Kostya Kortchinsky [Sun, 15 Aug 2021 16:09:46 +0000 (09:09 -0700)]
[scudo] Use stdint types for internal scudo types
`scudo::uptr` was defined as an `unsigned long` on 32-b platform,
while a `uintptr_t` is usually defined as an `unsigned int`.
This worked, this was not consistent, particularly with regard to
format string specifiers.
As suggested by Vitaly, since we are including `stdint.h`, define
the internal `scudo` integer types to those.
Differential Revision: https://reviews.llvm.org/D108089
Rong Xu [Mon, 16 Aug 2021 17:15:30 +0000 (10:15 -0700)]
[NFC] Fix typos
s/senstive/senstive/g
Jordan Rupprecht [Mon, 16 Aug 2021 17:04:32 +0000 (10:04 -0700)]
[NFC][AArch64] Fix unused var in release build
Shilei Tian [Mon, 16 Aug 2021 16:49:59 +0000 (12:49 -0400)]
[NFC] Trim trailing whitespaces in `llvm/CMakeLists.txt`
Simon Pilgrim [Mon, 16 Aug 2021 16:25:34 +0000 (17:25 +0100)]
[X86] Add i128 funnel shift tests
Test coverage for D108058
Paul Robinson [Mon, 16 Aug 2021 16:27:48 +0000 (09:27 -0700)]
Fix whitespace typo in 94b4598
Nikita Popov [Sun, 15 Aug 2021 21:43:34 +0000 (23:43 +0200)]
[profile] Don't use pragma comment linker on mingw
At least when compiling with gcc, this is not supported and will
result in errors when linking against the profiler runtime. Only
use the pragma comment linker based code with MSVC, but not with
a mingw toolchain. This also undoes D107620, which shouldn't be
relevant anymore.
Differential Revision: https://reviews.llvm.org/D108095
Paul Robinson [Mon, 16 Aug 2021 15:42:21 +0000 (08:42 -0700)]
[PS4] stp[n]cpy not available on PS4
Raphael Isemann [Mon, 16 Aug 2021 15:39:10 +0000 (17:39 +0200)]
[lldb] Add tests for TypeSystemMap::GetTypeSystemForLanguage
Regression tests for D108088 .
Reviewed By: mib
Differential Revision: https://reviews.llvm.org/D108121
Craig Topper [Mon, 16 Aug 2021 15:42:00 +0000 (08:42 -0700)]
[TypePromotion] Don't mutate the result type of SwitchInst.
SwitchInst should have a void result type.
Add a check to the verifier to catch this error.
Reviewed By: samparker
Differential Revision: https://reviews.llvm.org/D108084
Louis Dionne [Mon, 16 Aug 2021 15:42:30 +0000 (11:42 -0400)]
[libc++][NFC] Replace uses of 'constexpr friend' by 'friend constexpr'
This is done for consistency, since that's what we do everywhere else
in the library.
Dimitry Andric [Sun, 15 Aug 2021 15:59:32 +0000 (17:59 +0200)]
[lldb] Avoid unhandled Error in TypeSystemMap::GetTypeSystemForLanguage
When assertions are turned off, the `llvm::Error` value created at the
start of this function is overwritten using the move-assignment
operator, but the success value is never checked. Whenever a TypeSystem
cannot be found or created, this can lead to lldb core dumping with:
Program aborted due to an unhandled Error:
Error value was Success. (Note: Success values must still be checked prior to being destroyed).
Fix this by not creating a `llvm::Error` value in advance, and directly
returning the result of `llvm::make_error` instead, whenever an error is
encountered.
See also: <https://bugs.freebsd.org/253881> and
<https://bugs.freebsd.org/257829>.
Reviewed By: teemperor
Differential Revision: https://reviews.llvm.org/D108088
Simon Pilgrim [Mon, 16 Aug 2021 14:53:31 +0000 (15:53 +0100)]
[DAG] Fold concat_vectors(concat_vectors(x,y),concat_vectors(a,b)) -> concat_vectors(x,y,a,b)
Follow-up to D107068, attempt to fold nested concat_vectors/undefs, as long as both the vector and inner subvector types are legal.
This exposed the same issue in ARM's MVE LowerCONCAT_VECTORS_i1 (raised as PR51365) and AArch64's performConcatVectorsCombine which both assumed concat_vectors only took 2 subvector operands.
Differential Revision: https://reviews.llvm.org/D107597
Peyton, Jonathan L [Mon, 28 Jun 2021 21:27:27 +0000 (16:27 -0500)]
[OpenMP] Add a few small fixes
* Add comment to help ensure new construct data are added in two places
* Check for division by zero in the loop worksharing code
* Check for syntax errors in parrange parsing
Differential Revision: https://reviews.llvm.org/D105929
Peyton, Jonathan L [Tue, 13 Jul 2021 19:42:27 +0000 (14:42 -0500)]
[OpenMP] Fix incorrect parameters to sscanf_s call
On Windows, the documentation states that when using sscanf_s,
each %c and %s specifier must also have additional size parameter.
This patch adds the size parameter in the one place where %c is
used.
Differential Revision: https://reviews.llvm.org/D105931
Jeremy Morse [Mon, 16 Aug 2021 14:43:38 +0000 (15:43 +0100)]
Revert
54a61c94f93 and its follow up in
547b712500e
These were part of D107823, however asan has found something excitingly
wrong happening:
https://lab.llvm.org/buildbot/#/builders/5/builds/10543/steps/13/logs/stdio
Sanjay Patel [Mon, 16 Aug 2021 14:28:46 +0000 (10:28 -0400)]
[InstCombine] allow for constant-folding in GEP transform
This would crash the reduced test or as described in
https://llvm.org/PR51485
...because we can't mark a constant (-expression) with 'inbounds'.
Renato Golin [Mon, 16 Aug 2021 10:53:30 +0000 (11:53 +0100)]
Fix type in DenseMap<SmallBitVector, *> to match V.size()
Differential Revision: https://reviews.llvm.org/D108124
Jeremy Morse [Mon, 16 Aug 2021 14:28:56 +0000 (15:28 +0100)]
Suppress signedness-comparison warning
This is a follow-up to
54a61c94f93.
Kristóf Umann [Mon, 5 Jul 2021 13:11:46 +0000 (15:11 +0200)]
[analyzer] MallocChecker: Add a visitor to leave a note on functions that could have, but did not change ownership on leaked memory
This is a rather common feedback we get from out leak checkers: bug reports are
really short, and are contain barely any usable information on what the analyzer
did to conclude that a leak actually happened.
This happens because of our bug report minimizing effort. We construct bug
reports by inspecting the ExplodedNodes that lead to the error from the bottom
up (from the error node all the way to the root of the exploded graph), and mark
entities that were the cause of a bug, or have interacted with it as
interesting. In order to make the bug report a bit less verbose, whenever we
find an entire function call (from CallEnter to CallExitEnd) that didn't talk
about any interesting entity, we prune it (click here for more info on bug
report generation). Even if the event to highlight is exactly this lack of
interaction with interesting entities.
D105553 generalized the visitor that creates notes for these cases. This patch
adds a new kind of NoStateChangeVisitor that leaves notes in functions that
took a piece of dynamically allocated memory that later leaked as parameter,
and didn't change its ownership status.
Differential Revision: https://reviews.llvm.org/D105553
Jeremy Morse [Mon, 16 Aug 2021 13:46:52 +0000 (14:46 +0100)]
[DebugInfo][InstrRef] Honour too-much-debug-info cutouts
VarLoc based LiveDebugValues will abandon variable location propagation if
there are too many blocks and variable assignments in the function. If it
didn't, and we had (say) 1000 blocks and 1000 variables in scope, we'd end
up with 1 million DBG_VALUEs just at the start of blocks.
Instruction-referencing LiveDebugValues should honour this limitation too
(because the same limitation applies to it). Hoist the relevant command
line options into LiveDebugValues.cpp and pass it down into the
implementation classes as an argument to ExtendRanges. I've duplicated all
the run-lines in live-debug-values-cutoffs.mir to have an
instruction-referencing flavour.
Differential Revision: https://reviews.llvm.org/D107823
Riccardo Mori [Mon, 16 Aug 2021 13:52:24 +0000 (15:52 +0200)]
[Polly][Isl] Move to the new-polly-generator branch version of isl-noexceptions.h. NFCI
This is part of an effort to reduce the differences between the custom C++ bindings used right now by polly in `lib/External/isl/include/isl/isl-noxceptions.h` and the official isl C++ interface.
With this commit we are moving from the `polly-generator` branch to the `new-polly-generator` branch that is more mantainable and is based on the official C++ interface `cpp-checked.h`.
Changes made:
- There are now many sublcasses for `isl::ast_node` representing different isl types. Use `isl::ast_node_for`, `isl::ast_node_user`, `isl::ast_node_block` and `isl::ast_node_mark` where needed.
- There are now many sublcasses for `isl::schedule_node` representing different isl types. Use `isl::schedule_node_mark`, `isl::schedule_node_extension`, `isl::schedule_node_band` and `isl::schedule_node_filter` where needed.
- Replace the `isl::*::dump` with `dumpIslObj` since the isl dump method is not exposed in the C++ interface.
- `isl::schedule_node::get_child` has been renamed to `isl::schedule_node::child`
- `isl::pw_multi_aff::get_pw_aff` has been renamed to `isl::pw_multi_aff::at`
- The constructor `isl::union_map(isl::union_pw_multi_aff)` has been replaced with the static method `isl::union_map::from()`
- Replace usages of `isl::val::add_ui` with `isl::val::add`
- `isl::union_set_list::alloc` is now a constructor
- All the `isl_size` values are now wrapped inside the class `isl::size` use `isl::size::release` to get the internal `isl_size` value where needed.
- `isl-noexceptions.h` has been generated by https://github.com/patacca/isl/commit/
73f5ed1f4d1f72582f731590ef9e43d9ab1956ad
No functional change intended.
Reviewed By: Meinersbur
Differential Revision: https://reviews.llvm.org/D107225
Kristóf Umann [Wed, 7 Jul 2021 09:58:26 +0000 (11:58 +0200)]
[analyzer][NFC] Split the main logic of NoStoreFuncVisitor to an abstract NoStateChangeVisitor class
Preceding discussion on cfe-dev: https://lists.llvm.org/pipermail/cfe-dev/2021-June/068450.html
NoStoreFuncVisitor is a rather unique visitor. As VisitNode is invoked on most
other visitors, they are looking for the point where something changed -- change
on a value, some checker-specific GDM trait, a new constraint.
NoStoreFuncVisitor, however, looks specifically for functions that *didn't*
write to a MemRegion of interesting. Quoting from its comments:
/// Put a diagnostic on return statement of all inlined functions
/// for which the region of interest \p RegionOfInterest was passed into,
/// but not written inside, and it has caused an undefined read or a null
/// pointer dereference outside.
It so happens that there are a number of other similar properties that are
worth checking. For instance, if some memory leaks, it might be interesting why
a function didn't take ownership of said memory:
void sink(int *P) {} // no notes
void f() {
sink(new int(5)); // note: Memory is allocated
// Well hold on, sink() was supposed to deal with
// that, this must be a false positive...
} // warning: Potential memory leak [cplusplus.NewDeleteLeaks]
In here, the entity of interest isn't a MemRegion, but a symbol. The property
that changed here isn't a change of value, but rather liveness and GDM traits
managed by MalloChecker.
This patch moves some of the logic of NoStoreFuncVisitor to a new abstract
class, NoStateChangeFuncVisitor. This is mostly calculating and caching the
stack frames in which the entity of interest wasn't changed.
Descendants of this interface have to define 3 things:
* What constitutes as a change to an entity (this is done by overriding
wasModifiedBeforeCallExit)
* What the diagnostic message should be (this is done by overriding
maybeEmitNoteFor.*)
* What constitutes as the entity of interest being passed into the function (this
is also done by overriding maybeEmitNoteFor.*)
Differential Revision: https://reviews.llvm.org/D105553
Simon Pilgrim [Mon, 16 Aug 2021 12:12:13 +0000 (13:12 +0100)]
[X86] Add PR46315 test case
tashuang.zk [Mon, 16 Aug 2021 11:41:55 +0000 (13:41 +0200)]
[MLIR][DISC] Revise ParallelLoopTilingPass with inbound_check mode
Expand ParallelLoopTilingPass with an inbound_check mode.
In default mode, the upper bound of the inner loop is from the min op; in
inbound_check mode, the upper bound of the inner loop is the step of the outer
loop and an additional inbound check will be emitted inside of the inner loop.
This was 'FIXME' in the original codes and a typical usage is for GPU backends,
thus the outer loop and inner loop can be mapped to blocks/threads in seperate.
Differential Revision: https://reviews.llvm.org/D105455
Simon Pilgrim [Mon, 16 Aug 2021 11:57:05 +0000 (12:57 +0100)]
[InstCombine] Add PR38021 nuw test case.
Simon Pilgrim [Mon, 16 Aug 2021 11:30:52 +0000 (12:30 +0100)]
[InstCombine] Regenerate AddOverFlow.ll test checks.
Roman Lebedev [Mon, 16 Aug 2021 11:30:00 +0000 (14:30 +0300)]
Revert "[NFCI][IndVars] rewriteLoopExitValues(): nowadays SCEV should not change `GEP` base pointer"
https://bugs.llvm.org/show_bug.cgi?id=51490 was filed.
This reverts commit
35a8bdc775817ce13a6c9b5cf81502052634aa1f.
Dmitry Vyukov [Mon, 16 Aug 2021 11:08:35 +0000 (13:08 +0200)]
tsan: fix unused var warnings in a test
Reviewed By: melver
Differential Revision: https://reviews.llvm.org/D108118
Florian Hahn [Mon, 16 Aug 2021 10:48:25 +0000 (11:48 +0100)]
[LoopPeel] Add test with multiple exit blocks branching to unreachable.
Add test as suggested by @ebedev.ri in D108108.
AndreyChurbanov [Mon, 16 Aug 2021 10:39:23 +0000 (13:39 +0300)]
[OpenMP] libomp: cleanup: minor fixes to silence static analyzer.
Added couple more checks to silence KlocWork static code analyzer.
Differential Revision: https://reviews.llvm.org/D107348
Kazushi (Jam) Marukawa [Fri, 13 Aug 2021 10:30:44 +0000 (19:30 +0900)]
[CLANG][PATCH][FPEnv] Add support for option -ffp-eval-method and extend #pragma float_control similarly
Need to update a clang regression test for VE after
https://reviews.llvm.org/D93769.
Reviewed By: simoll
Differential Revision: https://reviews.llvm.org/D108069
David Sherwood [Mon, 16 Aug 2021 08:21:39 +0000 (09:21 +0100)]
[NFC] Remove unused code in llvm::createSimpleTargetReduction
Florian Hahn [Fri, 13 Aug 2021 10:23:58 +0000 (11:23 +0100)]
[PhaseOrdering] Add test for missed vectorization with vector::at calls.
This test illustrates missed vectorization of loops with multiple
std::vector::at calls, like
int sum(std::vector<int> *A, std::vector<int> *B, int N) {
int cost = 0;
for (int i = 0; i < N; ++i)
cost += A->at(i) + B->at(i);
return cost;
}
https://clang.godbolt.org/z/KbYoaPhvq
Florian Hahn [Fri, 13 Aug 2021 09:39:03 +0000 (10:39 +0100)]
[LoopUnroll] Add peeling tests with unreachable exits.
Dmitry Vyukov [Thu, 5 Aug 2021 15:18:17 +0000 (17:18 +0200)]
tsan: add new trace
Add structures for the new trace format,
functions that serialize and add events to the trace
and trace replaying logic.
Differential Revision: https://reviews.llvm.org/D107911
Florian Hahn [Mon, 16 Aug 2021 07:58:37 +0000 (08:58 +0100)]
Revert "[sanitizer] Define 32bit uptr as uint"
This reverts commit
45138f788c9b3c4ac5d9ae4479841c411c15190e.
It looks like this breaks building sanitizers on Darwin platforms on
Green Dragon
https://green.lab.llvm.org/green/job/clang-stage1-RA/23332/console
FAILED: lib/sanitizer_common/CMakeFiles/RTSanitizerCommonSymbolizerNoHooks.ios.dir/sanitizer_stacktrace.cpp.o
/Users/buildslave/jenkins/workspace/clang-stage1-RA@2/clang-build/./bin/clang++ -DHAVE_RPC_XDR_H=0 -I/Users/buildslave/jenkins/workspace/clang-stage1-RA@2/llvm-project/compiler-rt/lib/sanitizer_common/.. -Wall -std=c++14 -Wno-unused-parameter -O2 -g -DNDEBUG -arch armv7 -arch armv7s -arch arm64 -arch armv7k -isysroot /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.15.sdk -stdlib=libc++ -miphoneos-version-min=9.0 -isysroot /Applications/Xcode.app/Contents/Developer/Platforms/iPhoneOS.platform/Developer/SDKs/iPhoneOS13.5.sdk -fPIC -fno-builtin -fno-exceptions -funwind-tables -fno-stack-protector -fno-sanitize=safe-stack -fvisibility=hidden -fno-lto -Wthread-safety -Wthread-safety-reference -Wthread-safety-beta -O3 -g -Wno-gnu -Wno-variadic-macros -Wno-c99-extensions -Wno-format-pedantic -nostdinc++ -Wno-format -fno-rtti -Wframe-larger-than=570 -Wglobal-constructors -DSANITIZER_SUPPORTS_WEAK_HOOKS=0 -MD -MT lib/sanitizer_common/CMakeFiles/RTSanitizerCommonSymbolizerNoHooks.ios.dir/sanitizer_stacktrace.cpp.o -MF lib/sanitizer_common/CMakeFiles/RTSanitizerCommonSymbolizerNoHooks.ios.dir/sanitizer_stacktrace.cpp.o.d -o lib/sanitizer_common/CMakeFiles/RTSanitizerCommonSymbolizerNoHooks.ios.dir/sanitizer_stacktrace.cpp.o -c '/Users/buildslave/jenkins/workspace/clang-stage1-RA@2/llvm-project/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace.cpp'
In file included from /Users/buildslave/jenkins/workspace/clang-stage1-RA@2/llvm-project/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace.cpp:13:
In file included from /Users/buildslave/jenkins/workspace/clang-stage1-RA@2/llvm-project/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace.h:15:
/Users/buildslave/jenkins/workspace/clang-stage1-RA@2/llvm-project/compiler-rt/lib/sanitizer_common/sanitizer_common.h:1068:14: error: 'operator new' takes type size_t ('unsigned long') as first parameter
inline void *operator new(__sanitizer::operator_new_size_type size,
^
1 error generated.
Roman Lebedev [Mon, 16 Aug 2021 07:53:15 +0000 (10:53 +0300)]
Revert "Reland [SimplifyCFG] performBranchToCommonDestFolding(): form block-closed SSA form before cloning instructions (PR51125)"
This is still wrong, as failing bots suggest.
This reverts commit
3d9beefc7d713ad8462d92427ccd17b9532ce904.
Dmitry Vyukov [Thu, 12 Aug 2021 13:43:09 +0000 (15:43 +0200)]
sanitizers: compile with -O1 under debug
Tsan's check_memcpy.c test was disabled under debug because it failed.
But it points to real issues and does not help to just disable it.
I tried to enable it and see what fail and the first hit was default ctor for:
struct ChainedOriginDepotDesc {
u32 here_id;
u32 prev_id;
};
initializing these fields to 0's help partially,
but compiler still emits memset before calling ctor.
I did not try to see what's the next failure, because if it fails
on such small structs, it won't be realistic to fix everything
and keep working.
Compile runtimes with -O1 under debug instead.
It seems to fix all current failures. At least I run check-tsan
under clang/gcc x debug/non-debug and all combinations passed.
-O1 does not usually use too aggressive optimizations
and sometimes even makes debugging easier because machine code
is not exceedingly verbose.
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D107962
Cullen Rhodes [Mon, 16 Aug 2021 07:31:55 +0000 (07:31 +0000)]
[AArch64][SME] Disable NEON in streaming mode
In streaming mode most of the NEON instruction set is illegal, disable
NEON when compiling with `+streaming-sve`, unless NEON is explictly
requested.
Subsequent patches will add support for the small subset of NEON
instructions that are legal in streaming mode.
Reviewed By: paulwalker-arm, david-arm
Differential Revision: https://reviews.llvm.org/D107902
Christian Sigg [Thu, 10 Jun 2021 06:56:15 +0000 (08:56 +0200)]
Reset all options in cl::ResetCommandLineParser()
Reset cl::Positional, cl::Sink and cl::ConsumeAfter options as well in cl::ResetCommandLineParser().
Reviewed By: rriddle, sammccall
Differential Revision: https://reviews.llvm.org/D103356
Dmitry Vyukov [Mon, 16 Aug 2021 07:43:26 +0000 (09:43 +0200)]
sanitizer_common: fix format string in LibIgnore
uptr should be printed with %zu.
Differential Revision: https://reviews.llvm.org/D108106
Tres Popp [Tue, 10 Aug 2021 11:53:59 +0000 (13:53 +0200)]
[mlir] Set top-down traversal for LinalgElementwiseOpFusion
The primary pattern for this pass clones many operations from producers
to consumers. Doing this top down prevents duplicated work when a
producer has multiple consumers, if it also is consuming another
linalg.generic.
As an example, a chain of ~2600 generics that are fused into ~70
generics was resulting in 16255 pattern invocations. This took 14
seconds on one machine but takes only 0.3 seconds with top-down
traversal.
Differential Revision: https://reviews.llvm.org/D107818
Sylvestre Ledru [Mon, 16 Aug 2021 07:07:33 +0000 (09:07 +0200)]
Add support of the future Debian (Debian 12 - Bookworm)
https://wiki.debian.org/DebianBookworm
ETA: 2023
Nathan Ridge [Wed, 28 Jul 2021 05:58:28 +0000 (01:58 -0400)]
[clangd] Do not show inlay hints pertaining to code in other files
Fixes https://github.com/clangd/clangd/issues/817
Differential Revision: https://reviews.llvm.org/D106934
Craig Topper [Mon, 16 Aug 2021 04:31:12 +0000 (12:31 +0800)]
[X86] Support avx512fp16 compare instructions in the IntelInstPrinter.
This enables printing of the mnemonics that contain the predicate
in the Intel printer. This requires accounting for the memory size
that is explicitly printed in Intel syntax. Those changes have been
synced to the ATT printer as well.
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D108093
Fangrui Song [Mon, 16 Aug 2021 04:19:04 +0000 (21:19 -0700)]
[test] Change llvm-xray options to use the preferred double-dash forms and change -f= to -f
Fangrui Song [Mon, 16 Aug 2021 04:09:35 +0000 (21:09 -0700)]
[docs] Change llvm-xray options to use the preferred double-dash forms
Sanjay Patel [Sun, 15 Aug 2021 18:14:33 +0000 (14:14 -0400)]
[Analysis][SimplifyLibCalls] improve function signature check for memcmp
This would assert/crash as shown in:
https://llvm.org/PR50850
The matching for bcmp/bcopy should probably also be updated,
but that's another patch.
David Blaikie [Sun, 15 Aug 2021 19:30:52 +0000 (12:30 -0700)]
Add missing triple for test
Craig Topper [Sun, 15 Aug 2021 19:00:54 +0000 (12:00 -0700)]
[X86] Prevent accidentally accepting cmpeqsh as a valid mnemonic.
We should only accept as vcmpeqsh.
Same for all the other 31 comparison values.
Craig Topper [Sun, 15 Aug 2021 18:42:33 +0000 (11:42 -0700)]
[X86] Modify the commuted load isel pattern for VCMPSHZrm to match VCMPSSZrm/VCMPSDZrm.
This allows commuting any immediate value. The previous code only
commuted equality immediates. This was inherited from an earlier
version of VCMPSSZrm/VCMPSDZrm.
David Blaikie [Sat, 14 Aug 2021 19:46:08 +0000 (12:46 -0700)]
DWARFVerifier: Check section-relative references at the end of the section
This ensures that debug_types references aren't looked for in
debug_info section.
Behavior is still going to be questionable in an unlinked object file -
since cross-cu references could refer to symbols in another .debug_info
(or, in theory, .debug_types) chunk - but if a producer only uses
ref_addr to refer to things within the same .debug_info chunk in an
object file (eg: whole program optimization/LTO - producing two CUs into
a single .debug_info section in an object file - the ref_addrs there
could be resolved relative to that .debug_info chunk, not needing to
consider comdat (DWARFv5 type units or other creatures) chunks of
.debug_info, etc)
Craig Topper [Sun, 15 Aug 2021 18:35:20 +0000 (11:35 -0700)]
[X86] Add vcmpsh/vcmpph to X86InstrInfo::commuteInstructionImpl.
They were already added to findCommuteOpIndices, but they also
need to be in X86InstrInfo::commuteInstructionImpl in order
to adjust the immediate control.
Craig Topper [Sun, 15 Aug 2021 18:32:28 +0000 (11:32 -0700)]
[X86] Add some tests to show incorrect commuting of vcmpsh instructions.
Sanjay Patel [Sun, 15 Aug 2021 17:39:28 +0000 (13:39 -0400)]
[x86] split memcmp tests for 32/64-bit targets; NFC
memcmp is defined as taking a size_t length arg,
so that differs depending on pointer size of the
target.
We casually matched non-compliant function signatures
as memcmp, but that can cause crashing as seen with
PR50850.
If we fix that bug, these tests would no longer be
testing the expected behavior for a 32-bit target,
so I have duplicated all tests and adjusted them
to match the stricter definition of memcmp/bcmp
by changing the length arg to i32 on a 32-bit target.
Paul Walker [Sun, 15 Aug 2021 13:24:20 +0000 (14:24 +0100)]
[DAGCombiner] Stop visitEXTRACT_SUBVECTOR creating illegal BITCASTs post legalisation.
visitEXTRACT_SUBVECTOR can sometimes create illegal BITCASTs when
removing "redundant" INSERT_SUBVECTOR operations. This patch adds
an extra check to ensure such combines only occur after operation
legalisation if any resulting BITBAST is itself legal.
Differential Revision: https://reviews.llvm.org/D108086
Nico Weber [Sun, 15 Aug 2021 17:21:36 +0000 (13:21 -0400)]
[gn build] (manually) port
957334382cd1
Kazu Hirata [Sun, 15 Aug 2021 16:31:39 +0000 (09:31 -0700)]
[AsmParser] Remove MDSignedOrUnsignedField (NFC)
The last use was removed on Apr 18, 2020 in commit
aad3d578da0ddf6d0d3d95e5e09a32e47f6dfeb8.
David Green [Sun, 15 Aug 2021 16:25:16 +0000 (17:25 +0100)]
[InstCombine] Add call to matchSAddSubSat from min/max
This adds a call to matchSAddSubSat from smin/smax instrinsics, allowing
the same patterns to match if the canonical form of a min/max is an
intrinsics, not a icmp/select.
Differential Revision: https://reviews.llvm.org/D108077
Roman Lebedev [Sun, 15 Aug 2021 16:01:44 +0000 (19:01 +0300)]
Reland [SimplifyCFG] performBranchToCommonDestFolding(): form block-closed SSA form before cloning instructions (PR51125)
... with test change this time.
LLVM IR SSA form is "implicit" in `@pr51125`. While is a valid LLVM IR,
and does not require any PHI nodes, that completely breaks the further logic
in `CloneInstructionsIntoPredecessorBlockAndUpdateSSAUses()`
that updates the live-out uses of the bonus instructions.
What i believe we need to do, is to first make the SSA form explicit,
by inserting tautological PHI nodes, and rewriting the offending uses.
```
$ /builddirs/llvm-project/build-Clang12/bin/opt -load /repositories/alive2/build-Clang-release/tv/tv.so -load-pass-plugin /repositories/alive2/build-Clang-release/tv/tv.so -tv -simplifycfg -simplifycfg-require-and-preserve-domtree=1 -bonus-inst-threshold=10 -tv -o /dev/null /tmp/test.ll
----------------------------------------
@global_pr51125 = global 4 bytes, align 4
define i32 @pr51125() {
%entry:
br label %L
%L:
%ld = load i32, * @global_pr51125, align 4
%iszero = icmp eq i32 %ld, 0
br i1 %iszero, label %exit, label %L2
%L2:
store i32
4294967295, * @global_pr51125, align 4
%cmp = icmp eq i32 %ld,
4294967295
br i1 %cmp, label %L, label %exit
%exit:
%r = phi i32 [ %ld, %L2 ], [ %ld, %L ]
ret i32 %r
}
=>
@global_pr51125 = global 4 bytes, align 4
define i32 @pr51125() {
%entry:
%ld.old = load i32, * @global_pr51125, align 4
%iszero.old = icmp eq i32 %ld.old, 0
br i1 %iszero.old, label %exit, label %L2
%L2:
%ld2 = phi i32 [ %ld.old, %entry ], [ %ld, %L2 ]
store i32
4294967295, * @global_pr51125, align 4
%cmp = icmp ne i32 %ld2,
4294967295
%ld = load i32, * @global_pr51125, align 4
%iszero = icmp eq i32 %ld, 0
%or.cond = select i1 %cmp, i1 1, i1 %iszero
br i1 %or.cond, label %exit, label %L2
%exit:
%ld1 = phi i32 [ poison, %L2 ], [ %ld.old, %entry ]
%r = phi i32 [ %ld2, %L2 ], [ %ld.old, %entry ]
ret i32 %r
}
Transformation seems to be correct!
```
Fixes https://bugs.llvm.org/show_bug.cgi?id=51125
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D106317
Roman Lebedev [Sun, 15 Aug 2021 16:15:09 +0000 (19:15 +0300)]
Revert "[SimplifyCFG] performBranchToCommonDestFolding(): form block-closed SSA form before cloning instructions (PR51125)"
Forgot to stage the test change.
This reverts commit
78af5cb213b2f9fe3f47bf23947f14ac07024155.
Roman Lebedev [Sun, 15 Aug 2021 16:01:44 +0000 (19:01 +0300)]
[SimplifyCFG] performBranchToCommonDestFolding(): form block-closed SSA form before cloning instructions (PR51125)
LLVM IR SSA form is "implicit" in `@pr51125`. While is a valid LLVM IR,
and does not require any PHI nodes, that completely breaks the further logic
in `CloneInstructionsIntoPredecessorBlockAndUpdateSSAUses()`
that updates the live-out uses of the bonus instructions.
What i believe we need to do, is to first make the SSA form explicit,
by inserting tautological PHI nodes, and rewriting the offending uses.
```
$ /builddirs/llvm-project/build-Clang12/bin/opt -load /repositories/alive2/build-Clang-release/tv/tv.so -load-pass-plugin /repositories/alive2/build-Clang-release/tv/tv.so -tv -simplifycfg -simplifycfg-require-and-preserve-domtree=1 -bonus-inst-threshold=10 -tv -o /dev/null /tmp/test.ll
----------------------------------------
@global_pr51125 = global 4 bytes, align 4
define i32 @pr51125() {
%entry:
br label %L
%L:
%ld = load i32, * @global_pr51125, align 4
%iszero = icmp eq i32 %ld, 0
br i1 %iszero, label %exit, label %L2
%L2:
store i32
4294967295, * @global_pr51125, align 4
%cmp = icmp eq i32 %ld,
4294967295
br i1 %cmp, label %L, label %exit
%exit:
%r = phi i32 [ %ld, %L2 ], [ %ld, %L ]
ret i32 %r
}
=>
@global_pr51125 = global 4 bytes, align 4
define i32 @pr51125() {
%entry:
%ld.old = load i32, * @global_pr51125, align 4
%iszero.old = icmp eq i32 %ld.old, 0
br i1 %iszero.old, label %exit, label %L2
%L2:
%ld2 = phi i32 [ %ld.old, %entry ], [ %ld, %L2 ]
store i32
4294967295, * @global_pr51125, align 4
%cmp = icmp ne i32 %ld2,
4294967295
%ld = load i32, * @global_pr51125, align 4
%iszero = icmp eq i32 %ld, 0
%or.cond = select i1 %cmp, i1 1, i1 %iszero
br i1 %or.cond, label %exit, label %L2
%exit:
%ld1 = phi i32 [ poison, %L2 ], [ %ld.old, %entry ]
%r = phi i32 [ %ld2, %L2 ], [ %ld.old, %entry ]
ret i32 %r
}
Transformation seems to be correct!
```
Fixes https://bugs.llvm.org/show_bug.cgi?id=51125
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D106317
Roman Lebedev [Sun, 15 Aug 2021 16:02:32 +0000 (19:02 +0300)]
[NFC][SimplifyCFG] Autogenerate check lines in a test to declutter further update
Roman Lebedev [Sun, 15 Aug 2021 15:59:32 +0000 (18:59 +0300)]
[NFCI][IndVars] rewriteLoopExitValues(): nowadays SCEV should not change `GEP` base pointer
Currently/previously, while SCEV guaranteed that it produces the same value,
the way it was produced may be illegal IR, so we have an ugly check that
the replacement is valid.
But now that the SCEV strictness wrt the pointer/integer types has been improved,
i believe this invariant is already upheld by the SCEV itself, natively.
I think we should add an assertion, wait for a week, and then, if all is good,
rip out all this checking.
Or we could just do the latter directly i guess.
This reverts commit rL127839.
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D108043
Nikita Popov [Sun, 15 Aug 2021 14:47:27 +0000 (16:47 +0200)]
[IndVars] Don't check for pointer exit count (NFC)
After recent changes, exit counts and BE taken counts are always
integers, so convert these to assertions.
While here, also convert the loop invariance checks to asserts.
Exit counts are always loop invariant.
Qiu Chaofan [Sun, 15 Aug 2021 14:43:46 +0000 (22:43 +0800)]
[NFC] Simply update a FIXME comment
X86 overrided LowerOperationWrapper was moved to common implementation
in a7eae62.
Nikita Popov [Sun, 15 Aug 2021 13:46:25 +0000 (15:46 +0200)]
[FunctionImport] Fix build with old mingw (NFC)
std::errc::operation_not_supported is not universally supported.
Make use of LLVM's errc interoperability header, which lists
known-good errc values.
Harald van Dijk [Sun, 15 Aug 2021 12:35:53 +0000 (13:35 +0100)]
[ExecutionEngine] Check for libunwind before calling __register_frame
libgcc and libunwind have different flavours of __register_frame. Both
flavours are already correctly handled, except that the code to handle
the libunwind flavour is guarded by __APPLE__. This change uses the
presence of __unw_add_dynamic_fde in libunwind instead to detect whether
libunwind is used, rather than hardcoding it as Apple vs. non-Apple.
Fixes PR44074.
Thanks to Albert Jin <albert.jin@gmail.com> and Chris Schafmeister
<chris.schaf@verizon.net> for identifying the problem.
Reviewed By: lhames
Differential Revision: https://reviews.llvm.org/D106129
Dávid Bolvanský [Sun, 15 Aug 2021 11:54:58 +0000 (13:54 +0200)]
[Clang] Updated warning-wall.c test file
-Wbool-operation was moved to -Wall and test file needs to be adjusted.
Dávid Bolvanský [Sun, 15 Aug 2021 11:34:24 +0000 (13:34 +0200)]
[Clang] Put -Wbool-operation under -Wall
To keep compatibility with GCC.
Paul Walker [Fri, 13 Aug 2021 11:47:51 +0000 (12:47 +0100)]
[LoopVectorize] Don't emit remarks about lack of scalable vectors unless they're specifically requested.
Previously we emitted a "does not support scalable vectors"
remark for all targets whenever vectorisation is attempted. This
pollutes the output for architectures that don't support scalable
vectors and is likely confusing to the user.
Instead this patch introduces a debug message that reports when
scalable vectorisation is allowed by the target and only issues
the previous remark when scalable vectorisation is specifically
requested, for example:
#pragma clang loop vectorize_width(2, scalable)
Differential Revision: https://reviews.llvm.org/D108028
Nikita Popov [Sat, 14 Aug 2021 21:35:27 +0000 (23:35 +0200)]
[AArch64] Fix comparison peephole opt with non-0/1 immediate (PR51476)
This is a non-intrusive fix for
https://bugs.llvm.org/show_bug.cgi?id=51476 intended for backport
to the 13.x release branch. It expands on the current hack by
distinguishing between CmpValue of 0, 1 and 2, where 0 and 1 have
the obvious meaning and 2 means "anything else". The new optimization
from D98564 should only be performed for CmpValue of 0 or 1.
For main, I think we should switch the analyzeCompare() and
optimizeCompare() APIs to use int64_t instead of int, which is in
line with MachineOperand's notion of an immediate, and avoids this
problem altogether.
Differential Revision: https://reviews.llvm.org/D108076
Dávid Bolvanský [Sun, 15 Aug 2021 09:44:13 +0000 (11:44 +0200)]
Revert "[Remarks] Emit optimization remarks for atomics generating CAS loop"
This reverts commit
435785214f73ff0c92e97f2ade6356e3ba3bf661. Still same compile time issues for -O0 -g, eg. +1.3% for sqlite3.
Andrzej Warzynski [Fri, 13 Aug 2021 13:03:21 +0000 (13:03 +0000)]
[flang][nfc] Move `Semantics` from `FrontendAction` to `CompilerInstance`
`CompilerInstance` is a more appropriate place for a key component of
the frontend like `Semantics`.
This change opens a path for us to introduce new frontend actions that
will also run semantics, but for which inheriting from
`PrescanAndSemaAction` wouldn't make much sense. For example, for
code-gen actions we plan to introduce a dedicate hierarchy of action
classes.
I've also added a doxyment for `CompilerInstance` to add a bit of
context for this change (and also make future refactoring more informed).
As `CompilerInstance` in Flang has been inspired by its counterpart in
Clang, this comment is roughly a verbatim copy of the comment in Clang
(with some adjustments from me). Credits to Daniel Dunbar for the great
design and the original comment.
Differential Revision: https://reviews.llvm.org/D108035
Rainer Orth [Sun, 15 Aug 2021 07:21:08 +0000 (09:21 +0200)]
[asan][test] Un-xfail Posix/unpoison-alternate-stack.cpp on Solaris again
`Posix/unpoison-alternate-stack.cpp` currently `XPASS`es on Solaris. The
`XFAIL` had already been removed in D97933
<https://reviews.llvm.org/D97933>, but reintroduced by commit
f03d29601e0951da2c88f07d4234128e14e87870
<https://reviews.llvm.org/rGf03d29601e0951da2c88f07d4234128e14e87870> which
was never posted or justified.
Given the `XPASS`, this obviously wasn't NFC, so I suggest to remove it again.
Tested on `amd64-pc-solaris2.11` and `x86_64-pc-linux-gnu`.
Differential Revision: https://reviews.llvm.org/D108030
Anshil Gandhi [Sun, 15 Aug 2021 05:37:15 +0000 (23:37 -0600)]
[Remarks] Emit optimization remarks for atomics generating CAS loop
Implements ORE in AtomicExpand pass to report atomics generating
a compare and swap loop.
Differential Revision: https://reviews.llvm.org/D106891
Itay Bookstein [Sun, 15 Aug 2021 05:01:10 +0000 (22:01 -0700)]
[Linker] Import GlobalIFunc when importing symbols from another module
Reviewed By: MaskRay
Differential Revision: https://reviews.llvm.org/D107988
Vitaly Buka [Sun, 15 Aug 2021 01:42:05 +0000 (18:42 -0700)]
[msan] Fix ppc64 format warning
Vitaly Buka [Sun, 15 Aug 2021 01:33:03 +0000 (18:33 -0700)]
[sanitizer] Improve VSNPrintf internal diagnostics
Craig Topper [Sun, 15 Aug 2021 00:24:21 +0000 (17:24 -0700)]
[X86] Add parentheses around casts in X86 intrinsic headers.
Fixes PR51324.
Dmitry Vyukov [Sun, 15 Aug 2021 00:41:14 +0000 (17:41 -0700)]
sanitizer_common: support %l in format strings
Currently we only support %z and %ll width modifiers,
but surprisingly not %l. This makes it impossible to print longs
(sizeof(long) not necessary equal to sizeof(size_t)).
We had some printf's that printed longs with %zu,
but that's wrong and now with __attribute__((format)) in place
they are flagged by compiler. So we either have a choice of
doing static_cast<uptr>(long) everywhere or add %l.
Adding %l looks better, that's a standard modifier.
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D108066
Wang, Pengfei [Sun, 15 Aug 2021 00:17:30 +0000 (08:17 +0800)]
[X86] AVX512FP16 instructions enabling 2/6
Enable FP16 binary operator instructions.
Ref.: https://software.intel.com/content/www/us/en/develop/download/intel-avx512-fp16-architecture-specification.html
Reviewed By: LuoYuanke
Differential Revision: https://reviews.llvm.org/D105264
Vitaly Buka [Sat, 14 Aug 2021 23:51:10 +0000 (16:51 -0700)]
[sanitizer] Define 32bit uptr as uint
This makes it consistent with uintptr_t.
Vitaly Buka [Sat, 14 Aug 2021 23:26:06 +0000 (16:26 -0700)]
[sanitizer] Fix format string
Craig Topper [Sat, 14 Aug 2021 23:41:50 +0000 (16:41 -0700)]
[X86] Use a do {} while (0) in the _MM_EXTRACT_FLOAT implementation.
Previously we just used {}, but that doesn't work in situations
like this.
if (1)
_MM_EXTRACT_FLOAT(d, x, n);
else
...
The semicolon would terminate the if.
Craig Topper [Sat, 14 Aug 2021 23:34:52 +0000 (16:34 -0700)]
[X86] Use __builtin_bit_cast _mm_extract_ps instead of type punning through a union. NFC
Rainer Orth [Sat, 14 Aug 2021 22:20:47 +0000 (00:20 +0200)]
[test] Avoid unportable echo in Other/lit-quoting.txt
`LLVM :: Other/lit-quoting.txt` currently `FAIL`s on Solaris:
llvm/test/Other/lit-quoting.txt:8:9: error: CHECK2: expected string not found in input
CHECK2: {{^a\[b\\c$}}
^
<stdin>:1:1: note: scanning from here
a[b
^
This happens because echo with backslashes or special characters is
unportable, as extensively documented in the Autoconf manual. In the case
at hand, `echo 'a[b\c'` yields `a[b\c` on Linux, but `a[b` (no newline) on
Solaris.
This patch fixes this by using the portable alternative suggested in the
Autoconf manual.
Tested on `amd64-pc-solaris2.11`, `sparcv9-sun-solaris2.11`, and
`x86_64-pc-linux-gnu`.
Differential Revision: https://reviews.llvm.org/D108031
Nico Weber [Sat, 14 Aug 2021 21:58:21 +0000 (17:58 -0400)]
Simplify a .mailmap entry
Only one person committed with these email addresses, so there's no need to use
the map-different-names-for-one-email-address syntax.
No behavior change.
Differential Revision: https://reviews.llvm.org/D108007
David Green [Sat, 14 Aug 2021 21:48:10 +0000 (22:48 +0100)]
[InstCombine] Extend sadd.sat tests to include min/max patterns. NFC
This tests code starting from smin/smax, as opposed to the icmp/select
form. Also adds a ARM MVE phase ordering test for vectorizing to
sadd.sat from the original IR.
Stephen Neuendorffer [Thu, 24 Sep 2020 18:54:46 +0000 (11:54 -0700)]
[MLIR] Move TestDialect to ::test namespace
While the changes are extensive, they basically fall into a few
categories:
1) Moving the TestDialect itself.
2) Updating C++ code in tablegen to explicitly use ::mlir, since it
will be put in a headers that shouldn't expect a 'using'.
3) Updating some generic MLIR Interface definitions to do the same thing.
4) Updating the Tablegen generator in a few places to be explicit about
namespaces
5) Doing the same thing for llvm references, since we no longer pick
up the definitions from mlir/Support/LLVM.h
Differential Revision: https://reviews.llvm.org/D88251
Nikita Popov [Sat, 14 Aug 2021 18:54:19 +0000 (20:54 +0200)]
[Tests] Remove explicit -enable-mssa-loop-dependency options (NFC)
This is enabled by default. Drop explicit uses in preparation for
removing the option.
Also drop RUN lines that are now the same (typically modulo a
-verify-memoryssa option).
luxufan [Sat, 14 Aug 2021 16:30:42 +0000 (00:30 +0800)]
[JITLink] Unify x86-64 MachO and ELF 's optimize GOT/Stub function
This patch unify optimizeELF_x86_64_GOTAndStubs and optimizeMachO_x86_64_GOTAndStubs into a pure optimize_x86_64_GOTAndStubs
Reviewed By: lhames
Differential Revision: https://reviews.llvm.org/D108025
Kazu Hirata [Sat, 14 Aug 2021 15:49:40 +0000 (08:49 -0700)]
[Aarch64] Remove redundant c_str (NFC)
Identified with readability-redundant-string-cstr.
Owen [Thu, 12 Aug 2021 13:12:25 +0000 (06:12 -0700)]
[clang-format] Distinguish K&R C function definition and attribute
This is a follow-up to https://reviews.llvm.org/D107950 which
missed user-defined types in K&R C.
Differential Revision: https://reviews.llvm.org/D107961
eopXD [Sat, 14 Aug 2021 07:58:05 +0000 (15:58 +0800)]
[LoopIdiom] let the pass deal with runtime memset size
The current LIR does not deal with runtime-determined memset-size. This patch
utilizes SCEV and check if the PointerStrideSCEV and the MemsetSizeSCEV are equal.
Before comparison the pass would try to fold the expression that is already
protected by the loop guard.
Testcase file `memset-runtime.ll`, `memset-runtime-debug.ll` added.
This patch deals with proper loop-idiom. Proceeding patch wants to deal with SCEV-s
that are inequal after folding with the loop guards.
Reviewed By: lebedev.ri, Whitney
Differential Revision: https://reviews.llvm.org/D107353
mydeveloperday [Sat, 14 Aug 2021 11:05:21 +0000 (12:05 +0100)]
[clang-tidy] [PR50069] readability-braces-around-statements doesn't work well with [[likely]] [[unlikely]]
https://bugs.llvm.org/show_bug.cgi?id=50069
When clang-tidy sees:
```
if (true) [[unlikely]] {
...
}
```
It thinks the braces are missing and add them again.
```
if (true) { [[unlikely]] {
...
}
}
```
This revision aims to prevent that incorrect code generation
Reviewed By: aaron.ballman
Differential Revision: https://reviews.llvm.org/D105479
mydeveloperday [Sat, 14 Aug 2021 09:29:07 +0000 (10:29 +0100)]
[clang-format] NFC update the ClangFormatStyleOption.rst following previous change
clang/docs/tool/dump_format_style.py was not run as part of {D99840}
Bring ClangFormatStyleOptions.rst back in line.
Reviewed By: HazardyKnusperkeks
Differential Revision: https://reviews.llvm.org/D107958
Dawid Jurczak [Tue, 10 Aug 2021 10:56:44 +0000 (12:56 +0200)]
[NFC][DSE] Clean up KnownNoReads and MemorySSAScanLimit in DSE
Another simple cleanups set in DSE. CheckCache is removed since
1f1145006b32 and in consequence KnownNoReads is useless.
Also update description of MemorySSAScanLimit which default value is 150 instead 100.
Differential Revision: https://reviews.llvm.org/D107812
Lang Hames [Sat, 14 Aug 2021 08:27:16 +0000 (18:27 +1000)]
[JITLink][x86-64] Rename *Relaxable edges to *REXRelaxable.
The existing relaxable edges all assume a REX prefix. ELF includes non-REX
relaxations, so rename these edges to make room for the new kinds.