Tom Eccles [Fri, 17 Mar 2023 09:26:09 +0000 (09:26 +0000)]
[flang][hlfir] add a test chaining intrinsics
This test checks all of the parts of intrinsic lowering work together,
and makes sure that we can pass the hlfir.expr result of one intrinsic
as an argument to another.
Depends on D145503
Reviewed By: jeanPerier
Differential Revision: https://reviews.llvm.org/D145504
Tom Eccles [Fri, 17 Mar 2023 09:26:02 +0000 (09:26 +0000)]
[flang][hlfir] clean up intrinsic arg extended values
Depends on D145502
Reviewed By: jeanPerier
Differential Revision: https://reviews.llvm.org/D145503
Tom Eccles [Fri, 17 Mar 2023 09:25:52 +0000 (09:25 +0000)]
[flang][hlfir] move intrinsic lowering out of BufferizeHLFIR
This move is useful for a few reasons:
- It is easier to see what the intrinsic lowering is doing when the
operations it creates are not immediately lowered
- When lowering a HLFIR intrinsic generates an operation which needs
to be lowered by another pattern matcher in the same pass, MLIR will
run that other substitution before validating and finalizing the
original changes. This means that the erasure of operations is not
yet visible to subsequent matchers, which hugely complicates
transformations (in this case, hlfir.exprs cannot be rewritten
because they are still used by the now-erased HLFIR intrinsic op.
Reviewed By: jeanPerier
Differential Revision: https://reviews.llvm.org/D145502
Dhruva Chakrabarti [Fri, 17 Mar 2023 08:55:27 +0000 (09:55 +0100)]
[OpenMP] [OMPT] [8/8] Added lit tests for OMPT target callbacks
Added a new target ompt mode that depends on libomptarget OMPT support.
Added tests that verify callbacks for target regions, kernel launch,
and data transfer operations. All of them should pass on amdgpu using
make check-libomptarget.
Reviewed By: jplehr
Differential Revision: https://reviews.llvm.org/D127372
Nikita Popov [Thu, 16 Mar 2023 10:27:25 +0000 (11:27 +0100)]
[SCEV] Recognize vscale intrinsics
Now that SCEV has a dedicated vscale node type, we should also map
vscale intrinsics to it. To make sure this does not regress ranges
(which were KnownBits based previously), add support for vscale to
getRangeRef() as well.
Differential Revision: https://reviews.llvm.org/D146226
Nikita Popov [Thu, 16 Mar 2023 11:42:59 +0000 (12:42 +0100)]
[ValueTracking] Support vscale in computeConstantRange()
Add support for vscale in computeConstantRange(), based on
vscale_range attributes. This allows simplifying based on the
precise range, rather than a KnownBits approximation (which will
be off by a factor of two for the usual case of a power of two
upper bound).
Differential Revision: https://reviews.llvm.org/D146217
Nikita Popov [Thu, 16 Mar 2023 08:57:44 +0000 (09:57 +0100)]
[PassBuilder] Support O0 in default pipelines
The default and pre-link pipeline builders currently require you to
call a separate method for optimization level O0, even though they
have perfectly well-defined O0 optimization pipelines.
Accept O0 optimization level and call buildO0DefaultPipeline()
internally, so all consumers don't need to repeat this.
Differential Revision: https://reviews.llvm.org/D146200
Bjorn Pettersson [Thu, 16 Mar 2023 18:48:06 +0000 (19:48 +0100)]
[Analysis] Make order of analysis executions more stable
When debugging and using debug-pass-manager (e.g. in regression tests)
we prefer a consistent order in which analysis passes are executed.
But when for example doing
return MyClass(AM.getResult<LoopAnalysis>(F),
AM.getResult<DominatorTreeAnalysis>(F));
then the order in which LoopAnalysis and DominatorTreeAnalysis isn't
guaranteed, and might for example depend on which compiler that is
used when building LLVM.
I've not scanned the full source tree, but this fixes some occurances
of the above pattern found in lib/Analysis.
This problem was discussed briefly in review for D146206.
Bjorn Pettersson [Thu, 16 Mar 2023 01:00:47 +0000 (02:00 +0100)]
[LAA] Fix transitive analysis invalidation bug by implementing LoopAccessInfoManager::invalidate
The default invalidate method for analysis results is just looking
at the preserved state of the pass itself. It does not consider if
the analysis has an internal state that depend on other analyses.
Thus, we need to implement LoopAccessInfoManager::invalidate in order
to catch if LoopAccessAnalysis needs to be invalidated due to
transitive analyses such as AAManager is being invalidated. Otherwise
we might end up having references to an AAManager that is stale.
Fixes https://github.com/llvm/llvm-project/issues/61324
Differential Revision: https://reviews.llvm.org/D146206
Bjorn Pettersson [Thu, 16 Mar 2023 00:57:33 +0000 (01:57 +0100)]
[test][MemDep] Correct the 'NOT' checks in the invalidation.ll test case. NFC
The MemoryDependenceAnalysis/invalidation.ll test case was using
; CHECK-NOT-AA-INVALIDATE:
but I think the intention was to use
; CHECK-AA-INVALIDATE-NOT:
Simply changing the checks like that would make the test fail.
The old statement that AA being stateless would result in nothing
to invalidate when doing invalidate<aa> is not true afaict.
It would be different if for example doing invalidate<basic-aa>, then
the AAManager isn't invalidated (and then neither memdep would be
invalidated). But when the AAManager itself is invalidated then we
should expect to find both "Invalidating analysis: AAManager" and
"Invalidating analysis: MemoryDependenceAnalysis" in the output.
Differential Revision: https://reviews.llvm.org/D146205
Xi Ruoyao [Fri, 17 Mar 2023 08:27:41 +0000 (09:27 +0100)]
[libunwind][AArch64] Unbreak building with GNU assembler
GNU assembler mandates armv8.5-a for memtag instructions. Maybe
we should remove this restriction in GNU assembler, but let's work
around it for current GNU Binutils releases.
Differential Revision: https://reviews.llvm.org/D146109
Hans Wennborg [Thu, 16 Mar 2023 10:23:31 +0000 (11:23 +0100)]
Make globals used for array initialization codegen constant
As pointed out in D133835 these globals will never be written to
(they're only used for trivially copyable types), so they can always be
constant.
Differential revision: https://reviews.llvm.org/D146211
Chuanqi Xu [Fri, 17 Mar 2023 07:53:24 +0000 (15:53 +0800)]
[NFC] [C++20] [Modules] Tests that the duplicated declarations in GMF
won't get generated again
As the test shows, we don't want to see the specialized function bodies
if it is already contained in the imported modules. So we can save a lot
of compiling time then.
Francesco Petrogalli [Wed, 15 Mar 2023 12:09:03 +0000 (13:09 +0100)]
[Target][RISCV] Update SubtargetFeature definition for RV32/RV64 (NFCI).
This is done for consistency with other Predicate/Subtargetfeature
pairs, where the second parameter of the SubtargetFeature correspond
to the NAME of the def of the Predicate associated to the
SubtargetFeature.
Differential Revision: https://reviews.llvm.org/D146129
dbakunevich [Tue, 7 Mar 2023 10:34:00 +0000 (17:34 +0700)]
[Tools] Changed the number printing format for sorting
As part of this patch, I added the ability to add leading zeros.
This is necessary so that the generated files are sorted in ascending order.
Reviewed By: yrouban
Differential Revision: https://reviews.llvm.org/D145484
Wang, Xin10 [Fri, 17 Mar 2023 07:19:51 +0000 (03:19 -0400)]
[X86][MC]Fix wrong action for encode movdir64b
Movdir64b is special for its mem operand, 67 prefex can not only modify its add size,
so it's mem base and index reg should be the same type as source reg, such as
movdir64b (%rdx), rcx, and could not be movdir64b (%edx), rcx.
Now llvm-mc can encode the asm 'movdir64b (%edx), rcx' but the result is the same as
'movdir64b (%edx), ecx', which offend users' intention, while gcc will object this
action and give a warning.
I add 3 new mem descriptions to let llvm-mc to report the same error.
Reviewed By: skan, craig.topper
Differential Revision: https://reviews.llvm.org/D145893
Ethan Luis McDonough [Fri, 17 Mar 2023 07:26:51 +0000 (02:26 -0500)]
[flang] Feature list plugin
Plugin that counts the number of times each tree node occurs in a given program. Used for test coverage.
> One thing we need...is a way to determine what features a code uses. Preferably we would also be able to determine if they are implemented or not. Just the former could be done with a visitor for the parse tree. For the latter we would continue compilation and somehow ignore todo errors but collect them - @jdoerfert
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D143704
Karl-Johan Karlsson [Fri, 17 Mar 2023 06:28:38 +0000 (07:28 +0100)]
[compiler-rt] Avoid signed overflow in floatdidf.c and floatdisf.c
When compiling compiler-rt with -fsanitize=undefined and running testcases you
end up with the following warning:
UBSan: floatdisf.c:27:15: signed integer overflow:
9223372036854775807 - -1 cannot be represented in type 'di_int' (aka 'long long')
This can be avoided by doing the subtraction in a matching unsigned variant of
the type, given that the overflow is the expected result of the subtraction.
The same kind of pattern exists in floatdidf.c
This was found in an out of tree target.
Reviewed By: phosek
Differential Revision: https://reviews.llvm.org/D146135
Chuanqi Xu [Fri, 17 Mar 2023 07:10:26 +0000 (15:10 +0800)]
[Modules] [doc] Document the problem that we can't include headers before import declarations
Wang, Xin10 [Fri, 17 Mar 2023 06:16:58 +0000 (02:16 -0400)]
[X86]add assert to confirm not-null ptr in getArithmeticReductionCost
For the function getArithmeticReductionCost, it receive a ptr and dereferce it without check,
It is called many times in getTypeBasedIntrinsicInstrCost, the ptr passed to it is inited
from line 1709.
From the code, we can not ensure the ptr VecOpTy is inited when Tys is empty or Tys[VecTyIndex]
is not a VectorType, so that the getArithmeticReductionCost will do an undefined behavior.
I add assert to it, found the ptr passed to it in llvm tests are all not nullptr, but I think the check is
still meaningful for us.
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D146118
Vitaly Buka [Fri, 17 Mar 2023 06:13:32 +0000 (23:13 -0700)]
Revert "[RFC][X86][MemFold] Upgrade the mechanism of auto-generated Memory Folding Table"
X86FoldTablesEmitter.cpp:386:26: runtime error: shift exponent -1 is negative
This reverts commit
ca4c5331823755d5f7ae28a8bcaaa0fed21cf762.
Chuanqi Xu [Fri, 17 Mar 2023 05:37:51 +0000 (13:37 +0800)]
Remove the outdated feature macro '__cpp_coroutines'
The feature macro '__cpp_coroutines' is for coroutines TS. And the
coroutines TS is deprecated. So we should remove the feature macro too.
BTW, the corresponding feature macro for standard c++ coroutines is
'__cpp_impl_coroutine'.
Craig Topper [Fri, 17 Mar 2023 04:37:35 +0000 (21:37 -0700)]
Recommit "[Sema] Fix null pointer dereference handleAlwaysInlineAttr."
Remove use of constexpr if that failed on the build bots.
Original commit message:
It's possible for `getCalleeDecl()` to return a null pointer.
This was encountered by a user of our downstream compiler.
The case involved a DependentScopeDeclRefExpr.
Since this seems to only be for a warning diagnostic, I skipped
the diagnostic check if it returned null. But mabye there's a
different way to fix this.
Wang, Xin10 [Fri, 17 Mar 2023 02:55:08 +0000 (22:55 -0400)]
[X86] Remove unreachable code in X86TargetTransformInfo.cpp
In Function getVectorInstrCost, situation Opcode == Instruction::ExtractElement
and Opcode == Instruction::InsertElement are all handled in the first 2 if-statements,
So we have no chance for the code in line 4401.
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D145908
Zhangyin [Thu, 9 Mar 2023 06:14:43 +0000 (14:14 +0800)]
[libcxx] Add status paper for Parallelism-TS
Reviewed By: #libc, philnik, Mordante
Differential Revision: https://reviews.llvm.org/D145652
Matt Arsenault [Fri, 24 Feb 2023 14:57:39 +0000 (10:57 -0400)]
Attributor: Start inferring nofpclass
Some of this is boilerplate which doesn't do anything yet.
Matt Arsenault [Fri, 3 Mar 2023 14:12:23 +0000 (10:12 -0400)]
ValueTracking: Extract fcmpToClassTest out of InstCombine
Also update unsigned to FPClassTest
Matt Arsenault [Fri, 24 Feb 2023 14:08:46 +0000 (10:08 -0400)]
ValueTracking: Handle nofpclass in computeKnownFPClass
Matt Arsenault [Wed, 1 Feb 2023 14:44:06 +0000 (10:44 -0400)]
ValueTracking: Add start of computeKnownFPClass API
Add a new compute-known-bits like function to compute all
the interesting floating point properties at once.
Eventually this should absorb all the various floating point
queries we already have.
Craig Topper [Fri, 17 Mar 2023 03:05:33 +0000 (20:05 -0700)]
Revert "[Sema] Fix null pointer dereference handleAlwaysInlineAttr."
This reverts commit
10297470e953f4f3968c54c851c8af82b07af00b.
This is failing on a build bot.
Craig Topper [Fri, 17 Mar 2023 02:34:16 +0000 (19:34 -0700)]
[RISCV] Add MULW to RISCVStripWSuffix.
This converts MULW to MUL if the upper bits aren't used.
This will give more opportunities to use c.mul with Zcb.
Vitaly Buka [Fri, 17 Mar 2023 02:16:16 +0000 (19:16 -0700)]
Revert "[LSAN][HWASAN] Turn on leak sanitizer in HWASAN for Linux"
Breaks x86_64 qemu tests.
This reverts commit
6c756486d44914bac6fb20513b8c41f1a5726c4c.
Vitaly Buka [Fri, 17 Mar 2023 02:03:46 +0000 (19:03 -0700)]
Revert "[AMDGPUUnifyDivergentExitNodes] Add NewPM support"
Introduces nullptr dereference.
This reverts commit
a5455e32b364dabe499ec11722626d4bbaf047ba.
Tue Ly [Fri, 17 Mar 2023 01:17:20 +0000 (21:17 -0400)]
[libc][Obvious] Reduce the size of the tests for hypot(f).
Craig Topper [Thu, 16 Mar 2023 22:52:39 +0000 (15:52 -0700)]
[Sema] Fix null pointer dereference handleAlwaysInlineAttr.
It's possible for `getCalleeDecl()` to return a null pointer.
This was encountered by a user of our downstream compiler.
The case involved a DependentScopeDeclRefExpr.
Since this seems to only be for a warning diagnostic, I skipped
the diagnostic check if it returned null. But mabye there's a
different way to fix this.
Reviewed By: erichkeane
Differential Revision: https://reviews.llvm.org/D146089
Nikolas Klauser [Fri, 17 Feb 2023 10:31:41 +0000 (11:31 +0100)]
[runtimes] Synchronize warnings flags between libc++/libc++abi/libunwind
This mostly keeps the same warning flags. The most important exceptions are `-Wpedantic` and `-Wconversion`, which are now removed from libc++abi and libunwind.
Reviewed By: ldionne, #libunwind, #libc, #libc_abi
Spies: mikhail.ramalho, phosek, libcxx-commits
Differential Revision: https://reviews.llvm.org/D144252
Daniel Thornburgh [Thu, 16 Mar 2023 23:15:38 +0000 (16:15 -0700)]
Remove #if guards from AppendBounds to fix compile on old POSIX.
Stanislav Mekhanoshin [Thu, 16 Mar 2023 23:14:02 +0000 (16:14 -0700)]
[AMDGPU] Simplify AGPR reservation. NFC.
The intent of the code was to reserve all AGPRs on a target without
AGPRs, so just do exactly that. Existing code was unsound.
Alex Zinenko [Thu, 16 Mar 2023 16:58:53 +0000 (16:58 +0000)]
[mlir][transform] change RaggedArray internals
Change the internal storage scheme from storing a MutableArrayRef to
storing an explicit offset+length pair. Storing an ArrayRef is dangerous
because it contains the pointer to the first element in the range, but
the entire storage vector may be reallocated, making the pointer
dangling. We don't know when the reallocation happends, so we can't
update the ArrayRefs. Store the explicit offset instead and construct
ArrayRefs on-the-fly.
Reviewed By: Peiming
Differential Revision: https://reviews.llvm.org/D146239
Shoaib Meenai [Thu, 16 Mar 2023 22:43:36 +0000 (15:43 -0700)]
[TableGen] Fix test on macOS
macOS cmp requires the option to appear before the file names.
Juergen Ributzka [Thu, 16 Mar 2023 22:28:48 +0000 (15:28 -0700)]
[clang] Unconditionally add autolink hints for frameworks.
Clang infers framework autolink hints when parsing a modulemap. In order to do
so, it checks if the module is a framework and if there is a framework binary
or TBD file in the SDK. Only when Clang finds the filei, then the autolink hint
is added to the module metadata.
During a project build many clang processes perform this check, which causes
many stat calls - even for modules/frameworks that are not even used.
The linker is already resilient to non-existing framework links that come from
the autolink metadata, so there is no need for Clang to do this check.
Instead the autolink hints are now added unconditionally and the linker only
needs to do the check once. This reduces the overall number of stat calls.
This fixes rdar://
106578342.
Differential Revision: https://reviews.llvm.org/D146255
Alex Langford [Tue, 7 Mar 2023 18:12:14 +0000 (10:12 -0800)]
[lldb] Make MemoryCache::Read more resilient
MemoryCache::Read is not resilient to partial reads when reading memory
chunks less than or equal in size to L2 cache lines. There have been
attempts in the past to fix this but nothing really solved the root of
the issue.
I first created a test exercising MemoryCache's implementation and
documenting how I believe MemoryCache::Read should behave. I then
rewrote the implementation of MemoryCache::Read as needed to make sure
that the different scenarios behaved correctly.
rdar://
105407095
Differential Revision: https://reviews.llvm.org/D145624
Craig Topper [Thu, 16 Mar 2023 22:20:25 +0000 (15:20 -0700)]
[RISCV] Remove ExtName from RISCVExtensionInfo.
This field is never used in the compiler and was only used in
unit tests added recently.
It's only used as the value in a map where the extension name
is the key. If we need the string we can get it from the key.
Reviewed By: asb
Differential Revision: https://reviews.llvm.org/D146197
Douglas Yung [Thu, 16 Mar 2023 22:01:58 +0000 (15:01 -0700)]
Mark test added in D145155 as requiring asserts since it uses the "-debug-only" option.
This should fix the test failure in Release builds.
River Riddle [Tue, 14 Mar 2023 00:32:10 +0000 (17:32 -0700)]
[mlir][StorageUniquer] Use allocators per thread instead of per shard
This greatly reduces the number of allocators we create, while still
retaining thread safety. Reducing the number of allocators is much
better for locality and memory usage; this revision drops memory
usage for some MLIR heavy workloads (with lots of attributes/types)
by >=5%. This is due to the observation that the number of threads
is effectively always smaller than the number of parametric attributes/types.
Differential Revision: https://reviews.llvm.org/D145991
Kazu Hirata [Thu, 16 Mar 2023 21:55:28 +0000 (14:55 -0700)]
[Inliner] clang-format InlineCost.cpp and Inliner.cpp (NFC)
Kirill Stoimenov [Thu, 16 Mar 2023 21:50:59 +0000 (21:50 +0000)]
[HWASAN] Remove RUN statements, which have to effect on the test
Philip Reames [Thu, 16 Mar 2023 21:03:27 +0000 (14:03 -0700)]
[LSR] Add tests which demonstrate miscompiles in the current term-fold code
Mircea Trofin [Fri, 3 Mar 2023 02:45:34 +0000 (18:45 -0800)]
[DAE] Don't DAE if we musttail call a "live" (non-DAE-able) function
There are 2 such base cases: indirect calls and calls to functions external
to the module; and then any musttail calls to live functions (because of
the first 2 reasons or otherwise).
The IR validator reports, in these cases, that it "cannot guarantee tail
call due to mismatched parameter counts".
The fix is two-fold: first, we mark as "live" (i.e. non-DAE-able)
functions that make an indirect musttail call.
Then, we propagate live-ness to musttail callers of live functions.
Declared functions are already marked "live".
Differential Revision: https://reviews.llvm.org/D145209
Ilia Diachkov [Mon, 27 Feb 2023 18:16:48 +0000 (21:16 +0300)]
[SPIRV] fix several issues in builds with expensive checks
The patch fixes "Virtual register does not match instruction constraint"
and partly "Illegal virtual register for instruction" fails in the SPIRV
backend builds with LLVM_ENABLE_EXPENSIVE_CHECKS enabled. As a result,
the number of passed LIT tests with enabled checks is doubled.
Also, support for ndrange_*D builtins is placed in a separate function.
Differential Revision: https://reviews.llvm.org/D144897
Peiming Liu [Wed, 18 Jan 2023 19:24:42 +0000 (19:24 +0000)]
[mlir][sparse] add new sparisification option for dependent index reduction-based codegen
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D142927
Peiming Liu [Thu, 16 Mar 2023 17:57:02 +0000 (17:57 +0000)]
[mlir][sparse] add a coo slices matmul integration test.
Reviewed By: wrengr
Differential Revision: https://reviews.llvm.org/D146248
Kazu Hirata [Thu, 16 Mar 2023 20:02:55 +0000 (13:02 -0700)]
[mlir] Fix a warning
This patch fixes:
mlir/lib/Dialect/Linalg/TransformOps/LinalgTransformOps.cpp:479:13:
error: unused variable 'tiledOp' [-Werror,-Wunused-variable]
zhijian [Thu, 16 Mar 2023 19:26:12 +0000 (15:26 -0400)]
[AIX] unset bit "IsBackChainStored" of traceback table for leaf functions with no stack frame
Summary:
In function PPCAIXAsmPrinter::emitTracebackTable() ,the bit "IsBackChainStored" of traceback
table always set true, it will cause aix debug tools "dbx" emit an error info
"libdebug assertion "(framep->getGpr(STKP, &addr) == DB_SUCCESS && *nextStkpp == addr)"
when debug a leaf functions with no stack frame.
If a a leaf functions with no stack frame , the bit IsBackChainStored should be unset.
Reviewers: ChenZheng
Differential Revision: https://reviews.llvm.org/D146071
Sam James [Thu, 16 Mar 2023 19:03:42 +0000 (19:03 +0000)]
[libc] [test] Add missing <cstdint> include
Fixes build w/ GCC 13:
```
[587/7274] Building CXX object projects/libc/test/src/__support/CMakeFiles/libc_system_str_to_float_comparison_test.dir/str_to_float_comparison_test.cpp.o
FAILED: projects/libc/test/src/__support/CMakeFiles/libc_system_str_to_float_comparison_test.dir/str_to_float_comparison_test.cpp.o
/usr/lib/llvm/16/bin/clang++ -D_DEBUG -D_GNU_SOURCE -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS -I/home/sam/git/llvm-project/build/projects/libc/test/src/__support -I/home/sam/git/llvm-project/libc/test/src/__support -I/home/sam/git/llvm-project/build/include -I/home/sam/git/llvm-project/llvm/include -fPIC -fno-semantic-interposition -fvisibility-inlines-hidden -Werror=date-time -Werror=unguarded-availability-new -Wall -Wextra -Wno-unused-parameter -Wwrite-strings -Wcast-qual -Wmissing-field-initializers -pedantic -Wno-long-long -Wc++98-compat-extra-semi -Wimplicit-fallthrough -Wcovered-switch-default -Wno-noexcept-type -Wnon-virtual-dtor -Wdelete-non-virtual-dtor -Wsuggest-override -Wstring-conversion -Wmisleading-indentation -Wctad-maybe-unsupported -fdiagnostics-color -g -std=c++17 -MD -MT projects/libc/test/src/__support/CMakeFiles/libc_system_str_to_float_comparison_test.dir/str_to_float_comparison_test.cpp.o -MF projects/libc/test/src/__support/CMakeFiles/libc_system_str_to_float_comparison_test.dir/str_to_float_comparison_test.cpp.o.d -o projects/libc/test/src/__support/CMakeFiles/libc_system_str_to_float_comparison_test.dir/str_to_float_comparison_test.cpp.o -c /home/sam/git/llvm-project/libc/test/src/__support/str_to_float_comparison_test.cpp
/home/sam/git/llvm-project/libc/test/src/__support/str_to_float_comparison_test.cpp:34:15: error: unknown type name 'uint32_t'
static inline uint32_t hexCharToU32(char in) {
^
/home/sam/git/llvm-project/libc/test/src/__support/str_to_float_comparison_test.cpp:39:15: error: unknown type name 'uint32_t'
static inline uint32_t fastHexToU32(const char *inStr) {
^
```
Reviewed By: michaelrj
Differential Revision: https://reviews.llvm.org/D146111
Volodymyr Sapsai [Tue, 28 Feb 2023 03:35:44 +0000 (19:35 -0800)]
[Driver] Allow to collect `-save-stats` data to a file specified in the environment variable.
Using two environment variables `CC_PRINT_INTERNAL_STAT` and
`CC_PRINT_INTERNAL_STAT_FILE` to work like `CC_PRINT_PROC_STAT`.
The purpose of the change is to allow collecting the internal stats
without modifying the build scripts. Write all stats to a single file
to simplify aggregating the data.
Differential Revision: https://reviews.llvm.org/D144981
alosharm [Thu, 3 Nov 2022 05:19:28 +0000 (10:49 +0530)]
[DebugInfo][SROA] Correct debug info for global variables spanning multiple fragments in case of SROA
The function transferSRADebugInfo is modified to include missing cases.
Reviewed By: aprantl
Differential Revision: https://reviews.llvm.org/D138992
Fabio D'Urso [Thu, 16 Mar 2023 18:07:32 +0000 (18:07 +0000)]
[scudo] Test ScudoReleaseTest.BufferPool with a buffer pool allocated on the heap
The previous code resulted in 252 KiB being allocated on the stack,
which caused a stack overflow on Fuchsia.
Reviewed By: Chia-hungDuan
Differential Revision: https://reviews.llvm.org/D146229
Mahesh Ravishankar [Thu, 16 Mar 2023 18:01:03 +0000 (18:01 +0000)]
Fix asan error uncovered after landing D145133.
Differential Revision: https://reviews.llvm.org/D146249
Aaron Ballman [Thu, 16 Mar 2023 18:03:04 +0000 (14:03 -0400)]
Fix LLVM Sphinx build
This addresses the issue found by:
https://lab.llvm.org/buildbot/#/builders/30/builds/33006
Kirill Stoimenov [Thu, 16 Mar 2023 18:01:14 +0000 (18:01 +0000)]
[HWASAN] Disable leak checking for forced_unwind2.pass.cpp
Noah Goldstein [Wed, 15 Mar 2023 21:53:13 +0000 (16:53 -0500)]
[InstCombine] Add constant combines for `(urem/srem (mul X, Y), (mul X, Z))`
We can handle the following cases + some `nsw`/`nuw` flags:
`(srem (mul X, Y), (mul X, Z))`
[If `srem(Y, Z) == 0`]
-> 0
- https://alive2.llvm.org/ce/z/PW4XZ-
[If `srem(Y, Z) == Y`]
-> `(mul nuw nsw X, Y)`
- https://alive2.llvm.org/ce/z/DQe9Ek
-> `(mul nsw X, Y)`
- https://alive2.llvm.org/ce/z/Nr_MdH
[If `Y`/`Z` are constant]
-> `(mul/shl nuw nsw X, (srem Y, Z))`
- https://alive2.llvm.org/ce/z/ccTFj2
- https://alive2.llvm.org/ce/z/i_UQ5A
-> `(mul/shl nsw X, (srem Y, Z))`
- https://alive2.llvm.org/ce/z/mQKc63
- https://alive2.llvm.org/ce/z/uERkKH
`(urem (mul X, Y), (mul X, Z))`
[If `urem(Y, Z) == 0`]
-> 0
- https://alive2.llvm.org/ce/z/LL7UVR
[If `srem(Y, Z) == Y`]
-> `(mul nuw nsw X, Y)`
- https://alive2.llvm.org/ce/z/9Kgs_i
-> `(mul nuw X, Y)`
- https://alive2.llvm.org/ce/z/ow9i8u
[If `Y`/`Z` are constant]
-> `(mul nuw nsw X, (srem Y, Z))`
- https://alive2.llvm.org/ce/z/mNnQqJ
- https://alive2.llvm.org/ce/z/Bj_DR-
- https://alive2.llvm.org/ce/z/X6ZEtQ
-> `(mul nuw X, (srem Y, Z))`
- https://alive2.llvm.org/ce/z/SJYtUV
The rationale for doing this all in `InstCombine` rather than handling
the constant `mul` cases in `InstSimplify` is we often create a new
instruction because we are able to deduce more `nsw`/`nuw` flags than
the original instruction had.
Reviewed By: MattDevereau, sdesmalen
Differential Revision: https://reviews.llvm.org/D143014
Noah Goldstein [Tue, 31 Jan 2023 17:33:11 +0000 (11:33 -0600)]
[InstCombine] Add tests for combining (urem/srem (mul/shl X, Y), (mul/shl X, Z)); NFC
Differential Revision: https://reviews.llvm.org/D143013
Mark de Wever [Wed, 15 Mar 2023 17:19:25 +0000 (18:19 +0100)]
[NFC][libc++] Formats granularized charconv.
Reviewed By: #libc, Mordante
Differential Revision: https://reviews.llvm.org/D146153
Siva Chandra Reddy [Thu, 16 Mar 2023 17:47:31 +0000 (17:47 +0000)]
[libc][Obvious] Remove the STARTUP attribute from the api test listing.
Arthur Eubanks [Thu, 16 Mar 2023 17:05:34 +0000 (10:05 -0700)]
[docs] Mention removal of legacy optimization pipeline and inliner alloca merging
Reviewed By: asbirlea
Differential Revision: https://reviews.llvm.org/D146243
Shao-Ce SUN [Thu, 16 Mar 2023 13:01:57 +0000 (21:01 +0800)]
[NFC][ADT] Tweaked some functions in StringRef
Reduced the usage of variables in the `rfind` and `count` functions.
Replaced calls to `drop_front` with direct use of `substr` in `consume_front`,
since `starts_with` already performs length check on the string. Did the same
for other functions.
Using the `std::clamp` in `slice` makes the code more straightforward.
Reviewed By: bkramer
Differential Revision: https://reviews.llvm.org/D144787
Joseph Huber [Thu, 16 Mar 2023 17:41:28 +0000 (12:41 -0500)]
[libc] Remove unused startup source file
Summary:
This was not removed from the previous patch. Fix that.
Kirill Stoimenov [Thu, 16 Mar 2023 17:36:53 +0000 (17:36 +0000)]
[HWASAN] Disable leak checking for test_vector2.pass.cpp
Joseph Huber [Thu, 16 Mar 2023 16:50:39 +0000 (11:50 -0500)]
[libc] Remove startup option from integration tests
This generalizes handling of the integration tests. We now implicitly
depend on the `libc.startup.${LIBC_TARGET_OS}.crt1` file rather than
passing it in manually. This simplifies the interface.
Reviewed By: sivachandra
Differential Revision: https://reviews.llvm.org/D146237
Valery N Dmitriev [Fri, 24 Feb 2023 01:39:23 +0000 (17:39 -0800)]
[TTI] Add X86 target specific version of getPointersChainCost.
When all the pointers are off the same base address and have known
distances to each other these differences can be encoded into displacements
in x86 arch. So the only cost that matters is cost of the base GEP.
Differential Revision: https://reviews.llvm.org/D146102
Mirko Brkusanin [Thu, 16 Mar 2023 13:44:26 +0000 (14:44 +0100)]
[AMDGPU] Select flat atomic fmin/fmax
Also disables global atomic fmin/fmax x2 patterns on gfx11
Differential Revision: https://reviews.llvm.org/D146137
Florian Hahn [Thu, 16 Mar 2023 17:06:44 +0000 (17:06 +0000)]
[VPlan] Consolidate replicate region optimizations (NFC).
As suggested in D143865, consolidate replicate region creation and
optimization in a single helper that's exposed and used by LV.
Arthur Eubanks [Tue, 14 Mar 2023 20:52:57 +0000 (13:52 -0700)]
[Inliner] Remove alloca merging
See https://discourse.llvm.org/t/rfc-we-should-stop-merging-allocas-in-the-inliner/41975 for the RFC from 2016 around this.
Backend stack coloring takes care of what this initially set out to help with.
This only affects the legacy inliner, which for the most part doesn't matter anymore. The new PM inliner never looked at this flag.
There are some backends running the legacy always-inline pass, but that's unlikely to matter.
Reviewed By: asbirlea
Differential Revision: https://reviews.llvm.org/D146091
Arthur Eubanks [Tue, 14 Mar 2023 23:23:53 +0000 (16:23 -0700)]
[opt] Rename -enable-new-pm -> -bugpoint-enable-legacy-pm
-enable-new-pm is no longer necessary except for bugpoint. Make the name more clunky so it hopefully won't be used.
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D146103
Arthur Eubanks [Thu, 16 Mar 2023 16:47:21 +0000 (09:47 -0700)]
[Debugify] Use ModuleAnalysisManager in instrumentation
In preparation for doing module checks of PreservedAnalyses.
Jessica Clarke [Thu, 16 Mar 2023 16:47:07 +0000 (16:47 +0000)]
workflows: Don't try and run llvm-bugs in forks
Most forks won't have an LLVM_BUGS_KEY secret, so Mailgun will error out
with a missing key, and if forks do happen to have such a secret then we
still don't want to be getting emails to LLVM's mailing lists for
downstream forks.
Reviewed By: tstellar
Differential Revision: https://reviews.llvm.org/D146235
Peiming Liu [Tue, 14 Mar 2023 18:18:20 +0000 (18:18 +0000)]
[mlir][sparse] add test cases for sparse tensor slices on COO tensors
Reviewed By: aartbik, bixia
Differential Revision: https://reviews.llvm.org/D146074
Arthur Eubanks [Thu, 16 Mar 2023 16:38:35 +0000 (09:38 -0700)]
[HWAsan] Fix returned PreservedAnalyses
Initialization modifies the module.
Arthur Eubanks [Thu, 16 Mar 2023 16:37:26 +0000 (09:37 -0700)]
[FunctionImport] Fix returned PreservedAnalyses
Arthur Eubanks [Thu, 16 Mar 2023 16:36:48 +0000 (09:36 -0700)]
[Annotation2Metadata] Fix returned PreservedAnalyses
David Spickett [Thu, 16 Mar 2023 16:29:18 +0000 (16:29 +0000)]
[flang] Make tests depend on llvm-readobj not llvm-readelf
d4320cb2a5ef1680e519fa6b7cfd3a2f88cfed16 fixed a regular build
issue by making the flang tests depend on llvm-readelf.
Turns out llvm-readelf is a symlink to llvm-readobj and for
whatever reason, doesn't have the right exports at the moment
to be picked up in a standalone build.
https://lab.llvm.org/buildbot/#/builders/175/builds/26448
Maybe we should fix that, but for now, just require llvm-readobj
instead.
Arthur Eubanks [Thu, 16 Mar 2023 16:35:52 +0000 (09:35 -0700)]
[AMDGPUCtorDtorLowering] Fix returned PreservedAnalyses
Mikhail R. Gadelha [Thu, 16 Mar 2023 16:24:22 +0000 (13:24 -0300)]
[RISCV] Fix missing addi in test to validate lower inline asm m with offset
Wei Wang [Thu, 23 Feb 2023 00:58:23 +0000 (16:58 -0800)]
[Coroutines] Add remarks in CoroSplit and CoroElide passes
Add remarks to show frame size and alignment.
Reviewed By: ChuanqiXu
Differential Revision: https://reviews.llvm.org/D146175
Valentin Clement [Thu, 16 Mar 2023 16:16:18 +0000 (17:16 +0100)]
[flang] Do not try to rebox for assumed type
Just use conversion from assuemd type to assumed type
because a rebox needs more information that are not available
with assumed type.
Also use `isAssumedType` instead of `isBoxNone` since assumed
type can have sequence information.
Depends on D146207
Reviewed By: PeteSteinfeld
Differential Revision: https://reviews.llvm.org/D146209
Anshil Gandhi [Thu, 16 Mar 2023 16:09:42 +0000 (16:09 +0000)]
[AMDGPUUnifyDivergentExitNodes] Add NewPM support
Meanwhile, use UniformityAnalysis instead of LegacyDivergenceAnalysis to collect divergence info.
Reviewed By: arsenm, sameerds
Differential Revision: https://reviews.llvm.org/D141355
Mikhail R. Gadelha [Thu, 16 Mar 2023 16:06:50 +0000 (13:06 -0300)]
[RISCV] Added tests to validate lower inline asm m and A with offsets
Valentin Clement [Thu, 16 Mar 2023 16:10:26 +0000 (17:10 +0100)]
[flang] Embox argument for assumed type dummy argument to !fir.box<none>
When passing an argument to an assumed type dummy argument, embox
it directly to a !fir.box<none> box.
Reviewed By: PeteSteinfeld
Differential Revision: https://reviews.llvm.org/D146207
Nikita Popov [Thu, 16 Mar 2023 15:58:24 +0000 (16:58 +0100)]
[LICM] Add additional single thread promotion test (NFC)
Luke Lau [Thu, 9 Mar 2023 16:18:28 +0000 (16:18 +0000)]
[RISCV] Enable interleaved access vectorization
The loop vectorizer supports generating interleaved loads and stores via
shuffle patterns for fixed length vectors.
This enables it for RISC-V, since interleaved shuffle patterns can be
lowered to vlseg/vsseg in https://reviews.llvm.org/D145022
Reviewed By: reames
Differential Revision: https://reviews.llvm.org/D145155
Luke Lau [Wed, 15 Mar 2023 22:47:40 +0000 (22:47 +0000)]
[RISCV] Don't accidentally match deinterleave masks as interleaves
Consider a shuffle mask of <0, 2>:
This is one of two deinterleave masks to deinterleave a vector of 4
elements with factor 2.
Unfortunately, this is also technically an interleave mask, where
two subvectors of length 1 at indexes 0 and 2 will be interleaved.
This is because a mask can interleave non-contiguous subvectors:
e.g. <0, 6, 4, 1, 7, 5> on a vector of size 8:
```
<0 1 2 3 4 5 6 7> indices
^ ^ ^ ^ ^ ^
0 0 2 2 1 1 deinterleaved subvector
```
This means that deinterleaving shuffles can accidentally be costed as
interleaves.
And it's incorrect in the context of interleaves, because the
only interleave shuffles we model at the moment are single permutation
shuffles, i.e. we are interleaving the first vector below and ignoring
the second:
shufflevector <2 x i32> %v0, <2 x i32> poison, <2 x i32> <i32 0, i32 2>
A mask of <0, 2> interleaves across both vectors.
The fix here is to set NumInputElts correctly: We were setting it to
twice the mask length, i.e. using both input vectors. But in fact we're
actually only using the first vector here, and isInterleaveMask actually
already has logic to ensure that the mask indices stay within the bounds
of the input vectors.
This lacks a test case due to how we're unable to test deinterleave
shuffles (because they are length changing), but is covered in the tests
in D145155
Reviewed By: reames
Differential Revision: https://reviews.llvm.org/D146176
Alex Bradbury [Thu, 16 Mar 2023 15:02:53 +0000 (15:02 +0000)]
[RISCV][MC][NFC] Refactor RISCVAsmPrinter::PrintAsmMemoryOperand to use early return
Diego Caballero [Thu, 16 Mar 2023 04:14:16 +0000 (04:14 +0000)]
[mlir][Vector] Canonicalize create_mask(transpose)
When applying vector masking we may create a mask and then transpose it.
Transpositions are extremely expensive so this patch introduces a new
canonicalization pattern that remove the tranpose operation and create a
new transposed mask.
Differential Revision: https://reviews.llvm.org/D146193
Simon Pilgrim [Thu, 16 Mar 2023 14:32:58 +0000 (14:32 +0000)]
[TLI] getMemValueType - break apart if-else chain and use auto with dyn_cast (style). NFC.
More closely matches getValueType
Mahesh Ravishankar [Thu, 2 Mar 2023 00:33:14 +0000 (16:33 -0800)]
[mlir][TilingInterface] Modify `TilingInterface` methods to better return the state of the transformed IR.
Currently the `getTiledImplementation` and `generateResultTileValue`
return just `SmallVector<Operation *>` and `FailureOr<Value>`.
- For `getTiledImplementation` returning empty implies tiling wasnt
done. There is also an implicit assumption that the tiled operation
results correspond to the tiled values of the result of the original
operation. This cannot handle cases where the tiled implementation
might use multiple operations to compute the tiled value for the
results of the untiled operation. Sometimes, the tiled operation
might not directly give the tiled values, and might require casts,
etc to get a replacement.
- For `generateResultTileValue`, it is assumed that the op defining
the returned `Value` is the operation that represents the tiled
computation. Again presence of casts, etc violate this.
Instead make these methods return
```
struct TilingResult {
SmallVector<Operation *> tiledOps;
SmallVector<Value> tiledValues;
};
```
The `tiledOps` represent the operations generated that are relevant
for subsequent transformations. The `tiledValues` represent the tiled
values for the results of the original operation. This better
transmits the state of the transformed IR.
As a consequence the following methods also return `FailureOr<TilingResult>`
- `tensor::replaceExtractSliceWithTiledProducer`
- `tensor::bubbleUpPadSlice`
Differential Revision: https://reviews.llvm.org/D145133
Jakub Kuderski [Thu, 16 Mar 2023 14:16:14 +0000 (10:16 -0400)]
[mlir][vector] Add mask fold test for gather lowering
Check that `scf.if` checks are folded when the mask is all set / not
set.
This to address post-commit feedback for
https://reviews.llvm.org/D145942.
Reviewed By: dcaballe
Differential Revision: https://reviews.llvm.org/D146144
Joseph Huber [Thu, 16 Mar 2023 14:12:36 +0000 (09:12 -0500)]
[libc] Add missing dependencies to RISC-V startup implementation
Summary:
Just like the last patch, the threads and envrion dependencies were
missing. This lead to linker failures when building the tests.
Simon Pilgrim [Thu, 16 Mar 2023 14:06:23 +0000 (14:06 +0000)]
[TLI] Bring isZExtFree declarations together. NFC.
Don't spread them out over the header.
Tim Northover [Tue, 7 Feb 2023 12:26:29 +0000 (12:26 +0000)]
Recommit DwarfEHPrepare: insert extra unwind paths for stack protector to instrument
This is a mitigation patch for
https://bugs.chromium.org/p/llvm/issues/detail?id=30, where existing stack
protection is skipped if a function is returned through by an unwinder rather
than the normal call/return path. The recent patch D139254 added the ability to
instrument a visible unwind path, at least in the IR case (I'm working on the
SelectionDAG instrumentation too) but there are still invisible unwinds it
can't reach.
So this patch adds logic to DwarfEHPrepare that goes through a function,
converting any call that might throw into an invoke to a simple resume cleanup,
and adding cleanup clauses to existing landingpads that lack them. Obviously we
don't really want to do this if it's wasted effort, so I also exposed
requiresStackProtector from the actual StackProtector code to skip the extra
paths if they won't be used.
Changes:
* Move test to AArch64 directory as it relies on target presence.
* Re-add Dominator-tree maintenance. Accidentally cherry-picked wrong patch.
* Skip adding paths on Windows EH functions.
https://reviews.llvm.org/D143637