Sanjay Patel [Fri, 22 May 2020 15:37:58 +0000 (11:37 -0400)]
[InstCombine] (A + B) + B --> A + (B << 1)
This eliminates a use of 'B', so it can enable follow-on transforms
as well as improve analysis/codegen.
The PhaseOrdering test was added for D61726, and that shows
the limits of instcombine vs. real reassociation. We would
need to run some form of CSE to collapse that further.
The intermediate variable naming here is intentional because
there's a test at llvm/test/Bitcode/value-with-long-name.ll
that would break with the usual nameless value. I'm not sure
how to improve that test to be more robust.
The naming may also be helpful to debug regressions if this
change exposes weaknesses in the reassociation pass for example.
Sanjay Patel [Fri, 22 May 2020 14:20:43 +0000 (10:20 -0400)]
[InstCombine] add tests for adds with common operand; NFC
Denis Antrushin [Wed, 13 May 2020 17:55:07 +0000 (20:55 +0300)]
[SCEV] Constant fold MultExpr before applying depth limit.
Summary:
Users of SCEV reasonably assume that multiplication of two constant
SCEVs will in turn be constant.
However, that is not always the case:
First, we can get here with reached depth limit, and will create
MultExpr SCEV `C1 * C2` and cache it.
Then, we can get here with the same operands, but with small depth
level. But this time we will find existing MultExpr SCEV and return
it, instead of expected constant SCEV.
This patch changes getMultExpr to not apply depth limit to all constant
operands expression, allowing them to be folded.
Reviewers: reames, mkazantsev
Subscribers: hiraditya, javed.absar, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D79893
Xiangling_Liao [Fri, 22 May 2020 12:20:35 +0000 (08:20 -0400)]
[NFC][AIX] Remove spaces after the comma for '.csect' directive
To be consistent with other directives like '.comm', '.lcomm', we remove
the spaces after the comma for '.csect' on AIX.
Differential Revision: https://reviews.llvm.org/D80247
Matt Arsenault [Sun, 17 May 2020 14:51:22 +0000 (10:51 -0400)]
AMDGPU/GlobalISel: Fix masked control flow with fallthrough blocks
Unlike SelectionDAGBuilder, IRTranslator omits the unconditional
branch in fallthrough cases. Confusingly, the control flow pseudos
function in the opposite way the intrinsics are used, and the branch
targets always need to be swapped. We're inverting the target blocks,
so we need to figure out the old fallthrough block and insert a branch
to the original unconditional branch target.
Sanjay Patel [Fri, 22 May 2020 13:54:00 +0000 (09:54 -0400)]
[PhaseOrdering] regenerate test checks; NFC
Remove some redundant/unnecessary bits too.
Anh Tuyen Tran [Fri, 22 May 2020 13:30:56 +0000 (13:30 +0000)]
Title: [LV] Handle Fold-Tail of loops with vectorizarion factor equal to 1
Summary:
When handling loops whose VF is 1, fold-tail vectorization sets the
backedge taken count of the original loop with a vector of a single
element. This causes type-mismatch during instruction generartion.
The purpose of this patch is toto address the case of VF==1.
Reviewer: Ayal (Ayal Zaks), bmahjour (Bardia Mahjour), fhahn (Florian Hahn), gilr (Gil Rapaport), rengolin (Renato Golin)
Reviewed By: Ayal (Ayal Zaks), bmahjour (Bardia Mahjour), fhahn (Florian Hahn)
Subscribers: Ayal (Ayal Zaks), rkruppe (Hanna Kruppe), bmahjour (Bardia Mahjour), rogfer01 (Roger Ferrer Ibanez), vkmr (Vineet Kumar), bollu (Siddharth Bhat), hiraditya (Aditya Kumar), llvm-commits (Mailing List llvm-commits)
Tag: LLVM
Differential Revision: https://reviews.llvm.org/D79976
Simon Pilgrim [Fri, 22 May 2020 13:22:56 +0000 (14:22 +0100)]
TargetLowering.h - remove unnecessary includes. NFC.
Replace with forward declarations and move SizeOpts.h down to TargetLoweringBase.cpp
Simon Pilgrim [Fri, 22 May 2020 12:18:06 +0000 (13:18 +0100)]
[TargetLowering] Move TargetLoweringBase::isJumpTableRelative() implementation into TargetLoweringBase.cpp. NFC.
This will help with reducing header dependencies in TargetLowering.h in a future patch.
Louis Dionne [Fri, 22 May 2020 13:23:32 +0000 (09:23 -0400)]
[libc++] NFC: Inline array<T,N>::at methods inside the class
All other methods are defined in the class, so this increases consistency.
Sanjay Patel [Fri, 22 May 2020 13:12:06 +0000 (09:12 -0400)]
[SLP] fix verification check for valid IR
This is a fix for PR45965 - https://bugs.llvm.org/show_bug.cgi?id=45965 -
which was left out of D80106 because of a test failure.
SLP does its own mini-CSE after potentially creating redundant instructions,
so we need to wait for that to complete before running the verifier.
Otherwise, we will see a test failure for
test/Transforms/SLPVectorizer/X86/crash_vectorizeTree.ll (not changed here)
because a phi temporarily has identical but different incoming values for
the same incoming block.
A related, but independent, test that would have been altered here was
fixed with:
rG880df55
The test was escaping verification in SLP without this change because we
were not running verifyFunction() unless SLP actually changed the IR.
Differential Revision: https://reviews.llvm.org/D80401
Artem Dergachev [Fri, 22 May 2020 11:53:59 +0000 (14:53 +0300)]
[analyzer] LoopUnrolling: fix crash when a parameter is a loop counter.
When loop counter is a function parameter "isPossiblyEscaped" will not find
the variable declaration which lead to hitting "llvm_unreachable".
Parameters of reference type should be escaped like global variables;
otherwise treat them as unescaped.
Patch by Abbas Sabra!
Differential Revision: https://reviews.llvm.org/D80171
Sanjay Patel [Fri, 22 May 2020 12:39:22 +0000 (08:39 -0400)]
[SLP] fix test to have valid IR; NFC
This test was failing verification because the
metadata is ill-formed. This commit is split
from D80401 because it is an independent fix
(although the test would break with that change).
Nemanja Ivanovic [Fri, 22 May 2020 12:59:12 +0000 (07:59 -0500)]
[PowerPC] Treat 'Z' inline asm constraint as a true memory constraint
We currently emit incorrect codegen for this constraint because we set it as a
constraint that allows registers. This will cause the value to be copied to the
stack and that address to be passed as the address. This is not what we want.
Fixes: https://bugs.llvm.org/show_bug.cgi?id=42762
Differential revision: https://reviews.llvm.org/D77542
Nemanja Ivanovic [Fri, 22 May 2020 12:48:09 +0000 (07:48 -0500)]
[PowerPC] Add missing handling for half precision
The fix for PR39865 took care of some of the handling for half precision
but it missed a number of issues that still exist. This patch fixes the
remaining issues that cause crashes in the PPC back end.
Fixes: https://bugs.llvm.org/show_bug.cgi?id=45776
Differential revision: https://reviews.llvm.org/D79283
Marek Kurdej [Fri, 22 May 2020 12:33:16 +0000 (14:33 +0200)]
[CMake] Fix typos. NFC
Marek Kurdej [Fri, 22 May 2020 12:32:39 +0000 (14:32 +0200)]
[Target] Fix typos. NFC
Matt Arsenault [Thu, 21 May 2020 16:31:28 +0000 (12:31 -0400)]
InstCombine: Add constant folding/simplify for amdgcn.ldexp intrinsic
This really belongs in InstructionSimplify since it doesn't introduce
new instructions. Put it in instcombine to avoid increasing the number
of passes considering target intrinsics.
I also noticed that we seem to now be interpreting strictfp attributes
on call sites, so try to handle that.
Jon Roelofs [Fri, 22 May 2020 11:36:15 +0000 (05:36 -0600)]
Revert "[llvm][test] Add COM: directives before colon-less non-CHECKs in comments. NFC"
This reverts commit
183d6af081899973f00fc24aeafcfc32de732f02.
Revert pending further consensus building: https://reviews.llvm.org/D79963#2050521
Simon Pilgrim [Fri, 22 May 2020 11:28:26 +0000 (12:28 +0100)]
[AVR] Remove unsigned <= 0 checks. NFCI.
D77207 changed the bounds checks resulting in tests for positive unsigned values - dropping the superfluous check to fix gcc+Werror "error: comparison of unsigned expression >= 0 is always true [-Werror=type-limits]" warning.
Dmitry Preobrazhensky [Fri, 22 May 2020 10:54:59 +0000 (13:54 +0300)]
[AMDGPU][MC][GFX8+] Enabled clamp for v_mul_i32_i24_e64 and v_mul_u32_u24_e64
See bug 45925: https://bugs.llvm.org/show_bug.cgi?id=45925
Reviewers: arsenm, rampitec
Differential Revision: https://reviews.llvm.org/D80287
Simon Pilgrim [Fri, 22 May 2020 11:04:59 +0000 (12:04 +0100)]
Remove superfluous semicolon to stop Wpedantic warning. NFCI.
Simon Pilgrim [Fri, 22 May 2020 10:59:19 +0000 (11:59 +0100)]
Replace dyn_cast<> with isa<> to fix unused variable warning. NFCI.
QingShan Zhang [Fri, 22 May 2020 10:56:46 +0000 (10:56 +0000)]
[NFC][Test] Add test coverage for fsqrt on PowerPC
Valeriy Savchenko [Fri, 22 May 2020 09:51:25 +0000 (12:51 +0300)]
[analyzer] SumTimerInfo.py: Partially modernize
Differential Revision: https://reviews.llvm.org/D80427
Valeriy Savchenko [Fri, 22 May 2020 09:38:03 +0000 (12:38 +0300)]
[analyzer] SATestUpdateDiffs.py: Refactor and add type annotations
Differential Revision: https://reviews.llvm.org/D80426
Valeriy Savchenko [Fri, 22 May 2020 08:59:39 +0000 (11:59 +0300)]
[analyzer] SATestUtils.py: Refactor and add type annotations
Differential Revision: https://reviews.llvm.org/D80424
Valeriy Savchenko [Thu, 21 May 2020 15:28:36 +0000 (18:28 +0300)]
[analyzer] SATestBuild.py: Refactor and add type annotations
Summary:
SATest scripts should be more python-style than they are now.
This includes better architecture, type annotations, naming
convesions, and up-to-date language features. This commit starts
with two scripts SATestBuild and SATestAdd.
Differential Revision: https://reviews.llvm.org/D80423
Pavel Labath [Fri, 22 May 2020 10:41:46 +0000 (12:41 +0200)]
[lldb] Increase timeout in TestExitDuringExpression
200 microseconds is not enough time for any expression to execute
reliably. On linux, calling pthread_exit can result in call to dlopen,
which cannot complete in that time, particularly when running under a
debugger.
On linux, this test failed all the time, on macos, about two thirds of
runs were failing. This patch increases the timeout to 100ms, which is
enough to get it passing reliably on linux, though I wouldn't be
surprised if an even bigger timeout would be needed for remote test
runs.
Alexey Bader [Wed, 20 May 2020 19:22:30 +0000 (22:22 +0300)]
[SYCL] Prohibit arithmetic operations for incompatible pointers
Summary:
This change enables OpenCL diagnostics for the pointers annotated with
address space attribute SYCL mode.
Move `isAddressSpaceOverlapping` method from PointerType to QualType.
Reviewers: Anastasia, rjmccall
Reviewed By: rjmccall
Subscribers: rjmccall, jeroen.dobbelaere, Fznamznon, yaxunl, ebevhan, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D80317
Victor Campos [Fri, 22 May 2020 09:03:18 +0000 (10:03 +0100)]
Revert "[ARM] Improve codegen of volatile load/store of i64"
This reverts commit
8a12553223180246eeafaa0fa7bfa11e834d34b6.
A bug has been found when generating code for Thumb2. In some very
specific cases, the prologue/epilogue emitter generates erroneous stack
offsets for the new LDRD instructions that access the stack.
This bug does not seem to be caused by the reverted patch though. Likely
the latter has made an undiscovered issue emerge in the
prologue/epilogue emission pass. Nevertheless, this reversion is
necessary since it is blocking users of the ARM backend.
Simon Pilgrim [Fri, 22 May 2020 09:48:21 +0000 (10:48 +0100)]
StatepointLowering.h - remove unused includes. NFC.
Simon Pilgrim [Fri, 22 May 2020 09:47:12 +0000 (10:47 +0100)]
Fix msvc "switch statement contains 'default' but no 'case' labels" warning. NFC.
Stripped out the switch statement, but kept case labels as comments for future reference.
Simon Pilgrim [Thu, 21 May 2020 16:12:15 +0000 (17:12 +0100)]
MILexer.h/cpp - remove unused includes. NFC.
Remove duplicates in MILexer.cpp that are already included in MILexer.h.
Raphael Isemann [Fri, 22 May 2020 09:42:32 +0000 (11:42 +0200)]
[lldb] Enable C++14 when evaluating expressions in a C++14 frame
Summary:
Currently we never enable C++14 in the expression evaluator. This enables it when the language of the program is C++14.
It seems C++17 and so on isn't yet in any of the language enums (and the DWARF standard it seems), so C++17 support will be a follow up patch.
Reviewers: labath, JDevlieghere
Reviewed By: labath, JDevlieghere
Subscribers: aprantl
Differential Revision: https://reviews.llvm.org/D80308
Raphael Isemann [Fri, 22 May 2020 07:36:37 +0000 (09:36 +0200)]
[lldb][NFC] Overload raw_ostream operator << for ConstString
Summary: We are not doing this very often, but sometimes it's convenient when I can just << ConstStrings into llvm::errs() during testing.
Reviewers: labath, JDevlieghere
Reviewed By: labath, JDevlieghere
Subscribers: JDevlieghere
Differential Revision: https://reviews.llvm.org/D80310
Roman Lebedev [Thu, 21 May 2020 15:11:03 +0000 (18:11 +0300)]
[NFC] InstCombineNegator: use auto where type is obvious from the cast
Max Kazantsev [Fri, 22 May 2020 07:32:21 +0000 (14:32 +0700)]
[InstCombine] Sink pure instructions down to return and unreachable blocks
If the only user of `Instr` is in a return or unreachable block, we can
sink `Instr` to the`User` safely (unless it reads/writes memory).
Return or unreachable blocks are guaranteed to execute zero
or one time, and `Instr` always dominates `User`, so they either will
be executed together (execution of `User` always implies execution
of `Instr`) or not executed at all.
Differential Revision: https://reviews.llvm.org/D80120
Reviewed By: asbirlea, jdoerfert
Jonas Devlieghere [Fri, 22 May 2020 06:22:16 +0000 (23:22 -0700)]
[lldb/REPL] Fix unhandled switch case
Fix warning: enumeration value 'eExpressionThreadVanished' not handled
in switch [-Wswitch]
LLVM GN Syncbot [Fri, 22 May 2020 04:58:19 +0000 (04:58 +0000)]
[gn build] Port
2e40cf06df0
Craig Topper [Fri, 22 May 2020 01:16:02 +0000 (18:16 -0700)]
[X86] Allow avx512vp2intersect to be used with __builtin_cpu_supports.
compiler-rt and trunk libgcc support this now.
Craig Topper [Fri, 22 May 2020 01:12:21 +0000 (18:12 -0700)]
[X86] Add avx512vp2intersect feature to compiler-rt's feature detection to match libgcc.
Lang Hames [Fri, 22 May 2020 04:44:00 +0000 (21:44 -0700)]
[JITLink] Initial implementation of ELF / x86-64 support for JITLink.
This initial implementation supports section and symbol parsing, but no
relocation support. It enables JITLink to link and execute ELF relocatable
objects that do not require relocations.
Patch by Jared Wyles. Thanks Jared!
Differential Revision: https://reviews.llvm.org/D79832
ISHIGURO, Hiroshi [Fri, 22 May 2020 02:34:49 +0000 (11:34 +0900)]
[OPENMP] Fix mixture of omp and clang pragmas
Fixes PR45753
When a program that contains a loop to which both `omp parallel for`
pragma and `clang loop` pragma are associated is compiled with the
-fopenmp option, `clang loop` pragma did not take effect. The example
below should not be vectorized by the `clang loop` pragma but it was
actually vectorized. The cause is that `llvm.loop.vectorize.width`
was not output to the IR when -fopenmp is specified.
The fix attaches attributes if they exist for the loop.
[example.c]
```
int a[100], b[100];
void foo() {
#pragma omp parallel for
#pragma clang loop vectorize(disable)
for (int i = 0; i < 100; i++)
a[i] += b[i] * i;
}
```
[compile]
```
$ clang -O2 -fopenmp example.c -c -Rpass=vect
example.c:3:11: remark: vectorized loop (vectorization width: 4, interleaved count: 2) [-Rpass=loop-vectorize]
#pragma omp parallel for
^
```
[IR with -fopenmp]
```
$ clang -O2 exmaple.c -S -emit-llvm -mllvm -disable-llvm-optzns -o - -fopenmp | grep 'vectorize\.width'
```
[IR with -fno-openmp]
```
$ clang -O2 example.c -S -emit-llvm -mllvm -disable-llvm-optzns -o - -fno-openmp | grep 'vectorize\.width'
!7 = !{!"llvm.loop.vectorize.width", i32 1}
```
Differential Revision: https://reviews.llvm.org/D79921
Jinsong Ji [Fri, 22 May 2020 03:10:34 +0000 (03:10 +0000)]
[docs][llvm-extract] Add missing alias/bb options
llvm-extract get serveral new options, but we forgot to update doc.
This patch update the doc.
Reviewed By: volkan
Differential Revision: https://reviews.llvm.org/D80413
Jonas Devlieghere [Fri, 22 May 2020 03:36:39 +0000 (20:36 -0700)]
[lldb/Reproducers] Skip test that changes the source file while debugging
The VFS is a snapshot and cannot capture changes to the file system.
Jonas Devlieghere [Fri, 22 May 2020 03:34:39 +0000 (20:34 -0700)]
[lldb/Test] Fix replay with TestSetWatchpoint.py
The reproducers only track the creation of objects and not their
destruction. Therefore it keeps all objects alive indefinitely.
Nico Weber [Fri, 22 May 2020 03:04:41 +0000 (23:04 -0400)]
[gn build] (manually) port
49a4f3f7d88
Jessica Paquette [Fri, 22 May 2020 01:05:37 +0000 (18:05 -0700)]
[AArch64][GlobalISel] Add a post-legalizer combiner with a very simple combine.
(This patch is by Jessica, I'm just committing it on her behalf because I need
a post-legalizer combiner for something else).
This supersedes D77250, which did equivalent work in the selector. This can be
done pre-legalization or post-legalization. Post-legalization is more likely to
hit, since G_IMPLICIT_DEFs tend to appear during legalization. There's no reason
to not do it pre-legalization though-- if it can be caught earlier, great.
(I also think that it might be worth reimplementing D78769 using a
target-specific post-legalization combine too after thinking about it for a
while.)
Differential Revision: https://reviews.llvm.org/D78852
Paula Toth [Fri, 22 May 2020 00:39:03 +0000 (17:39 -0700)]
[libc] Make clang-tidy use host compiler's resource dir.
Summary: When building llvm-libc with linting enabled, clang-tidy would use the resource dir of the monorepo rather then the host compiler's resource dir. This presented issues when including headers from the host compiler e.g. for sanitizers. Therefore this patch explicitly tells clang-tidy to use the host compiler's resource dir.
Reviewers: sivachandra
Reviewed By: sivachandra
Subscribers: mgorny, tschuett, ecnelises, libc-commits
Tags: #libc-project
Differential Revision: https://reviews.llvm.org/D80265
Jim Ingham [Fri, 22 May 2020 01:38:49 +0000 (18:38 -0700)]
Maybe I need ENABLE_THREADS in the Makefile.
Jim Ingham [Fri, 22 May 2020 01:26:01 +0000 (18:26 -0700)]
This very simple .c file is failing on the Debian bot wit the error
undefined reference to pthread_create
I skipped the test till I can figure out why this didn't build.
Dominic Chen [Thu, 21 May 2020 00:05:40 +0000 (20:05 -0400)]
[llvm-extract] Fix basic block extraction by delaying search until the function is materialized
Summary: When I try to extract a basic block using llvm-extract, it erroneously reports that the named basic block doesn't exist. After digging into the code, it appears to be iterating over an unmaterialized function, which will fail to match any basic blocks.
Reviewers: volkan, qcolombet
Subscribers: llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D80346
Jim Ingham [Thu, 21 May 2020 01:00:56 +0000 (18:00 -0700)]
Handle the case where a thread exits while we are running a function on it.
Tony [Thu, 21 May 2020 21:20:06 +0000 (17:20 -0400)]
[AMDGPU] Correct formatting typos in documentation
Summary:
- Correct missing space in some "note" and "TODO" directives in
AMDGPUUsage.rst
- Correct warning for heading underline being too short in
BitCodeFormat.rst
Subscribers: kzhuravl, jvesely, wdng, nhaehnle, yaxunl, dstuttard, tpr, kerbowa, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D80407
Jonas Devlieghere [Fri, 22 May 2020 00:31:51 +0000 (17:31 -0700)]
[lldb/Test] Cleanup TestSymbolContext.py
Remove commented out code, fix the indentation and always use the full
path to the executable. The latter is necessary for the test to pass
from reproducer replay.
Heejin Ahn [Thu, 21 May 2020 06:02:41 +0000 (23:02 -0700)]
[WebAssembly] Warn on exception spec only when Wasm EH is used
Summary:
In D80061 we added warning for exception specifications with types (such
as `throw(int)`), but it was enabled every time the target was wasm,
which means it warned (and ignored) exception specifications even if
wasm EH was not used. This fixes it and we only have the warning when we
enable `-fwasm-exceptions`.
Reviewers: dschuff
Subscribers: sbc100, jgravelle-google, sunfish, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D80362
Zequan Wu [Thu, 21 May 2020 23:44:30 +0000 (16:44 -0700)]
[clang] Add nomerge function attribute to clang
Differential Revision: https://reviews.llvm.org/D79121
Jonas Devlieghere [Thu, 21 May 2020 21:52:12 +0000 (14:52 -0700)]
[lldb/Reproducers] Skip lldb-vscode category when lldb-run-with-repro is set.
This skips all the lldb-vscode tests when running the test suite with
reproducers.
LLVM GN Syncbot [Thu, 21 May 2020 23:43:28 +0000 (23:43 +0000)]
[gn build] Port
54e91a3c701
Jan Korous [Thu, 21 May 2020 21:37:41 +0000 (14:37 -0700)]
Reland "[Analyzer][WebKit] RefCntblBaseVirtualDtorChecker"
This reverts commit
1108f5c737dbdab0277874a7e5b237491839c43a.
Adrian Prantl [Thu, 21 May 2020 23:22:01 +0000 (16:22 -0700)]
Rename FunctionOptimizationWarning to the more generic FrameSelectedCallback (NFC)
Zequan Wu [Thu, 21 May 2020 23:12:03 +0000 (16:12 -0700)]
Revert "Add nomerge function attribute to clang"
This reverts commit
307e85395485e1eff9533b2d7952b16f33ceae38.
Vedant Kumar [Wed, 20 May 2020 01:03:22 +0000 (18:03 -0700)]
[Instruction] Set metadata uses to undef on deletion
Summary:
Replace any extant metadata uses of a dying instruction with undef to
preserve debug info accuracy. Some alternatives include:
- Treat Instruction like any other Value, and point its extant metadata
uses to an empty ValueAsMetadata node. This makes extant dbg.value uses
trivially dead (i.e. fair game for deletion in many passes), leading to
stale dbg.values being in effect for too long.
- Call salvageDebugInfoOrMarkUndef. Not needed to make instruction removal
correct. OTOH results in wasted work in some common cases (e.g. when all
instructions in a BasicBlock are deleted).
This came up while discussing some basic cases in
https://reviews.llvm.org/D80052.
Reviewers: jmorse, TWeaver, aprantl, dexonsmith, jdoerfert
Subscribers: jholewinski, qcolombet, hiraditya, jfb, sstefan1, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D80264
Alexey Lapshin [Wed, 20 May 2020 18:45:39 +0000 (21:45 +0300)]
[AARCH64][NEON] Allow to sink operands of aarch64_neon_pmull64.
Summary:
This patch fixes a problem when pmull2 instruction is not
generated for vmull_high_p64 intrinsic.
ISel has a pattern for int_aarch64_neon_pmull64 intrinsic to generate
PMULL2 instruction. That pattern assumes that extraction operations
are located in the same basic block. We need to sink them
if they are not. Handle operands of int_aarch64_neon_pmull64
into AArch64TargetLowering::shouldSinkOperands.
Reviewed by: efriedma
Differential Revision: https://reviews.llvm.org/D80320
Zequan Wu [Fri, 8 May 2020 18:36:34 +0000 (11:36 -0700)]
Add nomerge function attribute to clang
Craig Topper [Thu, 21 May 2020 22:23:00 +0000 (15:23 -0700)]
[Target] Use Align in TargetLoweringObjectFile::getSectionForConstant.
Differential Revision: https://reviews.llvm.org/D80363
Arthur Eubanks [Wed, 13 May 2020 23:33:09 +0000 (16:33 -0700)]
Don't jump to landing pads in Control Flow Optimizer
Summary: Likely fixes https://bugs.llvm.org/show_bug.cgi?id=45858.
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D80047
Jinsong Ji [Thu, 21 May 2020 22:06:10 +0000 (22:06 +0000)]
[docs] Fix buildbot failures
Buildbot has been failing since
http://lab.llvm.org:8011/builders/llvm-sphinx-docs/builds/44711
This patch fix the minor issues that cause warnings.
Ben Dunbobbin [Wed, 20 May 2020 10:36:10 +0000 (11:36 +0100)]
[PS4] Enable relaxed relocations by default
PS4 supports these relocations types; so, we want to enable them by default.
Differential Revision: https://reviews.llvm.org/D79980
Dominic Chen [Thu, 21 May 2020 01:11:23 +0000 (21:11 -0400)]
llvm-diff: Avoid crash with complex expressions
Summary: With complex IR, the difference engine can enter an inconsistent state where the instruction and block difference engines return different results. Prevent a crash by refusing to pop beyond the end of the vector.
Reviewers: rjmccall
Subscribers: llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D80351
Stephen Kelly [Fri, 10 Jan 2020 19:19:23 +0000 (19:19 +0000)]
Make the ExprMutationAnalyzer explicit about how it traverses the AST
Reviewers: aaron.ballman
Subscribers: baloghadamsoftware, a.sidorin, Szelethus, donat.nagy, dkrupp, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D72532
Stephen Kelly [Tue, 12 Nov 2019 13:29:40 +0000 (13:29 +0000)]
Set traversal explicitly where needed in tests
Reviewers: aaron.ballman, shafik
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D72531
Stephen Kelly [Tue, 12 Nov 2019 15:15:56 +0000 (15:15 +0000)]
Set traversal explicitly where needed in clang-tidy
Reviewers: aaron.ballman
Subscribers: nemanjai, kbarton, arphaman, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D72530
Tim Renouf [Sat, 25 Apr 2020 11:14:58 +0000 (12:14 +0100)]
[AMDGPU] Fixed incorrect PAL metadata register naming
This only affects assembly and -filetype=asm codegen of PAL metadata.
Differential Revision: https://reviews.llvm.org/D78860
Change-Id: I7b822e1917bf7b403486820d31afc483be207652
Tim Renouf [Mon, 18 May 2020 13:13:57 +0000 (14:13 +0100)]
[MsgPack] Added convenience assignment to MsgPackDocument
This commit increases the convenience of using the MsgPackDocument API,
especially when creating a document for writing out.
It adds direct assignment of bool, integer and string types to a
DocNode, as long as that DocNode is already inside a document, e.g. the
result of a map lookup. It also adds map lookup given an integer type
(it already had that for string).
So, to assign a string to a map element whose key is an int, you can
now write
MyMap[42] = "towel";
instead of
MyMap[MyMap.getDocument()->getNode(42)] =
MyMap.getDocument()->getNode("towel");
Also added MapDocNode::erase methods.
Differential Revision: https://reviews.llvm.org/D80121
Change-Id: I17301fa15bb9802231c52542798af5b54beb583e
Jean-Michel Gorius [Thu, 21 May 2020 14:30:48 +0000 (16:30 +0200)]
[CodeGen] Add support for multiple memory operands in MachineInstr::mayAlias
Summary:
To support all targets, the mayAlias member function needs to support instructions with multiple operands.
This revision also changes the order of the emitted instructions in some test cases.
Reviewers: efriedma, hfinkel, craig.topper, dmgreen
Reviewed By: efriedma
Subscribers: MatzeB, dmgreen, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D80161
Stanislav Mekhanoshin [Wed, 20 May 2020 23:24:06 +0000 (16:24 -0700)]
[AMDGPU] Promote alloca to vector in opt
Promote alloca to vector before SROA and loop unroll. If we manage
to eliminate allocas before unroll we may choose to unroll less.
Differential Revision: https://reviews.llvm.org/D80386
Sean Silva [Thu, 21 May 2020 20:10:38 +0000 (13:10 -0700)]
[NFC] Make assertion more informative.
This assert just caught me, and this improved message would have saved
me some time.
Eli Friedman [Thu, 21 May 2020 00:52:36 +0000 (17:52 -0700)]
[AArch64][SVE] Fill out missing unpredicated load/store patterns.
The set of patterns for unpredicated load/store was incomplete: it only
included non-extending stores. Fill out the remaining patterns for
extending stores, and add the corresponding support to frame offset
lowering.
Differential Revision: https://reviews.llvm.org/D80349
Tim Renouf [Thu, 21 May 2020 19:31:24 +0000 (20:31 +0100)]
[MsgPack] MsgPackDocument::readFromBlob now merges
The readFromBlob method can now be used to read MsgPack into a Document
that already contains something, merging the two. There is a new Merger
argument to readFromBlob, a callback function to resolve conflicts.
Differential Revision: https://reviews.llvm.org/D79671
Change-Id: Icf3e959217fe33cd907a41516c0386aef2847c0c
Hendrik Greving [Thu, 21 May 2020 20:17:18 +0000 (13:17 -0700)]
[ModuloSchedule] Add missing comma.
This is a test commit as per Chris to verify commit access.
Thanks!
Yitzhak Mandelbaum [Thu, 21 May 2020 19:49:23 +0000 (15:49 -0400)]
[libTooling][NFC]Fix typo in comment in RangeSelectors
Fixes transposed names in comment.
Marcello Maggioni [Wed, 29 Apr 2020 01:25:58 +0000 (01:25 +0000)]
[SelectionDAG] Add the option of disabling generic combines.
Summary:
For some targets generic combines don't really do much and they
consume a disproportionate amount of time.
There's not really a mechanism in SDISel to tactically disable
combines, but we can have a switch to disable all of them and
let the targets just implement what they specifically need.
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D79112
Marek Kurdej [Thu, 21 May 2020 19:48:51 +0000 (21:48 +0200)]
Reland [libc++] [LWG3321] Mark "year_month_day_last::day() specification does not cover !ok() values" issue as "Nothing to do", but add assertion.
Summary:
This LWG issue states that the result of `year_month_day_last::day()` is implementation defined if `ok()` is `false`.
However, from user perspective, calling `day()` in this situation will lead to a (possibly difficult to find) crash.
Hence, I have added an assertion to warn user at least when assertions are enabled.
Reviewed By: ldionne, #libc
Differential Revision: https://reviews.llvm.org/D70346
Nico Weber [Thu, 21 May 2020 19:49:46 +0000 (15:49 -0400)]
Revert "[Analyzer][WebKit] RefCntblBaseVirtualDtorChecker"
This reverts commit
f7c7e8a523f56b0ed1b14c0756ba4e5d1ccb48d2.
Breaks build everywhere.
Jonas Devlieghere [Thu, 21 May 2020 19:27:39 +0000 (12:27 -0700)]
[dsymutil] Add llvm_unreachable to silence warning
Fixes warning: control reaches end of non-void function [-Wreturn-type]
Thomas Raoux [Thu, 21 May 2020 18:59:31 +0000 (11:59 -0700)]
[mlir][spirv] Enable composite instructions for cooperative matrix type.
Enable inset/extract/construct composite ops as well as access chain for
cooperative matrix. ConstantComposite requires more change and will be done in
a separate patch. Also fix the getNumElements function for coopMatrix per
feedback from Jeff Bolz. The number of element is implementation dependent so
it cannot be known at compile time.
Differential Revision: https://reviews.llvm.org/D80321
Stanislav Mekhanoshin [Thu, 21 May 2020 18:58:35 +0000 (11:58 -0700)]
[AMDGPU] Added opt pipeline test. NFC.
Thomas Raoux [Thu, 21 May 2020 18:35:32 +0000 (11:35 -0700)]
[mlir][spirv] Add remaining cooperative matrix instructions
Adds support for cooperative matrix support for arithmetic and cast
instructions. It also adds cooperative matrix store, muladd and matrixlength
instructions which are part of the extension.
Differential Revision: https://reviews.llvm.org/D80181
Jan Korous [Thu, 21 May 2020 18:52:12 +0000 (11:52 -0700)]
[Analyzer][WebKit] RefCntblBaseVirtualDtorChecker
Differential Revision: https://reviews.llvm.org/D77177
Louis Dionne [Thu, 21 May 2020 18:33:32 +0000 (14:33 -0400)]
[libc++] Link back-deployment tests against the latest libc++ and libc++abi
Instead of linking the tests against a library in some version of the
SDK, always link against the latest library, but still run against the
specified back-deployment target dylib.
This makes more sense since what we're really trying to test is that
the current library can be used to produce binaries that run on some
deployment target -- not that linking against the library in some
previous SDK makes that possible.
This solves an additional issue that when linking against a system dylib,
the -rpath argument given to the tests is ignored because the install_name
of the system library we link against is absolute.
rdar://
63241847
Jonas Devlieghere [Thu, 21 May 2020 18:41:20 +0000 (11:41 -0700)]
[dsymutil] Fix conversion between unique_ptr and Expected
Reproducer.cpp:70:12: error: could not convert ‘Repro’ from
‘std::unique_ptr<llvm::dsymutil::ReproducerGenerate,
std::default_delete<llvm::dsymutil::ReproducerGenerate> >’ to
‘llvm::Expected<std::unique_ptr<llvm::dsymutil::Reproducer> >’
Hiroshi Yamauchi [Wed, 13 May 2020 19:14:33 +0000 (12:14 -0700)]
[IR] Make Module::setProfileSummary to replace an existing ProfileSummary flag.
Summary:
Module::setProfileSummary currently calls addModuelFlag. This prevents from
updating the ProfileSummary metadata in the module and results in a second
ProfileSummary added instead of replacing an existing one. I don't think this is
the expected behavior. It prevents updating the ProfileSummary and it does not
make sense to have more than one. To address this, add Module::setModuleFlag and
use it from setProfileSummary.
Reviewers: davidxl
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D79902
Thomas Lively [Thu, 21 May 2020 18:33:25 +0000 (11:33 -0700)]
[lld][WebAssembly] Do not emit initialization for .bss segments
Summary:
This patch fixes a bug where initialization code for .bss segments was
emitted in the memory initialization function even though the .bss
segments were discounted in the datacount section and omitted in the
data section. This was producing invalid binaries due to out-of-bounds
segment indices on the memory.init and data.drop instructions that
were trying to operate on the nonexistent .bss segments.
Reviewers: sbc100
Subscribers: dschuff, jgravelle-google, aheejin, sunfish, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D80354
LLVM GN Syncbot [Thu, 21 May 2020 18:20:26 +0000 (18:20 +0000)]
[gn build] Port
92fd3971e0d
Jonas Devlieghere [Thu, 21 May 2020 18:19:57 +0000 (11:19 -0700)]
[dsymutil] Fix include-style
Jonas Devlieghere [Thu, 21 May 2020 17:57:53 +0000 (10:57 -0700)]
[dsymutil] Add reproducers to dsymutil
Add support for generating a dsymutil reproducer. The result is a folder
containing all the object files for linking.
When --gen-reproducer is passed, dsymutil uses a FileCollectorFileSystem
which keeps track of all the files used by dsymutil. These files are
copied into a temporary directory when dsymutil exists.
When this path is passed to --use-reproducer, dsymutil uses a
RedirectingFileSystem that will use the files from the reproducer
directory instead of the actual paths. This means you don't need to mess
with the OSO path prefix.
Differential revision: https://reviews.llvm.org/D79398
Simon Atanasyan [Thu, 21 May 2020 13:14:24 +0000 (16:14 +0300)]
[mips] Reorganize check directives in the test. NFC
Adrian Prantl [Thu, 21 May 2020 17:42:26 +0000 (10:42 -0700)]
Move decorator to the correct function.