Alex Bradbury [Tue, 8 Aug 2017 16:20:39 +0000 (16:20 +0000)]
[RISCV] Fix warning about unused getSubtargetFeatureName()
llvm-svn: 310375
Roman Gareev [Tue, 8 Aug 2017 16:15:33 +0000 (16:15 +0000)]
Do not use isl_set_project_out to get all loop prefixes
Currently, only convex isolation sets can be efficiently processed by isl.
Consequently, as a temporary solution, we use a different algorithm for partial
tile isolation that helps to build convex isolation sets in some cases.
Reviewed-by: Tobias Grosser <tobias@grosser.es>
Differential Revision: https://reviews.llvm.org/D36278
llvm-svn: 310374
Nuno Lopes [Tue, 8 Aug 2017 16:13:24 +0000 (16:13 +0000)]
BasicAA: aliasGEP shouldn't get a PartialAlias response here
add an assert() to ensure that's the case (as I'm not convinced it won't happen)
llvm-svn: 310373
Simon Pilgrim [Tue, 8 Aug 2017 16:10:33 +0000 (16:10 +0000)]
[DAGCombiner] simplifyShuffleMask - handle UNDEF inputs from shuffles as well as BUILD_VECTOR
Minor extension to D36393
llvm-svn: 310372
Gabor Horvath [Tue, 8 Aug 2017 15:33:48 +0000 (15:33 +0000)]
[clang-tidy] Add new readability non-idiomatic static access check
Patch by: Lilla Barancsuk
Differential Revision: https://reviews.llvm.org/D35937
llvm-svn: 310371
Kamil Rytarowski [Tue, 8 Aug 2017 15:25:26 +0000 (15:25 +0000)]
Enable COMPILER_RT_HAS_SANITIZER_COMMON on NetBSD
Summary:
Temporarily keep disabled COMPILER_RT_HAS_ASAN on NetBSD.
Part of the code inspired by the original work on libsanitizer in GCC 5.4 by Christos Zoulas.
Sponsored by <The NetBSD Foundation>
Reviewers: joerg, filcab, kcc, vitalybuka
Reviewed By: vitalybuka
Subscribers: srhines, mgorny, #sanitizers, llvm-commits
Tags: #sanitizers
Differential Revision: https://reviews.llvm.org/D36312
llvm-svn: 310370
Tobias Grosser [Tue, 8 Aug 2017 15:14:46 +0000 (15:14 +0000)]
[RegisterPasses] Run polly-simplify also right after scop modeling
This allows us to get rid of stores that are overwritten within the very same
basic block, without ever being read beforehand. This simplification is
necessary for delicm to run on pb4's correlation.
llvm-svn: 310369
Gheorghe-Teodor Bercea [Tue, 8 Aug 2017 15:13:07 +0000 (15:13 +0000)]
Non-functional change. Fix test for D29654.
llvm-svn: 310368
Martin Probst [Tue, 8 Aug 2017 15:00:58 +0000 (15:00 +0000)]
clang-format: [JS] fix union type spacing in object & array types.
Summary:
Previously, clang-format would insert whitespace in union types nested in object
and array types, as it wouldn't recognize those as a type operator:
const x: {foo: number | null};
const x: [number | null];
While this is correct for actual binary operators, clang-format should not
insert whitespace into union and intersection types to mark those:
const x: {foo: number|null};
const x: [number|null];
This change propagates that the context is not an expression by inspecting
the preceding token and marking as non-expression if it was a type colon.
Reviewers: djasper
Subscribers: klimek
Differential Revision: https://reviews.llvm.org/D36136
llvm-svn: 310367
Alexander Kornienko [Tue, 8 Aug 2017 14:53:52 +0000 (14:53 +0000)]
[clang-tidy] 'implicit cast' -> 'implicit conversion'
Summary:
This patch renames checks, check options and changes messages to use correct
term "implicit conversion" instead of "implicit cast" (which has been in use in
Clang AST since ~10 years, but it's still technically incorrect w.r.t. C++
standard).
* performance-implicit-cast-in-loop -> performance-implicit-conversion-in-loop
* readability-implicit-bool-cast -> readability-implicit-bool-conversion
- readability-implicit-bool-cast.AllowConditionalIntegerCasts ->
readability-implicit-bool-conversion.AllowIntegerConditions
- readability-implicit-bool-cast.AllowConditionalPointerCasts ->
readability-implicit-bool-conversion.AllowPointerConditions
Reviewers: hokein, jdennett
Reviewed By: hokein
Subscribers: mgorny, JDevlieghere, xazax.hun, cfe-commits
Differential Revision: https://reviews.llvm.org/D36456
llvm-svn: 310366
Martin Probst [Tue, 8 Aug 2017 14:52:42 +0000 (14:52 +0000)]
clang-format: [JS] handle single lines comments ending in `\\`.
Summary:
Previously, clang-format would consider the following code line to be part of
the comment and incorrectly format the rest of the file.
Reviewers: djasper
Subscribers: klimek, cfe-commits
Differential Revision: https://reviews.llvm.org/D36159
llvm-svn: 310365
Alexey Bataev [Tue, 8 Aug 2017 14:44:43 +0000 (14:44 +0000)]
Revert "[OPENMP][DEBUG] Set proper address space info if required by target."
This reverts commit r310360.
llvm-svn: 310364
Alex Bradbury [Tue, 8 Aug 2017 14:43:36 +0000 (14:43 +0000)]
[RISCV] Add basic RISCVAsmParser (missing files)
This commit adds the files missing from rL310361. Apologies for the noise.
Differential Revision: https://reviews.llvm.org/D23563
llvm-svn: 310363
Gheorghe-Teodor Bercea [Tue, 8 Aug 2017 14:33:05 +0000 (14:33 +0000)]
[OpenMP] OpenMP device offloading code generation produces a cubin file which is then integrated in the host binary using the host linker.
Diff: D29654
llvm-svn: 310362
Alex Bradbury [Tue, 8 Aug 2017 14:32:35 +0000 (14:32 +0000)]
[RISCV] Add basic RISCVAsmParser
This doesn't yet support parsing things like %pcrel_hi(foo), but will handle
basic instructions with register or immediate operands.
Differential Revision: https://reviews.llvm.org/D23563
llvm-svn: 310361
Alexey Bataev [Tue, 8 Aug 2017 14:25:14 +0000 (14:25 +0000)]
[OPENMP][DEBUG] Set proper address space info if required by target.
Arguments, passed to the outlined function, must have correct address
space info for proper Debug info support. Patch sets global address
space for arguments that are mapped and passed by reference.
Also, cuda-gdb does not handle reference types correctly, so reference
arguments are represented as pointers.
llvm-svn: 310360
Nikolai Bozhenov [Tue, 8 Aug 2017 14:13:50 +0000 (14:13 +0000)]
[libclang] Fix PR34055 (incompatible update of clang-c/Index.h)
Fixes a regression introduced by r308218.
llvm-svn: 310359
Nemanja Ivanovic [Tue, 8 Aug 2017 13:52:45 +0000 (13:52 +0000)]
[PowerPC] Don't crash on larger splats achieved through 1-byte splats
We've implemented a 1-byte splat using XXSPLTISB on P9. However, LLVM will
produce a 1-byte splat even for wider element BUILD_VECTOR nodes. This patch
prevents crashing in that situation.
Differential Revision: https://reviews.llvm.org/D35650
llvm-svn: 310358
Daniel Sanders [Tue, 8 Aug 2017 13:21:26 +0000 (13:21 +0000)]
[globalisel][tablegen] Remove unnecessary ; to satisfy ubuntu-gcc7.1-werror.
llvm-svn: 310357
Nemanja Ivanovic [Tue, 8 Aug 2017 12:41:56 +0000 (12:41 +0000)]
Appease compilers that have the -Wcovered-switch-default switch.
llvm-svn: 310356
Siddharth Bhat [Tue, 8 Aug 2017 12:26:37 +0000 (12:26 +0000)]
[NFC] [PPCGCodeGen] Add missing REQUIRES: pollyacc line.
llvm-svn: 310354
Siddharth Bhat [Tue, 8 Aug 2017 12:26:32 +0000 (12:26 +0000)]
[ScopInfo] [NFC] Typo fix.
"to conservative" -> "too conservative".
llvm-svn: 310353
Amjad Aboud [Tue, 8 Aug 2017 12:17:56 +0000 (12:17 +0000)]
[X86] Improved X86::CMOV to Branch heuristic.
Resolved PR33954.
This patch contains two more constraints that aim to reduce the noise cases where we convert CMOV into branch for small gain, and end up spending more cycles due to overhead.
Differential Revision: https://reviews.llvm.org/D36081
llvm-svn: 310352
Kamil Rytarowski [Tue, 8 Aug 2017 12:10:08 +0000 (12:10 +0000)]
Reuse interception_linux for NetBSD
Summary:
Part of the code inspired by the original work on libsanitizer in GCC 5.4 by Christos Zoulas.
Sponsored by <The NetBSD Foundation>
Reviewers: joerg, kcc, vitalybuka, filcab
Reviewed By: vitalybuka
Subscribers: llvm-commits, #sanitizers
Tags: #sanitizers
Differential Revision: https://reviews.llvm.org/D36321
llvm-svn: 310351
Siddharth Bhat [Tue, 8 Aug 2017 12:00:59 +0000 (12:00 +0000)]
[Polly] [PPCGCodeGeneration] Handle failing of invariant load hoisting gracefully.
To do this, we replicate what `CodeGeneration` does. We expose
`markNodeUnreachable` from `CodeGeneration` to `PPCGCodeGeneration`.
Differential Revision: https://reviews.llvm.org/D36457
llvm-svn: 310350
Kamil Rytarowski [Tue, 8 Aug 2017 11:40:15 +0000 (11:40 +0000)]
Define OFF_T as 64-bit integer on NetBSD
Summary:
All 32 and 64 bit NetBSD platforms define off_t as 64-bit integer.
Part of the code inspired by the original work on libsanitizer in GCC 5.4 by Christos Zoulas.
Sponsored by <The NetBSD Foundation>
Reviewers: joerg, filcab, kcc, vitalybuka
Reviewed By: vitalybuka
Subscribers: emaste, kubamracek, llvm-commits
Tags: #sanitizers
Differential Revision: https://reviews.llvm.org/D35553
llvm-svn: 310349
Michael Kruse [Tue, 8 Aug 2017 11:27:12 +0000 (11:27 +0000)]
[DeLICM] Properly handle PHI writes becoming empty partial writes.
It is possible that partial writes are empty (write is never executed).
In this case, when in PHINode's incoming edge is never taken such that
the incoming write becomes an empty partial write, if enabled. The
issue is that when converting the union_map to an map, it's space
cannot be derived from the union_map itself. Rather, we need to
determine its space independently.
This fixes test-suite's MultiSource/Benchmarks/ASC_Sequoia/CrystalMk.
llvm-svn: 310348
Alex Lorenz [Tue, 8 Aug 2017 11:22:21 +0000 (11:22 +0000)]
Darwin's toolchain should be initialized before openmp offloading
is processed
This fixes an 'openmp-offload.c' test failure introduced by r310263.
llvm-svn: 310347
Nemanja Ivanovic [Tue, 8 Aug 2017 11:20:44 +0000 (11:20 +0000)]
[PowerPC] Eliminate compares - add i32 sext/zext handling for SETLE/SETGE
Adds handling for SETLE/SETGE comparisons on i32 values. Furthermore, it adds
the handling for the special case where RHS == 0.
Differential Revision: https://reviews.llvm.org/D34048
llvm-svn: 310346
Alex Lorenz [Tue, 8 Aug 2017 11:20:17 +0000 (11:20 +0000)]
Revert r310291, r310300 and r310332 because of test failure on Darwin
The commit r310291 introduced the failure. r310332 was a test fix commit and
r310300 was a followup commit. I reverted these two to avoid merge conflicts
when reverting.
The 'openmp-offload.c' test is failing on Darwin because the following
run lines:
// RUN: touch %t1.o
// RUN: touch %t2.o
// RUN: %clang -### -no-canonical-prefixes -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -save-temps -no-canonical-prefixes %t1.o %t2.o 2>&1 \
// RUN: | FileCheck -check-prefix=CHK-TWOCUBIN %s
trigger the following assertion:
Driver.cpp:3418:
assert(CachedResults.find(ActionTC) != CachedResults.end() &&
"Result does not exist??");
llvm-svn: 310345
Simon Pilgrim [Tue, 8 Aug 2017 11:03:30 +0000 (11:03 +0000)]
[DAGCombiner] Simplify shuffle mask index if the referenced input element is UNDEF
Fixes one of the cases in PR34041.
Differential Revision: https://reviews.llvm.org/D36393
llvm-svn: 310344
Daniel Sanders [Tue, 8 Aug 2017 10:44:31 +0000 (10:44 +0000)]
[globalisel][tablegen] Add support for importing 'imm' operands.
Summary:
This patch enables the import of rules containing 'imm' operands that do not
constrain the acceptable values using predicates. Support for ImmLeaf will
arrive in a later patch.
Depends on D35681
Reviewers: ab, t.p.northover, qcolombet, rovka, aditya_nandakumar
Reviewed By: rovka
Subscribers: kristof.beyls, javed.absar, igorb, llvm-commits
Differential Revision: https://reviews.llvm.org/D35833
llvm-svn: 310343
Chandler Carruth [Tue, 8 Aug 2017 10:13:23 +0000 (10:13 +0000)]
[PM] Fix a likely more critical infloop bug in the CGSCC pass manager.
This was just a bad oversight on my part. The code in question should
never have worked without this fix. But it turns out, there are
relatively few places that involve libfunctions that participate in
a single SCC, and unless they do, this happens to not matter.
The effect of not having this correct is that each time through this
routine, the edge from write_wrapper to write was toggled between a call
edge and a ref edge. First time through, it becomes a demoted call edge
and is turned into a ref edge. Next time it is a promoted call edge from
a ref edge. On, and on it goes forever.
I've added the asserts which should have always been here to catch silly
mistakes like this in the future as well as a test case that will
actually infloop without the fix.
The other (much scarier) infinite-inlining issue I think didn't actually
occur in practice, and I simply misdiagnosed this minor issue as that
much more scary issue. The other issue *is* still a real issue, but I'm
somewhat relieved that so far it hasn't happened in real-world code
yet...
llvm-svn: 310342
Abhishek Aggarwal [Tue, 8 Aug 2017 09:25:50 +0000 (09:25 +0000)]
Checking in files accidentally missed in later diffs of revision r310261
-- 2 files were missing in this commit which should have been there.
These files were submitted initially for review and were reviewed.
However, while updating the revision with newer diffs, I accidentally
forgot to include them in newer diffs. So commiting now.
llvm-svn: 310341
Siddharth Bhat [Tue, 8 Aug 2017 09:03:27 +0000 (09:03 +0000)]
[NFC] [GPUJIT] Print line number & size information on allocateMemoryForDeviceCuda failure
- It's useful to know the amount of memory asked for since, for example,
asking for `0` bytes of memory is illegal.
- Line number is helpful since we print the same message in the function
at different points.
llvm-svn: 310340
Craig Topper [Tue, 8 Aug 2017 06:19:24 +0000 (06:19 +0000)]
[InstCombine] Cast to BinaryOperator earlier in foldSelectIntoOp to simplify the code.
We no longer need the explicit operand count check or the later dynamic cast.
llvm-svn: 310339
Tobias Grosser [Tue, 8 Aug 2017 06:15:15 +0000 (06:15 +0000)]
[ScheduleOptimizer] Make matmul pattern detection work with delicm output
In certain cases delicm might decide to not leave the original array write in
the loop body, but to remove it and instead leave a transformed phi node as
write access. This commit teached the matmul pattern detection to order the
memory accesses according to when the access actually happens and use this
information to detect the new pattern. This makes pattern based matmul
optimization work for 2mm and 3mm in polybench 4 after
polly-position=before-vectorizer has been enabled.
llvm-svn: 310338
Tom Stellard [Tue, 8 Aug 2017 05:52:00 +0000 (05:52 +0000)]
AMDGPU: Fix warnings introduced by r310336
llvm-svn: 310337
Tom Stellard [Tue, 8 Aug 2017 04:57:55 +0000 (04:57 +0000)]
AMDGPU: Move R600 parts of AMDGPUISelDAGToDAG into their own class
Summary: This refactoring is required in order to split the R600 and GCN tablegen files.
Reviewers: arsenm
Subscribers: kzhuravl, wdng, nhaehnle, yaxunl, dstuttard, tpr, llvm-commits, t-tye
Differential Revision: https://reviews.llvm.org/D36286
llvm-svn: 310336
Konstantin Zhuravlyov [Tue, 8 Aug 2017 04:28:31 +0000 (04:28 +0000)]
AMDGPU: Also remove SI from docs
Differential Revision: https://reviews.llvm.org/D36424
llvm-svn: 310335
Chandler Carruth [Tue, 8 Aug 2017 02:27:49 +0000 (02:27 +0000)]
[PM] Relax the spelling of a pass name slightly in this test.
I forgot that MSVC doesn't preserve this typedef, my bad.
llvm-svn: 310334
Chandler Carruth [Tue, 8 Aug 2017 02:24:20 +0000 (02:24 +0000)]
[PM] Fix new LoopUnroll function pass by invalidating loop analysis
results when a loop is completely removed.
This is very hard to manifest as a visible bug. You need to arrange for
there to be a subsequent allocation of a 'Loop' object which gets the
exact same address as the one which the unroll deleted, and you need the
LoopAccessAnalysis results to be significant in the way that they're
stale. And you need a million other things to align.
But when it does, you get a deeply mysterious crash due to actually
finding a stale analysis result. This fixes the issue and tests for it
by directly checking we successfully invalidate things. I have not been
able to get *any* test case to reliably trigger this. Changes to LLVM
itself caused the only test case I ever had to cease to crash.
I've looked pretty extensively at less brittle ways of fixing this and
they are actually very, very hard to do. This is a somewhat strange and
unusual case as we have a pass which is deleting an IR unit, but is not
running within that IR unit's pass framework (which is what handles this
cleanly for the normal loop unroll). And where there isn't a definitive
way to clear *all* of the stale cache entries. And where the pass *is*
updating the core analysis that provides the IR units!
For example, we don't have any of these problems with Function analyses
because it is easy to clear out function analyses when the functions
themselves may have been deleted -- we clear an entire module's worth!
But that is too heavy of a hammer down here in the LoopAnalysisManager
layer.
A better long-term solution IMO is to require that AnalysisManager's
make their keys durable to this kind of thing. Specifically, when
caching an analysis for one IR unit that is conceptually "owned" by
a higher level IR unit, the AnalysisManager should incorporate this into
its data structures so that we can reliably clear these results without
having to teach each and every pass to do so manually as we do here. But
that is a change for another day as it will be a fairly invasive change
to the AnalysisManager infrastructure. Until then, this fortunately
seems to be quite rare.
llvm-svn: 310333
Reid Kleckner [Tue, 8 Aug 2017 01:36:16 +0000 (01:36 +0000)]
Fix openmp-offload.c test on Windows
llvm-svn: 310332
Reid Kleckner [Tue, 8 Aug 2017 01:33:53 +0000 (01:33 +0000)]
[codeview] Fix class name formatting
In particular, removes spaces between template arguments of class
templates to better match VS type visualizers.
llvm-svn: 310331
Vitaly Buka [Tue, 8 Aug 2017 01:01:59 +0000 (01:01 +0000)]
[asan] Restore dead-code-elimination optimization for Fuchsia
Summary:
r310244 fixed a bug introduced by r309914 for non-Fuchsia builds.
In doing so it also reversed the intended effect of the change for
Fuchsia builds, which was to allow all the AllocateFromLocalPool
code and its variables to be optimized away entirely.
This change restores that optimization for Fuchsia builds, but
doesn't have the original change's bug because the comparison
arithmetic now takes into account the size of the elements.
Submitted on behalf of Roland McGrath.
Reviewers: vitalybuka, alekseyshl
Reviewed By: alekseyshl
Subscribers: llvm-commits, kubamracek
Tags: #sanitizers
Differential Revision: https://reviews.llvm.org/D36430
llvm-svn: 310330
Shoaib Meenai [Tue, 8 Aug 2017 00:54:33 +0000 (00:54 +0000)]
[libc++abi] Use proper calling convention for TLS destructor
This is needed when using Windows threading.
llvm-svn: 310329
Eugene Zelenko [Tue, 8 Aug 2017 00:47:13 +0000 (00:47 +0000)]
[AMDGPU] Fix some Clang-tidy modernize-use-using and Include What You Use warnings; other minor fixes (NFC).
llvm-svn: 310328
Petr Hosek [Tue, 8 Aug 2017 00:37:59 +0000 (00:37 +0000)]
[CMake] Allow overriding lib dir suffix independently from LLVM
This matches the options already supported by libc++ and libc++abi.
Differential Revision: https://reviews.llvm.org/D36383
llvm-svn: 310327
Kostya Serebryany [Tue, 8 Aug 2017 00:17:20 +0000 (00:17 +0000)]
[libFuzzer] simplify code, NFC
llvm-svn: 310326
Kostya Serebryany [Tue, 8 Aug 2017 00:14:49 +0000 (00:14 +0000)]
[libFuzzer] remove stale code
llvm-svn: 310325
Kostya Serebryany [Tue, 8 Aug 2017 00:12:09 +0000 (00:12 +0000)]
[libFuzzer] simplify the implementation of -print_coverage=1
llvm-svn: 310324
Kamil Rytarowski [Mon, 7 Aug 2017 23:38:14 +0000 (23:38 +0000)]
Fix asan_test.cc build on NetBSD
Summary:
Include <stdarg.h> for variable argument list macros (va_list, va_start etc).
Add fallback definition of _LIBCPP_GET_C_LOCALE, this is required for
GNU libstdc++ compatibility. Define new macro SANITIZER_GET_C_LOCALE.
This value is currently required for FreeBSD and NetBSD for printf_l(3) tests.
Sponsored by <The NetBSD Foundation>
Reviewers: joerg, kcc, vitalybuka, filcab, fjricci
Reviewed By: vitalybuka
Subscribers: llvm-commits, emaste, kubamracek, #sanitizers
Tags: #sanitizers
Differential Revision: https://reviews.llvm.org/D36406
llvm-svn: 310323
Kamil Rytarowski [Mon, 7 Aug 2017 23:34:45 +0000 (23:34 +0000)]
Add NetBSD support in asan_stack.h
Summary:
Part of the code inspired by the original work on libsanitizer in GCC 5.4 by Christos Zoulas.
Sponsored by <The NetBSD Foundation>
Reviewers: joerg, kcc, vitalybuka, filcab, fjricci
Reviewed By: vitalybuka
Subscribers: davide, kubamracek, llvm-commits, #sanitizers
Tags: #sanitizers
Differential Revision: https://reviews.llvm.org/D36377
llvm-svn: 310322
Craig Topper [Mon, 7 Aug 2017 22:35:55 +0000 (22:35 +0000)]
[KnownBits] Fix copy pasto in comment. NFC
llvm-svn: 310320
Tobias Grosser [Mon, 7 Aug 2017 22:33:34 +0000 (22:33 +0000)]
Change Polly's position to "before-vectorizer"
Polly has traditionally always been executed at the beginning of the pass
pipeline as LLVM's inliner and DeLICM passes introduced plenty of scalar
dependences which prevented any kind of useful high-level loop optimizations
later in the pass pipeline. With DeLICM now being available, Polly can also
run optimizations when folded into the pass pipeline. This has the benefit
that Polly should now be more effective on C++ code and as an additional bonus,
no additional early canonicalization phase must be run. As a result, Polly
touches the code only if it applies a transformation. Code that does not
benefit from Polly is not touched and consequently will have the very same
execution time as without Polly enabled. Random performance changes, as could
sometimes be observed with polly-position=early are consequently not possible
any more. If performance is changed, this is due to Polly is choosing to
perform a transformation. If this choice is wrong, it can be fixed directly
in Polly.
http://polly.llvm.org/docs/Architecture.html#polly-in-the-llvm-pass-pipeline
llvm-svn: 310319
Sean Callanan [Mon, 7 Aug 2017 22:27:30 +0000 (22:27 +0000)]
This adds the argument --dump-ir to clang-import-test, which allows
viewing of the final IR. This is useful for confirming that
structure layout was correct.
I've added two tests:
- A test that checks that structs in top-level code are completed
correctly during struct layout (they are)
- A test that checks that structs defined in function bodies are
cpmpleted correctly during struct layout (currently they are not,
so this is XFAIL).
The second test fails because LookupSameContext()
(ExternalASTMerger.cpp) can't find the struct. This is an issue I
intend to resolve separately.
Differential Revision: https://reviews.llvm.org/D36429
llvm-svn: 310318
Simon Pilgrim [Mon, 7 Aug 2017 22:20:06 +0000 (22:20 +0000)]
[X86][AVX] Added test for broadcast shuffle from binary sources with undefs (D36393)
llvm-svn: 310317
Tobias Grosser [Mon, 7 Aug 2017 22:10:23 +0000 (22:10 +0000)]
[test] Add some missing options that become necessary after the recent default changes
llvm-svn: 310315
Tobias Grosser [Mon, 7 Aug 2017 22:06:07 +0000 (22:06 +0000)]
[DeLICM] Enable partial writes
This allows us to remove more scalar dependences. While this feature is still
rather experimental, we want to give it sufficient test coverage.
llvm-svn: 310314
Tobias Grosser [Mon, 7 Aug 2017 22:04:20 +0000 (22:04 +0000)]
Enable delicm to automatically remove scalar loop carried dependences
While this code is still rather we enable it by default to get better test
coverage.
llvm-svn: 310313
Tobias Grosser [Mon, 7 Aug 2017 22:02:06 +0000 (22:02 +0000)]
[test] Add one more test case for the previous commit
llvm-svn: 310312
Tobias Grosser [Mon, 7 Aug 2017 22:01:29 +0000 (22:01 +0000)]
[ZoneAlgo] Allow two writes that write identical values into same array slot
Two write statements which write into the very same array slot generally are
conflicting. However, in case the value that is written is identical, this
does not cause any problem. Hence, allow such write pairs in this specific
situation.
llvm-svn: 310311
Matt Arsenault [Mon, 7 Aug 2017 22:00:58 +0000 (22:00 +0000)]
AMDGPU: Implement getMinimumNopSize
llvm-svn: 310310
Reid Kleckner [Mon, 7 Aug 2017 21:23:38 +0000 (21:23 +0000)]
[Object] Initialize LoadConfig member to null
Executables may not contain a load config, and clients should be able to
test for nullability. Previously we'd return uninitialized memory. Now
getLoadConfig32/64 return valid pointers or null.
Fixes PR34108
llvm-svn: 310308
Gheorghe-Teodor Bercea [Mon, 7 Aug 2017 21:11:10 +0000 (21:11 +0000)]
[OpenMP] Error when trying to offload to an unsupported architecture
Summary: Throw an error when offloading is unsupported for a particular target architecture.
Reviewers: sfantao, caomhin, carlo.bertolli, ABataev, Hahnfeld
Reviewed By: ABataev
Subscribers: cfe-commits, rengolin
Differential Revision: https://reviews.llvm.org/D32035
llvm-svn: 310307
Gheorghe-Teodor Bercea [Mon, 7 Aug 2017 20:57:59 +0000 (20:57 +0000)]
[OpenMP] Prevent emission of exception handling code when using OpenMP to offload to NVIDIA devices.
Summary: For the OpenMP toolchain which offloads to NVIDIA GPUs make sure that no exception handling code is emitted.
Reviewers: arpith-jacob, sfantao, caomhin, carlo.bertolli, ABataev, Hahnfeld, hfinkel, tstellar
Reviewed By: ABataev, Hahnfeld
Subscribers: rengolin, Hahnfeld, cfe-commits
Differential Revision: https://reviews.llvm.org/D29904
llvm-svn: 310306
George Karpenkov [Mon, 7 Aug 2017 20:56:11 +0000 (20:56 +0000)]
Do not instrument libFuzzer itself when built with -DLLVM_USE_SANITIZE_COVERAGE
Fixes regression from https://reviews.llvm.org/D36295
Differential Revision: https://reviews.llvm.org/D36428
llvm-svn: 310305
Andreas Simbuerger [Mon, 7 Aug 2017 20:54:20 +0000 (20:54 +0000)]
[Polly] Fully-Indexed static expansion
This commit implements the initial version of fully-indexed static
expansion.
```
for(int i = 0; i<Ni; i++)
for(int j = 0; j<Ni; j++)
S: B[j] = j;
T: A[i] = B[i]
```
After the pass, we want this :
```
for(int i = 0; i<Ni; i++)
for(int j = 0; j<Ni; j++)
S: B[i][j] = j;
T: A[i] = B[i][i]
```
For now we bail (fail) in the following cases:
- Scalar access
- Multiple writes per SAI
- MayWrite Access
- Expansion that leads to an access to the original array
Furthermore: We still miss checks for escaping references to the array
base pointers. A future commit will add the missing escape-checks to
stay correct in those cases. The expansion is still locked behind a
CLI-Option and should not yet be used.
Patch contributed by: Nicholas Bonfante <bonfante.nicolas@gmail.com>
Reviewers: simbuerg, Meinersbur, bollu
Reviewed By: Meinersbur
Subscribers: mgorny, llvm-commits, pollydev
Differential Revision: https://reviews.llvm.org/D34982
llvm-svn: 310304
Martin Storsjo [Mon, 7 Aug 2017 20:52:04 +0000 (20:52 +0000)]
[AArch64] Ignore stdcall and similar on aarch64/windows
This is similar to what's done on arm and x86_64, where
these calling conventions are silently ignored, as in
SVN r245076.
Differential Revision: https://reviews.llvm.org/D36105
llvm-svn: 310303
Gheorghe-Teodor Bercea [Mon, 7 Aug 2017 20:31:51 +0000 (20:31 +0000)]
[OpenMP] Make OpenMP generated code for the NVIDIA device relocatable by default
Summary: When device offloading is enabled and the device is an NVIDIA GPU, OpenMP target regions must be compiled with relocation enabled by passing the "-c" flag to the PTXAS invocation.
Reviewers: arpith-jacob, caomhin, carlo.bertolli, ABataev, Hahnfeld, jlebar, hfinkel, tstellar
Reviewed By: Hahnfeld
Subscribers: Hahnfeld, rengolin, mkuron, cfe-commits
Differential Revision: https://reviews.llvm.org/D29642
llvm-svn: 310300
George Burgess IV [Mon, 7 Aug 2017 20:26:33 +0000 (20:26 +0000)]
Mark static variables static; NFC.
llvm-svn: 310299
Zachary Turner [Mon, 7 Aug 2017 20:24:01 +0000 (20:24 +0000)]
[llvm-pdbutil] Don't crash when a section contrib's isect is invalid.
llvm-svn: 310298
Zachary Turner [Mon, 7 Aug 2017 20:23:45 +0000 (20:23 +0000)]
[lld] Write the DataCRC to the output PDB.
llvm-svn: 310297
Dehao Chen [Mon, 7 Aug 2017 20:23:20 +0000 (20:23 +0000)]
Move the SampleProfileLoader right after EarlyFPM.
Summary: SampleProfileLoader pass do need to happen after some early cleanup passes so that inlining can happen correctly inside the SampleProfileLoader pass.
Reviewers: chandlerc, davidxl, tejohnson
Reviewed By: chandlerc, tejohnson
Subscribers: sanjoy, mehdi_amini, eraman, llvm-commits
Differential Revision: https://reviews.llvm.org/D36333
llvm-svn: 310296
Gheorghe-Teodor Bercea [Mon, 7 Aug 2017 20:19:23 +0000 (20:19 +0000)]
[OpenMP] Pass -v to PTXAS if it was passed to the driver.
Summary: When compiling code being offloaded by OpenMP to an NVIDIA GPU, pass the -v to PTXAS if it was passed to the CLANG driver.
Reviewers: arpith-jacob, caomhin, carlo.bertolli, ABataev, jlebar, hfinkel, tstellar
Reviewed By: jlebar
Subscribers: Hahnfeld, rengolin, cfe-commits
Differential Revision: https://reviews.llvm.org/D29644
llvm-svn: 310295
Gheorghe-Teodor Bercea [Mon, 7 Aug 2017 20:01:48 +0000 (20:01 +0000)]
[OpenMP] Integrate OpenMP target region cubin into host binary
Summary: OpenMP device offloading code generation produces a cubin file which is then integrated in the host binary using the host linker.
Reviewers: arpith-jacob, caomhin, carlo.bertolli, ABataev, Hahnfeld, jlebar, rnk, hfinkel, tstellar
Reviewed By: hfinkel
Subscribers: sfantao, rnk, rengolin, cfe-commits
Differential Revision: https://reviews.llvm.org/D29654
llvm-svn: 310291
Shoaib Meenai [Mon, 7 Aug 2017 19:59:58 +0000 (19:59 +0000)]
[libc++] Don't hardcode namespace in manual mangling
libc++'s inline namespace can change depending on the ABI version.
Instead of hardcoding __1 in the manual Microsoft ABI manglings for the
iostream globals, stringify _LIBCPP_NAMESPACE and use that instead, to
work across all ABI versions.
llvm-svn: 310290
Evgeny Stupachenko [Mon, 7 Aug 2017 19:56:34 +0000 (19:56 +0000)]
Reapply fix PR23384 (part 3 of 3) r304824 (was reverted in r305720).
The root cause of reverting was fixed - PR33514.
Summary:
The patch makes instruction count the highest priority for
LSR solution for X86 (previously registers had highest priority).
Reviewers: qcolombet
Differential Revision: http://reviews.llvm.org/D30562
From: Evgeny Stupachenko <evstupac@gmail.com>
<evgeny.v.stupachenko@intel.com>
llvm-svn: 310289
Aaron Ballman [Mon, 7 Aug 2017 19:26:17 +0000 (19:26 +0000)]
Removing an unused variable that was missed with the refactoring in r310272; NFC.
llvm-svn: 310285
Tobias Grosser [Mon, 7 Aug 2017 19:20:57 +0000 (19:20 +0000)]
[GPGPU] Remove redundant constructors
llvm-svn: 310284
Connor Abbott [Mon, 7 Aug 2017 19:10:56 +0000 (19:10 +0000)]
[AMDGPU] Add pseudo "old" source to all DPP instructions
Summary:
All instructions with the DPP modifier may not write to certain lanes of
the output if bound_ctrl=1 is set or any bits in bank_mask or row_mask
aren't set, so the destination register may be both defined and modified.
The right way to handle this is to add a constraint that the destination
register is the same as one of the inputs. We could tie the destination
to the first source, but that would be too restrictive for some use-cases
where we want the destination to be some other value before the
instruction executes. Instead, add a fake "old" source and tie it to the
destination. Effectively, the "old" source defines what value unwritten
lanes will get. We'll expose this functionality to users with a new
intrinsic later.
Also, we want to use DPP instructions for computing derivatives, which
means we need to set WQM for them. We also need to enable the entire
wavefront when using DPP intrinsics to implement nonuniform subgroup
reductions, since otherwise we'll get incorrect results in some cases.
To accomodate this, add a new operand to all DPP instructions which will
be interpreted by the SI WQM pass. This will be exposed with a new
intrinsic later. We'll also add support for Whole Wavefront Mode later.
I also fixed llvm.amdgcn.mov.dpp to overwrite the source and fixed up
the test. However, I could also keep the old behavior (where lanes that
aren't written are undefined) if people want it.
Reviewers: tstellar, arsenm
Subscribers: kzhuravl, wdng, nhaehnle, yaxunl, dstuttard, tpr, t-tye
Differential Revision: https://reviews.llvm.org/D34716
llvm-svn: 310283
Gheorghe-Teodor Bercea [Mon, 7 Aug 2017 18:43:37 +0000 (18:43 +0000)]
Non-functional change. Fix previous patch D34784.
llvm-svn: 310282
Michael Kruse [Mon, 7 Aug 2017 18:40:29 +0000 (18:40 +0000)]
[ForwardOpTree] Use known array content analysis to forward load instructions.
This is an addition to the -polly-optree pass that reuses the array
content analysis from DeLICM to find array elements that contain the
same value as the value loaded when the target statement instance
is executed.
The analysis is now enabled by default.
The known content analysis could also be used to rematerialize any
llvm::Value that was written to some array element, but currently
only loads are forwarded.
Differential Revision: https://reviews.llvm.org/D36380
llvm-svn: 310279
Benjamin Kramer [Mon, 7 Aug 2017 18:31:01 +0000 (18:31 +0000)]
Add some missing -no-canonical-prefixes.
llvm-svn: 310278
Matt Arsenault [Mon, 7 Aug 2017 18:30:35 +0000 (18:30 +0000)]
AMDGPU: Remove -mcpu=SI
Leftover from before amdgcn/r600 split.
llvm-svn: 310277
Kuba Mracek [Mon, 7 Aug 2017 18:12:59 +0000 (18:12 +0000)]
[tsan] Fix format string in WriteMemoryProfile
The current format string is broken and fails to parse.
Differential Revision: https://reviews.llvm.org/D36192
llvm-svn: 310276
Matt Arsenault [Mon, 7 Aug 2017 18:12:48 +0000 (18:12 +0000)]
AMDGPU: Remove redundant opt level check
addOptimizedRegAlloc isn't used for -O0 already.
llvm-svn: 310275
Matt Arsenault [Mon, 7 Aug 2017 18:12:47 +0000 (18:12 +0000)]
AMDGPU: Remove FixControlFlowLiveIntervals pass
This hasn't done anything in a long time. This was
running after the the control flow pseudos were expanded,
so this would never find them. The control flow pseudo
expansion was moved to solve the problem this pass was
supposed to solve in the first place, except handling
it earlier also fixes it for fast regalloc which doesn't
use LiveIntervals.
Noticed by checking LCOV reports.
llvm-svn: 310274
Kuba Mracek [Mon, 7 Aug 2017 18:12:01 +0000 (18:12 +0000)]
[asan] Make dump_registers.cc more stable
Differential Revision: https://reviews.llvm.org/D36231
llvm-svn: 310273
Craig Topper [Mon, 7 Aug 2017 18:10:39 +0000 (18:10 +0000)]
[InstCombine] Support (X | C1) & C2 --> (X & C2^(C1&C2)) | (C1&C2) for vector splats
Note the original code I deleted incorrectly listed this as (X | C1) & C2 --> (X & C2^(C1&C2)) | C1 Which is only valid if C1 is a subset of C2. This relied on SimplifyDemandedBits to remove any extra bits from C1 before we got to that code.
My new implementation avoids relying on that behavior so that it can be naively verified with alive.
Differential Revision: https://reviews.llvm.org/D36384
llvm-svn: 310272
Kuba Mracek [Mon, 7 Aug 2017 18:07:20 +0000 (18:07 +0000)]
[sanitizer] Remove use of task_for_pid from sanitizer_stoptheworld_mac.cc
Using task_for_pid to get the "self" task is not necessary, and it can fail (e.g. for sandboxed processes). Let's just use mach_task_self().
Differential Revision: https://reviews.llvm.org/D36284
llvm-svn: 310271
Abhishek Aggarwal [Mon, 7 Aug 2017 17:15:26 +0000 (17:15 +0000)]
Fixed build failure for revision r310261
-- Was failing for Linux
llvm-svn: 310270
Matt Arsenault [Mon, 7 Aug 2017 17:08:44 +0000 (17:08 +0000)]
AMDGPU: Use a custom areInlineCompatible
Fixes not inlining OpenCL library functions on AMDGPU,
which don't have an explicitly set target-cpu.
llvm-svn: 310269
Simon Pilgrim [Mon, 7 Aug 2017 16:49:09 +0000 (16:49 +0000)]
[X86][AVX] Add full test coverage of subvector_broadcasts from registers
X86SubVBroadcast is for memory subvector broadcasts, but we must test that it handles all cases without the load as well just in case.
This was noticed while I was triaging the test cases from PR34041.
llvm-svn: 310268
Simon Dardis [Mon, 7 Aug 2017 16:08:11 +0000 (16:08 +0000)]
[DebugInfo][DWARF] Address paulr's comment on rL310253.
llvm-svn: 310267
Abhishek Aggarwal [Mon, 7 Aug 2017 15:53:30 +0000 (15:53 +0000)]
Fixed build failure for revision r310261
-- Build was failing for freebsd
llvm-svn: 310266
Simon Pilgrim [Mon, 7 Aug 2017 15:50:43 +0000 (15:50 +0000)]
[X86][AVX] Cleanup subvector broadcast tests - remove old prefixes.
llvm-svn: 310265
Sanjay Patel [Mon, 7 Aug 2017 15:47:48 +0000 (15:47 +0000)]
[x86] revert r310208 to investigate test-suite failures (PR34105 / PR34097)
llvm-svn: 310264
Gheorghe-Teodor Bercea [Mon, 7 Aug 2017 15:39:11 +0000 (15:39 +0000)]
[OpenMP] Add flag for specifying the target device architecture for OpenMP device offloading
Summary:
OpenMP has the ability to offload target regions to devices which may have different architectures.
A new -fopenmp-target-arch flag is introduced to specify the device architecture.
In this patch I use the new flag to specify the compute capability of the underlying NVIDIA architecture for the OpenMP offloading CUDA tool chain.
Only a host-offloading test is provided since full device offloading capability will only be available when [[ https://reviews.llvm.org/D29654 | D29654 ]] lands.
Reviewers: hfinkel, Hahnfeld, carlo.bertolli, caomhin, ABataev
Reviewed By: hfinkel
Subscribers: guansong, cfe-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D34784
llvm-svn: 310263
Simon Dardis [Mon, 7 Aug 2017 15:37:57 +0000 (15:37 +0000)]
[DebugInfo][DWARF] Correct some usages of PRIx32 to PRIx64
These lead to tests failing spuriously as the values after being rendered to a
string were incorrect.
Reviewers: clayborg
Differential Revision: https://reviews.llvm.org/D36319
llvm-svn: 310262