Nikita Popov [Fri, 9 Sep 2022 14:24:14 +0000 (16:24 +0200)]
[AST] Use BatchAA in aliasesPointer() (NFC)
Nico Weber [Thu, 29 Sep 2022 00:18:27 +0000 (20:18 -0400)]
[lldb] Fix deprecation warning for using std::iterator
std::iterator was deprecated in C++17.
No behavior change.
Differential Revision: https://reviews.llvm.org/D134844
Pierre van Houtryve [Thu, 29 Sep 2022 08:30:20 +0000 (08:30 +0000)]
[AMDGPU][GISel] Update `isCanonicalized`
Recognize more opcodes in the function.
Fixes some regressions introduced in D134857 for fdiv.f16 too.
Depends on D134857
Reviewed By: arsenm, foad
Differential Revision: https://reviews.llvm.org/D134862
Pierre van Houtryve [Thu, 29 Sep 2022 07:57:29 +0000 (07:57 +0000)]
[GISel] Add more cases to isKnownNeverNaN
Make it even with the DAG implementation as of D134854
Reviewed By: arsenm, foad
Differential Revision: https://reviews.llvm.org/D134857
Nikita Popov [Fri, 9 Sep 2022 14:31:36 +0000 (16:31 +0200)]
[LICM] Collect more scalar promotion stats (NFC)
Collect more statistics for scalar promotion. In particular,
keep track of how many promotion candidates there were, and
whether it is a load or a load/store promotion.
Alex Brachet [Fri, 30 Sep 2022 14:06:51 +0000 (14:06 +0000)]
[Driver][Fuchsia] Add default linker flags
Differential Revision: https://reviews.llvm.org/D132416
Pierre van Houtryve [Fri, 30 Sep 2022 13:44:30 +0000 (13:44 +0000)]
[AMDGPU][GISel] Add Identity BUILD_VECTOR Combines
Folds-away BUILD_VECTOR-related noops in the post-legalizer combiner.
Depends on D134433
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D134953
Pierre van Houtryve [Fri, 30 Sep 2022 11:08:29 +0000 (11:08 +0000)]
[AMDGPU][GISel] Legalize V2S16 G_BUILD_VECTOR
Preparation patch for D134354 to make V2S16 G_BUILD_VECTOR legal.
Also removes RegBankInfo's scalarization of small BUILD_VECTORs,
replacing it with InstructionSelector logic instead.
This allows for V2S16 BUILD_VECTOR instructions to survive
all the way to ISel so we can select FMA/MAD_MIX instructions
in D134354.
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D134433
Sanjay Patel [Fri, 30 Sep 2022 12:58:13 +0000 (08:58 -0400)]
[SCCP] add tests for sitofp; NFC
Adapted from the existing tests for ashr, sdiv, srem.
Ivan Kosarev [Fri, 30 Sep 2022 13:24:08 +0000 (14:24 +0100)]
[AMDGPU][SetWavePriority] Fix dealing with MBBInfo records.
Happened earlier than I anticipated. :)
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D134726
Jitka Plesnikova [Wed, 21 Sep 2022 09:42:46 +0000 (11:42 +0200)]
[lldb] Fix 'error: non-const lvalue...' caused by SWIG 4.1.0
Fix the failure caused by change in SwigValueWraper for C++11 and later
for improved move semantics in SWIG commit.
https://github.com/swig/swig/commit/
d1055f4b3d51cb8060893f8036846ac743302dab
Amaury Séchet [Fri, 30 Sep 2022 12:30:27 +0000 (12:30 +0000)]
[NFC] Fix erroneous indentation.
Sanjay Patel [Fri, 30 Sep 2022 11:56:15 +0000 (07:56 -0400)]
[InstSimplify] look through vector select (shuffle) in min/max fold
This is an extension of the existing min/max+select fold (which already
has a very large number of variations) to allow a vector shuffle because
that's what we have in the motivating example from issue #42100.
A couple of Alive2 checks of variants (I don't know how to generalize
these in Alive):
https://alive2.llvm.org/ce/z/jUFAqT
And verify the PR42100 test:
https://alive2.llvm.org/ce/z/3EcASf
It's possible there is some generalization of the fold or a
VectorCombine/SLP answer for the motivating test, but I haven't found a
better/smaller solution yet.
We can also add even more variants here as follow-up patches. For example,
we can have shuffle followed by min/max; we also don't have this
canonicalization or the reverse:
https://alive2.llvm.org/ce/z/StHD9f
Differential Revision: https://reviews.llvm.org/D134879
Simon Pilgrim [Fri, 30 Sep 2022 12:26:24 +0000 (13:26 +0100)]
[SLP][X86] Add test coverage for Issue #58054
Nikita Popov [Fri, 30 Sep 2022 12:24:58 +0000 (14:24 +0200)]
[InstCombine] Add test for PR57488 (NFC)
Dmitry Vyukov [Fri, 30 Sep 2022 11:58:27 +0000 (13:58 +0200)]
tsan: fix deadlock/crash in signal handling
We set in_blocking_func around some blocking C functions so that we don't
delay signal infinitely (if in_blocking_func is set we deliver signals
synchronously).
However, pthread_join is blocking but also call munmap/free to free thread
resources. If we are inside the munmap/free interceptors called from
pthread_join and deliver a signal synchronously, it can lead to deadlocks
and crashes since we re-enter runtime and try to lock the same mutexes
or use the same per-thread data structures.
If we re-enter runtime via an interceptor when in_blocking_func is set,
temporary reset in_blocking_func around the interceptor and restore it back
when we return from the recursive interceptor.
Also move in_blocking_func from ThreadSignalContext to ThreadContext
so that we can CHECK that it's not set in SlotLocker ctor.
Fixes https://github.com/google/sanitizers/issues/1540
Reviewed By: melver
Differential Revision: https://reviews.llvm.org/D127845
Alex Zinenko [Fri, 30 Sep 2022 11:07:39 +0000 (11:07 +0000)]
[mlir] fix DiagnosedSilenceableFailure::takeDiagnostics
This function was returning an rvalue reference to an object that was
also cleared via RAII when the function returned, making it always
return an empty object. Make it accept the mutable reference to the
object instead to avoid this dangerous behavior.
Reviewed By: guraypp
Differential Revision: https://reviews.llvm.org/D134948
Zain Jaffal [Fri, 30 Sep 2022 11:32:25 +0000 (12:32 +0100)]
[AArch64] Add additional tests for SMULL instruction selection
Add tests where the operands are switched and where the top bit of the operand is set to 1
Reviewed By: fhahn
Differential Revision: https://reviews.llvm.org/D134867
Tonko Sabolčec [Fri, 30 Sep 2022 11:10:32 +0000 (11:10 +0000)]
[lldb] Fix member access in GetExpressionPath
This change fixes two issues in ValueObject::GetExpressionPath method:
1. Accessing members of struct references used to produce expression
paths such as "str.&str.member" (instead of the expected
"str.member"). This is fixed by assigning the flag tha the child
value is a dereference when calling Dereference() on references
and adjusting logic in expression path creation.
2. If the parent of member access is dereference, the produced
expression path was "*(ptr).member". This is incorrect, since it
dereferences the member instead of the pointer. This is fixed by
wrapping dereference expression into parenthesis, resulting with
"(*(ptr)).member".
Reviewed By: werat, clayborg
Differential Revision: https://reviews.llvm.org/D132734
Nikita Popov [Fri, 30 Sep 2022 11:06:53 +0000 (13:06 +0200)]
[InstCombine] Add test for phi translation during select of phi fold (NFC)
The phi translation performed during this fold is important for
correctness, but was apparently untested.
Alexander Kornienko [Fri, 30 Sep 2022 11:04:07 +0000 (13:04 +0200)]
Cleanup: avoid referring to std::vector<T> members when T is incomplete.
This is not legal according to the C++ standard, and causes build errors in
particular in C++20 mode. Fix it by defining the vector's type before using the
vector.
Patch by poompatai@google.com.
Nikita Popov [Fri, 30 Sep 2022 11:04:23 +0000 (13:04 +0200)]
[InstCombine] Regenerate test checks (NFC)
Alvin Wong [Wed, 28 Sep 2022 09:58:51 +0000 (12:58 +0300)]
[lldb][Windows] Always call SetExecutableModule on debugger connected
In `ProcessWindows::OnDebuggerConnected` (triggered from
`CREATE_PROCESS_DEBUG_EVENT`), we should always call
`Target::SetExecutableModule` regardless of whether LLDB has already
preloaded the executable modules. `SetExecutableModule` has the side
effect of clearing the module list of the Target, which help make sure
that module #0 is the executable module and the rest of the modules are
listed according to the DLL load order in the process (technically this
has no real consequences but it seems to make more sense anyway.) It
also fixes an issue where the modules preloaded by LLDB will be
duplicated when the debuggee process actually loads the DLL.
Reviewed By: labath
Differential Revision: https://reviews.llvm.org/D134636
Adrian Kuegel [Fri, 30 Sep 2022 10:47:32 +0000 (12:47 +0200)]
[mlir][SCF] Apply ClangTidyPerformance finding (NFC)
Simon Pilgrim [Fri, 30 Sep 2022 10:19:30 +0000 (11:19 +0100)]
Revert rG1b7089fe67b924bdd5ecef786a34bdba7a88778f "[SLP] Add ScalarizationOverheadBuilder helper to track vector extractions"
Revert rGef89409a59f3b79ae143b33b7d8e6ee6285aa42f "Fix 'unused-lambda-capture' gcc warning. NFCI."
Revert rG926ccfef032d206dcbcdf74ca1e3a9ebf4d1be45 "[SLP] ScalarizationOverheadBuilder - demand all elements for scalarization if the extraction index is unknown / out of bounds"
Revert ScalarizationOverheadBuilder sequence from D134605 - when accumulating extraction costs by Type (instead of specific Value), we are not distinguishing enough when they are coming from the same source or not, and we always just count the cost once. This needs addressing before we can use getScalarizationOverhead properly.
Florian Hahn [Fri, 30 Sep 2022 10:21:58 +0000 (11:21 +0100)]
[LoopDeletion] Clear block & loop dispo cache after breaking backedge.
breakLoopBackedge may remove blocks and loops. Also clear block &
loop disposition to avoid the cache containing invalid blocks and loops.
The coverage for the change is provided when using an ASAN build of opt
to run the LoopDeletion unit tests; without the fix, pointers to invalid
objects would be used.
Reviewed By: nikic
Differential Revision: https://reviews.llvm.org/D134663
Simon Pilgrim [Fri, 30 Sep 2022 10:07:46 +0000 (11:07 +0100)]
[SLP][X86] Add test case for crash reported on D134605
Simon Pilgrim [Fri, 30 Sep 2022 10:03:14 +0000 (11:03 +0100)]
[SLP][AArch64] Add test case for vectorization regression case reported on D134605
Oleg Shyshkov [Fri, 30 Sep 2022 10:03:00 +0000 (10:03 +0000)]
[mlir] Rename getTied* methods to getMatching* in LinalgInterface.
Summary:
As mentioned in the comment to https://reviews.llvm.org/D134444, the term `tied`
is a misnomer in this context and `matching` sounds much better.
Differential Revision: https://reviews.llvm.org/D134534
Zain Jaffal [Fri, 30 Sep 2022 10:03:48 +0000 (11:03 +0100)]
[AArch64] Add support for 128-bit non temporal loads.
Adding to the work done in `D131773` here we add support to 128-bit loads.
Reviewed By: fhahn
Differential Revision: https://reviews.llvm.org/D132559
Juan Manuel MARTINEZ CAAMAÑO [Tue, 27 Sep 2022 12:15:36 +0000 (12:15 +0000)]
[DebugInfo][LICM] Drop DebugLoc from IntrinsicInst when hoisting
The DebugLoc is conserved when hoisting function calls, to ensure the
DIScope is preserved if inlining occurs.
This commit drops the DebugLoc in the case the call is an intrinsic
call that won't be lowered into a function call.
Differential Revision: https://reviews.llvm.org/D134429
gonglingqin [Fri, 30 Sep 2022 06:49:35 +0000 (14:49 +0800)]
[LoongArch] Clean up redundant code introduced by conflict resolution. NFC
Roy Jacobson [Fri, 30 Sep 2022 08:24:46 +0000 (11:24 +0300)]
Revert "[Clang] Fix variant crashes from GH58028, GH57370"
This reverts commit
9706bb3165f5e508d5e2247ad8a3f45077df546d, some CI workers complain about the test.
Roy Jacobson [Thu, 29 Sep 2022 20:19:45 +0000 (23:19 +0300)]
[Clang] define __cpp_named_character_escapes
Define the feature test macro for named character escapes.
I assume this was not done because it was implemented before formally accepted, right? cxx_status says the paper is implemented.
Reviewed By: cor3ntin
Differential Revision: https://reviews.llvm.org/D134898
Roy Jacobson [Thu, 29 Sep 2022 15:38:46 +0000 (18:38 +0300)]
[Clang] Fix variant crashes from GH58028, GH57370
Fixes a null dereference in some diagnostic issuing code.
Closes https://github.com/llvm/llvm-project/issues/57370
Closes https://github.com/llvm/llvm-project/issues/58028
Reviewed By: shafik
Differential Revision: https://reviews.llvm.org/D134885
Siva Chandra Reddy [Wed, 7 Sep 2022 23:04:09 +0000 (23:04 +0000)]
[libc] Re-enable functions from signal.h and re-enable abort.
They were disabled because we were including linux/signal.h from our
signal.h. Linux's signal.h is not designed to be included from user
programs as it causes a lot of non-standard name pollution. Also, it is
not self-contained. This change defines types and macros relevant for
signal related syscalls within libc's headers and removes inclusion of
Linux headers.
This patch enables the funtions only for x86_64. They will be enabled
for aarch64 also in a follow up patch after testing.
Reviewed By: abrachet, lntue
Differential Revision: https://reviews.llvm.org/D134567
Pierre van Houtryve [Thu, 29 Sep 2022 08:11:09 +0000 (08:11 +0000)]
[TableGen] Add `countRendererFns` to `InstructionOperandMatcher`
Without it, the count of renderer functions is inaccurate and, in some
edge cases (like the patterns added in D134354), we can actually
go out of bounds (run out of pre-allocated renderer function spaces
in the GISel state)
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D134861
Michał Górny [Thu, 29 Sep 2022 20:38:52 +0000 (22:38 +0200)]
[clang] [Driver] Disable default configs via envvar during testing
Add support for a CLANG_NO_DEFAULT_CONFIG envvar that works like
the --no-default-config option when set to a non-empty value. Use it
to disable loading system configuration files during the test suite
runs.
Configuration files can change the driver behavior in extensive ways,
and it is neither really possible nor feasible to account for or undo
the effects of even the most common configuration uses. Therefore,
the most reasonable option seems to be to ignore configuration files
while running the majority of tests (with the notable exception of tests
for configuration file support).
Due to the diversity of ways that %clang is used in the test suite,
including using it to copy or symlink the clang executable, as well to
call -cc1 and -cc1as modes, it is not feasible to pass the explicit
options to disable config loading either. Using an environment variable
has the advantage of being easily applied across the test suite
and easily unset for default configuration file loading tests.
Differential Revision: https://reviews.llvm.org/D134905
serge-sans-paille [Thu, 29 Sep 2022 19:48:38 +0000 (21:48 +0200)]
[lldb] Get rid of __STDC_LIMIT_MACROS and __STDC_CONSTANT_MACROS
C++11 made the use of these macro obsolete, see https://sourceware.org/bugzilla/show_bug.cgi?id=15366
As a side effect this prevents https://github.com/swig/swig/issues/2193.
Differential Revision: https://reviews.llvm.org/D134877
Adrian Kuegel [Fri, 30 Sep 2022 06:36:32 +0000 (08:36 +0200)]
[mlir][Linalg] Reland Add ReduceOp to Linalg structured ops.
This op will allow to model (variadic) reductions with this special op
instead of using GenericOp.
This reverts commit
535fd753ef8dcfe5c73219ea7173e015a74e4f59.
Additional fix: implement a getLibraryName method.
Adrian Kuegel [Fri, 30 Sep 2022 06:33:10 +0000 (08:33 +0200)]
Vitaly Buka [Fri, 30 Sep 2022 06:32:16 +0000 (23:32 -0700)]
[sanitizer] Fix symbolizer build script
Timm Bäder [Fri, 16 Sep 2022 07:17:52 +0000 (09:17 +0200)]
[clang][Interp] Handle enums
Handle DeclRefExprs of enum types. They are otherwise handled like
integers.
Differential Revision: https://reviews.llvm.org/D134020
Michał Górny [Fri, 30 Sep 2022 05:59:41 +0000 (07:59 +0200)]
[clang] [test] Use --implicit-check-not in config-file3 tests
Thanks to Fangrui Song for the suggestion.
Phoebe Wang [Fri, 30 Sep 2022 06:00:42 +0000 (14:00 +0800)]
[Doc] Fix typo. NFC
Fixes #56429
Petr Hosek [Fri, 30 Sep 2022 05:52:49 +0000 (05:52 +0000)]
Revert "[CMake] Use libcxx targets for in-tree sanitizer C++ ABI"
This reverts commit
7dbdffefb768062f28599c256b5e013ee26ff04e since
it appears to have broken several ASan tests.
Kristof Beyls [Fri, 30 Sep 2022 05:42:06 +0000 (07:42 +0200)]
Fix spelling in Co-authored-by documentation
Fangrui Song [Fri, 30 Sep 2022 04:26:53 +0000 (21:26 -0700)]
[ELF] Reset verdefIndex for Defined preempting SharedSymbol
to avoid spurious "attempt to reassign symbol '...'" warning after
7a58dd1046a8052619d173b769f32f2df3aafbe8
Lang Hames [Fri, 30 Sep 2022 02:16:19 +0000 (19:16 -0700)]
[JITLink] Update external symbol scopes to reflect scopes of resolved defs.
This is a counterpart to
ffe2dda29f3, and does for scope what that commit did
for linkage.
Making the scope of external definitions visible to JITLink plugins will
allow us to distinguish hidden weak defs (which do not need to be tracked by
default) from default-scoped weak defs (which need to be updated to point at
a single chosen definition at runtime).
Yeting Kuo [Wed, 14 Sep 2022 08:29:53 +0000 (16:29 +0800)]
[SelectionDAG] Add helper function to check whether a SDValue is neutral element. NFC.
Using this helper makes work about neutral elements more easier. Although I only
find one case now, I think it will have more chance to be used since so many
combine works are related to neutral elements.
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D133866
Nirvedh Meshram [Thu, 29 Sep 2022 18:14:18 +0000 (11:14 -0700)]
[mlir][spirv] Handle dynamic/static cases differntly for kernel capability
Reviewed By: antiagainst
Differential Revision: https://reviews.llvm.org/D134908
LLVM GN Syncbot [Fri, 30 Sep 2022 01:33:02 +0000 (01:33 +0000)]
[gn build] Port
91c96a806cae
Petr Hosek [Thu, 29 Sep 2022 08:48:03 +0000 (08:48 +0000)]
[CMake] Use correct include flag for MSVC
When building using the MSVC driver, we need to use /imsvc rather than
-isystem for system headers.
Differential Revision: https://reviews.llvm.org/D134864
Petr Hosek [Thu, 29 Sep 2022 07:53:06 +0000 (07:53 +0000)]
[CMake] Use libcxx targets for in-tree sanitizer C++ ABI
When in-tree libcxx is selected as the sanitizer C++ ABI, use libcxx
targets rather than libcxxabi and libunwind, since those may not be
available on all platforms (Windows for example).
Differential Revision: https://reviews.llvm.org/D134855
Mircea Trofin [Fri, 30 Sep 2022 01:26:40 +0000 (18:26 -0700)]
Revert "[MLGO] ML Regalloc Priority Advisor"
This reverts commit
8f4f26ba5bd04f7b335836021e5e63b4236c0305.
Buildbot failures, e.g. https://lab.llvm.org/buildbot/#/builders/6/builds/14041
wren romano [Fri, 30 Sep 2022 00:24:01 +0000 (17:24 -0700)]
[mlir][sparse] Fixing case coverage warning
Followup to D133835 for fixing the warning on LLVM's Windows buildbot
Reviewed By: aartbik, Peiming, stella.stamenova
Differential Revision: https://reviews.llvm.org/D134925
Amaury Séchet [Fri, 30 Sep 2022 00:58:04 +0000 (00:58 +0000)]
[DAG] Simplify the select of constant combine code. NFC
Amaury Séchet [Sat, 6 Aug 2022 02:10:22 +0000 (02:10 +0000)]
[DAG] select Cond, C, -1 --> or (sext (not Cond)), C when C is MVT::i1
In the spirit of D130765 . Get rid of cbranches and/or cmov. Usually shorter, but sometime not, becaus eit's hard to prededict when dependency breaking xor will be introduced.
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D134736
Peiming Liu [Thu, 29 Sep 2022 21:59:02 +0000 (21:59 +0000)]
[mlir][sparse] refactoring: split translateIndices.
TranslateIndicesArray take an array of SSA value and convert them into another array of SSA values based on reassociation. Which makes it easier to be reused by `foreach` operator (as the indices array are given as an array of SSA values).
Reviewed By: aartbik, bixia
Differential Revision: https://reviews.llvm.org/D134918
Michael Jones [Tue, 27 Sep 2022 23:52:21 +0000 (16:52 -0700)]
[libc][windows] rename fenv internals for windows
On windows, including math.h causes macros for "OVERFLOW" and
"UNDERFLOW" to be defined. This patch renames some variables internal to
FEnvImpl.h to avoid colliding with those.
Reviewed By: sivachandra
Differential Revision: https://reviews.llvm.org/D134775
bixia1 [Thu, 29 Sep 2022 21:50:42 +0000 (14:50 -0700)]
[mlir][sparse] Allow the push_back operator to skip capacity check and reallocation.
Add UnitAttr `inbounds` for this purpose.
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D134913
Ben Dunbobbin [Thu, 29 Sep 2022 23:21:31 +0000 (00:21 +0100)]
[IR] Don't allow DLL storage-class and local linkage
Disallow this meaningless combination. Doing so simplifies analysis
of LLVM code w.r.t t DLL storage-class, and prevents mistakes with
DLL storage class.
- Change the assembler to reject DLL storage class on symbols with
local linkage.
- Change the bitcode reader to clear the DLL Storage class when the
linkage is local for auto-upgrading
- Update LangRef.
There is an existing restriction on non-default visibility and local
linkage which this is modelled on.
Differential Review: https://reviews.llvm.org/D134784
Aart Bik [Sat, 17 Sep 2022 00:55:50 +0000 (17:55 -0700)]
[mlir][sparse] implement singleton dimension level type
This is a first step towards fully implementing the new dimension
level types and properties, illustrating with a fully functional
sorted COO of any dimension. Note that the sparsification part is
pretty complete. The required parts in the runtime support library
have been kept to a minimum, to avoid huge conflicts with Wren's
ongoing refactoring. The missing parts will be filled in later.
Reviewed By: Peiming
Differential Revision: https://reviews.llvm.org/D134096
Steven Wu [Thu, 29 Sep 2022 23:20:47 +0000 (16:20 -0700)]
[MC] Properly disable debug-frame test on AppleSilicon Mac
Disable debug-frame tests on AppleSilicon Mac to improve some of the fix
in PR46647.
Reviewed By: aprantl
Differential Revision: https://reviews.llvm.org/D134896
Jeff Niu [Thu, 29 Sep 2022 20:45:44 +0000 (13:45 -0700)]
[mlir] Use `interleave` in `printOperands` (NFC)
Instead of reimplementing it.
Reviewed By: rriddle
Differential Revision: https://reviews.llvm.org/D134904
Jeff Niu [Thu, 29 Sep 2022 21:43:08 +0000 (14:43 -0700)]
[mlir] Add `OpAsmPrinter::printOptionalLocationSpecifier`
This is the corresponding method to
`OpAsmParser::parseOptionalLocationSpecifier` that prints a location
`loc(...)` based on the op printing flags. Together, these two functions
allow propagating user-level location info outside of their usual spots.
Reviewed By: rriddle
Differential Revision: https://reviews.llvm.org/D134910
Adrian Prantl [Thu, 29 Sep 2022 22:56:31 +0000 (15:56 -0700)]
add missing textual header to modulemap
Rahul Joshi [Thu, 29 Sep 2022 22:33:49 +0000 (15:33 -0700)]
Include <cmath> in FileUtilities.cpp for std:abs(double)
wren romano [Thu, 29 Sep 2022 20:13:48 +0000 (13:13 -0700)]
[mlir][sparse] Improve sparse_tensor::detail::readCOOValue template
This is a followup to the refactoring of D133462, D133830, D133831, and D133833.
Depends On D133833
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D133836
LLVM GN Syncbot [Thu, 29 Sep 2022 22:13:14 +0000 (22:13 +0000)]
[gn build] Port
8f4f26ba5bd0
wren romano [Thu, 29 Sep 2022 20:35:17 +0000 (13:35 -0700)]
[mlir][sparse] optimizing permutation validity check in toMLIRSparseTensor
The previous sorting-based check was O(n*log n). The new implementation is O(n).
This is a followup to the refactoring of D133462, D133830, D133831, and D133833.
Depends On D133833
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D133838
Siva Chandra [Thu, 29 Sep 2022 21:41:46 +0000 (14:41 -0700)]
[libc][Obvious] Enable some of the recently added functions on aarch64.
bixia1 [Thu, 29 Sep 2022 21:09:04 +0000 (14:09 -0700)]
[mlir][sparse] Move the implementation of sparse_tensor.push_back to the buffer rewriter.
Reviewed By: aartbik, Peiming
Differential Revision: https://reviews.llvm.org/D134777
wren romano [Thu, 29 Sep 2022 20:39:30 +0000 (13:39 -0700)]
[mlir][sparse] Adding isSorted bit to SparseTensorCOO
This is a followup to the refactoring of D133462, D133830, D133831, and D133833.
Depends On D133833
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D133839
wren romano [Thu, 29 Sep 2022 20:31:14 +0000 (13:31 -0700)]
[mlir][sparse] Cleaning up SparseTensorFile::readMMEHeader
This is a followup to the refactoring of D133462, D133830, D133831, and D133833.
Depends On D133833
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D133837
Eric Wang [Fri, 9 Sep 2022 19:20:27 +0000 (14:20 -0500)]
[MLGO] ML Regalloc Priority Advisor
The bulk of the implementation is common between 'release' mode (==AOT-ed
model) and 'development' mode (for training), the main difference is
that in development mode, we may also log features (for training logs),
inject scoring information and then produce the log file.
Differential Revision: https://reviews.llvm.org/D133616
Amara Emerson [Thu, 29 Sep 2022 16:22:25 +0000 (17:22 +0100)]
[AArch64][GlobalISel] Implement another combine for shufflevector->AArch64 G_EXT.
This is a port of an existing optimization in AArch64 ISelLowering, handling
a case when the same input vector can be used for both ext inputs.
Differential Revision: https://reviews.llvm.org/D134891
wren romano [Thu, 29 Sep 2022 20:09:24 +0000 (13:09 -0700)]
[mlir][sparse] Factoring out SparseTensorFile::canReadAs predicate
This is a followup to the refactoring of D133462, D133830, D133831, and D133833.
Depends On D133833
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D133835
wren romano [Thu, 29 Sep 2022 20:04:28 +0000 (13:04 -0700)]
[mlir][sparse] refactoring SparseTensorUtils: (4 of 4) documentation
Previously, the SparseTensorUtils.cpp library contained a C++ core implementation, but hid it in an anonymous namespace and only exposed a C-API for accessing it. Now we are factoring out that C++ core into a standalone C++ library so that it can be used directly by downstream clients (per request of one such client). This refactoring has been decomposed into a stack of differentials in order to simplify the code review process, however the full stack of changes should be considered together.
* D133462: Part 1: split one file into several
* D133830: Part 2: Reorder chunks within files
* D133831: Part 3: General code cleanup
* (this): Part 4: Update documentation
This part updates existing documentation, adds new documentation, and reflows the test for some existing documentation.
Depends On D133831
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D133833
wren romano [Thu, 29 Sep 2022 19:55:32 +0000 (12:55 -0700)]
[mlir][sparse] refactoring SparseTensorUtils: (3 of 4) code-cleanup
Previously, the SparseTensorUtils.cpp library contained a C++ core implementation, but hid it in an anonymous namespace and only exposed a C-API for accessing it. Now we are factoring out that C++ core into a standalone C++ library so that it can be used directly by downstream clients (per request of one such client). This refactoring has been decomposed into a stack of differentials in order to simplify the code review process, however the full stack of changes should be considered together.
* D133462: Part 1: split one file into several
* D133830: Part 2: Reorder chunks within files
* (this): Part 3: General code cleanup
* D133833: Part 4: Update documentation
This part performs some general code cleanup including:
* making more things `const`, especially for the targets of pointers
* using preincrement wherever possible ([[ https://llvm.org/docs/CodingStandards.html#prefer-preincrement | per LLVM style guide ]])
* adding messages to most `assert` statments.
* moving argument casting from the core function/method definitions to the CPP wrappers
Depends On D133830
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D133831
wren romano [Thu, 29 Sep 2022 18:55:51 +0000 (11:55 -0700)]
[mlir][sparse] refactoring SparseTensorUtils: (2 of 4) reordering
Previously, the SparseTensorUtils.cpp library contained a C++ core implementation, but hid it in an anonymous namespace and only exposed a C-API for accessing it. Now we are factoring out that C++ core into a standalone C++ library so that it can be used directly by downstream clients (per request of one such client). This refactoring has been decomposed into a stack of differentials in order to simplify the code review process, however the full stack of changes should be considered together.
* D133462: Part 1: split one file into several
* (this): Part 2: Reorder chunks within files
* D133831: Part 3: General code cleanup
* D133833: Part 4: Update documentation
This part moves chunks of code within files, but again aims to make no other changes. Many of these movements are part of a stylistic shift to reorder the components of class definitions as follows: data members, ctors/factories, getters, other public methods, private methods.
Depends On D133462
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D133830
wren romano [Thu, 29 Sep 2022 18:40:40 +0000 (11:40 -0700)]
[mlir][sparse] refactoring SparseTensorUtils: (1 of 4) file-splitting
Previously, the SparseTensorUtils.cpp library contained a C++ core implementation, but hid it in an anonymous namespace and only exposed a C-API for accessing it. Now we are factoring out that C++ core into a standalone C++ library so that it can be used directly by downstream clients (per request of one such client). This refactoring has been decomposed into a stack of differentials in order to simplify the code review process, however the full stack of changes should be considered together.
* (this): Part 1: split one file into several
* D133830: Part 2: Reorder chunks within files
* D133831: Part 3: General code cleanup
* D133833: Part 4: Update documentation
This part aims to make no changes other than the 1:N file splitting, and things which are forced to accompany that change.
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D133462
Xiang Li [Thu, 29 Sep 2022 21:26:01 +0000 (14:26 -0700)]
[NFC] Use %clang_cc1
Xiang Li [Thu, 29 Sep 2022 21:19:12 +0000 (14:19 -0700)]
[NFC] Move cc1 test out of Driver test
Joshua Batista [Thu, 29 Sep 2022 20:43:24 +0000 (13:43 -0700)]
[HLSL] add ceil library function
This change exposes the ceil library function for HLSL,
excluding long, int, and long long doubles.
Ceil is supported for all scalar, vector, and matrix types.
Long and long long double support is missing in this patch because those types
don't exist in HLSL. Int is missing because the ceil function only works on floating type arguments.
The full documentation of the HLSL ceil function is available here:
https://docs.microsoft.com/en-us/windows/win32/direct3dhlsl/dx-graphics-hlsl-ceil
Reviewed By: python3kgae
Differential Revision: https://reviews.llvm.org/D134319
Philip Reames [Thu, 29 Sep 2022 21:01:14 +0000 (14:01 -0700)]
[RISCV] Minor code motion in InsertVSETVLI [nfc]
Alex Brachet [Thu, 29 Sep 2022 20:57:10 +0000 (20:57 +0000)]
[scudo][Fuchsia] Increase MaxNumCachedHint
This brings the value closer to other platforms and allows
for future improvements, see D133897.
Siva Chandra Reddy [Thu, 29 Sep 2022 07:57:52 +0000 (07:57 +0000)]
[libc] Add GNU extension functions sched_getaffinity and sched_setaffinity.
Reviewed By: michaelrj
Differential Revision: https://reviews.llvm.org/D134858
Siu Chi Chan [Tue, 13 Sep 2022 17:42:43 +0000 (13:42 -0400)]
[HIP] stop forcing the lang std in the driver
D103221 changed HIP's default to C++14, removing the driver logic to
force it into a different std.
Change-Id: I9f5220a7456687039b0bd3b3574f3124d3cc7665
Differential Revision: https://reviews.llvm.org/D134314
Change-Id: I40513f2ebe93ee53ea95c8bb3cc704487d970263
Roy Jacobson [Thu, 29 Sep 2022 20:16:17 +0000 (23:16 +0300)]
[Clang][NFC] Fix broken link in release notes
Roy Jacobson [Thu, 29 Sep 2022 15:45:03 +0000 (18:45 +0300)]
[Clang] P1169R4: static operator()
Implements 'P1169R4: static operator()' from C++2b.
Reviewed By: #clang-language-wg, aaron.ballman
Differential Revision: https://reviews.llvm.org/D133659
Florian Hahn [Thu, 29 Sep 2022 19:49:55 +0000 (20:49 +0100)]
[SCEVExpander] Move LCSSA fixup to ::expand.
Move LCSSA fixup from ::expandCodeForImpl to ::expand(). This has
the advantage that we directly preserve LCSSA nodes here instead of
relying on doing so in rememberInstruction. It also ensures that we
don't add the non-LCSSA-safe value to InsertedExpressions.
Alternative to D132704.
Fixes #57000.
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D134739
Bjorn Pettersson [Thu, 29 Sep 2022 16:22:56 +0000 (18:22 +0200)]
[X86] Avoid miscompile in combineOr (X86ISelLowering.cpp)
In combineOr (X86ISelLowering.cpp) there is a DAG combine that rewrite
a "(0 - SetCC) | C" pattern into something simpler given that a LEA
can be used. Another requirement is that C has some specific value,
for example 1 or 7. When checking those requirements the code used a
32-bit unsigned variable to store the value of C. So for a 64-bit OR
this could miscompile in case any of the 32 most significant bits in
C were non zero.
This patch adds fixes the bug by using a large enough type for the
C value.
The faulty code seem to have been introduced by commit
9bceb8981d32fe
(D131358).
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D134892
Bjorn Pettersson [Thu, 29 Sep 2022 16:14:15 +0000 (18:14 +0200)]
[X86] Pre-commit test case showing bug in combineOr (X86ISelLowering.cpp)
In combineOr (X86ISelLowering.cpp) there is a DAG combine that rewrite
a "(0 - SetCC) | C" pattern into something simpler given that a LEA
can be used. Another requirement is that C has some specific value,
for example 1 or 7. When doing that check it is using a 32-bit
unsigned variable to store the value of C. So for a 64-bit OR this
could miscompile in case any of the 32 most significant bits in C
are set.
This patch adds a test case to show this miscompile bug.
Differential Revision: https://reviews.llvm.org/D134890
Fangrui Song [Thu, 29 Sep 2022 19:21:57 +0000 (12:21 -0700)]
[Driver][test] Restore %clang -cc1 in test/Driver
This partially reverts commit
1609a5d7715c06ff52c13af8b20ee64811a8ec7b (the
test/Driver part). We want to discourage %clang_cc1 and clang -cc1 in
test/Driver. The clang -cc1 uses in hlsl/offload/etc are not good examples.
Aart Bik [Thu, 29 Sep 2022 19:09:09 +0000 (12:09 -0700)]
[mlir][sparse] fix build breakage due to Arith name change
Reviewed By: Peiming
Differential Revision: https://reviews.llvm.org/D134899
Louis Dionne [Wed, 28 Sep 2022 20:07:43 +0000 (16:07 -0400)]
[libc++] Include skipped headers in the --trace-includes output
By default, Clang does not include headers that are skipped due to
the include guard optimization in the --trace-includes output, which
breaks the use case that we were trying to use it for.
However, Clang does support the -fshow-skipped-includes flag, which
does exactly what we need and will result in an accurate include
graph.
As a fly-by fix, make sure that our includes don't differ between
-fexceptions and -fno-exceptions.
Differential Revision: https://reviews.llvm.org/D134829
Michał Górny [Thu, 29 Sep 2022 18:58:47 +0000 (20:58 +0200)]
[clang] [test] Use %clang_cc1 substitution consistently
Use the `%clang_cc1` substitution consistently across the test suite,
replacing inline `%clang -cc1` invocations, except for one Preprocessor
test where this is causing breakage. This is necessary to ensure that
additional parameters passed via `%clang` do not interfere with `-cc1`
that must always be passed as the first command-line argument.
Remove the additional substitution blocking `%clang_cc1` use in Driver
tests. It has been added in 2013 and was supposed to prevent tests
calling `clang -cc1` from being added to Driver. The state of the test
suite proves that it did not succeed at all.
Differential Revision: https://reviews.llvm.org/D134880
Michał Górny [Thu, 29 Sep 2022 18:58:43 +0000 (20:58 +0200)]
[llvm] [lit] Move %clang_dxc substitution from clang/test
Move the `%clang_dxc` substitution from local definition in clang/test
to lit's `llvm/config.py` module where all other driver definitions
are found. This improves consistency and makes it easier to control
global clang options.
Differential Revision: https://reviews.llvm.org/D134871
Michał Górny [Thu, 29 Sep 2022 18:58:35 +0000 (20:58 +0200)]
[clang] [Driver] More flexible rules for loading default configs
Change the default config file loading logic to be more flexible
and more readable at the same time. The new algorithm focuses on four
locations, in order:
1. <triple>-<mode>.cfg using real driver mode
2. <triple>-<mode>.cfg using executable suffix
3. <triple>.cfg + <mode>.cfg using real driver mode
4. <triple>.cfg + <mode>.cfg using executable suffix
This is meant to preserve reasonable level of compatibility with
the existing use, while introducing more flexibility and making the code
simpler. Notably:
1. In this layout, the actual target triple is normally respected,
and e.g. in `-m32` build the `x86_64-*` configs will never be used.
2. Both real driver mode (preferable) and executable suffix are
supported. This permits correctly handling calls with explicit
`--driver-mode=` while at the same time preserving compatibility
with the existing code.
3. The first two locations provide users with the ability to override
configuration per specific target+mode combinaton, while the next two
make it possible to independently specify per-target and per-mode
configuration.
4. All config file locations are applicable independently of whether
clang is started via a prefixed executable, or bare `clang`.
5. If the target is not explicitly specified and the executable prefix
does not name a valid triple, it is used instead of the actual target
triple for backwards compatibility.
This is particularly meant to address Gentoo's use case for
configuration files: to configure the default runtimes (i.e. `-rtlib=`,
`-stdlib=`) and `--gcc-install-dir=` for all the relevant drivers,
as well as to make it more convenient for users to override `-W` flags
to test compatibility with future versions of Clang easier.
Differential Revision: https://reviews.llvm.org/D134337