This also fixes an issue where the benchmarks were being built with C++14
instead of C++17, as they should be.

This reverts commit 44587e2f. Sanitizer tests need to be updated.

Similarly as for pointers, even for integers a == b is usually false.

GCC also uses this heuristic.

Reviewed By: ebrevnov

Differential Revision: https://reviews.llvm.org/D85781

For HADD/HSUB/PACKS ops with repeated operands the lower/upper half element of each lane are known to be equivalent

Differential Revision: https://reviews.llvm.org/D85724

This patch changes SplitVecOp_EXTRACT_VECTOR_ELT to work correctly
for scalable vectors and also fixes an a bug in DAGCombiner where
the scalable property is dropped in visitTRUNCATE when attempting
to fold an extract + a truncate.

Reviewed By: efriedma

Differential Revision: https://reviews.llvm.org/D85754

Fix to include non-predicated version of write-back gather in special case
treatment for deducting the instruction type.
(This is fixing https://reviews.llvm.org/D85138 for corner cases)

Differential Revision: https://reviews.llvm.org/D85889

We are re-using tryToMergePartialOverlappingStores, which requires
earlier to domiante Later. In the long run,
tryToMergeParialOverlappingStores should be re-written using MemorySSA.

Fixes PR46513.

`getElfSegmentType` and `getElfPtType` are methods that are used for printing
segment types for LLVM and GNU styles accordingly.

This patch does a cleanup and simplification that allows to avoid
the code duplication and to get rid of one macro.

Differential revision: https://reviews.llvm.org/D85830

Differential Revision: https://reviews.llvm.org/D85831

This reverts commit 385c9d67.

Differential Revision: https://reviews.llvm.org/D85640

Specifying --headers is equivalent to setting --file-headers,
--program-headers and --section-headers at the same time.

The existent test case uses a precompiled object and doesn't test the
output properly. This patch fixes it.

Differential revision: https://reviews.llvm.org/D85832

Similarly as for pointers, even for integers a == b is usually false.

GCC also uses this heuristic.

Reviewed By: ebrevnov

Differential Revision: https://reviews.llvm.org/D85781

Don't duplicate includes already provided by the module header.

Don't duplicate includes already provided by the module header.

This change replaces the InitialLength of pub-tables with Format and
Length. All the InitialLength fields have been removed.

Reviewed By: jhenderson

Differential Revision: https://reviews.llvm.org/D85880

This lets basic AST-consuming actions be defined outside clangdserver.
(it essentially exposes TUScheduler::runWithAST).

Two use cases for now:
 - replace ClangdServer::dumpAST, which doesn't really belong in the public
   interface (a followup patch will do this)
 - allow embedders to add/experiment with extra features
   (e.g. I know some who want to try crazy AST+ML code completion...)

Maybe in the future we should use this sort of mechanism to make ClangdServer
less of a monolith, but that's not in scope for now.
This would probably require a richer and more carefully-thought-out API.

Differential Revision: https://reviews.llvm.org/D85883

In DAGTypeLegalizer::GenWidenVectorStores the algorithm assumes it only
ever deals with fixed width types, hence the offsets for each individual
store never take 'vscale' into account. I've changed the main loop in
that function to use TypeSize instead of unsigned for tracking the
remaining store amount and offset increment. In addition, I've changed
the loop to use the new IncrementPointer helper function for updating
the addresses in each iteration, since this handles scalable vector
types.

Whilst fixing this function I also fixed a minor issue in
IncrementPointer whereby we were not adding the no-unsigned-wrap flag
for the add instruction in the same way as the fixed width case does.

Also, I've added a report_fatal_error in GenWidenVectorTruncStores,
since this code currently uses a sequence of element-by-element scalar
stores.

I've added new tests in

  CodeGen/AArch64/sve-intrinsics-stores.ll
  CodeGen/AArch64/sve-st1-addressing-mode-reg-imm.ll

for the changes in GenWidenVectorStores.

Differential Revision: https://reviews.llvm.org/D84937

In narrowExtractedVectorLoad there is an optimisation that tries to
combine extract_subvector with a narrowing vector load. At the moment
this produces warnings due to the incorrect calls to
getVectorNumElements() for scalable vector types. I've got this
working for scalable vectors too when the extract subvector index
is a multiple of the minimum number of elements. I have added a
new variant of the function:

  MachineFunction::getMachineMemOperand

that copies an existing MachineMemOperand, but replaces the pointer
info with a null version since we cannot currently represent scaled
offsets.

I've added a new test for this particular case in:

  CodeGen/AArch64/sve-extract-subvector.ll

Differential Revision: https://reviews.llvm.org/D83950

Currently only two test failures remain on Sparc, both
`sparcv9-sun-solaris2.11` and `sparc64-unknown-linux-gnu`:

  LLVM :: DebugInfo/Generic/debug-label-inline.ll
  LLVM :: Linker/subprogram-linkonce-weak.ll

They seem related in that debug info isn't generated for instruction
bundles (like `retl+add` in the delay slot).

I've filed separate bugs for both files (Bug 47129 and 47131), though it's
probably the same issue.

This patch `XFAIL`s the tests.

Tested on `sparcv9-sun-solaris2.11` and `amd64-pc-solaris2.11`.

Differential Revision: https://reviews.llvm.org/D85827

When building on `sparc64-unknown-linux-gnu`, I found that a large number
of `SanitizerCommon-asan-sparc*-Linux` tests were `FAIL`ing, like

   SanitizerCommon-asan-sparc-Linux :: Linux/aligned_alloc-alignment.cpp
  [...]
   SanitizerCommon-asan-sparcv9-Linux :: Linux/aligned_alloc-alignment.cpp
  [...]

many of them due to

  fatal error: error in backend: Function "_Z14User_OnSIGSEGViP9siginfo_tPv": over-aligned dynamic alloca not supported.

which breaks ASan on Sparc.  Currently ASan is only built for the benefit
of `gcc` where it does work.  However, when enabling the compilation in
`compiler-rt` to make certain it continues to build, I missed
`compiler-rt/test/sanitizer_common` when disabling ASan testing on Sparc
(it's not yet enabled on Solaris).

This patch fixes the issue.

Tested on `sparcv9-sun-solaris2.11` with the `sanitizer_comon` testsuite enabled.

Differential Revision: https://reviews.llvm.org/D85732

We do that because:
* Big tests generated big tree dumps that could hardly serve as documentation.
* In most cases the tests didn't share setup, thus there was not much addition in lines of code.

We split tests for:
* `UserDefinedLiteral`
* `NestedBinaryOperator`
* `UserDefinedBinaryOperator`
* `UserDefinedPrefixOperator`
* `QualifiedId`

Differential Revision: https://reviews.llvm.org/D85819

unlink() was changed to llvm::sys::fs::remove() in f18.cpp, but the same
change also has to be applied to f18-parse-demo.cpp.

This is motivated by tooling (clangd, libclang etc) - headers without
declarations are legitimate even if they're not valid TUs.

The other use -x c-header cases (PCH/modules) are nonstandard anyway and this
warning doesn't seem necessary there either.

Differential Revision: https://reviews.llvm.org/D85789

Many OpenMP Clang tests do not RUN for version 4.5 and the default
version. This first patch in the series only handles test cases
which do not require any modifications in the CHECK lines after
adding RUN lines for default version.

Reviewed By: ABataev

Differential Revision: https://reviews.llvm.org/D84844

Use functions for process launching, temporary file creation and file deletion from LLVM support library instead of POSIX-specific ones.

This patch is part of the series to [[ http://lists.llvm.org/pipermail/flang-dev/2020-July/000448.html | make flang compilable with MS Visual Studio ]].

Reviewed By: DavidTruby

Differential Revision: https://reviews.llvm.org/D85660

This reverts commit e441b7a7.

This patch causes a compile error in tensorflow opensource project. The stack trace looks like:

Point of crash:
llvm/include/llvm/Analysis/LoopInfoImpl.h : line 35

(gdb) ptype *this
type = const class llvm::LoopBase<llvm::BasicBlock, llvm::Loop> [with BlockT = llvm::BasicBlock, LoopT = llvm::Loop]

(gdb) p *this
$1 = {ParentLoop = 0x0, SubLoops = std::vector of length 0, capacity 0, Blocks = std::vector of length 0, capacity 1,
  DenseBlockSet = {<llvm::SmallPtrSetImpl<llvm::BasicBlock const*>> = {<llvm::SmallPtrSetImplBase> = {<llvm::DebugEpochBase> = {Epoch = 3}, SmallArray = 0x1b2bf6c8, CurArray = 0x1b2bf6c8,
        CurArraySize = 8, NumNonEmpty = 0, NumTombstones = 0}, <No data fields>}, SmallStorage = {0xfffffffffffffffe, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}}, IsInvalid = true}

(gdb) p *this->DenseBlockSet->CurArray
$2 = (const void *) 0xfffffffffffffffe

I will try to get a case from tensorflow or use creduce to get a small case.

The method optionMatches() constructs 9865 std::string instances when comparing different
options. Many of these instances exceed the size of the internal storage and force memory
allocations. This patch adds an early exit check that eliminates most of the string allocations
while keeping the code simple.

Example inputs:
Prefix: /, Name: Fr
Prefix: -, Name: Fr
Prefix: -, Name: fsanitize-address-field-padding=
Prefix: -, Name: fsanitize-address-globals-dead-stripping
Prefix: -, Name: fsanitize-address-poison-custom-array-cookie
Prefix: -, Name: fsanitize-address-use-after-scope
Prefix: -, Name: fsanitize-address-use-odr-indicator
Prefix: -, Name: fsanitize-blacklist=

Differential Revision: D85538