Craig Topper [Tue, 2 Apr 2019 00:50:58 +0000 (00:50 +0000)]
[X86] Use unsigned type for opcodes throughout X86FixupLEAs.
All of the interfaces related to opcode in MachineInstr and MCInstrInfo refer to opcodes as unsigned.
llvm-svn: 357444
Matt Arsenault [Tue, 2 Apr 2019 00:46:19 +0000 (00:46 +0000)]
InstSimplify: Add missing case from r357386
llvm-svn: 357443
Michael Liao [Tue, 2 Apr 2019 00:36:37 +0000 (00:36 +0000)]
[AMDGPU] Add more test cases of D59608.
Summary: - Add more test cases.
Reviewers: arsenm
Subscribers: kzhuravl, jvesely, wdng, nhaehnle, yaxunl, dstuttard, tpr, t-tye, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D60071
llvm-svn: 357442
Matt Arsenault [Tue, 2 Apr 2019 00:36:04 +0000 (00:36 +0000)]
AMDGPU: Fix test filename
llvm-svn: 357441
Eli Friedman [Tue, 2 Apr 2019 00:01:23 +0000 (00:01 +0000)]
[ARM] Optimize expressions like "return x != 0;" for Thumb1.
There's an existing optimization for x != C, but somehow it was missing
a special case for 0.
While I'm here, also cleaned up the code/comments a bit: the second
value produced by the MERGE_VALUES was actually dead, since a CMOV only
produces one result.
Differential Revision: https://reviews.llvm.org/D59616
llvm-svn: 357437
Eli Friedman [Mon, 1 Apr 2019 23:55:57 +0000 (23:55 +0000)]
[ARM] Don't try to create "push {r12, lr}" in Thumb1 at -Oz.
It's a little tricky to make this issue show up because
prologue/epilogue emission normally likes to push at least two
registers... but it doesn't when lr is force-spilled due to function
length. Not sure if that really makes sense, but I decided not to touch
it for now.
Differential Revision: https://reviews.llvm.org/D59385
llvm-svn: 357436
Philip Reames [Mon, 1 Apr 2019 22:39:54 +0000 (22:39 +0000)]
[LoopPred] Rename a variable to simply a future patch [NFC]
llvm-svn: 357433
Jessica Paquette [Mon, 1 Apr 2019 22:19:13 +0000 (22:19 +0000)]
[AArch64][GlobalISe] Select STRQui for stores into v264s instead of scalarizing
This improves selection for vector stores into v2s64s. Before we just
scalarized them, but we can just use a STRQui instead.
Differential Revision: https://reviews.llvm.org/D60083
llvm-svn: 357432
Jonas Devlieghere [Mon, 1 Apr 2019 22:03:04 +0000 (22:03 +0000)]
[CMake] Only the Python scirpt interpreter should link against Python.
This patch removes spurious links against Python.
llvm-svn: 357431
Reid Kleckner [Mon, 1 Apr 2019 21:16:17 +0000 (21:16 +0000)]
Fix clangd unittest _WIN32 ifdef
WIN32 is not defined, _WIN32 is, use that instead.
llvm-svn: 357429
Jonas Devlieghere [Mon, 1 Apr 2019 20:39:03 +0000 (20:39 +0000)]
[Process] Use early returns in Process::WriteMemory (NFC)
I found the code of Process::WriteMemory particularly hard to follow
when reviewing Ismail's change in D60022. This simplifies the code and
hopefully prevents similar oversights in the future.
Differential revision: https://reviews.llvm.org/D60092
llvm-svn: 357428
Nick Lewycky [Mon, 1 Apr 2019 20:37:56 +0000 (20:37 +0000)]
[NFC] Remove dead parameter "FreeInLoop", fix some typos and trailing whitespace.
Differential Revision: https://reviews.llvm.org/D60084
llvm-svn: 357427
Jorge Gorbe Moya [Mon, 1 Apr 2019 20:37:22 +0000 (20:37 +0000)]
[lldb-vscode] Add logic to handle EOF when reading from lldb-vscode stdout.
Summary:
This change prevents the lldb-vscode test harness from hanging up waiting for
new messages when the lldb-vscode subprocess crashes.
Now, when an EOF from the subprocess pipe is detected we enqueue a `None` packet
in the received packets list. Then, during the message processing loop, we can
use this `None` packet to tell apart the case where lldb-vscode has terminated
unexpectedly from the normal situation where no pending messages means blocking
and waiting for more data.
I believe this should be enough to fix the issues with these tests hanging on
multiple platforms. Once this lands, I'll prepare and test a separate change
removing the @skipIfLinux annotations.
Reviewers: clayborg, zturner
Subscribers: lldb-commits
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D59849
llvm-svn: 357426
Nick Lewycky [Mon, 1 Apr 2019 20:03:16 +0000 (20:03 +0000)]
Not all blocks are reachable from entry. Don't assume they are.
Fixes a bug in isPotentiallyReachable, noticed by inspection.
llvm-svn: 357425
Jonas Devlieghere [Mon, 1 Apr 2019 19:56:15 +0000 (19:56 +0000)]
[API] Add SBReproducer to LLDB.h
llvm-svn: 357424
Louis Dionne [Mon, 1 Apr 2019 19:53:44 +0000 (19:53 +0000)]
[libcxx] Make sure reference_wrapper works with incomplete types
Summary: Completes P0357R3, which was merged into the C++20 Working Draft in San Diego.
Reviewers: EricWF, mclow.lists
Subscribers: christof, jkorous, dexonsmith, libcxx-commits
Differential Revision: https://reviews.llvm.org/D54722
llvm-svn: 357423
Louis Dionne [Mon, 1 Apr 2019 19:34:09 +0000 (19:34 +0000)]
[pstl] Mangle the parallel_strict_scan backend function
llvm-svn: 357422
Matthew Voss [Mon, 1 Apr 2019 19:23:56 +0000 (19:23 +0000)]
[NFC][LLD] Specify namespaces explicity to fix build failure on GCC 5 after r357383
llvm-svn: 357421
Med Ismail Bennani [Mon, 1 Apr 2019 19:08:47 +0000 (19:08 +0000)]
[Process] Fix WriteMemory return value
Summary:
In case of a breakpoint site overlapping with the destination address,
the WriteMemory method reported an incorrect memory size.
Instead of returning the right amount of bytes written, it falls through
the scope and returned 0.
Signed-off-by: Med Ismail Bennani <medismail.bennani@gmail.com>
Reviewers: jasonmolenda, friss, jingham
Subscribers: JDevlieghere, davide, lldb-commits, #lldb
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D60022
llvm-svn: 357420
Craig Topper [Mon, 1 Apr 2019 19:08:15 +0000 (19:08 +0000)]
[X86] Classify the AVX512 rounding control operand as X86::OPERAND_ROUNDING_CONTROL instead of MCOI::OPERAND_IMMEDIATE. Add an assert on legal values of rounding control in the encoder and remove an explicit mask.
This should allow llvm-exegesis to intelligently constrain the rounding mode.
The mask in the encoder shouldn't be necessary any more. We used to allow codegen to use 8-11 for rounding mode and the assembler would use 0-3 to mean the same thing so we masked here and in the printer. Codegen now matches the assembler and the printer was updated, but I forgot to update the encoder.
llvm-svn: 357419
Yi Kong [Mon, 1 Apr 2019 18:12:43 +0000 (18:12 +0000)]
[llvm-objcopy] Add --keep-symbols option
Differential Revision: https://reviews.llvm.org/D60054
llvm-svn: 357418
Peter Collingbourne [Mon, 1 Apr 2019 18:01:18 +0000 (18:01 +0000)]
ELF: Perform per-section .ARM.exidx processing during combineEhFrameSections(). NFCI.
And rename the function to combineEhSections(). This makes the processing
of .ARM.exidx even more similar to .eh_frame and means that we can avoid an
additional loop over InputSections.
Differential Revision: https://reviews.llvm.org/D60026
llvm-svn: 357417
Simon Pilgrim [Mon, 1 Apr 2019 17:48:03 +0000 (17:48 +0000)]
[SLP] getVectorElementSize and isTreeTinyAndNotFullyVectorizable are const methods. NFCI.
llvm-svn: 357416
Michael Kruse [Mon, 1 Apr 2019 17:47:41 +0000 (17:47 +0000)]
[CodeGen] Generate follow-up metadata for loops with more than one transformation.
Before this patch, CGLoop would dump all transformations for a loop into
a single LoopID without encoding any order in which to apply them.
rL348944 added the possibility to encode a transformation order using
followup-attributes.
When a loop has more than one transformation, use the follow-up
attribute define the order in which they are applied. The emitted order
is the defacto order as defined by the current LLVM pass pipeline,
which is:
LoopFullUnrollPass
LoopDistributePass
LoopVectorizePass
LoopUnrollAndJamPass
LoopUnrollPass
MachinePipeliner
This patch should therefore not change the assembly output, assuming
that all explicit transformations can be applied, and no implicit
transformations in-between. In the former case,
WarnMissedTransformationsPass should emit a warning (except for
MachinePipeliner which is not implemented yet). The latter could be
avoided by adding 'llvm.loop.disable_nonforced' attributes.
Because LoopUnrollAndJamPass processes a loop nest, generation of the
MDNode is delayed to after the inner loop metadata have been processed.
A temporary LoopID is therefore used to annotate instructions and
RAUW'ed by the actual LoopID later.
Differential Revision: https://reviews.llvm.org/D57978
llvm-svn: 357415
Simon Pilgrim [Mon, 1 Apr 2019 17:32:46 +0000 (17:32 +0000)]
[SLP] getGatherCost and isFullyVectorizableTinyTree are const methods. NFCI.
llvm-svn: 357414
Marshall Clow [Mon, 1 Apr 2019 17:23:30 +0000 (17:23 +0000)]
Fix PR#41323 'Race condition in steady_clock::now for _LIBCPP_WIN32API'. thanks to Ivan Afanasyev for the report.
llvm-svn: 357413
Alexey Bataev [Mon, 1 Apr 2019 16:56:59 +0000 (16:56 +0000)]
[OPENMP]Allocate clause allocator in target region.
According to OpenMP 5.0, 2.11.4 allocate Clause, Restrictions, allocate
clauses that appear on a target construct or on constructs in a target
region must specify an allocator expression unless a requires directive
with the dynamic_allocators clause is present in the same compilation
unit. Patch adds a check for this restriction.
llvm-svn: 357412
Louis Dionne [Mon, 1 Apr 2019 16:39:34 +0000 (16:39 +0000)]
[libc++] Declare std::tuple_element as struct instead of class
Similarly to https://reviews.llvm.org/rL350972, this revision changes
std::tuple_element from class to struct.
Fixes PR41331.
Thanks to Jan Wilken Dörrie for the patch.
Differential Revision: https://reviews.llvm.org/D60069
llvm-svn: 357411
Marshall Clow [Mon, 1 Apr 2019 16:38:02 +0000 (16:38 +0000)]
Fix PR41130 - 'operator/ of std::chrono::duration and custom type'. Thanks to Zulan for the report, and Howard for the direction of the fix.
llvm-svn: 357410
Caroline Tice [Mon, 1 Apr 2019 16:29:40 +0000 (16:29 +0000)]
Commit accidentally omitted test case.
This test case was approved as part of
https://reviews.llvm.org/D49434, but was accidentally
omitted from the final commit.
llvm-svn: 357409
Philip Reames [Mon, 1 Apr 2019 16:26:08 +0000 (16:26 +0000)]
[LoopPred] Be uniform about proving generated conditions
We'd been optimizing the case where the predicate was obviously true, do the same for the false case. Mostly just for completeness sake, but also may improve compile time in loops which will exit through the guard. Such loops are presumed rare in fastpath code, but may be present down untaken paths, so optimizing for them is still useful.
llvm-svn: 357408
Bixia Zheng [Mon, 1 Apr 2019 16:10:26 +0000 (16:10 +0000)]
[NVPTX] Fix the codegen for llvm.round.
Summary:
Previously, we translate llvm.round to PTX cvt.rni, which rounds to the
even interger when the source is equidistant between two integers. This
is not correct as llvm.round should round away from zero. This change
replaces llvm.round with a round away from zero implementation through
target specific custom lowering.
Modify a few affected tests to not check for cvt.rni. Instead, we check
for the use of a few constants used in implementing round. We are also
adding CUDA runnable tests to check for the values produced by
llvm.round to test-suites/External/CUDA.
Reviewers: tra
Subscribers: jholewinski, sanjoy, jlebar, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D59947
llvm-svn: 357407
Philip Reames [Mon, 1 Apr 2019 16:05:15 +0000 (16:05 +0000)]
[LoopPred] Delete the old condition expressions if unused
LoopPredication was replacing the original condition, but leaving the instructions to compute the old conditions around. This would get cleaned up by other passes of course, but we might as well do it eagerly. That also makes the test output less confusing.
llvm-svn: 357406
Gabor Marton [Mon, 1 Apr 2019 15:48:29 +0000 (15:48 +0000)]
Attempt to fix failing buildbot (ppc64le)
llvm-svn: 357405
Philip Reames [Mon, 1 Apr 2019 15:35:30 +0000 (15:35 +0000)]
[Tests] Autogen all the LoopPredication tests
I'm about to make some changes to the pass which cause widespread - but uninteresting - test diffs. Prepare the tests for easy updating.
llvm-svn: 357404
Philip Reames [Mon, 1 Apr 2019 15:32:07 +0000 (15:32 +0000)]
[Tests] Add tests for a possible loop predication transform variant
As highlighted by tests, if one of the operands is loop variant, but guaranteed to have the same value on all iterations, we have a missed oppurtunity.
llvm-svn: 357403
Gabor Marton [Mon, 1 Apr 2019 15:29:55 +0000 (15:29 +0000)]
[ASTImporter] Convert ODR diagnostics inside ASTImporter implementation
Summary:
ASTStructuralEquivalence uses a flag to indicate whether ODR diagnostics
should be considered errors or warnings as module Sema is more strict than
ASTMerge. The implementation of ASTImporter should allso follow
along the same lines.
Reviewers: martong, a.sidorin, shafik, a_sidorin
Reviewed By: shafik, a_sidorin
Subscribers: rnkovacs, martong, dkrupp, Szelethus, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D59761
Patch by Endre Fulop!
llvm-svn: 357402
Louis Dionne [Mon, 1 Apr 2019 15:21:46 +0000 (15:21 +0000)]
[pstl] Indent preprocessor directives as part of the clang-format rules
Summary:
Indenting preprocessor directives provides a significant gain in
readability. We do it for normal if statements, and it makes sense
to do it for preprocessor ifs too.
Reviewers: rodgert, MikeDvorskiy
Subscribers: jkorous, dexonsmith, jdoerfert, libcxx-commits
Differential Revision: https://reviews.llvm.org/D59767
llvm-svn: 357401
Neil Henning [Mon, 1 Apr 2019 15:19:52 +0000 (15:19 +0000)]
[AMDGPU] Pre-allocate WWM registers to reduce VGPR pressure.
This change incorporates an effort by Connor Abbot to change how we deal
with WWM operations potentially trashing valid values in inactive lanes.
Previously, the SIFixWWMLiveness pass would work out which registers
were being trashed within WWM regions, and ensure that the register
allocator did not have any values it was depending on resident in those
registers if the WWM section would trash them. This worked perfectly
well, but would cause sometimes severe register pressure when the WWM
section resided before divergent control flow (or at least that is where
I mostly observed it).
This fix instead runs through the WWM sections and pre allocates some
registers for WWM. It then reserves these registers so that the register
allocator cannot use them. This results in a significant register
saving on some WWM shaders I'm working with (130 -> 104 VGPRs, with just
this change!).
Differential Revision: https://reviews.llvm.org/D59295
llvm-svn: 357400
Michal Gorny [Mon, 1 Apr 2019 15:08:24 +0000 (15:08 +0000)]
[lldb] [Process/elf-core] Support aarch64 NetBSD core dumps
Include support for NetBSD core dumps from evbarm/aarch64 system,
and matching test cases for them.
Based on earlier work by Kamil Rytarowski.
Differential Revision: https://reviews.llvm.org/D60034
llvm-svn: 357399
Nico Weber [Mon, 1 Apr 2019 14:59:50 +0000 (14:59 +0000)]
gn build: Merge r357383
llvm-svn: 357398
David Spickett [Mon, 1 Apr 2019 14:56:37 +0000 (14:56 +0000)]
[AArch64] Add v8.5-a Memory Tagging STZGM instruction
This instruction writes a block of allocation tags
and stores zero to the associated data locations.
It differs from STGM by 1 bit and has the same
arguments.
The specification can be found here:
https://developer.arm.com/docs/ddi0596/c
Differential Revision: https://reviews.llvm.org/D60065
llvm-svn: 357397
Alex Bradbury [Mon, 1 Apr 2019 14:53:17 +0000 (14:53 +0000)]
[RISCV] Attach VK_RISCV_CALL to symbols upon creation
This patch replaces the addition of VK_RISCV_CALL in RISCVMCCodeEmitter by
creating the RISCVMCExpr when tail/call are parsed, or in the codegen case
when the callee symbols are created.
This required adding a new CallSymbol operand to allow only adding
VK_RISCV_CALL to tail/call instructions.
This patch will allow further expansion of parsing and codegen to easily
include PLT symbols which must generate the R_RISCV_CALL_PLT relocation.
Differential Revision: https://reviews.llvm.org/D55560
Patch by Lewis Revill.
llvm-svn: 357396
David Spickett [Mon, 1 Apr 2019 14:52:18 +0000 (14:52 +0000)]
[AArch64] Add v8.5-a Memory Tagging STGM/LDGM instructions
The STGV/LDGV instructions were replaced with
STGM/LDGM. The encodings remain the same but there
is no longer writeback so there are no unpredictable
encodings to check for.
The specfication can be found here:
https://developer.arm.com/docs/ddi0596/c
Differential Revision: https://reviews.llvm.org/D60064
llvm-svn: 357395
Gabor Marton [Mon, 1 Apr 2019 14:46:53 +0000 (14:46 +0000)]
[ASTImporter] Make ODR error handling configurable
Summary:
ODR errors are not necessarily true errors during the import of ASTs.
ASTMerge and CrossTU should use the warning equivalent of every CTU error,
while Sema should emit errors as before.
Reviewers: martong, a_sidorin, shafik, a.sidorin
Reviewed By: a_sidorin
Subscribers: rnkovacs, dkrupp, Szelethus, jdoerfert, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D58897
Patch by Endre Fulop!
llvm-svn: 357394
Alex Bradbury [Mon, 1 Apr 2019 14:42:56 +0000 (14:42 +0000)]
[RISCV] Generate address sequences suitable for mcmodel=medium
This patch adds an implementation of a PC-relative addressing sequence to be
used when -mcmodel=medium is specified. With absolute addressing, a 'medium'
codemodel may cause addresses to be out of range. This is because while
'medium' implies a 2 GiB addressing range, this 2 GiB can be at any offset as
opposed to 'small', which implies the first 2 GiB only.
Note that LLVM/Clang currently specifies code models differently to GCC, where
small and medium imply the same functionality as GCC's medlow and medany
respectively.
Differential Revision: https://reviews.llvm.org/D54143
Patch by Lewis Revill.
llvm-svn: 357393
David Spickett [Mon, 1 Apr 2019 14:41:14 +0000 (14:41 +0000)]
[AArch64] Add v8.5-a Memory Tagging GMID_EL1 register
The latest version of the MTE spec added a system
register 'GMID_EL1'. It contains the block size used
by the LDGM and STGM instructions and is read only.
The specification can be found here:
https://developer.arm.com/docs/ddi0596/c
llvm-svn: 357392
Alexandre Ganea [Mon, 1 Apr 2019 14:37:36 +0000 (14:37 +0000)]
Fix builder.
http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-fuzzer/builds/24702/steps/check-fuzzer/logs/stdio
llvm-svn: 357391
Alexey Bataev [Mon, 1 Apr 2019 14:25:31 +0000 (14:25 +0000)]
[OPENMP] Check that allocated variables are used in private clauses.
According to OpenMP 5.0 standard, 2.11.4 allocate Clause, Restrictions,
For any list item that is specified in the allocate clause on a
directive, a data-sharing attribute clause that may create a private
copy of that list item must be specified on the same directive. Patch
adds the checks for this restriction.
llvm-svn: 357390
Mikael Holmen [Mon, 1 Apr 2019 14:10:10 +0000 (14:10 +0000)]
[InstCombine] Handle vector gep with scalar argument in evaluateInDifferentElementOrder
Summary:
This fixes PR41270.
The recursive function evaluateInDifferentElementOrder expects to be called
on a vector Value, so when we call it on a vector GEP's arguments, we must
first check that the argument is indeed a vector.
Reviewers: reames, spatel
Reviewed By: spatel
Subscribers: llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D60058
llvm-svn: 357389
Matt Arsenault [Mon, 1 Apr 2019 14:08:26 +0000 (14:08 +0000)]
X86: Fix override warning
llvm-svn: 357388
Mikael Holmen [Mon, 1 Apr 2019 14:06:45 +0000 (14:06 +0000)]
Revert "[InstCombine] Handle vector gep with scalar argument in evaluateInDifferentElementOrder"
This reverts commit
75216a6dbcfe5fb55039ef06a07e419fa875f4a5.
I'll recommit with a better commit message with reference to the
phabricator review.
llvm-svn: 357387
Matt Arsenault [Mon, 1 Apr 2019 14:03:44 +0000 (14:03 +0000)]
InstSimplify: Add baseline test for upcoming change
llvm-svn: 357386
Mikael Holmen [Mon, 1 Apr 2019 13:48:56 +0000 (13:48 +0000)]
[InstCombine] Handle vector gep with scalar argument in evaluateInDifferentElementOrder
This fixes PR41270.
The recursive function evaluateInDifferentElementOrder expects to be called
on a vector Value, so when we call it on a vector GEP's arguments, we must
first check that the argument is indeed a vector.
llvm-svn: 357385
Clement Courbet [Mon, 1 Apr 2019 13:48:50 +0000 (13:48 +0000)]
[X86] Make post-ra scheduling macrofusion-aware.
Subscribers: MatzeB, arsenm, jvesely, nhaehnle, hiraditya, javed.absar, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D59688
llvm-svn: 357384
Alexandre Ganea [Mon, 1 Apr 2019 13:36:59 +0000 (13:36 +0000)]
[LLD][COFF] Early dependency detection
We introduce a new class hierarchy for debug types merging (in DebugTypes.h). The end-goal is to parallelize the type merging - please see the plan in D59226.
Previously, dependency discovery was done on the fly, much later, during the type merging loop. Unfortunately, parallelizing the type merging requires the dependencies to be merged in first, before any dependent ObjFile, thus this early discovery.
The overall intention for this path is to discover debug information dependencies at a much earlier stage, when processing input files. Currently, two types of dependency are supported: PDB type servers (when compiling with MSVC /Zi) and precompiled headers OBJs (when compiling with MSVC /Yc and /Yu). Once discovered, an explicit link is added into the dependent ObjFile, through the new debug types class hierarchy introduced in DebugTypes.h.
Differential Revision: https://reviews.llvm.org/D59053
llvm-svn: 357383
Sanjay Patel [Mon, 1 Apr 2019 13:36:40 +0000 (13:36 +0000)]
[InstCombine] eliminate commuted select-shuffles + binop (PR41304)
If we have a commutable vector binop with inverted select-shuffles,
we don't care about the order of the operands in each vector lane:
LHS = shuffle V1, V2, <0, 5, 6, 3>
RHS = shuffle V2, V1, <0, 5, 6, 3>
LHS + RHS --> <V1[0]+V2[0], V2[1]+V1[1], V2[2]+V1[2], V1[3]+V2[3]> --> V1 + V2
PR41304:
https://bugs.llvm.org/show_bug.cgi?id=41304
...is currently titled as an SLP enhancement, but at least for the
given example, we can reduce that in instcombine because we are just
eliminating shuffles.
As noted in the TODO, this could be generalized, but I haven't thought
through those patterns completely, so this is limited to what appears
to be always safe.
Differential Revision: https://reviews.llvm.org/D60048
llvm-svn: 357382
Clement Courbet [Mon, 1 Apr 2019 13:18:34 +0000 (13:18 +0000)]
[X86MacroFusion][NFC] Add more tests.
In preparation for D59688.
llvm-svn: 357381
Krasimir Georgiev [Mon, 1 Apr 2019 11:42:54 +0000 (11:42 +0000)]
[X86] Fix a test from r357317
Summary:
The missing `<` causes the lld command to override the test file, which fails in
environments marking the test files as readonly.
Reviewers: bkramer
Reviewed By: bkramer
Subscribers: llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D60060
llvm-svn: 357380
Simon Pilgrim [Mon, 1 Apr 2019 10:54:04 +0000 (10:54 +0000)]
[X86][SSE] Add fcmp constant folding tests
Initial test coverage for D60006
llvm-svn: 357379
Luis Marques [Mon, 1 Apr 2019 09:54:14 +0000 (09:54 +0000)]
[RISCV] Add seto pattern expansion
Adds a `seto` pattern expansion. Without it the lowerings of `fcmp one` and
`fcmp ord` would be inefficient due to an unoptimized double negation.
Differential Revision: https://reviews.llvm.org/D59699
llvm-svn: 357378
Fangrui Song [Mon, 1 Apr 2019 08:16:08 +0000 (08:16 +0000)]
[ELF] Rename SyntheticSection::empty to more appropriate isNeeded() with opposite meaning
Summary:
Some synthetic sections can be empty while still being needed, thus they
can't be removed by removeUnusedSyntheticSections(). Rename this member
function to more appropriate isNeeded() with the opposite meaning.
No functional change intended.
Reviewers: ruiu, espindola
Reviewed By: ruiu
Subscribers: jhenderson, grimar, emaste, arichardson, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D59982
llvm-svn: 357377
Pavel Labath [Mon, 1 Apr 2019 08:11:46 +0000 (08:11 +0000)]
[Linux/x86] Fix writing of non-gpr registers on newer processors
Summary:
We're using ptrace(PTRACE_SETREGSET, NT_X86_XSTATE) to write all non-gpt
registers on x86 linux. Unfortunately, this method has a quirk, where
the kernel rejects all attempts to write to this area if one supplies a
buffer which is smaller than the area size (even though the kernel will
happily accept partial reads from it).
This means that if the CPU supports some new registers/extensions that
we don't know about (in my case it was the PKRU extension), we will fail
to write *any* non-gpr registers, even those that we know about.
Since this is a situation that's likely to appear again and again, I add
code to NativeRegisterContextLinux_x86_64 to detect the runtime size of
the area, and allocate an appropriate buffer. This does not mean that we
will start automatically supporting all new extensions, but it does mean
that the new extensions will not prevent the old ones from working.
This fixes tests attempting to write to non-gpr registers on new intel
processors (cca Kaby Lake Refresh).
Reviewers: jankratochvil, davezarzycki
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D59991
llvm-svn: 357376
Craig Topper [Mon, 1 Apr 2019 05:26:12 +0000 (05:26 +0000)]
[X86] Use ISD::INTRINSIC_VOID in getTgtMemIntrinsic for truncating stores and scatter intrinsics.
This is the appropriate opcode for only having a chain output. Though I'm not
sure it matters much.
llvm-svn: 357375
Alex Bradbury [Mon, 1 Apr 2019 02:38:27 +0000 (02:38 +0000)]
[RISCV] Don't evaluatePCRelLo if a relocation will be forced (e.g. due to linker relaxation)
A pcrel_lo will point to the associated pcrel_hi fixup which in turn points to
the real target. RISCVMCExpr::evaluatePCRelLo will work around this
indirection in order to allow the fixup to be evaluate properly. However, if
relocations are forced (e.g. due to linker relaxation is enabled) then its
evaluation is undesired and will result in a relocation with the wrong target.
This patch modifies evaluatePCRelLo so it will not try to evaluate if the
fixup will be forced as a relocation. A new helper method is added to
RISCVAsmBackend to query this.
Differential Revision: https://reviews.llvm.org/D59686
llvm-svn: 357374
Rui Ueyama [Mon, 1 Apr 2019 00:25:17 +0000 (00:25 +0000)]
Simplify. NFC.
llvm-svn: 357373
Rui Ueyama [Mon, 1 Apr 2019 00:11:24 +0000 (00:11 +0000)]
Replace `typedef A B` with `using B = A`. NFC.
I did this using Perl.
Differential Revision: https://reviews.llvm.org/D60003
llvm-svn: 357372
Sylvestre Ledru [Sun, 31 Mar 2019 21:53:00 +0000 (21:53 +0000)]
Spelling correction for docs for cppcoreguidelines-owning-memory
Summary: There's a typo in the docs, as mentioned in the title. Please see the diff.
Reviewers: JonasToth
Subscribers: sylvestre.ledru, nemanjai, kbarton, cfe-commits
Tags: #clang-tools-extra, #clang
Differential Revision: https://reviews.llvm.org/D60050
llvm-svn: 357371
Eric Fiselier [Sun, 31 Mar 2019 20:49:06 +0000 (20:49 +0000)]
Make common_type's implementation common
Summary:
Currently the C++03 implementation of common_type has much different behavior than the C++11 one. This causes bugs, including inside `<chrono>`.
This patch unifies the two implementations as best it can. The more code they share, the less their behavior can diverge.
Reviewers: mclow.lists, ldionne, sbenza
Reviewed By: mclow.lists, ldionne
Subscribers: libcxx-commits
Tags: #libc
Differential Revision: https://reviews.llvm.org/D59678
llvm-svn: 357370
Nico Weber [Sun, 31 Mar 2019 16:49:54 +0000 (16:49 +0000)]
gn build: Add build files for most clang-tools-extra unit tests
Differential Revision: https://reviews.llvm.org/D60038
llvm-svn: 357369
Sanjay Patel [Sun, 31 Mar 2019 15:45:47 +0000 (15:45 +0000)]
[InstCombine] add tests for inverted select-shuffles + binop (PR41304); NFC
llvm-svn: 357368
Sanjay Patel [Sun, 31 Mar 2019 15:11:34 +0000 (15:11 +0000)]
[x86] allow movmsk with 2-element reductions
One motivation for making this change is that the lack of using movmsk is likely
a main source of perf difference between clang and gcc on the C-Ray benchmark as
shown here:
https://www.phoronix.com/scan.php?page=article&item=gcc-clang-2019&num=5
...but this change alone isn't enough to solve that problem.
The 'all-of' examples show what is likely the worst case trade-off: we end up with
an extra instruction (or 2 if we count the 'xor' register clearing). The 'any-of'
examples look clearly better using movmsk because we've traded 2 vector instructions
for 2 scalar instructions, and movmsk may have better timing than the generic 'movq'.
If we examine the llvm-mca output for these cases, it appears that even though the
'all-of' movmsk variant looks worse on paper, it would perform better on both
Haswell and Jaguar.
$ llvm-mca -mcpu=haswell no_movmsk.s -timeline
Iterations: 100
Instructions: 400
Total Cycles: 504
Total uOps: 400
Dispatch Width: 4
uOps Per Cycle: 0.79
IPC: 0.79
Block RThroughput: 1.0
$ llvm-mca -mcpu=haswell movmsk.s -timeline
Iterations: 100
Instructions: 600
Total Cycles: 358
Total uOps: 600
Dispatch Width: 4
uOps Per Cycle: 1.68
IPC: 1.68
Block RThroughput: 1.5
$ llvm-mca -mcpu=btver2 no_movmsk.s -timeline
Iterations: 100
Instructions: 400
Total Cycles: 407
Total uOps: 400
Dispatch Width: 2
uOps Per Cycle: 0.98
IPC: 0.98
Block RThroughput: 2.0
$ llvm-mca -mcpu=btver2 movmsk.s -timeline
Iterations: 100
Instructions: 600
Total Cycles: 311
Total uOps: 600
Dispatch Width: 2
uOps Per Cycle: 1.93
IPC: 1.93
Block RThroughput: 3.0
Finally, there may be CPUs where movmsk is horribly slow (old AMD small cores?), but if
that's true, then we're also almost certainly making the wrong transform already for
reductions with >2 elements, so that should be fixed independently.
Differential Revision: https://reviews.llvm.org/D59997
llvm-svn: 357367
Sanjay Patel [Sun, 31 Mar 2019 15:01:30 +0000 (15:01 +0000)]
[InstCombine] canonicalize select shuffles by commuting
In PR41304:
https://bugs.llvm.org/show_bug.cgi?id=41304
...we have a case where we want to fold a binop of select-shuffle (blended) values.
Rather than try to match commuted variants of the pattern, we can canonicalize the
shuffles and check for mask equality with commuted operands.
We don't produce arbitrary shuffle masks in instcombine, but select-shuffles are a
special case that the backend is required to handle because we already canonicalize
vector select to this shuffle form.
So there should be no codegen difference from this change. It's possible that this
improves CSE in IR though.
Differential Revision: https://reviews.llvm.org/D60016
llvm-svn: 357366
Liang Zou [Sun, 31 Mar 2019 14:49:00 +0000 (14:49 +0000)]
fix typo: "\t" => " "
Reviewers: llvm.org, Jim
Reviewed By: Jim
Subscribers: arsenm, jvesely, nhaehnle, rupprecht, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D59983
llvm-svn: 357365
David Chisnall [Sun, 31 Mar 2019 11:22:33 +0000 (11:22 +0000)]
[gnustep-objc] Make the GNUstep v2 ABI work for Windows DLLs.
Summary:
Based on a patch by Dustin Howett, modified to not change the ABI for
ELF platforms.
Use more Windows-like section names.
This also makes things more readable by PE/COFF debug tools that assume
sections fit in the first header.
With these changes in, it is now possible to build a working WinObjC
with clang and the WinObjC version of GNUstep libobjc (upstream GNUstep
libobjc + a work around for incremental linking, which can be removed
once LINK.EXE gains a feature to opt sections out of receiving extra
padding during an incremental link).
Patch by Dustin Howett!
Reviewers: DHowett-MSFT
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D58724
llvm-svn: 357364
David Chisnall [Sun, 31 Mar 2019 11:22:26 +0000 (11:22 +0000)]
COMDAT-fold block descriptors.
Without this change, linking multiple objects containing block
descriptors together on Windows will generate duplicate symbol errors.
Patch by Dustin Howett!
Differential Revision: https://reviews.llvm.org/D58807
llvm-svn: 357363
David Chisnall [Sun, 31 Mar 2019 11:22:19 +0000 (11:22 +0000)]
[objc-gnustep] Use .init_array not .ctors when requested.
This doesn't make a difference most of the time but FreeBSD/ARM doesn't
run anything in the .ctors array.
llvm-svn: 357362
Fedor Sergeev [Sun, 31 Mar 2019 10:15:39 +0000 (10:15 +0000)]
SafepointIRVerifier port to new Pass Manager
Straightforward port of StatepointIRVerifier pass to new Pass Manager framework.
Fix By: skatkov
Reviewed By: fedor.sergeev
Differential Revision: https://reviews.llvm.org/D59825
This is a re-land of r357147/r357148 with LLVM_ENABLE_MODULES build fixed.
Adding IR/SafepointIRVerifier.h into its own module.
llvm-svn: 357361
Luqman Aden [Sun, 31 Mar 2019 08:58:50 +0000 (08:58 +0000)]
[NFC][InstCombine] Add tests for combining icmp of no-wrap sub w/ constant.
llvm-svn: 357360
Fangrui Song [Sun, 31 Mar 2019 08:48:19 +0000 (08:48 +0000)]
Range-style std::find{,_if} -> llvm::find{,_if}. NFC
llvm-svn: 357359
Nico Weber [Sun, 31 Mar 2019 00:03:37 +0000 (00:03 +0000)]
gn build: Merge r357340
llvm-svn: 357358
Nico Weber [Sun, 31 Mar 2019 00:01:24 +0000 (00:01 +0000)]
gn build: Merge r357326
llvm-svn: 357357
Nico Weber [Sat, 30 Mar 2019 23:09:10 +0000 (23:09 +0000)]
Rename IncludeFixerTests to ClangIncludeFixerTests and ChangeNamespaceTests to ClangChangeNamespaceTests
Follow-up to r356897 and r356254.
llvm-svn: 357356
Simon Pilgrim [Sat, 30 Mar 2019 20:24:26 +0000 (20:24 +0000)]
[SystemZ] Remove fcmp undef from reduced test
Pre-commit for D60006 (Add fcmp UNDEF handling to SelectionDAG::FoldSetCC)
Approved by @uweigand (Ulrich Weigand)
llvm-svn: 357355
Simon Pilgrim [Sat, 30 Mar 2019 20:16:16 +0000 (20:16 +0000)]
[MIPS] Remove fcmp undef from reduced test
Pre-commit for D60006 (Add fcmp UNDEF handling to SelectionDAG::FoldSetCC)
Approved by @atanasyan (Simon Atanasyan)
llvm-svn: 357354
Craig Topper [Sat, 30 Mar 2019 18:59:17 +0000 (18:59 +0000)]
[X86] Teach isel for RMW binops to handle negate
Negate updates flags like a subtract. We should be able to use the flags from the RMW form of negate when we have (store (X86ISD::SUB 0, load A), A)
Differential Revision: https://reviews.llvm.org/D60007
llvm-svn: 357353
Alex Bradbury [Sat, 30 Mar 2019 17:59:30 +0000 (17:59 +0000)]
[RISCV] Add codegen support for ilp32f, ilp32d, lp64f, and lp64d ("hard float") ABIs
This patch adds support for the RISC-V hard float ABIs, building on top of
rL355771, which added basic target-abi parsing and MC layer support. It also
builds on some re-organisations and expansion of the upstream ABI and calling
convention tests which were recently committed directly upstream.
A number of aspects of the RISC-V float hard float ABIs require frontend
support (e.g. flattening of structs and passing int+fp for fp+fp structs in a
pair of registers), and will be addressed in a Clang patch.
As can be seen from the tests, it would be worthwhile extending
RISCVMergeBaseOffsets to handle constant pool as well as global accesses.
Differential Revision: https://reviews.llvm.org/D59357
llvm-svn: 357352
Simon Pilgrim [Sat, 30 Mar 2019 17:12:29 +0000 (17:12 +0000)]
[X86][SSE] detectAVGPattern - Match zext(or(x,y)) 'add like' patterns (PR41316)
Fixes PR41316 where the expanded PAVG intrinsic had had one of its ADDs turned into an OR due to its operands having no conflicting bits.
llvm-svn: 357351
Alex Bradbury [Sat, 30 Mar 2019 15:53:38 +0000 (15:53 +0000)]
[RISCV] Add RV64 CHECK lines to test/CodeGen/RISCV/vararg.ll and prepare for hard float tests
vararg.ll previously missed RV64 tests. This patch also prepares for using
vararg.ll to test handling of varargs for the ilp32f/ilp32d/lp64f/lp64d hard
float ABIs. In these ABIs, varargs are passed as in either the ilp32 or lp64
ABI. Due to some slight codegen differences, different check lines are needed
for when RV32D is enabled.
llvm-svn: 357350
Simon Pilgrim [Sat, 30 Mar 2019 15:31:53 +0000 (15:31 +0000)]
[X86][SSE] detectAVGPattern - begin generalizing ADD matches
Move the ADD matching into a helper - first NFC stage towards supporting 'ADD like' cases such as in PR41316
llvm-svn: 357349
Fangrui Song [Sat, 30 Mar 2019 14:38:51 +0000 (14:38 +0000)]
[cmake] Change deprecated $<CONFIG> to $<CONFIGURATION>. NFC
See rL357338 for a similar change. The informational expression
$<CONFIGURATION> has been deprecated since CMake 3.0
llvm-svn: 357348
Fangrui Song [Sat, 30 Mar 2019 14:08:59 +0000 (14:08 +0000)]
[llvm-objcopy] Replace the size() helper with SectionTableRef::size
Summary:
BTW, STLExtras.h provides llvm::size() which is similar to std::size()
for random access iterators. However, if we prefer qualified
llvm::size(), the member function .size() will be more convenient.
Reviewers: jhenderson, jakehehrlich, rupprecht, grimar, alexshap, espindola
Reviewed By: grimar
Subscribers: emaste, arichardson, jdoerfert, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D60028
llvm-svn: 357347
Simon Pilgrim [Sat, 30 Mar 2019 13:53:11 +0000 (13:53 +0000)]
[X86][SSE] Add PAVG test case from PR41316
llvm-svn: 357346
Paul Hoad [Sat, 30 Mar 2019 13:05:40 +0000 (13:05 +0000)]
[clang-format] [PR41187] moves Java import statements to the wrong location if code contains statements that start with the word import
Summary:
Import sorting of java file, incorrectly move import statement to after a function beginning with the word import.
Make 1 character change to regular expression to ensure there is always at least one space/tab after the word import
Previously clang-format --style="LLVM" would format
```
import X;
class C {
void m() {
importFile();
}
}
```
as
```
class C {
void m() {
importFile();
import X;
}
}
```
Reviewers: djasper, klimek, reuk, JonasToth
Reviewed By: klimek
Subscribers: cfe-commits
Tags: #clang-tools-extra
Differential Revision: https://reviews.llvm.org/D59684
llvm-svn: 357345
Reuben Thomas [Sat, 30 Mar 2019 12:32:35 +0000 (12:32 +0000)]
[clang-format]: Add NonEmptyParentheses spacing option
This patch aims to add support for the following rules from the JUCE coding standards:
- Always put a space before an open parenthesis that contains text - e.g. foo (123);
- Never put a space before an empty pair of open/close parenthesis - e.g. foo();
Patch by Reuben Thomas
Differential Revision: https://reviews.llvm.org/D55170
llvm-svn: 357344
Heejin Ahn [Sat, 30 Mar 2019 11:04:48 +0000 (11:04 +0000)]
[WebAssembly] Fix unwind destination mismatches in CFG stackify
Summary:
Linearing the control flow by placing `try`/`end_try` markers can create
mismatches in unwind destinations. This patch resolves these mismatches
by wrapping those instructions with an incorrect unwind destination with
a nested `try`/`catch`/`end_try` and branching to the right destination
within the new catch block.
Reviewers: dschuff
Subscribers: sunfish, sbc100, jgravelle-google, chrib, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D48345
llvm-svn: 357343
Heejin Ahn [Sat, 30 Mar 2019 09:29:57 +0000 (09:29 +0000)]
[WebAssembly] Run ExplicitLocals pass after CFGStackify
Summary:
While this does not change any final output, this will greatly simplify
ixing unwind destination mismatches in CFGStackify (D48345), because we
have to create some new registers there.
Reviewers: dschuff
Subscribers: sunfish, sbc100, jgravelle-google, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D59652
llvm-svn: 357342
Alex Bradbury [Sat, 30 Mar 2019 09:15:47 +0000 (09:15 +0000)]
[RISCV] Add DAGCombine for (SplitF64 (ConstantFP x))
The SplitF64 node is used on RV32D to convert an f64 directly to a pair of i32
(necessary as bitcasting to i64 isn't legal). When performed on a ConstantFP,
this will result in a FP load from the constant pool followed by a store to
the stack and two integer loads from the stack (necessary as there is no way
to directly move between f64 FPRs and i32 GPRs on RV32D). It's always cheaper
to just materialise integers for the lo and hi parts of the FP constant, so do
that instead.
llvm-svn: 357341
Anton Afanasyev [Sat, 30 Mar 2019 08:42:48 +0000 (08:42 +0000)]
Adds `-ftime-trace` option to clang that produces Chrome `chrome://tracing` compatible JSON profiling output dumps.
This change adds hierarchical "time trace" profiling blocks that can be visualized in Chrome, in a "flame chart" style. Each profiling block can have a "detail" string that for example indicates the file being processed, template name being instantiated, function being optimized etc.
This is taken from GitHub PR: https://github.com/aras-p/llvm-project-
20170507/pull/2
Patch by Aras Pranckevičius.
Differential Revision: https://reviews.llvm.org/D58675
llvm-svn: 357340
Alex Bradbury [Sat, 30 Mar 2019 05:24:42 +0000 (05:24 +0000)]
[RISCV][NFC] Remove floating point operations from test/CodeGen/RISCV/vararg.ll
This minimises differences in output when compiling with hardware floating
point support, which will be done in a future patch (to demonstrate the same
vararg calling convention is used).
llvm-svn: 357339