platform/upstream/llvm.git
18 months ago[clang][AST] Fix `-Wuninitialized`. NFC
Michael Liao [Sun, 9 Apr 2023 19:55:38 +0000 (15:55 -0400)]
[clang][AST] Fix `-Wuninitialized`. NFC

- Adjust the declaration order as non-static member are initialized in
  order of declaration in the class definition.

18 months ago[LV] Update tests checking VPlans to use patterns for VPValues.
Florian Hahn [Sun, 9 Apr 2023 19:32:09 +0000 (20:32 +0100)]
[LV] Update tests checking VPlans to use patterns for VPValues.

This makes the tests more robust to changes in value numbering for
VPValues.

18 months ago[TableGen] Remove TypeSetByHwMode::isDefaultOnly(). Use InfoByHwMode<T>::isSimple...
Craig Topper [Sun, 9 Apr 2023 17:56:06 +0000 (10:56 -0700)]
[TableGen] Remove TypeSetByHwMode::isDefaultOnly(). Use InfoByHwMode<T>::isSimple(). NFC

InfoByHwMode is the base class of TypeSetByHwMode. The two methods
did the same thing. No reason to have two ways to do it.

Also use the getSimple() access instead of Map.begin()->second.

18 months ago[gn build] Port e8cfbfd05a95
LLVM GN Syncbot [Sun, 9 Apr 2023 15:57:26 +0000 (15:57 +0000)]
[gn build] Port e8cfbfd05a95

18 months ago[libc++] Granularize system_error.
Mark de Wever [Sat, 8 Apr 2023 15:29:31 +0000 (17:29 +0200)]
[libc++] Granularize system_error.

Reviewed By: #libc, philnik

Differential Revision: https://reviews.llvm.org/D147853

18 months ago[clang-tidy] fix false positve for namespace with attrs in modernize-concat-nested...
Congcong Cai [Sun, 9 Apr 2023 14:23:28 +0000 (16:23 +0200)]
[clang-tidy] fix false positve for namespace with attrs in modernize-concat-nested-namespaces

Fixed https://github.com/llvm/llvm-project/issues/57530
Add pre check to avoid false positive for namespace with attributes like
```
namespace [[deprecated]] ns {}
```

Reviewed By: PiotrZSL

Differential Revision: https://reviews.llvm.org/D147857

18 months ago[libc++] Implements isblank.
Mark de Wever [Sat, 8 Apr 2023 12:59:09 +0000 (14:59 +0200)]
[libc++] Implements isblank.

This omission seems to be there for a long time, it's in the initial
libc++ import. This was discovered while working on the std modules.

Reviewed By: #libc, philnik

Differential Revision: https://reviews.llvm.org/D147850

18 months ago[libc++][format] Addresses LWG3720.
Mark de Wever [Fri, 17 Feb 2023 20:27:08 +0000 (21:27 +0100)]
[libc++][format] Addresses LWG3720.

  LWG3720 Restrict the valid types of arg-id for width and precision in
  std-format-spec

Depends on D144325

Reviewed By: #libc, ldionne

Differential Revision: https://reviews.llvm.org/D144326

18 months ago[libc++][format] range-default-formatter for strings.
Mark de Wever [Thu, 5 May 2022 16:57:32 +0000 (18:57 +0200)]
[libc++][format] range-default-formatter for strings.

Implements the range-default-formatter specialization range_format::string
and range_format::debug_string.

Implements parts of
- P2286R8 Formatting Ranges
- P2585R0 Improving default container formatting

Depends on D145847

Reviewed By: ldionne, #libc

Differential Revision: https://reviews.llvm.org/D145853

18 months agoRevert (3) "[Assignment Tracking] Enable by default"
OCHyams [Sun, 9 Apr 2023 10:21:55 +0000 (11:21 +0100)]
Revert (3) "[Assignment Tracking] Enable by default"

Revert D146987.

This reverts commit f11e1475c97c1ba6b418838f3592de930677c3d0 which
reapplies D146987.

Buildbot: https://lab.llvm.org/buildbot/#/builders/70/builds/36103

18 months agoReapply (2) "[Assignment Tracking] Enable by default"
OCHyams [Sun, 9 Apr 2023 08:59:26 +0000 (09:59 +0100)]
Reapply (2) "[Assignment Tracking] Enable by default"

Re-land D146987.

This reverts commit 64fba207a683a355d9059cd57adb8a139c2f5dda
which reverts D146987.

18 months ago[Assignment Tracking] Fix fragments for assignments to variables smaller than the...
OCHyams [Sun, 9 Apr 2023 08:47:09 +0000 (09:47 +0100)]
[Assignment Tracking] Fix fragments for assignments to variables smaller than the alloca

Prior to this patch the trackAssignments function would attribute all stores to
an alloca to all variables linked to the alloca. This is wrong in the case
where the alloca contains variables which are smaller than the alloca, and
caused erroneous fragment information to be generated.

Now stores outside the variable bounds are discarded, and we check whether a
fragment is needed based on whether the store covers the entire variable as
opposed to whether it covers the entire alloca (except for variables of unknown
size).

Note that trackAssignments doesn't yet understand whole variables sitting at
anything other than offset 0 in an alloca - those variables are still tracked
using dbg.declares.

Fixes https://lab.llvm.org/buildbot/#/builders/70/builds/36007

Reviewed By: jmorse

Differential Revision: https://reviews.llvm.org/D147777

18 months ago[clang-tidy] Fix AST Library documentation link
Jorge Pinto Sousa [Sun, 9 Apr 2023 08:45:07 +0000 (08:45 +0000)]
[clang-tidy] Fix AST Library documentation link

The link was pointing to the 'Lexer and Preprocessor Library'
instead.

Reviewed By: PiotrZSL

Differential Revision: https://reviews.llvm.org/D147874

18 months ago[VPlan] Don't add live-outs if scalar epilogue is required.
Florian Hahn [Sun, 9 Apr 2023 08:15:42 +0000 (09:15 +0100)]
[VPlan] Don't add live-outs if scalar epilogue is required.

Instead of clearing live outs when a scalar epilogue is required late,
don't add live outs during VPlan construction if a scalar epilogue is
required.

This enables more VPlan-based DCE (if the live out would be the only
user in the plan) and is a step towards removing an access of the cost
model in fixedVectorizedLoop (which is after VPlan execution).

Depends on D147468.

Reviewed By: Ayal

Differential Revision: https://reviews.llvm.org/D147471

18 months ago[TableGen] Reorder some checks in TreePatternNode::isIsomorphicTo to speedup -gen...
Craig Topper [Sun, 9 Apr 2023 07:42:35 +0000 (00:42 -0700)]
[TableGen] Reorder some checks in TreePatternNode::isIsomorphicTo to speedup -gen-dag-isel for RISC-V.

Comparing types is quite expensive when hardware modes are being
used. Checking the operator first can let us detect mismatches
earlier without checking types.

18 months ago[libc++][spaceship] Implement `operator<=>` for `time_point`
Hristo Hristov [Sat, 1 Apr 2023 08:51:03 +0000 (11:51 +0300)]
[libc++][spaceship] Implement `operator<=>` for `time_point`

Depends on D145881

Implements parts of **P1614R2**: https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2019/p1614r2.html
Implements `operator<=>` for `time_point`

Reviewed By: #libc, Mordante

Differential Revision: https://reviews.llvm.org/D146250

18 months ago[TableGen] Slightly improve the efficiency of InfoByHwMode::get.
Craig Topper [Sun, 9 Apr 2023 01:54:56 +0000 (18:54 -0700)]
[TableGen] Slightly improve the efficiency of InfoByHwMode::get.

This avoids some double map lookups. Also use Map.begin() to find
the default mode instead of going through find.

18 months ago[X86][mem-fold][NFC] Simplify code by transform A != 0 - > A
Shengchen Kan [Sun, 9 Apr 2023 01:29:41 +0000 (09:29 +0800)]
[X86][mem-fold][NFC] Simplify code by transform A != 0 - > A

18 months ago[libc++] Remove <cstdlib> includes
Nikolas Klauser [Sun, 8 Jan 2023 15:47:53 +0000 (16:47 +0100)]
[libc++] Remove <cstdlib> includes

We changed the `abort` calls when trying to throw exceptions in `-fno-exceptions` mode to `__verbose_abort` calls, which removes the dependency in most files.

Reviewed By: ldionne, #libc

Spies: dim, emaste, mikhail.ramalho, smeenai, libcxx-commits

Differential Revision: https://reviews.llvm.org/D146076

18 months ago[ORC][MachO] Remove unused variables.
Lang Hames [Sat, 8 Apr 2023 23:48:13 +0000 (16:48 -0700)]
[ORC][MachO] Remove unused variables.

18 months ago[CodeGen] Use range-based for loops (NFC)
Kazu Hirata [Sat, 8 Apr 2023 23:22:39 +0000 (16:22 -0700)]
[CodeGen] Use range-based for loops (NFC)

18 months ago[ORC] Fix reference to struct member in f448d44663a.
Lang Hames [Sat, 8 Apr 2023 23:09:45 +0000 (16:09 -0700)]
[ORC] Fix reference to struct member in f448d44663a.

A map value was converted from a tuple to a struct during development, but I
missed a use in an assert.

18 months ago[SCEV][NFC] Fix `Do not use 'else' after 'return'`
Joshua Cao [Sat, 8 Apr 2023 22:55:21 +0000 (15:55 -0700)]
[SCEV][NFC] Fix `Do not use 'else' after 'return'`

Follow LLVM coding standards and make clangd emit less warnings.

18 months ago[ORC][ORC-RT][MachO] Use _objc_(map|load)_images for ObjC & Swift registration.
Lang Hames [Fri, 7 Apr 2023 21:40:05 +0000 (14:40 -0700)]
[ORC][ORC-RT][MachO] Use _objc_(map|load)_images for ObjC & Swift registration.

This patch drops the individual registration calls to the ObjC and Swift
runtimes (for selectors, classes, etc.), and instead creates a Mach header and
load commands that can be passed to _objc_map_images and _objc_load_images to
trigger registration and execution of +load methods. This approach supports
categories (for which there is no current registration API), and more closely
follows dyld's ObjC & Swift registration path.

18 months ago[mlir] Use DenseMap::contains (NFC)
Kazu Hirata [Sat, 8 Apr 2023 22:44:18 +0000 (15:44 -0700)]
[mlir] Use DenseMap::contains (NFC)

18 months ago[SCEV] Add test for huge trip multiples
Joshua Cao [Sat, 8 Apr 2023 22:14:12 +0000 (15:14 -0700)]
[SCEV] Add test for huge trip multiples

18 months ago[AMDGPU] Update mcp-overlap-after-propagation.mir test
Sergei Barannikov [Sat, 8 Apr 2023 22:08:41 +0000 (01:08 +0300)]
[AMDGPU] Update mcp-overlap-after-propagation.mir test

The test was added in D69953, but since D67794 landed a bit later it no
longer catches the bug. Update the test to be representative again.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D146934

18 months ago[ORC][LLJIT] Improve debugging output, tighten assertion.
Lang Hames [Sat, 8 Apr 2023 17:38:52 +0000 (10:38 -0700)]
[ORC][LLJIT] Improve debugging output, tighten assertion.

Clients can provide an ExecutionSession or an ExecutorProcessControl object to
LLJITBuilder, but should not provide both.

18 months ago[TableGen] Reimplement union_modes for InfoByHwMode to be avoid a secondary set.
Craig Topper [Sat, 8 Apr 2023 19:46:15 +0000 (12:46 -0700)]
[TableGen] Reimplement union_modes for InfoByHwMode to be avoid a secondary set.

Previously we collected both modes in a set and iterated over that set.

Instead, iterate over the two maps in parallel and detect the differences
as we go.

18 months ago[TableGen] Use InfoByHwMode::getSimple() in place of *InfoByHwMode::begin(). NFC
Craig Topper [Sat, 8 Apr 2023 18:55:55 +0000 (11:55 -0700)]
[TableGen] Use InfoByHwMode::getSimple() in place of *InfoByHwMode::begin(). NFC

18 months ago[clang-tidy] fix hint use correct range to replace last NamespaceDecl
Congcong Cai [Sat, 8 Apr 2023 19:03:58 +0000 (21:03 +0200)]
[clang-tidy] fix hint use correct range to replace last NamespaceDecl

range of replacing last namespace decl should be from last non nested namespace to last namespace

Reviewed By: PiotrZSL

Differential Revision: https://reviews.llvm.org/D147843

18 months ago[libc++][NFC] Only run the preprocessor in transitive_includes.sh.cpp
Nikolas Klauser [Fri, 31 Mar 2023 12:47:33 +0000 (14:47 +0200)]
[libc++][NFC] Only run the preprocessor in transitive_includes.sh.cpp

This cuts the time it takes to run the test in half.

Reviewed By: Mordante, #libc

Spies: libcxx-commits

Differential Revision: https://reviews.llvm.org/D147320

18 months ago[M68k] Basic FP register files and FMOVE implementations
Min-Yih Hsu [Fri, 7 Apr 2023 17:57:32 +0000 (10:57 -0700)]
[M68k] Basic FP register files and FMOVE implementations

  - FP registers and register classes definitions
  - MC support for FP registers
  - MC support for the FPDR -> FPDR variant of FMOVE/FSMOVE/FDSMOVE

Note that we don't implement 32/64-bit FP registers with subregisters since
f32/f64 values will be converted to f80 upon storing into a register,
rather than occupying part of the register.
So we only use register classes with different value types to distinguish
them.

Differential Revision: https://reviews.llvm.org/D147480

18 months ago[clang-tidy] Fix extern fixes in readability-redundant-declaration
Piotr Zegar [Sat, 8 Apr 2023 17:17:12 +0000 (17:17 +0000)]
[clang-tidy] Fix extern fixes in readability-redundant-declaration

Currently check does not look into LinkageSpecDecl
when removing redundant variable re-declarations.
This were leaving code in non-compiling state.
Fix patch fixes this and adds removal also of 'extern C'.

Fixes: https://github.com/llvm/llvm-project/issues/42068

Reviewed By: carlosgalvezp

Differential Revision: https://reviews.llvm.org/D146904

18 months ago[libc++][format] Implements formatter thread::id.
Mark de Wever [Sat, 18 Feb 2023 17:30:56 +0000 (18:30 +0100)]
[libc++][format] Implements formatter thread::id.

Since stacktrace header is WIP and it's not sure that will be done
before LLVM17 update the documentation. When the header is implemented
implementing the formatter is trivial, so that can be done quickly
afterwards.

Implements parts of:
 - P2693R1 Formatting thread::id and stacktrace

Reviewed By: #libc, ldionne

Differential Revision: https://reviews.llvm.org/D144331

18 months ago[clang][Interp] Add missing static_assert message
Timm Bäder [Sat, 8 Apr 2023 13:44:11 +0000 (15:44 +0200)]
[clang][Interp] Add missing static_assert message

18 months ago[clang][Interp][NFC] Add test for e7f55bb
Timm Bäder [Sat, 8 Apr 2023 13:26:46 +0000 (15:26 +0200)]
[clang][Interp][NFC] Add test for e7f55bb

18 months ago[clang][Interp] Fix a crash when calling invalid constexpr functions
Timm Bäder [Sat, 8 Apr 2023 09:59:59 +0000 (11:59 +0200)]
[clang][Interp] Fix a crash when calling invalid constexpr functions

Differential Revision: https://reviews.llvm.org/D147845

18 months ago[libc++][chrono] Add sys_time formatter.
Mark de Wever [Sun, 20 Mar 2022 12:40:02 +0000 (13:40 +0100)]
[libc++][chrono] Add sys_time formatter.

Partially implements:
- P1361 Integration of chrono with text formatting
- P2372 Fixing locale handling in chrono formatters

Reviewed By: ldionne, #libc

Differential Revision: https://reviews.llvm.org/D144739

18 months ago[libc++][format] Implements LWG3892.
Mark de Wever [Fri, 3 Mar 2023 19:23:51 +0000 (20:23 +0100)]
[libc++][format] Implements LWG3892.

This LWG issue is based on the discussion regarding

  P2733R3 Fix handling of empty specifiers in std::format

This paper was disussed and changed a few times in LEWG during the
Issaquah meeting. The paper was not voted in, instead LEWG asked for
a DR against C++26.

This LWG issue contains the direction voted by LEWG. This issue has not
been voted in yet. However it fixes some of the defencies on the
container based formatting. Without this fix the range-default-formatter
for strings looks bad when used in containers.

The changes of this issue match the intended changes of P27333.

type                      fmt     before     after (if changed)
---------------------------------------------------------------
char                      {}      a
char                      {:?}    'a'
array<char, 1>            {}      ['a']
array<char, 1>            {::}    [a]
array<char, 1>            {::c}   [a]
array<char, 1>            {::?}   ['a']
map<char, char>           {}      {a: a}     -> {'a': 'a'}
map<char, char>           {::}    {'a': 'a'}
set<char>                 {}      {'a'}
set<char>                 {::}    {a}
set<char>                 {::c}   {a}
set<char>                 {::?}   {'a'}
tuple<char>               {}      ('a')
stack<char>               {}      ['a']
stack<char>               {::}    [a]
stack<char>               {::c}   [a]
stack<char>               {::?}   ['a']
array<array<char, 1>, 1>  {}      [[a]]      -> {'a': 'a'}
array<array<char, 1>, 1>  {::}    [['a']]
array<array<char, 1>, 1>  {:::}   [[a]]
array<array<char, 1>, 1>  {:::c}  [[a]]
array<array<char, 1>, 1>  {:::?}  [['a']]
array<tuple<char>, 1>     {}      [(a)]      -> [('a')]
tuple<tuple<char>>        {}      ((a))      -> (('a'))
tuple<array<char, 1>>     {}      ([a])      -> (['a'])

Note the optimization text as mentioned in the tuple formatter can't be
done. The call to parse may affect the formatter so its state needs to
be preserved.

Reviewed By: ldionne, #libc, EricWF

Differential Revision: https://reviews.llvm.org/D145847

18 months ago[VPlan] Require VFRange.End to be a power-of-2. (NFCI)
Florian Hahn [Sat, 8 Apr 2023 12:04:08 +0000 (13:04 +0100)]
[VPlan] Require VFRange.End to be a power-of-2. (NFCI)

This removes the need to convert the end of the range to the next
power-of-2 for the end iterator after 4bd3fda5124962 and was suggested
as follow-up TODO in D147468.

18 months agoAMDGPU: Add baseline test for issue 61083
Matt Arsenault [Sun, 5 Mar 2023 19:53:06 +0000 (15:53 -0400)]
AMDGPU: Add baseline test for issue 61083

18 months agollvm-reduce: Actually change to the temporary directory in test
Matt Arsenault [Sat, 8 Apr 2023 11:07:29 +0000 (07:07 -0400)]
llvm-reduce: Actually change to the temporary directory in test

Hopefully fixes issue #61132

18 months agoAMDGPU: Fix LiveVariables verifier error for values defined before SI_END_CF
Matt Arsenault [Sun, 5 Mar 2023 11:07:23 +0000 (07:07 -0400)]
AMDGPU: Fix LiveVariables verifier error for values defined before SI_END_CF

GlobalISel happens to insert some constant materializes before SI_END_CF
in one test. These need to be excluded from AliveBlocks since they
are defined in the original block and used in the split block,
so they aren't fully alive through either block.

The case where the value defined in the first block which was originally used
in a later block is still broken.

Avoids a verifier error in a future patch.

18 months agoAMDGPU: Fix missing MIR serialization for PSInputAddr/PSInputEnable
Matt Arsenault [Sun, 5 Mar 2023 01:14:08 +0000 (21:14 -0400)]
AMDGPU: Fix missing MIR serialization for PSInputAddr/PSInputEnable

Resuming any mir test for a pixel shader would assert in the AsmPrinter.

18 months ago[VPlan] Add VFRange::begin() and end() iterators. (NFCI)
Florian Hahn [Sat, 8 Apr 2023 09:19:56 +0000 (10:19 +0100)]
[VPlan] Add VFRange::begin() and end() iterators. (NFCI)

Add an iterator to iterate over all VFs in VFRange. This simplifies some
existing code and allows using all_of,any_of and none_of on a VFRange.

Reviewed By: Ayal

Differential Revision: https://reviews.llvm.org/D147468

18 months ago[Docs][llvm-remarkutil] Fix dangling reference in documentation
Aiden Grossman [Sat, 8 Apr 2023 07:45:55 +0000 (07:45 +0000)]
[Docs][llvm-remarkutil] Fix dangling reference in documentation

D147710 introduced a new annotation-count subcommand to llvm-remarkutil
and added in documentation. However, the reference added under the
subcommands list never actually pointed to anything. This patch adds a
marker for the reference to point to so that the link works and the
sphinx build finishes without any errors.

18 months ago[AMDGPU] Fix a case of updating LiveIntervals in SIOptimizeExecMaskingPreRA
Jay Foad [Thu, 6 Apr 2023 16:26:05 +0000 (17:26 +0100)]
[AMDGPU] Fix a case of updating LiveIntervals in SIOptimizeExecMaskingPreRA

This was causing two test failures when I applied D129208 to enable
extra verification of LiveIntervals:

  LLVM :: CodeGen/AMDGPU/optimize-negated-cond-exec-masking-wave32.mir
  LLVM :: CodeGen/AMDGPU/optimize-negated-cond-exec-masking.mir

Differential Revision: https://reviews.llvm.org/D147721

18 months ago[TableGen] Simplify TypeInfer::expandOverloads. NFC
Craig Topper [Sat, 8 Apr 2023 06:38:04 +0000 (23:38 -0700)]
[TableGen] Simplify TypeInfer::expandOverloads. NFC

Directly test the 5 overloaded types instead of doing extra
set creation and iteration.

18 months ago[clang][Sema][NFC] Save token name instead of the full token
Timm Bäder [Wed, 5 Apr 2023 11:46:11 +0000 (13:46 +0200)]
[clang][Sema][NFC] Save token name instead of the full token

Differential Revision: https://reviews.llvm.org/D147615

18 months ago[clang][Interp][NFCI] Call* ops don't modify the PC
Timm Bäder [Fri, 3 Feb 2023 11:20:15 +0000 (12:20 +0100)]
[clang][Interp][NFCI] Call* ops don't modify the PC

This caused the reported errors from the Call*() handlers to report the
wrong source location.

Fixes: https://github.com/llvm/llvm-project/issues/62002

18 months ago[Outliner] Add an option to only enable outlining of patterns above a certain threshold
Nathan Lanza [Sat, 8 Apr 2023 06:10:27 +0000 (02:10 -0400)]
[Outliner] Add an option to only enable outlining of patterns above a certain threshold

Outlining isn't always a win when the saved instruction count is >= 1.
The overhead of representing a new function in the binary depends on
exception metadata and alignment. So parameterize this for local tuning.

Reviewed By: paquette

Differential Revision: https://reviews.llvm.org/D136774

18 months ago[clang-tidy] fix concat-nest-namespace fix hint remove the macro
Congcong Cai [Sat, 8 Apr 2023 04:24:25 +0000 (06:24 +0200)]
[clang-tidy] fix concat-nest-namespace fix hint remove the macro

Partially fixed [#60035](https://github.com/llvm/llvm-project/issues/60035)
This patch refactor the FixHint for concat-nest-namespace.
1. remove each namespace except the last non-nest namespace.
2. replace the last non-nest namespace with the new name.

It can remain the comment / pragma / macro between namespace and update the close comment.

Reviewed By: PiotrZSL

Differential Revision: https://reviews.llvm.org/D147194

18 months ago[ORC] Drop tests for OrcV2CBindingsReflectProcessSymbols after 231107a8b5b.
Lang Hames [Sat, 8 Apr 2023 03:00:49 +0000 (03:00 +0000)]
[ORC] Drop tests for OrcV2CBindingsReflectProcessSymbols after 231107a8b5b.

This test was removed as LLJIT now reflects process symbols by default.

18 months ago[MergeICmps] Attach metadata to new created loads
Zhongyunde [Sat, 8 Apr 2023 02:45:39 +0000 (10:45 +0800)]
[MergeICmps] Attach metadata to new created loads

Use clone to keep the metadata, the issue is reported
by aeubanks on D141188.

Reviewed By: nikic, paulwalker-arm

Differential Revision: https://reviews.llvm.org/D146702

18 months agoRe-apply "[ORC] LLJIT updates: ExecutorNativePlatform, default ..." with fixes.
Lang Hames [Fri, 7 Apr 2023 17:06:20 +0000 (17:06 +0000)]
Re-apply "[ORC] LLJIT updates: ExecutorNativePlatform, default ..." with fixes.

This reapplies 371cb1af61d, which was reverted in 0b2240eda01 due to bot
failures.

The clang-repl test failure is fixed by dropping the process symbols definition
generator that was manually attached to the main JITDylib, since LLJIT now
exposes process symbols by default. (The bug was triggered when JIT'd code used
the process atexit provided by the generator, rather than the JIT atexit which
has been moved into the platform JITDylib).

Any LLJIT clients that see crashes in static destructors should likewise remove
any process symbol generators attached to their main JITDylib.

18 months ago[bazel build] pull in cubin only when needed
Aart Bik [Fri, 7 Apr 2023 23:24:27 +0000 (16:24 -0700)]
[bazel build] pull in cubin only when needed

Reviewed By: Peiming

Differential Revision: https://reviews.llvm.org/D147826

18 months agoFix build failure introduced in https://reviews.llvm.org/D145208
Augusto Noronha [Fri, 7 Apr 2023 23:36:11 +0000 (16:36 -0700)]
Fix build failure introduced in https://reviews.llvm.org/D145208

18 months agoAMDGPU: Make myself code owner
Matt Arsenault [Tue, 21 Feb 2023 18:50:11 +0000 (14:50 -0400)]
AMDGPU: Make myself code owner

18 months agollvm-reduce: Try running test in temporary directory
Matt Arsenault [Mon, 3 Apr 2023 13:29:44 +0000 (09:29 -0400)]
llvm-reduce: Try running test in temporary directory

Attempt to fix issue #61761

18 months ago[llvm-remarkutil] Add missing new line for `llvm/docs/CommandGuide/llvm-remarkutil...
Zain Jaffal [Fri, 7 Apr 2023 22:45:45 +0000 (23:45 +0100)]
[llvm-remarkutil] Add missing new line for `llvm/docs/CommandGuide/llvm-remarkutil.rst`

18 months agoRecommit "Add an option to print out annotation remark count."
Zain Jaffal [Fri, 7 Apr 2023 22:42:39 +0000 (23:42 +0100)]
Recommit  "Add an option to print out annotation remark count."

Add missing new line for `llvm/docs/CommandGuide/llvm-remarkutil.rst`

This reverts commit 0f7fcb4c670fbef2c25b835fdfdd29598c6c13ae.

18 months agoRevert "Add an option to print out annotation remark count."
Zain Jaffal [Fri, 7 Apr 2023 22:41:53 +0000 (23:41 +0100)]
Revert "Add an option to print out annotation remark count."

This reverts commit 7cc80ef5fa359b68ee85033f98b1bef1f37fb21c.

18 months ago[NFC] Add a new test case in add-and-not.ll
Amaury Séchet [Fri, 7 Apr 2023 22:34:53 +0000 (22:34 +0000)]
[NFC] Add a new test case in add-and-not.ll

18 months ago[clang:diagnostics] Turning off warn_self_assignment_overloaded for user-defined...
Xiang Li [Fri, 7 Apr 2023 15:34:56 +0000 (11:34 -0400)]
[clang:diagnostics] Turning off warn_self_assignment_overloaded for user-defined compound assignments

Fixes 42469 https://github.com/llvm/llvm-project/issues/42469

Only check self assignment on BO_Assign when BuildOverloadedBinOp.

18 months agoAdd an option to print out annotation remark count.
Zain Jaffal [Thu, 6 Apr 2023 14:09:24 +0000 (15:09 +0100)]
Add an option to print out annotation remark count.

This adds a `annotation-count` option to llvm-remarkutil.

```
llvm-remarkutil annotation-count -remark=REMARK
```
This will print out the remark count for a pass that uses annotation remarks.

Differential Revision: https://reviews.llvm.org/D147710

18 months ago[-Wunsafe-buffer-usage] FixableGadget for handling stand alone pointers under UPC
MalavikaSamak [Fri, 7 Apr 2023 22:11:05 +0000 (15:11 -0700)]
[-Wunsafe-buffer-usage] FixableGadget for handling stand alone pointers under UPC

This patch introduces UPCStandalonePointerGadget, a FixableGadget that emits fixits to
handle cases where a pointer identified as unsafe is simply referenced. An example of
such a case is when the pointer is input as an argument to a method call, where we can
not change the type of the argument. For cases where the strategy for the unsafe pointer is
to use std::span, the idea is to extract the underlying pointer by invoking the "data()"
method on the span instance.

For example, the gadget emits a fixit for S3, where S1, S2 are handled by other gadgets:
  S1: int *ptr = new int[10];
  S2: int val1 = ptr[k]; // Unsafe operation on ptr
  S3: foo(ptr); // Some method that accepts raw pointer => FIXIT: foo(ptr.data());

Reviewed by: NoQ, ziqingluo-90, jkorous

Differential revision: https://reviews.llvm.org/D143676

18 months ago[mlir][math]Expand Fused math.fmaf to a multiply-add
Balaji V. Iyer [Fri, 7 Apr 2023 21:47:20 +0000 (21:47 +0000)]
[mlir][math]Expand Fused math.fmaf to a multiply-add

Fused multiply and add are being pushed directly to the libm. This is problematic
for situations where libm is not available. This patch will break down a fused multiply and
add into a multiply followed by an add.

Reviewed By: rsuderman

Differential Revision: https://reviews.llvm.org/D147811

18 months agoRemove redundant `get()` call and use auto on LHS of make_unique
David Blaikie [Fri, 7 Apr 2023 22:12:31 +0000 (22:12 +0000)]
Remove redundant `get()` call and use auto on LHS of make_unique

18 months ago[Format/ObjC] Support NS_ERROR_ENUM in ObjC language guesser
Ben Hamilton [Tue, 4 Apr 2023 21:27:42 +0000 (15:27 -0600)]
[Format/ObjC] Support NS_ERROR_ENUM in ObjC language guesser

Apple added a new NS_ERROR_ENUM macro to help define enums for
NSError codes.

This updates libformat's Objective-C language-guessing heuristic
to detect the new macro as well as related NSError types.

Tested: New tests added.

Reviewed By: MyDeveloperDay

Differential Revision: https://reviews.llvm.org/D147577

18 months ago[Sanitizers][Atos] Remove null-ing of atos process pointer
Blue Gaston [Thu, 6 Apr 2023 17:33:40 +0000 (10:33 -0700)]
[Sanitizers][Atos] Remove null-ing of atos process pointer

Currently, when we send an address to atos to be symbolized, it is
expected that atos returns with more than it was sent, i.e. symbol
information for that address. In the case where only the address is
returned, we currently null the pointer to the atos process. Typically,
for modules where no symbolication is expected, we do not send the
address to atos.

However, in new simulators there is an early call that atos does not
return any symbol information for. And in this case, because we have
gotten rid of the pointer to the process, no subsequent frames are
symbolicated, even tho atos is still working/running.

This patch removes the nulling of the pointer to the process. This
allows subsequent calls to atos even after an unexpected result.
It also now Reports what has happened and the address this occurred.

This will improve symbolication in cases where we get an unepxected
result, and will make it easier to diagnose atos if it is not
symbolicating as expected.

Filed a radar about the change of behavior 107621524

rdar://107169715

Differential Revision: https://reviews.llvm.org/D147725

18 months ago[flang][NFC] Add select type test to check correct ordering
Valentin Clement [Fri, 7 Apr 2023 21:18:48 +0000 (14:18 -0700)]
[flang][NFC] Add select type test to check correct ordering

Somehow this test has been left behind in my sandbox. This
patch adds a lowering test for fir.select_type operation and
makes sure the dynamic type comaparison is done in the right
order when we have multiple CLASS IS type guard statement
for types that are linked.

This should have been posted with D138280.

Reviewed By: PeteSteinfeld

Differential Revision: https://reviews.llvm.org/D147807

18 months ago[COFF] Add MC support for emitting IMAGE_WEAK_EXTERN_ANTI_DEPENDENCY symbols
Eli Friedman [Fri, 3 Mar 2023 02:09:14 +0000 (18:09 -0800)]
[COFF] Add MC support for emitting IMAGE_WEAK_EXTERN_ANTI_DEPENDENCY symbols

This is mostly useful for ARM64EC, which uses such symbols extensively.

One interesting quirk of ARM64EC is that we need to be able to emit weak
symbols that point at each other (so if either symbol is defined
elsewhere, both symbols point at the definition).  This required a few
changes to the way we handle weak symbols on Windows.

Differential Revision: https://reviews.llvm.org/D145208

18 months ago[JSON] Fix uninitialized variable warning
Jonas Devlieghere [Fri, 7 Apr 2023 20:54:45 +0000 (13:54 -0700)]
[JSON] Fix uninitialized variable warning

Fix uninitialized variable warning when deserializing a std::optional<E>
where is an enum type.

JSON.h:771:20: warning: variable 'Result' is uninitialized when used here [-Wuninitialized]
  if (!fromJSON(E, Result, P))
                   ^~~~~~

18 months ago[lldb] Add unittests for a few FileSpec methods
Alex Langford [Fri, 7 Apr 2023 18:41:54 +0000 (11:41 -0700)]
[lldb] Add unittests for a few FileSpec methods

This adds tests for:
- FileSpec::TestFileNameExtensions
- FileSpec::TestFileNameStrippingExtension
- FileSpec::IsSourceImplementationFile

This additionally updates incorrect documentation.

Differential Revision: https://reviews.llvm.org/D147801

18 months ago[NFC][clang] Fix static analyzer tool remarks about large copies by values
Manna, Soumi [Fri, 7 Apr 2023 20:37:56 +0000 (16:37 -0400)]
[NFC][clang] Fix static analyzer tool remarks about large copies by values

        Reported by Coverity:

        Big parameter passed by value
        Copying large values is inefficient, consider passing by reference; Low, medium, and high size thresholds for detection can be adjusted.

       1. Inside "SemaConcept.cpp" file, in subsumes<clang::Sema::MaybeEmitAmbiguousAtomicConstraintsDiagnostic(clang::NamedDecl *, llvm::ArrayRef<clang::Expr const *>, clang::NamedDecl *, llvm::ArrayRef<clang::Expr const *>)::[lambda(clang::AtomicConstraint const &, clang::AtomicConstraint const &) (instance 2)]>(llvm::SmallVector<llvm::SmallVector<clang::AtomicConstraint *, 2u>, 4u>, llvm::SmallVector<llvm::SmallVector<clang::AtomicConstraint *, 2u>, 4u>, T1): A large function call parameter exceeding the low threshold is passed by value.

        i. pass_by_value: Passing parameter PDNF of type NormalForm (size 144 bytes) by value, which exceeds the low threshold of 128 bytes.

        ii. pass_by_value: Passing parameter QCNF of type NormalForm (size 144 bytes) by value, which exceeds the low threshold of 128 bytes.

        2. Inside "CodeGenAction.cpp" file, in clang::reportOptRecordError(llvm::Error, clang::DiagnosticsEngine &, clang::CodeGenOptions): A very large function call parameter exceeding the high threshold is passed by value.

        i. pass_by_value: Passing parameter CodeGenOpts of type clang::CodeGenOptions const (size 1560 bytes) by value, which exceeds the high threshold of 512 bytes.

        3. Inside "SemaCodeComplete.cpp" file, in HandleCodeCompleteResults(clang::Sema *, clang::CodeCompleteConsumer *, clang::CodeCompletionContext, clang::CodeCompletionResult *, unsigned int): A large function call parameter exceeding the low threshold is passed by value.

        i. pass_by_value: Passing parameter Context of type clang::CodeCompletionContext (size 200 bytes) by value, which exceeds the low threshold of 128 bytes.

        4. Inside "SemaConcept.cpp" file, in <unnamed>::SatisfactionStackRAII::SatisfactionStackRAII(clang::Sema &, clang::NamedDecl const *, llvm::FoldingSetNodeID): A large function call parameter exceeding the low threshold is passed by value.

        i. pass_by_value: Passing parameter FSNID of type llvm::FoldingSetNodeID (size 144 bytes) by value, which exceeds the low threshold of 128 bytes.

        Reviewed By: erichkeane, aaron.ballman

        Differential Revision: https://reviews.llvm.org/D147708

18 months ago[SLP][NFC]Make clusterSortPtrAccesses static.
Alexey Bataev [Fri, 7 Apr 2023 20:22:48 +0000 (13:22 -0700)]
[SLP][NFC]Make clusterSortPtrAccesses static.

18 months ago[InstCombine] Improve transforms for `(mul X, Y)` -> `(shl X, log2(Y)`
Noah Goldstein [Fri, 7 Apr 2023 19:35:37 +0000 (14:35 -0500)]
[InstCombine] Improve transforms for `(mul X, Y)` -> `(shl X, log2(Y)`

Using the more robust log2 search allows us to fold more cases (same
logic as exists for idiv/irem).

Reviewed By: nikic

Differential Revision: https://reviews.llvm.org/D146347

18 months ago[InstCombine] Add tests for indirect mul by power of 2; NFC
Noah Goldstein [Fri, 17 Mar 2023 06:32:32 +0000 (01:32 -0500)]
[InstCombine] Add tests for indirect mul by power of 2; NFC

Differential Revision: https://reviews.llvm.org/D146346

18 months ago[Clang][Sema] Add MultiLevelTemplateArgumentList::dump
Alexander Shaposhnikov [Fri, 7 Apr 2023 19:34:52 +0000 (19:34 +0000)]
[Clang][Sema] Add MultiLevelTemplateArgumentList::dump

Add MultiLevelTemplateArgumentList::dump
(similarly to TemplateArgument::dump).

Differential revision: https://reviews.llvm.org/D147744

18 months ago[mlir][sparse] Simplifying Merger::expContainsTensor
wren romano [Wed, 29 Mar 2023 23:36:14 +0000 (16:36 -0700)]
[mlir][sparse] Simplifying Merger::expContainsTensor

`expContainsTensor` used to call `expIsTensor` to short-circuit the recursive calls; however, the very first thing `expContainsTensor` does is to check `expIsTensor`, so the short-circuiting code just causes the function to check that condition redundantly.

Depends On D146684

Reviewed By: aartbik

Differential Revision: https://reviews.llvm.org/D146688

18 months ago[libc] Add baremetal abort
Alex Brachet [Fri, 7 Apr 2023 19:25:27 +0000 (19:25 +0000)]
[libc] Add baremetal abort

Differential Revision: https://reviews.llvm.org/D147795

18 months ago[lldb] Move "SelectMostRelevantFrame" from Thread::WillStop
Jim Ingham [Fri, 7 Apr 2023 19:17:57 +0000 (12:17 -0700)]
[lldb] Move "SelectMostRelevantFrame" from Thread::WillStop

SelectMostRelevantFrame triggers the StackFrameRecognizer construction,
which can run arbitrary Python code, call expressions etc. WillStop gets
called on every private stop while the recognizers are a user-facing
feature, so first off doing this work on every stop is inefficient. But
more importantly, you can get in to situations where the recognizer
causes an expression to get run, then when we fetch the stop event at
the end of the expression evaluation, we call WillStop again on the
expression handling thread, which will do the same StackFrameRecognizer
work again. If anyone is locking along that path, you will end up with a
deadlock between the two threads.

The example that brought this to my attention was the
objc_exception_throw recognizer which can cause the objc runtime
introspection functions to get run, and those take a lock in
AppleObjCRuntimeV2::DynamicClassInfoExtractor::UpdateISAToDescriptorMap
along this path, so the second thread servicing the expression deadlocks
against the first thread waiting for the expression to complete.

It makes more sense to have the frame recognizers run on demand, either
when someone asks for the variables for the frame, or when someone does
GetSelectedFrame. The former already worked that way, the only reason
this was being done in WillStop was because the StackFrameRecognizers
can change the SelectedFrame, so you needed to run them before the
anyone requested the SelectedFrame.

This patch moves SelectMostRelevantFrame to StackFrameList, and runs it
when GetSelectedFrame is called for the first time on a given stop. If
you call SetSelectedFrame before GetSelectedFrame, then you should NOT
run the recognizer & change the frame out from under you. This patch
also makes that work. There were already tests for this behavior, and
for the feature that caused the hang, but the hang is racy, and it
doesn't trigger all the time, so I don't have a way to test that
explicitly.

One more detail: it's actually pretty easy to end up calling
GetSelectedFrame, for instance if you ask for the best ExecutionContext
from an ExecutionContextRef it will fill the StackFrame with the result
of GetSelectedFrame and that would still have the same problems if this
happens on the Private State Thread. So this patch also short-circuits
SelectMostRelevantFrame if run on the that thread. I can't think of any
reason the computations that go on on the Private State Thread would
actually want the SelectedFrame - that's a user-facing concept, so
avoiding that complication is the best way to go.

rdar://107643231

Differential revision: https://reviews.llvm.org/D147753

18 months ago[SLP][NFC]Improve SLP time by precomputing value<->gather nodes
Alexey Bataev [Fri, 7 Apr 2023 18:22:27 +0000 (11:22 -0700)]
[SLP][NFC]Improve SLP time by precomputing value<->gather nodes
dependencies.

Improved compiled time by the precomputing the mapping between gathered
scalars and their gather/buildvector nodes for later use in
isGatherShuffledEntry to avoid recomputing this map each time this
function is called.

18 months ago[libc++] Reject standard attributes which are extensions in libcpp-uglify-attributes
Nikolas Klauser [Tue, 7 Mar 2023 16:00:01 +0000 (17:00 +0100)]
[libc++] Reject standard attributes which are extensions in libcpp-uglify-attributes

This adds a list of attributes which can be pretty to be able to reject attributes which were introduced in a later C++ standard.
Fixes #61196

Reviewed By: Mordante, #libc

Spies: mikhail.ramalho, jdoerfert, libcxx-commits

Differential Revision: https://reviews.llvm.org/D145508

18 months ago[libc][obvious] fix phrasing in assert comment
Michael Jones [Fri, 7 Apr 2023 18:50:43 +0000 (11:50 -0700)]
[libc][obvious] fix phrasing in assert comment

18 months ago[flang] Handle correctly user defined assignment for allocatable component
Valentin Clement [Fri, 7 Apr 2023 18:47:35 +0000 (11:47 -0700)]
[flang] Handle correctly user defined assignment for allocatable component

In the Fortran standard 2018 section 10.2.1.3 (13), it is mentioned
that all noncoarray allocatable component must follow this sequence of
operations:

1) If the component of the variable is allocated, it is deallocated.
2) If the component of the value of expr is allocated, the corresponding
   component of the variable is allocated with the same dynamic type and
   type parameters as the component of the value of expr. If it is an
   array, it is allocated with the same bounds. The value of the
   component of the value of expr is then assigned to the corresponding
   component of the variable using defined assignment if the declared type
   of the component has a type-bound defined assignment consistent with the
   component, and intrinsic assignment for the dynamic type of that component
   otherwise.

This patch updates the code to make use of the user defined assignment for
allocatable component and make sure the component is allocated correctly.

Reviewed By: klausler

Differential Revision: https://reviews.llvm.org/D147797

18 months agoRevert "[Clang][Sema] Fix comparison of constraint expressions"
Alexander Shaposhnikov [Fri, 7 Apr 2023 18:32:49 +0000 (18:32 +0000)]
Revert "[Clang][Sema] Fix comparison of constraint expressions"

This temporarily reverts commit
60bee9ff544541e83ffbd4be31923d0e8b644690.
The diff will be recommitted once the newly discovered
regressions are fixed.

18 months ago[Clang] Improve designated inits diagnostic location
Bill Wendling [Fri, 7 Apr 2023 18:35:29 +0000 (11:35 -0700)]
[Clang] Improve designated inits diagnostic location

A "null" designator won't have a valid location. Try to approximate this
location as best we can in that situation.

Closes 61118
Closes 46132

Reviewed By: shafik

Differential Revision: https://reviews.llvm.org/D147673

18 months ago[mlir] Add tests accidentally removed from arith expand-ops
Robert Suderman [Fri, 7 Apr 2023 18:12:12 +0000 (18:12 +0000)]
[mlir] Add tests accidentally removed from arith expand-ops

Some tests were accidentally removed due to debug code being included.
Readding the tests to guarantee coverage.

Reviewed By: NatashaKnk

Differential Revision: https://reviews.llvm.org/D147796

18 months ago[gn build] Port 376168babb51
LLVM GN Syncbot [Fri, 7 Apr 2023 17:38:29 +0000 (17:38 +0000)]
[gn build] Port 376168babb51

18 months ago[clang-tidy] Add modernize-type-traits check
Nathan James [Fri, 7 Apr 2023 17:38:13 +0000 (18:38 +0100)]
[clang-tidy] Add modernize-type-traits check

This check will look for usages of standard library type traits of the form `traits<...>::type` and `traits<...>::value` and convert them into `traits_t<...>` and `traits_v<...>` respectively.
This expands on the work in D135404 by supporting dependent traits with no instantiations as well as types.

Differential Revision: https://reviews.llvm.org/D137302

18 months ago[support] Revise ScopedPrinter formatting tests for floats
Paul Kirth [Fri, 24 Mar 2023 22:10:11 +0000 (22:10 +0000)]
[support] Revise ScopedPrinter formatting tests for floats

Previously there were several attempts to make the format checks for NaN
and Inf work across platforms, like AIX and Solaris, that print these
values slightly differently. This resulted in a number of forward fixes,
until we finally disabled the tests for NaN and Inf. This change should
make the test robust across different platforms, and reduce the overall
amount of code by delegating to helper functions that use the same
format strings as the implementations used by PrintNumber().

This additionally reverts commit 5a9bad171be5dfdf9430a0f6cbff14d29ca54181
and fa56e362af475e0758cfb41c42f78db50da7235c.

Reviewed By: jhenderson

Differential Revision: https://reviews.llvm.org/D146851

18 months ago[libc++] Module fixes for __synth_three_way.
Mark de Wever [Tue, 21 Mar 2023 17:05:35 +0000 (18:05 +0100)]
[libc++] Module fixes for __synth_three_way.

These changes make it possible to use __synth_three_way in modules. The
change from a lambda to a function is a Clang issue.

The change is list was needed since the compiler couldn't deduce the
comparison template argument.

Adds a few missing includes too.

Reviewed By: #libc, ldionne

Differential Revision: https://reviews.llvm.org/D146545

18 months ago[Clang] Fix buildbots after Fix filtering of inline namespaces for friend functions...
Shafik Yaghmour [Fri, 7 Apr 2023 16:18:15 +0000 (09:18 -0700)]
[Clang] Fix buildbots after Fix filtering of inline namespaces for friend functions commit

In commit cffadbd951e9 the test I added was using a C++17 feature and this
breaking some build bots. I don't need the feature and so I will modify
the test to not use it.

18 months ago[Clang] Fix filtering of inline namespaces for friend functions
Shafik Yaghmour [Fri, 7 Apr 2023 15:57:24 +0000 (08:57 -0700)]
[Clang] Fix filtering of inline namespaces for friend functions

PR D135370 implemented a performance improvement but it restricted the filtering
of declaration from inline namespace too much. In particular it did not filter
for the function template case.

This led to a regression and this PR removes that check.

This fixes: https://github.com/llvm/llvm-project/issues/61851

Differential Revision: https://reviews.llvm.org/D147762

18 months ago[libc++][format][chrono] Addresses LWG3842.
Mark de Wever [Sat, 18 Feb 2023 13:54:29 +0000 (14:54 +0100)]
[libc++][format][chrono] Addresses LWG3842.

  LWG3842 Unclear wording for precision in chrono-format-spec

Note there is nothing to do, the issue clarifies the wording in the Standard.
The new wording matches my interpretation of the previous wording and this has
already been implemented in libc++.

Reviewed By: #libc, ldionne

Differential Revision: https://reviews.llvm.org/D144328

18 months ago[libc++][format] Use granularized charconv.
Mark de Wever [Thu, 16 Mar 2023 16:59:42 +0000 (17:59 +0100)]
[libc++][format] Use granularized charconv.

This reduces the number of transitive includes when using format.

Reviewed By: #libc, ldionne

Differential Revision: https://reviews.llvm.org/D146240

18 months ago[libc++] Qualifies intmax_t and uintmax_t.
Mark de Wever [Fri, 24 Mar 2023 16:18:11 +0000 (17:18 +0100)]
[libc++] Qualifies intmax_t and uintmax_t.

This has been done using the following command

  find libcxx/test -type f -exec perl -pi -e 's|^([^/]+?)((?<!::)(?<!::u)u?intmax_t)|\1std::\2|' \{} \;

The std module doesn't export declarations in the global namespaace.
This is a preparation for that module.

Reviewed By: #libc, ldionne

Differential Revision: https://reviews.llvm.org/D146821

18 months ago[Doc][NFC] Fix a typo
Phoebe Wang [Fri, 7 Apr 2023 15:34:26 +0000 (23:34 +0800)]
[Doc][NFC] Fix a typo