Craig Topper [Sat, 19 Sep 2020 01:00:06 +0000 (18:00 -0700)]
[X86] Fix copy paste mistake in @ccnp flag.
We were treating @ccp and @ccnp the same.
Craig Topper [Sat, 19 Sep 2020 04:05:50 +0000 (21:05 -0700)]
[X86] Invert the compares in inline-asm-flag-output.ll so that the setcc instruction condition matches the test name. NFC
Also add nounwind to the tests to remove cfi directives.
David Blaikie [Sat, 19 Sep 2020 02:29:37 +0000 (19:29 -0700)]
DebugInfo: Cleanup RLE dumping, using a length-constrained DataExtractor rather than carrying the end offset separately
Nico Weber [Sat, 19 Sep 2020 01:27:42 +0000 (21:27 -0400)]
[gn build] (manually) port
5495b691646
Alexander Shaposhnikov [Sat, 19 Sep 2020 01:22:48 +0000 (18:22 -0700)]
[llvm-objcopy][MachO] Clean up the interface of Object
Remove the method addLoadCommand which was used only in a single place.
NFC.
Test plan: make check-all
Nico Weber [Sat, 19 Sep 2020 01:11:56 +0000 (21:11 -0400)]
Alexander Shaposhnikov [Sat, 19 Sep 2020 01:11:22 +0000 (18:11 -0700)]
[llvm-objcopy][MachO] Add llvm-bitcode-strip driver
This diff adds llvm-bitcode-strip driver to llvm-objcopy.
In the future this will enable us to build a replacement for the tool bitcode_strip.
Test plan: make check-all
Differential revision: https://reviews.llvm.org/D87212
Eric Christopher [Sat, 19 Sep 2020 01:00:02 +0000 (18:00 -0700)]
Temporarily Revert "RegAllocFast: Rewrite and improve"
as it's breaking a few tests in the lldb test suite.
Bot: http://lab.llvm.org:8011/builders/lldb-arm-ubuntu/builds/4226/steps/test/logs/stdio
This reverts commit
c8757ff3aa7dd7a25a6343f6ef74a70c7be04325.
Alexander Shaposhnikov [Sat, 19 Sep 2020 00:50:08 +0000 (17:50 -0700)]
[llvm-install-name-tool] Validate -id value early
The code which validates the value of -id is moved into the function parseInstallNameToolOptions.
Test plan: make check-all
Differential revision: https://reviews.llvm.org/D87855
Fangrui Song [Sat, 19 Sep 2020 00:23:46 +0000 (17:23 -0700)]
[LiveDebugValues] Add `#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)` to suppress -Wunused-function
Vitaly Buka [Sat, 19 Sep 2020 00:22:35 +0000 (17:22 -0700)]
[NFC][LSan] Add REQUIRES: linux
Additional registers scaning is only implemented for x86 linux.
Amara Emerson [Fri, 18 Sep 2020 23:46:02 +0000 (16:46 -0700)]
[AArch64][GlobalISel] Add tests for pre-existing selection support for <4 x s16> arithmetic/bitwise ops.
Amara Emerson [Fri, 18 Sep 2020 23:45:12 +0000 (16:45 -0700)]
[AArch64][GlobalISel] Legalize arithmetic ops for <4 x s16>
Vitaly Buka [Sat, 19 Sep 2020 00:08:21 +0000 (17:08 -0700)]
[NFC][StackSafety] Replace auto with type
Fixes static analyzer is warning.
Vitaly Buka [Fri, 18 Sep 2020 23:46:20 +0000 (16:46 -0700)]
[NFC][Asan] Fix test broken by RegAllocFast
The test worked only because by coincidence register with pointer was
clobbered.
After D52010 value is still preserved.
Fangrui Song [Fri, 18 Sep 2020 23:16:14 +0000 (16:16 -0700)]
[SCEV] Fix an unused variable in -DLLVM_ENABLE_ASSERTIONS=off build
Amara Emerson [Fri, 18 Sep 2020 22:13:12 +0000 (15:13 -0700)]
[GlobalISel] Add lowering support for G_ABS and use for AArch64.
Differential Revision: https://reviews.llvm.org/D87952
Amy Kwan [Fri, 18 Sep 2020 22:38:08 +0000 (17:38 -0500)]
[PowerPC] Implement Move to VSR Mask builtins in LLVM/Clang
This patch implements the vec_gen[b|h|w|d|q]m function prototypes in altivec.h
in order to utilize the move to VSR with mask instructions introduced in Power10.
Differential Revision: https://reviews.llvm.org/D82725
Fangrui Song [Fri, 18 Sep 2020 23:10:39 +0000 (16:10 -0700)]
[lld][WebAssembly] Fix -Wunused-variable after D87663
Vitaly Buka [Fri, 18 Sep 2020 22:24:46 +0000 (15:24 -0700)]
[NFC][sanitizer] Don't use ::testing::internal
Snehasish Kumar [Fri, 18 Sep 2020 21:08:21 +0000 (14:08 -0700)]
[clang] Remove profile available check for fsplit-machine-functions.
Enforcing a profile available check in the driver does not work with
incremental LTO builds where the LTO backend invocation does not include
the profile flags. At this point the profiles have already been consumed
and the IR contains profile metadata. Instead we always pass through the
-fsplit-machine-functions flag on user request. The pass itself contains
a check to return early if no profile information is available.
Differential Revision: https://reviews.llvm.org/D87943
Zixu Wang [Tue, 16 Jun 2020 23:39:50 +0000 (16:39 -0700)]
[clang][module] Improve incomplete-umbrella warning
Change the warning message for -Wincomplete-umbrella to report the location of the umbrella header;
Differential Revision: https://reviews.llvm.org/D82118
Philip Reames [Fri, 18 Sep 2020 21:53:29 +0000 (14:53 -0700)]
[instcombine][x86] Converted pdep/pext with shifted mask to simple arithmetic
If the mask of a pdep or pext instruction is a shift masked (i.e. one contiguous block of ones) we need at most one and and one shift to represent the operation without the intrinsic. One all platforms I know of, this is faster than the pdep/pext.
The cost modelling for multiple contiguous blocks might be worth exploring in a follow up, but it's not relevant for my current use case. It would almost certainly be a win on AMDs where these are really really slow though.
Differential Revision: https://reviews.llvm.org/D87861
Arthur Eubanks [Fri, 18 Sep 2020 18:26:58 +0000 (11:26 -0700)]
[test][InstrProf] Fix always_inline.ll under NPM
NPM's inliner does not clean up dead functions.
Differential Revision: https://reviews.llvm.org/D87922
Eric Christopher [Fri, 18 Sep 2020 21:47:43 +0000 (14:47 -0700)]
Temporarily Revert "[clangd] Add Random Forest runtime for code completion."
as a header doesn't appear to have made it into the commit.
This reverts commit
9b6765e784b39c88cb8cdb85ab083e6c95a997ed and followup
Lei Zhang [Fri, 18 Sep 2020 21:39:53 +0000 (17:39 -0400)]
[spirv] Move device info from resource limit into target env
Vendor/device information are not resource limits. Moving to
target environment directly for better organization.
Reviewed By: mravishankar
Differential Revision: https://reviews.llvm.org/D87911
Reid Kleckner [Fri, 18 Sep 2020 20:43:13 +0000 (13:43 -0700)]
[COFF] Move per-global .drective emission from AsmPrinter to TLOFCOFF
This changes the order of output sections and the output assembly, but
is otherwise NFC.
It simplifies the TLOF interface by removing two COFF-only methods.
Vedant Kumar [Fri, 18 Sep 2020 20:43:49 +0000 (13:43 -0700)]
[llvm-cov] Allow commas in filenames passed to `-object` flag
Currently, -object takes a comma separated list of objects as an
argument, which prevents it working with path names that contain a
comma. Drop comma-separated support, which requires to set pass the
-object flag multiple times to set multiple objects.
Patch by Andrew Gallagher!
Differential Revision: https://reviews.llvm.org/D87003
Vitaly Buka [Fri, 18 Sep 2020 10:00:50 +0000 (03:00 -0700)]
[msan][asan] Add runtime flag intercept_strcmp
Can be used to disable interceptor to workaround issues of
non-instrumented code.
Reviewed By: morehouse, eugenis
Differential Revision: https://reviews.llvm.org/D87897
Joseph Huber [Wed, 16 Sep 2020 21:15:56 +0000 (17:15 -0400)]
[OpenMP] Initial Support for OpenMP Webpage Documentation
Summary:
Adding support for generated html documentation for OpenMP. Changing
Cmake files to build the documentation and adding the base templates for
future documentation to be added.
Reviewers: jdoerfert
Subscribers: aaron.ballman arphaman guansong mgorny openmp-commits sstefan1 yaxunl
Tags: #OpenMP
Differential Revision: https://reviews.llvm.org/D87797
Nico Weber [Fri, 18 Sep 2020 20:24:54 +0000 (16:24 -0400)]
CompletionModelCodegen: Remove unused import
The unused import is 3.4+, so it also breaks py2.7 compat.
But this is easy to fix :)
Nico Weber [Fri, 18 Sep 2020 20:15:24 +0000 (16:15 -0400)]
clang: Make changes in
7c8bb409f31e py2.7-compatible
Sanjay Patel [Fri, 18 Sep 2020 14:36:24 +0000 (10:36 -0400)]
[InstSimplify] add tests for constant folding fmin/fmax with undef op; NFC
Nico Weber [Fri, 18 Sep 2020 20:01:00 +0000 (16:01 -0400)]
[gn build] add file i forgot to add in
929d91a55616
Nico Weber [Fri, 18 Sep 2020 19:56:15 +0000 (15:56 -0400)]
[gn build] (manually) port
9b6765e784b3 more
Eric Christopher [Fri, 18 Sep 2020 19:33:12 +0000 (12:33 -0700)]
Temporarily Revert "[SLP] Allow reordering of vectorization trees with reused instructions."
as it's infinite looping on occasion.
This reverts commit
455ca0ebb69210046928fedffe292420a30f89ad.
Miklos Vajna [Fri, 18 Sep 2020 19:42:38 +0000 (21:42 +0200)]
[clang-format] Recognize "hxx" as a C++ header in clang-format-diff.py
And shift "proto" to the next line to avoid a too long line.
Reviewed By: MyDeveloperDay
Differential Revision: https://reviews.llvm.org/D87931
Nico Weber [Fri, 18 Sep 2020 19:34:21 +0000 (15:34 -0400)]
[gn build] Do not sync filenames containing variable references
Krzysztof Parzyszek [Fri, 18 Sep 2020 19:28:11 +0000 (14:28 -0500)]
Pre-commit test for CSEing masked loads/stores
Nico Weber [Fri, 18 Sep 2020 19:26:52 +0000 (15:26 -0400)]
[gn build] (manually) port
9b6765e784b3
Nico Weber [Fri, 18 Sep 2020 19:25:53 +0000 (15:25 -0400)]
clangd: Make ompletionModelCodegen.py tpy2.7 compatible
LLVM still supports Python 2.7, so unbreak bots that still run that.
In a separate commit so that this is easy to revert once we drop
support :)
Stella Laurenzo [Fri, 18 Sep 2020 07:21:09 +0000 (00:21 -0700)]
First pass on MLIR python context lifetime management.
* Per thread https://llvm.discourse.group/t/revisiting-ownership-and-lifetime-in-the-python-bindings/1769
* Reworks contexts so it is always possible to get back to a py::object that holds the reference count for an arbitrary MlirContext.
* Retrofits some of the base classes to automatically take a reference to the context, elimintating keep_alives.
* More needs to be done, as discussed, when moving on to the operations/blocks/regions.
Differential Revision: https://reviews.llvm.org/D87886
Peter Collingbourne [Thu, 10 Sep 2020 02:15:26 +0000 (19:15 -0700)]
scudo: Add an API for disabling memory initialization per-thread.
Here "memory initialization" refers to zero- or pattern-init on
non-MTE hardware, or (where possible to avoid) memory tagging on MTE
hardware. With shared TSD the per-thread memory initialization state
is stored in bit 0 of the TLS slot, similar to PointerIntPair in LLVM.
Differential Revision: https://reviews.llvm.org/D87739
Simon Pilgrim [Fri, 18 Sep 2020 18:51:45 +0000 (19:51 +0100)]
[X86][AVX] lowerBuildVectorAsBroadcast - improve BROADCASTM lowering on non-VLX targets
Broadcast to a ZMM type then extract the low subvector.
Arthur Eubanks [Fri, 18 Sep 2020 18:32:58 +0000 (11:32 -0700)]
[test][TSan] Fix tests under NPM
Under NPM, the TSan passes are split into a module and function pass. A
couple tests were testing for inserted module constructors, which is
only part of the module pass.
Huihui Zhang [Fri, 18 Sep 2020 18:26:28 +0000 (11:26 -0700)]
[InstCombine][SVE] Skip scalable type for InstCombiner::getFlippedStrictnessPredicateAndConstant.
We cannot iterate on scalable vector, the number of elements is unknown at compile-time.
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D87918
David Blaikie [Fri, 18 Sep 2020 18:22:37 +0000 (11:22 -0700)]
Linewrap & remove some dead typedefs from previous commit
Cleanup for
51a505340dfdfdfd9ab32c7267a74db3cdeefa56
David Blaikie [Fri, 18 Sep 2020 18:17:56 +0000 (11:17 -0700)]
DebugInfo: Simplify line table parsing to take all the units together, rather than CUs and TUs separately
James Y Knight [Thu, 17 Sep 2020 22:10:19 +0000 (18:10 -0400)]
PR47468: Fix findPHICopyInsertPoint, so that copies aren't incorrectly inserted after an INLINEASM_BR.
findPHICopyInsertPoint special cases placement in a block with a
callbr or invoke in it. In that case, we must ensure that the copy is
placed before the INLINEASM_BR or call instruction, if the register is
defined prior to that instruction, because it may jump out of the
block.
Previously, the code placed it immediately after the last def _or
use_. This is wrong, if the use is the instruction which may jump. We
could correctly place it immediately after the last def (ignoring
uses), but that is non-optimal for register pressure.
Instead, place the copy after the last def, or before the
call/inlineasm_br, whichever is later.
Differential Revision: https://reviews.llvm.org/D87865
Simon Pilgrim [Fri, 18 Sep 2020 18:11:09 +0000 (19:11 +0100)]
[X86][AVX] Add missing non AVX512VL broadcastm test coverage
Matt Arsenault [Fri, 11 Sep 2020 15:42:44 +0000 (11:42 -0400)]
CodeGen: Move split block utility to MachineBasicBlock
AMDGPU needs this in several places, so consolidate them here.
Matt Arsenault [Mon, 14 Sep 2020 16:48:12 +0000 (12:48 -0400)]
RegAllocFast: Rewrite and improve
This rewrites big parts of the fast register allocator. The basic
strategy of doing block-local allocation hasn't changed but I tweaked
several details:
Track register state on register units instead of physical
registers. This simplifies and speeds up handling of register aliases.
Process basic blocks in reverse order: Definitions are known to end
register livetimes when walking backwards (contrary when walking
forward then uses may or may not be a kill so we need heuristics).
Check register mask operands (calls) instead of conservatively
assuming everything is clobbered. Enhance heuristics to detect
killing uses: In case of a small number of defs/uses check if they are
all in the same basic block and if so the last one is a killing use.
Enhance heuristic for copy-coalescing through hinting: We check the
first k defs of a register for COPYs rather than relying on there just
being a single definition. When testing this on the full llvm
test-suite including SPEC externals I measured:
average 5.1% reduction in code size for X86, 4.9% reduction in code on
aarch64. (ranging between 0% and 20% depending on the test) 0.5%
faster compiletime (some analysis suggests the pass is slightly slower
than before, but we more than make up for it because later passes are
faster with the reduced instruction count)
Also adds a few testcases that were broken without this patch, in
particular bug 47278.
Patch mostly by Matthias Braun
Matt Arsenault [Tue, 15 Sep 2020 13:16:14 +0000 (09:16 -0400)]
Reapply "RegAllocFast: Record internal state based on register units"
The regressions this caused should be fixed when
https://reviews.llvm.org/D52010 is applied.
This reverts commit
a21387c65470417c58021f8d3194a4510bb64f46.
Zequan Wu [Thu, 17 Sep 2020 02:05:51 +0000 (19:05 -0700)]
[CodeGen] emit CG profile for COFF object file
I forgot to add emission of CG profile for COFF object file, when adding the support (https://reviews.llvm.org/D81775)
Differential Revision: https://reviews.llvm.org/D87811
Arthur Eubanks [Fri, 18 Sep 2020 17:55:28 +0000 (10:55 -0700)]
[test][HWAsan] Fix kernel-inline.ll under NPM
David Blaikie [Fri, 18 Sep 2020 17:54:02 +0000 (10:54 -0700)]
DebugInfo: Tidy up initializing multi-section contributions in DWARFContext
Raul Tambre [Fri, 18 Sep 2020 17:07:05 +0000 (20:07 +0300)]
[Sema] Handle objc_super special lookup when checking builtin compatibility
objc_super is special and needs LookupPredefedObjCSuperType() called before performing builtin type comparisons.
This fixes an error when compiling macOS headers. A test is added.
Differential Revision: https://reviews.llvm.org/D87917
Arthur Eubanks [Fri, 18 Sep 2020 17:50:17 +0000 (10:50 -0700)]
[ASan][NewPM] Fix byref-args.ll under NPM
peter klausler [Thu, 17 Sep 2020 19:19:42 +0000 (12:19 -0700)]
[flang] Rework preprocessing of stringification
Hew more closely to the C17 standard; perform macro replacement
of arguments to function-like macros unless they're being stringified
or pasted. Test with a model "assert" macro idiom that exposed
the problem.
Differential Revision: https://reviews.llvm.org/D87650
Matt Arsenault [Thu, 10 Sep 2020 19:49:09 +0000 (15:49 -0400)]
AMDGPU: Don't sometimes allow instructions before lowered si_end_cf
Since
6524a7a2b9ca072bd7f7b4355d1230e70c679d2f, this would sometimes
not emit the or to exec at the beginning of the block, where it really
has to be. If there is an instruction that defines one of the source
operands, split the block and turn the si_end_cf into a terminator.
This avoids regressions when regalloc fast is switched to inserting
reloads at the beginning of the block, instead of spills at the end of
the block.
In a future change, this should always split the block.
Amara Emerson [Fri, 18 Sep 2020 17:23:35 +0000 (10:23 -0700)]
[AArch64][GlobalISel] Make <8 x s8> of G_BUILD_VECTOR legal.
Utkarsh Saxena [Tue, 14 Jul 2020 21:12:45 +0000 (23:12 +0200)]
[clangd] Add Random Forest runtime for code completion.
Summary:
[WIP]
- Proposes a json format for representing Random Forest model.
- Proposes a way to test the generated runtime using a test model.
TODO:
- Add generated source code snippet for easier review.
- Fix unused label warning.
- Figure out required using declarations for CATEGORICAL columns from Features.json.
- Necessary Google3 internal modifications for blaze before landing.
- Add documentation for format of the model.
- Document more.
Subscribers: mgorny, ilya-biryukov, MaskRay, jkorous, arphaman, kadircet, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D83814
Sean Silva [Fri, 18 Sep 2020 01:16:41 +0000 (18:16 -0700)]
[mlir][shape] Extend shape.cstr_require with a message.
I realized when using this that one can't get very good error messages
without an additional message attribute.
Differential Revision: https://reviews.llvm.org/D87875
mydeveloperday [Fri, 18 Sep 2020 17:16:02 +0000 (18:16 +0100)]
[clang-format] NFC ensure the clang-format tests remain clang-formatted
mydeveloperday [Fri, 18 Sep 2020 17:11:33 +0000 (18:11 +0100)]
[clang-format] Add a option for the position of Java static import
Some Java style guides and IDEs group Java static imports after
non-static imports. This patch allows clang-format to control
the location of static imports.
Patch by: @bc-lee
Reviewed By: MyDeveloperDay, JakeMerdichAMD
Differential Revision: https://reviews.llvm.org/D87201
JonChesterfield [Fri, 18 Sep 2020 17:10:27 +0000 (18:10 +0100)]
[libomptarget] Disable build of amdgpu plugin as it doesn't build with rocm.
Francis Visoiu Mistrih [Thu, 17 Sep 2020 22:41:01 +0000 (15:41 -0700)]
[NFC][ScheduleDAG] Remove unused EntrySU SUnit
EntrySU doesn't seem to be used at all when building the ScheduleDAG.
Differential Revision: https://reviews.llvm.org/D87867
Jianzhou Zhao [Fri, 18 Sep 2020 16:41:18 +0000 (16:41 +0000)]
Use one more byte to silence a warning from Vistual C++
Jon Roelofs [Fri, 18 Sep 2020 15:47:37 +0000 (08:47 -0700)]
Extending Baremetal toolchain's support for the rtlib option.
Differential Revision: https://reviews.llvm.org/D87164
Patch by Manuel Carrasco!
Andy Ly [Fri, 18 Sep 2020 15:18:12 +0000 (15:18 +0000)]
[MLIR][ODS] Add constBuilderCall for TypeArrayAttr
constBuilderCall was not defined for TypeArrayAttr, resulting in tblgen not emitting the correct code when TypeArrayAttr is used with a default valued attribute.
Reviewed By: antiagainst
Differential Revision: https://reviews.llvm.org/D87907
Simon Pilgrim [Fri, 18 Sep 2020 15:35:39 +0000 (16:35 +0100)]
[X86][AVX] lowerBuildVectorAsBroadcast - improve i64 BROADCASTM lowering on 32-bit targets
We already handle the the cases where we have a 'zero extended splat' build vector (a, 0, 0, 0, a, 0, 0, 0, ...) but were missing the case where the 'a' scalar was zero-extended as well - such as i64 -> vXi64 splat cases on 32-bit targets.
Matt Morehouse [Fri, 18 Sep 2020 15:47:31 +0000 (08:47 -0700)]
[DFSan] Add strpbrk wrapper.
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D87849
ergawy [Fri, 18 Sep 2020 15:47:37 +0000 (11:47 -0400)]
[MLIR][SPIRV] Create new ctx for deserialization in roundtrips.
Roundtripping SPIR-V modules used the same MLIRContext object for both
ways of the trip. This resulted in deserialization using a context
object already containing Types constructed during serialization.
This commit rectifies that by creating a new MLIRContext during
deserialization.
Reviewed By: mravishankar, antiagainst
Differential Revision: https://reviews.llvm.org/D87692
Valentin Clement [Fri, 18 Sep 2020 15:52:02 +0000 (11:52 -0400)]
[mlir][openacc] Add missing operands for acc.data operation
Add missing operands to represent copyin with readonly modifier, copyout with zero modifier
and create with zero modifier.
Reviewed By: ftynse
Differential Revision: https://reviews.llvm.org/D87874
Valentin Clement [Fri, 18 Sep 2020 15:37:34 +0000 (11:37 -0400)]
[mlir][openacc] Support Index and AnyInteger in loop op
Following patch D87712, this patch switch AnyInteger for operands gangNum, gangStatic,
workerNum, vectoreLength and tileOperands to Index and AnyInteger.
Reviewed By: ftynse
Differential Revision: https://reviews.llvm.org/D87848
Simon Pilgrim [Fri, 18 Sep 2020 15:10:09 +0000 (16:10 +0100)]
[X86][AVX] Add missing i686 broadcastm test coverage
Simon Pilgrim [Fri, 18 Sep 2020 14:11:13 +0000 (15:11 +0100)]
[DAG] BuildVectorSDNode::getSplatValue - pull out repeated getNumOperands() calls. NFCI.
David Tenty [Thu, 30 Jul 2020 12:31:54 +0000 (08:31 -0400)]
[AIX] Enable large code model when building with clang
Adam Czachorowski [Tue, 15 Sep 2020 17:47:50 +0000 (19:47 +0200)]
[clangd] Add option for disabling AddUsing tweak on some namespaces.
For style guides forbid "using" declarations for namespaces like "std".
With this new config option, AddUsing can be selectively disabled on
those.
Differential Revision: https://reviews.llvm.org/D87775
Hanhan Wang [Fri, 18 Sep 2020 14:07:10 +0000 (07:07 -0700)]
[mlir][StandardToSPIRV] Handle vector of i1 case for lowering zexti to SPIR-V.
Reviewed By: mravishankar
Differential Revision: https://reviews.llvm.org/D87887
Sanjay Patel [Fri, 18 Sep 2020 13:53:06 +0000 (09:53 -0400)]
[InstSimplify] fix fmin/fmax miscompile for partial undef vectors (PR47567)
It would also be correct to return the variable operand in these cases,
but eliminating a variable use is probably better for optimization.
Matt Arsenault [Thu, 17 Sep 2020 21:50:42 +0000 (17:50 -0400)]
IR: Move denormal mode parsing from MachineFunction to Function
This was just inspecting the IR to begin with, and is useful to check
in some places in the IR.
Matt Arsenault [Fri, 18 Sep 2020 13:28:57 +0000 (09:28 -0400)]
emacs: Add nofree and willreturn to list of attributes
Matt Arsenault [Thu, 17 Sep 2020 16:07:59 +0000 (12:07 -0400)]
Revert "[amdgpu] Lower SGPR-to-VGPR copy in the final phase of ISel."
This reverts commit
c3492a1aa1b98c8d81b0969d52cea7681f0624c2.
I think this is the wrong strategy and wrong place to do this
transform anyway. Also reverts follow up commit
7d593d0d6905b55ca1124fca5e4d1ebb17203138.
Alexey Bataev [Thu, 17 Sep 2020 15:24:00 +0000 (11:24 -0400)]
[SLP] Allow reordering of vectorization trees with reused instructions.
If some leaves have the same instructions to be vectorized, we may
incorrectly evaluate the best order for the root node (it is built for the
vector of instructions without repeated instructions and, thus, has less
elements than the root node). In this case we just can not try to reorder
the tree + we may calculate the wrong number of nodes that requre the
same reordering.
For example, if the root node is \<a+b, a+c, a+d, f+e\>, then the leaves
are \<a, a, a, f\> and \<b, c, d, e\>. When we try to vectorize the first
leaf, it will be shrink to \<a, b\>. If instructions in this leaf should
be reordered, the best order will be \<1, 0\>. We need to extend this
order for the root node. For the root node this order should look like
\<3, 0, 1, 2\>. This patch allows extension of the orders of the nodes
with the reused instructions.
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D45263
Mirko Brkusanin [Fri, 18 Sep 2020 13:19:54 +0000 (15:19 +0200)]
[AMDGPU] Set DS alignment requirements to be more strict
Alignment requirements for ds_read/write_b96/b128 for gfx9 and onward are
now the same as for other GCN subtargets. This way we can avoid any
unintentional use of these instructions on systems that do not support dword
alignment and instead require natural alignment.
This also makes 'SH_MEM_CONFIG.alignment_mode == STRICT' the default.
Differential Revision: https://reviews.llvm.org/D87821
Sanjay Patel [Thu, 17 Sep 2020 21:20:17 +0000 (17:20 -0400)]
[InstSimplify] add another test for NaN propagation; NFC
Daniel Kiss [Wed, 16 Sep 2020 21:03:19 +0000 (23:03 +0200)]
[libunwind] Support for leaf function unwinding.
Unwinding leaf function is useful in cases when the backtrace finds a
leaf function for example when it caused a signal.
This patch also add the support for the DW_CFA_undefined because it marks
the end of the frames.
Ryan Prichard provided code for the tests.
Reviewed By: #libunwind, mstorsjo
Differential Revision: https://reviews.llvm.org/D83573
Reland with limit the test to the x86_64-linux target.
Xing GUO [Fri, 18 Sep 2020 12:20:58 +0000 (20:20 +0800)]
[DWARFYAML] Make the include_directories, file_names and opcodes fields of the line table optional.
This patch makes the include_directories, file_names and opcodes fields
of the line table optional. This helps us simplify some tests.
Reviewed By: jhenderson
Differential Revision: https://reviews.llvm.org/D87878
Xing GUO [Fri, 18 Sep 2020 12:18:48 +0000 (20:18 +0800)]
[DWARFYAML][test] Use 'CHECK-NEXT:' to make checkers stricter. NFC.
This patch makes checkers stricter so that we are able to avoid
some potential problems earlier.
Reviewed By: jhenderson, MaskRay
Differential Revision: https://reviews.llvm.org/D87876
David Greene [Mon, 13 Jan 2020 18:16:35 +0000 (12:16 -0600)]
[UpdateCCTestChecks] Include generated functions if asked
Add the --include-generated-funcs option to update_cc_test_checks.py so that any
functions created by the compiler that don't exist in the source will also be
checked.
We need to maintain the output order of generated function checks so that
CHECK-LABEL works properly. To do so, maintain a list of functions output for
each prefix in the order they are output. Use this list to output checks for
generated functions in the proper order.
Differential Revision: https://reviews.llvm.org/D83004
Max Kazantsev [Fri, 18 Sep 2020 10:38:08 +0000 (17:38 +0700)]
[Test] Missing range check removal opportunity
Florian Hahn [Fri, 18 Sep 2020 10:40:45 +0000 (11:40 +0100)]
[Polly] Update map passed to SCEVParameterReweriter.
The type of the map the SCEVParameterRewriter takes has been changed in
4635f6050b10.
Update the single use in polly to use SCEV* as type of the values.
Nicolas Vasilache [Fri, 18 Sep 2020 10:13:25 +0000 (06:13 -0400)]
[mlir][Linalg] Evolve named ops to use assembly form and support linalg on tensors.
This revision allows representing a reduction at the level of linalg on tensors for named ops. When a structured op has a reduction and returns tensor(s), new conventions are added and documented.
As an illustration, the syntax for a `linalg.matmul` writing into a buffer is:
```
linalg.matmul ins(%a, %b : memref<?x?xf32>, tensor<?x?xf32>)
outs(%c : memref<?x?xf32>)
```
, whereas the syntax for a `linalg.matmul` returning a new tensor is:
```
%d = linalg.matmul ins(%a, %b : tensor<?x?xf32>, memref<?x?xf32>)
init(%c : memref<?x?xf32>)
-> tensor<?x?xf32>
```
Other parts of linalg will be extended accordingly to allow mixed buffer/tensor semantics in the presence of reductions.
Florian Hahn [Fri, 18 Sep 2020 07:54:09 +0000 (08:54 +0100)]
Recommit "[DSE] Switch to MemorySSA-backed DSE by default."
This switches to using DSE + MemorySSA by default again, after
fixing the issues reported after the first commit.
Notable fixes
fc8200633122,
a0017c2bc258.
This reverts commit
3a59628f3cc26eb085acfc9cbdc97243ef71a6c5.
Daniel Kiss [Fri, 18 Sep 2020 09:37:54 +0000 (11:37 +0200)]
Revert "[libunwind] Support for leaf function unwinding."
This reverts commit
23bef7ee9923b1262326981960397e8cd95d6923.
Florian Hahn [Fri, 18 Sep 2020 08:50:01 +0000 (09:50 +0100)]
[SCEV] Generalize SCEVParameterRewriter to accept SCEV expression as target.
This patch extends SCEVParameterRewriter to support rewriting unknown
epxressions to arbitrary SCEV expressions. It will be used by further
patches.
Reviewed By: reames
Differential Revision: https://reviews.llvm.org/D67176
Gabriel Hjort Ã…kerlund [Fri, 18 Sep 2020 08:08:32 +0000 (10:08 +0200)]
[TableGen][GlobalISel] Fix handling of zero_reg
When generating matching tables for GlobalISel, TableGen would output
"::zero_reg" whenever encountering the zero_reg, which in turn would
result in compilation error. This patch fixes that by instead outputting
NoRegister (== 0), which is the same result that TableGen produces when
generating matching tables for ISelDAG.
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D86215
Tim Northover [Tue, 8 Sep 2020 10:08:25 +0000 (11:08 +0100)]
AArch64: make sure jump table entries can reach entire image
This turns all jump table entries into deltas within the target
function because in the small memory model all code & static data must
be in a 4GB block somewhere in memory.
When the entries were a delta between the table location and a basic
block, the 32-bit signed entries are not enough to guarantee
reachability.
https://reviews.llvm.org/D87286
Vitaly Buka [Fri, 18 Sep 2020 08:17:54 +0000 (01:17 -0700)]
[NFC][gotsan] Fix 'format' error