Sanjay Patel [Sun, 2 Aug 2020 17:40:23 +0000 (13:40 -0400)]
[InstCombine] add tests for xor-of-ors; NFC
Simon Tatham [Mon, 3 Aug 2020 12:30:48 +0000 (13:30 +0100)]
[clang][ARM] Add name-mangling test for direct __fp16 arguments.
`clang/test/CodeGenCXX/fp16-mangle.cpp` tests pointers to __fp16, but
if you give the `-fallow-half-arguments-and-returns` option, then
clang can also leave an __fp16 unmodified as a function argument or
return type. This regression test checks the name-mangling of that.
Reviewed By: miyuki
Differential Revision: https://reviews.llvm.org/D85010
Nicholas Guy [Wed, 1 Jul 2020 10:35:58 +0000 (11:35 +0100)]
[ARM] Fix IT block generation after Thumb2SizeReduce with -Oz
Fixes a regression caused by D82439, in which IT blocks were no longer being
generated when -Oz is present. This was due to the CPSR register being marked as
dead, while this case was not accounted for.
Differential Revision: https://reviews.llvm.org/D83667
Haojian Wu [Sat, 25 Jul 2020 19:52:33 +0000 (21:52 +0200)]
[clangd] Support new/deleta operator in TargetFinder.
Differential Revision: https://reviews.llvm.org/D85028
Denys Petrov [Fri, 31 Jul 2020 12:54:46 +0000 (15:54 +0300)]
[analyzer] Introduce minor refactoring of SVal::getSubKind function
Summary: `BaseMask` occupies the lowest bits. Effect of applying the mask is neutralized by right shift operation, thus making it useless.
Fix: Remove a redundant bitwise operation.
Differential Revision: https://reviews.llvm.org/D85026
Denys Petrov [Fri, 31 Jul 2020 15:57:04 +0000 (18:57 +0300)]
[analyzer] Simplify function SVal::getAsSymbolicExpression and similar ones
Summary: Simplify functions SVal::getAsSymbolicExpression SVal::getAsSymExpr and SVal::getAsSymbol. After revision I concluded that `getAsSymbolicExpression` and `getAsSymExpr` repeat functionality of `getAsSymbol`, thus them can be removed.
Fix: Remove functions SVal::getAsSymbolicExpression and SVal::getAsSymExpr.
Differential Revision: https://reviews.llvm.org/D85034
Kadir Cetinkaya [Sat, 1 Aug 2020 17:03:40 +0000 (19:03 +0200)]
[clang][Tooling] Optimize addTargetAndMode in case of invalid modes
This skips searching for `target` related flags in the existing args if
we don't have a valid target to insert.
Depends on D85076
Differential Revision: https://reviews.llvm.org/D85077
Georgii Rymar [Thu, 23 Jul 2020 14:10:31 +0000 (17:10 +0300)]
[llvm-readobj] - Don't stop dumping when the name of a relocation section can't be read.
This removes undesired `unwrapOrError` calls from printRelocations() methods.
Differential revision: https://reviews.llvm.org/D84408
Georgii Rymar [Thu, 30 Jul 2020 14:55:47 +0000 (17:55 +0300)]
[llvm-readobj] - Massive test cases cleanup.
This patch does the following:
1) Starts using YAML macro to reduce the number of YAML documents in tests.
2) Adds `#` before 'RUN'/`CHECK` lines in a few tests where it is missing.
3) Removes unused YAML keys.
4) Starts using `ENTSIZE=<none>` to simplify tests (see D84526).
5) Removes trailing white spaces in a few places.
Differential revision: https://reviews.llvm.org/D85013
Alex Richardson [Mon, 3 Aug 2020 10:18:01 +0000 (11:18 +0100)]
Fix update_cc_test_checks.py --llvm-bin after D78478
Not passing --clang would result in a python exception after this change:
(TypeError: expected str, bytes or os.PathLike object, not NoneType)
because the --clang argument default was only being populated in the
initial argument parsing pass but not later on.
Fix this by adding an argparse callback to set the default values.
Reviewed By: vitalybuka, MaskRay
Differential Revision: https://reviews.llvm.org/D84511
Tatyana Krasnukha [Thu, 30 Jul 2020 17:13:23 +0000 (20:13 +0300)]
[lldb/Process/Windows] Attempting to kill exited/detached process in not an error
The lldb test-suite on Windows reports a 'CLEANUP ERROR' when attempting to kill
an exited/detached process. This change makes ProcessWindows consistent with
the other processes which only log the error. After this change a number of
'CLEANUP ERROR' messages are now removed.
Differential Revision: https://reviews.llvm.org/D84957
Tatyana Krasnukha [Fri, 17 Jul 2020 12:58:28 +0000 (15:58 +0300)]
[cmake] Make MSVC generate appropriate __cplusplus macro definition
Differential Revision: https://reviews.llvm.org/D84023
Alex Richardson [Mon, 3 Aug 2020 09:51:14 +0000 (10:51 +0100)]
Execute llvm-lit with the python found by CMake by default
The check-* targets run ${Python3_EXECUTABLE} $BUILD/bin/llvm-lit, but
running `./bin/llvm-lit $ARGS` from the build directory currently always
uses "python" to run llvm-lit. On most systems this will be python2.7 even
if we found python3 at CMake time.
Reviewed By: compnerd
Differential Revision: https://reviews.llvm.org/D84625
Alex Richardson [Mon, 3 Aug 2020 09:31:31 +0000 (10:31 +0100)]
[msan] Compile the libatomic.c test with a C compiler
Otherwise we end up compiling in C++ mode and on FreeBSD
/usr/include/stdatomic.h is not compatible with C++ since it uses _Bool.
Reviewed By: guiand, eugenis, vitalybuka, emaste
Differential Revision: https://reviews.llvm.org/D84510
Alex Richardson [Mon, 3 Aug 2020 09:04:01 +0000 (10:04 +0100)]
[asan][tsan] Mark tests failing with debug checks as XFAIL
See https://llvm.org/PR46862. This does not fix the underlying issue but at
least it allows me to run check-all again without having to disable
building compiler-rt.
Reviewed By: #sanitizers, vitalybuka
Differential Revision: https://reviews.llvm.org/D84650
Kadir Cetinkaya [Sat, 1 Aug 2020 16:55:30 +0000 (18:55 +0200)]
[clang][Tooling] Fix addTargetAndModeForProgramName to use correct flag names
The logic was using incorrect flag versions. For example:
- `-target=` can't be a prefix, it must be `--target=`.
- `--driver-mode` can't appear on its own, value must be attached to it.
While fixing those, also changes the append logic to make use of new
`--target=X` format instead of the legacy `-target X` version.
In addition to that makes use of the OPTTable instead of hardcoded strings to
make sure helper also gets updated if clang's options are modified.
Differential Revision: https://reviews.llvm.org/D85076
Georgii Rymar [Mon, 27 Jul 2020 13:03:03 +0000 (16:03 +0300)]
[llvm-readobj] - Don't call `unwrapOrErr` in `findSectionByName`.
We have a `findSectionByName` helper that tries to find a section
by it name. It is used in a few places, but never tested.
I'd like to reuse this helper for a different place.
For this, I've changed it to return Expected<> and now it
doesn't use `unwrapOrErr` anymore. It also now a member of
Dumper class and might report warnings.
Differential revision: https://reviews.llvm.org/D84651
Ilya Golovenko [Mon, 3 Aug 2020 09:34:14 +0000 (11:34 +0200)]
Add document outline symbols from unnamed contexts, e.g. extern "C".
It is necessary to traverse children of unnamed declaration contexts
to get symbols which are currently missing in document outline, e.g.:
extern "C" {
void foo();
}
Reviewed By: kadircet
Differential Revision: https://reviews.llvm.org/D84839
Georgii Rymar [Thu, 23 Jul 2020 12:26:23 +0000 (15:26 +0300)]
[yaml2obj] - Add a support for "<none>" value for all optional fields.
It implements an approach suggested in the D84398 thread.
With it the following:
```
Sections:
- Name: .bar
Type: SHT_PROGBITS
Offset: [[MACRO=<none>]]
```
works just like the `Offset` key was not specified.
It is useful for tests that want to have a default value for a field and to
have a way to override it at the same time.
Differential revision: https://reviews.llvm.org/D84526
Julian Gross [Fri, 31 Jul 2020 09:42:31 +0000 (11:42 +0200)]
[mlir] Extended Buffer Assignment to support AllocaOps.
Added support for AllocaOps in Buffer Assignment.
Differential Revision: https://reviews.llvm.org/D85017
Luboš Luňák [Sun, 2 Aug 2020 15:44:24 +0000 (17:44 +0200)]
fix lldb test on lib64 systems
Differential revision: https://reviews.llvm.org/D85096
Florian Hahn [Mon, 3 Aug 2020 08:47:16 +0000 (09:47 +0100)]
[LV] Do not check widening decision for instrs outside of loop.
No widening decisions will be computed for instructions outside the
loop. Do not try to get a widening decision. The load/store will be just
a scalar load, so treating at as normal should be fine I think.
Fixes PR46950.
Reviewed By: dmgreen
Differential Revision: https://reviews.llvm.org/D85087
Nicolas Vasilache [Mon, 3 Aug 2020 08:39:18 +0000 (04:39 -0400)]
[mlir][Vector] Add transformation + pattern to split vector.transfer_read into full and partial copies.
This revision adds a transformation and a pattern that rewrites a "maybe masked" `vector.transfer_read %view[...], %pad `into a pattern resembling:
```
%1:3 = scf.if (%inBounds) {
scf.yield %view : memref<A...>, index, index
} else {
%2 = vector.transfer_read %view[...], %pad : memref<A...>, vector<...>
%3 = vector.type_cast %extra_alloc : memref<...> to
memref<vector<...>> store %2, %3[] : memref<vector<...>> %4 =
memref_cast %extra_alloc: memref<B...> to memref<A...> scf.yield %4 :
memref<A...>, index, index
}
%res= vector.transfer_read %1#0[%1#1, %1#2] {masked = [false ... false]}
```
where `extra_alloc` is a top of the function alloca'ed buffer of one vector.
This rewrite makes it possible to realize the "always full tile" abstraction where vector.transfer_read operations are guaranteed to read from a padded full buffer.
The extra work only occurs on the boundary tiles.
Differential Revision: https://reviews.llvm.org/D84631
Raphael Isemann [Mon, 3 Aug 2020 08:23:48 +0000 (10:23 +0200)]
[debugserver] Fix that is_dot_app is producing unused warnings
Some build configurations don't use this static function.
Frederik Gossen [Mon, 3 Aug 2020 08:18:48 +0000 (08:18 +0000)]
[MLIR][Shape] Lower `shape.broadcast` to `scf`
Differential Revision: https://reviews.llvm.org/D85027
Xing GUO [Mon, 3 Aug 2020 08:14:47 +0000 (16:14 +0800)]
[MachOYAML] Remove redundant variable initialization. NFC.
The value of `is64Bit` is initialized in the constructor body.
Shinji Okumura [Mon, 3 Aug 2020 08:02:49 +0000 (17:02 +0900)]
[Attributor] Check nonnull attribute violation in AAUndefinedBehavior
This patch makes it possible to handle nonnull attribute violation at callsites in AAUndefinedBehavior.
If null pointer is passed to callee at a callsite and the corresponding argument of callee has nonnull attribute, the behavior of the callee is undefined.
In this patch, violations of argument nonnull attributes is only handled.
But violations of returned nonnull attributes can be handled and I will implement that in a follow-up patch.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D84733
Igor Kudrin [Mon, 3 Aug 2020 08:04:15 +0000 (15:04 +0700)]
[DebugInfo] Make DIEDelta::SizeOf() more explicit. NFCI.
The patch restricts DIEDelta::SizeOf() to accept only DWARF forms that
are actually used in the LLVM codebase. This should make the use of the
class more explicit and help to avoid issues similar to fixed in D83958
and D84094.
Differential Revision: https://reviews.llvm.org/D84095
Igor Kudrin [Mon, 3 Aug 2020 08:04:08 +0000 (15:04 +0700)]
[DebugInfo] Fix misleading using of DWARF forms with DIELabel. NFCI.
DIELabel can emit only 32- or 64-bit values, while it was created in
some places with DW_FORM_udata, which implies emitting uleb128.
Nevertheless, these places also expected to emit U32 or U64, but just
used a misleading DWARF form. The patch updates those places to use more
appropriate DWARF forms and restricts DIELabel::SizeOf() to accept only
forms that are actually used in the LLVM codebase.
Differential Revision: https://reviews.llvm.org/D84094
Igor Kudrin [Mon, 3 Aug 2020 08:04:00 +0000 (15:04 +0700)]
[DebugInfo] Fix a comment and a variable name. NFC.
DebugLocListIndex keeps the index of an entry list, not the offset.
Differential Revision: https://reviews.llvm.org/D84093
Igor Kudrin [Mon, 3 Aug 2020 08:03:37 +0000 (15:03 +0700)]
[DebugInfo] Make DIELocList::SizeOf() more explicit. NFCI.
DIELocList is used with a limited number of DWARF forms, see the only
place where it is instantiated, DwarfCompileUnit::addLocationList().
The patch marks the unexpected execution path in DIELocList::SizeOf()
as unreachable, to reduce ambiguity.
Differential Revision: https://reviews.llvm.org/D84092
Daniel Kiss [Mon, 3 Aug 2020 07:40:35 +0000 (09:40 +0200)]
[libunwind] Make the test depend on the libunwind explicitly.
Before this patch the `ninja check-unwind` won't rebuild the unwind library.
Reviewed By: jroelofs
Differential Revision: https://reviews.llvm.org/D85004
Djordje Todorovic [Mon, 3 Aug 2020 07:01:37 +0000 (09:01 +0200)]
[NFC] [MIR] Document the reg state flags
This patch adds documentation for the RegState enumeration.
Differential Revision: https://reviews.llvm.org/D84634
George Mitenkov [Mon, 3 Aug 2020 06:31:08 +0000 (09:31 +0300)]
[MLIR][SPIRV] Control attributes support for loop and selection
This patch handles loopControl and selectionControl in parsing and
printing. In order to reuse the functionality, and avoid handling cases when
`{` of the region is parsed as a dictionary attribute, `control` keyword was
introduced.`None` is a default control attribute. This functionality can be
later extended to `spv.func`.
Also, loopControl and selectionControl can now be (de)serialized.
Reviewed By: antiagainst
Differential Revision: https://reviews.llvm.org/D84175
Fangrui Song [Mon, 3 Aug 2020 06:05:50 +0000 (23:05 -0700)]
[CMake] Default ENABLE_X86_RELAX_RELOCATIONS to ON
This makes clang default to -Wa,-mrelax-relocations=yes, which enables
R_386_GOT32X (GNU as enables it regardless of -mrelax-relocations=) and
R_X86_64_[REX_]GOTPCRELX in MC. The produced object files require GNU ld>=2.26
to link. binutils 2.26 is considered a very old release today.
LLVM GN Syncbot [Mon, 3 Aug 2020 05:55:14 +0000 (05:55 +0000)]
[gn build] Port
160ff83765a
Saiyedul Islam [Mon, 3 Aug 2020 05:29:48 +0000 (05:29 +0000)]
[OpenMP][AMDGCN] Support OpenMP offloading for AMDGCN architecture - Part 3
Provides AMDGCN and NVPTX specific specialization of getGPUWarpSize,
getGPUThreadID, and getGPUNumThreads methods. Adds tests for AMDGCN
codegen for these methods in generic and simd modes. Also changes the
precondition in InitTempAlloca to be slightly more permissive. Useful for
AMDGCN OpenMP codegen where allocas are created with a cast to an
address space.
Reviewed By: ABataev
Differential Revision: https://reviews.llvm.org/D84260
Fangrui Song [Mon, 3 Aug 2020 05:13:59 +0000 (22:13 -0700)]
[MC] Default MCAsmBackend::mayNeedRelaxation() to false
compinder [Mon, 3 Aug 2020 03:27:17 +0000 (08:57 +0530)]
[FLANG] Fix issues in SELECT TYPE construct when intrinsic type specification is specified in TYPE GUARD statement.
Fix of PR46789 and PR46830.
Differential Revision: https://reviews.llvm.org/D84290
QingShan Zhang [Mon, 3 Aug 2020 03:25:20 +0000 (03:25 +0000)]
[NFC][PowerPC] Add a multiclass for fsetcc to define them in a uniform way
This is a refactor patch to prepare for adding the support for strict-fsetcc
in PowerPC backend. We want to move their definition into a uniform way so that,
we could add the strict node easier.
Reviewed By: shchenz
Differential Revision: https://reviews.llvm.org/D81712
StephenFan [Fri, 31 Jul 2020 05:31:48 +0000 (13:31 +0800)]
[RISCV] eliminate the repetition declare of SDLoc DL
Differential revision: https://reviews.llvm.org/D85002
Fangrui Song [Mon, 3 Aug 2020 01:05:17 +0000 (18:05 -0700)]
Reland D64327 [MC][ELF] Allow STT_SECTION referencing SHF_MERGE on REL targets
This drops a GNU gold workaround and reverts the revert commit rL366708.
Before binutils 2.34, gold -O2 and above did not correctly handle R_386_GOTOFF to
SHF_MERGE|SHF_STRINGS sections: https://sourceware.org/bugzilla/show_bug.cgi?id=16794
From the original review:
... it reduced the size of a big ARM-32 debug image by 33%. It contained ~68M
of relocations symbols out of total ~71M symbols (96% of symbols table was
generated for relocations with symbol).
-Wl,-O2 (and -Wl,-O3) is so rare that we should just lower the
optimization level for LLVM_LINKER_IS_GOLD rather than pessimizing all users.
Florian Hahn [Sun, 2 Aug 2020 21:36:53 +0000 (22:36 +0100)]
[Bindings] Remove ipc_propagation.
IPConstantPropagation has been removed, also remove the bindings.
Florian Hahn [Sun, 2 Aug 2020 21:04:44 +0000 (22:04 +0100)]
Recommit "[IPConstProp] Remove and move tests to SCCP."
This reverts commit
59d6e814ce0e7b40b7cc3ab136b9af2ffab9c6f8.
The cause for the revert (3 clang tests running opt -ipconstprop) was
fixed by removing those lines.
Vitaly Buka [Sun, 2 Aug 2020 20:56:38 +0000 (13:56 -0700)]
[StackSafety, NFC] Don't insert empty objects into the map
Result should be the same but it makes generateParamAccessSummary 5x
faster.
Florian Hahn [Sun, 2 Aug 2020 20:45:42 +0000 (21:45 +0100)]
[Clang] Remove run-lines which use opt to run -ipconstprop.
ipconstprop is going to get removed and checking opt with specific
passes makes the tests more fragile.
The tests retain the important checks that !callback metadata is created
correctly.
Jan Kratochvil [Sun, 2 Aug 2020 20:41:02 +0000 (22:41 +0200)]
[lldb] [test] Fix DW_TAG_GNU_call_site-DW_AT_low_pc.s relocation
I have made the DW_FORM_ref4 relative. One could also use relocated
DW_FORM_ref_addr instead.
Tested with:
echo 'void f(){}'|clang -o 1.o -c -Wall -g -x c -;./bin/clang -o 1 1.o ../llvm-monorepo/lldb/test/Shell/SymbolFile/DWARF/DW_TAG_GNU_call_site-DW_AT_low_pc.s;./bin/lldb --no-lldbinit ./1 -o r -o 'p p' -o exit
Craig Topper [Sun, 2 Aug 2020 16:58:55 +0000 (09:58 -0700)]
[X86] Use parity flag from byte test/cmp instruction for __builtin_parity when input fits in 8 bits.
If the upper bits of the __builtin_parity idiom are known to be
0 we were previously emitting an xor with 0 to get the parity flag.
But we can use cmp/test instead which may expose opportunities for
load folding or combining an AND.
Craig Topper [Sun, 2 Aug 2020 07:20:11 +0000 (00:20 -0700)]
[X86] Add test cases for missed opportunity to use a byte test instruction instead of an xor with 0 in parity patterns.
If the input to the ctpop fits in 8 bits, we can use the parity
flag from a TEST instruction, but we're currently XORing with 0.
Simon Pilgrim [Sun, 2 Aug 2020 17:11:01 +0000 (18:11 +0100)]
[AMDGPU] Regenerate tests to fix whitespace indentations
Noticed while updating D77804
Mehdi Amini [Sun, 2 Aug 2020 16:57:02 +0000 (16:57 +0000)]
Remove debug flags from test (NFC)
Simon Pilgrim [Sun, 2 Aug 2020 15:55:16 +0000 (16:55 +0100)]
[IR] Add IRBuilderBase::CreateVectorSplat(ElementCount EC) variant
As discussed on D81500, this adds a more general ElementCount variant of the build helper and converts the (non-scalable) unsigned NumElts variant to use it internally.
Sanjay Patel [Sun, 2 Aug 2020 15:11:05 +0000 (11:11 -0400)]
[InstSimplify] fold max (max X, Y), X --> max X, Y
https://alive2.llvm.org/ce/z/VGgG3M
Sanjay Patel [Sun, 2 Aug 2020 14:53:51 +0000 (10:53 -0400)]
[InstSimplify] add tests for max(max x,y), x) and variants; NFC
Matt Arsenault [Mon, 15 Jun 2020 15:54:49 +0000 (11:54 -0400)]
GlobalISel: Implement bitcast action for G_EXTRACT_VECTOR_ELEMENT
For AMDGPU, vectors with elements < 32 bits should be indexed in
32-bit elements and the desired bits extracted from there. For
elements > 64-bits, these should be reduce to 64/32 elements to enable
the normal dynamic indexing paths.
In the dynamic index cases, this produces shorter code most of the
time. This does immediately regress the constant index cases, but this
should be fixed once we have the most basic of shift combines.
The element size > 64 case is pretty much ported from the exisiting
DAG implementation for extract element promote. The increasing element
size case is new.
Simon Pilgrim [Sun, 2 Aug 2020 14:21:52 +0000 (15:21 +0100)]
X86InstrInfo.cpp - fix include ordering. NFCI.
Simon Pilgrim [Sun, 2 Aug 2020 14:16:00 +0000 (15:16 +0100)]
Use merge null and isa<> tests into isa_and_nonnull<>. NFCI.
Simon Pilgrim [Sun, 2 Aug 2020 14:31:36 +0000 (15:31 +0100)]
[DAG] TargetLowering::expandMUL_LOHI - pass SDLoc as const&
Try to be more consistent with the SDLoc param in the TargetLowering methods.
This also exposes an issue where we were passing a SDNode as a SDLoc, relying on the implicit SDLoc(SDNode) constructor.
Simon Pilgrim [Sun, 2 Aug 2020 14:03:04 +0000 (15:03 +0100)]
[DAG] TargetLowering::LowerAsmOutputForConstraint - pass SDLoc as const&
Try to be more consistent with the SDLoc param in the TargetLowering methods.
Simon Pilgrim [Sun, 2 Aug 2020 13:57:39 +0000 (14:57 +0100)]
Remove unused param tag to fix Wdocumentation warning. NFC.
Shinji Okumura [Sun, 2 Aug 2020 13:48:28 +0000 (22:48 +0900)]
Revert "[Attributor] AAPotentialValues Interface"
The commit cause build failure.
Nikita Popov [Sun, 2 Aug 2020 13:45:24 +0000 (15:45 +0200)]
[InstSimplify] Reduce code duplication in icmp of binop folds (NFC)
For folds where we check for the binop on both the LHS and RHS,
extract a function that expects it on the LHS and call it with
swapped order.
Xing GUO [Sun, 2 Aug 2020 13:37:11 +0000 (21:37 +0800)]
[DWARFYAML][debug_aranges] Make the 'Descriptors' field optional.
Simon Pilgrim [Sun, 2 Aug 2020 13:32:05 +0000 (14:32 +0100)]
[X86] Use const APInt& in for-range loop to avoid unnecessary copies. NFCI.
Fixes clang-tidy warning.
Simon Pilgrim [Sun, 2 Aug 2020 13:23:00 +0000 (14:23 +0100)]
[X86] Pass SDLoc by const reference. NFCI.
Simon Pilgrim [Sun, 2 Aug 2020 13:15:15 +0000 (14:15 +0100)]
[X86] Use const APInt& in for-range loop to avoid unnecessary copies. NFCI.
Fixes clang-tidy warning.
Simon Pilgrim [Sun, 2 Aug 2020 13:13:13 +0000 (14:13 +0100)]
[X86] combineX86ShuffleChain - pull out repeated RootVT.getSizeInBits() calls. NFCI.
Shinji Okumura [Sun, 2 Aug 2020 09:37:07 +0000 (18:37 +0900)]
[Attributor] AAPotentialValues Interface
This is a split patch of D80991.
This patch introduces AAPotentialValues and its interface only.
For more detail of AAPotentialValues abstract attribute, see the original patch.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D83283
Florian Hahn [Fri, 31 Jul 2020 21:10:39 +0000 (22:10 +0100)]
[SCEV] Precommit tests with signed counting down loop.
From PR46939.
Michał Górny [Thu, 30 Jul 2020 22:11:40 +0000 (00:11 +0200)]
[CMake] Pass bugreport URL to standalone clang build
BUG_REPORT_URL is currently used both in LLVM and in Clang but declared
only in the latter. This means that it's missing in standalone clang
builds and the driver ends up outputting:
PLEASE submit a bug report to and include [...]
(note the missing URL)
To fix this, include LLVM_PACKAGE_BUGREPORT in LLVMConfig.cmake
(similarly to how we pass PACKAGE_VERSION) and use it to fill
BUG_REPORT_URL when building clang standalone.
Differential Revision: https://reviews.llvm.org/D84987
Craig Topper [Sun, 2 Aug 2020 05:05:10 +0000 (22:05 -0700)]
[X86] Improve parity idiom recognition to handle (and (truncate (ctpop X)), 1).
Fixes part of PR46954
Craig Topper [Sun, 2 Aug 2020 04:57:24 +0000 (21:57 -0700)]
[X86] Add parity test cases for PR46954.
This adds tests cases where the parity idiom of (and (ctpop X, 1))
has a truncate in the middle.
AK [Sun, 2 Aug 2020 05:34:44 +0000 (22:34 -0700)]
[HotColdSplit] Add test case for unlikely attribute in outlined function
Authored by: rjf (Ruijie Fang)
Reviewed by: hiraditya,tejohnson,fhahn
Differential Revision: https://reviews.llvm.org/D69384
AK [Sun, 2 Aug 2020 02:15:05 +0000 (19:15 -0700)]
Outline non returning functions unless a longjmp
__assert_fail, abort, exit etc. are cold.
TODO: outline throw
Authored by: rjf (Ruijie Fang)
Reviewed by: hiraditya,tejohnson,fhahn
Differential Revision: https://reviews.llvm.org/D69257
Kazu Hirata [Sun, 2 Aug 2020 04:49:38 +0000 (21:49 -0700)]
Use llvm::is_contained where appropriate (NFC)
Use llvm::is_contained where appropriate (NFC)
Reviewed By: kazu
Differential Revision: https://reviews.llvm.org/D85083
Lang Hames [Sun, 2 Aug 2020 00:44:34 +0000 (17:44 -0700)]
[llvm-jitlink] Add -phony-externals option to suppress unresolved externals.
The -phony-externals option adds a generator which explicitly defines any
otherwise unresolved externals as null. This transforms link-time
unresolved-symbol errors into potential runtime null pointer accesses
(if an unresolved external is actually accessed during execution).
This option can be useful in -harness mode to avoid having to mock a
large number of symbols that are not reachable at runtime (e.g. unused
methods referenced by a class vtable).
Lang Hames [Sat, 1 Aug 2020 04:32:27 +0000 (21:32 -0700)]
[llvm-jitlink] Support promotion of ODR weak symbols in -harness mode.
This prevents weak symbols from being immediately dead-stripped when not
directly referenced from the test harneess, enabling use of weak symbols
from the code under test.
Fangrui Song [Sun, 2 Aug 2020 01:19:14 +0000 (18:19 -0700)]
[ELF] --wrap: set isUsedInRegularObj of __wrap_ only if it is defined
Fixes PR46169
Evgenii Stepanov [Sat, 1 Aug 2020 00:18:06 +0000 (17:18 -0700)]
[msan] Respect no_huge_pages_for_shadow.
Disable huge pages in the MSan shadow region when
no_huge_pages_for_shadow == true (default).
Differential Revision: https://reviews.llvm.org/D85061
Andrei Lebedev [Sat, 1 Aug 2020 22:54:11 +0000 (15:54 -0700)]
Updated the -I option description.
Nikita Popov [Thu, 30 Jul 2020 17:14:02 +0000 (19:14 +0200)]
[NewPM][LVI] Abandon LVI after CVP
As mentioned on D70376, LVI can currently cause performance issues
when running under NewPM. The problem is that, unlike the legacy
pass manager, NewPM will not immediately discard the LVI analysis
if the following pass does not need it. This is a problem, because
LVI has a high memory requirement, and mass invalidation of LVI
values is very inefficient. LVI should only be alive during passes
that actively interact with it.
This patch addresses the issue by explicitly abandoning LVI after CVP,
which gets us back to the LegacyPM behavior.
Differential Revision: https://reviews.llvm.org/D84959
Craig Topper [Sat, 1 Aug 2020 20:25:18 +0000 (13:25 -0700)]
[X86] Add assembler support for {disp8} and {disp32} to control the size of displacement used for memory operands.
These prefixes should override the default behavior and force a larger immediate size. I don't believe gas issues any warning if you use {disp8} when a 32-bit displacement is already required. And this patch doesn't either.
This completes the {disp8} and {disp32} support from PR46650.
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D84793
Craig Topper [Sat, 1 Aug 2020 17:01:58 +0000 (10:01 -0700)]
[InstSimplify] Fold abs(abs(x)) -> abs(x)
It's always safe to pick the earlier abs regardless of the nsw flag. We'll just lose it if it is on the outer abs but not the inner abs.
Differential Revision: https://reviews.llvm.org/D85053
Craig Topper [Sat, 1 Aug 2020 16:59:09 +0000 (09:59 -0700)]
[InstCombine] Fold abs(-x) -> abs(x)
Negating the input doesn't matter. I left a FIXME to copy the nsw flag if its present on the neg but not on the abs.
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D85055
Florian Hahn [Sat, 1 Aug 2020 19:58:05 +0000 (20:58 +0100)]
[PPC] Adjust run line for hardware-loops-crash.ll
Looks like %s was accidentally dropped.
Florian Hahn [Sat, 1 Aug 2020 19:38:38 +0000 (20:38 +0100)]
[LCSSA] Provide option for caller to clean up unused PHIs.
formLCSSAForInstructions is used by SCEVExpander, which tracks all
inserted instructions including LCSSA phis using asserting value
handles. This means cleanup needs to happen in the caller.
Extend formLCSSAForInstructions to take an optional pointer to a
vector. If this argument is non-nullptr, instead of directly deleting
the phis, add them to the vector, so the caller can process them.
This should address various PPC buildbot failures, including
http://lab.llvm.org:8011/builders/clang-ppc64be-linux-lnt/builds/40567
Simon Pilgrim [Sat, 1 Aug 2020 19:34:39 +0000 (20:34 +0100)]
[X86][AVX512] Fold concat(and(x,y),and(z,w)) -> and(concat(x,z),concat(y,w)) for 512-bit vectors
Helps vpternlog folding on non-AVX512BW targets
zoecarver [Sat, 1 Aug 2020 19:06:31 +0000 (12:06 -0700)]
[libcxx] Add compatible with constraint tests for some shared_ptr constructors.
Add shared_ptr tests where the element type and pointer type aren't 'convertible' but are 'compatible'.
Responding to a comment from D81414.
Differential Revision: https://reviews.llvm.org/D81532
Simon Pilgrim [Sat, 1 Aug 2020 13:54:21 +0000 (14:54 +0100)]
[X86][AVX] Ensure we only combine to PSHUFLW/PSHUFHW on supporting targets
Noticed while investigating combining from concatenated shuffle vectors, we weren't checking that PSHUFLW/PSHUFHW was legal - we were depending on lowering splitting to subvectors.
Florian Hahn [Sat, 1 Aug 2020 16:54:23 +0000 (17:54 +0100)]
[LCSSA] Use IRBuilder for PHI creation.
Use IRBuilder instead PHINode::Create. This should not impact the
generated code, but IRBuilder provides a way to register callbacks for
inserted instructions, which is convenient for some users.
Reviewed By: lebedev.ri
Differential Revision: https://reviews.llvm.org/D85037
Sanjay Patel [Sat, 1 Aug 2020 12:38:28 +0000 (08:38 -0400)]
[VectorCombine] add tests for non-zero gep offsets; NFC
Luofan Chen [Sat, 1 Aug 2020 13:27:16 +0000 (21:27 +0800)]
[Attributor][NFC] Update description for the dependency graph
The word "dependency graph" is a bit misleading. When there is an
edge from node A to B (A -> B), it actually mean that B depends on
A and when the state of A is updated, B should also be updated. So
I update the comment to make the description clearer.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D85065
David Green [Sat, 1 Aug 2020 13:01:18 +0000 (14:01 +0100)]
[ARM] Distribute post-inc for Thumb2 sign/zero extending loads/stores
This adds sign/zero extending scalar loads/stores to the MVE
instructions added in D77813, allowing us to create up more post-inc
instructions. These are comparatively simple, compared to LDR/STR (which
may be better turned into an LDRD/LDM), but still require some additions
over MVE instructions. Because there are i12 and i8 variants of the
offset loads/stores dealing with different signs, we may need to convert
an i12 address to a i8 negative instruction. t2LDRBi12 can also be
shrunk to a tLDRi under the right conditions, so we need to be careful
with codesize too.
Differential Revision: https://reviews.llvm.org/D78625
Benjamin Kramer [Sat, 1 Aug 2020 12:48:42 +0000 (14:48 +0200)]
[mlir][Vector] Simplify code a bit. NFCI.
Sanjay Patel [Sat, 1 Aug 2020 11:46:23 +0000 (07:46 -0400)]
[InstSimplify] simplify abs if operand is known non-negative
abs() should be rare enough that using value tracking is not going
to be a compile-time cost burden, so use it to reduce a variety of
potential patterns. We do this in DAGCombiner too.
Differential Revision: https://reviews.llvm.org/D85043
Sanjay Patel [Sat, 1 Aug 2020 11:44:33 +0000 (07:44 -0400)]
[InstSimplify] add abs test with assume; NFC
Simon Pilgrim [Sat, 1 Aug 2020 11:28:10 +0000 (12:28 +0100)]
[X86][AVX] Extend v2f64 BROADCAST(LOAD) -> BROADCAST_LOAD to v2i64/v4f32/v4i32
Minor precursor fix for D66004, but helps the SSE41 tests as well as they run with -disable-peephole
Nathan James [Sat, 1 Aug 2020 10:04:29 +0000 (11:04 +0100)]
[clang-tidy][NFC] Small refactor
Evgeny Leviant [Sat, 1 Aug 2020 09:58:52 +0000 (12:58 +0300)]
[MachineVerifier] Refactor calcRegsPassed. NFC
Patch improves performance of verify-machineinstrs pass up to 10x.
Differential revision: https://reviews.llvm.org/D84105
Nathan James [Sat, 1 Aug 2020 09:35:13 +0000 (10:35 +0100)]
[clang-tidy] readability-identifier-naming checks configs for included files
When checking for the style of a decl that isn't in the main file, the check will now search for the configuration that the included files uses to gather the style for its decls.
This can be useful to silence warnings in header files that follow a different naming convention without using header-filter to silence all warnings(even from other checks) in the header file.
Reviewed By: aaron.ballman, gribozavr2
Differential Revision: https://reviews.llvm.org/D84814