Amara Emerson [Tue, 16 Mar 2021 06:27:10 +0000 (23:27 -0700)]
[AArch64][GlobalISel] Fix crash on lowering <1 x half> types.
Yvan Roux [Tue, 16 Mar 2021 06:00:19 +0000 (07:00 +0100)]
[AArch64][ASAN] Disable fgets_fputs.cpp test.
This test is failing for long a time on AArch64 bots, disable it for now
to keep the bots green while investigating it.
Pushpinder Singh [Tue, 16 Mar 2021 04:54:21 +0000 (04:54 +0000)]
[OpenMP][AMDGPU] Skip backend and assemble phases for amdgcn
Remove emit-llvm-bc from addClangTargetOptions as it conflicts with -E for save-temps.
AMDGCN does not yet support linking object files so backend and assemble actions are
skipped, leaving LLVM IR as the output format.
Reviewed By: JonChesterfield, ronlieb
Differential Revision: https://reviews.llvm.org/D96769
wlei [Thu, 4 Mar 2021 04:37:19 +0000 (20:37 -0800)]
[CSSPGO][llvm-profgen] Fix getCanonicalFnName usage in llvm-profgen
Previously we didn't support to keep the unique linkage name(-funique-internal-linkage-name) in llvm-profgen. As discussed in https://reviews.llvm.org/D96932, we choose to do canonicalization for it.
Now since "selected" is set as the default parameter of getCanonicalFnName in `D96932`, we don't need to add any attribute here for the previous usage and only fix the missing usage in the pseudo probe decoding.
Differential Revision: https://reviews.llvm.org/D98226
Johannes Doerfert [Tue, 16 Mar 2021 02:52:38 +0000 (21:52 -0500)]
[OpenMP][FIX] Repair accidental replacement of _shfl_sync with _shfl
This was broken accidentally in D95752.
Reviewed By: ye-luo
Differential Revision: https://reviews.llvm.org/D98677
Johannes Doerfert [Sun, 14 Mar 2021 17:27:42 +0000 (12:27 -0500)]
[NVPTX] CUDA does provide malloc/free since compute capability 2.X
https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#dynamic-global-memory-allocation-and-operations
Reviewed By: tra
Differential Revision: https://reviews.llvm.org/D98606
Josh Berdine [Sat, 13 Mar 2021 23:29:36 +0000 (23:29 +0000)]
[OCaml][test] Fix Bindings/OCaml/executionengine.ml test
It seems that at some point it became necessary to pass `-thread` to
the ocaml compiler for this test.
Differential Revision: https://reviews.llvm.org/D98593
LLVM GN Syncbot [Tue, 16 Mar 2021 02:41:16 +0000 (02:41 +0000)]
[gn build] Port
4f198b0c27b0
Bing1 Yu [Tue, 16 Mar 2021 01:44:42 +0000 (09:44 +0800)]
[X86] Pass to transform amx intrinsics to scalar operation.
This pass runs in any situations but we skip it when it is not O0 and the
function doesn't have optnone attribute. With -O0, the def of shape to amx
intrinsics is near the amx intrinsics code. We are not able to find a
point which post-dominate all the shape and dominate all amx intrinsics.
To decouple the dependency of the shape, we transform amx intrinsics
to scalar operation, so that compiling doesn't fail. In long term, we
should improve fast register allocation to allocate amx register.
Reviewed By: pengfei
Differential Revision: https://reviews.llvm.org/D93594
David Blaikie [Tue, 16 Mar 2021 00:08:37 +0000 (17:08 -0700)]
Skip path separators to make the test portable across Win/Linux
Aart Bik [Mon, 15 Mar 2021 22:32:57 +0000 (15:32 -0700)]
[mlir][amx] Add Intel AMX dialect (architectural-specific vector dialect)
The Intel Advanced Matrix Extensions (AMX) provides a tile matrix
multiply unit (TMUL), a tile control register (TILECFG), and eight
tile registers TMM0 through TMM7 (TILEDATA). This new MLIR dialect
provides a bridge between MLIR concepts like vectors and memrefs
and the lower level LLVM IR details of AMX.
Reviewed By: nicolasvasilache
Differential Revision: https://reviews.llvm.org/D98470
Walter Erquinigo [Mon, 15 Mar 2021 23:35:49 +0000 (16:35 -0700)]
Fix
34885bffdf43920c0f011e17a65fd678100240dd
It failed https://lab.llvm.org/buildbot/#/builders/17/builds/5262 and
the fix is simply to relax a regex expression in a test.
Petr Hosek [Mon, 15 Mar 2021 23:26:18 +0000 (16:26 -0700)]
[CMake] Clean up unnecessary dependency
The LINK_COMPONENTS dependency between DebugInfoCodeView and
DebugInfoMSF is unnecessary. Breaking them would allow a more
fine-controlled distribution.
Patch By: dangyi
Differential Revision: https://reviews.llvm.org/D98465
Jon Chesterfield [Mon, 15 Mar 2021 23:17:49 +0000 (23:17 +0000)]
[libomptarget] Build amdgcn devicertl by default
[libomptarget] Build amdgcn devicertl by default
The cmake for this looks for an llvm install and does the right thing when
building as part of enable_runtimes. It will probably do the right thing
in other settings - at least, it won't try to build this with gcc.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D98658
LLVM GN Syncbot [Mon, 15 Mar 2021 23:01:19 +0000 (23:01 +0000)]
[gn build] Port
ecf6466f01c5
Lang Hames [Mon, 15 Mar 2021 21:56:29 +0000 (14:56 -0700)]
[JITLink][MachO][x86-64] Introduce generic x86-64 support.
This patch introduces generic x86-64 edge kinds, and refactors the MachO/x86-64
backend to use these edge kinds. This simplifies the implementation of the
MachO/x86-64 backend and makes it possible to write generic x86-64 passes and
utilities.
The new edge kinds are different from the original set used in the MachO/x86-64
backend. Several edge kinds that were not meaningfully distinguished in that
backend (e.g. the PCRelMinusN edges) have been merged into single edge kinds in
the new scheme (these edge kinds can be reintroduced later if we find a use for
them). At the same time, new edge kinds have been introduced to convey extra
information about the state of the graph. E.g. The Request*AndTransformTo**
edges represent GOT/TLVP relocations prior to synthesis of the GOT/TLVP
entries, and the 'Relaxable' suffix distinguishes edges that are candidates for
optimization from edges which should be left as-is (e.g. to enable runtime
redirection).
ELF/x86-64 will be refactored to use these generic edges at some point in the
future, and I anticipate a similar refactor to create a generic arm64 support
header too.
Differential Revision: https://reviews.llvm.org/D98305
Tim Keith [Mon, 15 Mar 2021 22:36:12 +0000 (15:36 -0700)]
[flang] Create intrinsics modules directory (contd.)
Use -module-dir rather than WORKING_DIRECTORY because we are potentially
creating the working directory in this custom command.
Amy Huang [Thu, 11 Mar 2021 00:36:07 +0000 (16:36 -0800)]
Emit inline implementation of __builtin__wmemchr on MSVCRT platforms.
The MSVC runtime library doesn't have a definition for wmemchr,
so provide an inline implementation.
Differential Revision: https://reviews.llvm.org/D98472
Nico Weber [Mon, 15 Mar 2021 22:20:15 +0000 (18:20 -0400)]
[gn build] merge
af2796c76d2f a bit more
The default is fine on non-Win, but on Win this needs an explicit
setting now that lit no longer has the right default.
Tim Keith [Mon, 15 Mar 2021 22:17:35 +0000 (15:17 -0700)]
[flang] Create intrinsics modules directory
A clean build fails using make because the intrinsics modules directory
doesn't exist. For some reason it works fine with ninja.
Walter Erquinigo [Mon, 15 Mar 2021 18:58:19 +0000 (11:58 -0700)]
[lldb-vscode] Handle request_evaluate's context attribute
Summary:
The request "evaluate" supports a "context" attribute, which is sent by VSCode. The attribute is defined here https://microsoft.github.io/debug-adapter-protocol/specification#Requests_Evaluate
The "clipboard" context is not yet supported by lldb-vscode, so we can forget about it for now. The 'repl' (i.e. Debug Console) and 'watch' (i.e. Watch Expression) contexts must use the expression parser in case the frame's variable path is not enough, as the user expects these expressions to never fail. On the other hand, the 'hover' expression is invoked whenever the user hovers on any keyword on the UI and the user is fine with the expression not being fully resolved, as they know that the 'repl' case is the fallback they can rely on.
Given that the 'hover' expression is invoked many many times without the user noticing it due to it being triggered by the mouse, I'm making it use only the frame's variable path functionality and not the expression parser. This should speed up tremendously the responsiveness of a debug session when the user only sets source breakpoints and inspect local variables, as the entire debug info is not needed to be parsed.
Regarding tests, I've tried to be as comprehensive as possible considering a multi-file project. Fortunately, the results from the "hover" case are enough most of the times.
Differential Revision: https://reviews.llvm.org/D98656
Peyton, Jonathan L [Mon, 15 Mar 2021 21:48:22 +0000 (16:48 -0500)]
[OpenMP] Remove unused cpu_stackoffset member
Alexander Yermolovich [Mon, 15 Mar 2021 20:46:41 +0000 (13:46 -0700)]
[DWARF] Check for AddrOffsetSectionBase to work with DWO Units.
Context: https://lists.llvm.org/pipermail/llvm-dev/2021-February/148521.html
A fix for llvm-symbolizer, and other tools like BOLT, that allows retrieving address when built with -gsplit-dwarf=single mode.
Reviewed By: dblaikie
Differential Revision: https://reviews.llvm.org/D96827
diggerlin [Mon, 15 Mar 2021 21:32:03 +0000 (17:32 -0400)]
[AIX][XCOFF] Fixed the test case which failed at aix OS because enable -mignore-xcoff-visibility by default.
Summary:
because we enable -mignore-xcoff-visibility by default when there is no -fvisibility option in the clang in AIX OS
it will cause some test case fail at aix os. in order to let the -mignore-xcoff-visibility to be disable, we need to add the -fvisibility=default for those test case.
Reviewers: hubert.reinterpretcast daltenty
Differential Revision: https://reviews.llvm.org/D98660
Artem Belevich [Fri, 5 Mar 2021 02:34:39 +0000 (18:34 -0800)]
[NVPTX] Avoid temp copy of byval kernel parameters.
Avoid making a temporary copy of byval argument if all accesses are loads and
therefore the pointer to the parameter can not escape.
This avoids excessive global memory accesses when each kernel makes its own
copy.
Differential revision: https://reviews.llvm.org/D98469
Nick Lewycky [Mon, 15 Mar 2021 21:26:39 +0000 (14:26 -0700)]
NFC: Formatting changes.
Run clang-format over these files.
Capitalize some variable names per clang-tidy's request.
Pulled out to simplify review of D98302.
peter klausler [Fri, 12 Mar 2021 21:40:31 +0000 (13:40 -0800)]
[flang] Runtime: implement INDEX intrinsic function
Implement INDEX in the runtime, reusing some infrastructure
(with generalization and renaming as needed) put into place
for its cousins SCAN and VERIFY.
I did not implement full Boyer-Moore substring searching
for the forward case, but did accelerate some advancement on
mismatches.
I (re)implemented unit testing for INDEX in the new gtest
framework, combining it with the tests that have recently
been ported to gtest for SCAN and VERIFY.
Differential Revision: https://reviews.llvm.org/D98553
Stanislav Mekhanoshin [Fri, 12 Mar 2021 19:33:19 +0000 (11:33 -0800)]
[AMDGPU] Fix copyPhysReg to not produce unalined vgpr access
RA can insert something like a sub1_sub2 COPY of a wide VGPR
tuple which results in the unaligned acces with v_pk_mov_b32
after the copy is expanded. This is regression after D97316.
Differential Revision: https://reviews.llvm.org/D98549
Florian Hahn [Mon, 15 Mar 2021 21:08:47 +0000 (21:08 +0000)]
[AnnotationRemarks] Remove unneeded Function.h include (NFC).
Nico Weber [Mon, 15 Mar 2021 21:04:50 +0000 (17:04 -0400)]
[gn build] merge
9bcf0eff99
Jonas Paulsson [Thu, 4 Mar 2021 01:52:05 +0000 (02:52 +0100)]
[SystemZ] Test for isinf and isfinite in testFPKind().
Recognize BI__builtin_isinf and BI__builtin_isfinite (and a few other opcodes
for finite) in testFPKind() and handle with TDC.
Review: Ulrich Weigand.
Differential Revision: https://reviews.llvm.org/D97901
Nico Weber [Mon, 15 Mar 2021 21:00:09 +0000 (17:00 -0400)]
[gn build] kind of merge
af2796c76d2f
Good enough for now. If we need more, we'll do the usual
platform-dependent hardcoding that in practice works for everything else
too.
Stanislav Mekhanoshin [Mon, 15 Mar 2021 20:54:33 +0000 (13:54 -0700)]
[AMDGPU] Fixed msan failure with uninitialized value
Jon Chesterfield [Mon, 15 Mar 2021 20:23:52 +0000 (20:23 +0000)]
[libomptarget][nfc][amdgcn] Use precise triple for devicertl build
Stefan Pintilie [Mon, 15 Mar 2021 17:33:49 +0000 (12:33 -0500)]
[PowerPC] Add __PCREL__ when PC Relative is enabled.
This patch adds the `__PCREL__` define when PC Relative addressing is enabled.
Reviewed By: nemanjai, #powerpc
Differential Revision: https://reviews.llvm.org/D98546
Jon Chesterfield [Mon, 15 Mar 2021 20:12:49 +0000 (20:12 +0000)]
[libomptarget][nfc] Drop unused DEVICE macro
[libomptarget][nfc] Drop unused DEVICE macro
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D98655
Jon Chesterfield [Mon, 15 Mar 2021 20:12:01 +0000 (20:12 +0000)]
[libomptarget] Build amdgpu plugin by default
[libomptarget] Build amdgpu plugin by default
This will build the amdgpu plugin if cmake is able to find the hsa
runtime library, which will be the case if rocm is installed or if
the hsa library has been installed somewhere cmake looks.
Reviewed By: tianshilei1992
Differential Revision: https://reviews.llvm.org/D98654
Kirill Bobyrev [Mon, 15 Mar 2021 20:02:24 +0000 (21:02 +0100)]
[clangd] Optionally add reflection for clangd-index-server
This was originally landed without the optional part and reverted later:
https://github.com/llvm/llvm-project/commit/
8080ea4c4b8c456c72c617587cc32f174b3105c1
Reviewed By: kadircet
Differential Revision: https://reviews.llvm.org/D98404
Markus Böck [Mon, 15 Mar 2021 20:03:27 +0000 (21:03 +0100)]
Sanjay Patel [Mon, 15 Mar 2021 20:01:52 +0000 (16:01 -0400)]
[SLP] update stale test comments; NFC
These bugs were fixed with
0a8e7ca402eb
Stanislav Mekhanoshin [Tue, 9 Feb 2021 00:36:10 +0000 (16:36 -0800)]
[AMDGPU] Use single cache policy operand
Replace individual operands GLC, SLC, and DLC with a single cache_policy
bitmask operand. This will reduce the number of operands in MIR and I hope
the amount of code. These operands are mostly 0 anyway.
Additional advantage that parser will accept these flags in any order unlike
now.
Differential Revision: https://reviews.llvm.org/D96469
Markus Böck [Mon, 15 Mar 2021 19:56:08 +0000 (20:56 +0100)]
[test] Add ability to get error messages from CMake for errc substitution
Visual Studios implementation of the C++ Standard Library does not use strerror to produce a message for std::error_code unlike other standard libraries such as libstdc++ or libc++ that might be used.
This patch adds a cmake script that through running a C++ program gets the error messages for the POSIX error codes and passes them onto lit through an optional config parameter.
If the config parameter is not set, or getting the messages failed, due to say a cross compiling configuration without an emulator, it will fall back to using pythons strerror functions.
Differential Revision: https://reviews.llvm.org/D98278
Jon Chesterfield [Mon, 15 Mar 2021 19:50:21 +0000 (19:50 +0000)]
[libomptarget] Fix devicertl build
[libomptarget] Fix devicertl build
The target specific functions in target_interface are extern C, but the
implementations for nvptx were mostly C++ mangling. That worked out as
a quirk of DEVICE macro expanding to nothing, except for shuffle.h which
only forward declared the functions with C++ linkage.
Also implements GetWarpSize, as used by shuffle, and includes target_interface
in nvptx target_impl.cu to help catch future divergence between interface and
implementation.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D98651
Michael Kruse [Mon, 15 Mar 2021 19:27:48 +0000 (14:27 -0500)]
[Polly] Fix deprecation warning. NFC.
IRBuilder::CreateLoad without type parameter was deprecated in
6312c538
to prepare for opaque pointers.
Wenlei He [Sat, 13 Mar 2021 21:55:28 +0000 (13:55 -0800)]
[CSSPGO] Load context profile for external functions in PreLink and populate ThinLTO import list
For ThinLTO's prelink compilation, we need to put external inline candidates into an import list attached to function's entry count metadata. This enables ThinLink to treat such cross module callee as hot in summary index, and later helps postlink to import them for profile guided cross module inlining.
For AutoFDO, the import list is retrieved by traversing the nested inlinee functions. For CSSPGO, since profile is flatterned, a few things need to happen for it to work:
- When loading input profile in extended binary format, we need to load all child context profile whose parent is in current module, so context trie for current module includes potential cross module inlinee.
- In order to make the above happen, we need to know whether input profile is CSSPGO profile before start reading function profile, hence a flag for profile summary section is added.
- When searching for cross module inline candidate, we need to walk through the context trie instead of nested inlinee profile (callsite sample of AutoFDO profile).
- Now that we have more accurate counts with CSSPGO, we swtiched to use entry count instead of total count to decided if an external callee is potentially beneficial to inline. This make it consistent with how we determine whether call tagert is potential inline candidate.
Differential Revision: https://reviews.llvm.org/D98590
Jianzhou Zhao [Mon, 15 Mar 2021 19:10:34 +0000 (19:10 +0000)]
[dfsan] Updated check_custom_wrappers.sh to dedup function names
The origin wrappers added by https://reviews.llvm.org/D98359 reuse
those __dfsw_ functions.
Fangrui Song [Mon, 15 Mar 2021 19:05:33 +0000 (12:05 -0700)]
Change void getNoop(MCInst &NopInst) to MCInst getNop()
Prefer (self-documenting) return values to output parameters (which are
liable to be used).
While here, rename Noop to Nop which is more widely used and improves
consistency with hasEmitNops/setEmitNops/emitNop/etc.
Jez Ng [Mon, 15 Mar 2021 18:54:18 +0000 (14:54 -0400)]
[lld-macho] Place LC_FUNCTION_STARTS data at the right position
This pleases the codesign
(Otherwise it complains about "function starts data out of place")
Reviewed By: #lld-macho, smeenai
Differential Revision: https://reviews.llvm.org/D98648
Jianzhou Zhao [Mon, 15 Mar 2021 18:54:16 +0000 (18:54 +0000)]
[dfsan] Do not check dfsan_get_origin by check_custom_wrappers.sh
It is implemented like dfsan_get_label, and does not any code
in dfsan_custome.cpp.
Craig Topper [Mon, 15 Mar 2021 18:44:59 +0000 (11:44 -0700)]
[RISCV] Add RISCVISD::BR_CC similar to RISCVISD::SELECT_CC.
This allows me to introduce similar combines for branches as
we have recently added for SELECT_CC. Some of them are less
useful for standalone setccs and only help branch instructions.
By having a BR_CC node its easier to only affect branches.
I'm using CondCodeSDNode to make isel patterns easier to
write so we can refer to the codes by name. SELECT_CC uses a
constant instead.
I've translated the condition code just like SELECT_CC so
we need less patterns for the swapped conditions. This
includes special cases for X < 1 and X > -1 that get translated
to blez and bgez by using a 0 constant.
computeKnownBitsForTargetNode support for SELECT_CC is added
to allow MaskedValueIsZero to work for cases where the true
and false values of the SELECT_CC are setccs and the
result of the SELECT_CC is used by a BR_CC. This was needed
to avoid regressions in some of the overflow tests.
Reviewed By: luismarques
Differential Revision: https://reviews.llvm.org/D98159
Jon Chesterfield [Mon, 15 Mar 2021 18:50:09 +0000 (18:50 +0000)]
[libomptarget] Drop assert.h, use freestanding for amdgcn devicertl
[libomptarget] Drop assert.h, use freestanding for amdgcn devicertl
Promotes the runtime assert to a link time error for the unimplemented
fallback functions. Enables amdgcn to build with only clang provided
headers, which makes it less likely to break other builds when enabled.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D98649
Philipp Tomsich [Mon, 15 Mar 2021 18:31:05 +0000 (11:31 -0700)]
[RISCV] Add isel-patterns to optimize (a < 1) into blez (a <= 0)
The following code-sequence showed up in a testcase (isolated from
SPEC2017) for if-conversion and vectorization when searching for the
maximum in an array:
addi a2, zero, 1
blt a1, a2, .LBB0_5
which can be expressed as `bge zero,a1,.LBB0_5`/`blez a1,/LBB0_5`.
More generally, we want to express (a < 1) as (a <= 0).
This adds the required isel-pattern and updates the testcases.
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D98449
Michael Kruse [Mon, 15 Mar 2021 17:13:21 +0000 (12:13 -0500)]
[Polly][Optimizer] Apply user-directed unrolling.
Make Polly look for unrolling metadata (https://llvm.org/docs/TransformMetadata.html#loop-unrolling) that is usually only interpreted by the LoopUnroll pass and apply it to the SCoP's schedule.
While not that useful by itself (there already is an unroll pass), it introduces mechanism to apply arbitrary loop transformation directives in arbitrary order to the schedule. Transformations are applied until no more directives are found. Since ISL's rescheduling would discard the manual transformations and it is assumed that when the user specifies the sequence of transformations, they do not want any other transformations to apply. Applying user-directed transformations can be controlled using the `-polly-pragma-based-opts` switch and is enabled by default.
This does not influence the SCoP detection heuristic. As a consequence, loop that do not fulfill SCoP requirements or the initial profitability heuristic will be ignored. `-polly-process-unprofitable` can be used to disable the latter.
Other than manually editing the IR, there is currently no way for the user to add loop transformations in an order other than the order in the default pipeline, or transformations other than the one supported by clang's LoopHint. See the `unroll_double.ll` test as example that clang currently is unable to emit. My own extension of `#pragma clang loop` allowing an arbitrary order and additional transformations is available here: https://github.com/meinersbur/llvm-project/tree/pragma-clang-loop. An effort to upstream this functionality as `#pragma clang transform` (because `#pragma clang loop` has an implicit transformation order defined by the loop pipeline) is D69088.
Additional transformations from my downstream pragma-clang-loop branch are tiling, interchange, reversal, unroll-and-jam, thread-parallelization and array packing. Unroll was chosen because it uses already-defined metadata and does not require correctness checks.
Reviewed By: sebastiankreutzer
Differential Revision: https://reviews.llvm.org/D97977
Stelios Ioannou [Tue, 9 Mar 2021 14:56:49 +0000 (14:56 +0000)]
[AArch64] Implement __rndr, __rndrrs intrinsics
This patch implements the __rndr and __rndrrs intrinsics to provide access to the random
number instructions introduced in Armv8.5-A. They are only defined for the AArch64
execution state and are available when __ARM_FEATURE_RNG is defined.
These intrinsics store the random number in their pointer argument and return a status
code if the generation succeeded. The difference between __rndr __rndrrs, is that the latter
intrinsic reseeds the random number generator.
The instructions write the NZCV flags indicating the success of the operation that we can
then read with a CSET.
[1] https://developer.arm.com/docs/101028/latest/data-processing-intrinsics
[2] https://bugs.llvm.org/show_bug.cgi?id=47838
Differential Revision: https://reviews.llvm.org/D98264
Change-Id: I8f92e7bf5b450e5da3e59943b53482edf0df6efc
Alex Zinenko [Mon, 15 Mar 2021 17:35:40 +0000 (18:35 +0100)]
[mlir] fix SPIR-V CPU and Vulkan runners after
e2310704d890ad252aeb1ca28b4b84d29514b1d1
The commit in question changed the syntax but did not update the runner
tests. This also required registering the MemRef dialect for custom
parser to work correctly.
serge-sans-paille [Wed, 3 Mar 2021 08:58:31 +0000 (09:58 +0100)]
Allow __ieee128 as an alias to __float128 on ppc
This matches gcc behavior.
Differential Revision: https://reviews.llvm.org/D97846
serge-sans-paille [Mon, 15 Mar 2021 17:18:55 +0000 (18:18 +0100)]
[NFC] Use higher level constructs to check for whitespace/newlines in the lexer
It turns out that according to valgrind and perf, it's also slightly faster.
Differential Revision: https://reviews.llvm.org/D98637
Luke Drummond [Wed, 10 Mar 2021 18:14:42 +0000 (18:14 +0000)]
[OpenCL] Respect calling convention for builtin
`__translate_sampler_initializer` has a calling convention of
`spir_func`, but clang generated calls to it using the default CC.
Instruction Combining was lowering these mismatching calling conventions
to `store i1* undef` which itself was subsequently lowered to a trap
instruction by simplifyCFG resulting in runtime `SIGILL`
There are arguably two bugs here: but whether there's any wisdom in
converting an obviously invalid call into a runtime crash over aborting
with a sensible error message will require further discussion. So for
now it's enough to set the right calling convention on the runtime
helper.
Reviewed By: svenh, bader
Differential Revision: https://reviews.llvm.org/D98411
Andrzej Warzynski [Mon, 15 Mar 2021 17:24:35 +0000 (17:24 +0000)]
[flang][docs] Fix the time for the new Flang driver call
Martin Storsjö [Fri, 26 Feb 2021 13:15:36 +0000 (15:15 +0200)]
[libcxx] [test] Fix the temp_directory_path test for windows
Check a different set of env vars, don't check the exact value
of the fallback path. (GetTempPath falls back to returning the Windows
folder if nothing better is available in env vars.)
The test still fails one check on windows (due to relying on perms::none),
which will be addressed separately.
Differential Revision: https://reviews.llvm.org/D98139
Juneyoung Lee [Tue, 16 Mar 2021 00:20:49 +0000 (09:20 +0900)]
[AssumeBundles] Add nonnull/align to op bundle if noundef exists
This is a patch to add nonnull and align to assume's operand bundle
only if noundef exists.
Since nonnull and align in fn attr have poison semantics, they should be
paired with noundef or noundef-implying attributes to be immediate UB.
Reviewed By: jdoerfert, Tyker
Differential Revision: https://reviews.llvm.org/D98228
Fraser Cormack [Fri, 12 Mar 2021 11:22:29 +0000 (11:22 +0000)]
[CodeGen] Fix issues with scalable-vector INSERT/EXTRACT_SUBVECTORs
This patch addresses a few issues when dealing with scalable-vector
INSERT_SUBVECTOR and EXTRACT_SUBVECTOR nodes.
When legalizing in DAGTypeLegalizer::SplitVecRes_INSERT_SUBVECTOR, we
store the low and high halves to the stack separately. The offset for
the high half was calculated incorrectly.
Additionally, we can optimize this process when we can detect that the
subvector is contained entirely within the low/high split vector type.
While this optimization is valid on scalable vectors, when performing
the 'high' optimization, the subvector must also be a scalable vector.
Note that the 'low' optimization is still conservative: it may be
possible to insert v2i32 into the low half of a split nxv1i32/nxv1i32,
but we can't guarantee it. It is always possible to insert v2i32 into
nxv2i32 or v2i32 into nxv4i32+2 as we know vscale is at least 1.
Lastly, in SelectionDAG::isSplatValue, we early-exit on the extracted subvector value
type being a scalable vector, forgetting that we can also extract a
fixed-length vector from a scalable one.
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D98495
Kevin Zhou [Mon, 15 Mar 2021 13:15:02 +0000 (08:15 -0500)]
[Polly] Refactoring astScheduleDimIsParallel to take the C++ wrapper object. NFC
Polly currently needs to be slowly refactor to use the C++ wrapper objects to handle the reference counters automatically.
I took the function of astScheduleDimIsParallel and refactored it so that it uses the C++ wrapper function as much as possible.
There are some problems with the IsParallel since it expects the C objects, so the C++ wrapper functions must be .release() and .get() first before they are able to be used with IsParallel.
When checking the ReductionDependencies Parallelism with the Build's Schedule, I opted to keep the union map as a C object rather than a C++ object. Eventually, changes will need to be made to IsParallel to refactor it to the C++ wrappers. When this is done, this function will also need to be slightly refactored to not use the C object.
Reviewed By: Meinersbur
Differential Revision: https://reviews.llvm.org/D98455
Jon Chesterfield [Mon, 15 Mar 2021 16:54:56 +0000 (16:54 +0000)]
[libomptarget][amdgcn] Drop use of inttypes.h, moving closer to freestanding
[libomptarget][amdgcn] Drop use of inttypes.h, moving closer to freestanding
The glibc headers are a periodic source of problems compiling the devicertl.
This patch resolves the following error run into while building llvm on a slightly
different linux system.
```
In file included from .../lib/clang/13.0.0/include/inttypes.h:21:
In file included from /usr/include/inttypes.h:25:
/usr/include/features.h:461:12: fatal error: 'sys/cdefs.h' file not found
# include <sys/cdefs.h>
^~~~~~~~~~~~~
```
As a second patch, removing assert.h from shuffle will let amdgcn build as
-ffreestanding, at which point only the headers that clang itself provides are
used and interactions with the host glibc are eliminated. Doing the same for
nvptx is complicated by printf handling but also seems worthwhile.
Reviewed By: tianshilei1992
Differential Revision: https://reviews.llvm.org/D98565
Martin Storsjö [Thu, 25 Feb 2021 10:43:36 +0000 (12:43 +0200)]
[libcxx] [test] Fix the get_temp_file_name() function for mingw
Add the missing includes for getting the defines and functions used
in the mingw version of get_temp_file_name().
This fixes 31 tests when built in a mingw configuration.
Also remove a redundant ifdef; _WIN32 is defined in mingw targets too.
Differential Revision: https://reviews.llvm.org/D97456
Martin Storsjö [Thu, 25 Feb 2021 21:06:18 +0000 (23:06 +0200)]
[libcxx] [test] Disable some allocation checks in class.path tests on windows
On windows, the path internal representation is wchar_t, and
input/output often goes through utf8 inbetween, which causes extra
allocations.
MS STL also fails a number of strict allocation checks, so this
shouldn't be a standards compliance issue.
Differential Revision: https://reviews.llvm.org/D98398
Nico Weber [Mon, 15 Mar 2021 16:50:45 +0000 (12:50 -0400)]
[gn build] (semi-manually) port
b136a74efc54
Christopher Tetreault [Mon, 15 Mar 2021 16:33:31 +0000 (09:33 -0700)]
[CMake] Require python 3.6 if enabling LLVM test targets
The lit test suite uses python 3.6 features. Rather than a strange
python syntax error upon running the lit tests, we will require the
correct version in CMake.
Reviewed By: serge-sans-paille, yln
Differential Revision: https://reviews.llvm.org/D95635
Craig Topper [Mon, 15 Mar 2021 16:20:26 +0000 (09:20 -0700)]
[RISCV] Improve legalization of i32 UADDO/USUBO on RV64.
The default legalization uses zero extends that require pair of shifts
on RISCV. Instead we can take advantage of the fact that unsigned
compares work equally well on sign extended inputs. This allows
us to use addw/subw and sext.w.
Reviewed By: luismarques
Differential Revision: https://reviews.llvm.org/D98233
Simon Pilgrim [Mon, 15 Mar 2021 16:15:05 +0000 (16:15 +0000)]
[X86][SSE] isHorizontalBinOp - ensure we clear any unused source operands to improve HADD/SUB matching
Our shuffle matching for HADD/SUB patterns wasn't clearing repeated ops in 'fake unary' style shuffle masks (unpack(x,x) etc.), preventing matching of add(fakeunary(),fakeunary()) style patterns.
Alex Zinenko [Mon, 15 Mar 2021 16:19:37 +0000 (17:19 +0100)]
[mlir] make memref.cast implement ViewLikeOpInterface
This was seemingly dropped in
e2310704d890ad252aeb1ca28b4b84d29514b1d1,
potentially due to a misrebase. The absence of this trait makes aliasing
analysis incorrect, leading to, e.g., buffer deallocation pass inserting
deallocations too early.
Jianzhou Zhao [Wed, 10 Mar 2021 06:31:56 +0000 (06:31 +0000)]
[dfsan] Add origin ABI wrappers for thread/signal/fork
This is a part of https://reviews.llvm.org/D95835.
See https://github.com/llvm/llvm-project/commit/
bb91e02efd00eda04296069a83228c8d9db105b7 about the similar issue of fork in MSan's origin tracking.
Reviewed By: morehouse
Differential Revision: https://reviews.llvm.org/D98359
Zahira Ammarguellat [Mon, 15 Mar 2021 16:15:34 +0000 (12:15 -0400)]
[NFC] Fix "unused parameter" error revealed in the Linux self-build.
Melanie Blower [Mon, 15 Mar 2021 14:06:33 +0000 (10:06 -0400)]
[clang][patch] Solve PR49479, File scope fp pragma should propagate to functions nested in struct, and initialization expressions
Previously, the CurFPFeatures state was set to command line settings before
semantic analysis of the nested member functions and initialization
expressions, that's not correct, it should use the pragma state which
is in effect at the lexical position.
Reviewed By: Erich Keane, Aaron Ballman
Differential Revision: https://reviews.llvm.org/D98211
Sanjay Patel [Mon, 15 Mar 2021 15:44:53 +0000 (11:44 -0400)]
[InstSimplify] ctlz({signbit} >>u x) --> x
The motivating pattern was handled in
0a2d69480d ,
but we should have this for symmetry.
But this really highlights that we could generalize for
any shifted constant if we match this in instcombine.
https://alive2.llvm.org/ce/z/MrmVNt
Sanjay Patel [Mon, 15 Mar 2021 15:20:27 +0000 (11:20 -0400)]
[InstSimplify] add tests for ctlz of shifted constant; NFC
Edward Jones [Thu, 12 Nov 2020 16:33:03 +0000 (16:33 +0000)]
[RISCV][compiler-rt] Add support for save-restore
This adds the compiler-rt entry points required by the
-msave-restore option.
Differential Revision: https://reviews.llvm.org/D91717
Thomas Preud'homme [Wed, 9 Dec 2020 22:18:32 +0000 (22:18 +0000)]
Stop traping on sNaN in __builtin_isinf
__builtin_isinf currently generates a floating-point compare operation
which triggers a trap when faced with a signaling NaN in StrictFP mode.
This commit uses integer operations instead to not generate any trap in
such a case.
Reviewed By: mibintc
Differential Revision: https://reviews.llvm.org/D97125
Martin Storsjö [Sun, 21 Feb 2021 23:20:28 +0000 (01:20 +0200)]
[libcxx] [docs] Update docs about how to build for Windows
Refresh the existing paragraphs on building in MSVC configurations,
add a sample of one working configuration for MinGW, and add more
details on what's necessary to run the tests these days.
Differential Revision: https://reviews.llvm.org/D97166
LLVM GN Syncbot [Mon, 15 Mar 2021 15:24:41 +0000 (15:24 +0000)]
[gn build] Port
13e49dcee48f
Jon Chesterfield [Mon, 15 Mar 2021 15:24:00 +0000 (15:24 +0000)]
[amdgpu] Implement lower function LDS pass
[amdgpu] Implement lower function LDS pass
Local variables are allocated at kernel launch. This pass collects global
variables that are used from non-kernel functions, moves them into a new struct
type, and allocates an instance of that type in every kernel. Uses are then
replaced with a constantexpr offset.
Prior to this pass, accesses from a function are compiled to trap. With this
pass, most such accesses are removed before reaching codegen. The trap logic
is left unchanged by this pass. It is still reachable for the cases this pass
misses, notably the extern shared construct from hip and variables marked
constant which survive the optimizer.
This is of interest to the openmp project because the deviceRTL runtime library
uses cuda shared variables from functions that cannot be inlined. Trunk llvm
therefore cannot compile some openmp kernels for amdgpu. In addition to the
unit tests attached, this patch applied to ROCm llvm with fixed-abi enabled
and the function pointer hashing scheme deleted passes the openmp suite.
This lowering will use more LDS than strictly necessary. It is intended to be
a functionally correct fallback for cases that are difficult to target from
future optimisation passes.
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D94648
Kostya Kortchinsky [Thu, 11 Mar 2021 22:42:30 +0000 (14:42 -0800)]
[scudo][standalone] Add shared library to makefile
Since we are looking to remove the old Scudo, we have to have a .so for
parity purposes as some platforms use it.
I tested this on Fuchsia & Linux, not on Android though.
Differential Revision: https://reviews.llvm.org/D98456
Tim Keith [Mon, 15 Mar 2021 15:02:58 +0000 (08:02 -0700)]
[flang] Build intrinsic .mod files in include/flang
The build was putting .mod files for intrinsic modules in
tools/flang/include/flang but the install puts them in include/flang,
as does the out-of-tree build. This confused things for the driver.
This change makes the build consistent with the install and simplifies
the flang script accordingly.
Also, clean up the cmake commands for building the .mod files.
Differential Revision: https://reviews.llvm.org/D98522
Simon Pilgrim [Mon, 15 Mar 2021 14:50:26 +0000 (14:50 +0000)]
[X86][SSE] canonicalizeShuffleWithBinOps - handle target shuffles.
Fold SHUFFLE(BINOP(SHUFFLE(X),SHUFFLE(Y))) -> BINOP(SHUFFLE'(X),SHUFFLE'(Y)) style patterns as well as the existing shuffles of constants.
Vy Nguyen [Sat, 13 Mar 2021 02:00:11 +0000 (21:00 -0500)]
[asan] Fixed test failing on windows due to different printf behaviour.
%p reported prints upper case hex chars on Windows.
The fix is to switch to using %#lx
Differential Revision: https://reviews.llvm.org/D98570
David Green [Mon, 15 Mar 2021 14:56:20 +0000 (14:56 +0000)]
[AArch64] Zero extended extract_vector_elt pattern
This adds a pattern for i64 zext_inreg(i32 extract_vector_elt X),
producing a single UMOVvi16 instruction that is already expected to
clear the top bits. The exact pattern that this matches is
and(anyext(vector_extract X, lane), 0xff), similar to the sext patterns
higher up in the same file.
Differential Revision: https://reviews.llvm.org/D98599
Dmitry Polukhin [Fri, 26 Feb 2021 17:07:20 +0000 (09:07 -0800)]
[clang-tidy] Enable modernize-concat-nested-namespaces also on headers
For some reason the initial implementation of the check had an explicit check
for the main file to avoid being applied in headers. This diff removes this
check and add a test for the check on a header.
Similar approach was proposed in D61989 but review there got stuck.
Test Plan: added new test case
Differential Revision: https://reviews.llvm.org/D97563
Nathan James [Mon, 15 Mar 2021 14:20:47 +0000 (14:20 +0000)]
[clang-tidy] Fix readability-identifer-naming duplicating prefix or suffix for replacements.
If a identifier has a correct prefix/suffix but a bad case, the fix won't strip them when computing the correct case, leading to duplication when the are added back.
Reviewed By: aaron.ballman
Differential Revision: https://reviews.llvm.org/D98521
Nathan James [Mon, 15 Mar 2021 14:16:51 +0000 (14:16 +0000)]
[ASTMatchers] Don't forward matchers in MapAnyOf
Forwarding these means that if an r-value reference is passed, the matcher will be moved. However it appears this happens for each mapped node matcher, resulting in use-after-move issues.
Reviewed By: steveire
Differential Revision: https://reviews.llvm.org/D98497
Jan Svoboda [Mon, 15 Mar 2021 14:07:53 +0000 (15:07 +0100)]
[clang][lex] Speculative fix for buffer overrun on raw string parse
This attempts to fix a (non-deterministic) buffer overrun when parsing raw string literals during modular build.
Similar fix to
4e5b5c36f47c9a406ea7f6b4f89fae477693973a.
Reviewed By: beccadax
Differential Revision: https://reviews.llvm.org/D94950
Amy Kwan [Mon, 15 Mar 2021 06:07:42 +0000 (01:07 -0500)]
[NFC][PowerPC] Add additional load/store test cases
This patch adds additional load/store test cases involving scalars, vectors,
and PC-Rel in preparation for the refactored load and store implementation
introduced in D93370.
Differential Revision: https://reviews.llvm.org/D97391
Alex Zinenko [Mon, 15 Mar 2021 13:40:01 +0000 (14:40 +0100)]
[mlir] fix integration tests post
e2310704d890ad252aeb1ca28b4b84d29514b1d1
The commit in question moved some ops across dialects but did not update
some of the target-specific integration tests that use these ops,
presumably because the corresponding target hardware was not available.
Fix these tests.
Wael Yehia [Mon, 15 Mar 2021 13:34:20 +0000 (09:34 -0400)]
[PATCH] fix location of test case
from D97507.
Muiez Ahmed [Mon, 15 Mar 2021 13:23:36 +0000 (09:23 -0400)]
[SystemZ][z/OS] Define _LIBCPP_ELAST
The aim is to define _LIBCPP_ELAST for z/OS libc++ since strerror/strerror_r can't handle out-of-range errno values.
Differential Revision: https://reviews.llvm.org/D98541
Alex Zinenko [Mon, 15 Mar 2021 13:06:25 +0000 (14:06 +0100)]
[mlir] enable Python bindings for the MemRef dialect
A previous commit moved multiple ops from Standard to MemRef dialect.
Some of these ops are exercised in Python bindings. Enable bindings for
the newly created MemRef dialect and update a test accordingly.
Nathan James [Mon, 15 Mar 2021 13:07:39 +0000 (13:07 +0000)]
[ASTMatchers][Dynamic] Add missing matchers from Registry
Add the `fixedPointLiteral`, `hasAnyBody` and `templateArgumentLoc` to the dynamic matcher registry.
Reviewed By: steveire
Differential Revision: https://reviews.llvm.org/D98556
Nathan James [Mon, 15 Mar 2021 13:06:48 +0000 (13:06 +0000)]
[ASTMatchers] Fix documentation for hasAnyBody matcher
Looks like a oversight when the matcher was added.
Reviewed By: steveire
Differential Revision: https://reviews.llvm.org/D98583
Luís Marques [Mon, 15 Mar 2021 13:01:26 +0000 (13:01 +0000)]
Revert "[Sanitizer][RISCV][AArch64][Android] Adjust allocator tests"
This reverts commit
a00347b56e36b30b26f9e54e947b965acf2b7eec due to a
test failure on the `sanitizer-x86_64-linux-android` buildbot.
Anton Afanasyev [Mon, 15 Mar 2021 12:52:07 +0000 (15:52 +0300)]
[SLP][Test] Precommit test for PR40522
Nico Weber [Mon, 15 Mar 2021 12:44:05 +0000 (08:44 -0400)]
[ast-dump-tool] fix regression if --empty-implementation but --json-input-path is not
Looks like this broke in one of the relands of
https://reviews.llvm.org/D93164