Jonas Hahnfeld [Mon, 7 Nov 2016 15:58:36 +0000 (15:58 +0000)]
[OpenMP] Enable ThreadSanitizer to check OpenMP programs
This patch allows ThreadSanitizer (Tsan) to verify OpenMP programs.
It means that no false positive will be reported by Tsan when
verifying an OpenMP programs.
This patch introduces annotations within the OpenMP runtime module to
provide information about thread synchronization to the Tsan runtime.
In order to enable the Tsan support when building the runtime, you must
enable the TSAN_SUPPORT option with the following environment variable:
-DLIBOMP_TSAN_SUPPORT=TRUE
The annotations will be enabled in the main shared library
(same mechanism of OMPT).
Patch by Simone Atzeni and Joachim Protze!
Differential Revision: https://reviews.llvm.org/D13072
llvm-svn: 286115
Yaxun Liu [Mon, 7 Nov 2016 15:55:51 +0000 (15:55 +0000)]
[OpenCL] Remove redundant test for OpenCL header file
Differential Revision: https://reviews.llvm.org/D26302
llvm-svn: 286114
Sanjay Patel [Mon, 7 Nov 2016 15:52:45 +0000 (15:52 +0000)]
[InstCombine] allow splat vector folds in adjustMinMax() (retry r285732)
This was reverted at r285866 because there was a crash handling a scalar
select of vectors. I added a check for that pattern and a test case based
on the example provided in the post-commit thread for r285732.
llvm-svn: 286113
Jonas Paulsson [Mon, 7 Nov 2016 15:45:06 +0000 (15:45 +0000)]
[SystemZ] Correct the SchedModel regarding vector unit / instructions.
* Use a generic vector unit to model the issue unit more accurately.
* Update some vector instructions that actually use the vector unit for more
than one cycle.
Review: Ulrich Weigand
llvm-svn: 286112
Amara Emerson [Mon, 7 Nov 2016 15:42:12 +0000 (15:42 +0000)]
This patch adds support for 16 bit floating point registers to the inline asm register selection on AArch64.
Without this patch, register allocation for the example below fails.
define half @test(half %a1, half %a2) #0 {
entry:
%0 = tail call half asm "sqrshl ${0:h}, ${1:h}, ${2:h}", "=w,w,w" (half %a1, half %a2) #1
ret half %0
}
Patch by Florian Hahn.
Differential Revision: https://reviews.llvm.org/D25080
llvm-svn: 286111
Chad Rosier [Mon, 7 Nov 2016 15:27:22 +0000 (15:27 +0000)]
[AArch64] Removed the narrow load merging code in the ld/st optimizer.
This feature has been disabled for some time now, so remove cruft.
Differential Revision: https://reviews.llvm.org/D26248
llvm-svn: 286110
Jonas Paulsson [Mon, 7 Nov 2016 14:47:25 +0000 (14:47 +0000)]
[SystemZ] Fixes in SchedModels for older subtargets.
IssueWidth updated to reflect the capacity of the issue unit correctly.
Correct number of FX and LS units modelled (2, was 1).
Review: Ulrich Weigand
llvm-svn: 286109
Chad Rosier [Mon, 7 Nov 2016 14:11:45 +0000 (14:11 +0000)]
[AliasSetTracker] Make AST smarter about assume intrinsics that don't actually affect memory.
Differential Revision: https://reviews.llvm.org/D26252
llvm-svn: 286108
James Molloy [Mon, 7 Nov 2016 13:38:21 +0000 (13:38 +0000)]
[Thumb1] Move padding earlier when synthesizing TBBs off of the PC
When the base register (register pointing to the jump table) is the PC, we expect the jump table to directly follow the jump sequence with no intervening padding.
If there is intervening padding, the calculated offsets will not be correct. One solution would be to account for any padding in the emitted LDRB instruction, but at the moment we don't support emitting MCExprs for the load offset.
In the meantime, it's correct and only a slight amount worse to just move the padding up, from just before the jump table to just before the jump instruction sequence. We can do that by emitting code alignment before the jump sequence, as we know the number of instructions in the sequence is always 4.
llvm-svn: 286107
Alexey Bataev [Mon, 7 Nov 2016 13:37:34 +0000 (13:37 +0000)]
[OPENMP] Fixed test on MIPS-based buildbots.
llvm-svn: 286106
Simon Pilgrim [Mon, 7 Nov 2016 13:34:29 +0000 (13:34 +0000)]
[X86][AVX512] Add AVX512VL/AVX512BWVL vector truncation tests
llvm-svn: 286105
Simon Pilgrim [Mon, 7 Nov 2016 13:28:20 +0000 (13:28 +0000)]
[X86][SSE] Drop unnecessary -mcpu argument from trunc tests
cpu/triple duplication
llvm-svn: 286104
Alexey Bataev [Mon, 7 Nov 2016 11:16:04 +0000 (11:16 +0000)]
[OPENMP] Fixed capturing of VLA variables.
After some changes in codegen capturing of VLA variables in OpenMP regions was broken, causing compiler crash. Patch fixes this issue.
llvm-svn: 286103
Diana Picus [Mon, 7 Nov 2016 10:01:43 +0000 (10:01 +0000)]
Revert "[OPENMP] Fixed capturing of VLA variables."
This reverts commit r286098 because the modified test breaks on many of the
buildbots.
llvm-svn: 286102
Pavel Labath [Mon, 7 Nov 2016 09:58:58 +0000 (09:58 +0000)]
Make space for FreeBSD core file tests
This renames the functionalities/postmortem/linux-core to elf-core and puts the
"linux" part into the individual names of the core files. Since the tests for
linux and freebsd core files are going to be very similar, having them close
together means they can reuse most of the plumbing.
llvm-svn: 286101
Eugene Leviant [Mon, 7 Nov 2016 09:04:06 +0000 (09:04 +0000)]
[ELF] Make InputSection<ELFT>::writeTo virtual
Differential revision: https://reviews.llvm.org/D26281
llvm-svn: 286100
Roger Ferrer Ibanez [Mon, 7 Nov 2016 08:23:59 +0000 (08:23 +0000)]
Protect std::experimental::optional tests under libcpp-no-exceptions
In these tests there are some paths that explicitly throw, so use
the TEST_THROW macro that was proposed for this and then skip the tests
that may enter the throwing path.
Differential Revision: https://reviews.llvm.org/D26142
llvm-svn: 286099
Alexey Bataev [Mon, 7 Nov 2016 08:07:25 +0000 (08:07 +0000)]
[OPENMP] Fixed capturing of VLA variables.
After some changes in codegen capturing of VLA variables in OpenMP
regions was broken, causing compiler crash. Patch fixes this issue.
llvm-svn: 286098
Craig Topper [Mon, 7 Nov 2016 07:01:09 +0000 (07:01 +0000)]
[AVX-512][Sema] Add more intrinsics to the checks for valid immediates for embedded rounding control arguments.
llvm-svn: 286097
Eric Liu [Mon, 7 Nov 2016 06:08:23 +0000 (06:08 +0000)]
Deduplicate replacements by FileEntry instead of file names.
Summary:
The current version does not deduplicate equivalent file paths correctly.
For example, a relative path and an absolute path are considered inequivalent.
Comparing FileEnry addresses these issues.
Reviewers: djasper
Subscribers: alexshap, klimek, cfe-commits
Differential Revision: https://reviews.llvm.org/D26288
llvm-svn: 286096
Dylan McKay [Mon, 7 Nov 2016 06:02:55 +0000 (06:02 +0000)]
[AVR] Enable the ISel, frame analyzer, and alloca passes
llvm-svn: 286095
Brian Gesiak [Mon, 7 Nov 2016 02:43:01 +0000 (02:43 +0000)]
Update CommandLine.rst getRegisteredOptions example
Summary: Update the docs to match the changes in http://reviews.llvm.org/D7132
Reviewers: beanz, llvm-commits, modocache
Differential Revision: https://reviews.llvm.org/D26296
llvm-svn: 286094
Vedant Kumar [Mon, 7 Nov 2016 02:39:37 +0000 (02:39 +0000)]
[lldb] Fix -Waggressive-loop-optimizations warning
We shouldn't access past the end of an array, even if we think that the
layout of the struct containing the array is always what we expect. The
compiler is free to optimize away the stores as undefined behavior, and
in fact, GCC 6.2.1 claims it will do exactly this.
llvm-svn: 286093
Craig Topper [Mon, 7 Nov 2016 02:12:57 +0000 (02:12 +0000)]
[AVX-512] Remove masked pmovzx/pmovsx builtins and autoupgrade them to selects and native zext/sext.
This mostly reuses earlier autoupgrade support for the sse and avx equivalents. Just needed to add the code to add the select.
llvm-svn: 286092
Craig Topper [Mon, 7 Nov 2016 00:13:46 +0000 (00:13 +0000)]
[X86] Remove GCCBuiltins from cvtsi2ss/cvtsi2sd/cvtss2sd intrinsics as they aren't used by clang. Add TODOs to remove these and some other unused intrinsics.
llvm-svn: 286091
Craig Topper [Mon, 7 Nov 2016 00:13:42 +0000 (00:13 +0000)]
[X86] Use StringRef::startswith to reduce a few compares in the intrinsic autoupgrade code.
llvm-svn: 286090
Craig Topper [Mon, 7 Nov 2016 00:13:39 +0000 (00:13 +0000)]
[AVX-512] Remove 128/256 masked pshufb intrinsics. Autoupgrade them to legacy intrinsics and a select.
llvm-svn: 286089
Zachary Turner [Mon, 7 Nov 2016 00:07:25 +0000 (00:07 +0000)]
Convert some helper functions to use StringRef.
I will probably submit a lot of small trivial changes
like this over the coming weeks. The end goal is to
convert Options::SetOptionValue to take the option arg
as a StringRef, but doing so in one pass would be a huge
mess of disparate changes just to satisfy the compiler,
and with a high risk of breaking. So I'm going to do
this in small pieces, changing seemingly random things
here and there until the final change to the signature
of Options::SetOptionValue() can be done trivially.
llvm-svn: 286088
Daniel Jasper [Sun, 6 Nov 2016 23:21:22 +0000 (23:21 +0000)]
Also delete the go side of the bindings deleted in r286085/r286086.
Also delete a comment I forgot to delete.
llvm-svn: 286087
Daniel Jasper [Sun, 6 Nov 2016 21:20:45 +0000 (21:20 +0000)]
Delete one more function that needs to go after r286062.
llvm-svn: 286086
Daniel Jasper [Sun, 6 Nov 2016 21:02:14 +0000 (21:02 +0000)]
Remove functions from go bindings that have been deleted in r286062.
llvm-svn: 286085
Krzysztof Parzyszek [Sun, 6 Nov 2016 20:55:57 +0000 (20:55 +0000)]
Reapply r286080 with a phony change in Hexagon's CMakeLists.txt
Cmake has not recognized that Hexagon.td has a new dependency in
HexagonPatterns.td. All changes to that file were not visible to
the build bots.
llvm-svn: 286084
Davide Italiano [Sun, 6 Nov 2016 20:55:40 +0000 (20:55 +0000)]
[RandomNumberGenerator] Remove MSVC 2013 compatibility hack.
llvm-svn: 286083
Saleem Abdulrasool [Sun, 6 Nov 2016 19:46:54 +0000 (19:46 +0000)]
ARM: lower fpowi appropriately for Windows ARM
This handles the last case of the builtin function calls that we would
generate code which differed from Microsoft's ABI. Rather than
generating a call to `__pow{d,s}i2` we now promote the parameter to a
float or double and invoke `powf` or `pow` instead.
Addresses PR30825!
llvm-svn: 286082
Krzysztof Parzyszek [Sun, 6 Nov 2016 19:36:09 +0000 (19:36 +0000)]
Revert r286080: it breaks build bots
llvm-svn: 286081
Krzysztof Parzyszek [Sun, 6 Nov 2016 19:03:38 +0000 (19:03 +0000)]
[Hexagon] Remove redundant custom selection code
The clr/set/toggle-bit instructions (with the bit index given as an
immediate operand) had both, custom selection code that generated them,
and selection patterns at the same time. The selection patterns were
not used, because the custom selection code was executed first.
This patch removes the custom code in favor of the selection patterns.
The custom code handled 64-bit registers as well with an immediate bit
index, and so new patterns were added to implement that.
It was also the same case for the instruction "Rd += asr(Rs, Rt)",
except that the custom code did not offer any additional functionality,
and was simply removed.
llvm-svn: 286080
Krzysztof Parzyszek [Sun, 6 Nov 2016 18:13:14 +0000 (18:13 +0000)]
[Hexagon] Round 5 of selection pattern simplifications
Remove unnecessary type casts in patterns.
llvm-svn: 286079
Krzysztof Parzyszek [Sun, 6 Nov 2016 18:09:56 +0000 (18:09 +0000)]
[Hexagon] Round 4 of selection pattern simplifications
Give simpler or more meaningful names to pat frags and xforms.
llvm-svn: 286078
Krzysztof Parzyszek [Sun, 6 Nov 2016 18:05:14 +0000 (18:05 +0000)]
[Hexagon] Round 3 of selection pattern simplifications
Remove unnecessary C++ functions for SDNode transforms. Move more
pat frags to files where they are used.
llvm-svn: 286077
Krzysztof Parzyszek [Sun, 6 Nov 2016 17:56:48 +0000 (17:56 +0000)]
[Hexagon] Round 2 of selection pattern simplifications
Add pat frags for any-, sign-, and zero-extensions.
llvm-svn: 286076
Simon Pilgrim [Sun, 6 Nov 2016 16:49:19 +0000 (16:49 +0000)]
[SelectionDAG] Add support for vector demandedelts in XOR opcodes
llvm-svn: 286075
Simon Pilgrim [Sun, 6 Nov 2016 16:36:29 +0000 (16:36 +0000)]
[X86] Add knownbits vector xor test
In preparation for demandedelts support
llvm-svn: 286074
Craig Topper [Sun, 6 Nov 2016 16:29:19 +0000 (16:29 +0000)]
[AVX-512] Remove intrinsics for 128/256-bit masked variable shift. Instead upgrade them to a select and the older AVX2 intrinsic.
llvm-svn: 286073
Craig Topper [Sun, 6 Nov 2016 16:29:14 +0000 (16:29 +0000)]
[AVX-512] Remove intrinsics for 128/256-bit masked shift by immediate. Instead upgrade them to a select and the older SSE/AVX2 intrinsic.
llvm-svn: 286072
Simon Pilgrim [Sun, 6 Nov 2016 16:29:09 +0000 (16:29 +0000)]
[SelectionDAG] Add support for vector demandedelts in OR opcodes
llvm-svn: 286071
Craig Topper [Sun, 6 Nov 2016 16:29:08 +0000 (16:29 +0000)]
[AVX-512] Remove intrinsics for 128/256-bit masked shift by single element in xmm. Instead upgrade them to a select and the older SSE/AVX2 intrinsic.
llvm-svn: 286070
Craig Topper [Sun, 6 Nov 2016 16:29:03 +0000 (16:29 +0000)]
[AVX-512] Remove a 512-bit test cases from the avx512vl test file. It already exists in the avx512f test file.
llvm-svn: 286069
Simon Pilgrim [Sun, 6 Nov 2016 16:05:59 +0000 (16:05 +0000)]
[X86] Add knownbits vector or test
In preparation for demandedelts support
llvm-svn: 286068
Simon Pilgrim [Sun, 6 Nov 2016 14:20:29 +0000 (14:20 +0000)]
[X86][SSE] Reuse zeroable element mask in lowerVectorShuffleAsElementInsertion. NFCI
Don't regenerate a zeroable element mask with computeZeroableShuffleElements when its already available.
llvm-svn: 286067
Benjamin Kramer [Sun, 6 Nov 2016 13:26:39 +0000 (13:26 +0000)]
[BitcodeWriter] Replace a manual byteswap with read32be.
No functional change intended.
llvm-svn: 286066
George Rimar [Sun, 6 Nov 2016 08:39:46 +0000 (08:39 +0000)]
[ELF] - Change argument of computeHash to MutableArrayRef. NFC.
This fixes casting warning and removes the need of that cast at all.
llvm-svn: 286065
George Rimar [Sun, 6 Nov 2016 08:26:53 +0000 (08:26 +0000)]
[ELF] - Fix linux BB after r286061.
llvm-svn: 286064
Craig Topper [Sun, 6 Nov 2016 07:50:25 +0000 (07:50 +0000)]
[X86] Add a few more fptoui test cases to the vec_fp_to_int.ll. The codegen for these test cases will be improved for AVX512 in a future commit.
llvm-svn: 286063
Amaury Sechet [Sun, 6 Nov 2016 07:48:46 +0000 (07:48 +0000)]
Kill deprecated attribute API
Summary:
This kill various depreacated API related to attribute :
- The deprecated C API attribute based on LLVMAttribute enum.
- The Raw attribute set format (planned to be removed in 4.0).
Reviewers: bkramer, echristo, mehdi_amini, void
Subscribers: mehdi_amini, llvm-commits
Differential Revision: https://reviews.llvm.org/D23039
llvm-svn: 286062
George Rimar [Sun, 6 Nov 2016 07:42:55 +0000 (07:42 +0000)]
[ELF] - Implemented threaded --build-id computation
Patch switches computing of --build-id hash to tree.
This is the way when input data is splitted by chunks,
hash is computed for each one in threaded/non-threaded way.
At the end hash is conputed for result tree.
With or without -threads the result hash is the same.
Differential revision: https://reviews.llvm.org/D26199
llvm-svn: 286061
Tim Shen [Sun, 6 Nov 2016 07:38:37 +0000 (07:38 +0000)]
[APFloat] Make functions that produce APFloaat objects use correct semantics.
Summary:
Fixes PR30869.
In D25977 I meant to change all functions that care about lifetime. I
changed constructors, factory functions, but I missed member/free
functions that return new instances. This patch changes them.
Reviewers: hfinkel, kbarton, echristo, joerg
Subscribers: llvm-commits, mehdi_amini
Differential Revision: https://reviews.llvm.org/D26269
llvm-svn: 286060
Craig Topper [Sun, 6 Nov 2016 04:12:52 +0000 (04:12 +0000)]
[AVX-512] Add missing EVEX version of pattern for (v2f64 (extloadv2f32 addr:)) -> VCVTPS2PDZ128rm
llvm-svn: 286059
Craig Topper [Sun, 6 Nov 2016 04:12:49 +0000 (04:12 +0000)]
[AVX-512] Add avx512vl command line to the fpext test and add -show-mc-encoding to show where we aren't using EVEX instructions.
llvm-svn: 286058
Craig Topper [Sun, 6 Nov 2016 04:12:46 +0000 (04:12 +0000)]
[AVX-512] Lower AVX cvtpd2ps intrinsic to ISD::FP_ROUND so it can use EVEX instruction when available.
llvm-svn: 286057
Craig Topper [Sun, 6 Nov 2016 04:12:42 +0000 (04:12 +0000)]
[AVX-512] Lower SSE/AVX cvtdq2ps intrinsics directly to ISD::SINT_TO_FP so they can use EVEX instructions when available.
llvm-svn: 286056
Craig Topper [Sun, 6 Nov 2016 02:03:58 +0000 (02:03 +0000)]
[AVX-512] Add -show-mc-encoding to legacy vector intrinsic tests so we can see when VEX or EVEX encoded instructions are being emitted. Make sure the tests all have an avx2 command line and an skx command line.
llvm-svn: 286055
Rui Ueyama [Sat, 5 Nov 2016 23:14:54 +0000 (23:14 +0000)]
Fix build breakage.
llvm-svn: 286054
Rui Ueyama [Sat, 5 Nov 2016 23:05:47 +0000 (23:05 +0000)]
Rewrite CommonInputSection as a synthetic input section.
A CommonInputSection is a section containing all common symbols.
That was an input section but was abstracted in a different way
than the synthetic input sections because it was written before
the synthetic input section was invented.
This patch rewrites CommonInputSection as a synthetic input section
so that it behaves better with other sections.
llvm-svn: 286053
Simon Atanasyan [Sat, 5 Nov 2016 22:58:01 +0000 (22:58 +0000)]
[ELF][MIPS] N32 ABI support
In short the patch introduces support for linking object file conform
MIPS N32 ABI [1]. This ABI is similar to N64 ABI but uses 32-bit
pointer size.
The most non-trivial requirement of this ABI is one more relocation
packing format. N64 ABI puts multiple relocation type into the single
relocation record. The N32 ABI uses series of successive relocations
with the same offset for this purpose. In this patch, new function
`mergeMipsN32RelTypes` handle this case and "convert" N32 relocation to
the N64 relocation so the rest of the code keep unchanged.
For now, linker does not support series of relocations applied to sections
without SHF_ALLOC bit. Probably later I will add the support or insert
some sort of assert into the `relocateNonAlloc` routine to catch this
case.
[1] ftp://www.linux-mips.org/pub/linux/mips/doc/ABI/MIPS-N32-ABI-Handbook.pdf
Differential revision: https://reviews.llvm.org/D26298
llvm-svn: 286052
Rui Ueyama [Sat, 5 Nov 2016 22:37:59 +0000 (22:37 +0000)]
Create a vector containing all input sections.
Previously, we do this piece of code to iterate over all input sections.
for (elf::ObjectFile<ELFT> *F : Symtab.getObjectFiles())
for (InputSectionBase<ELFT> *S : F->getSections())
It turned out that this mechanisms doesn't work well with synthetic
input sections because synthetic input sections don't belong to any
input file.
This patch defines a vector that contains all input sections including
synthetic ones.
llvm-svn: 286051
Tobias Grosser [Sat, 5 Nov 2016 21:46:01 +0000 (21:46 +0000)]
IslNodeBuilder: Ensure newly generated memory accesses are well-defined
Add some additional asserts that ensure newly code-generated memory accesses
are defined on all domain and schedule domain instances.
llvm-svn: 286050
Krzysztof Parzyszek [Sat, 5 Nov 2016 21:44:50 +0000 (21:44 +0000)]
[Hexagon] Relocate pattern-related bits to proper places
llvm-svn: 286049
Krzysztof Parzyszek [Sat, 5 Nov 2016 21:02:54 +0000 (21:02 +0000)]
[Hexagon] Round 1 of selection pattern simplifications
Consistently use register class pat frags instead of spelling out
the type and class each time.
llvm-svn: 286048
Tobias Grosser [Sat, 5 Nov 2016 21:02:43 +0000 (21:02 +0000)]
ScopInfo: Ensure copy statement memory accesses are correct
Add asserts that verify that the memory accesses of a new copy statement
are defined for all domain instances the copy statement is defined for.
llvm-svn: 286047
Boris Ulasevich [Sat, 5 Nov 2016 20:25:18 +0000 (20:25 +0000)]
test commit
llvm-svn: 286046
Simon Pilgrim [Sat, 5 Nov 2016 18:31:57 +0000 (18:31 +0000)]
[X86][SSE] Reuse zeroable element mask in lowerVectorShuffleAsBlend. NFCI
Don't regenerate a zeroable element mask with computeZeroableShuffleElements when its already available.
llvm-svn: 286045
Simon Pilgrim [Sat, 5 Nov 2016 18:22:13 +0000 (18:22 +0000)]
[X86][SSE] Reuse zeroable element mask in lowerVectorShuffleAsZeroOrAnyExtend. NFCI
Don't regenerate a zeroable element mask with computeZeroableShuffleElements when its already available.
llvm-svn: 286044
Simon Pilgrim [Sat, 5 Nov 2016 18:05:13 +0000 (18:05 +0000)]
[X86][SSE] Reuse zeroable element mask in SSE4A EXTRQ/INSERTQ vector shuffle lowering. NFCI
Don't regenerate a zeroable element mask with computeZeroableShuffleElements when its already available.
llvm-svn: 286043
Simon Pilgrim [Sat, 5 Nov 2016 17:53:27 +0000 (17:53 +0000)]
[X86][SSE] Reuse zeroable element mask in PSHUFB vector shuffle lowering. NFCI
Don't regenerate a zeroable element mask with computeZeroableShuffleElements when its already available.
llvm-svn: 286042
Daniel Jasper [Sat, 5 Nov 2016 17:43:16 +0000 (17:43 +0000)]
Simon Pilgrim [Sat, 5 Nov 2016 17:27:48 +0000 (17:27 +0000)]
[X86][SSE] Reuse zeroable element mask in lowerVectorShuffleAsInsertPS. NFCI
Don't regenerate a zeroable element mask with computeZeroableShuffleElements when its already available.
llvm-svn: 286040
Simon Pilgrim [Sat, 5 Nov 2016 17:12:19 +0000 (17:12 +0000)]
[X86][SSE] Reuse zeroable element mask in lowerVectorShuffleAsBitMask. NFCI
Don't regenerate a zeroable element mask with computeZeroableShuffleElements when its already available.
llvm-svn: 286039
Justin Lebar [Sat, 5 Nov 2016 16:47:25 +0000 (16:47 +0000)]
[LoopStrengthReduce] Don't use a DenseSet<int64_t> when we might add any valid int64_t to the set.
Summary:
SmallSetVector uses DenseSet, but that means we need to reserve some
values for the empty and tombstone keys.
It seems to me we should have a general way to let us store full-range
ints inside of DenseSets, and furthermore that we probably shouldn't
silently let you add ints into DenseSets without explicitly promising
that they're in range. But that's a battle for another day; for now,
just fix this code, since we currently do something Very Bad when
compiling ffmpeg.
Fixes PR30914.
Reviewers: jeremyhu
Subscribers: llvm-commits, mzolotukhin
Differential Revision: https://reviews.llvm.org/D26323
llvm-svn: 286038
Simon Pilgrim [Sat, 5 Nov 2016 16:40:20 +0000 (16:40 +0000)]
[X86][SSE] Reuse zeroable element mask instead of regenerating it. NFCI
We are repeatedly calling computeZeroableShuffleElements in many shuffle lowering calls for the same shuffle mask/inputs.
This is a first step towards reusing the zeroable result, initially just for lowerVectorShuffleAsShift calls.
llvm-svn: 286037
Krzysztof Parzyszek [Sat, 5 Nov 2016 15:01:38 +0000 (15:01 +0000)]
[Hexagon] Split all selection patterns into a separate file
This is just the basic separation, without any cleanup. Further changes
will follow.
llvm-svn: 286036
Ed Maste [Sat, 5 Nov 2016 14:45:13 +0000 (14:45 +0000)]
make-core.sh: add FreeBSD support
Differential Revision: https://reviews.llvm.org/D26315
llvm-svn: 286035
Simon Pilgrim [Sat, 5 Nov 2016 14:43:04 +0000 (14:43 +0000)]
Strip trailing whitespace. NFCI.
llvm-svn: 286034
Craig Topper [Sat, 5 Nov 2016 05:35:23 +0000 (05:35 +0000)]
[AVX-512] Use an equality compare instead of StringRef::startswith in a few places in auto upgrade that were looking for the complete intrinsic name anyway.
llvm-svn: 286033
Alina Sbirlea [Sat, 5 Nov 2016 04:22:15 +0000 (04:22 +0000)]
Correct mprotect page boundries to round up end page. Fixes PR30905.
Summary:
Update the boundries for mprotect.
Patch by Andrew Adams. Fixes PR30905.
Reviewers: loladiro, andrew.w.kaylor, chandlerc
Subscribers: abadams, llvm-commits
Differential Revision: https://reviews.llvm.org/D26312
llvm-svn: 286032
Craig Topper [Sat, 5 Nov 2016 04:00:31 +0000 (04:00 +0000)]
[X86] Remove broken support for autoupgrading llvm.x86.fma4.* intrinsics to llvm.x86.fma.*.
It currently fires an assert if you even try. Looking back, I don't think it ever worked because it only changed the name of the function object, but not the intrinsic ID stored in it. Given that, I think it can be removed since no one has noticed or complained in the past 4 years.
llvm-svn: 286031
Eugene Zelenko [Sat, 5 Nov 2016 01:00:56 +0000 (01:00 +0000)]
Fix some Clang-tidy modernize-use-default and Include What You Use warnings; other minor fixes.
Differential revision: https://reviews.llvm.org/D26320
llvm-svn: 286030
Eric Fiselier [Sat, 5 Nov 2016 00:30:27 +0000 (00:30 +0000)]
Upgrade in-tree google benchmark to v1.1
llvm-svn: 286029
Rui Ueyama [Fri, 4 Nov 2016 23:26:29 +0000 (23:26 +0000)]
Fix buildbot.
llvm-svn: 286028
NAKAMURA Takumi [Fri, 4 Nov 2016 23:22:19 +0000 (23:22 +0000)]
Append a (uint32_t)pad into arm_thread_state64_t. Implicit pad isn't appended on i686.
llvm-svn: 286027
Lang Hames [Fri, 4 Nov 2016 22:29:34 +0000 (22:29 +0000)]
[ExecutionEngine] Document the insane ownership contract for
ExecutionEngine::removeModule.
llvm-svn: 286026
Rui Ueyama [Fri, 4 Nov 2016 22:25:39 +0000 (22:25 +0000)]
Don't gc .interp section.
This change fixes a bug that was introduced by r285851.
r285851 converted .interp section as an output section to an input
section. But I forgot to make it a "Live" section, so if -gc-section
is given, it was garbage collected.
llvm-svn: 286025
Justin Bogner [Fri, 4 Nov 2016 21:55:23 +0000 (21:55 +0000)]
cmake: When using LLVM_DISTRIBUTION_COMPONENTS, adjust LLVMExports accordingly
This Makes sure we only export targets that we're distributing, since
cmake will fail to import the file otherwise due to missing targets.
llvm-svn: 286024
Felix Berger [Fri, 4 Nov 2016 20:51:31 +0000 (20:51 +0000)]
[ClangTidy - performance-unnecessary-value-param] Only add "const" when current parameter is not already const qualified
Reviewers: alexfh, sbenza, aaron.ballman
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D26207
llvm-svn: 286010
Krzysztof Parzyszek [Fri, 4 Nov 2016 20:41:03 +0000 (20:41 +0000)]
[Hexagon] Account for <def,read-undef> when validating moves for predication
llvm-svn: 286009
Felix Berger [Fri, 4 Nov 2016 20:29:22 +0000 (20:29 +0000)]
[clang-tidy] Ignore incomplete types when determining whether they are expensive to copy
Summary: IsExpensiveToCopy can return false positives for incomplete types, so ignore them.
All existing ClangTidy tests that depend on this function still pass as the types are complete.
Reviewers: alexfh, aaron.ballman
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D26195
llvm-svn: 286008
Stephan T. Lavavej [Fri, 4 Nov 2016 20:26:59 +0000 (20:26 +0000)]
[libcxx] [test] Replace _LIBCPP_STD_VER with TEST_STD_VER.
This replaces every occurrence of _LIBCPP_STD_VER in the tests with
TEST_STD_VER. Additionally, for every affected
file, #include "test_macros.h" is being added explicitly if it wasn't
already there.
https://reviews.llvm.org/D26294
llvm-svn: 286007
Weiming Zhao [Fri, 4 Nov 2016 19:17:32 +0000 (19:17 +0000)]
Fix 24560: assembler does not share constant pool for same constants
Summary: This patch returns the same label if the CP entry with the same value has been created.
Reviewers: eli.friedman, rengolin, jmolloy
Subscribers: majnemer, jmolloy, llvm-commits
Differential Revision: https://reviews.llvm.org/D25804
llvm-svn: 286006
NAKAMURA Takumi [Fri, 4 Nov 2016 18:50:45 +0000 (18:50 +0000)]
llvm/test/Transforms/DCE/calls-errno.ll: Suppress checking @pow(+0,-1).
It depends on host's pow(3), and mingw's pow doesn't raise any errors, just returns +INF.
llvm-svn: 286005
Rui Ueyama [Fri, 4 Nov 2016 18:22:36 +0000 (18:22 +0000)]
Revert r285968: Always use parallel_for_each because it falls back to std::for_each.
It turned ou that we actually want to call std::for_each even if
threading is supported. Unless --thread is given, LLD shouldn't use
more than one threads.
llvm-svn: 286004
Enrico Granata [Fri, 4 Nov 2016 18:15:39 +0000 (18:15 +0000)]
Preliminary plumbing work to make 'parray' able to take offset and stride options
llvm-svn: 286003