Commit Graph

245465 Commits

Author SHA1 Message Date
Konstantin Zhuravlyov fda33eaf0c [AMDGPU] Perform uchar to float combine for ISD::SINT_TO_FP
This will prevent following regression when enabling i16 support (D18049):
  test/CodeGen/AMDGPU/cvt_f32_ubyte.ll

Differential Revision: https://reviews.llvm.org/D25805

llvm-svn: 284891
2016-10-21 22:10:03 +00:00
Richard Smith 5e9746f520 DR583, DR1512: Implement a rewrite to C++'s 'composite pointer type' rules.
This has two significant effects:

1) Direct relational comparisons between null pointer constants (0 and nullopt)
   and pointers are now ill-formed. This was always the case for C, and it
   appears that C++ only ever permitted by accident. For instance, cases like
     nullptr < &a
   are now rejected.

2) Comparisons and conditional operators between differently-cv-qualified
   pointer types now work, and produce a composite type that both source
   pointer types can convert to (when possible). For instance, comparison
   between 'int **' and 'const int **' is now valid, and uses an intermediate
   type of 'const int *const *'.

Clang previously supported #2 as an extension.

We do not accept the cases in #1 as an extension. I've tested a fair amount of
code to check that this doesn't break it, but if it turns out that someone is
relying on this, we can easily add it back as an extension.

This is a re-commit of r284800.

llvm-svn: 284890
2016-10-21 22:00:42 +00:00
Richard Smith 00c03c5de7 Remove 'misc-pointer-and-integral-operation' clang-tidy check. The only cases
it detects are ill-formed (some per C++ core issue 1512, others always have
been).

llvm-svn: 284888
2016-10-21 21:50:28 +00:00
Justin Lebar 5e83dfedb8 Switch SmallSetVector to use DenseSet when it overflows its inline space.
Summary:
SetVector already used DenseSet, but SmallSetVector used std::set.  This
leads to surprising performance differences.  Moreover, it means that
the set of key types accepted by SetVector and SmallSetVector are
quite different!

In order to make this change, we had to convert some callsites that used
SmallSetVector<std::string, N> to use SmallSetVector<CachedHashString, N>
instead.

Reviewers: timshen

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D25648

llvm-svn: 284887
2016-10-21 21:45:01 +00:00
Richard Smith 104887ed08 Fix typo (ordered comparison between pointer and 0).
llvm-svn: 284886
2016-10-21 21:37:18 +00:00
Malcolm Parsons 57c09c8e23 [Sema] Store a SourceRange for multi-token builtin types
Summary:
clang-tidy's modernize-use-auto check uses the SourceRange of a
TypeLoc when replacing the type with auto.
This was producing the wrong result for multi-token builtin types
like long long:

-long long *ll = new long long();
+auto long *ll = new long long();

Reviewers: alexfh, hokein, rsmith, Prazek, aaron.ballman

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D25363

llvm-svn: 284885
2016-10-21 21:13:56 +00:00
Justin Bogner 8659d16631 Support: Annotate Error and Expected<> with LLVM_NODISCARD
It's always a mistake to ignore return values with Error or Expected
types.

llvm-svn: 284884
2016-10-21 21:01:12 +00:00
Mandeep Singh Grang 3642dd3ca4 [polly] Change SmallPtrSet which is being iterated to SmallSetVector in ScopInfo.h
Summary: This will avoid non-deterministic iteration order.

Reviewers: grosser, jdoerfert, zinob, mgrang

Subscribers: #polly

Tags: #polly

Differential Revision: https://reviews.llvm.org/D25880

llvm-svn: 284883
2016-10-21 21:00:11 +00:00
Justin Lebar 4d38a5cf74 [CUDA] Simplify some repeated diagnostic expectations in CUDA tests.
Instead of repeating the diagnostic, use "expected-note N".

Test-only change.

llvm-svn: 284882
2016-10-21 20:50:47 +00:00
Tim Shen abd1a6e687 [libcxx] Support std::regex_constants::match_not_null
Summary: Fixes PR21597.

Reviewers: mclow.lists, EricWF

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D25595

llvm-svn: 284881
2016-10-21 20:41:47 +00:00
Bob Haarman 62c730c9bd fixed typo in InstrProf.h; NFC
llvm-svn: 284880
2016-10-21 20:38:37 +00:00
Artem Belevich 07db5cf6c8 Declare H and H new/delete.
llvm-svn: 284879
2016-10-21 20:34:05 +00:00
Rafael Espindola ca656239eb Compact SymbolBody from 56 to 48 bytes.
llvm-svn: 284878
2016-10-21 20:32:41 +00:00
Vassil Vassilev f054b8c378 Remove accidentally checked in assert.
Thanks to Manman for spotting this.

llvm-svn: 284877
2016-10-21 20:30:30 +00:00
Justin Lebar 3c37dc0cb2 [ADT] Get rid of use of LLVM_NOEXCEPT in CachedHashString.h.
LLVM_NOEXCEPT is no longer necessary (yay).

llvm-svn: 284876
2016-10-21 20:28:00 +00:00
Tom Stellard 6c7dd980e4 AMDGPU/SI: Fix crash caused by r284267
Reviewers: arsenm, nhaehnle

Subscribers: kzhuravl, wdng, nhaehnle, yaxunl, tony-tye, llvm-commits

Differential Revision: https://reviews.llvm.org/D25782

llvm-svn: 284875
2016-10-21 20:25:11 +00:00
Sanjay Patel ca92c36e01 [DAG] enhance computeKnownBits to handle SHL with vector splat constant
Also, use APInt to avoid crashing on types larger than vNi64.

llvm-svn: 284874
2016-10-21 20:16:27 +00:00
Justin Lebar ae90ad2065 [clang-tidy] Don't use a SmallSetVector of an enum.
Summary:
This doesn't work after converting SmallSetVector to use DenseSet.

Instead we can just use a SmallVector.

Reviewers: timshen

Subscribers: nemanjai, cfe-commits

Differential Revision: https://reviews.llvm.org/D25647

llvm-svn: 284873
2016-10-21 20:13:39 +00:00
Justin Lebar c9d66eb05b [ADT] Compare strings' hashes first before comparing their values.
Summary:
We already have the hashes in hand, and comparing hashes should be much
more discriminatory than comparing the StringRefs' sizes.

Reviewers: rafael

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D25705

llvm-svn: 284872
2016-10-21 20:10:51 +00:00
Justin Lebar d87ea9a1c9 [ADT] Add CachedHashString.
Summary:
This is like CachedHashStringRef, but owns its data.

This lets us use strings inside of DenseMaps.

Reviewers: timshen

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D25645

llvm-svn: 284871
2016-10-21 20:10:48 +00:00
Justin Lebar 245c3e75cd [IR] Add DenseMapInfo<CallSite>.
Summary:
A CallSite is basically an Instruction*, and you can put Instruction*s
into DenseMaps, so you should be able to do the same with CallSites.

This is used in a later patch.

Reviewers: timshen

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D25643

llvm-svn: 284870
2016-10-21 20:10:44 +00:00
Justin Lebar 6f72737c3b [CUDA] Use FunctionDeclAndLoc for the Sema::LocsWithCUDACallDiags hashtable.
Summary: NFC

Reviewers: rnk

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D25797

llvm-svn: 284869
2016-10-21 20:08:52 +00:00
Li Huang faa857dba7 [SCEV] Memoize visitMulExpr results in SCEVRewriteVisitor.
Summary:
When SCEVRewriteVisitor traverses the SCEV DAG, it may visit the same SCEV
multiple times if this SCEV is referenced by multiple other SCEVs. This has
exponential time complexity in the worst case. Memoizing the results will
avoid re-visiting the same SCEV. Add a map to save the results, and override
the visit function of SCEVVisitor. Now SCEVRewriteVisitor only visit each
SCEV once and thus returns the same result for the same input SCEV.

This patch fixes PR18606, PR18607.

Reviewers: Sanjoy Das, Mehdi Amini, Michael Zolotukhin

Differential Revision: https://reviews.llvm.org/D25810

llvm-svn: 284868
2016-10-21 20:05:21 +00:00
Kevin Enderby 2f9d8d0702 Fix a bug in the code of llvm-cxxdump in dumpArchive() when
iterating over an archive with object and non-object members that
would cause an Abort because to was not calling consumeError()
when the code was wanting to ignore a non-object file.

Found by Justin Bogner!

llvm-svn: 284867
2016-10-21 20:03:14 +00:00
Kostya Serebryany 42909a6f3a [libFuzzer] mention one more trophie
llvm-svn: 284866
2016-10-21 20:01:45 +00:00
Peter Collingbourne ecdd58f1d6 Analysis: Move llvm::getConstantRangeFromMetadata to IR library.
We're about to start using it there.

Differential Revision: https://reviews.llvm.org/D25877

llvm-svn: 284865
2016-10-21 19:59:26 +00:00
Peter Collingbourne e9bd49824d X86: Improve BT instruction selection for 64-bit values.
If a 64-bit value is tested against a bit which is known to be in the range
[0..31) (modulo 64), we can use the 32-bit BT instruction, which has a slightly
shorter encoding.

Differential Revision: https://reviews.llvm.org/D25862

llvm-svn: 284864
2016-10-21 19:57:55 +00:00
Simon Pilgrim ab48872313 [X86][AVX512BWVL] Added support for lowering v16i16 shuffles to AVX512BWVL vpermw
llvm-svn: 284863
2016-10-21 19:54:38 +00:00
Rafael Espindola 8b2c8536e5 Use a CachedHashString for comdats too.
We were already using it in the larger sets/maps. This provides about
1% speedup in linking xul and chromium.

llvm-svn: 284862
2016-10-21 19:49:42 +00:00
Bob Haarman 653baa2aaa [pdb] added support for dumping globals stream
Summary: This adds support for dumping the globals stream from PDB files using llvm-pdbdump, similar to the support we have for the publics stream.

Reviewers: ruiu, zturner

Subscribers: beanz, mgorny, modocache

Differential Revision: https://reviews.llvm.org/D25801

llvm-svn: 284861
2016-10-21 19:43:19 +00:00
Simon Pilgrim da814cba0d [X86][AVX512BWVL] Added support for combining target v16i16 shuffles to AVX512BWVL vpermw
llvm-svn: 284860
2016-10-21 19:40:29 +00:00
Haojian Wu 29c38f75a2 [clang-move] Support moving template class forward declarations.
llvm-svn: 284859
2016-10-21 19:26:43 +00:00
Simon Pilgrim 0109bf116f [X86][AVX512] Added support for combining target shuffles to AVX512 vpermpd/vpermq/vpermps/vpermd/vpermw
llvm-svn: 284858
2016-10-21 19:18:09 +00:00
Krzysztof Parzyszek 6e7fa99d3a [RDF] Use RegisterId typedef more consistently, NFC
llvm-svn: 284857
2016-10-21 19:12:13 +00:00
Benjamin Kramer 33e9760376 Remove move constructors that are identical to the generated default move ctor.
llvm-svn: 284856
2016-10-21 18:55:07 +00:00
Anna Thomas 0860259434 [StripGCRelocates] New pass to remove gc.relocates added by RS4GC
Summary:
Utility pass to remove gc.relocates created by rewrite statepoints for GC.
With respect to safepoint verification, the IR generated would be incorrect, and cannot run
as such.

This would be a single transformation on the final optimized IR.
The benefit of the pass is for easy analysis when the IRs are 'polluted' by too
many gc.relocates.
Added tests.

test run: All RS4GC tests with -verify option. Local downstream tests on large
IR files. This also works when the pointer being gc.relocated is another
gc.relocate.

Reviewers: sanjoy, reames

Subscribers: beanz, mgorny, llvm-commits

Differential Revision: https://reviews.llvm.org/D25096

llvm-svn: 284855
2016-10-21 18:43:16 +00:00
Chris Bieneman 1bd5f52e68 [CMake] Fix standalone build
not and FileCheck targets may not be defined in standalone builds.

llvm-svn: 284854
2016-10-21 18:38:44 +00:00
Kevin Enderby 41c9c00bf0 For llvm-objdump for Mach-O files add printing of
the ARM_THREAD_STATE in the same format as
otool-classic(1) on darwin.

Also remove an extra space in printing the initprot to make
the output match otool-classic(1) on darwin.

rdar://28851457

llvm-svn: 284852
2016-10-21 18:22:35 +00:00
Jim Ingham e3b7ba2ffe Add the new stdcpp formatters to the Xcode project.
llvm-svn: 284851
2016-10-21 18:18:25 +00:00
Eli Friedman 286c5a76ba [SCEVAffinator] Make precise modular math more correct.
Integer math in LLVM IR is modular. Integer math in isl is
arbitrary-precision. Modeling LLVM IR math correctly in isl requires
either adding assumptions that math doesn't actually overflow, or
explicitly wrapping the math. However, expressions with the "nsw" flag
are special; we can pretend they're arbitrary-precision because it's
undefined behavior if the result wraps. SCEV expressions based on IR
instructions with an nsw flag also carry an nsw flag (roughly; actually,
the real rule is a bit more complicated, but the details don't matter
here).

Before this patch, SCEV flags were also overloaded with an additional
function: the ZExt code was mutating SCEV expressions as a hack to
indicate to checkForWrapping that we don't need to add assumptions to
the operand of a ZExt; it'll add explicit wrapping itself. This kind of
works... the problem is that if anything else ever touches that SCEV
expression, it'll get confused by the incorrect flags.

Instead, with this patch, we make the decision about whether to
explicitly wrap the math a bit earlier, basing the decision purely on
the SCEV expression itself, and not its users.

Differential Revision: https://reviews.llvm.org/D25287

llvm-svn: 284848
2016-10-21 18:08:02 +00:00
Enrico Granata 5a75a0d8f5 I hadn't fixed all the instances of the old marker for scripted format strings; do so now
llvm-svn: 284847
2016-10-21 18:03:55 +00:00
Rui Ueyama fe4c9f96c6 Use FileCheck's -strict-whitespace option instead of sed.
llvm-svn: 284846
2016-10-21 17:50:26 +00:00
Mandeep Singh Grang 48e7add80f [polly] Change SmallPtrSet which are being iterated into SmallSetVector
Summary: Otherwise the lack of an iteration order results in non-determinism in codegen.

Reviewers: _jdoerfert, zinob, grosser

Tags: #polly

Differential Revision: https://reviews.llvm.org/D25863

llvm-svn: 284845
2016-10-21 17:29:10 +00:00
Sanjay Patel 81029f6a76 [DAG] fold negation of sign-bit
0 - X --> 0, if the sub is NUW
0 - X --> 0, if X is 0 or the minimum signed value and the sub is NSW
0 - X --> X, if X is 0 or the minimum signed value

This is the DAG equivalent of:
https://reviews.llvm.org/rL284649

plus the fold for the NUW case which already existed in InstSimplify.

Note that we miss a vector fold because of a deficiency in the DAG version of
computeKnownBits().

llvm-svn: 284844
2016-10-21 17:24:26 +00:00
Artem Belevich 9fb40e3bbd Removed unused function argument. NFC.
Differential Revision: https://reviews.llvm.org/D25839

llvm-svn: 284843
2016-10-21 17:15:46 +00:00
Sanjay Patel 97de8a4c3a [x86] add tests for potential negation folds
These are the backend equivalents for the tests added in r284627.
The patterns may emerge late, so we should have folds for these in the DAG too.

llvm-svn: 284842
2016-10-21 16:56:29 +00:00
Krzysztof Parzyszek b71085b547 [Hexagon] Handle spills of partially defined double vector registers
After register allocation it is possible to have a spill of a register
that is only partially defined. That in itself it fine, but creates a
problem for double vector registers. Stores of such registers are pseudo
instructions that are expanded into pairs of individual vector stores,
and in case of a partially defined source, one of the stores may use
an entirely undefined register. To avoid this, track the defined parts
and only generate actual stores for those.

llvm-svn: 284841
2016-10-21 16:38:29 +00:00
Derek Schuff 6f69783f1f [WebAssembly] Fix for 0xc call_indirect changes
Summary:
Need to reorder the operands to have the callee as the last argument.
Adds a pseudo-instruction, and a pass to lower it into a real
call_indirect.

This is the first of two options for how to fix the problem.

Reviewers: dschuff, sunfish

Subscribers: jfb, beanz, mgorny, llvm-commits

Differential Revision: https://reviews.llvm.org/D25708

llvm-svn: 284840
2016-10-21 16:38:07 +00:00
Abderrazek Zaafrani 9daf8110c8 Set the vectorizer MaxInterleaveFactor for Exynos.
llvm-svn: 284839
2016-10-21 16:28:27 +00:00
Reid Kleckner ac2a2a86e4 Fix -Wunused-variable warning in libFuzzer
llvm-svn: 284838
2016-10-21 16:26:27 +00:00