Raphael Isemann [Wed, 4 Mar 2020 20:57:05 +0000 (12:57 -0800)]
[lldb][NFC] Use C++11 for loop in ClangASTSource::FindExternalLexicalDecls
Zola Bridges [Wed, 4 Mar 2020 20:48:52 +0000 (21:48 +0100)]
[x86][SLH] Rm liveness check from data invariance check
SLH had two functions named isDataInvariant and isDataInvariantLoad that
checked whether the passed instruction was data invariant. For some instructions,
if the EFLAGS were dead then they were considered data invariant, otherwise
they were not considered data invariant.
In this patch, I extracted that EFLAGS liveness check and made it
explicit at every call to isDataInvariant and isDataInvariantLoad.
This makes the isDataInvariant function behave more generally
and preserves the liveness check behavior that SLH would like to have.
Tested via llvm-lit llvm/test/CodeGen/X86/speculative-load-hardening*
This is the first step in making these two data invariance checks
available for non-SLH passes. The second step is to move the passes from
SLH to X86InstrInfo.cpp. I'll follow up with a patch that does that.
Differential Revision: https://reviews.llvm.org/D70283
Eric Fiselier [Wed, 4 Mar 2020 20:48:33 +0000 (15:48 -0500)]
[libc++] Mark another test as flaky
Lang Hames [Wed, 4 Mar 2020 20:36:50 +0000 (12:36 -0800)]
[ORC] Add some debugging output for initializers.
This output can be useful in tracking down initialization failures in the JIT.
Fangrui Song [Wed, 4 Mar 2020 03:49:14 +0000 (19:49 -0800)]
[llvm-readelf] Make --all output order closer to GNU readelf
https://bugs.llvm.org/show_bug.cgi?id=43403
The new order makes it easy to compare the two tools' --all.
Reviewed By: grimar, rupprecht
Differential Revision: https://reviews.llvm.org/D75592
Lei Zhang [Wed, 4 Mar 2020 20:12:33 +0000 (15:12 -0500)]
[mlir][spirv] Support integer signedness
This commit updates SPIR-V dialect to support integer signedness
by relaxing various checks for signless to just normal integers.
The hack for spv.Bitcast can now be removed.
Differential Revision: https://reviews.llvm.org/D75611
Fangrui Song [Wed, 4 Mar 2020 00:59:33 +0000 (16:59 -0800)]
[llvm-objdump] --syms: print st_size as "%016" PRIx64 instead of "%08" PRIx64 for 64-bit objects
This is GNU objdump's behavior and it is reasonable to match.
Reviewed By: grimar
Differential Revision: https://reviews.llvm.org/D75588
Lei Zhang [Wed, 4 Mar 2020 19:41:27 +0000 (14:41 -0500)]
[mlir][ods] Improve integer signedness modelling
A previous commit added support for integer signedness in C++
IntegerType. This change introduces ODS definitions for
integer types and integer (element) attributes w.r.t. signedness.
This commit also updates various existing definitions' descriptions
to mention signless where suitable to make it more clear.
Positive and non-negative integer attributes are removed to avoid
the explosion of subclasses. Instead, one should use more atmoic
constraints together with Confined to model that. For example,
`Confined<..., [IntPositive]>`.
Differential Revision: https://reviews.llvm.org/D75610
Alexey Bataev [Wed, 4 Mar 2020 19:37:51 +0000 (14:37 -0500)]
[OPENMP50]Codegen for 'depend' clause in depobj directive.
Added codegen for 'depend' clause in depobj directive. The depend clause
is emitted as kmp_depend_info <deps>[<number_of_items_in_clause> + 1]. The
first element in this array is reserved for storing the number of
elements in this array: <deps>[0].base_addr =
<number_of_items_in_clause>;
This extra element is required to implement 'update' and 'destroy'
clauses. It is required to know the size of array to destroy it
correctly and to update depency kind.
Martijn Vels [Wed, 4 Mar 2020 19:51:21 +0000 (14:51 -0500)]
Revert "Partially inline basic_string copy constructor in UNSTABLE"
This reverts commit
8cf76e913b867a98a9843aa1b3d782632ed5d930.
We are investigating why this causes compilation issues under -O3
Sam Clegg [Wed, 4 Mar 2020 19:29:45 +0000 (11:29 -0800)]
[lld][WebAssembly] Add support for --rsp-quoting
This also changes to default style to match the host.
Reviewed By: ruiu
Differential Revision: https://reviews.llvm.org/D75577
Louis Dionne [Wed, 4 Mar 2020 19:38:51 +0000 (14:38 -0500)]
[libc++] Avoid deprecation warning in the implementation of __has_rebind
Wei Mi [Thu, 13 Feb 2020 17:23:27 +0000 (09:23 -0800)]
Generate Callee Saved Register (CSR) related cfi directives like .cfi_restore.
https://reviews.llvm.org/D42848 only handled CFA related cfi directives but
didn't handle CSR related cfi. The patch adds the CSR part. Basically it reuses
the framework created in D42848. For each basicblock, the patch tracks which
CSR set have been saved at its CFG predecessors's exits, and compare the CSR
set with the set at its previous basicblock's exit (The previous block is the
block laid before the current block). If the saved CSR set at its previous
basicblock's exit is larger, .cfi_restore will be inserted.
The patch also generates proper .cfi_restore in epilogue to make sure the
saved CSR set is consistent for the incoming edges of each block.
Differential Revision: https://reviews.llvm.org/D74303
Guozhi Wei [Wed, 4 Mar 2020 19:10:32 +0000 (11:10 -0800)]
[CodeGenPrepare] Handle ExtractValueInst in dupRetToEnableTailCallOpts
As the test case shows if there is an ExtractValueInst in the Ret block, function dupRetToEnableTailCallOpts can't duplicate it into the block containing call. So later no tail call is generated in CodeGen.
This patch adds the ExtractValueInst handling code in function dupRetToEnableTailCallOpts and FoldReturnIntoUncondBranch, and later tail call can be generated for this case.
Differential Revision: https://reviews.llvm.org/D74242
Raphael Isemann [Wed, 4 Mar 2020 19:05:48 +0000 (11:05 -0800)]
[lldb] Fix typo in ClangASTSource logging statement
Lei Zhang [Mon, 2 Mar 2020 22:27:05 +0000 (17:27 -0500)]
[mlir][spirv] Add get() method to TargetEnvAttr taking raw values
Also make the getResourceLimits() method more explicit about its return type.
Differential Revision: https://reviews.llvm.org/D75484
Eric Fiselier [Wed, 4 Mar 2020 18:54:04 +0000 (13:54 -0500)]
[libc++] Move std::string extern template declarations to end of class.
They need to appear before any member functions are ODR used, because
they change the visibility of many of these functions and otherwise
they could end up with hidden visibility in the DSO.
Louis Dionne [Wed, 4 Mar 2020 18:54:58 +0000 (13:54 -0500)]
[libc++] Revert to previous implementation of __has_rebind
The new implementation introduced in
5b1e5b43 broke the bot running GCC 5.
Fangrui Song [Wed, 4 Mar 2020 16:17:35 +0000 (08:17 -0800)]
[ELF][test] Improve llvm-objdump -t tests
In many cases, llvm-nm or llvm-readelf is more suitable.
David Green [Wed, 4 Mar 2020 11:45:41 +0000 (11:45 +0000)]
[LSR] Add masked load and store handling
This teaches Loop Strength Reduction the details about masked load and
store address operands, so that it can have a better time optimising
them as it would for normal loads and stores.
Differential Revision: https://reviews.llvm.org/D75371
Raphael Isemann [Wed, 4 Mar 2020 18:32:01 +0000 (10:32 -0800)]
[lldb] Fix several LLDB_LOGs with wrong indices in ClangASTSource.cpp
Mitch Phillips [Wed, 4 Mar 2020 18:04:11 +0000 (10:04 -0800)]
Revert "Fix GSYM tests to run the yaml files and fix test failures on some machines."
This reverts commit
8d41f1a02369537cae1a7d00c0fa717fc3aca575.
This change broke the MSan buildbots - see comments in
https://reviews.llvm.org/D75390 for more information.
Jonathan Coe [Wed, 4 Mar 2020 18:09:26 +0000 (18:09 +0000)]
[clang-format] Improve identification of C# nullables
Summary: Consider `? identifier =` and `? identifier;` to be nullable within function bodies.
Reviewers: krasimir
Reviewed By: krasimir
Subscribers: cfe-commits, MyDeveloperDay
Tags: #clang-format, #clang
Differential Revision: https://reviews.llvm.org/D75606
Jon Chesterfield [Wed, 4 Mar 2020 17:51:38 +0000 (17:51 +0000)]
[libomptarget] Implement most hip atomic functions in terms of intrinsics
Summary:
[libomptarget] Implement hip atomic functions in terms of intrinsics
All but atomicInc can be implemented using type generic clang intrinsics.
There is not yet a corresponding intrinsic for atomicInc in clang, only one in
LLVM. This patch leaves atomicInc as an unresolved symbol.
Reviewers: jdoerfert, ABataev, hfinkel, grokos, arsenm
Reviewed By: arsenm
Subscribers: sri, saiislam, wdng, jvesely, mgorny, jfb, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D73076
River Riddle [Wed, 4 Mar 2020 17:44:36 +0000 (09:44 -0800)]
[mlir][NFC] Sort the operations alphabetically and add header blocks
Summary:
The order of the operations has fallen out of sync as operations have been renamed and new ones have been added.
Differential Revision: https://reviews.llvm.org/D75540
Nikita Popov [Wed, 4 Mar 2020 17:33:00 +0000 (18:33 +0100)]
[InstCombine] Don't explicitly invoke const folding in shift combine
InstCombine uses an IRBuilder that automatically performs
target-dependent constant folding, so explicitly invoking it here
is not necessary.
Nikita Popov [Fri, 28 Feb 2020 21:29:54 +0000 (22:29 +0100)]
[InstCombine] Use IRBuilder to create bitcast
This makes sure that the constant expression bitcast goes through
target-dependent constant folding, and thus avoids an additional
iteration of InstCombine.
Nikita Popov [Sat, 29 Feb 2020 17:51:18 +0000 (18:51 +0100)]
[ConstProp] Add test for bitcast to gep fold; NFC
Nikita Popov [Sat, 29 Feb 2020 17:24:20 +0000 (18:24 +0100)]
[InstSimplify] Add additional icmp of gep folding test; NFC
Nikita Popov [Sat, 29 Feb 2020 17:22:41 +0000 (18:22 +0100)]
[InstSimplify] Regenerate compare.ll checks; NFC
Nikita Popov [Tue, 3 Mar 2020 18:16:11 +0000 (19:16 +0100)]
[ConstantFolding] Always return something from ConstantFoldConstant
Spin-off from D75407. As described there, ConstantFoldConstant()
currently returns null for non-ConstantExpr/ConstantVector inputs,
but otherwise always returns non-null, independently of whether
any folding has happened or not.
This is confusing and makes consumer code more complicated.
I would expect either that ConstantFoldConstant() returns only if
it actually folded something, or that it always returns non-null.
I'm going to the latter possibility here, which appears to be more
useful considering existing usage.
Differential Revision: https://reviews.llvm.org/D75543
Jeremy Morse [Wed, 4 Mar 2020 17:08:17 +0000 (17:08 +0000)]
[analyzer] decode() a bytes object to make Python3 happy
The output of subprocess.check_output is decode()'d through the rest of
this python program, but one appears to have been missed for the output
of the call to "clang -print-file-name=include".
On Windows, with Python 3.6, this leads to the 'args' array being a mix of
bytes and strings, which causes exceptions later down the line.
I can't easily test with python2 on Windows, but python2 on Ubuntu 18.04
was happy with this change.
Craig Topper [Wed, 4 Mar 2020 07:45:59 +0000 (23:45 -0800)]
[X86] Directly form VBROADCAST_LOAD in lowerShuffleAsBroadcast on AVX targets.
If we would emit a VBROADCAST node, we can instead directly emit
a VBROADCAST_LOAD. This allows us to get rid of the special case
to use an f64 load on 32-bit targets for vXi64.
I believe there is more cleanup we can do later in this function,
but I'll do that in follow ups.
Simon Pilgrim [Wed, 4 Mar 2020 17:02:20 +0000 (17:02 +0000)]
[X86] Add tests showing failure to combine consecutive loads + FSHR into a single load
Similar to some of the regressions seen in D75114
Simon Pilgrim [Wed, 4 Mar 2020 15:04:03 +0000 (15:04 +0000)]
[X86] Add tests showing failure to combine consecutive loads + FSHL into a single load
Similar to some of the regressions seen in D75114
Michael Park [Wed, 4 Mar 2020 16:27:14 +0000 (11:27 -0500)]
[libc++][P0174] Deprecated/removed parts of default allocator.
Differential Revision: https://reviews.llvm.org/D70117
Raphael Isemann [Wed, 4 Mar 2020 00:31:25 +0000 (16:31 -0800)]
[lldb][NFC] Modernize logging UserExpression
Raphael Isemann [Wed, 4 Mar 2020 17:01:45 +0000 (09:01 -0800)]
Fix modules build after MatrixBuilder patch
The addition of MatrixBuilder.h broke the modules build:
```
While building module 'LLVM_intrinsic_gen' imported from llvm-project/llvm/lib/IR/AbstractCallSite.cpp:19:
While building module 'LLVM_IR' imported from llvm-project/llvm/include/llvm/IR/Argument.h:19:
In file included from <module-includes>:6:
llvm-project/llvm/include/llvm/IR/MatrixBuilder.h:19:10: fatal error: cyclic dependency in module 'LLVM_intrinsic_gen': LLVM_intrinsic_gen -> LLVM_IR -> LLVM_intrinsic_gen
^
While building module 'LLVM_intrinsic_gen' imported from llvm-project/llvm/lib/IR/AbstractCallSite.cpp:19:
In file included from <module-includes>:1:
llvm-project/llvm/include/llvm/IR/Argument.h:19:10: fatal error: could not build module 'LLVM_IR'
~~~~~~~~^~~~~~~~~~~~~~~~~
llvm-project/llvm/lib/IR/AbstractCallSite.cpp:19:10: fatal error: could not build module 'LLVM_intrinsic_gen'
~~~~~~~~^~~~~~~~~~~~~~~~~~~~
```
Louis Dionne [Wed, 4 Mar 2020 16:32:49 +0000 (11:32 -0500)]
[libc++] Fix reverse_iterator test when UBSan is enabled
The goal of the test was only to check that we could access the
`this->current` member of std::reverse_iterator from a derived
class, but in doing so we incremented a null iterator, which is UB.
Sanjay Patel [Wed, 4 Mar 2020 16:08:28 +0000 (11:08 -0500)]
[PassManager] adjust VectorCombine placement
The initial placement of vector-combine in the opt pipeline revealed phase ordering bugs:
https://bugs.llvm.org/show_bug.cgi?id=45015
https://bugs.llvm.org/show_bug.cgi?id=42022
This patch contains a few independent changes:
1. Move the pass up in the pipeline, so it happens just after loop-vectorization.
This is only to keep vectorization passes together in the pipeline at the moment.
I don't have evidence of interaction between these yet.
2. Add an -early-cse pass after -vector-combine to clean up redundant ops. This was
partly proposed as far back as rL219644 (which is why it's effectively being moved
in the old PM code). This is important because the subsequent -instcombine doesn't
work as well without EarlyCSE. With the CSE, -instcombine is able to squash
shuffles together in 1 of the tests (because those are simple "select" shuffles).
3. Remove the -vector-combine pass that was running after SLP. We may want to do that
eventually, but I don't have a test case to support it yet.
Differential Revision: https://reviews.llvm.org/D75145
Balazs Benics [Wed, 4 Mar 2020 16:03:59 +0000 (17:03 +0100)]
[analyzer][NFC] Use CallEvent checker callback in GenericTaintChecker
Summary:
Intended to be a non-functional change but it turned out CallEvent handles
constructor calls unlike CallExpr which doesn't triggered for constructors.
All in all, this change shouldn't be observable since constructors are not
yet propagating taintness like functions.
In the future constructors should propagate taintness as well.
This change includes:
- NFCi change all uses of the CallExpr to CallEvent
- NFC rename some functions, mark static them etc.
- NFC omit explicit TaintPropagationRule type in switches
- NFC apply some clang-tidy fixits
Reviewers: NoQ, Szelethus, boga95
Reviewed By: Szelethus
Subscribers: martong, whisperity, xazax.hun, baloghadamsoftware, szepet,
a.sidorin, mikhail.ramalho, donat.nagy, dkrupp, Charusso, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D72035
Sanjay Patel [Wed, 4 Mar 2020 15:22:09 +0000 (10:22 -0500)]
[SDAG] simplify FP binops to undef
As discussed in the commit thread for rGa253a2a and D73978, we can do more undef folding for FP ops.
The nnan and ninf fast-math-flags specify that if an operand is the disallowed value, the result is
poison, so we can produce an undef result.
But this doesn't work as expected (the undef operand cases remain) because of a Flags propagation
problem in SelectionDAGBuilder.
I've added DAGCombiner calls to enable these for the other cases because we've shown in other
patches that (because of the limited way that SDAG iterates), it is possible to miss simplifications
like this if they are done only at node creation time.
Several potential follow-ups to expand on this patch are possible.
Differential Revision: https://reviews.llvm.org/D75576
Alex Zinenko [Wed, 4 Mar 2020 10:44:22 +0000 (11:44 +0100)]
[mlir] support conversion of parallel reduction loops to std
Recently introduced support for converting sequential reduction loops to
CFG of basic blocks in the Standard dialect makes it possible to perform
a staged conversion of parallel reduction loops into a similar CFG by
using sequential loops as an intermediate step. This is already the case
for parallel loops without reduction, so extend the pattern to support
an additional use case.
Differential Revision: https://reviews.llvm.org/D75599
Jeremy Morse [Wed, 4 Mar 2020 15:22:31 +0000 (15:22 +0000)]
Quote a python executable path
On my Windows machine at least, the path to python contains a space.
Alexey Bataev [Wed, 4 Mar 2020 14:29:50 +0000 (09:29 -0500)]
[OPENMP50]'source' and 'sink' kinds are not allowed in depobj.
Do not allow to use 'sink' and 'source' dependency kinds in 'depobj'
directive.
AndreyChurbanov [Wed, 4 Mar 2020 15:15:05 +0000 (18:15 +0300)]
[openmp] OpenMP 5.1 omp_display_env function implementation.
Patch by Michael Klemm.
Differential Revision: https://reviews.llvm.org/D74956
Louis Dionne [Wed, 4 Mar 2020 15:10:02 +0000 (10:10 -0500)]
[libc++] Mark the shared_future.wait_for test as being flaky
It is timing sensitive and it fails from time to time. If marking it as
flaky doesn't help, we can try tweaking the time outs.
John Brawn [Wed, 4 Mar 2020 13:25:06 +0000 (13:25 +0000)]
Handle PluginAttrInstances using ManagedStatic
We need to make sure that PluginAttrInstances is deleted before shared libraries
are unloaded, because otherwise when deleting its contents we'll try to access
a virtual destructor which no longer exists.
As shared libraries are managed using ManagedStatic we can do this by also using
ManagedStatic for PluginAttrInstances as ManagedStatics are deleted in reverse
order of construction and we know that PluginAttrInstances will only be
accessed, and thus constructed, after shared libraries have been loaded.
Mikhail Maltsev [Wed, 4 Mar 2020 14:54:24 +0000 (14:54 +0000)]
Revert "Promote nameless lambda used by dl_iterate_phdr to named function to clean up control flow inside findUnwindSections. Also, expose the data structure"
This reverts commit
d93371238e7f5d5f9c506a88cf4f05f8dea9d008.
The commit broke the build in several configurations (including
Windows and bare-metal). For details see comments in
https://reviews.llvm.org/D75480
Alexander Belyaev [Tue, 3 Mar 2020 23:45:45 +0000 (00:45 +0100)]
[MLIR] Add a comment to `requiredOperandCount` arg in LoopOps.cpp.
Differential Revision: https://reviews.llvm.org/D75578
Tatyana Krasnukha [Wed, 12 Feb 2020 10:50:34 +0000 (13:50 +0300)]
[lldb][NFC] Move local variables near to their usage
Tatyana Krasnukha [Tue, 3 Mar 2020 10:29:12 +0000 (13:29 +0300)]
[lldb] Make BreakpointResolver hold weak_ptr instead of raw pointer to breakpoint
This prevents calling Breakpoint::shared_from_this of an object that is not owned by any shared_ptr.
Differential Revision: https://reviews.llvm.org/D74557
David Green [Wed, 4 Mar 2020 12:56:42 +0000 (12:56 +0000)]
[ARM] Change all tests from "thumbv8.1-m.main" to "thumbv8.1m.main". NFC
Stephan Herhut [Wed, 4 Mar 2020 13:39:58 +0000 (14:39 +0100)]
[lld] Avoid creating files outside of work directory.
Summary:
A test is passing `-o -` to lld in the hope of writing the output to
standard out but that is not the case. Instead it creates a file named
`-.lto.o`. This fixes it by creating a temporary file in the work
directory.
Differential Revision: https://reviews.llvm.org/D75605
Evgeniy Brevnov [Wed, 4 Mar 2020 12:49:53 +0000 (19:49 +0700)]
Lost regression test from commit
5a63813dc7f.
Pavel Labath [Tue, 3 Mar 2020 16:03:55 +0000 (17:03 +0100)]
[lldb] Reduce duplication in CommandObjectDisassemble
This command had nearly identical code for the "then" and "else"
branches of the "if (m_options.num_instructions != 0)" condition.
This patch factors out the common parts of the two blocks to reduce
duplication.
Jonathan Coe [Wed, 4 Mar 2020 12:13:33 +0000 (12:13 +0000)]
[clang-format] parse C# object initialisers
Summary:
Treat C# object initializers as braced lists.
Allow lambdas inside C# braced lists.
Reviewers: krasimir
Reviewed By: krasimir
Subscribers: cfe-commits, MyDeveloperDay
Tags: #clang-format, #clang
Differential Revision: https://reviews.llvm.org/D75473
Adrian Kuegel [Fri, 28 Feb 2020 14:57:38 +0000 (15:57 +0100)]
Add rsqrt op to Standard dialect and lower it to LLVM dialect.
Summary:
This adds an rsqrt op to the standard dialect, and lowers
it as 1 / sqrt to the LLVM dialect.
Differential Revision: https://reviews.llvm.org/D75353
Pavel Labath [Tue, 3 Mar 2020 17:27:39 +0000 (18:27 +0100)]
Use new DWARFDataExtractor::getInitialLength in DWARFDebugFrame
Pavel Labath [Tue, 3 Mar 2020 17:51:12 +0000 (18:51 +0100)]
Use new DWARFDataExtractor::getInitialLength in DWARFDebugPubTable
Pavel Labath [Tue, 3 Mar 2020 17:52:56 +0000 (18:52 +0100)]
Use new DWARFDataExtractor::getInitialLength in DWARFUnit
Pavel Labath [Tue, 3 Mar 2020 17:57:18 +0000 (18:57 +0100)]
Use new DWARFDataExtractor::getInitialLength in DWARFVerifier
Pavel Labath [Mon, 2 Mar 2020 11:48:59 +0000 (12:48 +0100)]
Use DWARFDataExtractor::getInitialLength in debug_aranges
Summary:
getInitialLength is a *DWARF*DataExtractor method so I had to "upgrade"
some DataExtractors to be able to make use of it.
Reviewers: ikudrin, jhenderson, probinson
Subscribers: aprantl, hiraditya, llvm-commits, dblaikie
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D75535
Pavel Labath [Fri, 28 Feb 2020 09:03:00 +0000 (10:03 +0100)]
Use DWARFDataExtractor::getInitialLength in DWARFDebugAddr
Reviewers: ikudrin, jhenderson, probinson
Subscribers: hiraditya, dblaikie, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D75532
Kerry McLaughlin [Wed, 4 Mar 2020 11:30:20 +0000 (11:30 +0000)]
[AArch64][SVE] Add SVE2 intrinsic for xar
Summary: Implements the @llvm.aarch64.sve.xar intrinsic
Reviewers: andwar, c-rhodes, dancgr, efriedma, rengolin
Reviewed By: andwar
Subscribers: tschuett, kristof.beyls, hiraditya, rkruppe, psnobl, cfe-commits, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D75160
Kadir Cetinkaya [Wed, 4 Mar 2020 11:41:18 +0000 (12:41 +0100)]
[clangd] Fix buildbots
Evgeniy Brevnov [Wed, 4 Mar 2020 08:36:51 +0000 (15:36 +0700)]
[DependenceAnalysis] Dependecies for loads marked with "ivnariant.load" should not be shared with general accesses(PR42151).
Summary:
This is second attempt to fix the problem with incorrect dependencies reported in presence of invariant load. Initial fix (https://reviews.llvm.org/D64405) was reverted due to a regression reported in https://reviews.llvm.org/D70516.
The original fix changed caching behavior for invariant loads. Namely such loads are not put into the second level cache (NonLocalDepInfo). The problem with that fix is the first level cache (CachedNonLocalPointerInfo) still works as if invariant loads were in the second level cache. The solution is in addition to not putting dependence results into the second level cache avoid putting info about invariant loads into the first level cache as well.
Reviewers: jdoerfert, reames, hfinkel, efriedma
Reviewed By: jdoerfert
Subscribers: DaniilSuchkov, hiraditya, bmahjour, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D73027
Simon Pilgrim [Wed, 4 Mar 2020 11:25:34 +0000 (11:25 +0000)]
Fix some getAs/castAs null dereference static analyzer warnings.
Use castAs if we know the cast should succeed (or we're dereferencing without check), the castAs will assert as well so we can remove local non-null asserts.
Simon Pilgrim [Wed, 4 Mar 2020 10:53:27 +0000 (10:53 +0000)]
[AMDGPU] performCvtF32UByteNCombine - revisit node after src operand simplification.
If SimplifyDemandedBits succeeds in simplifying the byte src, add the CVT_F32_UBYTE node back to the worklist as we might be able to simplify further.
Yet another step towards removing SelectionDAG::GetDemandedBits.
Florian Hahn [Wed, 4 Mar 2020 11:13:48 +0000 (11:13 +0000)]
[Matrix] Add IR MatrixBuilder.
This builder provides a convenient way for targets to lower various matrix
operations to LLVM IR, making use of matrix intrinsics where available.
Reviewers: anemet, Gerolf, hfinkel, andrew.w.kaylor, LuoYuanke
Reviewed By: anemet
Differential Revision: https://reviews.llvm.org/D72280
Kadir Cetinkaya [Wed, 4 Mar 2020 11:07:13 +0000 (12:07 +0100)]
[clangd] Fix windows buildbots
gbreynoo [Wed, 4 Mar 2020 10:56:48 +0000 (10:56 +0000)]
[llvm-ar][test] Add to llvm-ar test coverage
- Added handling of thin archives to symtab.test.
- Added handling of newlines to response.test.
-
62fa3332c9c1af1e66dfecd40f5b4e78882998b2 exposed behaviour
regarding the use of -- on the command line. Added
double-hyphen.test to cover this.
Differential Revision: https://reviews.llvm.org/D73333
Ilya Biryukov [Mon, 2 Mar 2020 15:07:09 +0000 (16:07 +0100)]
[AST] Refactor propagation of dependency bits. NFC
Summary:
This changes introduces an enum to represent dependencies as a bitmask
and extract common patterns from code that computes dependency bits into
helper functions.
Reviewers: rsmith, martong, shafik, ilya-biryukov, hokein
Subscribers: hokein, sammccall, Mordante, riccibruno, merge_guards_bot, rnkovacs, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D71920
evgeny [Wed, 4 Mar 2020 10:05:02 +0000 (13:05 +0300)]
Attempt to fix buildbot after
497c110e
Kadir Cetinkaya [Wed, 19 Feb 2020 16:36:41 +0000 (17:36 +0100)]
[clangd] Make use of syntax tokens in ReplayPreamble
Summary: Replace usage of RawLexer with syntax tokens inside ReplayPreamble.
Reviewers: sammccall
Subscribers: ilya-biryukov, MaskRay, jkorous, arphaman, usaxena95, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D74842
Kadir Cetinkaya [Tue, 25 Feb 2020 08:33:52 +0000 (09:33 +0100)]
[clangd] Fix early selection for non-vardecl declarators
Summary:
Selection tree was performing an early claim only for VarDecls, but
there are other cases where we can have declarators, e.g. FieldDecls. This patch
extends the early claim logic to all types of declarators.
Fixes https://github.com/clangd/clangd/issues/292
Reviewers: sammccall
Subscribers: ilya-biryukov, MaskRay, jkorous, arphaman, usaxena95, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D75106
evgeny [Wed, 4 Mar 2020 09:46:31 +0000 (12:46 +0300)]
[lld][ELF][COFF] Fix archived bitcode files naming
Differential revision: https://reviews.llvm.org/D75422
Georgii Rymar [Wed, 4 Mar 2020 09:33:10 +0000 (12:33 +0300)]
[Object/ELF] - Fix the offset type used in ELFFile<ELFT>::getEntry().
We use size_t for a file offset what is wrong, because size_t is 32-bit
value on 32-bit platforms.
I was reported that after my
0b511c23021
"[llvm-readobj] - Report warnings instead of errors for broken relocations."
The following error is observed on 32-bit Arch Linux:
[100%] Running all regression tests
FAIL: LLVM :: tools/llvm-readobj/ELF/relocation-errors.test (52954 of 54768)
******************** TEST 'LLVM :: tools/llvm-readobj/ELF/relocation-errors.test' FAILED ***
...
llvm-project/llvm/test/tools/llvm-readobj/ELF/relocation-errors.test:9:14:error: LLVM-NEXT: expected string not found in input
# LLVM-NEXT: warning: '[[FILE]]': unable to print relocation 1 in section 3: unable to access section [index 6] data at 0x17e7e7e8b0: offset goes past the end of file
^
<stdin>:9:1: note: scanning from here
/llvm-project/build/bin/llvm-readobj: warning: 'llvm-project/build/test/tools/llvm-readobj/ELF/Output/relocation-errors.test.tmp64': unable to print relocation 1 in section 3: unable to access section [index 6] data at 0xe7e7e8b0: offset goes past the end of file
This patch should fix the issue.
Simon Tatham [Tue, 3 Mar 2020 17:37:14 +0000 (17:37 +0000)]
[ARM,MVE] Add the `vshlcq` intrinsics.
Summary:
The VSHLC instruction performs a left shift of a whole vector register
by an immediate shift count up to 32, shifting in new bits at the low
end from a GPR and delivering the shifted-out bits from the high end
back into the same GPR.
Since the instruction produces two outputs (the shifted vector
register and the output GPR of shifted-out bits), it has to be
instruction-selected in C++ rather than Tablegen.
Reviewers: MarkMurrayARM, dmgreen, miyuki, ostannard
Reviewed By: miyuki
Subscribers: kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D75445
Simon Tatham [Tue, 3 Mar 2020 17:37:05 +0000 (17:37 +0000)]
[ARM,MVE] Add the `vsbciq` intrinsics.
Summary:
These are exactly parallel to the existing `vadciq` intrinsics, which
we implemented last year as part of the original MVE intrinsics
framework setup.
Just like VADC/VADCI, the MVE VSBC/VSBCI instructions deliver two
outputs, both of which the intrinsic exposes: a modified vector
register and a carry flag. So they have to be instruction-selected in
C++ rather than Tablegen. However, in this case, that's trivial: the
same C++ isel routine we already have for VADC works unchanged, and
all we have to do is to pass it a different instruction id.
Reviewers: MarkMurrayARM, dmgreen, miyuki, ostannard
Reviewed By: miyuki
Subscribers: kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D75444
Craig Topper [Tue, 3 Mar 2020 05:59:44 +0000 (21:59 -0800)]
[X86] Directly form VBROADCAST_LOAD for BUILD_VECTOR of splat loads in lowerBuildVectorAsBroadcast.
Charusso [Wed, 4 Mar 2020 06:06:35 +0000 (07:06 +0100)]
[analyzer] AnalysisDeclContext: Refactor and documentation
Summary:
`ScopeContext` wanted to be a thing, but sadly it is dead code.
If you wish to continue the work in D19979, here was a tiny code which
could be reused, but that tiny and that dead, I felt that it is unneded.
Note: Other changes are truly uninteresting.
Reviewed By: NoQ
Differential Revision: https://reviews.llvm.org/D73519
Charusso [Wed, 4 Mar 2020 05:55:50 +0000 (06:55 +0100)]
[analyzer] AnalyzerOptions: Remove 'fixits-as-remarks'
Summary: The new way of checking fix-its is `%check_analyzer_fixit`.
Reviewed By: NoQ, Szelethus, xazax.hun
Differential Revision: https://reviews.llvm.org/D73729
Charusso [Wed, 4 Mar 2020 05:24:52 +0000 (06:24 +0100)]
[analyzer] FixItHint: Apply and test hints with the Clang-Tidy's script
Summary:
This patch introduces a way to apply the fix-its by the Analyzer:
`-analyzer-config apply-fixits=true`.
The fix-its should be testable, therefore I have copied the well-tested
`check_clang_tidy.py` script. The idea is that the Analyzer's workflow
is different so it would be very difficult to use only one script for
both Tidy and the Analyzer, the script would diverge a lot.
Example test: `// RUN: %check-analyzer-fixit %s %t -analyzer-checker=core`
When the copy-paste happened the original authors were:
@alexfh, @zinovy.nis, @JonasToth, @hokein, @gribozavr, @lebedev.ri
Reviewed By: NoQ, alexfh, zinovy.nis
Differential Revision: https://reviews.llvm.org/D69746
hsmahesha [Wed, 4 Mar 2020 05:23:32 +0000 (10:53 +0530)]
[HIP] Make sure, unused hip-pinned-shadow global var is kept within device code
Summary:
hip-pinned-shadow global var should remain in the final code object irrespective
of whether it is used or not within the code. Add it to used list, so that it
will not get eliminated when it is unused.
Reviewers: yaxunl, tra, hliao
Reviewed By: yaxunl
Subscribers: hliao, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D75402
Juneyoung Lee [Sat, 29 Feb 2020 06:45:04 +0000 (15:45 +0900)]
[ValueTracking] Let isGuaranteedNotToBeUndefOrPoison look into branch conditions of dominating blocks' terminators
Summary:
```
br i1 c, BB1, BB2:
BB1:
use1(c)
BB2:
use2(c)
```
In BB1 and BB2, c is never undef or poison because otherwise the branch would have triggered UB.
Checked with Alive2
Reviewers: xbolva00, spatel, lebedev.ri, reames, jdoerfert, nlopes, sanjoy
Reviewed By: reames
Subscribers: jdoerfert, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D75401
Amara Emerson [Tue, 3 Mar 2020 19:36:47 +0000 (11:36 -0800)]
[GlobalISel][Localizer] Enable intra-block localization of already-local uses.
This changes the localizer to attempt intra-block localizer of instructions
that have local uses. This is useful because sometimes the entry block itself
has many uses of constant-like instructions, which would benefit from shortening
live ranges. Previously if an inst had no non-local uses, we wouldn't add it to
the list of instructions to attempt further intra-block localization.
This gives a 0.7% geomean code size improvement on CTMark.
Differential Revision: https://reviews.llvm.org/D75555
Fangrui Song [Wed, 4 Mar 2020 01:14:25 +0000 (17:14 -0800)]
[MC][test] Improve some llvm-objdump -t tests
Delete two redundant tests.
Frank Laub [Wed, 4 Mar 2020 00:24:03 +0000 (16:24 -0800)]
[MLIR][NFC] Fix for VS2017 bug.
Apparently there is a bug in VS2017 concerning compile-time constant
expressions.
https://developercommunityapi.westus.cloudapp.azure.com/content/problem/110435/not-compile-time-constant-expression.html
This change works with any C++11 compiler (including VS2017) and is a little bit
easier to read.
Reviewers: stella.stamenova, nicolasvasilache
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D75584
Fangrui Song [Wed, 4 Mar 2020 00:37:52 +0000 (16:37 -0800)]
[gn build] Fix llvm-gsymutil after D75291
Fangrui Song [Tue, 3 Mar 2020 21:58:02 +0000 (13:58 -0800)]
[MCDwarf] Change emitListsTableHeaderStart to use a reference and fold Start/End symbols generation into it
Apply @dblaikie's suggestions in a post-commit review for D75375
Reviewed By: dblaikie
Differential Revision: https://reviews.llvm.org/D75568
Lang Hames [Wed, 4 Mar 2020 00:02:46 +0000 (16:02 -0800)]
[ORC] Skip ST_File symbols in MaterializationUnit interfaces / resolution.
ST_File symbols aren't relevant for linking purposes, but can end up shadowing
real symbols if they're not filtered.
No test case yet: The ideal testcase for this would be an ELF llvm-jitlink test,
but llvm-jitlink support for ELF is still under development. We should add a
testcase for this once support lands in tree.
Stefanos Baziotis [Wed, 4 Mar 2020 00:12:33 +0000 (02:12 +0200)]
[LoopTerminology][NFC] Fix typo
Richard Smith [Tue, 3 Mar 2020 23:56:08 +0000 (15:56 -0800)]
PR45087: Fix check for emptiness when determining whether a trivial copy
operation needs to read from its operand.
Greg Clayton [Tue, 3 Mar 2020 23:51:58 +0000 (15:51 -0800)]
Fix buildbots by including MC for StringTableBuilder.
Alexander Belyaev [Tue, 3 Mar 2020 23:37:50 +0000 (00:37 +0100)]
[MLIR] Expose makeCanonicalStridedLayoutExpr in StandardTypes.h.
Differential Revision: https://reviews.llvm.org/D75575
Richard Smith [Tue, 3 Mar 2020 23:19:13 +0000 (15:19 -0800)]
PR45083: Mark statement expressions as being dependent if they appear in
dependent contexts.
We previously assumed they were neither value- nor
instantiation-dependent under any circumstances, which would lead to
crashes and other misbehavior.
Martijn Vels [Tue, 3 Mar 2020 22:47:23 +0000 (17:47 -0500)]
Partially inline basic_string copy constructor in UNSTABLE
his change splits the copy constructor up inlining short initialization, and explicitly outlining long initialization into __init_copy_ctor_external() which is the externally instantiated slow path.
For unstable ABI, this has the following changes:
remove basic_string(const basic_string&)
remove basic_string(const basic_string&, const Allocator&)
add __init_copy_ctor_external(const value_type*, size_type)
Quick local benchmark for Copy:
Master
```
---------------------------------------------------------------
Benchmark Time CPU Iterations
---------------------------------------------------------------
BM_StringCopy_Empty 3.50 ns 3.51 ns
199326720
BM_StringCopy_Small 3.50 ns 3.51 ns
199510016
BM_StringCopy_Large 15.7 ns 15.7 ns
45230080
BM_StringCopy_Huge 1503 ns 1503 ns 464896
```
```
---------------------------------------------------------------
Benchmark Time CPU Iterations
---------------------------------------------------------------
BM_StringCopy_Empty 1.99 ns 2.00 ns
356471808
BM_StringCopy_Small 3.29 ns 3.30 ns
203425792
BM_StringCopy_Large 13.3 ns 13.3 ns
52948992
BM_StringCopy_Huge 1472 ns 1472 ns 475136
```
Author: Martijn Vels <martijn.vels@gmail.com>
Reviewers: EricWF, mclow.list
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D73223
Lang Hames [Tue, 3 Mar 2020 22:18:12 +0000 (14:18 -0800)]
[JITLink] Add a -slab-address option to llvm-jitlink.
This option can be used to for JITLink to link as-if the target memory slab were
allocated at a specific start address. This can be used to both verify that
cross-address space linking is working correctly, and to ensure that certain
address-sensitive optimizations (e.g. GOT and stub elimination) either do or do
not fire, depending on the requirements of the test case.
This argument is only valid for testing in conjunction with -noexec -slab-alloc,
and will produce an error if used without those arguments.
Jonathan Coe [Tue, 3 Mar 2020 22:21:33 +0000 (22:21 +0000)]
[clang-format] Do not format C# array subscript operators as attributes
Summary:
Fix misidentification of C# array subscript operators.
Reviewers: krasimir
Reviewed By: krasimir
Subscribers: cfe-commits, MyDeveloperDay
Tags: #clang-format, #clang
Differential Revision: https://reviews.llvm.org/D75517