Tres Popp [Fri, 5 Jun 2020 09:22:10 +0000 (11:22 +0200)]
[mlir] Move attr-dict to end of Shape dialect ops
Summary: This is to be consistent with most other assemblyFormat examles in MLIR.
Differential Revision: https://reviews.llvm.org/D81245
Lang Hames [Fri, 5 Jun 2020 21:08:15 +0000 (14:08 -0700)]
[ORC] Always use ObjectLinkingLayer/JITLink for MachO on x86-64 and arm64.
JITLink supports all code and relocation models, so there's no reason to
conditionalize using JITLink on the code or relocation model settings.
Clients wanting to use RTDyldObjectLinkingLayer/RuntimeDyld will now
need to use a custom object linking layer creator.
Wen-Heng (Jack) Chung [Fri, 5 Jun 2020 21:26:57 +0000 (16:26 -0500)]
[mlir][mlir-rocm-runner] Detect HIP version and AMD ISA version.
Summary:
Prior to ROCm / HIP 3.5, HSA code object V2 is the default binary format.
Starting from ROCm 3.5, HSA code object V3 becomes default.
Also invoke `rocm_agent_enumerator` to detect proper AMD ISA version on the
system. Use `gfx900` as the fallback value.
Reviewers: jerryyin yaxunl
Subscribers: mgorny, yaxunl, mehdi_amini, rriddle, jpienaar, shauheen, antiagainst, nicolasvasilache, arpith-jacob, mgester, lucyrfox, aartbik, liufengdb, stephenneuendorffer, Joonsoo, grosul1, frgossen, Kayjukh, jurahul, msifontes
Tags: #mlir
Differential Revision: https://reviews.llvm.org/D81309
LLVM GN Syncbot [Sat, 6 Jun 2020 01:50:36 +0000 (01:50 +0000)]
[gn build] Port
29203480632
Nico Weber [Sat, 6 Jun 2020 01:50:04 +0000 (21:50 -0400)]
Revert "recommit "[HIP] Add default header and include path""
This reverts commit
1fa43e0b34d9736f62c6c1b6c371a5e39cd1624d.
Still breaks tests on several bots, see https://reviews.llvm.org/D81176
Nico Weber [Sat, 6 Jun 2020 01:20:11 +0000 (21:20 -0400)]
Revert "Migrate the rest of COFFObjectFile to Error"
This reverts commit
b5289656b865d2a73cf90819e20a96fb8414ab0b.
__attribute__((optnone)) doesn't build with msvc, see
http://lab.llvm.org:8011/builders/clang-x64-windows-msvc/builds/16326
LLVM GN Syncbot [Sat, 6 Jun 2020 00:47:43 +0000 (00:47 +0000)]
[gn build] Port
1fa43e0b34d
Jessica Paquette [Fri, 5 Jun 2020 00:08:36 +0000 (17:08 -0700)]
[AArch64][GlobalISel] Move dup optimization into post-legalizer combiner
Since all of the other G_SHUFFLE_VECTOR transforms are going there, let's do
this with dup as well. This is nice, because it lets us split up the original
code into matching, register bank selection, and instruction selection.
- Create G_DUP, make it equivalent to AArch64dup
- Add a post-legalizer combine which is 90% a copy-and-paste from
tryOptVectorDup, except with shuffle matching closer to what SelectionDAG
does in `ShuffleVectorSDNode::isSplatMask`.
- Teach RegBankSelect about G_DUP. Since dup selection relies on the correct
register bank for FP/GPR dup selection, this is necessary.
- Kill `tryOptVectorDup`, since it's now entirely handled by G_DUP.
- Add testcases for the combine, RegBankSelect, and selection. The selection
test gives the same selection results as the old test.
Differential Revision: https://reviews.llvm.org/D81221
Diego Caballero [Sat, 6 Jun 2020 00:29:56 +0000 (17:29 -0700)]
[mlir] Fix representation of BF16 constants
This patch is a follow-up on https://reviews.llvm.org/D81127
BF16 constants were represented as 64-bit floating point values due to the lack
of support for BF16 in APFloat. APFloat was recently extended to support
BF16 so this patch is fixing the BF16 constant representation to be 16-bit.
Reviewed By: rriddle
Differential Revision: https://reviews.llvm.org/D81218
Yaxun (Sam) Liu [Fri, 5 Jun 2020 20:49:38 +0000 (16:49 -0400)]
recommit "[HIP] Add default header and include path"
recommit
11d06b9511bd25aabbfad10dff548b0ce29135a5 with
fix for lit tests.
Diego Caballero [Sat, 6 Jun 2020 00:06:42 +0000 (17:06 -0700)]
Fix convertBFloatAPFloatToAPInt for NaN/Inf values
Bfloat type has an 8-bit exponent so the exponent of NaN/Inf numbers
must be 0xff instead of 0x1f. This is probably a copy-paste mistake
from the half float type.
Reviewed By: lattner
Differential Revision: https://reviews.llvm.org/D81302
Richard Smith [Mon, 4 May 2020 23:56:47 +0000 (16:56 -0700)]
PR34581: Don't remove an 'if (p)' guarding a call to 'operator delete(p)' under -Oz.
Summary:
This transformation is correct for a builtin call to 'free(p)', but not
for 'operator delete(p)'. There is no guarantee that a user replacement
'operator delete' has no effect when called on a null pointer.
However, the principle behind the transformation *is* correct, and can
be applied more broadly: a 'delete p' expression is permitted to
unconditionally call 'operator delete(p)'. So do that in Clang under
-Oz where possible. We do this whether or not 'p' has trivial
destruction, since the destruction might turn out to be trivial after
inlining, and even for a class-specific (but non-virtual,
non-destroying, non-array) 'operator delete'.
Reviewers: davide, dnsampaio, rjmccall
Reviewed By: dnsampaio
Subscribers: hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D79378
stevewan [Fri, 5 Jun 2020 23:54:00 +0000 (19:54 -0400)]
[PowerPC] Do not special case Darwin on PowerPC in target cpu handling
Summary: This patch removes the special handling for Darwin on PowerPC in the default target cpu handling, because Darwin is no longer supported on the PowerPC platform.
Reviewers: hubert.reinterpretcast, daltenty
Reviewed By: hubert.reinterpretcast
Subscribers: wuzish, nemanjai, shchenz, steven.zhang, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D81115
Philip Reames [Fri, 5 Jun 2020 21:48:52 +0000 (14:48 -0700)]
[Tests] Migrate a number of tests to gc-live bundle representation
Reid Kleckner [Fri, 5 Jun 2020 20:56:51 +0000 (13:56 -0700)]
Migrate the rest of COFFObjectFile to Error
Reid Kleckner [Fri, 5 Jun 2020 23:12:38 +0000 (16:12 -0700)]
Re-land "Migrate Binary::checkOffset from error_code to Error, NFC"
This reverts commit
38f3ba591e3a64fa5bbe684b3171c7bda6c5b527.
Fix the XCOFF error handling. Unlike std::error_code, Error must be
consumed or handled.
Fangrui Song [Fri, 5 Jun 2020 23:16:49 +0000 (16:16 -0700)]
Fix -Wunused-variable in -DLLVM_ENABLE_ASSERTIONS=off builds after
ebcbd5ba39c017bb621eefa3175a224aae85ddc8
Richard Smith [Fri, 5 Jun 2020 02:16:05 +0000 (19:16 -0700)]
PR46209: properly determine whether a copy assignment operator is
trivial.
We previously took a shortcut by assuming that if a subobject had a
trivial copy assignment operator (with a few side-conditions), we would
always invoke it, and could avoid going through overload resolution.
That turns out to not be correct in the presenve of ref-qualifiers (and
also won't be the case for copy-assignments with requires-clauses
either). Use the same logic for lazy declaration of copy-assignments
that we use for all other special member functions.
Previously committed as
c57f8a3a20540fcf9fbf98c0a73f381ec32fce2a. This
now also includes an extension of LLDB's workaround for handling special
members without the help of Sema to cover copy assignments.
Richard Smith [Fri, 5 Jun 2020 02:15:22 +0000 (19:15 -0700)]
Set the captures on a CXXRecordDecl representing a lambda closure type
before marking it complete.
No functionality change intended.
Previously committed as
c13dd74e311d2ac70dd3ea663d800307d1aa5b6b.
Fangrui Song [Fri, 5 Jun 2020 22:59:34 +0000 (15:59 -0700)]
[ELF] Don't cause assertion failure if --dynamic-list or --version-script takes an empty file
Fixes PR46184
Report line 1 of the last memory buffer.
Alex Lorenz [Fri, 5 Jun 2020 22:40:00 +0000 (15:40 -0700)]
[arm][darwin] Don't generate libcalls for wide shifts on Darwin
Similar to
ceb801612a678bdffe7e7bf163bd0eed9c9b73a2.
Darwin doesn't always use compiler-rt, and so we can't assume that these
functions are available on arm.
Davide Italiano [Fri, 5 Jun 2020 21:51:58 +0000 (14:51 -0700)]
[IRExecutionUnit] Early returns for clarity. NFCI.
Nicolas Vasilache [Fri, 5 Jun 2020 21:44:02 +0000 (17:44 -0400)]
[mlir][Linalg] NFC - Make useFullTileBuffersByDefault option take a boolean.
Dan Gohman [Fri, 5 Jun 2020 21:27:52 +0000 (14:27 -0700)]
[WebAssembly] Improve clang diagnostics for wasm attributes
This patch addresses the review comments on r352930:
- Removes redundant diagnostic checking code
- Removes errnoneous use of diag::err_alias_is_definition, which
turned out to be ineffective anyway since functions can be defined later
in the translation unit and avoid detection.
- Adds a test for various invalid cases for import_name and import_module.
This reapplies D59520, with the addition of adding
`InGroup<IgnoredAttributes>` to the new warnings, to fix the
Misc/warning-flags.c failure.
Differential Revision: https://reviews.llvm.org/D59520
Craig Topper [Fri, 5 Jun 2020 21:27:04 +0000 (14:27 -0700)]
[X86] Add test cases for PR46203. NFC
Nicolas Vasilache [Fri, 5 Jun 2020 21:23:15 +0000 (17:23 -0400)]
[mlir] Fix spurious f64 -> f16 change in CPU runner test
Nikita Popov [Fri, 5 Jun 2020 21:18:26 +0000 (23:18 +0200)]
[CGP] Remove unnecessary MaybeAlign use (NFC)
Stores now always have an alignment.
Nikita Popov [Fri, 5 Jun 2020 21:11:04 +0000 (23:11 +0200)]
[LoopIdiomRecognize] Remove unnecessary MaybeAlign use (NFC)
Loads and stores always have an alignment now.
Marco Elver [Fri, 5 Jun 2020 20:56:24 +0000 (22:56 +0200)]
[ASan][Test] Fix globals test for Mach-O
Summary: Use a portable section name, as for the test's purpose any name will do.
Reviewers: nickdesaulniers, thakis
Reviewed By: thakis
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D81306
Nico Weber [Fri, 5 Jun 2020 21:00:20 +0000 (17:00 -0400)]
Revert "Migrate Binary::checkOffset from error_code to Error, NFC"
This reverts commit
74bd98829d82312676a60c5c2d142e20691b2f13.
Breaks LLVM::section-headers.test everywhere, see e.g.
http://lab.llvm.org:8011/builders/clang-x86_64-debian-fast/builds/29940/steps/test-check-all/logs/FAIL%3A%20LLVM%3A%3Asection-headers.test
Adrian Prantl [Fri, 5 Jun 2020 18:58:05 +0000 (11:58 -0700)]
Fix an oversight in GetXcodeContentsDirectory()
Since FindXcodeContentsDirectoryInPath expects the *.app/Contents and
DEVELOPER_DIR is supposed to point to Xcode.app, we need to append the
Contents path first.
Differential Revision: https://reviews.llvm.org/D81290
Adrian Prantl [Fri, 5 Jun 2020 18:38:59 +0000 (11:38 -0700)]
Factor out GetEnvDeveloperDir() (NFC)
Differential Revision: https://reviews.llvm.org/D81289
Craig Topper [Fri, 5 Jun 2020 18:46:07 +0000 (11:46 -0700)]
[X86] Fold undef elts to 0 in getTargetVShiftByConstNode.
Similar to D81212.
Differential Revision: https://reviews.llvm.org/D81292
Reid Kleckner [Fri, 5 Jun 2020 20:32:33 +0000 (13:32 -0700)]
Migrate Binary::checkOffset from error_code to Error, NFC
In my use case, this saved 100ms of time doing one-time-initialization
for std::error_code().
Stanislav Mekhanoshin [Wed, 27 May 2020 20:09:00 +0000 (13:09 -0700)]
AMDGPU/GlobalISel: cmp/select method for extract element
Differential Revision: https://reviews.llvm.org/D80749
Reid Kleckner [Tue, 2 Jun 2020 00:11:06 +0000 (17:11 -0700)]
[codeview] Put !heapallocsite on calls to operator new
Clang marks calls to operator new as heap allocation sites, but the
operator declared at global scope returns a void pointer. There is no
explicit cast in the code, so the compiler has to write down the
allocated type itself.
Also generalize a cast to use CallBase, so that we mark heap alloc sites
when exceptions are enabled.
Reviewed By: MaskRay
Differential Revision: https://reviews.llvm.org/D80966
LLVM GN Syncbot [Fri, 5 Jun 2020 19:48:20 +0000 (19:48 +0000)]
[gn build] Port
8a8c6913a93
Stanislav Mekhanoshin [Mon, 1 Jun 2020 21:35:12 +0000 (14:35 -0700)]
SROA: Remove pointer from visited along with instruction
If an instruction is erased we also need to remove it from
Visited set. There is a very small chance that an another
newly created instruction will be created with the same
pointer value in place of an erased one.
Differential Revision: https://reviews.llvm.org/D80958
Yaxun (Sam) Liu [Fri, 5 Jun 2020 19:41:00 +0000 (15:41 -0400)]
Revert "[HIP] Add default header and include path"
This reverts commit
11d06b9511bd25aabbfad10dff548b0ce29135a5.
Yaxun (Sam) Liu [Fri, 5 Jun 2020 19:31:11 +0000 (15:31 -0400)]
Attempt to fix hip-include-path.hip
Saleem Abdulrasool [Fri, 5 Jun 2020 00:40:41 +0000 (17:40 -0700)]
lld: use modern library search ordering
This merges the static and shared library and behaves as if
`-search_paths_first` was specified which is also the default behaviour
on ld64 (and now lld). Unify the paths, and use `llvm::sys::path` to
deal with the path to be truly agnostic to the host.
Alexandre Ganea [Thu, 4 Jun 2020 19:30:45 +0000 (15:30 -0400)]
[llvm-pdbutil] Add missing --id-stats to dump the IPI stream
Before this patch, llvm-pdbutil supported only --type-stats to dump stats about a PDB TPI stream.
Adding --id-stats for completion.
Matt Arsenault [Sun, 31 May 2020 00:28:36 +0000 (20:28 -0400)]
GlobalISel: Add helper for constructing load from offset
Adrian Prantl [Fri, 5 Jun 2020 02:02:11 +0000 (19:02 -0700)]
Teach GetXcodeSDK to look in the Xcode that contains LLDB
instead of preferring the one chosen with xcode-select.
<rdar://problem/
64000666>
Differential Revision: https://reviews.llvm.org/D81210
Adrian Prantl [Fri, 5 Jun 2020 00:57:11 +0000 (17:57 -0700)]
Move GetXcode*Directory into HostInfo (NFC)
These functions really don't belong into PlatformDarwin, since they
actualy query state of the Host and not of the remote platform.
Matt Arsenault [Thu, 4 Jun 2020 02:06:49 +0000 (22:06 -0400)]
GlobalISel: Make known bits/alignment API more consistent
Just computing the alignment makes sense without caring about the
general known bits, such as for non-integral pointers. Separate the
two and start calling into the TargetLowering hooks for frame indexes.
Start calling the TargetLowering implementation for FrameIndexes,
which improves the AMDGPU matching for stack addressing modes. Also
introduce a new hook for returning known alignment of target
instructions. For AMDGPU, it would be useful to report the known
alignment implied by certain intrinsic calls.
Also stop using MaybeAlign.
Matt Arsenault [Sat, 16 May 2020 14:41:35 +0000 (10:41 -0400)]
GlobalISel: Add G_EXTRACT/G_INSERT offset to legalize info
Immediate legalize fields were added for G_SEXT_INREG. Simiarly, these
are likely not legal except for certain offsets.
Dan Gohman [Fri, 5 Jun 2020 18:52:11 +0000 (11:52 -0700)]
Revert "[WebAssembly] Improve clang diagnostics for wasm attributes"
It broke clang-check.
This reverts commit
931fcd3ba011c27626a95e5e35b657763284882a.
Jay Foad [Fri, 5 Jun 2020 16:12:32 +0000 (17:12 +0100)]
Correctly report modified status for AMDGPUUnifyDivergentExitNodes
Related to https://reviews.llvm.org/D80916
Differential Revision: https://reviews.llvm.org/D81271
Fangrui Song [Fri, 5 Jun 2020 18:46:50 +0000 (11:46 -0700)]
Fix -Wunused-variable after D80515
Nico Weber [Fri, 5 Jun 2020 18:39:04 +0000 (14:39 -0400)]
[gn build] (manually) merge
5eedf07ab999d
Yaxun (Sam) Liu [Fri, 5 Jun 2020 18:36:17 +0000 (14:36 -0400)]
Attempt to fix hip-include-path.hip
Matt Arsenault [Fri, 5 Jun 2020 17:57:29 +0000 (13:57 -0400)]
AMDGPU: Fix incorrect selection of buffer atomic fadd
There were additional standalone patterns for these nodes which were
missing the subtarget predicate.
Matt Arsenault [Fri, 5 Jun 2020 17:48:59 +0000 (13:48 -0400)]
AMDGPU: Fix missing immarg on buffer.atomic.fadd intrinsic
Craig Topper [Fri, 5 Jun 2020 17:49:21 +0000 (10:49 -0700)]
[X86] Teach combineVectorShiftImm to constant fold undef elements to 0 not undef.
Shifts are supposed to always shift in zeros or sign bits regardless of their inputs. It's possible the input value may have been replaced with undef by SimplifyDemandedBits, but the shift in zeros are still demanded.
This issue was reported to me by ispc from 10.0. Unfortunately their failing test does not fail on trunk. Seems to be because the shl is optimized out earlier now and doesn't become VSHLI.
ispc bug https://github.com/ispc/ispc/issues/1771
Differential Revision: https://reviews.llvm.org/D81212
Dan Gohman [Sat, 21 Dec 2019 06:11:03 +0000 (22:11 -0800)]
[WebAssembly] Improve clang diagnostics for wasm attributes
This patch addresses the review comments on r352930:
- Removes redundant diagnostic checking code
- Removes errnoneous use of diag::err_alias_is_definition, which
turned out to be ineffective anyway since functions can be defined later
in the translation unit and avoid detection.
- Adds a test for various invalid cases for import_name and import_module.
Differential Revision: https://reviews.llvm.org/D59520
Evgenii Stepanov [Fri, 5 Jun 2020 17:57:44 +0000 (10:57 -0700)]
[hwasan] Disable malloc-fill by default.
Summary: Non-zero malloc fill is causing way too many hard to debug issues.
Reviewers: kcc, pcc, hctim
Subscribers: #sanitizers, llvm-commits
Tags: #sanitizers
Differential Revision: https://reviews.llvm.org/D81284
Marco Elver [Fri, 29 May 2020 10:52:27 +0000 (12:52 +0200)]
[KernelAddressSanitizer] Make globals constructors compatible with kernel
Summary:
This makes -fsanitize=kernel-address emit the correct globals
constructors for the kernel. We had to do the following:
- Disable generation of constructors that rely on linker features such
as dead-global elimination.
- Only emit constructors for globals *not* in explicit sections. The
kernel uses sections for special globals, which we should not touch.
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=203493
Tested:
1. With 'clang/test/CodeGen/asan-globals.cpp'.
2. With test_kasan.ko, we can see:
BUG: KASAN: global-out-of-bounds in kasan_global_oob+0xb3/0xba [test_kasan]
Reviewers: glider, andreyknvl
Reviewed By: glider
Subscribers: cfe-commits, nickdesaulniers, hiraditya, llvm-commits
Tags: #llvm, #clang
Differential Revision: https://reviews.llvm.org/D80805
Nikita Popov [Sat, 30 May 2020 14:07:44 +0000 (16:07 +0200)]
[LiveDebugValues] Fix output stream (NFC)
This should dump to the provided Out, rather than dbgs(), though
they coincide in current usage.
Nikita Popov [Sat, 30 May 2020 14:13:41 +0000 (16:13 +0200)]
[LiveDebugValues] Remove PendingInLocs (NFC)
PendingInLocs ends up having the same value as InLocs, just computed
a bit more indirectly. It is a leftover of a previous implementation
approach.
This patch drops PendingInLocs, as well as the Diff and Removed
calulations, which are no longer needed.
Differential Revision: https://reviews.llvm.org/D80868
Matt Arsenault [Sun, 31 May 2020 18:59:52 +0000 (14:59 -0400)]
AMDGPU/GlobalISel: Fix some broken YAML in MIR test
Different tools seem to be more or less strict with the markers
between different functions.
Matt Arsenault [Sun, 31 May 2020 18:19:46 +0000 (14:19 -0400)]
GlobalISel: Infer nofpexcept flag during selection for non-strict ops
Match SelectionDAG's behavior of adding nofpexcept to out instructions
that may raise fp exceptions that are selected from instructions that
do not.
Alex Lorenz [Fri, 5 Jun 2020 17:54:03 +0000 (10:54 -0700)]
[apple clang] disable in-process CC1 to preserve crashlog compatibility
The in-process CC1 currently doesn't interoperate with the macOS crash analytics,
which we would like to keep enabled for Apple clang. This commit restores the
out-of-process CC1 to the Apple clang CMake configuration for now.
Differential Revision: https://reviews.llvm.org/D80849
Sameer Arora [Fri, 5 Jun 2020 17:47:05 +0000 (10:47 -0700)]
[llvm-objcopy] Reorder --dump-section for MachO
Reorder `DumpSection` under `handleArgs` in file `MachOObjcopy.cpp`. The
operation to dump a section is now performed before both add and remove
section operations for MachO file format.
Change for the ELF format at D81097. Together fixes
https://bugs.llvm.org/show_bug.cgi?id=44283
Reviewed By: jhenderson
Differential Revision: https://reviews.llvm.org/D81123
Nicolas Vasilache [Fri, 5 Jun 2020 17:41:05 +0000 (13:41 -0400)]
[mlir][test][CPU] Reduce the size of mlir-cpu-runner-tests
Two tests regularly show up on the long tail when testing MLIR.
This revision reduces their size.
Sameer Arora [Fri, 5 Jun 2020 17:42:19 +0000 (10:42 -0700)]
[llvm-objcopy] Reorder --dump-section before --remove-section for ELF
Reorder `DumpSection` under `handleArgs` in file `ELFObjcopy.cpp`.
`DumpSection` is placed before `replaceAndRemoveSections` and is
therefore now the first operation under `handleArgs`. Thus, it is now
performed before both `add` and `remove` section operations.
Change for the MachO format at D81123. Together fixes https://bugs.llvm.org/show_bug.cgi?id=44283.
Reviewed By: alexshap, jhenderson, MaskRay
Differential Revision: https://reviews.llvm.org/D81097
Louis Dionne [Fri, 5 Jun 2020 17:39:49 +0000 (13:39 -0400)]
[libc++] Avoid passing -xc++ in Lit feature detection
Otherwise, if %{flags} contain other files like static libraries, those
files are treated as C++ source files instead of object files, and the
compiler gets all confused.
David Green [Fri, 5 Jun 2020 14:34:15 +0000 (15:34 +0100)]
[ARM] VQMOVN demand bits analysis
Similar to VMOVN, a VQMOVN will only demand the top/bottom lanes of it's
first input. However unlike VMOVN it will need access to the entire
second argument, as that value is saturated not just moved in place.
Differential Revision: https://reviews.llvm.org/D80515
Sameer Arora [Fri, 5 Jun 2020 17:29:24 +0000 (10:29 -0700)]
[llvm-ar] Update error messages and tests as per latest preferred style
It updates two error messages under `performOperation` in the file
llvm-ar.cpp. Furthermore, it also updates tests that print out these
error messages: `llvm/test/Object/ar-create.test` and
`llvm/test/tools/llvm-ar/print.test`.
Reviewed By: jhenderson, MaskRay
Differential Revision: https://reviews.llvm.org/D80846
Sameer Arora [Fri, 5 Jun 2020 17:29:18 +0000 (10:29 -0700)]
[llvm-ar] Add more tests for errors in opening archives
Add 2 more tests for the function performOperation under llvm-ar.cpp. It
tests the scenario when the archive could not be opened for reasons
other than `no_such_file_or_directory` In particular, it tests for the
cases `permission_denied` and `is_a_directory` for the target archive.
Reviewed By: alexshap, jhenderson, smeenai, MaskRay
Differential Revision: https://reviews.llvm.org/D80838
Nicolas Vasilache [Fri, 5 Jun 2020 17:20:59 +0000 (13:20 -0400)]
[mlir][Vector] Add folding of memref_cast into vector_transfer ops
Summary:
This revision adds a common folding pattern that starts appearing on
vector_transfer ops.
Differential Revision: https://reviews.llvm.org/D81281
LLVM GN Syncbot [Fri, 5 Jun 2020 17:22:19 +0000 (17:22 +0000)]
[gn build] Port
11d06b9511b
Yaxun (Sam) Liu [Fri, 5 Jun 2020 17:21:24 +0000 (13:21 -0400)]
Fix lit tests hip-include-path.hip and hip-host-cpu-macros.cu
Sander de Smalen [Fri, 5 Jun 2020 15:46:25 +0000 (16:46 +0100)]
Reland D80640: [CodeGen][SVE] Calculate correct type legalization for scalable vectors.
This reverts commit
9bcef270d7a319c6c0fdffc6c80984a8f0a30ecb.
Louis Dionne [Fri, 5 Jun 2020 17:06:37 +0000 (13:06 -0400)]
[libc++abi] Properly fix XFAILs for exception alignment
Since <unwind.h> is in the SDK, not in /usr/include, the XFAILs must
be predicated on the compiler version (ideally even on the SDK version)
instead of the target system version.
Nicolas Vasilache [Fri, 5 Jun 2020 17:02:13 +0000 (13:02 -0400)]
[mlir][Linalg] NFC - Cleanup debug, address post-commit review.
Jacques Pienaar [Fri, 5 Jun 2020 16:47:37 +0000 (09:47 -0700)]
[mlir] Add verify method to adaptor
This allows verifying op-indepent attributes (e.g., attributes that do not require the op to have been created) before constructing an operation. These include checking whether required attributes are defined or constraints on attributes (such as I32 attribute). This is not perfect (e.g., if one had a disjunctive constraint where one part relied on the op and the other doesn't, then this would not try and extract the op independent from the op dependent).
The next step is to move these out to a trait that could be verified earlier than in the generated method. The first use case is for inferring the return type while constructing the op. At that point you don't have an Operation yet and that ends up in one having to duplicate the same checks, e.g., verify that attribute A is defined before querying A in shape function which requires that duplication. Instead this allows one to invoke a method to verify all the traits and, if this is checked first during verification, then all other traits could use attributes knowing they have been verified.
It is a little bit funny to have these on the adaptor, but I see the adaptor as a place to collect information about the op before the op is constructed (e.g., avoiding stringly typed accessors, verifying what is possible to verify before the op is constructed) while being cheap to use even with constructed op (so layer of indirection between the op constructed/being constructed). And from that point of view it made sense to me.
Differential Revision: https://reviews.llvm.org/D80842
Louis Dionne [Fri, 5 Jun 2020 15:12:47 +0000 (11:12 -0400)]
[CMake] Do not append -lm to CMAKE_REQUIRED_LIBRARIES on Apple
On Apple platforms, linking against libSystem.dylib is sufficient, and
some Apple platforms don't provide libm.dylib. On those platforms, adding
-lm to CMAKE_REQUIRED_LIBRARIES causes all subsequent compile-flag checks
to fail due to the missing library.
Differential Revision: https://reviews.llvm.org/D81265
Yaxun (Sam) Liu [Wed, 3 Jun 2020 14:05:52 +0000 (10:05 -0400)]
[HIP] Add default header and include path
To support std::complex and some other standard C/C++ functions in HIP device code,
they need to be forced to be __host__ __device__ functions by pragmas. This is done
by some clang standard C++ wrapper headers which are shared between cuda-clang and hip-Clang.
For these standard C++ wapper headers to work properly, specific include path order
has to be enforced:
clang C++ wrapper include path
standard C++ include path
clang include path
Also, these C++ wrapper headers require device version of some standard C/C++ functions
must be declared before including them. This needs to be done by including a default
header which declares or defines these device functions. The default header is always
included before any other headers are included by users.
This patch adds the the default header and include path for HIP.
Differential Revision: https://reviews.llvm.org/D81176
Simon Pilgrim [Fri, 5 Jun 2020 16:40:19 +0000 (17:40 +0100)]
ScalarEvolutionNormalization.h - reduce ScalarEvolutionExpressions.h include to forward declaration. NFC.
Simon Pilgrim [Fri, 5 Jun 2020 16:27:55 +0000 (17:27 +0100)]
TypeMetadataUtils.h - reduce Instructions.h include to forward declaration. NFC.
Move implicit include dependencies down to source file.
Sanjay Patel [Fri, 5 Jun 2020 16:20:13 +0000 (12:20 -0400)]
[x86] form reduction intrinsics from vectorizers instead of raw IR
Motivating examples are seen in the PhaseOrdering tests based on:
https://bugs.llvm.org/show_bug.cgi?id=43953#c2 - if we have
intrinsics there, some pass can fold them.
The intrinsics are still named "experimental" at this point, but
if there is no fallout from this patch, that will be a good
indicator that it is safe to finalize them.
Differential Revision: https://reviews.llvm.org/D80867
Sanjay Patel [Fri, 5 Jun 2020 14:11:21 +0000 (10:11 -0400)]
[PhaseOrdering] add tests for reductions; NFC (PR43953)
Nico Weber [Fri, 5 Jun 2020 16:32:23 +0000 (12:32 -0400)]
[gn build] (manually) port
a6fcf5ca033
Arthur Eubanks [Thu, 4 Jun 2020 16:48:42 +0000 (09:48 -0700)]
Add ASan metadata globals to @llvm.compiler.used under COFF
Summary:
This matches ELF.
This makes the number of ASan failures under the new pass manager on
Windows go from 18 to 1.
Under the old pass manager, the ASan module pass was one of the very
last things run, so these globals didn't get removed due to GlobalOpt.
But with the NPM the ASan module pass that adds these globals are run
much earlier in the pipeline and GlobalOpt ends up removing them.
Reviewers: vitalybuka, hans
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D81175
Arthur Eubanks [Thu, 4 Jun 2020 20:08:35 +0000 (13:08 -0700)]
Mark lambda in use-after-scope-capture.cpp as noinline
Summary:
As explained in https://bugs.llvm.org/show_bug.cgi?id=46208,
symbolization on Windows after inlining and around
lambdas/std::functions doesn't work very well. Under the new pass
manager, there is inlining at -O1.
use-after-scope-capture.cpp checks that the symbolization points to the
line containing "return x;", but the combination of
Windows/inlining/lambdas makes the symbolization point to the line
"f = [&x]() {".
Mark the lambda as noinline since this test is not a test for
symbolization.
Reviewers: hans, dblaikie, vitalybuka
Subscribers: #sanitizers
Tags: #sanitizers
Differential Revision: https://reviews.llvm.org/D81193
Simon Pilgrim [Fri, 5 Jun 2020 15:53:10 +0000 (16:53 +0100)]
[X86][SSE] Simplify MOVMSK patterns based on comparison
An initial patch adding combineSetCCMOVMSK to simplify MOVMSK and its vector input based on the comparison of the MOVMSK result.
This first stage just adds support for some simple MOVMSK(PACKSSBW()) cases where we remove the PACKSS if we're comparing ne/eq zero (any_of patterns), allowing us to directly compare against the v8i16 source vector(s) bitcasted to v16i8, with suitable masking to take into account of which signbits are valid.
Future combines could peek through further PACKSS, target shuffles, handle all_of patterns (ne/eq -1), optimize to a PTEST op, etc.
Differential Revision: https://reviews.llvm.org/D81171
Xun Li [Fri, 5 Jun 2020 15:51:54 +0000 (08:51 -0700)]
Fix the roundtrip test under llvm-profdata
Summary:
According to the comments, we want to convert the profile into two binary formats, and then into the md5text format.
We seems to have ignored the intermediate files.
This patch uses them to complete the full roundtrips.
Reviewers: wmi, wenlei
Reviewed By: wmi
Subscribers: llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D81202
Pete Steinfeld [Wed, 3 Jun 2020 15:03:14 +0000 (08:03 -0700)]
[flang] Fix crash on erroneous expressions
Summary:
If you create an expression with parse errors, the `parser::Expr.typedExpr`
will be empty, which causes a compiler crash. The crash is caused by the
check in check-do-forall.cpp that scans all expresssions to see if `DO`
variables are being modified.
It turned out that the problem was that I was fetching subexpressions of type
`parser::Expr`, which are not guaranteed to have a non-null `typedExpr`. I
fixed this by only grabbing the top-level expression from which to gather
arguments as part of the DO loop analysis. This, in turn, exposed a problem
where I wasn't collecting all of the actual arguments in an expression. This
was caused by the fact that I wasn't recursing through the rest of the
expression after finding an argument. I fixed this by recursing through the
argument in the member function in `CollectActualArgumentsHelper`.
Reviewers: klausler, tskeith, DavidTruby
Subscribers: llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D81101
Fangrui Song [Fri, 5 Jun 2020 15:41:03 +0000 (08:41 -0700)]
[ELF] Handle -u before input files
If both a.a and b.so define foo
```
ld.bfd -u foo a.a b.so # foo is defined
ld.bfd a.a b.so -u foo # foo is defined
ld.bfd -u foo b.so a.a # foo is undefined (provided at runtime by b.so)
ld.bfd b.so a.a -u foo # foo is undefined (provided at runtime by b.so)
```
In all cases we make foo undefined in the output. I tend to think the
GNU ld behavior makes more sense.
* In their model, they have to treat -u as a fake object file with an
undefined symbol before all input files, otherwise the first archive would not be fetched.
* Following their behavior allows us to drop a --warn-backrefs special case.
Reviewed By: psmith
Differential Revision: https://reviews.llvm.org/D81052
Zinovy Nis [Sun, 31 May 2020 18:53:13 +0000 (21:53 +0300)]
Fix crash on misc-redundant-expression
Differential Revision: https://reviews.llvm.org/D80896
Jonas Devlieghere [Fri, 5 Jun 2020 15:39:02 +0000 (08:39 -0700)]
[dsymutil] Don't emit N_AST symbol entries in the Mach-O companion file
Treat N_AST symbol table entries like other debug entries and don't emit
them in the linked binary.
Differential revision: https://reviews.llvm.org/D81205
Nicolas Vasilache [Fri, 5 Jun 2020 15:23:43 +0000 (11:23 -0400)]
[mlir][SCF] Add single iteration scf.for promotion to the FuncOp level helper.
Previously only the Affine version would be folded.
Differential Revision: https://reviews.llvm.org/D81261
Julian Lettner [Wed, 8 Apr 2020 05:48:39 +0000 (22:48 -0700)]
[lit] Improve naming of test result categories
Improve consistency when printing test results:
Previously we were using different labels for group names (the header
for the list of, e.g., failing tests) and summary count lines. For
example, "Failing Tests"/"Unexpected Failures". This commit changes lit
to label things consistently.
Improve wording of labels:
When talking about individual test results, the first word in
"Unexpected Failures", "Expected Passes", and "Individual Timeouts" is
superfluous. Some labels contain the word "Tests" and some don't.
Let's simplify the names.
Before:
```
Failing Tests (1):
...
Expected Passes : 3
Unexpected Failures: 1
```
After:
```
Failed Tests (1):
...
Passed: 3
Failed: 1
```
Reviewed By: ldionne
Differential Revision: https://reviews.llvm.org/D77708
Sander de Smalen [Fri, 5 Jun 2020 15:03:13 +0000 (16:03 +0100)]
Revert "[CodeGen][SVE] Calculate correct type legalization for scalable vectors."
Seems to break some buildbots, reverting the patch for now.
This reverts commit
164f4b9d26fdf3cd640a09b63b5ec44d033cbe8a.
Haojian Wu [Tue, 2 Jun 2020 14:35:29 +0000 (16:35 +0200)]
[AST] Record SourceLocation for TypoExpr.
Reviewers: sammccall
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D81008
Kazushi (Jam) Marukawa [Fri, 5 Jun 2020 14:57:37 +0000 (16:57 +0200)]
[VE] Support logical operation instructions in MC layer
Summary:
Add regression tests of asmparser, mccodeemitter, and disassembler for
logical operation instructions. Also change asmparser to support CMOV
instruction. And, add new EQV/MRG/NND isntructions also.
Differential Revision: https://reviews.llvm.org/D81219
Wen-Heng (Jack) Chung [Wed, 20 May 2020 21:07:49 +0000 (16:07 -0500)]
[mlir][gpu] Introduce mlir-rocm-runner.
Summary:
`mlir-rocm-runner` is introduced in this commit to execute GPU modules on ROCm
platform. A small wrapper to encapsulate ROCm's HIP runtime API is also inside
the commit.
Due to behavior of ROCm, raw pointers inside memrefs passed to `gpu.launch`
must be modified on the host side to properly capture the pointer values
addressable on the GPU.
LLVM MC is used to assemble AMD GCN ISA coming out from
`ConvertGPUKernelToBlobPass` to binary form, and LLD is used to produce a shared
ELF object which could be loaded by ROCm HIP runtime.
gfx900 is the default target be used right now, although it could be altered via
an option in `mlir-rocm-runner`. Future revisions may consider using ROCm Agent
Enumerator to detect the right target on the system.
Notice AMDGPU Code Object V2 is used in this revision. Future enhancements may
upgrade to AMDGPU Code Object V3.
Bitcode libraries in ROCm-Device-Libs, which implements math routines exposed in
`rocdl` dialect are not yet linked, and is left as a TODO in the logic.
Reviewers: herhut
Subscribers: mgorny, tpr, dexonsmith, mehdi_amini, rriddle, jpienaar, shauheen, antiagainst, nicolasvasilache, csigg, arpith-jacob, mgester, lucyrfox, aartbik, liufengdb, stephenneuendorffer, Joonsoo, grosul1, frgossen, Kayjukh, jurahul, llvm-commits
Tags: #mlir, #llvm
Differential Revision: https://reviews.llvm.org/D80676
Sander de Smalen [Fri, 5 Jun 2020 11:54:39 +0000 (12:54 +0100)]
[CodeGen][SVE] Calculate correct type legalization for scalable vectors.
This patch updates TargetLoweringBase::computeRegisterProperties and
TargetLoweringBase::getTypeConversion to support scalable vectors,
and make the right calls on how to legalise them. These changes are required
to legalise both MVTs and EVTs.
Reviewers: efriedma, david-arm, ctetreau
Reviewed By: efriedma
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D80640
serge-sans-paille [Thu, 4 Jun 2020 21:33:03 +0000 (23:33 +0200)]
[SCCP] Report changes after removing stores to constant global
Differential Revision: https://reviews.llvm.org/D81228