Nico Weber [Tue, 20 Jul 2021 10:43:30 +0000 (06:43 -0400)]
[gn build] (manually) port
bc1a2979fc70
Chen Zheng [Tue, 20 Jul 2021 10:27:47 +0000 (10:27 +0000)]
[PowerPC][NFC] add more cases for lfiwzx/lfiwax
Jan Svoboda [Tue, 20 Jul 2021 10:12:40 +0000 (12:12 +0200)]
[clang][deps] Avoid minimizing PCH input files
This patch avoid minimizing input files that contributed to a PCH or its modules. This prevents the implicit modular build to fail on unexpected file size. Depends on D106146.
Reviewed By: dexonsmith
Differential Revision: https://reviews.llvm.org/D104536
Jan Svoboda [Tue, 20 Jul 2021 09:42:34 +0000 (11:42 +0200)]
[clang][deps] Separate filesystem caches for minimized and original files
This patch separates the local and global caches of `DependencyScanningFilesystem` into two buckets: minimized files and original files. This is necessary to deal with precompiled modules/headers.
Consider a single worker with its instance of filesystem:
1. Build system uses the worker to scan dependencies of module A => filesystem cache gets populated with minimized input files.
2. Build system uses the results to explicitly build module A => explicitly built module captures the state of the real filesystem (containing non-minimized input files).
3. Build system uses the prebuilt module A as an explicit precompiled dependency for another compile job B.
4. Build system uses the same worker to scan dependencies for job B => worker uses implicit modular build to discover dependencies, which validates the filesystem state embedded in the prebuilt module (non-minimized files) to the current view of the filesystem (minimized files), resulting in validation failures.
This problem can be avoided in step 4 by collecting input files from the precompiled module and marking them as "ignored" in the minimizing filesystem. This way, the validation should succeed, since we should be always dealing with the original (non-minized) input files. However, the filesystem already minimized the input files in step 1 and put it in the cache, which gets used in step 4 as well even though it's marked ignored (do not minimize). This patch essentially fixes this oversight by making the `"file is minimized"` part of the cache key (from high level).
Depends on D106064.
Reviewed By: dexonsmith
Differential Revision: https://reviews.llvm.org/D106146
Stanislav Mekhanoshin [Fri, 16 Jul 2021 16:15:10 +0000 (09:15 -0700)]
[AMDGPU] Disable LDS lowering for GFX shaders
Apparently these need external LDS symbols to remain.
Fixes: SC1-3279
Differential Revision: https://reviews.llvm.org/D106288
Dawid Jurczak [Mon, 5 Jul 2021 09:42:17 +0000 (11:42 +0200)]
[DSE] Transform memset + malloc --> calloc (PR25892)
After this change DSE can eliminate malloc + memset and emit calloc.
It's https://reviews.llvm.org/D101440 follow-up.
Differential Revision: https://reviews.llvm.org/D103009
Florian Mayer [Tue, 20 Jul 2021 09:36:26 +0000 (10:36 +0100)]
Revert "[hwasan] Use stack safety analysis."
This reverts commit
e9c63ed10b3bdf6eb3fa76d1a3eb403d6fc6a118.
Jan Svoboda [Tue, 20 Jul 2021 08:57:12 +0000 (10:57 +0200)]
[clang][deps] Normalize ignored filenames in minimizing file system
This patch normalizes filenames in `DependencyScanningWorkerFilesystem` so that lookup of ignored files works correctly on Windows (where `/` and `\` are equivalent).
Reviewed By: dexonsmith
Differential Revision: https://reviews.llvm.org/D106064
Florian Mayer [Mon, 19 Jul 2021 10:16:38 +0000 (11:16 +0100)]
[NFC] run clang-format on hwasan use-after-scope tests.
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D106259
Kai Luo [Tue, 20 Jul 2021 09:11:01 +0000 (09:11 +0000)]
[PowerPC] Add lit.local.cfg in AtomicExpand tests
Fixed build errors on other platforms.
Florian Mayer [Fri, 16 Jul 2021 15:30:37 +0000 (16:30 +0100)]
[hwasan] [NFC] copy and disable ASAN tests to hwasan.
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D106159
Florian Mayer [Mon, 19 Jul 2021 11:52:17 +0000 (12:52 +0100)]
[hwasan] Use stack safety analysis.
This avoids unnecessary instrumentation.
Reviewed By: eugenis, vitalybuka
Differential Revision: https://reviews.llvm.org/D105703
Sander de Smalen [Mon, 19 Jul 2021 09:48:42 +0000 (10:48 +0100)]
[AArch64][SVE][InstCombine] last{a,b} of a splat vector
Replace last{a,b}(splat(X)) with X, irrespective of the predicate.
Patch by/Committing on behalf of: Usman Nadeem (mnadeem)
Reviewed By: sdesmalen
Differential Revision: https://reviews.llvm.org/D105520
Benjamin Kramer [Tue, 20 Jul 2021 08:34:05 +0000 (10:34 +0200)]
[lldb] Remove unused variable. NFCI
Cullen Rhodes [Tue, 20 Jul 2021 07:19:10 +0000 (07:19 +0000)]
[AArch64][SME] Add system registers and related instructions
This patch adds the new system registers introduced in SME:
- ID_AA64SMFR0_EL1 (ro) SME feature identifier.
- SMCR_ELx (r/w) streaming mode control register for configuring
effective SVE Streaming SVE Vector length when the PE is in
Streaming SVE mode.
- SVCR (r/w) streaming vector control register, visible at all
exception levels. Provides access to PSTATE.SM and PSTATE.ZA
using MSR and MRS instructions.
- SMPRI_EL1 (r/w) streaming mode execution priority register.
- SMPRIMAP_EL2 (r/w) streaming mode priority mapping register.
- SMIDR_EL1 (ro) streaming mode identification register.
- TPIDR2_EL0 (r/w) for use by SME software to manage per-thread
SME context.
- MPAMSM_EL1 (r/w) MPAM (v8.4) streaming mode register, for
labelling memory accesses performed in streaming mode.
Also added in this patch are the SME mode change instructions.
Three MSR immediate instructions are implemented to set or clear
PSTATE.SM, PSTATE.ZA, or both respectively:
- MSR SVCRSM, #<imm1>
- MSR SVCRZA, #<imm1>
- MSR SVCRSMZA, #<imm1>
The following smstart/smstop aliases are also implemented for
convenience:
smstart -> MSR SVCRSMZA, #1
smstart sm -> MSR SVCRSM, #1
smstart za -> MSR SVCRZA, #1
smstop -> MSR SVCRSMZA, #0
smstop sm -> MSR SVCRSM, #0
smstop za -> MSR SVCRZA, #0
The reference can be found here:
https://developer.arm.com/documentation/ddi0602/2021-06
Reviewed By: david-arm
Differential Revision: https://reviews.llvm.org/D105576
Chia-hung Duan [Tue, 20 Jul 2021 07:26:23 +0000 (15:26 +0800)]
[mlir-tblgen] Fix failed matching when binds same operand of an op in different depth
For example, we will generate incorrect code for the pattern,
def : Pat<((FooOp (FooOp, $a, $b), $b)), (...)>;
We didn't allow $b to be bond twice with same operand of same op.
Reviewed By: jpienaar
Differential Revision: https://reviews.llvm.org/D105677
Jason Molenda [Tue, 20 Jul 2021 07:31:19 +0000 (00:31 -0700)]
Remove the DarwinLog functionality from debguserver
Remove the DarwinLog and qStructuredDataPlugins support
from debugserver. The DarwinLog plugin was never debugged
fully and made reliable, and the underlying private APIs
it uses have migrated since 2016 so none of them exist
any longer.
Differential Revision: https://reviews.llvm.org/D106324
rdar://
75073283
Amara Emerson [Tue, 20 Jul 2021 07:03:37 +0000 (00:03 -0700)]
[AArch64][GlobalISel] Don't form truncstores in postlegalizer-lowering for s128.
We don't support truncating s128 stores, so don't form them.
Johannes Doerfert [Wed, 14 Jul 2021 03:54:52 +0000 (22:54 -0500)]
[Attributor] Use set vector instead of vector to prevent duplicates
Johannes Doerfert [Tue, 13 Jul 2021 03:14:36 +0000 (22:14 -0500)]
[Attributor] Simplify to values in the genericValueTraversal
We already simplified to a constant, given the new interface we can also
simplify to a generic value.
Johannes Doerfert [Wed, 14 Jul 2021 03:50:57 +0000 (22:50 -0500)]
[Attributor] Use checkForAllUses instead of custom use tracking
AAMemoryBehaviorFloating used a custom use tracking mechanism even
though checkForAllUses exists and is already more powerful. Further,
AAMemoryBehaviorFloating uses AANoCapture to guarantee that there are no
aliases and following the uses is sufficient. This is an OK assumption
if checkForAllUses is used but custom tracking is easily out of sync
with AANoCapture and problems follow.
Dmitry Vyukov [Sun, 18 Jul 2021 08:19:37 +0000 (10:19 +0200)]
sanitizer_common: add new mutex
We currently have 3 different mutexes:
- RWMutex
- BlockingMutex
- __tsan::Mutex
RWMutex and __tsan::Mutex are roughly the same,
except that tsan version supports deadlock detection.
BlockingMutex degrades better under heavy contention
from lots of threads (blocks in OS), but much slower
for light contention and has non-portable performance
and has larger static size and is not reader-writer.
Add a new mutex that combines all advantages of these
mutexes: it's reader-writer, has fast non-contended path,
supports blocking to gracefully degrade under higher contention,
has portable size/performance.
For now it's named Mutex2 for incremental submission. The plan is to:
- land this change
- then move deadlock detection logic from tsan
- then rename it to Mutex and remove tsan Mutex
- then typedef RWMutex/BlockingMutex to this mutex
SpinMutex stays as separate type because it has faster fast path:
1 atomic RMW per lock/unlock as compared to 2 for this mutex.
Reviewed By: vitalybuka, melver
Differential Revision: https://reviews.llvm.org/D106231
Kai Luo [Tue, 20 Jul 2021 06:14:08 +0000 (06:14 +0000)]
[PowerPC] Fallback to base's implementation of shouldExpandAtomicCmpXchgInIR and shouldExpandAtomicCmpXchgInIR
If we can't decide `shouldExpandAtomicCmpXchgInIR` or `shouldExpandAtomicCmpXchgInIR` in PPC's implementation after https://reviews.llvm.org/rGb9c3941cd61de1e1b9e4f3311ddfa92394475f4b, resort to base's implementation.
This fixes internal build of OpenMP which uses atomic operations on float.
Reviewed By: jsji
Differential Revision: https://reviews.llvm.org/D106234
Petr Hosek [Tue, 20 Jul 2021 05:45:33 +0000 (22:45 -0700)]
[NFC][profile] Move writeMMappedFile to ELF ifdef block
This avoids the compiler warning on Darwin where that function is unused.
Craig Topper [Tue, 20 Jul 2021 05:52:05 +0000 (22:52 -0700)]
[RISCV] Add test cases to show an issue with our fcvt.wu isel patterns on RV64.
The pattern we match is (sext_inreg (assertzexti32 (fp_to_uint)), i32). If
the assertzexti32 has an additional user we'll end up emitting
an fcvt.wu and an fcvt.lu.
This can happen if the original fp_to_uint before type legalization
has one user that causes a sext_inreg to be emitted and one that
doesn't.
LLVM GN Syncbot [Tue, 20 Jul 2021 05:51:14 +0000 (05:51 +0000)]
[gn build] Port
adb55d7c3265
Michał Górny [Mon, 19 Jul 2021 11:17:52 +0000 (13:17 +0200)]
[lldb] Make WatchpointList iterable
Based on
de448c0a9e5088979526e2e67152fe547ae4ccf0.
Differential Revision: https://reviews.llvm.org/D106263
Dmitry Vyukov [Mon, 19 Jul 2021 14:19:07 +0000 (16:19 +0200)]
tsan: remove the stats subsystem
I don't think the stat subsystem was ever used since tsan
development in 2012. But it adds lots of code and this
effectively dead code needs to be updated if the runtime
code changes, which adds maintanance cost for no benefit.
Normal profiler usually gives enough info and that info
is more trustworthy.
Remove the stats subsystem.
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D106276
Johannes Doerfert [Tue, 20 Jul 2021 05:30:20 +0000 (00:30 -0500)]
[Attributor] Gracefully handle interprocedural reachability queries
This does ensure `InformationCache::getPotentiallyReachable` will not
crash/assert on instructions from different functions but simply return
that one is reachable, which is conservatively correct.
Johannes Doerfert [Thu, 15 Jul 2021 21:09:33 +0000 (16:09 -0500)]
[Attributor] Ensure to simplify operands in AAValueConstantRange
As with other patches before, the simplification callback interface
requires us to go through the Attributor::getAssumedSimplified API first
before we recurs.
It is unclear if the problem can be explicitly tested with our current
infrastructure.
Johannes Doerfert [Thu, 15 Jul 2021 20:41:46 +0000 (15:41 -0500)]
[Attributor] Extend the AAValueSimplify compare simplification logic
We first simplify the operands of a compare and then reason on the
simplified versions, e.g., with AANonNull.
This does improve the simplification capabilities but also fixes a
potential problem that has not yet been observed by simplifying the
operands first.
Johannes Doerfert [Tue, 13 Jul 2021 03:22:31 +0000 (22:22 -0500)]
[Attributor][NFCI] Expose `getAssumedUnderlyingObjects` API
Johannes Doerfert [Tue, 20 Jul 2021 05:26:52 +0000 (00:26 -0500)]
[OpenMP] Remove XFAIL and update check lines properly
Undo
15c5701c8324d2dea519fa379c04d5c619a570ab and update check lines.
Johannes Doerfert [Tue, 20 Jul 2021 04:50:20 +0000 (23:50 -0500)]
[Attributor][NFC] Fix function name spelling
Dmitry Vyukov [Mon, 19 Jul 2021 13:54:44 +0000 (15:54 +0200)]
tsan: add pragma line to buildgo.sh
Add pragma line so that errors messages point to the actual
source files rather than to the concatenated gotsan.cpp.
Reviewed By: vitalybuka, melver
Differential Revision: https://reviews.llvm.org/D106275
Johannes Doerfert [Tue, 20 Jul 2021 04:14:35 +0000 (23:14 -0500)]
[OpenMP][FIX] Temporarily XFAIL tests waiting for new check lines
The test is not wrong nor is the current main broken, it just an
interplay issue. Check lines will be updated in shortly.
Johannes Doerfert [Wed, 14 Jul 2021 21:06:57 +0000 (16:06 -0500)]
[Attributor][FIX] Do not simplify byval arguments
A byval argument is a different value in the caller and callee, we
cannot propagate the information as part of AAValueSimplify. Users that
want to deal with byval arguments need to specifically perform the
argument -> call site step. We do not do this for now.
Johannes Doerfert [Mon, 12 Jul 2021 00:27:26 +0000 (19:27 -0500)]
[Attributor] Introduce AAPointerInfo
This patch introduces AAPointerInfo which tracks the uses of a pointer
and places them in "bins" based on their offset from the base and access
size.
As with other AAs, any pointer can be tracked but it is up to the user
to make sense of the results. The user in this patch is AAValueSimplify
and AAPotentialValues which both utilize AAPointerInfo to determine the
value of a load. For now, this is restricted to loads of allocas and
internal globals. Through the use of AAPointerInfo and the "bins" we can
track struct members separately. The users also know that storing only
zeros (at unknown indices) will result in loading only 0 (from unknown
indices). Other than that, the users are flow and context insensitive
(for now).
To deal with the "bins" more easily, AAPointerInfo provides a
forallInterfearingAccesses that applies a callback on all accesses
that might interfere with a given load or store.
Differential Revision: https://reviews.llvm.org/D104432
Johannes Doerfert [Mon, 10 May 2021 06:11:06 +0000 (01:11 -0500)]
[Attributor] Simplify loads
As a first step to simplify loads we only handle `null` and `undef`
underlying objects, as well as objects that have the load as a single user.
Loads of those values can be replaced by the initializer, if any.
Proper reasoning is introduced in a follow up patch
Differential Revision: https://reviews.llvm.org/D103862
Johannes Doerfert [Thu, 15 Jul 2021 18:12:00 +0000 (13:12 -0500)]
[OpenMP] Fix carefully track SPMDCompatibilityTracker
We did not properly use SPMDCompatibilityTracker in various places.
This patch makes sure we look at the validity properly and also fix
the state if we can.
Differential Revision: https://reviews.llvm.org/D106085
Albion Fung [Tue, 20 Jul 2021 02:27:02 +0000 (21:27 -0500)]
Revert "[PowerPC] Extra test case for LDARX"
This reverts commit
1d3e77e7a8421a9d2dd13e3ef499ea967ea8f85c as
some buildbots seem to be unable to obtain the target
powerpc64le-unknown-linux-gnu.
Tony Tye [Tue, 20 Jul 2021 01:20:22 +0000 (01:20 +0000)]
[NFC] Correct documentation error in OpenMP release ReleaseNotes
Reviewed By: tianshilei1992
Differential Revision: https://reviews.llvm.org/D106330
Lang Hames [Mon, 19 Jul 2021 23:06:39 +0000 (09:06 +1000)]
[ORC-RT] Introduce a ORC_RT_JIT_DISPATCH_TAG macro.
This macro can be used to define tag variables for use with jit-dispatch.
Lang Hames [Mon, 19 Jul 2021 22:52:21 +0000 (08:52 +1000)]
[ORC-RT] Add ORC_RT prefix to WEAK_IMPORT macro.
Hsiangkai Wang [Mon, 19 Jul 2021 08:57:41 +0000 (16:57 +0800)]
[Clang][RISCV] Correct the alignment of stores generated by vlseg/vlsegff.
Differential Revision: https://reviews.llvm.org/D106255
Fangrui Song [Tue, 20 Jul 2021 01:13:52 +0000 (18:13 -0700)]
Revert D106128 "[lld-macho] Use DO_BIND_ADD_ADDR_IMM_SCALED for bind opcodes"
This reverts commit
321b2bef098553ec648e4174aae92c63a6e1a810.
`for (BindIR *p = &opcodes[0]; p->opcode != BIND_OPCODE_DONE; ++p) {` has a heap-buffer-overflow with test/MachO/bind-opcodes.
Shilei Tian [Tue, 20 Jul 2021 01:08:40 +0000 (21:08 -0400)]
[OpenMP] Fixed a segmentation fault when using taskloop and target nowait
The synchronization of task loop misses hidden helper tasks, causing segmentation
fault reported in https://bugs.llvm.org/show_bug.cgi?id=50002.
Reviewed By: ye-luo
Differential Revision: https://reviews.llvm.org/D106220
Matt Arsenault [Wed, 12 May 2021 21:10:24 +0000 (17:10 -0400)]
CodeGen: Make MachineOptimizationRemarkEmitterPass a CFG analysis
This avoids rerunning it a few times.
Albion Fung [Tue, 20 Jul 2021 01:03:22 +0000 (20:03 -0500)]
[PowerPC] Extra test case for LDARX
An extra test case added for the builtin __LDARX.
Differential revision: https://reviews.llvm.org/D105926
Matt Arsenault [Fri, 16 Jul 2021 23:59:22 +0000 (19:59 -0400)]
GlobalISel: Remove some mystery code that clears isReturned
I don't understand what this is going for, and haven't found an analog
in DAG code. No tests fail with this removed.
Matt Arsenault [Fri, 16 Jul 2021 13:54:52 +0000 (09:54 -0400)]
AArch64/GlobalISel: Preserve memory types
Fangrui Song [Tue, 20 Jul 2021 00:09:01 +0000 (17:09 -0700)]
Revert D105519 "[WebAssembly] Deduplicate imports of the same module name, field name, and type" and its followup
This reverts commit
4ae575b9997e0903d1c2ec01a43e3f3f2db5df16 and
9b965b37c75d626c01951184088314590e38d299.
There is an use-of-uninitialized-value bug in the `else` branch in ImportSection::addImport.
Joseph Huber [Mon, 19 Jul 2021 20:54:31 +0000 (16:54 -0400)]
[Libomptarget] Remove volatile from NVPTX work function
Currently the NPVTX work function is marked volatile. This prevents some
optimizations from using this value.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106310
Nico Weber [Mon, 19 Jul 2021 23:57:02 +0000 (19:57 -0400)]
[gn build] Fix llvm_build_instrumented_coverage=true builds with goma/rbe
Louis Dionne [Fri, 16 Jul 2021 19:52:42 +0000 (15:52 -0400)]
[libc++] Tidy-up instances of __STDCPP_DEFAULT_NEW_ALIGNMENT__ in the tests
See https://reviews.llvm.org/D105905 for context.
Differential Revision: https://reviews.llvm.org/D106182
Louis Dionne [Mon, 19 Jul 2021 16:34:56 +0000 (12:34 -0400)]
[libc++] Add a bunch of missing _LIBCPP_HIDE_FROM_ABI in <ranges>
We've been forgetting to add those to most of the <ranges> review.
To avoid forgetting in the future, I added an item in the pre-commit
checklist.
Differential Revision: https://reviews.llvm.org/D106287
Philip Reames [Mon, 19 Jul 2021 23:33:01 +0000 (16:33 -0700)]
[tests] Add a couple of tests for zero stride trip counts w/loop varying exit values
Vincent Lee [Fri, 16 Jul 2021 03:25:48 +0000 (20:25 -0700)]
[lld-macho] Use DO_BIND_ADD_ADDR_IMM_SCALED for bind opcodes
Implement pass 3 of bind opcodes from ld64 (which supports both 32-bit and 64-bit).
Pass 3 implementation condenses BIND_OPCODE_DO_BIND_ADD_ADDR_ULEB opcode
to BIND_OPCODE_DO_BIND_ADD_ADDR_IMM_SCALED. This change is already behind an
O2 flag so it shouldn't impact current performance. I verified ld64's output with x86_64 LLD
and they were both emitting the same optimized bind opcodes (although in a slightly different
order). Tested with arm64_32 LLD and compared that with x86 LLD that the order of the bind
opcodes are the same (offset values are different which should be expected).
Reviewed By: int3, #lld-macho
Differential Revision: https://reviews.llvm.org/D106128
Eli Friedman [Fri, 16 Jul 2021 23:21:20 +0000 (16:21 -0700)]
[ScalarEvolution] Refine computeMaxBECountForLT to be accurate in more cases.
Allow arbitrary strides, and make sure we return the correct result when
the backedge-taken count is zero.
Differential Revision: https://reviews.llvm.org/D106197
Yi Zhang [Mon, 19 Jul 2021 22:10:13 +0000 (15:10 -0700)]
Dyanamic shape support for memref reassociation reshape ops
Only memref with identity layout map is supported for now.
Reviewed By: mravishankar
Differential Revision: https://reviews.llvm.org/D106180
Philip Reames [Mon, 19 Jul 2021 22:11:38 +0000 (15:11 -0700)]
[SCEV] Add a clarifying comment in howManyLessThans
Wrap semantics are subtle when combined with multiple exits. This has caused several rounds of confusion during recent reviews, so try to document the subtly distinction between when wrap flags provide <u and <=u facts.
Arthur Eubanks [Mon, 19 Jul 2021 20:35:57 +0000 (13:35 -0700)]
[NewPM][opt] Add -debug-pass-manager=quiet to not print analysis info
Reviewed By: asbirlea
Differential Revision: https://reviews.llvm.org/D106307
Arthur Eubanks [Mon, 19 Jul 2021 20:20:57 +0000 (13:20 -0700)]
[NewPM] Bail out of devirtualization wrapper if the current SCC is invalidated
The specific case that triggered this was when inlining a recursive
internal function into itself caused the recursion to go away, allowing
the inliner to mark the function as dead. The inliner marks the SCC as
invalidated but does not provide a new SCC to continue with.
This matches the implementations of ModuleToPostOrderCGSCCPassAdaptor
and CGSCCPassManager.
Fixes PR50363.
Reviewed By: asbirlea
Differential Revision: https://reviews.llvm.org/D106306
Quinn Pham [Fri, 16 Jul 2021 16:48:00 +0000 (11:48 -0500)]
[PowerPC] swdiv_nochk Builtins for XL Compat
This patch is in a series of patches to provide builtins for
compatibility with the XL compiler. This patch adds software divide
builtins with no checking. These builtins are each emitted as a fast
fdiv.
Reviewed By: #powerpc, nemanjai
Differential Revision: https://reviews.llvm.org/D106150
MaheshRavishankar [Mon, 19 Jul 2021 21:35:20 +0000 (14:35 -0700)]
Split `InferShapedTypeOpInterface` to create `ReifyRankedShapedTypeInterface`.
The `reifyReturnTypeShapesPerResultDim` method supports shape
inference for rsults that are ranked types. These are used lower in
the codegeneration stack than its counter part `reifyReturnTypeShapes`
which also supports unranked types, and is more suited for use higher
up the compilation stack. To have separation of concerns, this method
is split into its own interface.
See discussion : https://llvm.discourse.group/t/better-layering-for-infershapedtypeopinterface/3823
Differential Revision: https://reviews.llvm.org/D106133
LLVM GN Syncbot [Mon, 19 Jul 2021 21:33:24 +0000 (21:33 +0000)]
[gn build] Port
08b289867b5a
Nathan Ridge [Mon, 21 Jun 2021 07:19:19 +0000 (03:19 -0400)]
[clang] Respect PrintingPolicy::FullyQualifiedName when printing a template-id
Fixes PR50774
Differential Revision: https://reviews.llvm.org/D104619
Sam Clegg [Thu, 15 Jul 2021 00:16:15 +0000 (17:16 -0700)]
[lld][WebAssembly] Cleanup duplicate fields in Symbols.h. NFC
This avoids duplication and simplifies the code in several places
without increasing the size of the symbol union (at least not
above the assert'd limit of 120 bytes).
Differential Revision: https://reviews.llvm.org/D106026
Jim Ingham [Mon, 19 Jul 2021 21:25:37 +0000 (14:25 -0700)]
Don't use !eStateRunning when you mean eStateStopped in DestroyImpl.
When we go to destroy the process, we first try to halt it, if
we succeeded and the target stopped, we want to clear out the
thread plans and breakpoints in case we still need to resume to complete
killing the process. If the target was exited or detached, it's
pointless but harmless to do this. But if the state is eStateInvalid -
for instance if we tried to interrupt the target to Halt it and that
fails - we don't want to keep trying to interact with the inferior,
so we shouldn't do this work.
This change explicitly checks eStateStopped, and only does the pre-resume
cleanup if we did manage to stop the process.
Artem Belevich [Mon, 19 Jul 2021 21:27:41 +0000 (14:27 -0700)]
Revert "[MemCpyOpt] Enable memcpy optimizations unconditionally."
This reverts commit
2c98298a7559dfe4a264ef1adaad0921526768cc which breaks
sanitizers.
thomasraoux [Mon, 19 Jul 2021 21:00:51 +0000 (14:00 -0700)]
[mlir] Fix bazel build
Differential Revision: https://reviews.llvm.org/D106311
Derek Schuff [Wed, 2 Jun 2021 21:37:22 +0000 (14:37 -0700)]
[WebAssembly] Generate R_WASM_FUNCTION_OFFSET relocs in debuginfo sections
Debug info sections need R_WASM_FUNCTION_OFFSET_I32 relocs (with FK_Data_4 fixup
kinds) to refer to functions (instead of R_WASM_TABLE_INDEX as is used in data
sections). Usually this is done in a convoluted way, with unnamed temp data
symbols which target the start of the function, in which case
WasmObjectWriter::recordRelocation converts it to use the section symbol
instead. However in some cases the function can actually be undefined; in this
case the dwarf generator uses the function symbol (a named undefined function
symbol) instead. In that case the section-symbol transform doesn't work and we
need to generate the correct reloc type a different way. In this change
WebAssemblyWasmObjectWriter::getRelocType takes the fixup section type into
account to choose the correct reloc type.
Fixes PR50408
Differential Revision: https://reviews.llvm.org/D103557
Mircea Trofin [Wed, 14 Jul 2021 22:03:14 +0000 (15:03 -0700)]
[MLGO] Use binary protobufs for improved training performance.
It turns out that during training, the time required to parse the
textual protobuf of a training log is about the same as the time it
takes to compile the module generating that log. Using binary protobufs
instead elides that cost almost completely.
Differential Revision: https://reviews.llvm.org/D106157
Nick Fitzgerald [Mon, 19 Jul 2021 20:02:46 +0000 (13:02 -0700)]
[WebAssembly] Deduplicate imports of the same module name, field name, and type
When two symbols import the same thing, only one import should be emitted in the Wasm file.
Fixes https://bugs.llvm.org/show_bug.cgi?id=50938
Reviewed By: sbc100
Differential Revision: https://reviews.llvm.org/D105519
Leonard Grey [Mon, 19 Jul 2021 20:44:15 +0000 (16:44 -0400)]
[lld/mac] Add test for --lto-O
This belongs to
fe08e9c4871, I (thakis) forgot to `git add` it back then.
Differential Revision: https://reviews.llvm.org/D105223
thomasraoux [Tue, 13 Jul 2021 03:49:21 +0000 (20:49 -0700)]
[mlir] Add software pipelining transformation for scf.For op
This is the first step to support software pipeline for scf.for loops.
This is only the transformation to create pipelined kernel and
prologue/epilogue.
The scheduling needs to be given by user as many different algorithm
and heuristic could be applied.
This currently doesn't handle loop arguments, this will be added in a
follow up patch.
Differential Revision: https://reviews.llvm.org/D105868
Nico Weber [Mon, 19 Jul 2021 18:38:15 +0000 (14:38 -0400)]
[lld/mac] Resolve defined symbols before undefined symbols
Ports https://reviews.llvm.org/D95985 to the MachO port.
Happens to fix PR51135; see that bug for details.
Also makes lld's behavior match ld64 for the included test case.
Differential Revision: https://reviews.llvm.org/D106293
owenca [Thu, 15 Jul 2021 23:16:49 +0000 (16:16 -0700)]
[clang-format] Break an unwrapped line at a K&R C parameter decl
Break an unwrapped line before the first parameter declaration in a
K&R C function definition.
This fixes PR51074.
Differential Revision: https://reviews.llvm.org/D106112
Stuart Brady [Mon, 7 Jun 2021 15:30:22 +0000 (16:30 +0100)]
[demangler] Fix demangling of 'half'
Demangle 'Dh' as 'half' (as per GCC), and not 'decimal16' (which doesn't
make sense, as there is no IEEE 754 decimal16 format).
The Itanium C++ ABI specification describes 'Dh' as:
> IEEE 754r half-precision floating point (16 bits)
(https://itanium-cxx-abi.github.io/cxx-abi/abi.html#mangling-builtin)
Reviewed By: ldionne, jyknight
Differential Revision: https://reviews.llvm.org/D103833
Teresa Johnson [Mon, 19 Jul 2021 19:35:45 +0000 (12:35 -0700)]
[LangRef] Clarify support for multiple metadata attachments with same id
As discussed on D105251, currently the compiler does not support
multiple metadata attachments on instructions having the same
identifier, whereas it does for global objects. Note this in the
Language Reference manual for clarity.
See D105251 for discussions of history behind this divergence, and the
complexities and possible approaches of adding this support to
instructions in the future.
Differential Revision: https://reviews.llvm.org/D106304
Tony Tye [Mon, 19 Jul 2021 02:21:11 +0000 (02:21 +0000)]
[AMDGPU] Reserve AMDGPU ELF e_flags machine 0x45
Reviewed By: rampitec
Differential Revision: https://reviews.llvm.org/D106249
MaheshRavishankar [Mon, 19 Jul 2021 19:50:30 +0000 (12:50 -0700)]
[mlir][Linalg] NFC: Rename FusionOfTensors pass to FusionOfElementwiseOps pass.
This makes it more explicit what the scope of this pass is. The name
of this pass predates fusion on tensors using tile + fuse, and hence
the confusion.
Differential Revision: https://reviews.llvm.org/D106132
LLVM GN Syncbot [Mon, 19 Jul 2021 19:24:16 +0000 (19:24 +0000)]
[gn build] Port
54902e00d128
Petr Hosek [Thu, 8 Jul 2021 20:44:05 +0000 (13:44 -0700)]
[InstrProfiling] Use weak alias for bias variable
We need the compiler generated variable to override the weak symbol of
the same name inside the profile runtime, but using LinkOnceODRLinkage
results in weak symbol being emitted in which case the symbol selected
by the linker is going to depend on the order of inputs which can be
fragile.
This change replaces the use of weak definition inside the runtime with
a weak alias. We place the compiler generated symbol inside a COMDAT
group so dead definition can be garbage collected by the linker.
We also disable the use of runtime counter relocation on Darwin since
Mach-O doesn't support weak external references, but Darwin already uses
a different continous mode that relies on overmapping so runtime counter
relocation isn't needed there.
Differential Revision: https://reviews.llvm.org/D105176
Artem Belevich [Thu, 15 Jul 2021 20:39:47 +0000 (13:39 -0700)]
[infer-address-spaces] Handle complex non-pointer constexpr arguments.
Fixes https://bugs.llvm.org/show_bug.cgi?id=51099
Differential Revision: https://reviews.llvm.org/D106098
Simon Pilgrim [Mon, 19 Jul 2021 18:28:42 +0000 (19:28 +0100)]
[SLP][X86] Add dot product tests based off PR51075
Nico Weber [Sat, 17 Jul 2021 15:18:48 +0000 (11:18 -0400)]
[lld/mac] Change load command order to be more like ld64
No meaningful behavior change. Makes diffing `otool -l` output a bit easier.
Differential Revision: https://reviews.llvm.org/D106219
Haowei Wu [Mon, 19 Jul 2021 18:59:04 +0000 (11:59 -0700)]
[ifs] Fix linking errors on some llvm builders
This change fixes linking errors on some llvm builders.
Artem Belevich [Wed, 30 Jun 2021 22:36:14 +0000 (15:36 -0700)]
[MemCpyOpt] Enable memcpy optimizations unconditionally.
The patch does not depend on the availability of the library functions for
memcpy/memset as it operates on LLVM intrinsics. The optimizations are useful
on the targets that have these functions disabled (e.g. NVPTX & AMDGPU).
Differential Revision: https://reviews.llvm.org/D104801
Geoffrey Martin-Noble [Mon, 19 Jul 2021 18:48:52 +0000 (11:48 -0700)]
[Bazel] Update for
6103fdfab4
Update Bazel config for
https://github.com/llvm/llvm-project/commit/
6103fdfab4
by deleting the llvm-elfabi target.
Differential Revision: https://reviews.llvm.org/D106295
David Carlier [Mon, 19 Jul 2021 18:46:35 +0000 (19:46 +0100)]
[Sanitizer] Intercepts flopen/flopenat on FreeBSD.
Reviewers: vitalybuka
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D106218
Louis Dionne [Mon, 19 Jul 2021 15:23:09 +0000 (11:23 -0400)]
[libc++] Disable #pragma system_header in the new testing configuration
The new testing configuration did not turn off #pragma system_header,
which means we were not seeing warnings in system headers.
Differential Revision: https://reviews.llvm.org/D106187
LLVM GN Syncbot [Mon, 19 Jul 2021 18:24:11 +0000 (18:24 +0000)]
[gn build] Port
8b4acb067fd3
LLVM GN Syncbot [Mon, 19 Jul 2021 18:24:10 +0000 (18:24 +0000)]
[gn build] Port
61fa9afe4c5b
LLVM GN Syncbot [Mon, 19 Jul 2021 18:24:09 +0000 (18:24 +0000)]
[gn build] Port
6103fdfab4e2
Haowei Wu [Fri, 2 Apr 2021 03:55:11 +0000 (20:55 -0700)]
[ifs][elfabi] Merge llvm-ifs/elfabi tools
This change merges llvm-elfabi and llvm-ifs tools.
Differential Revision: https://reviews.llvm.org/D100139
Haowei Wu [Wed, 7 Apr 2021 22:50:12 +0000 (15:50 -0700)]
[ifs] Prepare llvm-ifs for elfabi/ifs merging.
This diff changes llvm-ifs to use unified IFS file format
and perform other renaming changes in preparation for the
merging between elfabi/ifs.
Differential Revision: https://reviews.llvm.org/D99810
Haowei Wu [Wed, 31 Mar 2021 23:48:56 +0000 (16:48 -0700)]
[elfabi] Prepare elfabi/ifs merging.
This change implements unified text stub format and command line
interface proposed in the elfabi/ifs merge plan.
Differential Revision: https://reviews.llvm.org/D99399
Jonas Paulsson [Fri, 16 Jul 2021 09:23:46 +0000 (11:23 +0200)]
[SystemZ] Handle NoRegister in SystemZTargetLowering::emitMemMemWrapper().
Bugfix: The compiler should be able to generate a memset to nullptr.
Review: Ulrich Weigand
Amy Huang [Mon, 19 Jul 2021 17:42:28 +0000 (10:42 -0700)]
Revert "[llvm][sve] Lowering for VLS truncating stores" because it
causes a seg fault (see https://reviews.llvm.org/D104471).
This reverts commit
c305557acdaad453e32309d575fe9c6c7090c099.