platform/upstream/llvm.git
2 years ago[clang-format] Correctly format variable templates.
Marek Kurdej [Fri, 11 Mar 2022 11:51:15 +0000 (12:51 +0100)]
[clang-format] Correctly format variable templates.

Fixes https://github.com/llvm/llvm-project/issues/54257.

Reviewed By: MyDeveloperDay, HazardyKnusperkeks, owenpan

Differential Revision: https://reviews.llvm.org/D121456

2 years ago[clang-format] Add space to comments starting with '#'.
Marek Kurdej [Tue, 1 Mar 2022 08:42:28 +0000 (09:42 +0100)]
[clang-format] Add space to comments starting with '#'.

Fixes https://github.com/llvm/llvm-project/issues/35116.

Reviewed By: MyDeveloperDay, HazardyKnusperkeks, owenpan

Differential Revision: https://reviews.llvm.org/D121451

2 years ago[clang-format] Handle attributes before case label.
Marek Kurdej [Wed, 9 Mar 2022 10:05:34 +0000 (11:05 +0100)]
[clang-format] Handle attributes before case label.

Fixes https://github.com/llvm/llvm-project/issues/53110.

Reviewed By: MyDeveloperDay, HazardyKnusperkeks, owenpan

Differential Revision: https://reviews.llvm.org/D121450

2 years ago[X86] Update AVX512 intrinsic tests to avoid adds
Simon Pilgrim [Sun, 13 Mar 2022 17:28:18 +0000 (17:28 +0000)]
[X86] Update AVX512 intrinsic tests to avoid adds

As noticed in D119654, by adding the masked intrinsics results together we can end up with the selects being canonicalized away from the intrinsic - this isn't what we want to test here so replace with a insertvalue chain into a aggregate instead to retain all the results.

2 years ago[PhaseOrdering] Update naming of blocks after 95f76bff1c40.
Florian Hahn [Sun, 13 Mar 2022 17:23:25 +0000 (17:23 +0000)]
[PhaseOrdering] Update naming of blocks after 95f76bff1c40.

2 years ago[X86] Update AVX512VL intrinsic tests to avoid adds
Simon Pilgrim [Sun, 13 Mar 2022 17:20:12 +0000 (17:20 +0000)]
[X86] Update AVX512VL intrinsic tests to avoid adds

As noticed in D119654, by adding the masked intrinsics results together we can end up with the selects being canonicalized away from the intrinsic - this isn't what we want to test here so replace with a insertvalue chain into a aggregate instead to retain all the results.

2 years ago[LV] Create & use VPScalarIVSteps for all scalar users.
Florian Hahn [Sun, 13 Mar 2022 17:15:24 +0000 (17:15 +0000)]
[LV] Create & use VPScalarIVSteps for all scalar users.

This patch is a follow-up to D115953. It updates optimizeInductions
to also introduce new VPScalarIVStepsRecipes if an IV has both vector
and scalar uses.

It updates all uses that only need scalar values to use the newly
created recipe for the scalar steps.

This completes untangling of VPWidenIntOrFpInductionRecipe
code-generation. Now the recipe *only* creates the widened vector
values, as it says on the tin.

The code to genereate IR has been moved directly to
VPWidenIntOrFpInductionRecipe::execute.

Note that the recipe has been updated to hold a reference to
ScalarEvolution, which is needed to expand the step, until we can place
the corresponding SCEV expansion in the pre-header.

Depends on D120827.

Reviewed By: Ayal

Differential Revision: https://reviews.llvm.org/D120828

2 years ago[clang-tidy][docs] Fix wrong url in DontModifyStdNamespaceCheck
Balazs Benics [Sun, 13 Mar 2022 16:51:00 +0000 (17:51 +0100)]
[clang-tidy][docs] Fix wrong url in DontModifyStdNamespaceCheck

It was probably a copy-paste mistake.
The check was added as `cert-dcl58-cpp`, so the doc should point there.

Reviewed By: aaron.ballman

Differential Revision: https://reviews.llvm.org/D121373

2 years ago[clang-tidy][docs][NFC] Update URL and docs of PostfixOperatorCheck
Balazs Benics [Sun, 13 Mar 2022 16:51:00 +0000 (17:51 +0100)]
[clang-tidy][docs][NFC] Update URL and docs of PostfixOperatorCheck

The docs URL was dangling, and the docs suggested that it has no fixits,
but it actually had.

Reviewed By: aaron.ballman

Differential Revision: https://reviews.llvm.org/D121372

2 years ago[clang-tidy][docs][NFC] Refer to the CERT rule in bugprone-shared-ptr-array-mismatch...
Balazs Benics [Sun, 13 Mar 2022 16:51:00 +0000 (17:51 +0100)]
[clang-tidy][docs][NFC] Refer to the CERT rule in bugprone-shared-ptr-array-mismatch docs

Document the connection between this checker and the corresponding CERT
rule.

Reviewed By: aaron.ballman

Differential Revision: https://reviews.llvm.org/D121214

2 years ago[libc++] Replace _LIBCPP_HAS_NO_CONCEPTS with _LIBCPP_STD_VER > 17. NFCI.
Joe Loser [Sat, 12 Mar 2022 15:46:57 +0000 (10:46 -0500)]
[libc++] Replace _LIBCPP_HAS_NO_CONCEPTS with _LIBCPP_STD_VER > 17. NFCI.

All supported compilers that support C++20 now support concepts. So, remove
`_LIB_LIBCPP_HAS_NO_CONCEPTS` in favor of `_LIBCPP_STD_VER > 17`. Similarly in
the tests, remove `// UNSUPPORTED: libcpp-no-concepts`.

Differential Revision: https://reviews.llvm.org/D121528

2 years ago[SDAG] simplify bitwise logic with repeated operand
Sanjay Patel [Sun, 13 Mar 2022 15:06:53 +0000 (11:06 -0400)]
[SDAG] simplify bitwise logic with repeated operand

We do not have general reassociation here (and probably
do not need it), but I noticed these were missing in
patches/tests motivated by D111530, so we can at
least handle the simplest patterns.

The VE test diff looks correct, but we miss that
pattern in IR currently:
https://alive2.llvm.org/ce/z/u66_PM

2 years ago[AArch64] add tests for bitwise logic reassociation; NFC
Sanjay Patel [Sun, 13 Mar 2022 14:29:27 +0000 (10:29 -0400)]
[AArch64] add tests for bitwise logic reassociation; NFC

Chooses from a variety of scalar/vector/illegal types
because that should not inhibit any folds.

2 years ago[InstCombine] Add additional icmp eq/ne test coverage for Issue #32161
Simon Pilgrim [Sun, 13 Mar 2022 14:40:34 +0000 (14:40 +0000)]
[InstCombine] Add additional icmp eq/ne test coverage for Issue #32161

2 years ago[MLIR][Presburger][NFC] Fix doc for PresburgerSpace::numLocals
Groverkss [Sun, 13 Mar 2022 14:34:23 +0000 (20:04 +0530)]
[MLIR][Presburger][NFC] Fix doc for PresburgerSpace::numLocals

2 years ago[PhaseOrdering] loop-rotation-vs-common-code-hoisting.ll - merge equivalent check...
Simon Pilgrim [Sun, 13 Mar 2022 14:15:35 +0000 (14:15 +0000)]
[PhaseOrdering] loop-rotation-vs-common-code-hoisting.ll - merge equivalent check-prefixes

2 years ago[MLIR][NFC] Fix deprecation message.
Christian Sigg [Sun, 13 Mar 2022 13:17:08 +0000 (14:17 +0100)]
[MLIR][NFC] Fix deprecation message.

2 years ago[MLIR][Presburger] Add support for PresburgerRelation
Groverkss [Sun, 13 Mar 2022 11:36:10 +0000 (17:06 +0530)]
[MLIR][Presburger] Add support for PresburgerRelation

This patch adds supports for union of relations (PresburgerRelation).  Along
with this, support for PresburgerSet is also maintained.

This patch is part of a series of patches to add support for relations in
Presburger library.

Reviewed By: arjunp

Differential Revision: https://reviews.llvm.org/D121417

2 years ago[X86] combineCMP - peek through zero-extensions for X86cmp(zext(x0),0) zero tests...
Simon Pilgrim [Sun, 13 Mar 2022 11:38:40 +0000 (11:38 +0000)]
[X86] combineCMP - peek through zero-extensions for X86cmp(zext(x0),0) zero tests (PR38960)

If we're comparing a value against zero, strip away any zero-extension and perform the comparison on the pre-extended value

Fixes #38308

Differential Revision: https://reviews.llvm.org/D121472

2 years ago[RISCV][RVV] Fix vslide1up/down intrinsics overflow bug for SEW=64 on RV32
Lehua Ding [Sun, 13 Mar 2022 10:06:09 +0000 (18:06 +0800)]
[RISCV][RVV] Fix vslide1up/down intrinsics overflow bug for SEW=64 on RV32

Reviewed By: craig.topper, kito-cheng

Differential Revision: https://reviews.llvm.org/D120899

2 years ago[AMDGPU] Add llvm.amdgcn.s.setprio intrinsic
Austin Kerbow [Fri, 4 Mar 2022 07:59:26 +0000 (23:59 -0800)]
[AMDGPU] Add llvm.amdgcn.s.setprio intrinsic

Reviewed By: rampitec, arsenm

Differential Revision: https://reviews.llvm.org/D120976

2 years ago[NFC][flang][OpenMP] Fixes formatting issues with D110714
Nimish Mishra [Sun, 13 Mar 2022 11:30:23 +0000 (17:00 +0530)]
[NFC][flang][OpenMP] Fixes formatting issues with D110714

This NFC fixes formatting issues introduced with https://reviews.llvm.org/D110714

Reviewed By: peixin, shraiysh

Differential Revision: https://reviews.llvm.org/D121186

2 years ago[MLIR][Python] Add SCFIfOp Python binding
chhzh123 [Sun, 13 Mar 2022 05:24:00 +0000 (05:24 +0000)]
[MLIR][Python] Add SCFIfOp Python binding

Current generated Python binding for the SCF dialect does not allow
users to call IfOp to create if-else branches on their own.
This PR sets up the default binding generation for scf.if operation
to address this problem.

Reviewed By: ftynse

Differential Revision: https://reviews.llvm.org/D121076

2 years ago[RISCV] Add DAGCombine to fold (bitreverse (bswap X)) to brev8 with Zbkb.
Craig Topper [Sun, 13 Mar 2022 00:23:17 +0000 (16:23 -0800)]
[RISCV] Add DAGCombine to fold (bitreverse (bswap X)) to brev8 with Zbkb.

If the type is less than XLenVT, type legalization will turn this
into (srl (bitreverse (bswap (srl (bswap X), C))), C). We can't
completely recover from these shifts. They introduce zeros into
the upper bits of the result and we can't easily tell if they are
needed. By doing a DAG combine early, we avoid introducing these
shifts.

2 years ago[RISCV] Add Zbp command lines to bswap-bitreverse.ll. NFC
Craig Topper [Sun, 13 Mar 2022 00:13:54 +0000 (16:13 -0800)]
[RISCV] Add Zbp command lines to bswap-bitreverse.ll. NFC

2 years ago[flang] Improve runtime crash messages
Peter Steinfeld [Fri, 11 Mar 2022 22:01:07 +0000 (14:01 -0800)]
[flang] Improve runtime crash messages

Where possible, I added additional information to the messages to help
programmers figure out what went wrong.  I also removed all uses of the word
"bad" from the messages since (to me) that implies a moral judgement rather
than a programming error.  I replaced it with either "invalid" or "unsupported"
where appropriate.

Differential Revision: https://reviews.llvm.org/D121493

2 years ago[ELF] Simplify addCopyRelSymbol with invokeELFT. NFC
Fangrui Song [Sat, 12 Mar 2022 22:08:10 +0000 (14:08 -0800)]
[ELF] Simplify addCopyRelSymbol with invokeELFT. NFC

2 years ago[clang-format][docs] Fix incorrect 'clang-format 12' option markers
Krystian Kuzniarek [Sat, 12 Mar 2022 21:49:43 +0000 (22:49 +0100)]
[clang-format][docs] Fix incorrect 'clang-format 12' option markers

Introduced by 23a5090c6, some style option markers indicated
'clang-format 12', though their respective options were available in
earlier releases.

Differential Revision: https://reviews.llvm.org/D120631

2 years ago[clang-format] Handle builtins in constraint expression
Björn Schäpers [Tue, 1 Mar 2022 22:15:09 +0000 (23:15 +0100)]
[clang-format] Handle builtins in constraint expression

Fixes https://github.com/llvm/llvm-project/issues/54106

Differential Revision: https://reviews.llvm.org/D120774

2 years ago[DebugInfo] Include DW_TAG_skeleton_unit when looking for parent UnitDie
Wenlei He [Sat, 26 Feb 2022 07:09:45 +0000 (23:09 -0800)]
[DebugInfo] Include DW_TAG_skeleton_unit when looking for parent UnitDie

`DIE::getUnitDie` looks up parent DIE until compile unit or type unit is found. However for skeleton CU with debug fission, we would have DW_TAG_skeleton_unit instead of DW_TAG_compile_unit as top level DIE.

This change fixes the look up so we can get DW_TAG_skeleton_unit as UnitDie for skeleton CU.

Differential Revision: https://reviews.llvm.org/D120610

2 years ago[mlirTranslateMain] Add a customization callback.
Chris Lattner [Fri, 4 Mar 2022 05:11:37 +0000 (21:11 -0800)]
[mlirTranslateMain] Add a customization callback.

mlir-translate and related tools currently have a fixed set
of flags that are built into Translation.cpp.  This works for
simple cases, but some clients want to change the default
globally (e.g. default to allowing unregistered dialects
without a command line flag), or support dialect-independent
translations without having those translations register every
conceivable dialect they could be used with (breaking
modularity).

This approach could also be applied to mlirOptMain to reduce
the significant number of flags it has accumulated.

Differential Revision: https://reviews.llvm.org/D120970

2 years agoReplace forward declaration by actual declaration of MemoryBuffer in Object/Binary.h
serge-sans-paille [Sat, 12 Mar 2022 20:53:14 +0000 (21:53 +0100)]
Replace forward declaration by actual declaration of MemoryBuffer in Object/Binary.h

This is a partial undo of ed98c1b37661b0795a5e34517941485f0f0688d1, see
https://lab.llvm.org/buildbot#builders/37/builds/11529
for the actual error.

2 years agoAdd missing include under EXPENSIVE_CHECKS
serge-sans-paille [Sat, 12 Mar 2022 17:53:53 +0000 (18:53 +0100)]
Add missing include under EXPENSIVE_CHECKS

2 years ago[clang-format] Minimize the damage caused by AlignArrayOfStructures when working...
mydeveloperday [Sat, 12 Mar 2022 17:21:33 +0000 (17:21 +0000)]
[clang-format] Minimize the damage caused by AlignArrayOfStructures when working on non square arrays

I have lost count of the number of times this has been reported, but it fundamentally comes down to the fact that the "AlignArrayLeft/Right" function is fundamentally broken for non-square arrays.

As a result, a pointer can end up running off the end of the array structure, I've spent the last 2 weekends trying to rewrite this algorithm but I've struggled to get it aligned correctly.

This is an interim fix, that ignores all array that are non-square and leaves them alone. I think this can allow us to close out most of the crashes (if not all).

I think this can help reduce the number of bugs coming in that are duplicates.

https://github.com/llvm/llvm-project/issues/53748
https://github.com/llvm/llvm-project/issues/51767
https://github.com/llvm/llvm-project/issues/51277

Reviewed By: curdeius, HazardyKnusperkeks, feg208

Differential Revision: https://reviews.llvm.org/D121069

2 years ago[gn build] Port 999fa9f68788
LLVM GN Syncbot [Sat, 12 Mar 2022 17:10:56 +0000 (17:10 +0000)]
[gn build] Port 999fa9f68788

2 years ago[BOLT][NFC] Use getShortOpcodeArith in X86MCPlusBuilder
Amir Ayupov [Thu, 10 Mar 2022 20:21:39 +0000 (12:21 -0800)]
[BOLT][NFC] Use getShortOpcodeArith in X86MCPlusBuilder

Unify `llvm::X86::getRelaxedOpcodeArith` and `getShortArithOpcode` in
X86MCPlusBuilder.cpp.

Addresses https://lists.llvm.org/pipermail/llvm-dev/2022-January/154526.html

Reviewed By: rafauler

Differential Revision: https://reviews.llvm.org/D121404

2 years ago[X86][NFC] Move table from getRelaxedOpcodeArith into its own class
Amir Ayupov [Fri, 21 Jan 2022 07:34:53 +0000 (23:34 -0800)]
[X86][NFC] Move table from getRelaxedOpcodeArith into its own class

Move out the table and prepare the code to reuse it for the reverse mapping.
Follows the example of memory folding/unfolding tables in X86InstrFoldTables.cpp

Preparation step to unify `llvm::X86::getRelaxedOpcodeArith` and
`getShortArithOpcode` in BOLT X86MCPlusBuilder.cpp.

Addresses https://lists.llvm.org/pipermail/llvm-dev/2022-January/154526.html

Reviewed By: skan, MaskRay

Differential Revision: https://reviews.llvm.org/D121402

2 years agoCleanup includes: DebugInfo & CodeGen
serge-sans-paille [Wed, 9 Mar 2022 21:29:31 +0000 (22:29 +0100)]
Cleanup includes: DebugInfo & CodeGen

Discourse thread: https://discourse.llvm.org/t/include-what-you-use-include-cleanup
Differential Revision: https://reviews.llvm.org/D121332

2 years ago[clang][deps] Update helper script documentation
Jan Svoboda [Sat, 12 Mar 2022 15:55:38 +0000 (16:55 +0100)]
[clang][deps] Update helper script documentation

Since D118986, clang-scan-deps returns the whole command line even for the original translation units.

2 years ago[clang][deps] Simplify PREFIX definitions in tests
Jan Svoboda [Sat, 12 Mar 2022 15:44:29 +0000 (16:44 +0100)]
[clang][deps] Simplify PREFIX definitions in tests

Instead of outputting the test directory into the JSON result file, parsing it with `FileCheck` and then potentially stripping it, simply use `FileCheck`'s `-D` option.

Note that we use `%/t` instead of `%t` in order to normalize to forward slashes on Windows, which matches what we do with `sed 's:\\\\\?:/:g'`.

Differential Revision: https://reviews.llvm.org/D121516

2 years ago[MLIR] Fix block label parsing bug
Uday Bondhugula [Sat, 12 Mar 2022 02:08:10 +0000 (07:38 +0530)]
[MLIR] Fix block label parsing bug

Fix bug in `Block` label parsing:
https://github.com/llvm/llvm-project/issues/54343

The `parseOptionalBlockArgList` method was doing the wrong thing
(contrary to its doc comment) and its calling context was also
incorrect. This led to a parse failure for something like "^bb0()".

Fixes #54343

Differential Revision: https://reviews.llvm.org/D121503

2 years ago[libc++] Remove unneeded tests.
Mark de Wever [Sat, 12 Mar 2022 11:15:26 +0000 (12:15 +0100)]
[libc++] Remove unneeded tests.

As suggested in D120742.

2 years ago[libc++][NFC] Use TEST_HAS_NO_EXCEPTIONS.
Mark de Wever [Sat, 12 Mar 2022 11:11:00 +0000 (12:11 +0100)]
[libc++][NFC] Use TEST_HAS_NO_EXCEPTIONS.

The test should use the already available test macro.

2 years ago[libc++] Add a missing test case.
Mark de Wever [Sat, 12 Mar 2022 11:07:54 +0000 (12:07 +0100)]
[libc++] Add a missing test case.

2 years ago[clang][deps] Attempt to fix Windows test failure
Jan Svoboda [Sat, 12 Mar 2022 11:09:10 +0000 (12:09 +0100)]
[clang][deps] Attempt to fix Windows test failure

2 years ago[clang][deps] Remove '-fmodules-cache-path=' arguments
Jan Svoboda [Sat, 12 Mar 2022 10:42:04 +0000 (11:42 +0100)]
[clang][deps] Remove '-fmodules-cache-path=' arguments

With explicit modules build, the '-fmodules-cache-path=' argument is unused.

This patch removes the argument to avoid warnings or errors (with '-Werror') stemming from that.

Depends on D118915.

Reviewed By: dexonsmith

Differential Revision: https://reviews.llvm.org/D120474

2 years ago[clang][deps] Generate '-fmodule-file=' only for direct dependencies
Jan Svoboda [Sat, 12 Mar 2022 10:32:47 +0000 (11:32 +0100)]
[clang][deps] Generate '-fmodule-file=' only for direct dependencies

The `clang-scan-deps` tool currently generates `-fmodule-file=` command-line arguments for the whole transitive closure of modular dependencies. This is not necessary, we only need to provide the direct dependencies on the command line. Information about transitive dependencies is stored within the `.pcm` files of direct dependencies. This makes the command lines shorter, but should be a NFC otherwise (unless there are bugs in the loading mechanism for explicit modules).

Depends on D120465.

Reviewed By: Bigcheese

Differential Revision: https://reviews.llvm.org/D118915

2 years ago[clang][deps] Disable implicit module maps
Jan Svoboda [Sat, 12 Mar 2022 10:04:17 +0000 (11:04 +0100)]
[clang][deps] Disable implicit module maps

Since D113473, we don't report any module map files via `-fmodule-map-file=` in explicit builds. The ultimate goal here is to make sure Clang doesn't open/read/parse/evaluate unnecessary module maps.

However, implicit module maps still end up reading all reachable module maps. This patch disables implicit module maps in explicit builds.

Unfortunately, we still need to report some module map files that aren't encoded in PCM files of dependencies: module maps that are necessary to correctly evaluate includes in modules marked as `[no_undeclared_includes]`.

Depends on D120464.

Reviewed By: dexonsmith

Differential Revision: https://reviews.llvm.org/D120465

2 years ago[Attributor] Remove capture tracker usage and follow uses explicitly
Johannes Doerfert [Tue, 8 Mar 2022 21:49:26 +0000 (15:49 -0600)]
[Attributor] Remove capture tracker usage and follow uses explicitly

Before we used the capture tracker to follow pointer uses, now we do it
explicitly ourselves through the Attributor API. There are multiple
benefits: For one, the boilerplate is cut down by a lot. The class,
potential copies vector, etc. is all not needed anymore. We also do
avoid explicitly looking through memory here, something that was
duplicated and should only live in the `checkForAllUses~ helper. More
importantly, as we do simplifications we need to make sure all parties
are in sync when they reason about uses. The old way did not allow us to
do this but the new one does as every use visiting AA goes through
`checkForAllUses` now..

2 years ago[CaptureTracking][NFCI] Expose capture tracking logic
Johannes Doerfert [Tue, 8 Mar 2022 21:02:36 +0000 (15:02 -0600)]
[CaptureTracking][NFCI] Expose capture tracking logic

The logic exposed by this patch via `llvm::DetermineUseCaptureKind` was
part of `llvm::PointerMayBeCaptured`. In the Attributor we want to keep
track of the work list items but still reuse the logic if a use might
capture a value. A follow up for the Attributor removes ~100 lines of
code and complexity while making future handling of simplified values
possible.

Differential Revision: https://reviews.llvm.org/D121272

2 years ago[lsan] ppc64: dynamically pick address range for allocator
Fangrui Song [Sat, 12 Mar 2022 04:16:17 +0000 (20:16 -0800)]
[lsan] ppc64: dynamically pick address range for allocator

For ppc64 PIE, it seems that [0xa00000000000,0xc00000000000) may be occupied
which will lead to a segfault in certain kernel configurations
(clang-ppc64le-rhel). Use the `!kUsingConstantSpaceBeg` code path like Fuchsia.

Reviewed By: jsji

Differential Revision: https://reviews.llvm.org/D121257

2 years ago[MC] Fix letter case of some MCSection member functions
Fangrui Song [Sat, 12 Mar 2022 04:07:00 +0000 (20:07 -0800)]
[MC] Fix letter case of some MCSection member functions

2 years ago[Attributor] Avoid replacing return operands twice
Johannes Doerfert [Fri, 25 Feb 2022 22:57:18 +0000 (16:57 -0600)]
[Attributor] Avoid replacing return operands twice

As replacements will become more complex it is better to have a single
AA responsible for replacing a use. Before this patch AAValueSimplify*
and AAValueSimplifyReturned could both try to replace the returned
value. The latter was marginally better for the old pass manager
when a function was already carrying a `returned` attribute and when
the context of the return instruction was important. The second
shortcoming was resolved by looking for return attributes in the
AAValueSimplifyCallSiteReturned initialization. The old PM impact is
not concerning.

This is yet another step towards the removal of AAReturnedValues, the
very first AA we should now try to eliminate due to the overlapping
logic with value simplification.

2 years ago[MC] Destroy WasmAllocator in MCContext::reset
Fangrui Song [Sat, 12 Mar 2022 03:55:07 +0000 (19:55 -0800)]
[MC] Destroy WasmAllocator in MCContext::reset

To not leave lingering getWasmSection instances.

2 years ago[RISCV] Move GORCIW/GREVIW formation to isel patterns.
Craig Topper [Sat, 12 Mar 2022 01:10:03 +0000 (17:10 -0800)]
[RISCV] Move GORCIW/GREVIW formation to isel patterns.

Type legalize narrow RISCVISD::GREV/GORC with constant to a larger
type without switching to W. Detect sext_inreg+gorci/grevi with a
uimm5 immediate during isel to emit GREVIW/GORCIW.

This allows us to better propagate known bits information through
extended bits after type legalization. It will also simplify a
change I'm considering for BREV8 with Zbkb.

A future patch will add computeKnownBits support for GORC.

A further improvement here would be to use hasAllWUsers and
doPeepholeSExtW like we do for SLLIW, but I don't think we have
the test coverage for that yet.

2 years ago[RISCV] Add test cases that show that we're too aggressive about using greviw/gorciw...
Craig Topper [Fri, 11 Mar 2022 23:29:26 +0000 (15:29 -0800)]
[RISCV] Add test cases that show that we're too aggressive about using greviw/gorciw. NFC

We currently type legalize to the W form, but type legalization
doesn't place any requirements on the extended bits. So we are ok
to use GREVI/GORCI for type legalization as long as the control
doesn't cross any bits from the extended bits into the lower bits.

This can allow us to recognize cases where the extended bits end
up being all zeros and we can propagate that information through.

My plan is to move greviw/gorciw formation to isel patterns similar
to slliw.

2 years ago[gn build] Port ee0f8c401030
LLVM GN Syncbot [Sat, 12 Mar 2022 00:46:45 +0000 (00:46 +0000)]
[gn build] Port ee0f8c401030

2 years ago[libc++][ranges] Implement ranges::find{, _if, _if_not}
Nikolas Klauser [Sat, 12 Mar 2022 00:45:35 +0000 (01:45 +0100)]
[libc++][ranges] Implement ranges::find{, _if, _if_not}

Reviewed By: var-const, #libc, ldionne

Spies: ldionne, tcanens, libcxx-commits, mgorny

Differential Revision: https://reviews.llvm.org/D121248

2 years ago[Attributor][FIX] Make sure to not ignore non-load users of stores
Johannes Doerfert [Sat, 12 Mar 2022 00:39:21 +0000 (18:39 -0600)]
[Attributor][FIX] Make sure to not ignore non-load users of stores

When we look through memory for a store we used to allow any other use
of the memory that is reachable. This is generally OK but we need to
make sure to actually let the user look at these properly. For now,
we simply require loads (via exact reloads).

2 years ago[mlir] Update function referenced in comment (NFC)
Jacques Pienaar [Sat, 12 Mar 2022 00:31:54 +0000 (16:31 -0800)]
[mlir] Update function referenced in comment (NFC)

2 years ago[builtins] Do not force thumb mode directive in arm/sync-ops.h
Khem Raj [Sat, 12 Mar 2022 00:25:49 +0000 (16:25 -0800)]
[builtins] Do not force thumb mode directive in arm/sync-ops.h

.thumb_func was not switching mode until [1]
so it did not show up but now that .thumb_func (without argument) is
switching mode, its causing build failures on armv6 ( rpi0 ) even when
build is explicitly asking for this file to be built with -marm (ARM
mode), therefore use DEFINE_COMPILERRT_FUNCTION macro to add function
header which considers arch and mode from compiler cmdline to decide if
the function is built using thumb mode or arm mode.

[1] https://reviews.llvm.org/D101975

Note that it also needs https://reviews.llvm.org/D99282

Reviewed By: peter.smith, MaskRay

Differential Revision: https://reviews.llvm.org/D104183

2 years ago[gn build] Port b1b4b6f36695
LLVM GN Syncbot [Sat, 12 Mar 2022 00:10:18 +0000 (00:10 +0000)]
[gn build] Port b1b4b6f36695

2 years ago[Clang][VE] Add vector load intrinsics
Kazushi (Jam) Marukawa [Sat, 5 Mar 2022 16:17:04 +0000 (01:17 +0900)]
[Clang][VE] Add vector load intrinsics

Add vector load intrinsic instructions for VE.

Reviewed By: simoll

Differential Revision: https://reviews.llvm.org/D121049

2 years agoRename EHFrameSplitter to DWARFRecordSectionSplitter
Shubham Sandeep Rastogi [Fri, 11 Mar 2022 20:14:05 +0000 (12:14 -0800)]
Rename EHFrameSplitter to DWARFRecordSectionSplitter

EHFrameSplitter does the exact same work to split up the eh_frame as it would need for any section that follows the DWARF record, therefore this patch just changes the name of it to DWARFRecordSectionSplitter to be more general.

Differential Revision: https://reviews.llvm.org/D121486

2 years ago[CMake] Rename TARGET_TRIPLE to LLVM_TARGET_TRIPLE
Petr Hosek [Thu, 2 Dec 2021 09:04:46 +0000 (01:04 -0800)]
[CMake] Rename TARGET_TRIPLE to LLVM_TARGET_TRIPLE

This clarifies that this is an LLVM specific variable and avoids
potential conflicts with other projects.

Differential Revision: https://reviews.llvm.org/D119918

2 years ago[nfc][openmp] Swap arguments to remove noise from upcoming diff
Jon Chesterfield [Fri, 11 Mar 2022 23:07:11 +0000 (23:07 +0000)]
[nfc][openmp] Swap arguments to remove noise from upcoming diff

2 years ago[InlineCost] Add cl::opt for target attributes compatibility check. NFC
Anna Thomas [Fri, 11 Mar 2022 21:47:40 +0000 (16:47 -0500)]
[InlineCost] Add cl::opt for target attributes compatibility check. NFC

This patch adds a CL option for avoiding the attribute compatibility
check between caller and callee in TTI. TTI attribute compatibility
checks for target CPU and target features.
In our downstream compiler, this attribute always remains the same
between callee and caller. By avoiding the addition of this attribute to
each of our inline candidate (and then checking them here during inline
cost), we save some compile time.

The option is kept false, so this change is an NFC upstream.

2 years ago[NFC] Improve comment and rename test file
Julian Lettner [Fri, 11 Mar 2022 22:23:11 +0000 (14:23 -0800)]
[NFC] Improve comment and rename test file

2 years ago[Attributor] Allow not to default initialize AAs for live internal functions
Johannes Doerfert [Fri, 11 Mar 2022 18:51:24 +0000 (12:51 -0600)]
[Attributor] Allow not to default initialize AAs for live internal functions

Outside users of the Attributor, e.g., OpenMP-opt, want to seed AAs
themselves. We should not seed all default AAs one an internal function
becomes live. That said, there should be a callback such that they can
do lazy seeding as well.

Differential Revision: https://reviews.llvm.org/D121489

2 years ago[Attributor] Cleanup manifest and liveness for CGSCC passes
Johannes Doerfert [Thu, 10 Mar 2022 18:17:38 +0000 (12:17 -0600)]
[Attributor] Cleanup manifest and liveness for CGSCC passes

There was some ad-hoc handling of liveness and manifest to avoid
breaking CGSCC guarantees. Things always slipped through though.
This cleanup will:

1) Prevent us from manifesting any "information" outside the CGSCC.
   This might be too conservative but we need to opt-in to annotation
   not try to avoid some problematic ones.
2) Avoid running any liveness analysis outside the CGSCC. We did have
   some AAIsDeadFunction handling to this end but we need this for all
   AAIsDead classes. The reason is that AAIsDead information is only
   correct if we actually manifest it, since we don't (see point 1) we
   cannot actually derive/use it at all. We are currently trying to
   avoid running any AA updates outside the CGSCC but that seems to
   impact things quite a bit.
3) Assert, don't check, that our modifications (during cleanup) modifies
   only CGSCC functions.

2 years ago[MC][test] Improve compression.s
Fangrui Song [Fri, 11 Mar 2022 22:42:07 +0000 (14:42 -0800)]
[MC][test] Improve compression.s

* Actually test compressed .debug_line for 32-bit
* Test .debug_frame can be compressed
* Test a section not named .debug_* is not compressed

2 years ago[clang-format][NFC] Group QualifierAlignment with other C++ passes
owenca [Fri, 11 Mar 2022 02:16:12 +0000 (18:16 -0800)]
[clang-format][NFC] Group QualifierAlignment with other C++ passes

Also increases the initial size of Passes to 8 and move the
definition of BracesInserter up.

Differential Revision: https://reviews.llvm.org/D121434

2 years ago[clang-tidy] Refactor: Prefer single quoted strings over double quoted strings [NFC]
Richard [Mon, 3 Jan 2022 18:38:42 +0000 (11:38 -0700)]
[clang-tidy] Refactor: Prefer single quoted strings over double quoted strings [NFC]

2 years ago[MC] Remove unneeded zlib opt-out for .debug_frame
Fangrui Song [Fri, 11 Mar 2022 21:32:24 +0000 (13:32 -0800)]
[MC] Remove unneeded zlib opt-out for .debug_frame

The opt-out from rL236267 (2015) is untested and seems no longer needed (or not
needed when rL236267 was committed): there is nothing special with uncompressed
alignment. This brings us in line with GCC which compresses .debug_frame .

Checked that -g -fno-asynchronous-unwind-tables + objcopy
--decompress-debug-sections output is identical to -g
-fno-asynchronous-unwind-tables -gz + objcopy --decompress-debug-sections
output.

2 years ago[lldb] Removed scoped timer from ReadMemoryFromFileCache
Dave Lee [Fri, 11 Mar 2022 20:45:47 +0000 (12:45 -0800)]
[lldb] Removed scoped timer from ReadMemoryFromFileCache

`ReadMemoryFromFileCache` can be called at a high rate, and has fast execution.
Signposts for high rate & brief duration can have a negative impact on tracing;
emitting a high volume signposts can lead to blocking, affecting performance,
and total volume makes human review of the trace harder because of the noise.

Differential Revision: https://reviews.llvm.org/D121226

2 years ago[lldb] Move FindTypes scoped timer to SymbolFileDWARFDebugMap
Dave Lee [Fri, 11 Mar 2022 19:16:06 +0000 (11:16 -0800)]
[lldb] Move FindTypes scoped timer to SymbolFileDWARFDebugMap

2 years ago[lldb] Include lldb/Host/HostInfo.h in ProcessWindows.cpp
Jonas Devlieghere [Fri, 11 Mar 2022 20:34:41 +0000 (12:34 -0800)]
[lldb] Include lldb/Host/HostInfo.h in ProcessWindows.cpp

2 years ago[AMDGPU] Support v_mov_b64 in dpp combine
Stanislav Mekhanoshin [Thu, 10 Mar 2022 21:46:55 +0000 (13:46 -0800)]
[AMDGPU] Support v_mov_b64 in dpp combine

Differential Revision: https://reviews.llvm.org/D121411

2 years ago[AMDGPU] gfx940 MUBUF format changes
Stanislav Mekhanoshin [Tue, 8 Mar 2022 18:28:23 +0000 (10:28 -0800)]
[AMDGPU] gfx940 MUBUF format changes

Differential Revision: https://reviews.llvm.org/D121234

2 years ago[lldb] Fix typo in ProcessWindows.h
Jonas Devlieghere [Fri, 11 Mar 2022 19:32:21 +0000 (11:32 -0800)]
[lldb] Fix typo in ProcessWindows.h

2 years ago[mlir][spirv] Make EntryPointABIAttr.local_size optional
Ivan Butygin [Wed, 23 Feb 2022 11:12:07 +0000 (14:12 +0300)]
[mlir][spirv] Make EntryPointABIAttr.local_size optional

* It doesn't required by OpenCL/Intel Level Zero and can be set programmatically.
* Add GPU to spirv lowering in case when attribute is not present.
* Set higher benefit to WorkGroupSizeConversion pattern so it will always try to lower first from the attribute.

Differential Revision: https://reviews.llvm.org/D120399

2 years ago[CommandLine] Remove `may only occur zero or one times!` error
Fangrui Song [Fri, 11 Mar 2022 19:25:04 +0000 (11:25 -0800)]
[CommandLine] Remove `may only occur zero or one times!` error

Early adoption of new technologies or adjusting certain code generation/IR optimization thresholds
is often available through some cl::opt options (which have unstable surfaces).
Specifying such an option twice will lead to an error.

```
% clang -c a.c -mllvm -disable-binop-extract-shuffle -mllvm -disable-binop-extract-shuffle
clang (LLVM option parsing): for the --disable-binop-extract-shuffle option: may only occur zero or one times!
% clang -c a.c -mllvm -hwasan-instrument-reads=0 -mllvm -hwasan-instrument-reads=0
clang (LLVM option parsing): for the --hwasan-instrument-reads option: may only occur zero or one times!
% clang -c a.c -mllvm --scalar-evolution-max-arith-depth=32 -mllvm --scalar-evolution-max-arith-depth=16
clang (LLVM option parsing): for the --scalar-evolution-max-arith-depth option: may only occur zero or one times!
```

The option is specified twice, because there is sometimes a global setting and
a specific file or project may need to override (or duplicately specify) the
value.

The error is contrary to the common practice of getopt/getopt_long command line
utilities that let the last option win and the `getLastArg` behavior used by
Clang driver options. I have seen such errors for several times. I think the
error just makes users inconvenient, while providing very little value on
discouraging production usage of unstable surfaces (this goal is itself
controversial, because developers might not want to commit to a stable surface
too early, or there is just some subtle codegen toggle which is infeasible to
have a driver option). Therefore, I suggest we drop the diagnostic, at least
before the diagnostic gets sufficiently better support for the overridding needs.

Removing the error is a degraded error checking experience. I think this error
checking behavior, if desirable, should be enabled explicitly by tools. Users
preferring the behavior can figure out a way to do so.

Reviewed By: jhenderson, rnk

Differential Revision: https://reviews.llvm.org/D120455

2 years ago[lldb] Add a getter for the process' system architecture
Jonas Devlieghere [Fri, 11 Mar 2022 18:50:10 +0000 (10:50 -0800)]
[lldb] Add a getter for the process' system architecture

This patch adds a getter for the process' system architecture. I went
with Process::GetSystemArchitecture to match
Platform::GetSystemArchitecture.

Differential revision: https://reviews.llvm.org/D121443

2 years ago[lld-macho] Don't allocate memory in parallelForEach
Jez Ng [Fri, 11 Mar 2022 18:31:20 +0000 (13:31 -0500)]
[lld-macho] Don't allocate memory in parallelForEach

... since BumpPtrAllocator isn't thread-safe.

Reviewed By: #lld-macho, Roger

Differential Revision: https://reviews.llvm.org/D121458

2 years ago[mlir][memref] memref.reinterpret_cast folding
Ivan Butygin [Mon, 21 Feb 2022 12:16:34 +0000 (15:16 +0300)]
[mlir][memref] memref.reinterpret_cast folding

* reinterpret_cast(reinterpret_cast(x)) -> reinterpret_cast(x)
* reinterpret_cast(cast(x)) -> reinterpret_cast(x)
* reinterpret_cast(subview(x)) -> reinterpret_cast(x) if subview offsets are 0

Differential Revision: https://reviews.llvm.org/D120242

2 years ago[LoopUnswitch] Use SmallPtrSet instead of std::set. NFCI.
Benjamin Kramer [Fri, 11 Mar 2022 18:13:58 +0000 (19:13 +0100)]
[LoopUnswitch] Use SmallPtrSet instead of std::set. NFCI.

2 years ago[VPlan] Implement VPCanonicalIVPHIRecipe::onlyFirstLaneUsed.
Florian Hahn [Fri, 11 Mar 2022 17:56:54 +0000 (17:56 +0000)]
[VPlan] Implement VPCanonicalIVPHIRecipe::onlyFirstLaneUsed.

The recipe only uses the first lane of its operands.

Suggested & split off D120827.

2 years ago[RISCV] Optimize vfmv.s.f intrinsic with scalar 0.0 to vmv.s.x with x0.
Craig Topper [Fri, 11 Mar 2022 01:56:17 +0000 (17:56 -0800)]
[RISCV] Optimize vfmv.s.f intrinsic with scalar 0.0 to vmv.s.x with x0.

We already do this for RISCVISD::VFMV_S_F_VL and the vfmv.v.f
intrinsic.

Reviewed By: kito-cheng

Differential Revision: https://reviews.llvm.org/D121429

2 years ago[flang][NFC] rename IsKindParameterizedDerivedType and fix comment typos
Jean Perier [Fri, 11 Mar 2022 18:03:19 +0000 (19:03 +0100)]
[flang][NFC] rename IsKindParameterizedDerivedType and fix comment typos

Following post-review feedback on https://reviews.llvm.org/D120804 and
https://reviews.llvm.org/D120801 about type descriptor changes, fix typos in
comments and rename IsKindParameterizedDerivedType to
IsDerivedTypeWithKindParameter. Remove a useless `;`.

Reviewed By: clementval, PeteSteinfeld

Differential Revision: https://reviews.llvm.org/D121470

2 years agoReturn an error when dsymutil might produce an invalid mach-o file.
Greg Clayton [Thu, 10 Mar 2022 19:29:45 +0000 (11:29 -0800)]
Return an error when dsymutil might produce an invalid mach-o file.

64 bit mach-o files have sections that only have 32 bit file offsets. If dsymutil tries to produce an invalid mach-o file, then error out with a good error string.

Differential Revision: https://reviews.llvm.org/D121398

2 years ago[flang] Add tests for allocatable global
Valentin Clement [Fri, 11 Mar 2022 17:39:33 +0000 (18:39 +0100)]
[flang] Add tests for allocatable global

This patch adds couple of tests for allocatable
globals and missing lowering for them

This patch is part of the upstreaming effort from fir-dev branch.

Reviewed By: PeteSteinfeld

Differential Revision: https://reviews.llvm.org/D121473

Co-authored-by: Jean Perier <jperier@nvidia.com>
Co-authored-by: Eric Schweitz <eschweitz@nvidia.com>
2 years ago[flang] Lower of elemental calls in array expression
Valentin Clement [Fri, 11 Mar 2022 17:38:08 +0000 (18:38 +0100)]
[flang] Lower of elemental calls in array expression

This patch adds tests and missing lowering
code to lower elemental function/subroutine calls
in array expression

This patch is part of the upstreaming effort from fir-dev branch.

Reviewed By: PeteSteinfeld

Differential Revision: https://reviews.llvm.org/D121474

Co-authored-by: Jean Perier <jperier@nvidia.com>
Co-authored-by: Eric Schweitz <eschweitz@nvidia.com>
2 years ago[NFC][Clang][OpaquePtr] Remove calls to Address::deprecated in a couple
Akira Hatanaka [Fri, 11 Mar 2022 17:30:31 +0000 (09:30 -0800)]
[NFC][Clang][OpaquePtr] Remove calls to Address::deprecated in a couple
more files

Differential Revision: https://reviews.llvm.org/D121135

2 years ago[lldb] Don't overwrite the host arch with the process arch
Jonas Devlieghere [Fri, 11 Mar 2022 06:52:20 +0000 (22:52 -0800)]
[lldb] Don't overwrite the host arch with the process arch

Don't overwrite the host architecture (obtained from qHostInfo) with the
process info (obtained from qProcessInfo).

Differential revision: https://reviews.llvm.org/D121442

2 years ago[ASan] Export optimized callbacks.
Kirill Stoimenov [Fri, 11 Mar 2022 15:47:06 +0000 (15:47 +0000)]
[ASan] Export optimized callbacks.

This is a part of optimized callback reverts. This is needed to export the callbacks from the rt-asan libraries.

Reviewed By: kstoimenov

Differential Revision: https://reviews.llvm.org/D121464

2 years ago[X86] Add PR38960 comineqss test case
Simon Pilgrim [Fri, 11 Mar 2022 16:46:47 +0000 (16:46 +0000)]
[X86] Add PR38960 comineqss test case

This inverts the flags result compared to the original PR38960 comieqss test case

2 years ago[Bitcode] Don't confuse type attributes on declaration and call
Nikita Popov [Fri, 11 Mar 2022 16:30:34 +0000 (17:30 +0100)]
[Bitcode] Don't confuse type attributes on declaration and call

We should not be using APIs here that try to fetch the attribute
from both the call attributes and the function attributes. Otherwise
we'll try to upgrade a non-existent sret attribute on the call using
the attribute on the function.

2 years ago[AsmParser] Check forward reference type with opaque pointers
Nikita Popov [Fri, 11 Mar 2022 16:14:48 +0000 (17:14 +0100)]
[AsmParser] Check forward reference type with opaque pointers

While we don't need to check the element type in this case, we
do need to make sure that the pointers have the same address space,
otherwise RAUW will assert.

2 years ago[flang] Fix DYLIB builds
Andrzej Warzynski [Fri, 11 Mar 2022 14:02:30 +0000 (14:02 +0000)]
[flang] Fix DYLIB builds

https://reviews.llvm.org/D120568 broke builds that set
both `LLVM_BUILD_LLVM_DYLIB` and `LLVM_LINK_LLVM_DYLIB`. This patch
fixes that.

The build failure was caused by the fact that some LLVM libraries (which
are also LLVM components) were listed directly as link-time dependencies
instead of using `LINK_COMPONENTS` in CMake files. This lead to a linker
invocation like this (simplified version to demonstrate the problem):
```
ld lib/libLLVM.so lib/libLLVMAnalysis.a lib/libLLVMTarget.a
```
That's problematic and unnecessary because `libLLVM.so` incorporates
`libLLVMAnalysis` and `libLLVMTarget`. A correct invocation would look
like this (`LLVM_LINK_LLVM_DYLIB` _is not_ set):
```
ld  lib/libLLVMAnalysis.a lib/libLLVMTarget.a
```
or this (`LLVM_LINK_LLVM_DYLIB` _is_ set):
```
ld lib/libLLVM.so
```

Differential Revision: https://reviews.llvm.org/D121461

2 years ago[Attributor][FIX] Avoid double free (and useless state copy)
Johannes Doerfert [Fri, 11 Mar 2022 16:08:34 +0000 (10:08 -0600)]
[Attributor][FIX] Avoid double free (and useless state copy)

In an attempt to remove the memory leak we introduced a double free.
The problem was that we allowed a plain copy of the state and it was
actually used. The use was useless, so it is gone now. The copy
constructor is gone as well. The move constructor ensures the Accesses
pointers are owned by a single state, I hope.

Reported by: https://lab.llvm.org/buildbot/#/builders/16/builds/25820