Simon Pilgrim [Thu, 10 Sep 2020 14:12:05 +0000 (15:12 +0100)]
[X86] Use Register instead of unsigned. NFCI.
Fixes llvm-prefer-register-over-unsigned clang-tidy warnings.
Simon Pilgrim [Thu, 10 Sep 2020 12:09:48 +0000 (13:09 +0100)]
AArch64MachineFunctionInfo.h - remove unnecessary TargetFrameLowering.h include. NFCI.
Nico Weber [Thu, 3 Sep 2020 23:37:29 +0000 (19:37 -0400)]
Add an explicit toggle for the static analyzer in clang-tidy
Instead of using CLANG_ENABLE_STATIC_ANALYZER for use of the
static analyzer in both clang and clang-tidy, add a second
toggle CLANG_TIDY_ENABLE_STATIC_ANALYZER.
This allows enabling the static analyzer in clang-tidy while
disabling it in clang.
Differential Revison: https://reviews.llvm.org/D87118
Frederik Gossen [Thu, 10 Sep 2020 14:41:39 +0000 (14:41 +0000)]
[MLIR][Standard] Simplify `tensor_from_elements`
Define assembly format and add required traits.
Differential Revision: https://reviews.llvm.org/D87366
Yuriy Chernyshov [Thu, 10 Sep 2020 14:40:40 +0000 (16:40 +0200)]
[TableGen] Do not construct string from nullptr
While I am trying to forbid such usages systematically in
https://reviews.llvm.org/D79427 / P2166R0 to C++ standard,
this PR fixes this (definitelly incorrect) usage in llvm.
Differential Revision: https://reviews.llvm.org/D87185
Tim Keith [Thu, 10 Sep 2020 14:22:52 +0000 (07:22 -0700)]
[flang] Fix check for distinguishable operators/assignments
Change how generic operators and assignments are checked for
distinguishable procedures. Because of how they are invoked, available
type-bound generics and normal generics all have to be considered
together. This is different from how generic names are checked.
Move common part of checking into DistinguishabilityHelper so that it
can be used in both cases after the appropriate procedures have been
added.
Cache result of Procedure::Characterize(Symbol) in a map in
CheckHelper so that we don't have to worry about passing the
characterized Procedures around or the cost of recomputing them.
Add MakeOpName() to construct names for defined operators and assignment
for using in error messages. This eliminates the need for different
messages in those cases.
When the procedures for a defined operator or assignment are undistinguishable,
include the type name in the error message, otherwise it may be ambiguous.
Add missing check that procedures for defined operators are functions
and that their dummy arguments are INTENT(IN) or VALUE.
Differential Revision: https://reviews.llvm.org/D87341
Nikita Popov [Thu, 10 Sep 2020 14:16:44 +0000 (16:16 +0200)]
[InstCombine] Temporarily do not drop volatile stores before unreachable
See discussion in D87149. Dropping volatile stores here is legal
per LLVM semantics, but causes issues for real code and may result
in a change to LLVM volatile semantics. Temporarily treat volatile
stores as "not guaranteed to transfer execution" in just this place,
until this issue has been resolved.
Kamil Rytarowski [Thu, 10 Sep 2020 10:45:24 +0000 (12:45 +0200)]
[lldb] [netbsd] Avoid comparison of signed and unsigned integers
Cast ProcessID to ::pid_t.
LLVM GN Syncbot [Thu, 10 Sep 2020 13:33:11 +0000 (13:33 +0000)]
[gn build] Port
ebf496d8055
Roman Lebedev [Thu, 10 Sep 2020 13:31:56 +0000 (16:31 +0300)]
Revert "[clang-tidy] New check readability-prefer-member-initializer"
Either contains unbounded loops, or has *very* high runtime,
100+x of all the current clang-tidy checks.
This reverts commit
f5fd7486d6c0debb465de3e927fcc31884874280.
Jay Foad [Thu, 10 Sep 2020 12:49:33 +0000 (13:49 +0100)]
[TargetLowering] Fix comments describing XOR -> OR/AND transformations
Sanjay Patel [Thu, 10 Sep 2020 12:45:36 +0000 (08:45 -0400)]
[x86] add tests for fmax/fmin experimental intrinsics with 'fast' FMF; NFC
D87391 proposes to change the lowerings for 'nnan'-only FMF.
That's the minimal requirement to get good codegen for x86,
but currently we have bugs hindering that output unless the
full 'fast' FMF is applied. These tests provide coverage for
the ideal lowerings.
Guillaume Chatelet [Thu, 10 Sep 2020 12:27:32 +0000 (12:27 +0000)]
Fix invalid link format in Clang LanguageExtension
Guillaume Chatelet [Thu, 10 Sep 2020 12:01:18 +0000 (12:01 +0000)]
Fix broken link for Sphinx installation
Florian Hahn [Thu, 10 Sep 2020 12:09:25 +0000 (13:09 +0100)]
[DSE] Support eliminating memcpy.inline.
MemoryLocation has been taught about memcpy.inline, which means we can
get the memory locations read and written by it. This means DSE can
handle memcpy.inline
Louis Dionne [Wed, 9 Sep 2020 20:14:56 +0000 (16:14 -0400)]
[libcxx] Simplify back-deployment testing
The needs of back-deployment testing currently require two different
ways of running the test suite: one based on the deployment target,
and one based on the target triple. Since the triple includes all the
information we need, it's better to have just one way of doing things.
Furthermore, `--param platform=XXX` is also supersedded by using the
target triple. Previously, this parameter would serve the purpose of
controling XFAILs for availability markup errors, however it is possible
to achieve the same thing by using with_system_cxx_lib only and using
.verify.cpp tests instead, as explained in the documentation changes.
The motivation for this change is twofold:
1. This part of the Lit config has always been really confusing and
complicated, and it has been a source of bugs in the past. I have
simplified it iteratively in the past, but the complexity is still
there.
2. The deployment-target detection started failing in weird ways in
recent Clangs, breaking our CI. Instead of band-aid patching the
issue, I decided to remove the complexity altogether by using target
triples even on Apple platforms.
A follow-up to this commit will bring the test suite in line with
the recommended way of handling availability markup tests.
Max Kazantsev [Thu, 10 Sep 2020 11:45:12 +0000 (18:45 +0700)]
[NFC] Refactoring in SCEV: add missing `const` qualifiers
Simon Pilgrim [Thu, 10 Sep 2020 11:58:03 +0000 (12:58 +0100)]
[CodeGen][X86] Move x86 builtin intrinsic/codegen tests into X86 subfolder.
There are still plenty of tests that specify x86 as a triple but most shouldn't be doing anything very target specific - we can move any ones that I have missed on a case by case basis.
serge-sans-paille [Tue, 8 Sep 2020 12:38:16 +0000 (14:38 +0200)]
[clang-tidy] Fix reST syntax
Authored by Eisuke Kawashima [https://github.com/llvm/llvm-project/pull/245]
Simon Pilgrim [Thu, 10 Sep 2020 11:38:23 +0000 (12:38 +0100)]
[Codegen][X86] Move AMX specific codegen tests into X86 subfolder.
Simon Pilgrim [Thu, 10 Sep 2020 11:17:54 +0000 (12:17 +0100)]
[CostModel][X86] Add vXi32 division by uniform constant costs (PR47476)
Other types can be handled in future patches but their uniform / non-uniform costs are more similar and don't appear to cause many vectorization issues.
Simon Pilgrim [Thu, 10 Sep 2020 10:52:20 +0000 (11:52 +0100)]
[SLP][X86] Add division by uniform constant tests (PR47476)
Simon Pilgrim [Thu, 10 Sep 2020 10:29:06 +0000 (11:29 +0100)]
[X86][SSE] lowerShuffleAsSplitOrBlend always returns a shuffle.
lowerShuffleAsSplitOrBlend always returns a target shuffle result (and is the default operation for lowering some shuffle types), so we don't need to check for null.
Simon Pilgrim [Wed, 9 Sep 2020 16:48:22 +0000 (17:48 +0100)]
[X86] Remove WaitInsert::TTI member. NFCI.
This is only ever set/used inside WaitInsert::runOnMachineFunction so don't bother storing it in the class.
Gabor Marton [Thu, 10 Sep 2020 10:41:29 +0000 (12:41 +0200)]
[analyzer][StdLibraryFunctionsChecker] Add better diagnostics
Differential Revision: https://reviews.llvm.org/D79431
Kerry McLaughlin [Thu, 10 Sep 2020 08:55:54 +0000 (09:55 +0100)]
[SVE][CodeGen] Legalisation of truncate for scalable vectors
Truncating from an illegal SVE type to a legal type, e.g.
`trunc <vscale x 4 x i64> %in to <vscale x 4 x i32>`
fails after PromoteIntOp_CONCAT_VECTORS attempts to
create a BUILD_VECTOR.
This patch changes the promote function to create a sequence of
INSERT_SUBVECTORs if the return type is scalable, and replaces
these with UNPK+UZP1 for AArch64.
Reviewed By: paulwalker-arm
Differential Revision: https://reviews.llvm.org/D86548
Gabor Marton [Mon, 7 Sep 2020 14:56:36 +0000 (16:56 +0200)]
[analyzer][StdLibraryFunctionsChecker] Remove strcasecmp
There are 2 reasons to remove strcasecmp and strncasecmp.
1) They are also modeled in CStringChecker and the related argumentum
contraints are checked there.
2) The argument constraints are checked in CStringChecker::evalCall.
This is fundamentally flawed, they should be checked in checkPreCall.
Even if we set up CStringChecker as a weak dependency for
StdLibraryFunctionsChecker then the latter reports the warning always.
Besides, CStringChecker fails to discover the constraint violation
before the call, so, its evalCall returns with `true` and then
StdCLibraryFunctions also tries to evaluate, this causes an assertion
in CheckerManager.
Either we fix CStringChecker to handle the call prerequisites in
checkPreCall, or we must not evaluate any pure functions in
StdCLibraryFunctions that are also handled in CStringChecker.
We do the latter in this patch.
Differential Revision: https://reviews.llvm.org/D87239
Juneyoung Lee [Thu, 10 Sep 2020 10:05:24 +0000 (19:05 +0900)]
Enable InsertFreeze flag of JumpThreading when used in LTO
This patch enables inserting freeze when JumpThreading converts a select to
a conditional branch when it is run in LTO.
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D85534
Kamil Rytarowski [Thu, 10 Sep 2020 09:44:12 +0000 (11:44 +0200)]
[compiler-rt] [netbsd] Reintroduce __sanitizer_protoent
Partial revert of https://reviews.llvm.org/D82424
Sam Tebbs [Wed, 9 Sep 2020 13:01:02 +0000 (14:01 +0100)]
[ARM][LowOverheadLoops] Allow tail predication on predicated instructions with unknown lane
values
The effects of unpredicated vector instruction with unknown
lanes cannot be predicted and therefore cannot be tail predicated. This
does not apply to predicated vector instructions and so this patch
allows tail predication on them.
Differential Revision: https://reviews.llvm.org/D87376
Martin Storsjö [Tue, 8 Sep 2020 20:14:42 +0000 (23:14 +0300)]
[llvm-readobj] [ARMWinEH] Fix printing of exception handlers with packed epilogues
If there's a packed epilogue (indicated by the flag E), the EpilogueCount()
field actually should be interpreted as EpilogueOffset.
Differential Revision: https://reviews.llvm.org/D87365
Martin Storsjö [Tue, 8 Sep 2020 06:56:45 +0000 (09:56 +0300)]
[llvm-readobj] [ARMWinEH] Print set_fp/add_fp differently in epilogues
This matches how e.g. stp/ldp and other opcodes are printed differently
for epilogues.
Also add a missing --strict-whitespace in an existing test that
was added explicitly for testing vertical alignment, and change to
using temp files for the generated object files.
Differential Revision: https://reviews.llvm.org/D87363
aartbik [Wed, 9 Sep 2020 18:11:52 +0000 (11:11 -0700)]
[mlir] [VectorOps] Enable 32-bit index optimizations
Rationale:
After some discussion we decided that it is safe to assume 32-bit
indices for all subscripting in the vector dialect (it is unlikely
the dialect will be used; or even work; for such long vectors).
So rather than detecting specific situations that can exploit
32-bit indices with higher parallel SIMD, we just optimize it
by default, and let users that don't want it opt-out.
Reviewed By: nicolasvasilache, bkramer
Differential Revision: https://reviews.llvm.org/D87404
Sam Parker [Mon, 7 Sep 2020 11:06:02 +0000 (12:06 +0100)]
[SCEV] Constant expansion cost at minsize
As code size is the only thing we care about at minsize, query the
cost of materialising immediates when calculating the cost of a SCEV
expansion. We also modify the CostKind to TCK_CodeSize for minsize,
instead of RecipThroughput.
Differential Revision: https://reviews.llvm.org/D76434
Sam Parker [Wed, 9 Sep 2020 07:15:55 +0000 (08:15 +0100)]
[ARM] Tail predicate VQDMULH and VQRDMULH
Mark the family of instructions as valid for tail predication.
Differential Revision: https://reviews.llvm.org/D87348
Snehasish Kumar [Thu, 10 Sep 2020 00:57:03 +0000 (17:57 -0700)]
[clang] Disallow fbasic-block-sections on non-ELF, non-x86 targets.
Basic block sections is untested on other platforms and binary formats apart
from x86,elf. This patch emits a warning and drops the flag if the platform
and binary format are not compatible. Add a test to ensure that
specifying an incompatible target in the driver does not enable the
feature.
Differential Revision: https://reviews.llvm.org/D87426
Jakub Lichman [Thu, 10 Sep 2020 07:03:43 +0000 (07:03 +0000)]
[mlir][Linalg] Small refactoring of ConvOpVectorization
This commit addresses comments that were requested on D86619
after it was landed.
Differential Revision: https://reviews.llvm.org/D87354
Juneyoung Lee [Thu, 10 Sep 2020 06:49:04 +0000 (15:49 +0900)]
[JumpThreading] Conditionally freeze its condition when unfolding select
This patch fixes pr45956 (https://bugs.llvm.org/show_bug.cgi?id=45956 ).
To minimize its impact to the quality of generated code, I suggest enabling
this only for LTO as a start (it has two JumpThreading passes registered).
This patch contains a flag that makes JumpThreading enable it.
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D84940
Max Kazantsev [Thu, 10 Sep 2020 06:38:49 +0000 (13:38 +0700)]
[NFC] Rename variables to avoid name confusion
Name `LI` is used for loop info, loop and load inst at the same
function, which causes a lot of confusion.
Max Kazantsev [Thu, 10 Sep 2020 06:29:45 +0000 (13:29 +0700)]
[LoopLoadElim] Filter away candidates that stop being AddRecs after loop versioning. PR47457
The test in PR47457 demonstrates a situation when candidate load's pointer's SCEV
is no loger a SCEVAddRec after loop versioning. The code there assumes that it is
always a SCEVAddRec and crashes otherwise.
This patch makes sure that we do not consider candidates for which this requirement
is broken after the versioning.
Differential Revision: https://reviews.llvm.org/D87355
Reviewed By: asbirlea
Daniel Stone [Thu, 10 Sep 2020 03:15:41 +0000 (23:15 -0400)]
libclc/spirv: Add various functions
Adds fma,fmod,ldexp.
Reviewer: jenatali jvesely
Differential Revision: https://reviews.llvm.org/D85911
MaheshRavishankar [Thu, 10 Sep 2020 05:20:12 +0000 (22:20 -0700)]
[mlir][Linalg] Add Utility method to get loop ranges for a LinalgOp.
Also refactor the getViewSizes method to work on LinalgOp instead of
being a templated version. Keeping the templated version for
compatibility.
Differential Revision: https://reviews.llvm.org/D87303
Qiu Chaofan [Thu, 10 Sep 2020 05:28:09 +0000 (13:28 +0800)]
[PowerPC] [FPEnv] Disable strict FP mutation by default
22a0edd0 introduced a config IsStrictFPEnabled, which controls the
strict floating point mutation (transforming some strict-fp operations
into non-strict in ISel). This patch disables the mutation by default
since we've finished PowerPC strict-fp enablement in backend.
Reviewed By: uweigand
Differential Revision: https://reviews.llvm.org/D87222
Petr Hosek [Thu, 10 Sep 2020 05:03:13 +0000 (22:03 -0700)]
[lit] Use correct variable name for libxml2
This addresses an issue introduced in
c4d7536136b3.
Petr Hosek [Fri, 24 Jul 2020 22:47:38 +0000 (15:47 -0700)]
[CMake] Simplify CMake handling for libxml2
This matches the changes made to handling of zlib done in 10b1b4a
where we rely on find_package and the imported target rather than
manually appending the library and include paths. The use of
LLVM_LIBXML2_ENABLED has been replaced by LLVM_ENABLE_LIBXML2
thus reducing the number of variables.
Differential Revision: https://reviews.llvm.org/D84563
Jordan Rupprecht [Thu, 10 Sep 2020 03:23:59 +0000 (20:23 -0700)]
[NFC] Move definition of variable now only used in debug builds
David Blaikie [Thu, 10 Sep 2020 01:17:44 +0000 (18:17 -0700)]
mlir/Transforms/BufferPlacement.h: Add missing override
Fangrui Song [Thu, 10 Sep 2020 00:24:45 +0000 (17:24 -0700)]
[gcov] Delete flush_fn_list (unused since D83149)
Alexander Shaposhnikov [Thu, 10 Sep 2020 00:11:08 +0000 (17:11 -0700)]
[llvm-install-name-tool] Add -V flag
This diff adds -V alias for --version to make llvm-install-name-tool
consistent with other tools (llvm-objcopy, llvm-strip, etc).
Test plan: make check-all
Differential revision: https://reviews.llvm.org/D87264
Fangrui Song [Wed, 9 Sep 2020 23:57:33 +0000 (16:57 -0700)]
[gcov] Delete unused llvm_gcda_increment_indirect_counter
It has been unused since r157564 (2012).
Matt Arsenault [Wed, 9 Sep 2020 20:58:52 +0000 (16:58 -0400)]
AMDGPU: Hoist check for VGPRs
Matt Arsenault [Wed, 9 Sep 2020 22:08:48 +0000 (18:08 -0400)]
AMDGPU: Skip all meta instructions in hazard recognizer
This was not adding a necessary nop due to thinking the kill counted.
Matt Arsenault [Wed, 9 Sep 2020 14:24:35 +0000 (10:24 -0400)]
AMDGPU: Fix inserting waitcnts before kill uses
Vitaly Buka [Wed, 9 Sep 2020 23:17:37 +0000 (16:17 -0700)]
[Asan] Return nullptr for invalid chunks
CHUNK_ALLOCATED. CHUNK_QUARANTINE are only states
which make AsanChunk useful for GetAsanChunk callers.
In either case member of AsanChunk are not useful.
Fix few cases which didn't expect nullptr. Most of the callers are already
expects nullptr.
Reviewed By: morehouse
Differential Revision: https://reviews.llvm.org/D87135
Juneyoung Lee [Wed, 9 Sep 2020 17:55:06 +0000 (02:55 +0900)]
[ValueTracking] isKnownNonZero, computeKnownBits for freeze
This implements support for isKnownNonZero, computeKnownBits when freeze is involved.
```
br (x != 0), BB1, BB2
BB1:
y = freeze x
```
In the above program, we can say that y is non-zero. The reason is as follows:
(1) If x was poison, `br (x != 0)` raised UB
(2) If x was fully undef, the branch again raised UB
(3) If x was non-zero partially undef, say `undef | 1`, `freeze x` will return a nondeterministic value which is also non-zero.
(4) If x was just a concrete value, it is trivial
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D75808
Ryan Prichard [Wed, 9 Sep 2020 22:43:35 +0000 (15:43 -0700)]
[libunwind] Bare-metal DWARF: set dso_base to 0
Previously, DwarfFDECache::findFDE used 0 as a special value meaning
"search the entire cache, including dynamically-registered FDEs".
Switch this special value to -1, which doesn't make sense as a DSO
base.
Fixes PR47335.
Reviewed By: compnerd, #libunwind
Differential Revision: https://reviews.llvm.org/D86748
dfukalov [Fri, 4 Sep 2020 19:44:01 +0000 (22:44 +0300)]
[AMDGPU] Fix for folding v2.16 literals.
It was found some packed immediate operands (e.g. `<half 1.0, half 2.0>`) are
incorrectly processed so one of two packed values were lost.
Introduced new function to check immediate 32-bit operand can be folded.
Converted condition about current op_sel flags value to fall-through.
Fixes: SWDEV-247595
Reviewed By: rampitec
Differential Revision: https://reviews.llvm.org/D87158
Krzysztof Parzyszek [Wed, 9 Sep 2020 22:27:04 +0000 (17:27 -0500)]
Mark masked.{store,scatter,compressstore} intrinsics as write-only
Jessica Paquette [Wed, 9 Sep 2020 16:45:54 +0000 (09:45 -0700)]
[AArch64][GlobalISel] Share address mode selection code for memops
We were missing support for the G_ADD_LOW + ADRP folding optimization in the
manual selection code for G_LOAD, G_STORE, and G_ZEXTLOAD.
As a result, we were missing cases like this:
```
@foo = external hidden global i32*
define void @baz(i32* %0) {
store i32* %0, i32** @foo
ret void
}
```
https://godbolt.org/z/16r7ad
This functionality already existed in the addressing mode functions for the
importer. So, this patch makes the manual selection code use
`selectAddrModeIndexed` rather than duplicating work.
This is a 0.2% geomean code size improvement for CTMark at -O3.
There is one code size increase (0.1% on lencod) which is likely because
`selectAddrModeIndexed` doesn't look through constants.
Differential Revision: https://reviews.llvm.org/D87397
Florian Hahn [Wed, 9 Sep 2020 18:36:41 +0000 (19:36 +0100)]
[DSE,MemorySSA] Handle atomic stores explicitly in isReadClobber.
Atomic stores are modeled as MemoryDef to model the fact that they may
not be reordered, depending on the ordering constraints.
Atomic stores that are monotonic or weaker do not limit re-ordering, so
we do not have to treat them as potential read clobbers.
Note that llvm/test/Transforms/DeadStoreElimination/MSSA/atomic.ll
already contains a set of negative test cases.
Reviewed By: asbirlea
Differential Revision: https://reviews.llvm.org/D87386
Nikita Popov [Wed, 9 Sep 2020 20:35:02 +0000 (22:35 +0200)]
[DAGCombiner] Fold fmin/fmax of NaN
fminnum(X, NaN) is X, fminimum(X, NaN) is NaN. This mirrors the
behavior of existing InstSimplify folds.
This is expected to improve the reduction lowerings in D87391,
which use NaN as a neutral element.
Differential Revision: https://reviews.llvm.org/D87415
Nikita Popov [Wed, 9 Sep 2020 21:48:44 +0000 (23:48 +0200)]
[ARM] Add additional fmin/fmax with nan tests (NFC)
Adding these to ARM which has both FMINNUM and FMINIMUM.
Jonas Devlieghere [Wed, 9 Sep 2020 21:41:00 +0000 (14:41 -0700)]
[lldb] Pass the arch as part of the triple in the ARCH_CFLAGS
Amara Emerson [Wed, 9 Sep 2020 21:30:47 +0000 (14:30 -0700)]
Add REQUIRES: asserts to a test that uses an asserts only flag.
Amara Emerson [Fri, 28 Aug 2020 23:21:34 +0000 (16:21 -0700)]
[GlobalISel] Enable usage of BranchProbabilityInfo in IRTranslator.
We weren't using this before, so none of the MachineFunction CFG edges had the
branch probability information added. As a result, block placement later in the
pipeline was flying blind.
This is enabled only with optimizations enabled like SelectionDAG.
Differential Revision: https://reviews.llvm.org/D86824
Nikita Popov [Wed, 9 Sep 2020 20:35:56 +0000 (22:35 +0200)]
[X86] Add tests for minnum/maxnum with constant NaN (NFC)
Amara Emerson [Mon, 24 Aug 2020 21:10:38 +0000 (14:10 -0700)]
[GlobalISel][IRTranslator] Generate better conditional branch lowering.
This is a port of the functionality from SelectionDAG, which tries to find
a tree of conditions from compares that are then combined using OR or AND,
before using that result as the input to a branch. Instead of naively
lowering the code as is, this change converts that into a sequence of
conditional branches on the sub-expressions of the tree.
Like SelectionDAG, we re-use the case block codegen functionality from
the switch lowering utils, which causes us to generate some different code.
The result of which I've tried to mitigate in earlier combine patches.
Differential Revision: https://reviews.llvm.org/D86665
Amara Emerson [Mon, 24 Aug 2020 17:46:50 +0000 (10:46 -0700)]
[GlobalISel] Rewrite the elide-br-by-swapping-icmp-ops combine to do less.
This combine previously tried to take sequences like:
%cond = G_ICMP pred, a, b
G_BRCOND %cond, %truebb
G_BR %falsebb
%truebb:
...
%falsebb:
...
and by inverting the compare predicate and swapping branch targets, delete the
G_BR and instead have a single conditional branch to the falsebb. Since in an
earlier patch we have a combine to fold not(icmp) into just an inverted icmp,
we don't need this combine to do as much. This patch instead generalizes the
combine by just looking for:
G_BRCOND %cond, %truebb
G_BR %falsebb
%truebb:
...
%falsebb:
...
and then inverting the condition using a not (xor). The xor can be folded away
in a separate combine. This change also lets us avoid some optimization code
in the IRTranslator.
I also think that deleting G_BRs in the combiner is unnecessary. That's
something that targets can decide to do at selection time and could simplify
generic code in future.
Differential Revision: https://reviews.llvm.org/D86664
Siva Chandra Reddy [Wed, 9 Sep 2020 18:28:14 +0000 (11:28 -0700)]
[libc][MPFRWrapper] Provide a way to include MPFR header in downstream repos.
Reviewed By: asteinhauser
Differential Revision: https://reviews.llvm.org/D87412
Sean Silva [Tue, 8 Sep 2020 22:49:50 +0000 (15:49 -0700)]
Wordsmith RegionBranchOpInterface verification errors
I was having a lot of trouble parsing the messages. In particular, the
messages like:
```
<stdin>:3:8: error: 'scf.if' op along control flow edge from Region #0 to scf.if source #1 type '!npcomprt.tensor' should match input #1 type 'tensor<?xindex>'
```
In particular, one thing that kept catching me was parsing the "to scf.if
source #1 type" as one thing, but really it is
"to parent results: source type #1".
Differential Revision: https://reviews.llvm.org/D87334
Hiroshi Yamauchi [Fri, 21 Aug 2020 19:44:36 +0000 (12:44 -0700)]
[X86] Add support for using fast short rep mov for memcpy lowering.
Disabled by default behind an option.
Differential Revision: https://reviews.llvm.org/D86883
Tony [Sat, 5 Sep 2020 22:53:47 +0000 (22:53 +0000)]
[AMDGPU] Correct gfx1031 XNACK setting documentation
- gfx1031 does not support XNACK.
Differential Revision: https://reviews.llvm.org/D87198
Jian Cai [Wed, 9 Sep 2020 18:58:22 +0000 (11:58 -0700)]
[MC] Resolve the difference of symbols in consecutive MCDataFragements
Try to resolve the difference of two symbols in consecutive MCDataFragments.
This is important for an idiom like "foo:instr; .if . - foo; instr; .endif"
(https://bugs.llvm.org/show_bug.cgi?id=43795).
Reviewed By: nickdesaulniers
Differential Revision: https://reviews.llvm.org/D69411
Fangrui Song [Wed, 9 Sep 2020 19:31:25 +0000 (12:31 -0700)]
[gcov] Give the __llvm_gcov_ctr load instruction a name for more readable output
Krzysztof Parzyszek [Wed, 9 Sep 2020 19:28:00 +0000 (14:28 -0500)]
[Hexagon] Account for truncating pairs to non-pairs when widening truncates
Added missing selection patterns for vpackl.
Sanjay Patel [Tue, 8 Sep 2020 21:00:40 +0000 (17:00 -0400)]
[InstCombine] add tests for add/sub-of-shl; NFC
Fangrui Song [Wed, 9 Sep 2020 19:06:39 +0000 (12:06 -0700)]
[gcov] Don't split entry block; add a synthetic entry block instead
The entry block is split at the first instruction where `shouldKeepInEntry`
returns false. The created basic block has a br jumping to the original entry
block. The new basic block causes the function label line and the other entry
block lines to be covered by different basic blocks, which can affect line
counts with special control flows (fork/exec in the entry block requires
heuristics in llvm-cov gcov to get consistent line counts).
int main() { // BB0
return 0; // BB2 (due to entry block splitting)
}
// BB1 is the exit block (since gcov 4.8)
This patch adds a synthetic entry block (like PGOInstrumentation and GCC) and
inserts an edge from the synthetic entry block to the original entry block. We
can thus remove the tricky `shouldKeepInEntry` and entry block splitting. The
number of basic blocks does not change, but the emitted .gcno files will be
smaller because we can save one GCOV_TAG_LINES tag.
// BB0 is the synthetic entry block with a single edge to BB2
int main() { // BB2
return 0; // BB2
}
// BB1 is the exit block (since gcov 4.8)
Olivier Giroux [Wed, 9 Sep 2020 19:14:53 +0000 (12:14 -0700)]
Revert a test using padding bits in atomics
Guillaume Chatelet [Wed, 9 Sep 2020 19:10:30 +0000 (19:10 +0000)]
[NFC] Separate bitcode reading for FUNC_CODE_INST_CMPXCHG(_OLD)
This is preparatory work to unable storing alignment for AtomicCmpXchgInst.
See D83136 for context and bug: https://bugs.llvm.org/show_bug.cgi?id=27168
This is the fixed version of D83375, which was submitted and reverted.
Differential Revision: https://reviews.llvm.org/D87373
Mark de Wever [Wed, 9 Sep 2020 17:12:32 +0000 (19:12 +0200)]
Implements [[likely]] and [[unlikely]] in IfStmt.
This is the initial part of the implementation of the C++20 likelihood
attributes. It handles the attributes in an if statement.
Differential Revision: https://reviews.llvm.org/D85091
Krzysztof Parzyszek [Wed, 9 Sep 2020 18:44:29 +0000 (13:44 -0500)]
[DSE] Explicitly not use MSSA in testcase for now
It fails for some reason, but it shouldn't stop switching to MSSA in DSE.
Dave Lee [Tue, 8 Sep 2020 20:43:15 +0000 (13:43 -0700)]
[debugserver] Extract function for default launch flavor
Extract a function for turning `eLaunchFlavorDefault` into a concreate `eLaunchFlavor` value.
This new function encapsulates the few compile time variables involved, and also prevents clang unused code diagnostics.
Differential Revision: https://reviews.llvm.org/D87327
Krzysztof Parzyszek [Wed, 2 Sep 2020 19:06:58 +0000 (14:06 -0500)]
[DSE] Handle masked stores
Johannes Doerfert [Wed, 9 Sep 2020 15:09:30 +0000 (10:09 -0500)]
Revert "[Attributor] Re-enable a run line in noalias.ll"
The underlying issue is still there, just hides on most systems, even
some Windows builds :(
See:
http://lab.llvm.org:8011/builders/llvm-clang-x86_64-expensive-checks-win/builds/25479/steps/test-check-all/logs/FAIL%3A%20LLVM%3A%3Anoalias.ll
This reverts commit
2600c9e2efce1dc4c64870b00a45ae0082c685fc.
Eric Fiselier [Wed, 9 Sep 2020 18:25:17 +0000 (14:25 -0400)]
[libc++] Fix variant benchmark build for some configurations.
The benchmarks expect to be built in C++17 or newer, but this
isn't always how CMake configures the C++ dialect. Instead
we need to explicitly set the CXX_STANDARD target property.
Dave Lee [Wed, 9 Sep 2020 05:53:08 +0000 (22:53 -0700)]
[lldb/Docs] Correct LLDB_ENABLE_TESTS to LLDB_INCLUDE_TESTS
Fix references to LLDB_ENABLE_TESTS.
Differential Revision: https://reviews.llvm.org/D87345
Siva Chandra Reddy [Wed, 9 Sep 2020 17:19:37 +0000 (10:19 -0700)]
[libc][NFC] Add spec files as dependencies of integration test.
Ulrich Weigand [Wed, 9 Sep 2020 17:09:52 +0000 (19:09 +0200)]
[DAGCombine] Skip re-visiting EntryToken to avoid compile time explosion
During the main DAGCombine loop, whenever a node gets replaced, the new
node and all its users are pushed onto the worklist. Omit this if the
new node is the EntryToken (e.g. if a store managed to get optimized
out), because re-visiting the EntryToken and its users will not uncover
any additional opportunities, but there may be a large number of such
users, potentially causing compile time explosion.
This compile time explosion showed up in particular when building the
SingleSource/UnitTests/matrix-types-spec.cpp test-suite case on any
platform without SIMD vector support.
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D86963
Olivier Giroux [Wed, 9 Sep 2020 17:00:09 +0000 (10:00 -0700)]
Update atomic feature macros, synopsis, signatures to match C++20. Improve test coverage for non-lock-free atomics.
Mircea Trofin [Wed, 9 Sep 2020 16:55:06 +0000 (09:55 -0700)]
[NFC][MLInliner] Don't initialize in an assert.
Since the build bots have assertions enabled, this flew under the radar.
Siva Chandra Reddy [Wed, 9 Sep 2020 15:32:51 +0000 (08:32 -0700)]
[libc][obvious] Fix strtok_r signature in the spec.
Joseph Huber [Fri, 4 Sep 2020 19:03:49 +0000 (15:03 -0400)]
[OpenMP] Begin Printing Information Dumps In Libomptarget and Plugins
Summary:
This patch starts adding support for adding information dumps to libomptarget
and rtl plugins. The information printing is controlled by the
LIBOMPTARGET_INFO environment variable introduced in D86483. The goal of this
patch is to provide the user with additional information about the device
during kernel execution and providing the user with information dumps in the
case of failure. This patch added the ability to dump the pointer mapping table
as well as printing the number of blocks and threads in the cuda RTL.
Reviewers: jdoerfort gkistanova ye-luo
Subscribers: guansong openmp-commits sstefan1 yaxunl ye-luo
Tags: #OpenMP
Differential Revision: https://reviews.llvm.org/D87165
Simon Pilgrim [Wed, 9 Sep 2020 15:33:19 +0000 (16:33 +0100)]
X86CallFrameOptimization.cpp - use const references where possible. NFCI.
Krzysztof Parzyszek [Wed, 9 Sep 2020 15:23:00 +0000 (10:23 -0500)]
[DSE] Add testcase that uses masked loads and stores
Jakub Lichman [Tue, 8 Sep 2020 15:04:35 +0000 (15:04 +0000)]
[mlir][Linalg] Reduction dimensions specified in TC definition of ConvOps.
This commit specifies reduction dimensions for ConvOps. This prevents
running reduction loops in parallel and enables easier detection of kernel dimensions
which we will need later on.
Differential Revision: https://reviews.llvm.org/D87288
Simon Pilgrim [Wed, 9 Sep 2020 15:13:55 +0000 (16:13 +0100)]
X86FrameLowering::adjustStackWithPops - cleanup auto usage. NFCI.
Don't use auto for non-obvious types, and use const references.
Qiu Chaofan [Wed, 9 Sep 2020 14:38:58 +0000 (22:38 +0800)]
[PowerPC] Fix STRICT_FRINT/STRICT_FNEARBYINT lowering
In standard C library, both rint and nearbyint returns rounding result
in current rounding mode. But nearbyint never raises inexact exception.
On PowerPC, x(v|s)r(d|s)pic may modify FPSCR XX, raising inexact
exception. So we can't select constrained fnearbyint into xvrdpic.
One exception here is xsrqpi, which will not raise inexact exception, so
fnearbyint f128 is okay here.
Reviewed By: uweigand
Differential Revision: https://reviews.llvm.org/D87220
Jay Foad [Wed, 9 Sep 2020 13:55:48 +0000 (14:55 +0100)]
[AMDGPU] Simplify S_SETREG_B32 case in EmitInstrWithCustomInserter
NFC.
compinder [Wed, 9 Sep 2020 13:32:51 +0000 (19:02 +0530)]
[FLANG] Generate error for invalid selector.
Fix of PR47339
Differential Revision: https://reviews.llvm.org/D87073/new/
Dmitry Preobrazhensky [Wed, 9 Sep 2020 12:58:12 +0000 (15:58 +0300)]
[AMDGPU][MC] Improved diagnostic messages for invalid registers
Corrected parser to issue meaningful error messages for invalid and malformed registers.
See bug 41303: https://bugs.llvm.org/show_bug.cgi?id=41303
Reviewers: arsenm, rampitec
Differential Revision: https://reviews.llvm.org/D87234