbaa !0 + %i.next = add i32 %i, 1 + %cond = icmp eq i32 %i.next, %n + br i1 %cond, label %exit, label %loop + +exit: + ret void +} + +!0 = !{!1, !1, i64 0} +!1 = !{!"float", !2} +!2 = !{!"root"} +!5 = !{float 2.500000e+00} +!6 = !{i32 0, i32 100} diff --git a/llvm/test/Transforms/PhaseOrdering/AArch64/hoist-load-from-vector-loop.ll b/llvm/test/Transforms/PhaseOrdering/AArch64/hoist-load-from-vector-loop.ll new file mode 100644 index 0000000..a35bcf1 --- /dev/null +++ b/llvm/test/Transforms/PhaseOrdering/AArch64/hoist-load-from-vector-loop.ll @@ -0,0 +1,46 @@ +; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-globals none --version 6 +; RUN: opt -passes='default<O3>' -S %s | FileCheck %s + +target triple = "arm64-apple-macosx" + +%"class.dealii::VectorizedArray" = type { [4 x double] } + +define void @hoist_invariant_load(ptr %invariant_ptr, i64 %num_elements, ptr %array) { +; CHECK-LABEL: define void @hoist_invariant_load( +; CHECK-SAME: ptr readonly captures(none) [[INVARIANT_PTR:%.*]], i64 [[NUM_ELEMENTS:%.*]], ptr captures(none) [[ARRAY:%.*]]) local_unnamed_addr #[[ATTR0:[0-9]+]] { +; CHECK-NEXT: [[ENTRY:.*]]: +; CHECK-NEXT: [[CMP1_NOT:%.*]] = icmp eq i64 [[NUM_ELEMENTS]], 0 +; CHECK-NEXT: br i1 [[CMP1_NOT]], label %[[EXIT:.*]], label %[[LOOP_LATCH:.*]] +; CHECK: [[LOOP_LATCH]]: +; CHECK-NEXT: [[I2:%.*]] = phi i64 [ [[I_NEXT:%.*]], %[[LOOP_LATCH]] ], [ 0, %[[ENTRY]] ] +; CHECK-NEXT: [[GEP:%.*]] = getelementptr nusw %"class.dealii::VectorizedArray", ptr [[ARRAY]], i64 [[I2]] +; CHECK-NEXT: [[INVARIANT_VAL:%.*]] = load double, ptr [[INVARIANT_PTR]], align 8 +; CHECK-NEXT: [[ARRAY_VAL:%.*]] = load double, ptr [[GEP]], align 8 +; CHECK-NEXT: [[SUM:%.*]] = fadd double [[INVARIANT_VAL]], [[ARRAY_VAL]] +; CHECK-NEXT: store double [[SUM]], ptr [[GEP]], align 8 +; CHECK-NEXT: [[I_NEXT]] = add nuw i64 [[I2]], 1 +; CHECK-NEXT: [[EXITCOND_NOT:%.*]] = icmp eq i64 [[I_NEXT]], [[NUM_ELEMENTS]] +; CHECK-NEXT: br i1 [[EXITCOND_NOT]], label %[[EXIT]], label %[[LOOP_LATCH]] +; CHECK: [[EXIT]]: +; CHECK-NEXT: ret void +; +entry: + br label %loop.header + +loop.header: ; preds = %loop.latch, %entry + %i = phi i64 [ 0, %entry ], [ %i.next, %loop.latch ] + %cmp = icmp ult i64 %i, %num_elements + br i1 %cmp, label %loop.latch, label %exit + +loop.latch: ; preds = %loop.header + %gep = getelementptr nusw %"class.dealii::VectorizedArray", ptr %array, i64 %i + %invariant_val = load double, ptr %invariant_ptr, align 8 + %array_val = load double, ptr %gep, align 8 + %sum = fadd double %array_val, %invariant_val + store double %sum, ptr %gep, align 8 + %i.next = add i64 %i, 1 + br label %loop.header + +exit: ; preds = %loop.header + ret void +} diff --git a/llvm/test/Transforms/SLPVectorizer/AArch64/div-like-mixed-with-undefs.ll b/llvm/test/Transforms/SLPVectorizer/AArch64/div-like-mixed-with-undefs.ll index d16843c..6629b12 100644 --- a/llvm/test/Transforms/SLPVectorizer/AArch64/div-like-mixed-with-undefs.ll +++ b/llvm/test/Transforms/SLPVectorizer/AArch64/div-like-mixed-with-undefs.ll @@ -1,21 +1,21 @@ ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 6 ; RUN: opt -passes=slp-vectorizer -S -slp-threshold=-100 -mtriple=aarch64-unknown-linux-gnu < %s | FileCheck %s -define ptr @test(ptr %d) { +define ptr @test(ptr %d, i64 %v) { ; CHECK-LABEL: define ptr @test( -; CHECK-SAME: ptr [[D:%.*]]) { +; CHECK-SAME: ptr [[D:%.*]], i64 [[V:%.*]]) { ; CHECK-NEXT: [[ENTRY:.*:]] -; CHECK-NEXT: [[TMP0:%.*]] = load i8, ptr null, align 1 +; CHECK-NEXT: [[TMP0:%.*]] = load i8, ptr [[D]], align 1 ; CHECK-NEXT: [[CMP4_2:%.*]] = icmp eq i8 [[TMP0]], 0 -; CHECK-NEXT: [[TMP1:%.*]] = select i1 [[CMP4_2]], i64 0, i64 0 -; CHECK-NEXT: [[TMP2:%.*]] = xor i64 0, 0 -; CHECK-NEXT: [[TMP3:%.*]] = udiv i64 [[TMP2]], 0 -; CHECK-NEXT: [[TMP4:%.*]] = udiv i64 1, 0 +; CHECK-NEXT: [[TMP1:%.*]] = select i1 [[CMP4_2]], i64 0, i64 4 +; CHECK-NEXT: [[TMP2:%.*]] = xor i64 0, [[V]] +; CHECK-NEXT: [[TMP3:%.*]] = udiv i64 [[TMP2]], 3 +; CHECK-NEXT: [[TMP4:%.*]] = udiv i64 1, [[V]] ; CHECK-NEXT: [[TMP5:%.*]] = insertelement <6 x i64> poison, i64 [[TMP1]], i32 0 ; CHECK-NEXT: [[TMP6:%.*]] = insertelement <6 x i64> [[TMP5]], i64 [[TMP3]], i32 1 ; CHECK-NEXT: [[TMP7:%.*]] = insertelement <6 x i64> [[TMP6]], i64 [[TMP4]], i32 4 ; CHECK-NEXT: [[TMP8:%.*]] = shufflevector <6 x i64> [[TMP7]], <6 x i64> poison, <6 x i32> <i32 0, i32 1, i32 0, i32 1, i32 4, i32 4> -; CHECK-NEXT: [[TMP9:%.*]] = mul <6 x i64> [[TMP8]], <i64 2, i64 6, i64 1, i64 1, i64 1, i64 0> +; CHECK-NEXT: [[TMP9:%.*]] = mul <6 x i64> [[TMP8]], <i64 2, i64 6, i64 4, i64 3, i64 5, i64 4> ; CHECK-NEXT: [[TMP10:%.*]] = extractelement <6 x i64> [[TMP9]], i32 0 ; CHECK-NEXT: [[TMP11:%.*]] = getelementptr i8, ptr [[D]], i64 [[TMP10]] ; CHECK-NEXT: [[TMP12:%.*]] = extractelement <6 x i64> [[TMP9]], i32 1 @@ -31,23 +31,23 @@ define ptr @test(ptr %d) { ; CHECK-NEXT: ret ptr [[TMP20]] ; entry: - %0 = load i8, ptr null, align 1 + %0 = load i8, ptr %d, align 1 %cmp4.2 = icmp eq i8 %0, 0 - %1 = select i1 %cmp4.2, i64 0, i64 0 + %1 = select i1 %cmp4.2, i64 0, i64 4 %2 = shl i64 %1, 1 %3 = getelementptr i8, ptr %d, i64 %2 - %4 = xor i64 0, 0 - %5 = udiv i64 %4, 0 + %4 = xor i64 0, %v + %5 = udiv i64 %4, 3 %6 = mul i64 %5, 6 %7 = getelementptr i8, ptr %d, i64 %6 - %8 = shl i64 %1, 0 + %8 = shl i64 %1, 2 %scevgep42 = getelementptr i8, ptr %d, i64 %8 - %9 = mul i64 %5, 1 + %9 = mul i64 %5, 3 %10 = getelementptr i8, ptr %d, i64 %9 - %11 = udiv i64 1, 0 - %12 = mul i64 %11, 1 + %11 = udiv i64 1, %v + %12 = mul i64 %11, 5 %13 = getelementptr i8, ptr %d, i64 %12 - %14 = mul i64 %11, 0 + %14 = mul i64 %11, 4 %15 = getelementptr i8, ptr %d, i64 %14 ret ptr %15 } diff --git a/llvm/test/tools/llvm-config/paths.test b/llvm/test/tools/llvm-config/paths.test index 419f155..61d86f7 100644 --- a/llvm/test/tools/llvm-config/paths.test +++ b/llvm/test/tools/llvm-config/paths.test @@ -4,18 +4,34 @@ RUN: llvm-config --bindir 2>&1 | FileCheck --check-prefix=CHECK-BINDIR %s CHECK-BINDIR: {{.*}}{{/|\\}}bin CHECK-BINDIR-NOT: error: CHECK-BINDIR-NOT: warning +RUN: llvm-config --bindir --quote-paths 2>&1 | FileCheck --check-prefix=CHECK-BINDIR2 %s +CHECK-BINDIR2: {{.*}}{{/|\\\\}}bin +CHECK-BINDIR2-NOT: error: +CHECK-BINDIR2-NOT: warning RUN: llvm-config --includedir 2>&1 | FileCheck --check-prefix=CHECK-INCLUDEDIR %s CHECK-INCLUDEDIR: {{.*}}{{/|\\}}include CHECK-INCLUDEDIR-NOT: error: CHECK-INCLUDEDIR-NOT: warning +RUN: llvm-config --includedir --quote-paths 2>&1 | FileCheck --check-prefix=CHECK-INCLUDEDIR2 %s +CHECK-INCLUDEDIR2: {{.*}}{{/|\\\\}}include +CHECK-INCLUDEDIR2-NOT: error: +CHECK-INCLUDEDIR2-NOT: warning RUN: llvm-config --libdir 2>&1 | FileCheck --check-prefix=CHECK-LIBDIR %s CHECK-LIBDIR: {{.*}}{{/|\\}}lib{{.*}} CHECK-LIBDIR-NOT: error: CHECK-LIBDIR-NOT: warning +RUN: llvm-config --libdir --quote-paths 2>&1 | FileCheck --check-prefix=CHECK-LIBDIR2 %s +CHECK-LIBDIR2: {{.*}}{{/|\\\\}}lib{{.*}} +CHECK-LIBDIR2-NOT: error: +CHECK-LIBDIR2-NOT: warning RUN: llvm-config --cmakedir 2>&1 | FileCheck --check-prefix=CHECK-CMAKEDIR %s CHECK-CMAKEDIR: {{.*}}{{/|\\}}cmake{{/|\\}}llvm CHECK-CMAKEDIR-NOT: error: CHECK-CMAKEDIR-NOT: warning +RUN: llvm-config --cmakedir --quote-paths 2>&1 | FileCheck --check-prefix=CHECK-CMAKEDIR2 %s +CHECK-CMAKEDIR2: {{.*}}{{/|\\\\}}cmake{{/|\\\\}}llvm +CHECK-CMAKEDIR2-NOT: error: +CHECK-CMAKEDIR2-NOT: warning diff --git a/llvm/tools/bugpoint/ListReducer.h b/llvm/tools/bugpoint/ListReducer.h index 06f8ddb2..ceee853 100644 --- a/llvm/tools/bugpoint/ListReducer.h +++ b/llvm/tools/bugpoint/ListReducer.h @@ -32,7 +32,7 @@ template <typename ElTy> struct ListReducer { KeepPrefix // The prefix alone satisfies the predicate }; - virtual ~ListReducer() {} + virtual ~ListReducer() = default; /// This virtual function should be overriden by subclasses to implement the /// test desired. The testcase is only required to test to see if the Kept diff --git a/llvm/tools/bugpoint/ToolRunner.h b/llvm/tools/bugpoint/ToolRunner.h index c9da9af..9ff0663 100644 --- a/llvm/tools/bugpoint/ToolRunner.h +++ b/llvm/tools/bugpoint/ToolRunner.h @@ -105,7 +105,7 @@ public: createCustomExecutor(const char *Argv0, std::string &Message, const std::string &ExecCommandLine); - virtual ~AbstractInterpreter() {} + virtual ~AbstractInterpreter() = default; /// compileProgram - Compile the specified program from bitcode to executable /// code. This does not produce any output, it is only used when debugging diff --git a/llvm/tools/dsymutil/BinaryHolder.h b/llvm/tools/dsymutil/BinaryHolder.h index cb5bd95..27d7151 100644 --- a/llvm/tools/dsymutil/BinaryHolder.h +++ b/llvm/tools/dsymutil/BinaryHolder.h @@ -110,7 +110,7 @@ public: std::string Filename; TimestampTy Timestamp; - KeyTy() {} + KeyTy() = default; KeyTy(StringRef Filename, TimestampTy Timestamp) : Filename(Filename.str()), Timestamp(Timestamp) {} }; diff --git a/llvm/tools/lli/lli.cpp b/llvm/tools/lli/lli.cpp index 7fee06b..017e210 100644 --- a/llvm/tools/lli/lli.cpp +++ b/llvm/tools/lli/lli.cpp @@ -305,7 +305,7 @@ public: this->CacheDir[this->CacheDir.size() - 1] != '/') this->CacheDir += '/'; } - ~LLIObjectCache() override {} + ~LLIObjectCache() override = default; void notifyObjectCompiled(const Module *M, MemoryBufferRef Obj) override { const std::string &ModuleID = M->getModuleIdentifier(); diff --git a/llvm/tools/llvm-config/llvm-config.cpp b/llvm/tools/llvm-config/llvm-config.cpp index 020b1b5..5300c5c 100644 --- a/llvm/tools/llvm-config/llvm-config.cpp +++ b/llvm/tools/llvm-config/llvm-config.cpp @@ -24,6 +24,7 @@ #include "llvm/Config/config.h" #include "llvm/Support/FileSystem.h" #include "llvm/Support/Path.h" +#include "llvm/Support/Program.h" #include "llvm/Support/WithColor.h" #include "llvm/Support/raw_ostream.h" #include "llvm/TargetParser/Triple.h" @@ -232,6 +233,7 @@ Options:\n\ --link-static Link the component libraries statically.\n\ --obj-root Print the object root used to build LLVM.\n\ --prefix Print the installation prefix.\n\ + --quote-paths Quote and escape paths when needed.\n\ --shared-mode Print how the provided components can be collectively linked (`shared` or `static`).\n\ --system-libs System Libraries needed to link against LLVM components.\n\ --targets-built List of all targets currently built.\n\ @@ -324,7 +326,7 @@ int main(int argc, char **argv) { // information. std::string ActivePrefix, ActiveBinDir, ActiveIncludeDir, ActiveLibDir, ActiveCMakeDir; - std::string ActiveIncludeOption; + std::vector<std::string> ActiveIncludeOptions; if (IsInDevelopmentTree) { ActiveIncludeDir = std::string(LLVM_SRC_ROOT) + "/include"; ActivePrefix = CurrentExecPrefix; @@ -350,8 +352,8 @@ int main(int argc, char **argv) { } // We need to include files from both the source and object trees. - ActiveIncludeOption = - ("-I" + ActiveIncludeDir + " " + "-I" + ActiveObjRoot + "/include"); + ActiveIncludeOptions.push_back(ActiveIncludeDir); + ActiveIncludeOptions.push_back(ActiveObjRoot + "/include"); } else { ActivePrefix = CurrentExecPrefix; { @@ -370,7 +372,7 @@ int main(int argc, char **argv) { sys::path::make_absolute(ActivePrefix, Path); ActiveCMakeDir = std::string(Path); } - ActiveIncludeOption = "-I" + ActiveIncludeDir; + ActiveIncludeOptions.push_back(ActiveIncludeDir); } /// We only use `shared library` mode in cases where the static library form @@ -399,7 +401,9 @@ int main(int argc, char **argv) { llvm::replace(ActiveBinDir, '/', '\\'); llvm::replace(ActiveLibDir, '/', '\\'); llvm::replace(ActiveCMakeDir, '/', '\\'); - llvm::replace(ActiveIncludeOption, '/', '\\'); + llvm::replace(ActiveIncludeDir, '/', '\\'); + for (auto &Include : ActiveIncludeOptions) + llvm::replace(Include, '/', '\\'); } SharedDir = ActiveBinDir; StaticDir = ActiveLibDir; @@ -501,6 +505,32 @@ int main(int argc, char **argv) { }; raw_ostream &OS = outs(); + + // Check if we want quoting and escaping. + bool QuotePaths = std::any_of(&argv[0], &argv[argc], [](const char *Arg) { + return StringRef(Arg) == "--quote-paths"; + }); + + auto MaybePrintQuoted = [&](StringRef Str) { + if (QuotePaths) + sys::printArg(OS, Str, /*Quote=*/false); // only add quotes if necessary + else + OS << Str; + }; + + // Render include paths and associated flags + auto RenderFlags = [&](StringRef Flags) { + bool First = true; + for (auto &Include : ActiveIncludeOptions) { + if (!First) + OS << ' '; + std::string FlagsStr = "-I" + Include; + MaybePrintQuoted(FlagsStr); + First = false; + } + OS << ' ' << Flags << '\n'; + }; + for (int i = 1; i != argc; ++i) { StringRef Arg = argv[i]; @@ -509,24 +539,32 @@ int main(int argc, char **argv) { if (Arg == "--version") { OS << PACKAGE_VERSION << '\n'; } else if (Arg == "--prefix") { - OS << ActivePrefix << '\n'; + MaybePrintQuoted(ActivePrefix); + OS << '\n'; } else if (Arg == "--bindir") { - OS << ActiveBinDir << '\n'; + MaybePrintQuoted(ActiveBinDir); + OS << '\n'; } else if (Arg == "--includedir") { - OS << ActiveIncludeDir << '\n'; + MaybePrintQuoted(ActiveIncludeDir); + OS << '\n'; } else if (Arg == "--libdir") { - OS << ActiveLibDir << '\n'; + MaybePrintQuoted(ActiveLibDir); + OS << '\n'; } else if (Arg == "--cmakedir") { - OS << ActiveCMakeDir << '\n'; + MaybePrintQuoted(ActiveCMakeDir); + OS << '\n'; } else if (Arg == "--cppflags") { - OS << ActiveIncludeOption << ' ' << LLVM_CPPFLAGS << '\n'; + RenderFlags(LLVM_CPPFLAGS); } else if (Arg == "--cflags") { - OS << ActiveIncludeOption << ' ' << LLVM_CFLAGS << '\n'; + RenderFlags(LLVM_CFLAGS); } else if (Arg == "--cxxflags") { - OS << ActiveIncludeOption << ' ' << LLVM_CXXFLAGS << '\n'; + RenderFlags(LLVM_CXXFLAGS); } else if (Arg == "--ldflags") { - OS << ((HostTriple.isWindowsMSVCEnvironment()) ? "-LIBPATH:" : "-L") - << ActiveLibDir << ' ' << LLVM_LDFLAGS << '\n'; + std::string LDFlags = + HostTriple.isWindowsMSVCEnvironment() ? "-LIBPATH:" : "-L"; + LDFlags += ActiveLibDir; + MaybePrintQuoted(LDFlags); + OS << ' ' << LLVM_LDFLAGS << '\n'; } else if (Arg == "--system-libs") { PrintSystemLibs = true; } else if (Arg == "--libs") { @@ -580,7 +618,8 @@ int main(int argc, char **argv) { } else if (Arg == "--shared-mode") { PrintSharedMode = true; } else if (Arg == "--obj-root") { - OS << ActivePrefix << '\n'; + MaybePrintQuoted(ActivePrefix); + OS << '\n'; } else if (Arg == "--ignore-libllvm") { LinkDyLib = false; LinkMode = BuiltSharedLibs ? LinkModeShared : LinkModeAuto; @@ -590,6 +629,8 @@ int main(int argc, char **argv) { LinkMode = LinkModeStatic; } else if (Arg == "--help") { usage(false); + } else if (Arg == "--quote-paths") { + // Was already handled above this loop. } else { usage(); } @@ -682,26 +723,30 @@ int main(int argc, char **argv) { auto PrintForLib = [&](const StringRef &Lib) { const bool Shared = LinkMode == LinkModeShared; + std::string LibFileName; if (PrintLibNames) { - OS << GetComponentLibraryFileName(Lib, Shared); + LibFileName = GetComponentLibraryFileName(Lib, Shared); } else if (PrintLibFiles) { - OS << GetComponentLibraryPath(Lib, Shared); + LibFileName = GetComponentLibraryPath(Lib, Shared); } else if (PrintLibs) { // On Windows, output full path to library without parameters. // Elsewhere, if this is a typical library name, include it using -l. if (HostTriple.isWindowsMSVCEnvironment()) { - OS << GetComponentLibraryPath(Lib, Shared); + LibFileName = GetComponentLibraryPath(Lib, Shared); } else { + LibFileName = "-l"; StringRef LibName; if (GetComponentLibraryNameSlice(Lib, LibName)) { // Extract library name (remove prefix and suffix). - OS << "-l" << LibName; + LibFileName += LibName; } else { // Lib is already a library name without prefix and suffix. - OS << "-l" << Lib; + LibFileName += Lib; } } } + if (!LibFileName.empty()) + MaybePrintQuoted(LibFileName); }; if (LinkMode == LinkModeShared && LinkDyLib) diff --git a/llvm/tools/llvm-cov/CoverageExporter.h b/llvm/tools/llvm-cov/CoverageExporter.h index 751e55d..ba946a1 100644 --- a/llvm/tools/llvm-cov/CoverageExporter.h +++ b/llvm/tools/llvm-cov/CoverageExporter.h @@ -37,7 +37,7 @@ protected: : Coverage(CoverageMapping), Options(Options), OS(OS) {} public: - virtual ~CoverageExporter(){}; + virtual ~CoverageExporter() = default; /// Render the CoverageMapping object. virtual void renderRoot(const CoverageFilters &IgnoreFilters) = 0; diff --git a/llvm/tools/llvm-cov/CoverageFilters.h b/llvm/tools/llvm-cov/CoverageFilters.h index 5345b0c..3cee23a 100644 --- a/llvm/tools/llvm-cov/CoverageFilters.h +++ b/llvm/tools/llvm-cov/CoverageFilters.h @@ -28,7 +28,7 @@ struct FunctionRecord; /// Matches specific functions that pass the requirement of this filter. class CoverageFilter { public: - virtual ~CoverageFilter() {} + virtual ~CoverageFilter() = default; /// Return true if the function passes the requirements of this filter. virtual bool matches(const coverage::CoverageMapping &CM, diff --git a/llvm/tools/llvm-cov/SourceCoverageView.h b/llvm/tools/llvm-cov/SourceCoverageView.h index 43fb890a..bde187e 100644 --- a/llvm/tools/llvm-cov/SourceCoverageView.h +++ b/llvm/tools/llvm-cov/SourceCoverageView.h @@ -122,7 +122,7 @@ public: static std::unique_ptr<CoveragePrinter> create(const CoverageViewOptions &Opts); - virtual ~CoveragePrinter() {} + virtual ~CoveragePrinter() = default; /// @name File Creation Interface /// @{ @@ -288,7 +288,7 @@ public: create(StringRef SourceName, const MemoryBuffer &File, const CoverageViewOptions &Options, CoverageData &&CoverageInfo); - virtual ~SourceCoverageView() {} + virtual ~SourceCoverageView() = default; /// Return the source name formatted for the host OS. std::string getSourceName() const; diff --git a/llvm/tools/llvm-diff/lib/DiffConsumer.h b/llvm/tools/llvm-diff/lib/DiffConsumer.h index 08c3afc..d4f339b 100644 --- a/llvm/tools/llvm-diff/lib/DiffConsumer.h +++ b/llvm/tools/llvm-diff/lib/DiffConsumer.h @@ -49,7 +49,7 @@ class StringRef; virtual void logd(const DiffLogBuilder &Log) = 0; protected: - virtual ~Consumer() {} + virtual ~Consumer() = default; }; class DiffConsumer : public Consumer { diff --git a/llvm/tools/llvm-diff/lib/DifferenceEngine.h b/llvm/tools/llvm-diff/lib/DifferenceEngine.h index 436a355..b829b2c 100644 --- a/llvm/tools/llvm-diff/lib/DifferenceEngine.h +++ b/llvm/tools/llvm-diff/lib/DifferenceEngine.h @@ -54,7 +54,7 @@ namespace llvm { virtual bool operator()(const Value *L, const Value *R) = 0; protected: - virtual ~Oracle() {} + virtual ~Oracle() = default; }; DifferenceEngine(Consumer &consumer) diff --git a/llvm/tools/llvm-exegesis/lib/AArch64/Target.cpp b/llvm/tools/llvm-exegesis/lib/AArch64/Target.cpp index 2c13dd5..0e73ada 100644 --- a/llvm/tools/llvm-exegesis/lib/AArch64/Target.cpp +++ b/llvm/tools/llvm-exegesis/lib/AArch64/Target.cpp @@ -112,7 +112,7 @@ namespace { // Use X19 as the loop counter register since it's a callee-saved register // that's available for temporary use. -constexpr const MCPhysReg kDefaultLoopCounterReg = AArch64::X19; +constexpr MCPhysReg kDefaultLoopCounterReg = AArch64::X19; class ExegesisAArch64Target : public ExegesisTarget { public: diff --git a/llvm/tools/llvm-exegesis/lib/Analysis.cpp b/llvm/tools/llvm-exegesis/lib/Analysis.cpp index fb84328..f3bf9690 100644 --- a/llvm/tools/llvm-exegesis/lib/Analysis.cpp +++ b/llvm/tools/llvm-exegesis/lib/Analysis.cpp @@ -446,7 +446,7 @@ void Analysis::printClusterRawHtml(const BenchmarkClustering::ClusterId &Id, } // namespace exegesis -static constexpr const char kHtmlHead[] = R"( +static constexpr char kHtmlHead[] = R"( <head> <title>llvm-exegesis Analysis Results</title> <style> diff --git a/llvm/tools/llvm-exegesis/lib/Assembler.cpp b/llvm/tools/llvm-exegesis/lib/Assembler.cpp index fd7924d..163f141 100644 --- a/llvm/tools/llvm-exegesis/lib/Assembler.cpp +++ b/llvm/tools/llvm-exegesis/lib/Assembler.cpp @@ -44,8 +44,8 @@ namespace llvm { namespace exegesis { -static constexpr const char ModuleID[] = "ExegesisInfoTest"; -static constexpr const char FunctionID[] = "foo"; +static constexpr char ModuleID[] = "ExegesisInfoTest"; +static constexpr char FunctionID[] = "foo"; static const Align kFunctionAlignment(4096); // Fills the given basic block with register setup code, and returns true if diff --git a/llvm/tools/llvm-exegesis/lib/BenchmarkResult.cpp b/llvm/tools/llvm-exegesis/lib/BenchmarkResult.cpp index 1823a53..c6164b6 100644 --- a/llvm/tools/llvm-exegesis/lib/BenchmarkResult.cpp +++ b/llvm/tools/llvm-exegesis/lib/BenchmarkResult.cpp @@ -21,9 +21,9 @@ #include "llvm/Support/Format.h" #include "llvm/Support/raw_ostream.h" -static constexpr const char kIntegerPrefix[] = "i_0x"; -static constexpr const char kDoublePrefix[] = "f_"; -static constexpr const char kInvalidOperand[] = "INVALID"; +static constexpr char kIntegerPrefix[] = "i_0x"; +static constexpr char kDoublePrefix[] = "f_"; +static constexpr char kInvalidOperand[] = "INVALID"; namespace llvm { @@ -202,7 +202,7 @@ struct CustomMappingTraits<std::map<exegesis::ValidationEvent, int64_t>> { Io.setError("Key is not a valid validation event"); return; } - Io.mapRequired(KeyStr.str().c_str(), VI[*Key]); + Io.mapRequired(KeyStr, VI[*Key]); } static void output(IO &Io, std::map<exegesis::ValidationEvent, int64_t> &VI) { @@ -245,8 +245,8 @@ template <> struct SequenceElementTraits<exegesis::RegisterValue> { }; template <> struct ScalarTraits<exegesis::RegisterValue> { - static constexpr const unsigned kRadix = 16; - static constexpr const bool kSigned = false; + static constexpr unsigned kRadix = 16; + static constexpr bool kSigned = false; static void output(const exegesis::RegisterValue &RV, void *Ctx, raw_ostream &Out) { diff --git a/llvm/tools/llvm-exegesis/lib/BenchmarkRunner.cpp b/llvm/tools/llvm-exegesis/lib/BenchmarkRunner.cpp index 1fd0a15..12fad7d 100644 --- a/llvm/tools/llvm-exegesis/lib/BenchmarkRunner.cpp +++ b/llvm/tools/llvm-exegesis/lib/BenchmarkRunner.cpp @@ -155,7 +155,7 @@ private: #ifdef LLVM_ON_UNIX // See "Exit Status for Commands": // https://pubs.opengroup.org/onlinepubs/9699919799/xrat/V4_xcu_chap02.html - constexpr const int kSigOffset = 128; + constexpr int kSigOffset = 128; return make_error<SnippetSignal>(CRC.RetCode - kSigOffset); #else // The exit code of the process on windows is not meaningful as a @@ -877,7 +877,7 @@ Error BenchmarkRunner::getValidationCountersToRun( return Error::success(); } -BenchmarkRunner::FunctionExecutor::~FunctionExecutor() {} +BenchmarkRunner::FunctionExecutor::~FunctionExecutor() = default; } // namespace exegesis } // namespace llvm diff --git a/llvm/tools/llvm-exegesis/lib/BenchmarkRunner.h b/llvm/tools/llvm-exegesis/lib/BenchmarkRunner.h index e688b81..16d3c9c 100644 --- a/llvm/tools/llvm-exegesis/lib/BenchmarkRunner.h +++ b/llvm/tools/llvm-exegesis/lib/BenchmarkRunner.h @@ -73,8 +73,8 @@ public: // Scratch space to run instructions that touch memory. struct ScratchSpace { - static constexpr const size_t kAlignment = 1024; - static constexpr const size_t kSize = 1 << 20; // 1MB. + static constexpr size_t kAlignment = 1024; + static constexpr size_t kSize = 1 << 20; // 1MB. ScratchSpace() : UnalignedPtr(std::make_unique<char[]>(kSize + kAlignment)), AlignedPtr( diff --git a/llvm/tools/llvm-exegesis/lib/Clustering.h b/llvm/tools/llvm-exegesis/lib/Clustering.h index 9d6c110..2b0f5b4 100644 --- a/llvm/tools/llvm-exegesis/lib/Clustering.h +++ b/llvm/tools/llvm-exegesis/lib/Clustering.h @@ -67,11 +67,11 @@ public: ClusterId(size_t Id, bool IsUnstable = false) : Id_(Id), IsUnstable_(IsUnstable) {} - static constexpr const size_t kMaxValid = + static constexpr size_t kMaxValid = (std::numeric_limits<size_t>::max() >> 1) - 4; - static constexpr const size_t kNoise = kMaxValid + 1; - static constexpr const size_t kError = kMaxValid + 2; - static constexpr const size_t kUndef = kMaxValid + 3; + static constexpr size_t kNoise = kMaxValid + 1; + static constexpr size_t kError = kMaxValid + 2; + static constexpr size_t kUndef = kMaxValid + 3; size_t Id_ : (std::numeric_limits<size_t>::digits - 1); size_t IsUnstable_ : 1; diff --git a/llvm/tools/llvm-exegesis/lib/Error.h b/llvm/tools/llvm-exegesis/lib/Error.h index 9b71fe8..c899023 100644 --- a/llvm/tools/llvm-exegesis/lib/Error.h +++ b/llvm/tools/llvm-exegesis/lib/Error.h @@ -81,7 +81,7 @@ private: struct PerfCounterNotFullyEnabled : public ErrorInfo<PerfCounterNotFullyEnabled> { static char ID; - PerfCounterNotFullyEnabled() {} + PerfCounterNotFullyEnabled() = default; void log(raw_ostream &OS) const override; diff --git a/llvm/tools/llvm-exegesis/lib/ParallelSnippetGenerator.cpp b/llvm/tools/llvm-exegesis/lib/ParallelSnippetGenerator.cpp index 79a585e..aa40913 100644 --- a/llvm/tools/llvm-exegesis/lib/ParallelSnippetGenerator.cpp +++ b/llvm/tools/llvm-exegesis/lib/ParallelSnippetGenerator.cpp @@ -350,7 +350,5 @@ ParallelSnippetGenerator::generateCodeTemplates( return Result; } -constexpr const size_t ParallelSnippetGenerator::kMinNumDifferentAddresses; - } // namespace exegesis } // namespace llvm diff --git a/llvm/tools/llvm-exegesis/lib/ParallelSnippetGenerator.h b/llvm/tools/llvm-exegesis/lib/ParallelSnippetGenerator.h index 8a6b856..d3c85c0 100644 --- a/llvm/tools/llvm-exegesis/lib/ParallelSnippetGenerator.h +++ b/llvm/tools/llvm-exegesis/lib/ParallelSnippetGenerator.h @@ -28,7 +28,7 @@ public: generateCodeTemplates(InstructionTemplate Variant, const BitVector &ForbiddenRegisters) const override; - static constexpr const size_t kMinNumDifferentAddresses = 6; + static constexpr size_t kMinNumDifferentAddresses = 6; private: // Instantiates memory operands within a snippet. diff --git a/llvm/tools/llvm-exegesis/lib/SnippetRepetitor.cpp b/llvm/tools/llvm-exegesis/lib/SnippetRepetitor.cpp index 80f5ce4..37dcc7c 100644 --- a/llvm/tools/llvm-exegesis/lib/SnippetRepetitor.cpp +++ b/llvm/tools/llvm-exegesis/lib/SnippetRepetitor.cpp @@ -131,7 +131,7 @@ private: } // namespace -SnippetRepetitor::~SnippetRepetitor() {} +SnippetRepetitor::~SnippetRepetitor() = default; std::unique_ptr<const SnippetRepetitor> SnippetRepetitor::Create(Benchmark::RepetitionModeE Mode, diff --git a/llvm/tools/llvm-exegesis/lib/SubprocessMemory.h b/llvm/tools/llvm-exegesis/lib/SubprocessMemory.h index 572d108..52ee980 100644 --- a/llvm/tools/llvm-exegesis/lib/SubprocessMemory.h +++ b/llvm/tools/llvm-exegesis/lib/SubprocessMemory.h @@ -32,8 +32,8 @@ namespace exegesis { class SubprocessMemory { public: - static constexpr const size_t AuxiliaryMemoryOffset = 1; - static constexpr const size_t AuxiliaryMemorySize = 4096; + static constexpr size_t AuxiliaryMemoryOffset = 1; + static constexpr size_t AuxiliaryMemorySize = 4096; // Gets the thread ID for the calling thread. static long getCurrentTID(); diff --git a/llvm/tools/llvm-exegesis/lib/Target.cpp b/llvm/tools/llvm-exegesis/lib/Target.cpp index fc5f82f..2ad6c5a 100644 --- a/llvm/tools/llvm-exegesis/lib/Target.cpp +++ b/llvm/tools/llvm-exegesis/lib/Target.cpp @@ -23,7 +23,7 @@ cl::OptionCategory Options("llvm-exegesis options"); cl::OptionCategory BenchmarkOptions("llvm-exegesis benchmark options"); cl::OptionCategory AnalysisOptions("llvm-exegesis analysis options"); -ExegesisTarget::~ExegesisTarget() {} // anchor. +ExegesisTarget::~ExegesisTarget() = default; // anchor. static ExegesisTarget *FirstTarget = nullptr; @@ -215,7 +215,7 @@ const PfmCountersInfo &ExegesisTarget::getDummyPfmCounters() const { return PfmCountersInfo::Dummy; } -ExegesisTarget::SavedState::~SavedState() {} // anchor. +ExegesisTarget::SavedState::~SavedState() = default; // anchor. namespace { diff --git a/llvm/tools/llvm-exegesis/lib/UopsBenchmarkRunner.h b/llvm/tools/llvm-exegesis/lib/UopsBenchmarkRunner.h index ef47b7f..74a18da 100644 --- a/llvm/tools/llvm-exegesis/lib/UopsBenchmarkRunner.h +++ b/llvm/tools/llvm-exegesis/lib/UopsBenchmarkRunner.h @@ -30,7 +30,7 @@ public: ExecutionMode, ValCounters) {} ~UopsBenchmarkRunner() override; - static constexpr const size_t kMinNumDifferentAddresses = 6; + static constexpr size_t kMinNumDifferentAddresses = 6; private: Expected<std::vector<BenchmarkMeasure>> diff --git a/llvm/tools/llvm-exegesis/lib/X86/Target.cpp b/llvm/tools/llvm-exegesis/lib/X86/Target.cpp index b4437f7..6dc64765 100644 --- a/llvm/tools/llvm-exegesis/lib/X86/Target.cpp +++ b/llvm/tools/llvm-exegesis/lib/X86/Target.cpp @@ -278,9 +278,9 @@ static Expected<std::vector<CodeTemplate>> generateLEATemplatesCommon( assert(X86II::getMemoryOperandNo(Instr.Description.TSFlags) == 1 && "invalid LEA"); - constexpr const int kDestOp = 0; - constexpr const int kBaseOp = 1; - constexpr const int kIndexOp = 3; + constexpr int kDestOp = 0; + constexpr int kBaseOp = 1; + constexpr int kIndexOp = 3; auto PossibleDestRegs = Instr.Operands[kDestOp].getRegisterAliasing().sourceBits(); remove(PossibleDestRegs, ForbiddenRegisters); @@ -548,7 +548,7 @@ private: void initStack(unsigned Bytes); - static constexpr const unsigned kF80Bytes = 10; // 80 bits. + static constexpr unsigned kF80Bytes = 10; // 80 bits. APInt Constant_; std::vector<MCInst> Instructions; @@ -864,7 +864,7 @@ const MCPhysReg ExegesisX86Target::kUnavailableRegistersSSE[12] = { // We're using one of R8-R15 because these registers are never hardcoded in // instructions (e.g. MOVS writes to EDI, ESI, EDX), so they have less // conflicts. -constexpr const MCPhysReg kDefaultLoopCounterReg = X86::R8; +constexpr MCPhysReg kDefaultLoopCounterReg = X86::R8; } // namespace @@ -1110,9 +1110,9 @@ std::vector<MCInst> ExegesisX86Target::setRegTo(const MCSubtargetInfo &STI, #ifdef __linux__ #ifdef __arm__ -static constexpr const uintptr_t VAddressSpaceCeiling = 0xC0000000; +static constexpr uintptr_t VAddressSpaceCeiling = 0xC0000000; #else -static constexpr const uintptr_t VAddressSpaceCeiling = 0x0000800000000000; +static constexpr uintptr_t VAddressSpaceCeiling = 0x0000800000000000; #endif void generateRoundToNearestPage(unsigned int Register, diff --git a/llvm/tools/llvm-libtool-darwin/DependencyInfo.h b/llvm/tools/llvm-libtool-darwin/DependencyInfo.h index 784ec3f..80bad8f 100644 --- a/llvm/tools/llvm-libtool-darwin/DependencyInfo.h +++ b/llvm/tools/llvm-libtool-darwin/DependencyInfo.h @@ -18,7 +18,7 @@ public: explicit DependencyInfo(std::string DependencyInfoPath) : DependencyInfoPath(DependencyInfoPath) {} - virtual ~DependencyInfo(){}; + virtual ~DependencyInfo() = default; virtual void addMissingInput(llvm::StringRef Path) { NotFounds.insert(Path.str()); diff --git a/llvm/tools/llvm-mca/CodeRegionGenerator.cpp b/llvm/tools/llvm-mca/CodeRegionGenerator.cpp index f7f929e..1469071 100644 --- a/llvm/tools/llvm-mca/CodeRegionGenerator.cpp +++ b/llvm/tools/llvm-mca/CodeRegionGenerator.cpp @@ -26,7 +26,7 @@ namespace llvm { namespace mca { // This virtual dtor serves as the anchor for the CodeRegionGenerator class. -CodeRegionGenerator::~CodeRegionGenerator() {} +CodeRegionGenerator::~CodeRegionGenerator() = default; Expected<const CodeRegions &> AsmCodeRegionGenerator::parseCodeRegions( const std::unique_ptr<MCInstPrinter> &IP, bool SkipFailures) { diff --git a/llvm/tools/llvm-mca/CodeRegionGenerator.h b/llvm/tools/llvm-mca/CodeRegionGenerator.h index a48c67a..c30f67a 100644 --- a/llvm/tools/llvm-mca/CodeRegionGenerator.h +++ b/llvm/tools/llvm-mca/CodeRegionGenerator.h @@ -151,7 +151,7 @@ protected: bool SkipFailures) = 0; public: - CodeRegionGenerator() {} + CodeRegionGenerator() = default; virtual ~CodeRegionGenerator(); }; diff --git a/llvm/tools/llvm-objcopy/ObjcopyOptions.cpp b/llvm/tools/llvm-objcopy/ObjcopyOptions.cpp index 3d7f33c..8aa843b 100644 --- a/llvm/tools/llvm-objcopy/ObjcopyOptions.cpp +++ b/llvm/tools/llvm-objcopy/ObjcopyOptions.cpp @@ -811,12 +811,12 @@ objcopy::parseObjcopyOptions(ArrayRef<const char *> ArgsArr, .Case("boot_application", COFF::IMAGE_SUBSYSTEM_WINDOWS_BOOT_APPLICATION) .Case("console", COFF::IMAGE_SUBSYSTEM_WINDOWS_CUI) - .Cases("efi_application", "efi-app", + .Cases({"efi_application", "efi-app"}, COFF::IMAGE_SUBSYSTEM_EFI_APPLICATION) - .Cases("efi_boot_service_driver", "efi-bsd", + .Cases({"efi_boot_service_driver", "efi-bsd"}, COFF::IMAGE_SUBSYSTEM_EFI_BOOT_SERVICE_DRIVER) .Case("efi_rom", COFF::IMAGE_SUBSYSTEM_EFI_ROM) - .Cases("efi_runtime_driver", "efi-rtd", + .Cases({"efi_runtime_driver", "efi-rtd"}, COFF::IMAGE_SUBSYSTEM_EFI_RUNTIME_DRIVER) .Case("native", COFF::IMAGE_SUBSYSTEM_NATIVE) .Case("posix", COFF::IMAGE_SUBSYSTEM_POSIX_CUI) diff --git a/llvm/tools/llvm-objdump/SourcePrinter.h b/llvm/tools/llvm-objdump/SourcePrinter.h index 5c131a0..19acc87 100644 --- a/llvm/tools/llvm-objdump/SourcePrinter.h +++ b/llvm/tools/llvm-objdump/SourcePrinter.h @@ -34,7 +34,7 @@ public: LiveElement(const char *Name, DWARFUnit *Unit, const DWARFDie FuncDie) : Name(Name), Unit(Unit), FuncDie(FuncDie) {} - virtual ~LiveElement() {}; + virtual ~LiveElement() = default; const char *getName() const { return Name; } virtual bool liveAtAddress(object::SectionedAddress Addr) const = 0; diff --git a/llvm/tools/llvm-objdump/llvm-objdump.h b/llvm/tools/llvm-objdump/llvm-objdump.h index 3525be9..bac8589 100644 --- a/llvm/tools/llvm-objdump/llvm-objdump.h +++ b/llvm/tools/llvm-objdump/llvm-objdump.h @@ -84,7 +84,7 @@ protected: public: Dumper(const object::ObjectFile &O); - virtual ~Dumper() {} + virtual ~Dumper() = default; void reportUniqueWarning(Error Err); void reportUniqueWarning(const Twine &Msg); diff --git a/llvm/tools/llvm-pdbutil/DumpOutputStyle.cpp b/llvm/tools/llvm-pdbutil/DumpOutputStyle.cpp index b2362ec..d836d98 100644 --- a/llvm/tools/llvm-pdbutil/DumpOutputStyle.cpp +++ b/llvm/tools/llvm-pdbutil/DumpOutputStyle.cpp @@ -68,7 +68,7 @@ DumpOutputStyle::DumpOutputStyle(InputFile &File) RefTracker.reset(new TypeReferenceTracker(File)); } -DumpOutputStyle::~DumpOutputStyle() {} +DumpOutputStyle::~DumpOutputStyle() = default; PDBFile &DumpOutputStyle::getPdb() { return File.pdb(); } object::COFFObjectFile &DumpOutputStyle::getObj() { return File.obj(); } diff --git a/llvm/tools/llvm-pdbutil/DumpOutputStyle.h b/llvm/tools/llvm-pdbutil/DumpOutputStyle.h index 6714a6a..ea4a47f 100644 --- a/llvm/tools/llvm-pdbutil/DumpOutputStyle.h +++ b/llvm/tools/llvm-pdbutil/DumpOutputStyle.h @@ -29,7 +29,7 @@ class TypeReferenceTracker; struct StatCollection { struct Stat { - Stat() {} + Stat() = default; Stat(uint32_t Count, uint32_t Size) : Count(Count), Size(Size) {} uint32_t Count = 0; uint32_t Size = 0; diff --git a/llvm/tools/llvm-pdbutil/OutputStyle.h b/llvm/tools/llvm-pdbutil/OutputStyle.h index 8cc9016..a09fb82 100644 --- a/llvm/tools/llvm-pdbutil/OutputStyle.h +++ b/llvm/tools/llvm-pdbutil/OutputStyle.h @@ -17,7 +17,7 @@ namespace pdb { class OutputStyle { public: - virtual ~OutputStyle() {} + virtual ~OutputStyle() = default; virtual Error dump() = 0; }; diff --git a/llvm/tools/llvm-pdbutil/StreamUtil.h b/llvm/tools/llvm-pdbutil/StreamUtil.h index 9d6030c..6b8c13f 100644 --- a/llvm/tools/llvm-pdbutil/StreamUtil.h +++ b/llvm/tools/llvm-pdbutil/StreamUtil.h @@ -35,7 +35,7 @@ enum class StreamPurpose { struct StreamInfo { public: - StreamInfo() {} + StreamInfo() = default; uint32_t getModuleIndex() const { return *ModuleIndex; } StreamPurpose getPurpose() const { return Purpose; } diff --git a/llvm/tools/llvm-profgen/ProfiledBinary.cpp b/llvm/tools/llvm-profgen/ProfiledBinary.cpp index 94728ce..96db6a7 100644 --- a/llvm/tools/llvm-profgen/ProfiledBinary.cpp +++ b/llvm/tools/llvm-profgen/ProfiledBinary.cpp @@ -187,7 +187,7 @@ ProfiledBinary::ProfiledBinary(const StringRef ExeBinPath, load(); } -ProfiledBinary::~ProfiledBinary() {} +ProfiledBinary::~ProfiledBinary() = default; void ProfiledBinary::warnNoFuncEntry() { uint64_t NoFuncEntryNum = 0; diff --git a/llvm/tools/llvm-rc/ResourceScriptStmt.h b/llvm/tools/llvm-rc/ResourceScriptStmt.h index a81e384f..84da9be 100644 --- a/llvm/tools/llvm-rc/ResourceScriptStmt.h +++ b/llvm/tools/llvm-rc/ResourceScriptStmt.h @@ -242,9 +242,9 @@ public: virtual raw_ostream &log(raw_ostream &OS) const { return OS << "Base statement\n"; }; - RCResource() {} + RCResource() = default; RCResource(uint16_t Flags) : MemoryFlags(Flags) {} - virtual ~RCResource() {} + virtual ~RCResource() = default; virtual Error visit(Visitor *) const { llvm_unreachable("This is unable to call methods from Visitor base"); @@ -290,7 +290,7 @@ class OptionalStmtList : public OptionalStmt { std::vector<std::unique_ptr<OptionalStmt>> Statements; public: - OptionalStmtList() {} + OptionalStmtList() = default; raw_ostream &log(raw_ostream &OS) const override; void addStmt(std::unique_ptr<OptionalStmt> Stmt) { @@ -510,7 +510,7 @@ public: virtual raw_ostream &log(raw_ostream &OS) const { return OS << "Base menu definition\n"; } - virtual ~MenuDefinition() {} + virtual ~MenuDefinition() = default; virtual uint16_t getResFlags() const { return 0; } virtual MenuDefKind getKind() const { return MkBase; } @@ -818,7 +818,7 @@ public: enum StmtKind { StBase = 0, StBlock = 1, StValue = 2 }; virtual raw_ostream &log(raw_ostream &OS) const { return OS << "VI stmt\n"; } - virtual ~VersionInfoStmt() {} + virtual ~VersionInfoStmt() = default; virtual StmtKind getKind() const { return StBase; } static bool classof(const VersionInfoStmt *S) { diff --git a/llvm/tools/llvm-rc/ResourceVisitor.h b/llvm/tools/llvm-rc/ResourceVisitor.h index a121a0a..1815c6b 100644 --- a/llvm/tools/llvm-rc/ResourceVisitor.h +++ b/llvm/tools/llvm-rc/ResourceVisitor.h @@ -55,7 +55,7 @@ public: virtual Error visitVersionStmt(const VersionStmt *) = 0; virtual Error visitMenuStmt(const MenuStmt *) = 0; - virtual ~Visitor() {} + virtual ~Visitor() = default; }; } // namespace rc diff --git a/llvm/tools/llvm-readobj/ObjDumper.cpp b/llvm/tools/llvm-readobj/ObjDumper.cpp index 0b59dd4..20e027a 100644 --- a/llvm/tools/llvm-readobj/ObjDumper.cpp +++ b/llvm/tools/llvm-readobj/ObjDumper.cpp @@ -41,7 +41,7 @@ ObjDumper::ObjDumper(ScopedPrinter &Writer, StringRef ObjName) : W(Writer) { }; } -ObjDumper::~ObjDumper() {} +ObjDumper::~ObjDumper() = default; void ObjDumper::reportUniqueWarning(Error Err) const { reportUniqueWarning(toString(std::move(Err))); diff --git a/llvm/tools/llvm-readtapi/DiffEngine.h b/llvm/tools/llvm-readtapi/DiffEngine.h index 7ab57d4..b350ceef 100644 --- a/llvm/tools/llvm-readtapi/DiffEngine.h +++ b/llvm/tools/llvm-readtapi/DiffEngine.h @@ -39,7 +39,7 @@ enum DiffAttrKind { class AttributeDiff { public: AttributeDiff(DiffAttrKind Kind) : Kind(Kind){}; - virtual ~AttributeDiff(){}; + virtual ~AttributeDiff() = default; DiffAttrKind getKind() const { return Kind; } private: diff --git a/llvm/tools/llvm-xray/xray-graph.h b/llvm/tools/llvm-xray/xray-graph.h index fd96449..bf25f8d 100644 --- a/llvm/tools/llvm-xray/xray-graph.h +++ b/llvm/tools/llvm-xray/xray-graph.h @@ -86,7 +86,7 @@ public: }; GraphT G; - using VertexIdentifier = typename decltype(G)::VertexIdentifier; + using VertexIdentifier = decltype(G)::VertexIdentifier; using EdgeIdentifier = decltype(G)::EdgeIdentifier; /// Use a Map to store the Function stack for each thread whilst building the diff --git a/llvm/unittests/ADT/BreadthFirstIteratorTest.cpp b/llvm/unittests/ADT/BreadthFirstIteratorTest.cpp index 0cd7fd3..a737390 100644 --- a/llvm/unittests/ADT/BreadthFirstIteratorTest.cpp +++ b/llvm/unittests/ADT/BreadthFirstIteratorTest.cpp @@ -78,7 +78,7 @@ TEST(BreadthFristIteratorTest, Cycle) { static_assert( std::is_convertible_v<decltype(*std::declval<bf_iterator<Graph<3>>>()), - typename bf_iterator<Graph<3>>::reference>); + bf_iterator<Graph<3>>::reference>); // bf_iterator should be (at-least) a forward-iterator static_assert(std::is_base_of_v<std::forward_iterator_tag, diff --git a/llvm/unittests/ADT/ConcurrentHashtableTest.cpp b/llvm/unittests/ADT/ConcurrentHashtableTest.cpp index ee1ee41..1b82df1 100644 --- a/llvm/unittests/ADT/ConcurrentHashtableTest.cpp +++ b/llvm/unittests/ADT/ConcurrentHashtableTest.cpp @@ -21,7 +21,7 @@ using namespace parallel; namespace { class String { public: - String() {} + String() = default; const std::string &getKey() const { return Data; } template <typename AllocatorTy> diff --git a/llvm/unittests/ADT/DepthFirstIteratorTest.cpp b/llvm/unittests/ADT/DepthFirstIteratorTest.cpp index 95923b8..f792878 100644 --- a/llvm/unittests/ADT/DepthFirstIteratorTest.cpp +++ b/llvm/unittests/ADT/DepthFirstIteratorTest.cpp @@ -59,7 +59,7 @@ TEST(DepthFirstIteratorTest, ActuallyUpdateIterator) { static_assert( std::is_convertible_v<decltype(*std::declval<df_iterator<Graph<3>>>()), - typename df_iterator<Graph<3>>::reference>); + df_iterator<Graph<3>>::reference>); // df_iterator should be (at-least) a forward-iterator static_assert(std::is_base_of_v<std::forward_iterator_tag, diff --git a/llvm/unittests/ADT/DirectedGraphTest.cpp b/llvm/unittests/ADT/DirectedGraphTest.cpp index 49ccf06..82a631b 100644 --- a/llvm/unittests/ADT/DirectedGraphTest.cpp +++ b/llvm/unittests/ADT/DirectedGraphTest.cpp @@ -43,7 +43,7 @@ public: class DGTestGraph : public DGTestBase { public: DGTestGraph() = default; - ~DGTestGraph(){}; + ~DGTestGraph() = default; }; using EdgeListTy = SmallVector<DGTestEdge *, 2>; diff --git a/llvm/unittests/ADT/IListIteratorBitsTest.cpp b/llvm/unittests/ADT/IListIteratorBitsTest.cpp index 97c14265..b430bcb 100644 --- a/llvm/unittests/ADT/IListIteratorBitsTest.cpp +++ b/llvm/unittests/ADT/IListIteratorBitsTest.cpp @@ -93,8 +93,8 @@ TEST(IListIteratorBitsTest, ConsAndAssignment) { class dummy { // Test that we get an ilist_iterator_w_bits out of the node given that the // options are enabled. - using node_options = typename ilist_detail::compute_node_options< - Node, ilist_iterator_bits<true>>::type; + using node_options = + ilist_detail::compute_node_options<Node, ilist_iterator_bits<true>>::type; static_assert(std::is_same<Node::self_iterator, llvm::ilist_iterator_w_bits<node_options, false, false>>::value); @@ -102,7 +102,7 @@ class dummy { // Now test that a plain node, without the option, gets a plain // ilist_iterator. using plain_node_options = - typename ilist_detail::compute_node_options<PlainNode>::type; + ilist_detail::compute_node_options<PlainNode>::type; static_assert(std::is_same< PlainNode::self_iterator, llvm::ilist_iterator<plain_node_options, false, false>>::value); diff --git a/llvm/unittests/ADT/IListTest.cpp b/llvm/unittests/ADT/IListTest.cpp index 2fdc8e1..984014f 100644 --- a/llvm/unittests/ADT/IListTest.cpp +++ b/llvm/unittests/ADT/IListTest.cpp @@ -19,7 +19,7 @@ namespace { struct Node : ilist_node<Node> { int Value; - Node() {} + Node() = default; Node(int Value) : Value(Value) {} Node(const Node&) = default; ~Node() { Value = -1; } diff --git a/llvm/unittests/ADT/IteratorTest.cpp b/llvm/unittests/ADT/IteratorTest.cpp index 691fbce..b5d63ef 100644 --- a/llvm/unittests/ADT/IteratorTest.cpp +++ b/llvm/unittests/ADT/IteratorTest.cpp @@ -48,11 +48,10 @@ struct AdaptedIter : iterator_adaptor_base<AdaptedIter, WeirdIter> {}; // Test that iterator_adaptor_base forwards typedefs, if value_type is // unchanged. -static_assert(std::is_same_v<typename AdaptedIter::value_type, Shadow<0>>, ""); -static_assert(std::is_same_v<typename AdaptedIter::difference_type, Shadow<1>>, - ""); -static_assert(std::is_same_v<typename AdaptedIter::pointer, Shadow<2>>, ""); -static_assert(std::is_same_v<typename AdaptedIter::reference, Shadow<3>>, ""); +static_assert(std::is_same_v<AdaptedIter::value_type, Shadow<0>>, ""); +static_assert(std::is_same_v<AdaptedIter::difference_type, Shadow<1>>, ""); +static_assert(std::is_same_v<AdaptedIter::pointer, Shadow<2>>, ""); +static_assert(std::is_same_v<AdaptedIter::reference, Shadow<3>>, ""); // Ensure that pointe{e,r}_iterator adaptors correctly forward the category of // the underlying iterator. diff --git a/llvm/unittests/ADT/PostOrderIteratorTest.cpp b/llvm/unittests/ADT/PostOrderIteratorTest.cpp index 4c2a66e..838481f 100644 --- a/llvm/unittests/ADT/PostOrderIteratorTest.cpp +++ b/llvm/unittests/ADT/PostOrderIteratorTest.cpp @@ -44,7 +44,7 @@ TEST(PostOrderIteratorTest, Compiles) { static_assert( std::is_convertible_v<decltype(*std::declval<po_iterator<Graph<3>>>()), - typename po_iterator<Graph<3>>::reference>); + po_iterator<Graph<3>>::reference>); // Test post-order and reverse post-order traversals for simple graph type. TEST(PostOrderIteratorTest, PostOrderAndReversePostOrderTraverrsal) { diff --git a/llvm/unittests/ADT/SmallVectorTest.cpp b/llvm/unittests/ADT/SmallVectorTest.cpp index 1a01f30..74fc737 100644 --- a/llvm/unittests/ADT/SmallVectorTest.cpp +++ b/llvm/unittests/ADT/SmallVectorTest.cpp @@ -159,7 +159,7 @@ int Constructable::numCopyAssignmentCalls; int Constructable::numMoveAssignmentCalls; struct NonCopyable { - NonCopyable() {} + NonCopyable() = default; NonCopyable(NonCopyable &&) {} NonCopyable &operator=(NonCopyable &&) { return *this; } private: diff --git a/llvm/unittests/ADT/StringMapTest.cpp b/llvm/unittests/ADT/StringMapTest.cpp index 92ae364..1d92de4 100644 --- a/llvm/unittests/ADT/StringMapTest.cpp +++ b/llvm/unittests/ADT/StringMapTest.cpp @@ -367,7 +367,7 @@ TEST_F(StringMapTest, NonDefaultConstructable) { } struct Immovable { - Immovable() {} + Immovable() = default; Immovable(Immovable &&) = delete; // will disable the other special members }; diff --git a/llvm/unittests/ADT/TypeSwitchTest.cpp b/llvm/unittests/ADT/TypeSwitchTest.cpp index b801228..0a92717 100644 --- a/llvm/unittests/ADT/TypeSwitchTest.cpp +++ b/llvm/unittests/ADT/TypeSwitchTest.cpp @@ -167,7 +167,7 @@ TEST(TypeSwitchTest, DefaultNullptr) { TEST(TypeSwitchTest, DefaultNullptrForPointerLike) { struct Value { void *ptr; - Value(const Value &other) : ptr(other.ptr) {} + Value(const Value &other) = default; Value(std::nullptr_t) : ptr(nullptr) {} Value() : Value(nullptr) {} }; diff --git a/llvm/unittests/CodeGen/AsmPrinterDwarfTest.cpp b/llvm/unittests/CodeGen/AsmPrinterDwarfTest.cpp index af2d56d..d0991e6 100644 --- a/llvm/unittests/CodeGen/AsmPrinterDwarfTest.cpp +++ b/llvm/unittests/CodeGen/AsmPrinterDwarfTest.cpp @@ -383,7 +383,7 @@ class AsmPrinterHandlerTest : public AsmPrinterFixtureBase { public: TestHandler(AsmPrinterHandlerTest &Test) : Test(Test) {} - ~TestHandler() override {} + ~TestHandler() override = default; void setSymbolSize(const MCSymbol *Sym, uint64_t Size) override {} void beginModule(Module *M) override { Test.BeginCount++; } void endModule() override { Test.EndCount++; } diff --git a/llvm/unittests/CodeGen/MFCommon.inc b/llvm/unittests/CodeGen/MFCommon.inc index a86a68c..0180ba0 100644 --- a/llvm/unittests/CodeGen/MFCommon.inc +++ b/llvm/unittests/CodeGen/MFCommon.inc @@ -86,7 +86,7 @@ public: : TargetSubtargetInfo(Triple(""), "", "", "", {}, {}, {}, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr), FL(), TL(TM) {} - ~BogusSubtarget() override {} + ~BogusSubtarget() override = default; const TargetFrameLowering *getFrameLowering() const override { return &FL; } @@ -117,7 +117,7 @@ public: Reloc::Static, CodeModel::Small, CodeGenOptLevel::Default), ST(*this) {} - ~BogusTargetMachine() override {} + ~BogusTargetMachine() override = default; const TargetSubtargetInfo *getSubtargetImpl(const Function &) const override { return &ST; diff --git a/llvm/unittests/CodeGen/MachineOperandTest.cpp b/llvm/unittests/CodeGen/MachineOperandTest.cpp index 3f3f48f..0373c7a 100644 --- a/llvm/unittests/CodeGen/MachineOperandTest.cpp +++ b/llvm/unittests/CodeGen/MachineOperandTest.cpp @@ -424,4 +424,24 @@ TEST(MachineOperandTest, HashValue) { ASSERT_TRUE(MO1.isIdenticalTo(MO2)); } +TEST(MachineOperandTest, RegisterLiveOutHashValue) { + LLVMContext Ctx; + Module Mod("Module", Ctx); + auto MF = createMachineFunction(Ctx, Mod); + MachineBasicBlock *MBB = MF->CreateMachineBasicBlock(); + MCInstrDesc MCID = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}; + auto *MI1 = MF->CreateMachineInstr(MCID, DebugLoc()); + auto *MI2 = MF->CreateMachineInstr(MCID, DebugLoc()); + MBB->insert(MBB->begin(), MI1); + MBB->insert(MBB->begin(), MI2); + uint32_t Mask1 = 0; + uint32_t Mask2 = 0; + MI1->addOperand(*MF, MachineOperand::CreateRegLiveOut(&Mask1)); + MI2->addOperand(*MF, MachineOperand::CreateRegLiveOut(&Mask2)); + auto MO1 = MI1->getOperand(0); + auto MO2 = MI2->getOperand(0); + EXPECT_EQ(hash_value(MO1), hash_value(MO2)); + EXPECT_TRUE(MO1.isIdenticalTo(MO2)); +} + } // end namespace diff --git a/llvm/unittests/DebugInfo/CodeView/RandomAccessVisitorTest.cpp b/llvm/unittests/DebugInfo/CodeView/RandomAccessVisitorTest.cpp index 5c96199..fab40b9 100644 --- a/llvm/unittests/DebugInfo/CodeView/RandomAccessVisitorTest.cpp +++ b/llvm/unittests/DebugInfo/CodeView/RandomAccessVisitorTest.cpp @@ -84,7 +84,7 @@ public: class RandomAccessVisitorTest : public testing::Test { public: - RandomAccessVisitorTest() {} + RandomAccessVisitorTest() = default; static void SetUpTestCase() { GlobalState = std::make_unique<GlobalTestState>(); diff --git a/llvm/unittests/DebugInfo/CodeView/TypeIndexDiscoveryTest.cpp b/llvm/unittests/DebugInfo/CodeView/TypeIndexDiscoveryTest.cpp index b1f19e9..62b7591 100644 --- a/llvm/unittests/DebugInfo/CodeView/TypeIndexDiscoveryTest.cpp +++ b/llvm/unittests/DebugInfo/CodeView/TypeIndexDiscoveryTest.cpp @@ -21,7 +21,7 @@ using namespace llvm::codeview; class TypeIndexIteratorTest : public testing::Test { public: - TypeIndexIteratorTest() {} + TypeIndexIteratorTest() = default; void SetUp() override { Refs.clear(); diff --git a/llvm/unittests/DebugInfo/DWARF/DWARFDebugLineTest.cpp b/llvm/unittests/DebugInfo/DWARF/DWARFDebugLineTest.cpp index 2fe5260..aa5b292 100644 --- a/llvm/unittests/DebugInfo/DWARF/DWARFDebugLineTest.cpp +++ b/llvm/unittests/DebugInfo/DWARF/DWARFDebugLineTest.cpp @@ -864,7 +864,7 @@ TEST_F(DebugLineBasicFixture, CallbackUsedForUnterminatedSequence) { } struct AdjustAddressFixtureBase : public CommonFixture { - virtual ~AdjustAddressFixtureBase() {} + virtual ~AdjustAddressFixtureBase() = default; // Create and update the prologue as specified by the subclass, then return // the length of the table. diff --git a/llvm/unittests/ExecutionEngine/JITLink/JITLinkTestUtils.h b/llvm/unittests/ExecutionEngine/JITLink/JITLinkTestUtils.h index f03c82f..a2732e3 100644 --- a/llvm/unittests/ExecutionEngine/JITLink/JITLinkTestUtils.h +++ b/llvm/unittests/ExecutionEngine/JITLink/JITLinkTestUtils.h @@ -19,7 +19,7 @@ class MockJITLinkMemoryManager : public llvm::jitlink::JITLinkMemoryManager { public: class Alloc { public: - virtual ~Alloc() {} + virtual ~Alloc() = default; }; class SimpleAlloc : public Alloc { diff --git a/llvm/unittests/ExecutionEngine/Orc/LibraryResolverTest.cpp b/llvm/unittests/ExecutionEngine/Orc/LibraryResolverTest.cpp index 31f0cce..2a396da 100644 --- a/llvm/unittests/ExecutionEngine/Orc/LibraryResolverTest.cpp +++ b/llvm/unittests/ExecutionEngine/Orc/LibraryResolverTest.cpp @@ -30,7 +30,9 @@ using namespace llvm; using namespace llvm::orc; -#if defined(__APPLE__) || defined(__linux__) +// Disabled due to test setup issue — YAML to shared library creation seems +// invalid on some build bots. (PR #165360) Not related to code logic. +#if 0 // TODO: Add COFF (Windows) support for these tests. // this facility also works correctly on Windows (COFF), // so we should eventually enable and run these tests for that platform as well. diff --git a/llvm/unittests/MC/SystemZ/SystemZMCDisassemblerTest.cpp b/llvm/unittests/MC/SystemZ/SystemZMCDisassemblerTest.cpp index 87fad37..25c22d1 100644 --- a/llvm/unittests/MC/SystemZ/SystemZMCDisassemblerTest.cpp +++ b/llvm/unittests/MC/SystemZ/SystemZMCDisassemblerTest.cpp @@ -61,7 +61,7 @@ Context &getContext() { class SystemZMCSymbolizerTest : public MCSymbolizer { public: SystemZMCSymbolizerTest(MCContext &MC) : MCSymbolizer(MC, nullptr) {} - ~SystemZMCSymbolizerTest() override {} + ~SystemZMCSymbolizerTest() override = default; bool tryAddingSymbolicOperand([[maybe_unused]] MCInst &Inst, [[maybe_unused]] raw_ostream &CStream, diff --git a/llvm/unittests/MC/X86/X86MCDisassemblerTest.cpp b/llvm/unittests/MC/X86/X86MCDisassemblerTest.cpp index 286528f..6d44151 100644 --- a/llvm/unittests/MC/X86/X86MCDisassemblerTest.cpp +++ b/llvm/unittests/MC/X86/X86MCDisassemblerTest.cpp @@ -62,7 +62,7 @@ Context &getContext() { class X86MCSymbolizerTest : public MCSymbolizer { public: X86MCSymbolizerTest(MCContext &MC) : MCSymbolizer(MC, nullptr) {} - ~X86MCSymbolizerTest() override {} + ~X86MCSymbolizerTest() override = default; struct OpInfo { int64_t Value = 0; diff --git a/llvm/unittests/MIR/MachineMetadata.cpp b/llvm/unittests/MIR/MachineMetadata.cpp index 0f038d9..5875512 100644 --- a/llvm/unittests/MIR/MachineMetadata.cpp +++ b/llvm/unittests/MIR/MachineMetadata.cpp @@ -33,7 +33,7 @@ using namespace llvm; class MachineMetadataTest : public testing::Test { public: - MachineMetadataTest() {} + MachineMetadataTest() = default; protected: LLVMContext Context; diff --git a/llvm/unittests/MIR/MachineStableHashTest.cpp b/llvm/unittests/MIR/MachineStableHashTest.cpp index ea0de1a..bedecb1 100644 --- a/llvm/unittests/MIR/MachineStableHashTest.cpp +++ b/llvm/unittests/MIR/MachineStableHashTest.cpp @@ -22,7 +22,7 @@ using namespace llvm; class MachineStableHashTest : public testing::Test { public: - MachineStableHashTest() {} + MachineStableHashTest() = default; protected: LLVMContext Context; diff --git a/llvm/unittests/Object/XCOFFObjectFileTest.cpp b/llvm/unittests/Object/XCOFFObjectFileTest.cpp index f696cde..10217f6 100644 --- a/llvm/unittests/Object/XCOFFObjectFileTest.cpp +++ b/llvm/unittests/Object/XCOFFObjectFileTest.cpp @@ -18,10 +18,10 @@ using namespace llvm::XCOFF; TEST(XCOFFObjectFileTest, XCOFFObjectType) { // Create an arbitrary object of a non-XCOFF type and test that // dyn_cast<XCOFFObjectFile> returns null for it. - char Buf[sizeof(typename ELF64LE::Ehdr)] = {}; + char Buf[sizeof(ELF64LE::Ehdr)] = {}; memcpy(Buf, "\177ELF", 4); - auto *EHdr = reinterpret_cast<typename ELF64LE::Ehdr *>(Buf); + auto *EHdr = reinterpret_cast<ELF64LE::Ehdr *>(Buf); EHdr->e_ident[llvm::ELF::EI_CLASS] = llvm::ELF::ELFCLASS64; EHdr->e_ident[llvm::ELF::EI_DATA] = llvm::ELF::ELFDATA2LSB; diff --git a/llvm/unittests/Support/AlignOfTest.cpp b/llvm/unittests/Support/AlignOfTest.cpp index 979f2cf..53358a28 100644 --- a/llvm/unittests/Support/AlignOfTest.cpp +++ b/llvm/unittests/Support/AlignOfTest.cpp @@ -79,14 +79,14 @@ struct V8 : V5, virtual V6, V7 { double zz; double S6::f() { return 0.0; } float D2::g() { return 0.0f; } -V1::~V1() {} -V2::~V2() {} -V3::~V3() {} -V4::~V4() {} -V5::~V5() {} -V6::~V6() {} -V7::~V7() {} -V8::~V8() {} +V1::~V1() = default; +V2::~V2() = default; +V3::~V3() = default; +V4::~V4() = default; +V5::~V5() = default; +V6::~V6() = default; +V7::~V7() = default; +V8::~V8() = default; template <typename M> struct T { M m; }; diff --git a/llvm/unittests/Support/AllocatorTest.cpp b/llvm/unittests/Support/AllocatorTest.cpp index 1069e43..2337f34 100644 --- a/llvm/unittests/Support/AllocatorTest.cpp +++ b/llvm/unittests/Support/AllocatorTest.cpp @@ -235,7 +235,7 @@ class MockSlabAllocator { static size_t LastSlabSize; public: - ~MockSlabAllocator() { } + ~MockSlabAllocator() = default; void *Allocate(size_t Size, size_t /*Alignment*/) { // Allocate space for the alignment, the slab, and a void* that goes right diff --git a/llvm/unittests/Support/BinaryStreamTest.cpp b/llvm/unittests/Support/BinaryStreamTest.cpp index 70cd403..06ed12b 100644 --- a/llvm/unittests/Support/BinaryStreamTest.cpp +++ b/llvm/unittests/Support/BinaryStreamTest.cpp @@ -110,7 +110,7 @@ constexpr uint32_t NumStreams = 2 * NumEndians; class BinaryStreamTest : public testing::Test { public: - BinaryStreamTest() {} + BinaryStreamTest() = default; void SetUp() override { Streams.clear(); diff --git a/llvm/unittests/Support/Casting.cpp b/llvm/unittests/Support/Casting.cpp index 18327f6..7906750 100644 --- a/llvm/unittests/Support/Casting.cpp +++ b/llvm/unittests/Support/Casting.cpp @@ -23,7 +23,7 @@ template <typename T> IllegalCast *cast(...) { return nullptr; } // with conversion facility // struct bar { - bar() {} + bar() = default; bar(const bar &) = delete; struct foo *baz(); struct foo *caz(); @@ -36,7 +36,7 @@ struct foo { }; struct base { - virtual ~base() {} + virtual ~base() = default; }; struct derived : public base { @@ -375,12 +375,12 @@ namespace inferred_upcasting { class Base { public: // No classof. We are testing that the upcast is inferred. - Base() {} + Base() = default; }; class Derived : public Base { public: - Derived() {} + Derived() = default; }; // Even with no explicit classof() in Base, we should still be able to cast @@ -529,7 +529,7 @@ TEST(CastingTest, smart_dyn_cast_or_null) { #ifndef NDEBUG namespace assertion_checks { struct Base { - virtual ~Base() {} + virtual ~Base() = default; }; struct Derived : public Base { diff --git a/llvm/unittests/Support/InstructionCostTest.cpp b/llvm/unittests/Support/InstructionCostTest.cpp index efe8388..5392689 100644 --- a/llvm/unittests/Support/InstructionCostTest.cpp +++ b/llvm/unittests/Support/InstructionCostTest.cpp @@ -14,7 +14,7 @@ using namespace llvm; namespace { struct CostTest : public testing::Test { - CostTest() {} + CostTest() = default; }; } // namespace diff --git a/llvm/unittests/Support/OptimizedStructLayoutTest.cpp b/llvm/unittests/Support/OptimizedStructLayoutTest.cpp index e8cd5f4..0bcae0d 100644 --- a/llvm/unittests/Support/OptimizedStructLayoutTest.cpp +++ b/llvm/unittests/Support/OptimizedStructLayoutTest.cpp @@ -25,7 +25,7 @@ class LayoutTest { bool Verified = false; public: - LayoutTest() {} + LayoutTest() = default; LayoutTest(const LayoutTest &) = delete; LayoutTest &operator=(const LayoutTest &) = delete; ~LayoutTest() { assert(Verified); } diff --git a/llvm/unittests/Support/YAMLIOTest.cpp b/llvm/unittests/Support/YAMLIOTest.cpp index 283e5f8..7446c07 100644 --- a/llvm/unittests/Support/YAMLIOTest.cpp +++ b/llvm/unittests/Support/YAMLIOTest.cpp @@ -3221,12 +3221,12 @@ template <> struct TaggedScalarTraits<Scalar> { template <> struct CustomMappingTraits<Map> { static void inputOne(IO &IO, StringRef Key, Map &M) { - IO.mapRequired(Key.str().c_str(), M[Key]); + IO.mapRequired(Key, M[Key]); } static void output(IO &IO, Map &M) { for (auto &N : M) - IO.mapRequired(N.getKey().str().c_str(), N.getValue()); + IO.mapRequired(N.getKey(), N.getValue()); } }; diff --git a/llvm/unittests/Support/raw_ostream_proxy_test.cpp b/llvm/unittests/Support/raw_ostream_proxy_test.cpp index 864dda7..446e64a 100644 --- a/llvm/unittests/Support/raw_ostream_proxy_test.cpp +++ b/llvm/unittests/Support/raw_ostream_proxy_test.cpp @@ -40,8 +40,6 @@ public: bool IsDisplayed = false; }; -constexpr size_t BufferedNoPwriteSmallVectorStream::PreferredBufferSize; - TEST(raw_ostream_proxyTest, write) { // Besides confirming that "write" works, this test confirms that the proxy // takes on the buffer from the stream it's proxying, such that writes to the diff --git a/llvm/unittests/Target/AArch64/AArch64InstPrinterTest.cpp b/llvm/unittests/Target/AArch64/AArch64InstPrinterTest.cpp index 4dfc0bc..a835a34 100644 --- a/llvm/unittests/Target/AArch64/AArch64InstPrinterTest.cpp +++ b/llvm/unittests/Target/AArch64/AArch64InstPrinterTest.cpp @@ -36,10 +36,8 @@ static std::string AArch64InstPrinterTestPrintAlignedLabel(uint64_t value) { MCAsmInfo MAI; MCInstrInfo MII; MCRegisterInfo MRI; - MCSubtargetInfo STI(Triple(""), "", "", "", {}, - ArrayRef((SubtargetFeatureKV *)NULL, (size_t)0), - ArrayRef((SubtargetSubTypeKV *)NULL, (size_t)0), NULL, - NULL, NULL, NULL, NULL, NULL); + MCSubtargetInfo STI(Triple(""), "", "", "", {}, {}, {}, nullptr, nullptr, + nullptr, nullptr, nullptr, nullptr); MCContext Ctx(Triple(""), &MAI, &MRI, &STI); MCInst MI; diff --git a/llvm/unittests/TargetParser/TargetParserTest.cpp b/llvm/unittests/TargetParser/TargetParserTest.cpp index 759109a..0e5d40a 100644 --- a/llvm/unittests/TargetParser/TargetParserTest.cpp +++ b/llvm/unittests/TargetParser/TargetParserTest.cpp @@ -737,8 +737,7 @@ TEST(TargetParserTest, ARMFPUNeonSupportLevel) { for (ARM::FPUKind FK = static_cast<ARM::FPUKind>(0); FK <= ARM::FPUKind::FK_LAST; FK = static_cast<ARM::FPUKind>(static_cast<unsigned>(FK) + 1)) - if (FK == ARM::FK_LAST || - ARM::getFPUName(FK).find("neon") == std::string::npos) + if (FK == ARM::FK_LAST || !ARM::getFPUName(FK).contains("neon")) EXPECT_EQ(ARM::NeonSupportLevel::None, ARM::getFPUNeonSupportLevel(FK)); else EXPECT_NE(ARM::NeonSupportLevel::None, ARM::getFPUNeonSupportLevel(FK)); @@ -748,9 +747,8 @@ TEST(TargetParserTest, ARMFPURestriction) { for (ARM::FPUKind FK = static_cast<ARM::FPUKind>(0); FK <= ARM::FPUKind::FK_LAST; FK = static_cast<ARM::FPUKind>(static_cast<unsigned>(FK) + 1)) { - if (FK == ARM::FK_LAST || - (ARM::getFPUName(FK).find("d16") == std::string::npos && - ARM::getFPUName(FK).find("vfpv3xd") == std::string::npos)) + if (FK == ARM::FK_LAST || (!ARM::getFPUName(FK).contains("d16") && + !ARM::getFPUName(FK).contains("vfpv3xd"))) EXPECT_EQ(ARM::FPURestriction::None, ARM::getFPURestriction(FK)); else EXPECT_NE(ARM::FPURestriction::None, ARM::getFPURestriction(FK)); diff --git a/llvm/unittests/XRay/GraphTest.cpp b/llvm/unittests/XRay/GraphTest.cpp index 37f07cc..0d46a3d 100644 --- a/llvm/unittests/XRay/GraphTest.cpp +++ b/llvm/unittests/XRay/GraphTest.cpp @@ -23,8 +23,8 @@ struct EAttr { unsigned EA; }; typedef Graph<VAttr, EAttr, unsigned> GraphT; -typedef typename GraphT::VertexIdentifier VI; -typedef typename GraphT::EdgeIdentifier EI; +typedef GraphT::VertexIdentifier VI; +typedef GraphT::EdgeIdentifier EI; // Test Fixture template <typename T> class GraphTest : public testing::Test { @@ -56,8 +56,8 @@ private: typedef ::testing::Types<GraphT, const GraphT> GraphTestTypes; -using VVT = typename GraphT::VertexValueType; -using EVT = typename GraphT::EdgeValueType; +using VVT = GraphT::VertexValueType; +using EVT = GraphT::EdgeValueType; TYPED_TEST_SUITE(GraphTest, GraphTestTypes, ); diff --git a/llvm/unittests/tools/llvm-exegesis/AArch64/TargetTest.cpp b/llvm/unittests/tools/llvm-exegesis/AArch64/TargetTest.cpp index 7a66117b..5bc489b 100644 --- a/llvm/unittests/tools/llvm-exegesis/AArch64/TargetTest.cpp +++ b/llvm/unittests/tools/llvm-exegesis/AArch64/TargetTest.cpp @@ -28,7 +28,7 @@ using testing::IsEmpty; using testing::Not; using testing::NotNull; -constexpr const char kTriple[] = "aarch64-unknown-linux"; +constexpr char kTriple[] = "aarch64-unknown-linux"; class AArch64TargetTest : public ::testing::Test { protected: diff --git a/llvm/unittests/tools/llvm-exegesis/PowerPC/TargetTest.cpp b/llvm/unittests/tools/llvm-exegesis/PowerPC/TargetTest.cpp index 3708f18..0e90654 100644 --- a/llvm/unittests/tools/llvm-exegesis/PowerPC/TargetTest.cpp +++ b/llvm/unittests/tools/llvm-exegesis/PowerPC/TargetTest.cpp @@ -20,16 +20,13 @@ namespace llvm{ namespace exegesis { - -void InitializePowerPCExegesisTarget(); - namespace { using testing::NotNull; using testing::IsEmpty; using testing::Not; -constexpr const char kTriple[] = "powerpc64le-unknown-linux"; +constexpr char kTriple[] = "powerpc64le-unknown-linux"; class PowerPCTargetTest : public PPCTestBase { protected: diff --git a/llvm/unittests/tools/llvm-exegesis/RISCV/TargetTest.cpp b/llvm/unittests/tools/llvm-exegesis/RISCV/TargetTest.cpp index c86a436..13a1e5a 100644 --- a/llvm/unittests/tools/llvm-exegesis/RISCV/TargetTest.cpp +++ b/llvm/unittests/tools/llvm-exegesis/RISCV/TargetTest.cpp @@ -20,9 +20,6 @@ namespace llvm { namespace exegesis { - -void InitializeRISCVExegesisTarget(); - namespace { using testing::IsEmpty; diff --git a/llvm/unittests/tools/llvm-exegesis/X86/SnippetFileTest.cpp b/llvm/unittests/tools/llvm-exegesis/X86/SnippetFileTest.cpp index de883ab..755a748 100644 --- a/llvm/unittests/tools/llvm-exegesis/X86/SnippetFileTest.cpp +++ b/llvm/unittests/tools/llvm-exegesis/X86/SnippetFileTest.cpp @@ -23,9 +23,6 @@ namespace llvm { namespace exegesis { - -void InitializeX86ExegesisTarget(); - namespace { using testing::ElementsAre; diff --git a/llvm/unittests/tools/llvm-exegesis/X86/SnippetGeneratorTest.cpp b/llvm/unittests/tools/llvm-exegesis/X86/SnippetGeneratorTest.cpp index 60c7262..5953f4e 100644 --- a/llvm/unittests/tools/llvm-exegesis/X86/SnippetGeneratorTest.cpp +++ b/llvm/unittests/tools/llvm-exegesis/X86/SnippetGeneratorTest.cpp @@ -20,9 +20,6 @@ namespace llvm { namespace exegesis { - -void InitializeX86ExegesisTarget(); - namespace { using testing::AnyOf; diff --git a/llvm/unittests/tools/llvm-exegesis/X86/SnippetRepetitorTest.cpp b/llvm/unittests/tools/llvm-exegesis/X86/SnippetRepetitorTest.cpp index 41ee402..7a40901de 100644 --- a/llvm/unittests/tools/llvm-exegesis/X86/SnippetRepetitorTest.cpp +++ b/llvm/unittests/tools/llvm-exegesis/X86/SnippetRepetitorTest.cpp @@ -16,9 +16,6 @@ namespace llvm { namespace exegesis { - -void InitializeX86ExegesisTarget(); - namespace { using testing::ElementsAre; @@ -52,8 +49,8 @@ protected: Fill(Sink); } - static constexpr const unsigned kMinInstructions = 3; - static constexpr const unsigned kLoopBodySize = 5; + static constexpr unsigned kMinInstructions = 3; + static constexpr unsigned kLoopBodySize = 5; std::unique_ptr<TargetMachine> TM; std::unique_ptr<LLVMContext> Context; diff --git a/llvm/unittests/tools/llvm-exegesis/X86/SubprocessMemoryTest.cpp b/llvm/unittests/tools/llvm-exegesis/X86/SubprocessMemoryTest.cpp index a0cad28..08c18e4 100644 --- a/llvm/unittests/tools/llvm-exegesis/X86/SubprocessMemoryTest.cpp +++ b/llvm/unittests/tools/llvm-exegesis/X86/SubprocessMemoryTest.cpp @@ -29,7 +29,7 @@ namespace exegesis { // This needs to be updated anytime a test is added or removed from the test // suite. -static constexpr const size_t TestCount = 4; +static constexpr size_t TestCount = 4; class SubprocessMemoryTest : public X86TestBase { protected: diff --git a/llvm/unittests/tools/llvm-exegesis/X86/TargetTest.cpp b/llvm/unittests/tools/llvm-exegesis/X86/TargetTest.cpp index 846729c6..5a21a69 100644 --- a/llvm/unittests/tools/llvm-exegesis/X86/TargetTest.cpp +++ b/llvm/unittests/tools/llvm-exegesis/X86/TargetTest.cpp @@ -53,9 +53,6 @@ bool operator==(const MCInst &a, const MCInst &b) { namespace llvm { namespace exegesis { - -void InitializeX86ExegesisTarget(); - namespace { using testing::AllOf; @@ -585,7 +582,7 @@ TEST_F(X86Core2TargetTest, SetRegToDf0) { TEST_F(X86Core2Avx512TargetTest, FillMemoryOperands_ADD64rm) { const Instruction &I = getInstr(X86::ADD64rm); InstructionTemplate IT(&I); - constexpr const int kOffset = 42; + constexpr int kOffset = 42; State.getExegesisTarget().fillMemoryOperands(IT, X86::RDI, kOffset); // Memory is operands 2-6. EXPECT_THAT(IT.getValueFor(I.Operands[2]), IsReg(X86::RDI)); @@ -598,7 +595,7 @@ TEST_F(X86Core2Avx512TargetTest, FillMemoryOperands_ADD64rm) { TEST_F(X86Core2Avx512TargetTest, FillMemoryOperands_VGATHERDPSZ128rm) { const Instruction &I = getInstr(X86::VGATHERDPSZ128rm); InstructionTemplate IT(&I); - constexpr const int kOffset = 42; + constexpr int kOffset = 42; State.getExegesisTarget().fillMemoryOperands(IT, X86::RDI, kOffset); // Memory is operands 4-8. EXPECT_THAT(IT.getValueFor(I.Operands[4]), IsReg(X86::RDI)); @@ -628,9 +625,9 @@ TEST_F(X86Core2TargetTest, GenerateLowerMunmapTest) { } #ifdef __arm__ -static constexpr const uintptr_t VAddressSpaceCeiling = 0xC0000000; +static constexpr uintptr_t VAddressSpaceCeiling = 0xC0000000; #else -static constexpr const uintptr_t VAddressSpaceCeiling = 0x0000800000000000; +static constexpr uintptr_t VAddressSpaceCeiling = 0x0000800000000000; #endif TEST_F(X86Core2TargetTest, GenerateUpperMunmapTest) { diff --git a/llvm/unittests/tools/llvm-exegesis/X86/TestBase.h b/llvm/unittests/tools/llvm-exegesis/X86/TestBase.h index 4122726..b4c84d1 100644 --- a/llvm/unittests/tools/llvm-exegesis/X86/TestBase.h +++ b/llvm/unittests/tools/llvm-exegesis/X86/TestBase.h @@ -22,7 +22,7 @@ namespace exegesis { void InitializeX86ExegesisTarget(); -constexpr const char kTriple[] = "x86_64-unknown-linux"; +constexpr char kTriple[] = "x86_64-unknown-linux"; class X86TestBase : public ::testing::Test { protected: diff --git a/llvm/utils/FileCheck/FileCheck.cpp b/llvm/utils/FileCheck/FileCheck.cpp index 305c28b..a5473f9 100644 --- a/llvm/utils/FileCheck/FileCheck.cpp +++ b/llvm/utils/FileCheck/FileCheck.cpp @@ -193,7 +193,7 @@ struct MarkerStyle { std::string Note; /// Does this marker indicate inclusion by -dump-input-filter=error? bool FiltersAsError; - MarkerStyle() {} + MarkerStyle() = default; MarkerStyle(char Lead, raw_ostream::Colors Color, const std::string &Note = "", bool FiltersAsError = false) : Lead(Lead), Color(Color), Note(Note), FiltersAsError(FiltersAsError) { diff --git a/llvm/utils/TableGen/Basic/ARMTargetDefEmitter.cpp b/llvm/utils/TableGen/Basic/ARMTargetDefEmitter.cpp index 3f284ee..b63ce36 100644 --- a/llvm/utils/TableGen/Basic/ARMTargetDefEmitter.cpp +++ b/llvm/utils/TableGen/Basic/ARMTargetDefEmitter.cpp @@ -220,7 +220,7 @@ static void emitARMTargetDef(const RecordKeeper &RK, raw_ostream &OS) { ProfileLower + "'"); // Name of the object in C++ - const std::string CppSpelling = ArchInfoName(Major, Minor, ProfileUpper); + std::string CppSpelling = ArchInfoName(Major, Minor, ProfileUpper); OS << "inline constexpr ArchInfo " << CppSpelling << " = {\n"; CppSpellings.push_back(std::move(CppSpelling)); diff --git a/llvm/utils/TableGen/Basic/RuntimeLibcallsEmitter.cpp b/llvm/utils/TableGen/Basic/RuntimeLibcallsEmitter.cpp index ed802e2..6a36f47 100644 --- a/llvm/utils/TableGen/Basic/RuntimeLibcallsEmitter.cpp +++ b/llvm/utils/TableGen/Basic/RuntimeLibcallsEmitter.cpp @@ -154,7 +154,7 @@ public: Provides = ProvideMap.lookup(ProvidesDef); } - ~RuntimeLibcallImpl() {} + ~RuntimeLibcallImpl() = default; const Record *getDef() const { return TheDef; } diff --git a/llvm/utils/TableGen/Basic/TargetFeaturesEmitter.h b/llvm/utils/TableGen/Basic/TargetFeaturesEmitter.h index 99e4820..412f323 100644 --- a/llvm/utils/TableGen/Basic/TargetFeaturesEmitter.h +++ b/llvm/utils/TableGen/Basic/TargetFeaturesEmitter.h @@ -43,7 +43,7 @@ public: void printFeatureKeyValues(raw_ostream &OS, const FeatureMapTy &FeatureMap); void printCPUKeyValues(raw_ostream &OS, const FeatureMapTy &FeatureMap); virtual void run(raw_ostream &O); - virtual ~TargetFeaturesEmitter() {}; + virtual ~TargetFeaturesEmitter() = default; }; } // namespace llvm #endif diff --git a/llvm/utils/TableGen/Common/CodeGenRegisters.cpp b/llvm/utils/TableGen/Common/CodeGenRegisters.cpp index 8d0ec9a..2eb94b7e 100644 --- a/llvm/utils/TableGen/Common/CodeGenRegisters.cpp +++ b/llvm/utils/TableGen/Common/CodeGenRegisters.cpp @@ -1651,8 +1651,7 @@ template <> struct llvm::GraphTraits<SubRegIndexCompositionGraph> { struct ChildIteratorType : public iterator_adaptor_base< ChildIteratorType, CompMapIt, - typename std::iterator_traits<CompMapIt>::iterator_category, - NodeRef> { + std::iterator_traits<CompMapIt>::iterator_category, NodeRef> { ChildIteratorType(CompMapIt I) : ChildIteratorType::iterator_adaptor_base(I) {} diff --git a/llvm/utils/TableGen/Common/CodeGenTarget.cpp b/llvm/utils/TableGen/Common/CodeGenTarget.cpp index 3db0d07..1e93788 100644 --- a/llvm/utils/TableGen/Common/CodeGenTarget.cpp +++ b/llvm/utils/TableGen/Common/CodeGenTarget.cpp @@ -80,7 +80,7 @@ CodeGenTarget::CodeGenTarget(const RecordKeeper &records) MacroFusions = Records.getAllDerivedDefinitions("Fusion"); } -CodeGenTarget::~CodeGenTarget() {} +CodeGenTarget::~CodeGenTarget() = default; StringRef CodeGenTarget::getName() const { return TargetRec->getName(); } diff --git a/llvm/utils/TableGen/Common/DAGISelMatcher.h b/llvm/utils/TableGen/Common/DAGISelMatcher.h index f87de75..a19f444 100644 --- a/llvm/utils/TableGen/Common/DAGISelMatcher.h +++ b/llvm/utils/TableGen/Common/DAGISelMatcher.h @@ -105,7 +105,7 @@ protected: Matcher(KindTy K) : Kind(K) {} public: - virtual ~Matcher() {} + virtual ~Matcher() = default; unsigned getSize() const { return Size; } void setSize(unsigned sz) { Size = sz; } diff --git a/llvm/utils/TableGen/Common/GlobalISel/GlobalISelMatchTable.cpp b/llvm/utils/TableGen/Common/GlobalISel/GlobalISelMatchTable.cpp index 5d49715..7af757c 100644 --- a/llvm/utils/TableGen/Common/GlobalISel/GlobalISelMatchTable.cpp +++ b/llvm/utils/TableGen/Common/GlobalISel/GlobalISelMatchTable.cpp @@ -457,7 +457,7 @@ std::optional<LLTCodeGen> llvm::gi::MVTToLLT(MVT::SimpleValueType SVT) { void Matcher::optimize() {} -Matcher::~Matcher() {} +Matcher::~Matcher() = default; //===- GroupMatcher -------------------------------------------------------===// @@ -1150,11 +1150,11 @@ void RuleMatcher::insnmatchers_pop_front() { Matchers.erase(Matchers.begin()); } //===- PredicateMatcher ---------------------------------------------------===// -PredicateMatcher::~PredicateMatcher() {} +PredicateMatcher::~PredicateMatcher() = default; //===- OperandPredicateMatcher --------------------------------------------===// -OperandPredicateMatcher::~OperandPredicateMatcher() {} +OperandPredicateMatcher::~OperandPredicateMatcher() = default; bool OperandPredicateMatcher::isHigherPriorityThan( const OperandPredicateMatcher &B) const { @@ -1941,7 +1941,7 @@ bool InstructionOperandMatcher::isHigherPriorityThan( //===- OperandRenderer ----------------------------------------------------===// -OperandRenderer::~OperandRenderer() {} +OperandRenderer::~OperandRenderer() = default; //===- CopyRenderer -------------------------------------------------------===// diff --git a/llvm/utils/TableGen/Common/GlobalISel/GlobalISelMatchTable.h b/llvm/utils/TableGen/Common/GlobalISel/GlobalISelMatchTable.h index 0f1241e..84dfca4 100644 --- a/llvm/utils/TableGen/Common/GlobalISel/GlobalISelMatchTable.h +++ b/llvm/utils/TableGen/Common/GlobalISel/GlobalISelMatchTable.h @@ -621,7 +621,7 @@ public: DefinedInsnVariablesMap::const_iterator defined_insn_vars_end() const { return InsnVariableIDs.end(); } - iterator_range<typename DefinedInsnVariablesMap::const_iterator> + iterator_range<DefinedInsnVariablesMap::const_iterator> defined_insn_vars() const { return make_range(defined_insn_vars_begin(), defined_insn_vars_end()); } @@ -632,8 +632,7 @@ public: MutatableInsnSet::const_iterator mutatable_insns_end() const { return MutatableInsns.end(); } - iterator_range<typename MutatableInsnSet::const_iterator> - mutatable_insns() const { + iterator_range<MutatableInsnSet::const_iterator> mutatable_insns() const { return make_range(mutatable_insns_begin(), mutatable_insns_end()); } void reserveInsnMatcherForMutation(InstructionMatcher *InsnMatcher) { @@ -1375,7 +1374,7 @@ class InstructionPredicateMatcher : public PredicateMatcher { public: InstructionPredicateMatcher(PredicateKind Kind, unsigned InsnVarID) : PredicateMatcher(Kind, InsnVarID) {} - ~InstructionPredicateMatcher() override {} + ~InstructionPredicateMatcher() override = default; /// Compare the priority of this object and B. /// @@ -2319,7 +2318,7 @@ public: ActionKind getKind() const { return Kind; } - virtual ~MatchAction() {} + virtual ~MatchAction() = default; // Some actions may need to add extra predicates to ensure they can run. virtual void emitAdditionalPredicates(MatchTable &Table, diff --git a/llvm/utils/TableGen/Common/InfoByHwMode.cpp b/llvm/utils/TableGen/Common/InfoByHwMode.cpp index 4c8197d..2b3155c 100644 --- a/llvm/utils/TableGen/Common/InfoByHwMode.cpp +++ b/llvm/utils/TableGen/Common/InfoByHwMode.cpp @@ -174,7 +174,7 @@ bool RegSizeInfoByHwMode::hasStricterSpillThan( } void RegSizeInfoByHwMode::writeToStream(raw_ostream &OS) const { - typedef typename decltype(Map)::value_type PairType; + typedef decltype(Map)::value_type PairType; std::vector<const PairType *> Pairs; for (const auto &P : Map) Pairs.push_back(&P); diff --git a/llvm/utils/TableGen/FastISelEmitter.cpp b/llvm/utils/TableGen/FastISelEmitter.cpp index e0be104..c4dbb14 100644 --- a/llvm/utils/TableGen/FastISelEmitter.cpp +++ b/llvm/utils/TableGen/FastISelEmitter.cpp @@ -85,7 +85,7 @@ struct OperandsSignature { char Repr = OK_Invalid; public: - OpKind() {} + OpKind() = default; bool operator<(OpKind RHS) const { return Repr < RHS.Repr; } bool operator==(OpKind RHS) const { return Repr == RHS.Repr; } diff --git a/llvm/utils/TableGen/GlobalISelCombinerEmitter.cpp b/llvm/utils/TableGen/GlobalISelCombinerEmitter.cpp index 043bc628..50e63a4 100644 --- a/llvm/utils/TableGen/GlobalISelCombinerEmitter.cpp +++ b/llvm/utils/TableGen/GlobalISelCombinerEmitter.cpp @@ -2441,7 +2441,7 @@ public: explicit GICombinerEmitter(const RecordKeeper &RK, const CodeGenTarget &Target, StringRef Name, const Record *Combiner); - ~GICombinerEmitter() override {} + ~GICombinerEmitter() override = default; void run(raw_ostream &OS); }; diff --git a/llvm/utils/TableGen/RegisterBankEmitter.cpp b/llvm/utils/TableGen/RegisterBankEmitter.cpp index 61b0b66..60aa2d8 100644 --- a/llvm/utils/TableGen/RegisterBankEmitter.cpp +++ b/llvm/utils/TableGen/RegisterBankEmitter.cpp @@ -100,8 +100,7 @@ public: return RCsWithLargestRegSize[HwMode]; } - iterator_range<typename RegisterClassesTy::const_iterator> - register_classes() const { + iterator_range<RegisterClassesTy::const_iterator> register_classes() const { return RCs; } }; diff --git a/llvm/utils/TableGen/X86DisassemblerTables.cpp b/llvm/utils/TableGen/X86DisassemblerTables.cpp index 3414190..b8c3c02 100644 --- a/llvm/utils/TableGen/X86DisassemblerTables.cpp +++ b/llvm/utils/TableGen/X86DisassemblerTables.cpp @@ -708,7 +708,7 @@ DisassemblerTables::DisassemblerTables() { HasConflicts = false; } -DisassemblerTables::~DisassemblerTables() {} +DisassemblerTables::~DisassemblerTables() = default; void DisassemblerTables::emitModRMDecision(raw_ostream &o1, raw_ostream &o2, unsigned &i1, unsigned &i2, diff --git a/llvm/utils/TableGen/X86ModRMFilters.h b/llvm/utils/TableGen/X86ModRMFilters.h index 7bf111f..4eb57b0 100644 --- a/llvm/utils/TableGen/X86ModRMFilters.h +++ b/llvm/utils/TableGen/X86ModRMFilters.h @@ -28,7 +28,7 @@ class ModRMFilter { public: /// Destructor - Override as necessary. - virtual ~ModRMFilter() {} + virtual ~ModRMFilter() = default; /// isDumb - Indicates whether this filter returns the same value for /// any value of the ModR/M byte. diff --git a/llvm/utils/gn/secondary/clang-tools-extra/clang-tidy/bugprone/BUILD.gn b/llvm/utils/gn/secondary/clang-tools-extra/clang-tidy/bugprone/BUILD.gn index 0361233..f280f69 100644 --- a/llvm/utils/gn/secondary/clang-tools-extra/clang-tidy/bugprone/BUILD.gn +++ b/llvm/utils/gn/secondary/clang-tools-extra/clang-tidy/bugprone/BUILD.gn @@ -69,6 +69,7 @@ static_library("bugprone") { "ParentVirtualCallCheck.cpp", "PointerArithmeticOnPolymorphicObjectCheck.cpp", "PosixReturnCheck.cpp", + "RawMemoryCallOnNonTrivialTypeCheck.cpp", "RedundantBranchConditionCheck.cpp", "ReservedIdentifierCheck.cpp", "ReturnConstRefFromParameterCheck.cpp", diff --git a/llvm/utils/gn/secondary/clang-tools-extra/clang-tidy/cert/BUILD.gn b/llvm/utils/gn/secondary/clang-tools-extra/clang-tidy/cert/BUILD.gn index b097e13..3ad0a83 100644 --- a/llvm/utils/gn/secondary/clang-tools-extra/clang-tidy/cert/BUILD.gn +++ b/llvm/utils/gn/secondary/clang-tools-extra/clang-tidy/cert/BUILD.gn @@ -21,7 +21,6 @@ static_library("cert") { "FloatLoopCounter.cpp", "LimitedRandomnessCheck.cpp", "MutatingCopyCheck.cpp", - "NonTrivialTypesLibcMemoryCallsCheck.cpp", "ProperlySeededRandomGeneratorCheck.cpp", "ThrownExceptionTypeCheck.cpp", ] diff --git a/llvm/utils/gn/secondary/libcxx/include/BUILD.gn b/llvm/utils/gn/secondary/libcxx/include/BUILD.gn index c20d2aa..9e0b951 100644 --- a/llvm/utils/gn/secondary/libcxx/include/BUILD.gn +++ b/llvm/utils/gn/secondary/libcxx/include/BUILD.gn @@ -1524,6 +1524,7 @@ if (current_toolchain == default_toolchain) { "__type_traits/is_valid_expansion.h", "__type_traits/is_void.h", "__type_traits/is_volatile.h", + "__type_traits/is_within_lifetime.h", "__type_traits/lazy.h", "__type_traits/make_32_64_or_128_bit.h", "__type_traits/make_const_lvalue_ref.h", diff --git a/llvm/utils/gn/secondary/lldb/tools/lldb-dap/tool/BUILD.gn b/llvm/utils/gn/secondary/lldb/tools/lldb-dap/tool/BUILD.gn index 8b764843..1773fe3 100644 --- a/llvm/utils/gn/secondary/lldb/tools/lldb-dap/tool/BUILD.gn +++ b/llvm/utils/gn/secondary/lldb/tools/lldb-dap/tool/BUILD.gn @@ -5,7 +5,7 @@ import("//llvm/version.gni") tablegen("Options") { visibility = [ ":lldb-dap" ] args = [ "-gen-opt-parser-defs" ] - td_file = "../Options.td" + td_file = "Options.td" } if (host_os == "mac") { diff --git a/llvm/utils/gn/secondary/llvm/lib/ExecutionEngine/Orc/TargetProcess/BUILD.gn b/llvm/utils/gn/secondary/llvm/lib/ExecutionEngine/Orc/TargetProcess/BUILD.gn index c4ce990..937e81b 100644 --- a/llvm/utils/gn/secondary/llvm/lib/ExecutionEngine/Orc/TargetProcess/BUILD.gn +++ b/llvm/utils/gn/secondary/llvm/lib/ExecutionEngine/Orc/TargetProcess/BUILD.gn @@ -12,6 +12,8 @@ static_library("TargetProcess") { "JITLoaderGDB.cpp", "JITLoaderPerf.cpp", "JITLoaderVTune.cpp", + "LibraryResolver.cpp", + "LibraryScanner.cpp", "OrcRTBootstrap.cpp", "RegisterEHFrames.cpp", "SimpleExecutorDylibManager.cpp", diff --git a/llvm/utils/gn/secondary/llvm/unittests/ExecutionEngine/Orc/BUILD.gn b/llvm/utils/gn/secondary/llvm/unittests/ExecutionEngine/Orc/BUILD.gn index dfe6d6d..111e4c9 100644 --- a/llvm/utils/gn/secondary/llvm/unittests/ExecutionEngine/Orc/BUILD.gn +++ b/llvm/utils/gn/secondary/llvm/unittests/ExecutionEngine/Orc/BUILD.gn @@ -24,6 +24,7 @@ unittest("OrcJITTests") { "JITLinkRedirectionManagerTest.cpp", "JITTargetMachineBuilderTest.cpp", "LazyCallThroughAndReexportsTest.cpp", + "LibraryResolverTest.cpp", "LookupAndRecordAddrsTest.cpp", "MachOPlatformTest.cpp", "MapperJITLinkMemoryManagerTest.cpp", |