Igor Kudrin [Thu, 16 May 2019 05:23:25 +0000 (05:23 +0000)]
[LTO] Improve readability of module IDs
Module IDs can appear in diagnostic messages.
This patch adds some auxiliary symbols to improve their readability.
Differential Revision: https://reviews.llvm.org/D61857
llvm-svn: 360858
Igor Kudrin [Thu, 16 May 2019 05:23:13 +0000 (05:23 +0000)]
[IRMover] Improve diagnostic messages for conflicting metadata
This does the similar for error messages as rL344011 has done for warnings.
With llvm::lto::LTO, the error might appear when LTO::run() is executed.
In that case, the calling code cannot know which module causes the error
and, subsequently, cannot hint the user.
Differential Revision: https://reviews.llvm.org/D61880
llvm-svn: 360857
Matt Arsenault [Thu, 16 May 2019 04:09:06 +0000 (04:09 +0000)]
GlobalISel: Add buildFConstant for APFloat
llvm-svn: 360853
Matt Arsenault [Thu, 16 May 2019 04:08:55 +0000 (04:08 +0000)]
GlobalISel: Add some FP instructions to MachineIRBuilder
This makes FP legalization code more convenient.
llvm-svn: 360852
Matt Arsenault [Thu, 16 May 2019 04:08:46 +0000 (04:08 +0000)]
GlobalISel: Fix indentation
llvm-svn: 360851
Matt Arsenault [Thu, 16 May 2019 04:08:39 +0000 (04:08 +0000)]
GlobalISel: Add G_FCOPYSIGN
llvm-svn: 360850
Matt Arsenault [Thu, 16 May 2019 04:08:25 +0000 (04:08 +0000)]
Fix missing const
llvm-svn: 360849
Matt Morehouse [Thu, 16 May 2019 04:00:41 +0000 (04:00 +0000)]
[libFuzzer] Increase merge-sigusr sleep after sending signal.
Test is flaky on buildbot at least partially due to the fuzz target not
exiting before we read its output.
llvm-svn: 360848
Rui Ueyama [Thu, 16 May 2019 03:54:50 +0000 (03:54 +0000)]
Pemove SymbolTable::addBitcode as it is redundant.
Differential Revision: https://reviews.llvm.org/D61897
llvm-svn: 360846
Rui Ueyama [Thu, 16 May 2019 03:54:41 +0000 (03:54 +0000)]
Consistently return `Symbol *` from SymbolTable's add-family functions.
llvm-svn: 360845
Rui Ueyama [Thu, 16 May 2019 03:45:13 +0000 (03:45 +0000)]
De-template parseFile() and SymbolTable's add-family functions.
Differential Revision: https://reviews.llvm.org/D61896
llvm-svn: 360844
Yevgeny Rouban [Thu, 16 May 2019 03:39:09 +0000 (03:39 +0000)]
Fix prof branch_weights in entry_counts_missing_dbginfo.ll test
Removed extra parameter from !prof branch_weights metadata of
a call instruction according to the spec.
Differential Revision: https://reviews.llvm.org/D61932
llvm-svn: 360843
Kristina Brooks [Thu, 16 May 2019 03:30:08 +0000 (03:30 +0000)]
Revert r360833 until I can work out the issue with Win32 bots
This reverts "r360833: [Clang][PP] Add the __FILE_NAME__ builtin macro."
The tests are failing on Windows bots, reverting the patchset until I can
work out why.
llvm-svn: 360842
Rui Ueyama [Thu, 16 May 2019 03:29:03 +0000 (03:29 +0000)]
Introduce CommonSymbol.
Previously, we handled common symbols as a kind of Defined symbol,
but what we were doing for common symbols is pretty different from
regular defined symbols.
Common symbol and defined symbol are probably as different as shared
symbol and defined symbols are different.
This patch introduces CommonSymbol to represent common symbols.
After symbols are resolved, they are converted to Defined symbols
residing in a .bss section.
Differential Revision: https://reviews.llvm.org/D61895
llvm-svn: 360841
Matt Morehouse [Thu, 16 May 2019 03:04:44 +0000 (03:04 +0000)]
[libFuzzer] Also kill parent process in merge-siguser.test.
llvm-svn: 360840
Kristina Brooks [Thu, 16 May 2019 02:46:12 +0000 (02:46 +0000)]
Fix assumption about Win32 paths in r360833
Attempt to fix Windows buildbots due to differences in
path handling (caused by r360833).
llvm-svn: 360839
Rui Ueyama [Thu, 16 May 2019 02:14:00 +0000 (02:14 +0000)]
Simplify SymbolTable::add{Defined,Undefined,...} functions.
SymbolTable's add-family functions have lots of parameters because
when they have to create a new symbol, they forward given arguments
to Symbol's constructors. Therefore, the functions take at least as
many arguments as their corresponding constructors.
This patch simplifies the add-family functions. Now, the functions
take a symbol instead of arguments to construct a symbol. If there's
no existing symbol, a given symbol is memcpy'ed to the symbol table.
Otherwise, the functions attempt to merge the existing and a given
new symbol.
I also eliminated `CanOmitFromDynSym` parameter, so that the functions
take really one argument.
Symbol classes are trivially constructible, so looks like constructing
them to pass to add-family functions is as cheap as passing a lot of
arguments to the functions. A quick benchmark showed that this patch
seems performance-neutral.
This is a preparation for
http://lists.llvm.org/pipermail/llvm-dev/2019-April/131902.html
Differential Revision: https://reviews.llvm.org/D61855
llvm-svn: 360838
Richard Smith [Thu, 16 May 2019 02:06:16 +0000 (02:06 +0000)]
Fix regression in r360311 caused by reversed bool arguments.
llvm-svn: 360837
Matt Morehouse [Thu, 16 May 2019 01:56:11 +0000 (01:56 +0000)]
[libFuzzer] Fix typo in merge-sigusr.test.
llvm-svn: 360836
Matt Morehouse [Thu, 16 May 2019 01:32:39 +0000 (01:32 +0000)]
[libFuzzer] Use PID to send signals rather than process name.
pkill reads the process name as a pattern, not a raw name. This means
that if the process name contains + or other regex characters, pkill
fails.
llvm-svn: 360835
Eric Christopher [Thu, 16 May 2019 01:07:54 +0000 (01:07 +0000)]
Fix typo in comment of CSAction -> Action.
llvm-svn: 360834
Kristina Brooks [Thu, 16 May 2019 00:52:41 +0000 (00:52 +0000)]
[Clang][PP] Add the __FILE_NAME__ builtin macro.
This patch adds the `__FILE_NAME__` macro that expands to the
last component of the path, similar to `__FILE__` except with
a guarantee that only the last path component (without the
separator) will be rendered.
I intend to follow through with discussion of this with WG14
as a potential inclusion in the C standard or failing that,
try to discuss this with GCC developers since this extension
is desired by GCC and Clang users/developers alike.
Differential Revision: https://reviews.llvm.org/D61756
llvm-svn: 360833
Sam Clegg [Thu, 16 May 2019 00:45:50 +0000 (00:45 +0000)]
[ELF] Fix typo in header guard. NFC.
llvm-svn: 360832
Lang Hames [Thu, 16 May 2019 00:21:10 +0000 (00:21 +0000)]
[ORC] Modify DenseMap hashing for SymbolStringPtrs.
Modifies the DenseMapInfo<SymbolStringPtr>::getHashValue method to take its
argument by const-ref rather than by value (to avoid unnecessary ref-counting
operations) and to defer to DenseMapInfo<void*> for the hash value computation
(since SymbolStringPtrs are just pointers under the hood).
llvm-svn: 360831
Lang Hames [Thu, 16 May 2019 00:21:07 +0000 (00:21 +0000)]
[JITLink][MachO] Use getSymbol64TableEntry for 64-bit MachO files.
Fixes a think-o. No test case: The nlist and nlist64 data structures happen to
line up for this field, so there's no way to construct a failing test case.
llvm-svn: 360830
Vitaly Buka [Thu, 16 May 2019 00:19:37 +0000 (00:19 +0000)]
Fix GN build
llvm-svn: 360829
Richard Smith [Wed, 15 May 2019 23:36:14 +0000 (23:36 +0000)]
Make tentative parsing to detect template-argument-lists less aggressive
(and less wrong).
It's not correct to assume that X<something, Type> is always a
template-id; there are a few cases where the comma takes us into a
non-expression syntactic context in which 'Type' might be permissible.
Stop doing that.
This slightly regresses our error recovery on the cases where the
construct is intended to be a template-id. We typically do still manage
to diagnose a missing 'template' keyword, but we realize this too late
to properly recover from the error.
This fixes a regression introduced by r360308.
llvm-svn: 360827
Simon Atanasyan [Wed, 15 May 2019 22:27:19 +0000 (22:27 +0000)]
[mips] Always use _LARGEFILE_SOURCE / _FILE_OFFSET_BITS for building MIPS 32-bit
When MIPS 32-bit compiler-rt is building on 32-bit host or using 32-bit
`DLLVM_HOST_TRIPLE` the `_LARGEFILE_SOURCE` and the `_FILE_OFFSET_BITS=64`
macros defined by statements from the `HandleLLVMOptions.cmake`. In case
of building 32-bit libraries on 64-bit host using default host triple
these macros are not defined. As a result assertions check a consistency
between the `struct_kernel_stat_sz` constant and the `struct_kernel_stat_sz`
start to fail.
To resolve this problem and enable building both 32/64-bit versions
of MIPS compiler-rt libraries on 64-bit host at once always explicitly
define the `_LARGEFILE_SOURCE` and the `_FILE_OFFSET_BITS=64` macros
for MIPS 32-bit.
llvm-svn: 360825
Matt Morehouse [Wed, 15 May 2019 22:26:48 +0000 (22:26 +0000)]
[libFuzzer] Echo fuzzer output on sigusr tests.
Improves debuggability when the fuzz target crashes.
llvm-svn: 360824
Craig Topper [Wed, 15 May 2019 21:59:53 +0000 (21:59 +0000)]
[X86] Delay creating index register negations during address matching until after we know for sure the match will succeed
If we're trying to match an LEA, its possible the LEA match will be deemed unprofitable. In which case the negation we created in matchAddress would be left dangling in the SelectionDAG. This could artificially increase use counts for other nodes in the DAG. Though I don't have an example of that. But it just seems like bad form to have dangling nodes in isel.
Differential Revision: https://reviews.llvm.org/D61047
llvm-svn: 360823
Adrian Prantl [Wed, 15 May 2019 21:58:43 +0000 (21:58 +0000)]
Update LLDB license on the LLDB home page to Apache+LLVM.
llvm-svn: 360822
Adrian Prantl [Wed, 15 May 2019 21:49:00 +0000 (21:49 +0000)]
Get back the navigation sidebar on the LLDB website.
This returns the look & feel of the Sphinx-generated LLDB website to
the original pre-Sphinx layout.
Differential Revision: https://reviews.llvm.org/D61913
llvm-svn: 360819
Reid Kleckner [Wed, 15 May 2019 21:46:05 +0000 (21:46 +0000)]
[codeview] Fix SDNode representation of annotation labels
Before this change, they were erroneously constructed with the EH_LABEL
SDNode opcode, which caused other passes to interact with them in
incorrect ways. See the FIXME about fastisel that this addresses in the
existing test case.
Fixes PR41890
llvm-svn: 360818
Simon Atanasyan [Wed, 15 May 2019 21:26:25 +0000 (21:26 +0000)]
[mips] Use range-based `for` loops. NFC
llvm-svn: 360817
Mandeep Singh Grang [Wed, 15 May 2019 21:23:41 +0000 (21:23 +0000)]
[AArch64] only indicate CFI on Windows if we emitted CFI
Summary:
Otherwise, we emit directives for CFI without any actual CFI opcodes to
go with them, which causes tools to malfunction. The technique is
similar to what the x86 backend already does.
Fixes https://bugs.llvm.org/show_bug.cgi?id=40876
Patch by: froydnj (Nathan Froyd)
Reviewers: mstorsjo, eli.friedman, rnk, mgrang, ssijaric
Reviewed By: rnk
Subscribers: javed.absar, kristof.beyls, llvm-commits, dmajor
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D61960
llvm-svn: 360816
Craig Topper [Wed, 15 May 2019 21:16:28 +0000 (21:16 +0000)]
[X86] Strengthen type constraints on some specialized X86 ISD opcodes that don't have any flexibility. NFC
These particular instructions only operate on 128-bit vectors and have no wider equivalents. And the
element size is always known.
One could argue that MOVSS/MOVSD could be merged, but that's probably disruptive to code in
X86ISelLowering and probably low value.
llvm-svn: 360815
Roman Lebedev [Wed, 15 May 2019 21:15:44 +0000 (21:15 +0000)]
[NFC][InstCombine] Add some more tests for pulling binops through shifts
The ashr variant may see relaxation in https://reviews.llvm.org/D61938
llvm-svn: 360814
Reid Kleckner [Wed, 15 May 2019 20:53:39 +0000 (20:53 +0000)]
[codeview] Finish support for reading and writing S_ANNOTATION records
Implement dumping via llvm-pdbutil and llvm-readobj.
llvm-svn: 360813
Cameron McInally [Wed, 15 May 2019 20:48:50 +0000 (20:48 +0000)]
Revert llvm-svn: 360807
Somehow submitted this patch twice.
llvm-svn: 360812
David Bolvansky [Wed, 15 May 2019 20:29:33 +0000 (20:29 +0000)]
[clang-format] Fixed self assignment
Reviewers: MyDeveloperDay, RKSimon
Reviewed By: MyDeveloperDay
Subscribers: RKSimon, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D61281
llvm-svn: 360810
Gheorghe-Teodor Bercea [Wed, 15 May 2019 20:28:23 +0000 (20:28 +0000)]
[OpenMP][Bugfix] Move double and float versions of abs under c++ macro
Summary:
This is a fix for the reported bug:
[[ https://bugs.llvm.org/show_bug.cgi?id=41861 | 41861 ]]
abs functions need to be moved under the c++ macro to avoid conflicts with included headers.
Reviewers: tra, jdoerfert, hfinkel, ABataev, caomhin
Reviewed By: jdoerfert
Subscribers: guansong, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D61959
llvm-svn: 360809
Cameron McInally [Wed, 15 May 2019 20:27:37 +0000 (20:27 +0000)]
Pre-commit unary fneg tests to InstSimplify
llvm-svn: 360808
Cameron McInally [Wed, 15 May 2019 20:27:35 +0000 (20:27 +0000)]
Add unary fneg to InstSimplify/fp-nan.ll
llvm-svn: 360807
Richard Smith [Wed, 15 May 2019 20:22:21 +0000 (20:22 +0000)]
[c++20] For P1327R1: support dynamic_cast in constant expression
evaluation.
llvm-svn: 360806
Kristof Umann [Wed, 15 May 2019 20:19:51 +0000 (20:19 +0000)]
Revert "[analyzer] Add a test for plugins using checker dependencies"
Buildbots don't seem to find the new plugin.
llvm-svn: 360805
Gheorghe-Teodor Bercea [Wed, 15 May 2019 20:18:21 +0000 (20:18 +0000)]
[OpenMP][bugfix] Fix issues with C++ 17 compilation when handling math functions
Summary: In OpenMP device offloading we must ensure that unde C++ 17, the inclusion of cstdlib will works correctly.
Reviewers: ABataev, tra, jdoerfert, hfinkel, caomhin
Reviewed By: jdoerfert
Subscribers: Hahnfeld, guansong, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D61949
llvm-svn: 360804
Erik Pilkington [Wed, 15 May 2019 20:15:01 +0000 (20:15 +0000)]
[CodeGenObjC] invoke objc_autorelease, objc_retain when necessary
Any of these methods can be overridden, so we need to invoke these functions.
Differential revision: https://reviews.llvm.org/D61957
llvm-svn: 360802
Michal Gorny [Wed, 15 May 2019 20:03:31 +0000 (20:03 +0000)]
[lldb] [test] Skip one more TestMiBreak on NetBSD
llvm-svn: 360800
Kristof Umann [Wed, 15 May 2019 19:47:26 +0000 (19:47 +0000)]
[analyzer] Add a test for plugins using checker dependencies
Also, I moved the existing analyzer plugin to test/ as well, in order not to
give the illusion that the analyzer supports plugins -- it's capable of handling
them, but does not _support_ them.
Differential Revision: https://reviews.llvm.org/D59464
llvm-svn: 360799
Pete Couperus [Wed, 15 May 2019 19:46:17 +0000 (19:46 +0000)]
Uncomment LLVM_FALLTHROUGH.
llvm-svn: 360798
Cameron McInally [Wed, 15 May 2019 19:37:03 +0000 (19:37 +0000)]
Add unary fneg to InstSimplify/fp-nan.ll
llvm-svn: 360797
Taewook Oh [Wed, 15 May 2019 19:35:38 +0000 (19:35 +0000)]
[PredicateInfo] Do not process unreachable operands.
Summary: We should excluded unreachable operands from processing as their DFS visitation order is undefined. When `renameUses` function sorts `OpsToRename` (https://fburl.com/d2wubn60), the comparator assumes that the parent block of the operand has a corresponding dominator tree node. This is not the case for unreachable operands and crashes the compiler.
Reviewers: dberlin, mgrang, davide
Subscribers: efriedma, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D61154
llvm-svn: 360796
Jan Kratochvil [Wed, 15 May 2019 19:22:33 +0000 (19:22 +0000)]
Permit cross-CU references
So far dw_offset_t was global for the whole SymbolFileDWARF but with
.debug_types the same dw_offset_t may mean two different things depending on
its section (=CU). So references now return whole new referenced DWARFDIE
instead of just dw_offset_t.
This means that some functions have to now handle 16 bytes instead of 8 bytes
but I do not see that anywhere performance critical.
Differential Revision: https://reviews.llvm.org/D61502
llvm-svn: 360795
Sam Clegg [Wed, 15 May 2019 19:01:51 +0000 (19:01 +0000)]
[lld] Remove 2>&1 from --trace-symbol tests
The tracing goes to stdout so this is not needed.
Also remove the "not" from the final check in ELF/trace-symbols.s.
According the comment the check is that we don't crash, so we should
be checking for success here. Previously this step is error'ing with
undefined symbols because it didn't include all the needed objects.
Differential Revision: https://reviews.llvm.org/D61928
llvm-svn: 360794
Artem Dergachev [Wed, 15 May 2019 18:41:32 +0000 (18:41 +0000)]
[analyzer] RetainCount: Fix os_returns_retained_on_zero with weird return types.
The checker was crashing when it was trying to assume a structure
to be null or non-null so that to evaluate the effect of the annotation.
Differential Revision: https://reviews.llvm.org/D61958
llvm-svn: 360790
Nicolai Haehnle [Wed, 15 May 2019 17:48:10 +0000 (17:48 +0000)]
[MachineOperand] Add a ChangeToGA method
Summary:
Analogous to the other ChangeToXXX methods. See the next patch for a
use case.
Change-Id: I6548d614706834fb9109ab3c8fe915e9c6ece2a7
Reviewers: arsenm, kzhuravl
Subscribers: wdng, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D61651
llvm-svn: 360789
Don Hinton [Wed, 15 May 2019 17:47:51 +0000 (17:47 +0000)]
[clang-tidy] Recommit r360785 "modernize-loop-convert: impl const cast iter" with correct attribution
Summary:
modernize-loop-convert was not detecting implicit casts to
const_iterator as convertible to range-based loops:
std::vector<int> vec{1,2,3,4}
for(std::vector<int>::const_iterator i = vec.begin();
i != vec.end();
++i) { }
Thanks to Don Hinton for advice.
As well, this change adds a note for this check's applicability to code
targeting OpenMP prior version 5 as this check will continue breaking
compilation with `-fopenmp`. Thanks to Roman Lebedev for pointing this
out.
Fixes PR#35082
Patch by Torbjörn Klatt!
Reviewed By: hintonda
Tags: #clang-tools-extra, #clang
Differential Revision: https://reviews.llvm.org/D61827
llvm-svn: 360788
Don Hinton [Wed, 15 May 2019 17:36:54 +0000 (17:36 +0000)]
Revert [clang-tidy] modernize-loop-convert: impl const cast iter
This reverts r360785 (git commit
42d28be802fe5beab18bc1a27f89894c0a290d44)
llvm-svn: 360787
Nicolai Haehnle [Wed, 15 May 2019 17:29:58 +0000 (17:29 +0000)]
RegAlloc: try to fail more gracefully when out of registers
Summary:
The emitError path allows the program to continue, unlike report_fatal_error.
This is friendlier to use cases where LLVM is embedded in a larger program,
because the caller may be able to deal with the error somewhat gracefully.
Change the number of requested NOP bytes in the AArch64 and PowerPC
test cases to avoid triggering an unrelated assertion. The compilation
still fails, as verified by the test.
Change-Id: Iafb9ca341002a597b82e59ddc7a1f13c78758e3d
Reviewers: arsenm, MatzeB
Subscribers: qcolombet, nemanjai, wdng, javed.absar, kristof.beyls, kbarton, jsji, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D61489
llvm-svn: 360786
Don Hinton [Wed, 15 May 2019 16:58:58 +0000 (16:58 +0000)]
[clang-tidy] modernize-loop-convert: impl const cast iter
Summary:
modernize-loop-convert was not detecting implicit casts to
const_iterator as convertible to range-based loops:
std::vector<int> vec{1,2,3,4}
for(std::vector<int>::const_iterator i = vec.begin();
i != vec.end();
++i) { }
Thanks to Don Hinton for advice.
As well, this change adds a note for this check's applicability to code
targeting OpenMP prior to version 5 as this check will continue breaking
compilation with `-fopenmp`. Thanks to Roman Lebedev for pointing this
out.
Fixes PR#35082
Reviewed By: hintonda
Tags: #clang-tools-extra, #clang
Differential Revision: https://reviews.llvm.org/D61827
llvm-svn: 360785
Andrey Churbanov [Wed, 15 May 2019 16:53:45 +0000 (16:53 +0000)]
Fixed https://bugs.llvm.org/show_bug.cgi?id=41584.
Removed unconditional and unsafe decrement of counter
of active threads in pool at shutdown time.
Differential Revision: https://reviews.llvm.org/D61944
llvm-svn: 360784
Kevin Petit [Wed, 15 May 2019 16:39:28 +0000 (16:39 +0000)]
Test commit
Remove stray space.
Signed-off-by: Kevin Petit <kevin.petit@arm.com>
llvm-svn: 360783
Sam Clegg [Wed, 15 May 2019 16:03:28 +0000 (16:03 +0000)]
[WebAssembly] LTO: Honor comdat groups when loading bitcode files
But don't apply comdat groups when loading the LTO object files.
This is basically the same logic used by the ELF linker.
Differential Revision: https://reviews.llvm.org/D61924
llvm-svn: 360782
Thomas Preud'homme [Wed, 15 May 2019 15:20:45 +0000 (15:20 +0000)]
[FileCheck] Fix sphinx error: Make input be gas block
Summary:
Change example of input text from being llvm block to being gas block
since that text is made-up assembly.
Reviewers: jhenderson, jdenny, probinson, arichardson
Subscribers: llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D61893
llvm-svn: 360781
Hiroshi Yamauchi [Wed, 15 May 2019 15:15:16 +0000 (15:15 +0000)]
[JumpThreading] A bug fix for stale loop info after unfold select
Summary:
The return value of a TryToUnfoldSelect call was not checked, which led to an
incorrectly preserved loop info and some crash.
The original crash was reported on https://reviews.llvm.org/D59514.
Reviewers: davidxl, amehsan
Reviewed By: davidxl
Subscribers: fhahn, brzycki, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D61920
llvm-svn: 360780
Kristof Umann [Wed, 15 May 2019 15:06:25 +0000 (15:06 +0000)]
[clang-tidy] new check: bugprone-branch-clone
Implement a check for detecting if/else if/else chains where two or more
branches are Type I clones of each other (that is, they contain identical code)
and for detecting switch statements where two or more consecutive branches are
Type I clones of each other.
Patch by DonĂ¡t Nagy!
Differential Revision: https://reviews.llvm.org/D54757
llvm-svn: 360779
Ryan Taylor [Wed, 15 May 2019 14:43:55 +0000 (14:43 +0000)]
[AMDGPU] Increases available SGPR for Calling Convention
Summary:
SGPR in CC can be either hw initialized or set by other chained shaders
and so this increases the SGPR count availalbe to CC to 105.
Change-Id: I3dfadc750fe4a3e2bd07117a2899fd13f3e2fef3
Subscribers: arsenm, kzhuravl, jvesely, wdng, nhaehnle, yaxunl, dstuttard, tpr, t-tye, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D61261
llvm-svn: 360778
Cameron McInally [Wed, 15 May 2019 14:31:33 +0000 (14:31 +0000)]
Teach InstSimplify -X + X --> 0.0 about unary FNeg
Differential Revision: https://reviews.llvm.org/D61916
llvm-svn: 360777
Clement Courbet [Wed, 15 May 2019 14:21:59 +0000 (14:21 +0000)]
Revert r360771 "[MergeICmps] Simplify the code."
Breaks a bunch of builbdots.
llvm-svn: 360776
Clement Courbet [Wed, 15 May 2019 14:00:45 +0000 (14:00 +0000)]
[MergeICmps] Fix r360771.
Twine references a StringRef by reference, not value...
llvm-svn: 360775
Andrey Churbanov [Wed, 15 May 2019 13:45:36 +0000 (13:45 +0000)]
Introduce new OpenMP 5.0 depend object type.
The implementation should be done by compiler, user can only declare
objects of this type and use them in OpenMP directives.
Differential Revision: https://reviews.llvm.org/D61860
llvm-svn: 360774
Stephen Tozer [Wed, 15 May 2019 13:41:44 +0000 (13:41 +0000)]
Revert "[Salvage] Change salvage debug info implementation to use DW_OP_LLVM_convert where needed"
This reverts r360772 due to build issues.
Reverted commit:
17dd4d7403770bd683675e45f5517e0cdb8f9b2b.
llvm-svn: 360773
Stephen Tozer [Wed, 15 May 2019 13:15:48 +0000 (13:15 +0000)]
[Salvage] Change salvage debug info implementation to use DW_OP_LLVM_convert where needed
Fixes issue: https://bugs.llvm.org/show_bug.cgi?id=40645
Previously, LLVM had no functional way of performing casts inside of a
DIExpression(), which made salvaging cast instructions other than Noop
casts impossible. With the recent addition of DW_OP_LLVM_convert this
salvaging is now possible, and so can be used to fix the attached bug as
well as any cases where SExt instruction results are lost in the
debugging metadata. This patch introduces this fix by expanding the
salvage debug info method to cover these cases using the new operator.
Differential revision: https://reviews.llvm.org/D61184
llvm-svn: 360772
Clement Courbet [Wed, 15 May 2019 13:04:24 +0000 (13:04 +0000)]
[MergeICmps] Simplify the code.
Instead of patching the original blocks, we now generate new blocks and
delete the old blocks. This results in simpler code with a less twisted
control flow (see the change in `entry-block-shuffled.ll`).
This will make https://reviews.llvm.org/D60318 simpler by making it more
obvious where control flow created and deleted.
Reviewers: gchatelet
Subscribers: hiraditya, llvm-commits, spatel
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D61736
llvm-svn: 360771
Simon Pilgrim [Wed, 15 May 2019 13:03:10 +0000 (13:03 +0000)]
Revert rL360675 : [APFloat] APFloat::Storage::Storage - fix use after move
This was mentioned both in https://www.viva64.com/en/b/0629/ and by scan-build checks
........
There's concerns this may just introduce a use-after-free instead.....
llvm-svn: 360770
David Green [Wed, 15 May 2019 12:58:02 +0000 (12:58 +0000)]
[ARM] Don't use the Machine Scheduler for cortex-m at minsize
The new cortex-m schedule in rL360768 helps performance, but can increase the
amount of high-registers used. This, on average, ends up increasing the
codesize by a fair amount (because less instructions are converted from T2 to
T1). On cortex-m at -Oz, where we are quite size-paranoid, it is better to use
the existing DAG scheduler with the RegPressure scheduling preference (at least
until the issues around T2 vs T1 instructions can be improved).
I have also made sure that the Sched::RegPressure dag scheduler is always
chosen for MinSize.
The test shows one case where we increase the number of registers used.
Differential Revision: https://reviews.llvm.org/D61882
llvm-svn: 360769
David Green [Wed, 15 May 2019 12:41:58 +0000 (12:41 +0000)]
[ARM] Cortex-M4 schedule
This patch adds a simple Cortex-M4 schedule, renaming the existing M3
schedule to M4 and filling in the latencies as-per the Cortex-M4 TRM:
https://developer.arm.com/docs/ddi0439/latest
Most of these are 1, with the important exception being loads taking 2
cycles. A few others are also higher, but I don't believe they make a
large difference. I've repurposed the M3 schedule as the latencies are
mostly the same between the two cores, with the M4 having more FP and
DSP instructions. We also turn on MISched and UseAA for the cores that
now use this.
It also adds some schedule Write's to various instruction to make things
simpler.
Differential Revision: https://reviews.llvm.org/D54142
llvm-svn: 360768
Michal Gorny [Wed, 15 May 2019 12:13:20 +0000 (12:13 +0000)]
[lldb] [test] Mark frequently failing flaky tests skipped on NetBSD
llvm-svn: 360767
Nico Weber [Wed, 15 May 2019 12:08:45 +0000 (12:08 +0000)]
gn build: Merge r360671
llvm-svn: 360766
Simon Atanasyan [Wed, 15 May 2019 12:05:27 +0000 (12:05 +0000)]
[mips] LLVM and GAS now use same instructions for CFA Definition. NFCI
LLVM previously used `DW_CFA_def_cfa` instruction in .eh_frame to set
the register and offset for current CFA rule. We change it to
`DW_CFA_def_cfa_register` which is the same one used by GAS that only
changes the register but keeping the old offset.
Patch by Mirko Brkusanin.
Differential Revision: https://reviews.llvm.org/D61899
llvm-svn: 360765
Nico Weber [Wed, 15 May 2019 12:03:10 +0000 (12:03 +0000)]
gn build: Run `git ls-files '*.gn' '*.gni' | xargs llvm/utils/gn/gn.py format`
llvm-svn: 360764
Tim Northover [Wed, 15 May 2019 12:01:04 +0000 (12:01 +0000)]
arm64_32: add some unittests that were in the wrong commit.
Accidentally dropped them when committing the arm64_32 binutils support.
There's no change to real code.
llvm-svn: 360763
Fangrui Song [Wed, 15 May 2019 11:23:54 +0000 (11:23 +0000)]
Delete unnecessary copy ctors/copy assignment operators
It's the simplest and gives the cleanest semantics.
llvm-svn: 360762
Michal Gorny [Wed, 15 May 2019 10:48:55 +0000 (10:48 +0000)]
[lldb] [lit] Pass --mode=compile to fix compiler-full-path.test
Pass '--mode=compile' to fix compiler-full-path.test failure on NetBSD
buildbot (apparently due to lack of 'link' executable).
Fixes r360355. Acked by Pavel Labath.
llvm-svn: 360761
Gabor Marton [Wed, 15 May 2019 10:29:48 +0000 (10:29 +0000)]
[ASTImporter] Use llvm::Expected and Error in the importer API
Summary:
This is the final phase of the refactoring towards using llvm::Expected
and llvm::Error in the ASTImporter API.
This involves the following:
- remove old Import functions which returned with a pointer,
- use the Import_New functions (which return with Err or Expected) everywhere
and handle their return value
- rename Import_New functions to Import
This affects both Clang and LLDB.
Reviewers: shafik, teemperor, aprantl, a_sidorin, balazske, a.sidorin
Subscribers: rnkovacs, dkrupp, Szelethus, gamesh411, cfe-commits, lldb-commits
Tags: #clang, #lldb
Differential Revision: https://reviews.llvm.org/D61438
llvm-svn: 360760
Roman Lebedev [Wed, 15 May 2019 10:24:38 +0000 (10:24 +0000)]
[NFC][InstCombine] Regenerate trunc.ll test
llvm-svn: 360759
Florian Hahn [Wed, 15 May 2019 10:05:49 +0000 (10:05 +0000)]
[LV] Move getScalarizationOverhead and vector call cost computations to CM. (NFC)
This reduces the number of parameters we need to pass in and they seem a
natural fit in LoopVectorizationCostModel. Also simplifies things for
D59995.
As a follow up refactoring, we could only expose a expose a
shouldUseVectorIntrinsic() helper in LoopVectorizationCostModel, instead
of calling getVectorCallCost/getVectorIntrinsicCost in
InnerLoopVectorizer/VPRecipeBuilder.
Reviewers: Ayal, hsaito, dcaballe, rengolin
Reviewed By: rengolin
Differential Revision: https://reviews.llvm.org/D61638
llvm-svn: 360758
Fangrui Song [Wed, 15 May 2019 09:15:13 +0000 (09:15 +0000)]
Group forward declarations in one namespace lldb_private {}
llvm-svn: 360757
Stefan Granitz [Wed, 15 May 2019 08:59:02 +0000 (08:59 +0000)]
[CMake] Add error to clarify that lldb requires libcxx
Summary:
This adds a specific error message to clarify that lldb requires libcxx when
built together with clang on macOS. In addition, the lldb building docs are also
updated.
Fixes https://bugs.llvm.org/show_bug.cgi?id=41866
Reviewers: sgraenitz, JDevlieghere, EricWF
Reviewed By: sgraenitz
Subscribers: mgorny, lldb-commits
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D61877
llvm-svn: 360756
Clement Courbet [Wed, 15 May 2019 08:21:18 +0000 (08:21 +0000)]
[[DAGCombiner][NFC] Add a comment.
As suggested in D61846.
llvm-svn: 360755
Pavel Labath [Wed, 15 May 2019 07:45:40 +0000 (07:45 +0000)]
DWARF: s/CompileUnit/Unit/ in DWARFFormValue
The class has been converted to use DWARFUnit, but a number of uses of
the words compile unit remained. This removes all such references
Get/SetCompileUnit becomes Get/SetUnit, and m_cu becomes m_unit.
llvm-svn: 360754
Jonas Devlieghere [Wed, 15 May 2019 05:39:41 +0000 (05:39 +0000)]
Replace assert with static_assert here applicable.
Replaces assert() with static_assert() if the condition is can be
evaluated at compile time.
llvm-svn: 360753
Jonas Devlieghere [Wed, 15 May 2019 05:31:14 +0000 (05:31 +0000)]
Mark private unimplemented functions as deleted
Applies modernize-use-equals-delete to the LLDB code base and removes
the now redundant comments.
llvm-svn: 360751
Jonas Devlieghere [Wed, 15 May 2019 04:42:30 +0000 (04:42 +0000)]
[Docs] Document lldb-dotest
Document the lldb-dotest binary.
llvm-svn: 360748
Craig Topper [Wed, 15 May 2019 04:15:46 +0000 (04:15 +0000)]
[X86] Use OR32mi8Locked instead of LOCK_OR32mi8 in emitLockedStackOp.
They encode the same way, but OR32mi8Locked sets hasUnmodeledSideEffects set
which should be stronger than the mayLoad/mayStore on LOCK_OR32mi8. I think
this makes sense since we are using it as a fence.
This also seems to hide the operation from the speculative load hardening pass
so I've reverted r360511.
llvm-svn: 360747
Saleem Abdulrasool [Wed, 15 May 2019 03:57:07 +0000 (03:57 +0000)]
Revert "build: use the correct variable"
This reverts commit
b5a8abd57f23e2f621d5ceb0f64f1bb8f9579c3f. This
should not be needed as the lldb-server tool will add
`LLDB_CAN_USE_LLDB_SERVER` which will never be set to true on Windows.
llvm-svn: 360745
Michal Gorny [Wed, 15 May 2019 03:32:47 +0000 (03:32 +0000)]
[lldb] [lit] Fix whitespace in matches for remaining AVX512 tests
llvm-svn: 360744
Fangrui Song [Wed, 15 May 2019 03:08:21 +0000 (03:08 +0000)]
Fix 2-field llvm.global_ctors `REQUIRES: asserts` tests after rL360742
llvm-svn: 360743
Fangrui Song [Wed, 15 May 2019 02:35:32 +0000 (02:35 +0000)]
[IR] Disallow llvm.global_ctors and llvm.global_dtors of the 2-field form in textual format
The 3-field form was introduced by D3499 in 2014 and the legacy 2-field
form was planned to be removed in LLVM 4.0
For the textual format, this patch migrates the existing 2-field form to
use the 3-field form and deletes the compatibility code.
test/Verifier/global-ctors-2.ll checks we have a friendly error message.
For bitcode, lib/IR/AutoUpgrade UpgradeGlobalVariables will upgrade the
2-field form (add i8* null as the third field).
Reviewed By: rnk, dexonsmith
Differential Revision: https://reviews.llvm.org/D61547
llvm-svn: 360742
Alex Langford [Wed, 15 May 2019 01:46:45 +0000 (01:46 +0000)]
[Target] Generalize some behavior in Thread
Summary:
I don't think there's a good reason for this behavior to be considered
ObjC-specific. We can generalize this.
Differential Revision: https://reviews.llvm.org/D61776
llvm-svn: 360741