Shoaib Meenai [Sun, 14 May 2017 18:34:56 +0000 (18:34 +0000)]
[COFF] Gracefully handle empty .drectve sections
Running `llvm-readobj -coff-directives msvcrt.lib` resulted in this error:
Invalid data was encountered while parsing the file
This happened because some of the object files in the archive have empty
`.drectve` sections. These empty sections result in a `parse_failed` error being
returned from `COFFObjectFile::getSectionContents()`, which in turn caused
`llvm-readobj` to stop. With this change, `getSectionContents` now returns
success, and like before the resulting array is empty.
Patch by Dave Lee.
Differential Revision: https://reviews.llvm.org/D32652
llvm-svn: 303014
Simon Pilgrim [Sun, 14 May 2017 17:59:46 +0000 (17:59 +0000)]
[X86][XOP] XOP's general v16i8 shifts will be used instead of v8i16 shift + mask.
Tweak cost model to match what lowering actually does.
llvm-svn: 303013
Simon Pilgrim [Sun, 14 May 2017 17:36:07 +0000 (17:36 +0000)]
[X86][SSE] Account for cost of extract/insert of v32i8 vector shifts
llvm-svn: 303012
Zachary Turner [Sun, 14 May 2017 17:11:05 +0000 (17:11 +0000)]
[StringExtras] Add llvm::to_integer.
This is a very thin wrapper around StringRef::getAsInteger.
It serves three purposes.
1) It allows a cleaner syntax when you have something other than
a StringRef - for example, a std::string or an llvm::SmallString.
Previously, in this case you would have to write something like:
StringRef(MyStr).getAsInteger(0, Result)
by explicitly constructing a temporary StringRef. This can be
done implicitly however with the new function by just writing:
to_integer(MyStr, ...).
2) Correcting the travesty that is getAsInteger's return value.
This function returns true on success, and false on failure.
While this may cause confusion with people familiar with the
getAsInteger API, there seems to be widespread agreement that
the return semantics of getAsInteger was a mistake.
3) It allows the Radix to be deduced as a default argument by
putting it last in the parameter list. Most uses of getAsInteger
pass 0 for the first argument. With this syntax it can just be
omitted.
llvm-svn: 303011
Simon Pilgrim [Sun, 14 May 2017 13:38:53 +0000 (13:38 +0000)]
[X86][XOP] Account for cost of extract/insert of 256-bit vector shifts
llvm-svn: 303010
Simon Pilgrim [Sun, 14 May 2017 11:46:26 +0000 (11:46 +0000)]
[X86][AVX] Allow 32-bit targets to peek through subvectors to extract constant splats for vXi64 shifts.
llvm-svn: 303009
Simon Pilgrim [Sun, 14 May 2017 11:13:03 +0000 (11:13 +0000)]
[X86][AVX] Add additional 32-bit target vector shift tests
Shows issue with 32-bits not being able to peek through subvectors to extract constant splats
llvm-svn: 303008
Tobias Grosser [Sun, 14 May 2017 09:20:56 +0000 (09:20 +0000)]
[Simplify] Remove some leftover dead code
llvm-svn: 303007
Tobias Grosser [Sun, 14 May 2017 09:09:54 +0000 (09:09 +0000)]
[Polly] Fix code generation of llvm.expect intrinsic
At the time of code generation, an instruction with an llvm intrinsic is ignored
in copyBB. However, if the value of the instruction is used later in the
program, the value needs to be synthesized. However, this is causing some issues
with the instructions being generated in a hoisted basic block.
Removing llvm.expect from the list of ignored intrinsics fixes this bug.
This resolves http://llvm.org/PR32324.
Contributed-by: Annanay Agarwal <cs14btech11001@iith.ac.in>
Tags: #polly
Differential Revision: https://reviews.llvm.org/D32992
llvm-svn: 303006
Sylvestre Ledru [Sun, 14 May 2017 07:55:01 +0000 (07:55 +0000)]
update of the url
llvm-svn: 303005
Craig Topper [Sun, 14 May 2017 07:54:43 +0000 (07:54 +0000)]
[InstSimplify] Add patterns for folding (A & B) | (~A ^ B) -> (~A ^ B) and its commuted variants.
We already had (A & ~B) | (A ^ B), but we missed the cases where the not was part of the xor.
llvm-svn: 303004
Craig Topper [Sun, 14 May 2017 07:54:40 +0000 (07:54 +0000)]
foo
llvm-svn: 303003
Craig Topper [Sun, 14 May 2017 06:18:34 +0000 (06:18 +0000)]
[BasicAA] Alphabetize includes. NFC
llvm-svn: 303002
Yan Wang [Sun, 14 May 2017 04:14:59 +0000 (04:14 +0000)]
[clang-tidy] TwineLocalCheck: add param # checking
Summary:
The statement **getArg** tries to get the first one without checking, which may cause segmentation fault.
Reviewers: chh, bkramer
Reviewed By: bkramer
Subscribers: cfe-commits, xazax.hun
Tags: #clang-tools-extra
Differential Revision: https://reviews.llvm.org/D33103
llvm-svn: 303001
Xinliang David Li [Sun, 14 May 2017 02:58:39 +0000 (02:58 +0000)]
Renable test that was disabled due to cost analysis
llvm-svn: 303000
Xinliang David Li [Sun, 14 May 2017 02:54:02 +0000 (02:54 +0000)]
Fix test failure on windows -- do not return deleted func
llvm-svn: 302999
Zachary Turner [Sun, 14 May 2017 01:13:40 +0000 (01:13 +0000)]
[llvm-pdbdump] Add the option to sort functions and data.
llvm-svn: 302998
Simon Pilgrim [Sat, 13 May 2017 22:10:58 +0000 (22:10 +0000)]
[SelectionDAG] Added support for EXTRACT_SUBVECTOR/CONCAT_VECTORS demandedelts in ComputeNumSignBits
llvm-svn: 302997
Peter Collingbourne [Sat, 13 May 2017 22:10:13 +0000 (22:10 +0000)]
Add missing files
llvm-svn: 302996
Peter Collingbourne [Sat, 13 May 2017 22:06:46 +0000 (22:06 +0000)]
Move lib/LibDriver -> lib/ToolDrivers/llvm-lib. NFCI.
This reorganisation prevents us from cluttering up the top-level lib directory
with more driver libraries such as llvm-dlltool (see D29892).
llvm-svn: 302995
Simon Pilgrim [Sat, 13 May 2017 21:50:18 +0000 (21:50 +0000)]
[X86][SSE] Test showing missing EXTRACT_SUBVECTOR/CONCAT_VECTORS demandedelts support in ComputeNumSignBits
llvm-svn: 302994
Simon Pilgrim [Sat, 13 May 2017 19:57:10 +0000 (19:57 +0000)]
[SelectionDAG] Add VECTOR_SHUFFLE support to ComputeNumSignBits
llvm-svn: 302993
Simon Pilgrim [Sat, 13 May 2017 17:41:07 +0000 (17:41 +0000)]
[X86][SSE] Test showing inability of ComputeNumSignBits to resolve shuffles
llvm-svn: 302992
Craig Topper [Sat, 13 May 2017 17:22:16 +0000 (17:22 +0000)]
[ValueTracking] Remove const_casts on several calls to computeKnownBits and ComputeSignBit. NFC
llvm-svn: 302991
Justin Bogner [Sat, 13 May 2017 16:24:38 +0000 (16:24 +0000)]
MSan: Mark MemorySanitizer tests that use x86 intrinsics as REQUIRES: x86
Tests that use target intrinsics are inherently target specific. Mark
them as such.
llvm-svn: 302990
Simon Pilgrim [Sat, 13 May 2017 13:42:35 +0000 (13:42 +0000)]
[x86, SSE] AVX1 PR28129 (256-bit all-ones rematerialization)
Further perf tests on Jaguar indicate that:
vxorps %ymm0, %ymm0, %ymm0
vcmpps $15, %ymm0, %ymm0, %ymm0
is consistently faster (by about 9%) than:
vpcmpeqd %xmm0, %xmm0, %xmm0
vinsertf128 $1, %xmm0, %ymm0, %ymm0
Testing equivalent code on a SandyBridge (E5-2640) puts it slightly (~3%) faster as well.
Committed on behalf of @dtemirbulatov
Differential Revision: https://reviews.llvm.org/D32416
llvm-svn: 302989
Simon Pilgrim [Sat, 13 May 2017 13:25:57 +0000 (13:25 +0000)]
[LoopOptimizer][Fix]PR32859, PR24738
The Loop vectorizer pass introduced undef value while it is fixing output of LCSSA form.
Here it is:
before: %e.0.ph = phi i32 [ 0, %for.inc.2.i ]
after: %e.0.ph = phi i32 [ 0, %for.inc.2.i ], [ undef, %middle.block ]
and after this change we have:
%e.0.ph = phi i32 [ 0, %for.inc.2.i ]
%e.0.ph = phi i32 [ 0, %for.inc.2.i ], [ 0, %middle.block ]
Committed on behalf of @dtemirbulatov
Differential Revision: https://reviews.llvm.org/D33055
llvm-svn: 302988
Michael Kruse [Sat, 13 May 2017 12:20:57 +0000 (12:20 +0000)]
[Simplify] Remove identical write removal. NFC.
Removal of overwritten writes currently encompasses all the cases
of the identical write removal.
There is an observable behavioral change in that the last, instead
of the first, MemoryAccess is kept. This should not affect the
generated code, however.
Differential Revision: https://reviews.llvm.org/D33143
llvm-svn: 302987
Michael Kruse [Sat, 13 May 2017 11:49:34 +0000 (11:49 +0000)]
[Simplify] Remove writes that are overwritten.
Remove memory writes that are overwritten by later writes. This works
for StoreInsts:
store double 21.0, double* %A
store double 42.0, double* %A
scalar writes at the end of a statement and mixes of these.
Multiple writes can be the result of DeLICM, which might map multiple
writes to the same location when it knows that these do no conflict
(for instance because they write the same value). Such writes
interfere with pattern-matched optimization such as gemm and may not
get removed by other LLVM passes after code generation.
Differential Revision: https://reviews.llvm.org/D33142
llvm-svn: 302986
Vivek Pandya [Sat, 13 May 2017 10:59:05 +0000 (10:59 +0000)]
This reverts r302984
llvm-svn: 302985
Vivek Pandya [Sat, 13 May 2017 08:55:43 +0000 (08:55 +0000)]
Simplify MIR Output used for Codegen Testing
- MIRYamlMapping: Default value provided for fields which have optional
mappings. Implemented == operators for required classes. When a field's value is
same as default value specified YAML IO class will not print it.
- MIRPrinter: Above mentioned behaviour is not on by default. If -simplify-mir
option not specified, then make yaml::Output to print fields with default values
too.
Differential Revision: https://reviews.llvm.org/D32304
llvm-svn: 302984
Craig Topper [Sat, 13 May 2017 07:14:17 +0000 (07:14 +0000)]
[APInt] Use Lo_32/Hi_32/Make_64 in a few more places in the divide code. NFCI
llvm-svn: 302983
Craig Topper [Sat, 13 May 2017 06:56:04 +0000 (06:56 +0000)]
[InstCombine] Prevent InstCombine from triggering an extra iteration if something changed in the initial Worklist creation
Summary:
If the Worklist build causes an IR change this change flag currently factors into the flag for running another iteration of the iteration loop. But only changes during processing should trigger another loop.
This patch captures the worklist creation change flag into the outside the loop flag currently used for DbgDeclares and only sends that flag up to the caller. Rerunning the loop only depends on IC.run() now.
This uses the debug output of InstCombine to determine if one or two iterations run. I couldn't think of a better way to detect it since the second spurious iteration shoudn't make any visible changes. Just wasted computation.
I can do a pre-commit of the test case with the CHECK-NOT as a CHECK if this is an ok way to check this.
This is a subset of D31678 as I'm still not sure how to verify the analysis behavior for that.
Reviewers: davide, majnemer, spatel, chandlerc
Reviewed By: davide
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D32453
llvm-svn: 302982
Akira Hatanaka [Sat, 13 May 2017 06:28:17 +0000 (06:28 +0000)]
[libcxxabi] Do not align field unwindHeader when building for ARM EHABI.
For ARM EHABI, _Unwind_Exception is an alias of _Unwind_Control_Block,
which is not aligned, so we shouldn't align unwindHeader either.
rdar://problem/
25364625
llvm-svn: 302981
Justin Bogner [Sat, 13 May 2017 05:52:17 +0000 (05:52 +0000)]
ConstProp: Split x86 SSE intrinsic tests out of calls.ll
This allows us to mark this as `REQUIRES: x86`, since it uses x86
target specific intrinsics.
llvm-svn: 302980
Justin Bogner [Sat, 13 May 2017 05:39:46 +0000 (05:39 +0000)]
InstCombine: Move tests that use target intrinsics into subdirectories
Tests with target intrinsics are inherently target specific, so it
doesn't actually make sense to run them if we've excluded their
target.
llvm-svn: 302979
Akira Hatanaka [Sat, 13 May 2017 03:14:49 +0000 (03:14 +0000)]
[libcxxabi] Align unwindHeader on a double-word boundary.
r276215 made a change to annotate _Unwind_Exception with attribute
"aligned" so that an exception object following field __cxa_exception
is sufficiently aligned. This fix hasn't been incorporated to unwind.h
on Darwin since it is an ABI breaking change.
Instead of annotating struct _Unwind_Exception with the attribute, this
commit annotates field unwindHeader of __cxa_exception. This ensures the
exception object is sufficiently aligned without breaking the ABI.
This recommits r302763 with fixes to RUN lines in the test case.
rdar://problem/
25364625
Differential Revision: https://reviews.llvm.org/D33030
llvm-svn: 302978
NAKAMURA Takumi [Sat, 13 May 2017 03:05:38 +0000 (03:05 +0000)]
Disable llvm/test/Transforms/NewGVN/pr32934.ll while Davide is investigating.
llvm-svn: 302977
Davide Italiano [Sat, 13 May 2017 02:45:47 +0000 (02:45 +0000)]
[NewGVN] XFAIL a flaky test until I find out what's going on.
I bet the change is correct but this test seems to expose some underlying
problem that manifest only on some buildbots, and I'm not able to reproduce
locally. Unfortunately I can't debug right now but I don't want to annoy
people with spurious failures, so I'll XFAIL until I can take a look (over
the weekend).
llvm-svn: 302976
Sean Callanan [Sat, 13 May 2017 00:46:33 +0000 (00:46 +0000)]
[ASTImporter] Improve handling of incomplete types
ASTImporter has some bugs when it's importing types
that themselves come from an ExternalASTSource. This
is exposed particularly in the behavior when
comparing complete TagDecls with forward
declarations. This patch does several things:
- Adds a test case making sure that conflicting
forward-declarations are resolved correctly;
- Extends the clang-import-test harness to test
two-level importing, so that we make sure we
complete types when necessary; and
- Fixes a few bugs I found this way. Failure to
complete types was one; however, I also discovered
that complete RecordDecls aren't properly added to
the redecls chain for existing forward
declarations.
llvm-svn: 302975
Craig Topper [Sat, 13 May 2017 00:35:30 +0000 (00:35 +0000)]
[APInt] Fix typo in comment. NFC
llvm-svn: 302974
Dylan McKay [Sat, 13 May 2017 00:22:34 +0000 (00:22 +0000)]
[AVR] When lowering Select8/Select16, put newly generated MBBs in the same spot
Contributed by Dr. Gergő Érdi.
Fixes a bug.
Raised from (https://github.com/avr-rust/rust/issues/49).
llvm-svn: 302973
Justin Bogner [Sat, 13 May 2017 00:12:52 +0000 (00:12 +0000)]
AA: Use generic intrinsics for tests instead of target specific ones
Update a few tests to use llvm.masked.load/store instead of arm neon
vector loads and stores, and move the tests that are actually specific
to those arm intrinsics to their own files. This lets us mark the
tests that use target specific intrinsics as requiring those targets.
llvm-svn: 302972
Justin Bogner [Sat, 13 May 2017 00:11:39 +0000 (00:11 +0000)]
[CodeView] Silence some -Wsign-compare warnings
llvm-svn: 302971
Dylan McKay [Sat, 13 May 2017 00:00:26 +0000 (00:00 +0000)]
[AVR] Remove an unused variable
llvm-svn: 302970
Richard Smith [Sat, 13 May 2017 00:00:16 +0000 (00:00 +0000)]
Add LangOptions method to query whether we are tracking the owning module for a local declaration.
In preparation for expanding this behavior to cover additional cases.
llvm-svn: 302969
Xinliang David Li [Fri, 12 May 2017 23:41:43 +0000 (23:41 +0000)]
[PartialInlining] Profile based cost analysis
Implemented frequency based cost/saving analysis
and related options.
The pass is now in a state ready to be turne on
in the pipeline (in follow up).
Differential Revision: http://reviews.llvm.org/D32783
llvm-svn: 302967
Richard Smith [Fri, 12 May 2017 23:40:52 +0000 (23:40 +0000)]
Remove unused tracking of owning module for MacroInfo objects.
llvm-svn: 302966
Richard Smith [Fri, 12 May 2017 23:27:00 +0000 (23:27 +0000)]
[modules] When creating a declaration, cache its owning module immediately
rather than waiting until it's queried.
Currently this is only applied to local submodule visibility mode, as we don't
yet allocate storage for the owning module in non-local-visibility modules
compilations.
llvm-svn: 302965
Aditya Nandakumar [Fri, 12 May 2017 22:54:52 +0000 (22:54 +0000)]
[GISel]: Add a getConstantFPVRegVal utility
This might be useful across various GISel Passes
https://reviews.llvm.org/D33051
llvm-svn: 302964
Aditya Nandakumar [Fri, 12 May 2017 22:43:58 +0000 (22:43 +0000)]
[GISel]: Fix undefined behavior while accessing DefaultAction map
We end up dereferencing the end iterator here when the Aspect doesn't exist in the DefaultAction map.
Change the API to return Optional<LLT> and return None when not found.
Also update the callers to handle the None case
llvm-svn: 302963
Eugene Zelenko [Fri, 12 May 2017 22:26:42 +0000 (22:26 +0000)]
[XRay] Fix build with libc++ (NFC).
llvm-svn: 302962
Eugene Zelenko [Fri, 12 May 2017 22:25:07 +0000 (22:25 +0000)]
[IR] Fix some Clang-tidy modernize-use-using warnings; other minor fixes (NFC).
llvm-svn: 302961
Andrew Kaylor [Fri, 12 May 2017 22:11:26 +0000 (22:11 +0000)]
[TLI] Add mapping for various '__<func>_finite' forms of the math routines to SVML routines
Patch by Chris Chrulski
Differential Revision: https://reviews.llvm.org/D31789
llvm-svn: 302957
Andrew Kaylor [Fri, 12 May 2017 22:11:20 +0000 (22:11 +0000)]
[ConstantFolding] Add folding for various math '__<func>_finite' routines generated from -ffast-math
Patch by Chris Chrulski
Differential Revision: https://reviews.llvm.org/D31788
llvm-svn: 302956
Andrew Kaylor [Fri, 12 May 2017 22:11:12 +0000 (22:11 +0000)]
[TLI] Add declarations for various math header file routines from math-finite.h that create '__<func>_finite as functions
Patch by Chris Chrulski
Differential Revision: https://reviews.llvm.org/D31787
llvm-svn: 302955
Sean Callanan [Fri, 12 May 2017 21:53:44 +0000 (21:53 +0000)]
Fixed the OS X build after Error -> Status rename.
llvm-svn: 302954
Craig Topper [Fri, 12 May 2017 21:45:50 +0000 (21:45 +0000)]
[APInt] Add early outs for a division by 1 to udiv/urem/udivrem
We already counted the number of bits in the RHS so its pretty cheap to just check if the RHS is 1.
Differential Revision: https://reviews.llvm.org/D33154
llvm-svn: 302953
Craig Topper [Fri, 12 May 2017 21:45:44 +0000 (21:45 +0000)]
[APInt] In udivrem, remember the bit width in a local variable so we don't reread it from the LHS which might be aliased with Quotient or Remainder.
This helped the compiler generate better code for the single word case. It was able to remember that the bit width was still a single word when it created the Remainder APInt and not create code for it possibly being multiword.
llvm-svn: 302952
Adrian Prantl [Fri, 12 May 2017 21:38:32 +0000 (21:38 +0000)]
LTO: Don't verify modules twice in verifyMergedModuleOnce
Differential Revision: https://reviews.llvm.org/D33140
llvm-svn: 302951
Sanjay Patel [Fri, 12 May 2017 21:30:31 +0000 (21:30 +0000)]
[Doc] Document "Splat" in the lexicon
Patch by Wei-Ren Chen (陳韋任)!
Differential Revision: https://reviews.llvm.org/D32964
llvm-svn: 302950
Sanjay Patel [Fri, 12 May 2017 20:53:48 +0000 (20:53 +0000)]
[x86] add vector tests for demanded bits; NFC
llvm-svn: 302949
Eugene Zemtsov [Fri, 12 May 2017 20:44:09 +0000 (20:44 +0000)]
Fix build on Mac OS.
llvm-svn: 302948
Richard Smith [Fri, 12 May 2017 20:42:54 +0000 (20:42 +0000)]
Revert r302932, as it appears to be breaking stage2 for some of our modules-enabled buildbots.
llvm-svn: 302947
Casey Carter [Fri, 12 May 2017 20:33:46 +0000 (20:33 +0000)]
[test] Add specific test for P0138R2, direct-list-init of fixed enums from integers, part 3/3.
Test the C++17 Core Language feature independently from other std::byte tests.
llvm-svn: 302946
Casey Carter [Fri, 12 May 2017 20:33:41 +0000 (20:33 +0000)]
[test] Avoid P0138R2, direct-list-init of fixed enums from integers, part 2/3.
This C++17 Core Language feature isn't necessary when testing std::byte.
It's a minor convenience, but it limits test coverage to very new compilers.
This part activates the tests for more compilers.
llvm-svn: 302945
Casey Carter [Fri, 12 May 2017 20:33:32 +0000 (20:33 +0000)]
[test] Avoid P0138R2, direct-list-init of fixed enums from integers, part 1/3.
This C++17 Core Language feature isn't necessary when testing std::byte.
It's a minor convenience, but it limits test coverage to very new compilers.
This part changes the code.
Fixes D32386.
llvm-svn: 302944
Changpeng Fang [Fri, 12 May 2017 20:31:12 +0000 (20:31 +0000)]
AMDGPU/SI: Don't promote to vector if the load/store is volatile.
Summary:
We should not change volatile loads/stores in promoting alloca to vector.
Reviewers:
arsenm
Differential Revision:
http://reviews.llvm.org/D33107
llvm-svn: 302943
Simon Pilgrim [Fri, 12 May 2017 19:56:43 +0000 (19:56 +0000)]
[NVPTX] Don't flag StoreRetVal memory chain operands as ReadMem (PR32146)
This fixes 47 of the 75 NVPTX '-verify-machineinstrs with EXPENSIVE_CHECKS' errors in PR32146.
Differential Revision: https://reviews.llvm.org/D33147
llvm-svn: 302942
Simon Dardis [Fri, 12 May 2017 19:55:32 +0000 (19:55 +0000)]
[Sema] Silence buildbot failures introduced by r302935
Attempt to silence buildbot failures by pinning the test to a given
triple rather than the host's triple.
llvm-svn: 302941
Teresa Johnson [Fri, 12 May 2017 19:32:17 +0000 (19:32 +0000)]
Remove ignore-empty-index-file option
Summary:
Clang changes to remove this option and replace with a parameter
always set in the context of a ThinLTO distributed backend.
Depends on D33133.
Reviewers: pcc
Subscribers: mehdi_amini, eraman, cfe-commits
Differential Revision: https://reviews.llvm.org/D33134
llvm-svn: 302940
Teresa Johnson [Fri, 12 May 2017 19:32:11 +0000 (19:32 +0000)]
Remove ignore-empty-index-file option
Summary:
As discussed in the D32195 review thread and on IRC, remove this option
and replace with parameter, which will be set to true when invoked
from clang in the context of a ThinLTO distributed backend.
Reviewers: pcc
Subscribers: mehdi_amini, llvm-commits
Differential Revision: https://reviews.llvm.org/D33133
llvm-svn: 302939
Dehao Chen [Fri, 12 May 2017 19:29:27 +0000 (19:29 +0000)]
Add LiveRangeShrink pass to shrink live range within BB.
Summary: LiveRangeShrink pass moves instruction right after the definition with the same BB if the instruction and its operands all have more than one use. This pass is inexpensive and guarantees optimal live-range within BB.
Reviewers: davidxl, wmi, hfinkel, MatzeB, andreadb
Reviewed By: MatzeB, andreadb
Subscribers: hiraditya, jyknight, sanjoy, skatkov, gberry, jholewinski, qcolombet, javed.absar, krytarowski, atrick, spatel, RKSimon, andreadb, MatzeB, mehdi_amini, mgorny, efriedma, davide, dberlin, llvm-commits
Differential Revision: https://reviews.llvm.org/D32563
llvm-svn: 302938
Tim Shen [Fri, 12 May 2017 19:25:37 +0000 (19:25 +0000)]
[PPC] Move the combine "a << (b % (sizeof(a) * 8)) -> (PPCshl a, b)" to the backend. NFC.
Summary:
Eli pointed out that it's unsafe to combine the shifts to ISD::SHL etc.,
because those are not defined for b > sizeof(a) * 8, even after some of
the combiners run.
However, PPCISD::SHL defines that behavior (as the instructions themselves).
Move the combination to the backend.
The tests in shift_mask.ll still pass.
Reviewers: echristo, hfinkel, efriedma, iteratee
Subscribers: nemanjai, llvm-commits
Differential Revision: https://reviews.llvm.org/D33076
llvm-svn: 302937
Zachary Turner [Fri, 12 May 2017 19:18:12 +0000 (19:18 +0000)]
[CodeView] Add a random access type visitor.
This adds a visitor that is capable of accessing type
records randomly and caching intermediate results that it
learns about during partial linear scans. This yields
amortized O(1) access to a type stream even though type
streams cannot normally be indexed.
Differential Revision: https://reviews.llvm.org/D33009
llvm-svn: 302936
Simon Dardis [Fri, 12 May 2017 19:11:06 +0000 (19:11 +0000)]
[Sema] Support implicit scalar to vector conversions
This patch teaches clang to perform implicit scalar to vector conversions
when one of the operands of a binary vector expression is a scalar which
can be converted to the element type of the vector without truncation
following GCC's implementation.
If the (constant) scalar is can be casted safely, it is implicitly casted to the
vector elements type and splatted to produce a vector of the same type.
Contributions from: Petar Jovanovic
Reviewers: bruno, vkalintiris
Differential Revision: https://reviews.llvm.org/D25866
llvm-svn: 302935
Haojian Wu [Fri, 12 May 2017 19:01:02 +0000 (19:01 +0000)]
[include-fixer] Don't throw exception when parsing unknown arguments in
vim script.
llvm-svn: 302934
Geoff Berry [Fri, 12 May 2017 18:57:10 +0000 (18:57 +0000)]
[AArch64][Falkor] Refine modeling of multiply accumulate forwarding.
llvm-svn: 302933
Richard Smith [Fri, 12 May 2017 18:56:03 +0000 (18:56 +0000)]
[modules] Simplify module macro handling in non-local-submodule-visibility mode.
When reaching the end of a module, we used to convert its macros to
ModuleMacros but also leave them in the MacroDirective chain for the
identifier. This meant that every lookup of such a macro would find two
(identical) definitions. It also made it difficult to determine the correct
owner for a macro when reaching the end of a module: the most recent
MacroDirective in the chain could be from an #included submodule rather than
the current module.
Simplify this: whenever we convert a MacroDirective to a ModuleMacro when
leaving a module, clear out the MacroDirective chain for that identifier, and
just rely on the ModuleMacro to provide the macro definition information.
(We don't want to do this for local submodule visibility mode, because in that
mode we maintain a distinct MacroDirective chain for each submodule, and we
need to keep around the prior MacroDirective in case we re-enter the submodule
-- for instance, if its header is #included more than once in a module build,
we need the include guard directive to stick around. But the problem doesn't
arise in this case for the same reason: each submodule has its own
MacroDirective chain, so the macros don't leak out of submodules in the first
place.)
llvm-svn: 302932
Craig Topper [Fri, 12 May 2017 18:19:01 +0000 (18:19 +0000)]
[APInt] Add an assert to check for divide by zero in udivrem. NFC
udiv and urem already had the same assert.
llvm-svn: 302931
Craig Topper [Fri, 12 May 2017 18:18:57 +0000 (18:18 +0000)]
[APInt] Remove unnecessary checks of rhsWords==1 with lhsWords==1 from udiv and udivrem. NFC
At this point in the code rhsWords is guaranteed to be non-zero and less than or equal to lhsWords. So if lhsWords is 1, rhsWords must also be 1. urem alread had the check removed so this makes all 3 consistent.
llvm-svn: 302930
Jonathan Peyton [Fri, 12 May 2017 18:01:32 +0000 (18:01 +0000)]
Clang-format and whitespace cleanup of source code
This patch contains the clang-format and cleanup of the entire code base. Some
of clang-formats changes made the code look worse in places. A best effort was
made to resolve the bulk of these problems, but many remain. Most of the
problems were mangling line-breaks and tabbing of comments.
Patch by Terry Wilmarth
Differential Revision: https://reviews.llvm.org/D32659
llvm-svn: 302929
Tom Stellard [Fri, 12 May 2017 17:59:30 +0000 (17:59 +0000)]
AMDGPU: Add lit.local.cfg to disable global-isel tests when global-isel is disabled
This should fix bots broken by r302919.
llvm-svn: 302928
Simon Pilgrim [Fri, 12 May 2017 17:42:36 +0000 (17:42 +0000)]
Strip trailing whitespace. NFCI.
llvm-svn: 302927
Michael Kruse [Fri, 12 May 2017 17:23:07 +0000 (17:23 +0000)]
[Simplify] Reset all stats between runs.
llvm-svn: 302926
Craig Topper [Fri, 12 May 2017 17:20:30 +0000 (17:20 +0000)]
[KnownBits] Add bit counting methods to KnownBits struct and use them where possible
This patch adds min/max population count, leading/trailing zero/one bit counting methods.
The min methods return answers based on bits that are known without considering unknown bits. The max methods give answers taking into account the largest count that unknown bits could give.
Differential Revision: https://reviews.llvm.org/D32931
llvm-svn: 302925
Reid Kleckner [Fri, 12 May 2017 17:06:16 +0000 (17:06 +0000)]
[ubsan] Enable debug info in test binaries
This fixes tests that use debug info to check ubsan stack traces. One
was XFAILd on Windows and the other was actively failing for weeks.
llvm-svn: 302924
Reid Kleckner [Fri, 12 May 2017 17:02:40 +0000 (17:02 +0000)]
[codeview] Fix assertion failure introduced in r295354 refactoring
CodeViewDebug sets Asm to nullptr to disable debug info generation. You
can get a .ll file like no-cus.ll from 'clang -gcodeview -g0', which
happens in the ubsan test suite.
llvm-svn: 302923
Galina Kistanova [Fri, 12 May 2017 17:00:13 +0000 (17:00 +0000)]
Make sure we have actually written what is expected by the test.
llvm-svn: 302922
George Rimar [Fri, 12 May 2017 17:00:07 +0000 (17:00 +0000)]
[ELF] - Apply clang-format. NFC.
llvm-svn: 302921
Francis Ricci [Fri, 12 May 2017 16:52:19 +0000 (16:52 +0000)]
Disable two failing darwin lsan tests
These tests don't fail consistently in all cases, but they
fail most of the time on the buildbots. Mark as UNSUPPORTED for now to
avoid buildbots failing due to XPASS.
llvm-svn: 302920
Tom Stellard [Fri, 12 May 2017 16:46:46 +0000 (16:46 +0000)]
AMDGPU/GlobalISel: Mark 32-bit integer constants as legal
Reviewers: arsenm
Reviewed By: arsenm
Subscribers: kzhuravl, wdng, nhaehnle, yaxunl, rovka, kristof.beyls, igorb, dstuttard, tpr, t-tye, llvm-commits
Differential Revision: https://reviews.llvm.org/D33115
llvm-svn: 302919
Alex Lorenz [Fri, 12 May 2017 16:32:26 +0000 (16:32 +0000)]
[index] Index template specialization arguments for function templats
Also ensure that class template specialization arguments are covered
rdar://
31812032
llvm-svn: 302918
Reid Kleckner [Fri, 12 May 2017 16:30:56 +0000 (16:30 +0000)]
[asan/win] Re-enable Win64 asan tests on Win8+
Our theory is that reserving large amounts of shadow memory isn't
reliable on Win7 and earlier NT kernels. This affects the
clang-x64-ninja-win7 buildbot, which uses Windows 7.
llvm-svn: 302917
Craig Topper [Fri, 12 May 2017 16:28:21 +0000 (16:28 +0000)]
[APInt] Use MathExtras.h BitsToFloat/Double and Float/DoubleToBits instead of type punning through a union
The functions in MathExtras.h uses a safer memcpy instead of going through a union.
Differential Revision: https://reviews.llvm.org/D33116
llvm-svn: 302916
Adrian Prantl [Fri, 12 May 2017 16:23:53 +0000 (16:23 +0000)]
Simplify DINamespace caching in CGDebugInfo
This addresses review feedback from r302840.
By not canonicalizing namespace decls and using lexical decl context
instead of lookuing up the semantic decl context we can take advantage
of the fact that DINamespaces a reuniqued. This way non-module debug
info is unchanged and module debug info still gets distinct namespace
declarations when they ocur in different modules.
Thanks to Richard Smith for pointing this out!
llvm-svn: 302915
James Y Knight [Fri, 12 May 2017 16:01:23 +0000 (16:01 +0000)]
[SPARC] Support 'f' and 'e' inline asm constraints.
Patch by Patrick Boettcher.
Differential Revision: https://reviews.llvm.org/D29117
llvm-svn: 302913
Francis Ricci [Fri, 12 May 2017 16:01:15 +0000 (16:01 +0000)]
Disable two failing darwin lsan tests
These are causing buildbot failures, disable for now.
llvm-svn: 302912
James Y Knight [Fri, 12 May 2017 15:59:10 +0000 (15:59 +0000)]
[SPARC] Support 'f' and 'e' inline asm constraints.
Based on patch by Patrick Boettcher and Chris Dewhurst.
Differential Revision: https://reviews.llvm.org/D29116
llvm-svn: 302911
Sanjay Patel [Fri, 12 May 2017 15:56:39 +0000 (15:56 +0000)]
[x86] add tests for potential vector narrowing optimization (PR32790)
llvm-svn: 302910