Jake Ehrlich [Tue, 15 May 2018 20:53:53 +0000 (20:53 +0000)]
[llvm-objcopy] Add --only-keep-debug as a noop
This option just keeps being a problem and really needs to be implemented
in some fashion. Implementing it properly requires some kind of
"replaceSectionReference" method because all the existing links need to be
maintained. The desired behavior is just for allocated sections to become
NOBITS but actually implementing that is rather tricky due to the current
design of llvm-objcopy. However converting allocated sections to NOBITS is
just an optimization and not something debuggers need. Debuggers can debug
a stripped executable and take an unstripped executable for that stripped
executable as input. Additionally allocated sections account for a very
small part of debug binaries so this optimization is quite small. I propose
that for the time being we implement this as a NOP so that people can use
llvm-objcopy where they need to, just in a sub-optimal way.
This option has already blocked a lot of people and its currently blocking me.
llvm-svn: 332396
Evandro Menezes [Tue, 15 May 2018 20:41:12 +0000 (20:41 +0000)]
[AArch64] Improve single vector lane unscaled stores
When storing the 0th lane of a vector, use a simpler and usually more
efficient scalar store instead. In this case, also using the unscaled
offset.
Differential revision: https://reviews.llvm.org/D46762
llvm-svn: 332394
Sanjay Patel [Tue, 15 May 2018 20:34:09 +0000 (20:34 +0000)]
[InstCombine] add more tests for binop-shuffle; NFC
The splat pattern is part of PR37463:
https://bugs.llvm.org/show_bug.cgi?id=37463
llvm-svn: 332393
Nirav Dave [Tue, 15 May 2018 20:31:53 +0000 (20:31 +0000)]
[DAGCombine] Move load checks on store of loads into candidate
search. NFCI.
Migrate single-use and non-volatility, non-indexed requirements on
stores of immediate store values to candidate collection pass from
later stage.
llvm-svn: 332392
Peter Collingbourne [Tue, 15 May 2018 20:21:58 +0000 (20:21 +0000)]
Nios2: Unbreak build.
llvm-svn: 332391
Matt Davis [Tue, 15 May 2018 20:21:04 +0000 (20:21 +0000)]
[llvm-mca] Introduce a pipeline Stage class and FetchStage.
Summary:
This is just an idea, really two ideas. I expect some push-back,
but I realize that posting a diff is the most comprehensive way to express
these concepts.
This patch introduces a Stage class which represents the
various stages of an instruction pipeline. As a start, I have created a simple
FetchStage that is based on existing logic for how MCA produces
instructions, but now encapsulated in a Stage. The idea should become more concrete
once we introduce additional stages. The idea being, that when a stage completes,
the next stage in the pipeline will be executed. Stages are chained together
as a singly linked list to closely model a real pipeline. For now there is only one stage,
so the stage-to-stage flow of instructions isn't immediately obvious.
Eventually, Stage will also handle event notifications, but that functionality
is not complete, and not destined for this patch. Ideally, an interested party
can register for notifications from a particular stage. Callbacks will be issued to
these listeners at various points in the execution of the stage.
For now, eventing functionality remains similar to what it has been in mca::Backend.
We will be building-up the Stage class as we move on, such as adding debug output.
This patch also removes the unique_ptr<Instruction> return value from
InstrBuilder::createInstruction. An Instruction pointer is still produced,
but now it's up to the caller to decide how that item should be managed post-allocation
(e.g., smart pointer). This allows the Fetch stage to create instructions and
manage the lifetime of those instructions as it wishes, and not have to be bound to any
specific managed pointer type. Other callers of createInstruction might have different
requirements, and thus can manage the pointer to fit their needs. Another idea would be to push the
ownership to the RCU.
Currently, the FetchStage will wrap the Instruction
pointer in a shared_ptr. This allows us to remove the Instruction container in
Backend, which was probably going to disappear, or move, at some point anyways.
Note that I did run these changes through valgrind, to make sure we are not leaking
memory. While the shared_ptr comes with some additional overhead it relieves us
from having to manage a list of generated instructions, and/or make lookup calls
to remove the instructions.
I realize that both the Stage class and the Instruction pointer management
(mentioned directly above) are separate but related ideas, and probably should
land as separate patches; I am happy to do that if either idea is decent.
The main reason these two ideas are together is that
Stage::execute() can mutate an InstRef. For the fetch stage, the InstRef is populated
as the primary action of that stage (execute()). I didn't want to change the Stage interface
to support the idea of generating an instruction. Ideally, instructions are to
be pushed through the pipeline. I didn't want to draw too much of a
specialization just for the fetch stage. Excuse the word-salad.
Reviewers: andreadb, courbet, RKSimon
Reviewed By: andreadb
Subscribers: llvm-commits, mgorny, javed.absar, tschuett, gbedwell
Differential Revision: https://reviews.llvm.org/D46741
llvm-svn: 332390
Chandler Carruth [Tue, 15 May 2018 20:16:57 +0000 (20:16 +0000)]
[x86][eflags] Fix PR37431 by teaching the EFLAGS copy lowering to
specially handle SETB_C* pseudo instructions.
Summary:
While the logic here is somewhat similar to the arithmetic lowering, it
is different enough that it made sense to have its own function.
I actually tried a bunch of different optimizations here and none worked
well so I gave up and just always do the arithmetic based lowering.
Looking at code from the PR test case, we actually pessimize a bunch of
code when generating these. Because SETB_C* pseudo instructions clobber
EFLAGS, we end up creating a bunch of copies of EFLAGS to feed multiple
SETB_C* pseudos from a single set of EFLAGS. This in turn causes the
lowering code to ruin all the clever code generation that SETB_C* was
hoping to achieve. None of this is needed. Whenever we're generating
multiple SETB_C* instructions from a single set of EFLAGS we should
instead generate a single maximally wide one and extract subregs for all
the different desired widths. That would result in substantially better
code generation. But this patch doesn't attempt to address that.
The test case from the PR is included as well as more directed testing
of the specific lowering pattern used for these pseudos.
Reviewers: craig.topper
Subscribers: sanjoy, mcrosier, llvm-commits, hiraditya
Differential Revision: https://reviews.llvm.org/D46799
llvm-svn: 332389
Benjamin Kramer [Tue, 15 May 2018 20:08:15 +0000 (20:08 +0000)]
Use perfect forwarding to deduplicate code in unit test. NFC.
llvm-svn: 332388
Konstantin Zhuravlyov [Tue, 15 May 2018 19:32:47 +0000 (19:32 +0000)]
AMDGPU: Fix v_dot{4, 8}* instruction encoding
Differential Revision: https://reviews.llvm.org/D46848
llvm-svn: 332387
Martin Storsjo [Tue, 15 May 2018 19:21:28 +0000 (19:21 +0000)]
[llvm-rc] Add support for the optional CLASS statement for dialogs
Differential Revision: https://reviews.llvm.org/D46875
llvm-svn: 332386
Michael Zolotukhin [Tue, 15 May 2018 18:40:29 +0000 (18:40 +0000)]
[MemorySSA] Don't sort IDF blocks.
Summary:
After r332167 we started to sort the IDF blocks inside IDF calculation, so
there is no need to re-sort them on the user site. The test changes are due to
a slightly different order we're using now (originally we used DFSInNumber and
now the blocks are sorted by a pair (LevelFromRoot, DFSInNumber)).
Reviewers: dberlin, mgrang
Subscribers: Prazek, hiraditya, george.burgess.iv, llvm-commits
Differential Revision: https://reviews.llvm.org/D46899
llvm-svn: 332385
Sunil Srivastava [Tue, 15 May 2018 18:28:42 +0000 (18:28 +0000)]
Fixed some rtti-options tests.
Certain tests in rtti-options.cpp are not really testing anything because they are testing for the absence of -frtti option to the cc1 process. Since the cc1 process does not take -frtti option, these tests are passing tautologically.
The RTTI mode is enabled by default in cc1, and -fno-rtti disables it. Therefore the correct way to check for enabling of RTTI is to check for the absence of -fno-rtti to cc1, and the correct way to check for disabling of RTTI is to check for the presence of -fno-rtti to cc1.
This patch fixes those tests.
Differential Revision: https://reviews.llvm.org/D46836
llvm-svn: 332384
Krzysztof Parzyszek [Tue, 15 May 2018 18:15:59 +0000 (18:15 +0000)]
[Hexagon] Add driver options for subtarget features
llvm-svn: 332383
Yan Zhang [Tue, 15 May 2018 18:13:51 +0000 (18:13 +0000)]
add AR to acronyms of clang-tidy property check
Reviewers: hokein, benhamilton
Reviewed By: benhamilton
Subscribers: klimek, cfe-commits
Differential Revision: https://reviews.llvm.org/D46895
llvm-svn: 332382
Andrea Di Biagio [Tue, 15 May 2018 18:11:45 +0000 (18:11 +0000)]
[llvm-mca] use a formatted_raw_ostream to insert padding and get rid of tabs. NFC
llvm-svn: 332381
Alexey Bataev [Tue, 15 May 2018 18:01:01 +0000 (18:01 +0000)]
[OPENMP, NVPTX] Do not globalize variables with reference/pointer types.
In generic data-sharing mode we do not need to globalize
variables/parameters of reference/pointer types. They already are placed
in the global memory.
llvm-svn: 332380
Tom Stellard [Tue, 15 May 2018 17:57:09 +0000 (17:57 +0000)]
AMDGPU/GlobalISel: Implement select() for G_FCONSTANT
Summary: Also clean up G_CONSTANT selection.
Reviewers: arsenm, nhaehnle
Subscribers: kzhuravl, wdng, yaxunl, rovka, kristof.beyls, dstuttard, tpr, t-tye, llvm-commits
Differential Revision: https://reviews.llvm.org/D46170
llvm-svn: 332379
Sam McCall [Tue, 15 May 2018 17:43:27 +0000 (17:43 +0000)]
[clangd] Extract scoring/ranking logic, and shave yaks.
Summary:
Code completion scoring was embedded in CodeComplete.cpp, which is bad:
- awkward to test. The mechanisms (extracting info from index/sema) can be
unit-tested well, the policy (scoring) should be quantitatively measured.
Neither was easily possible, and debugging was hard.
The intermediate signal struct makes this easier.
- hard to reuse. This is a bug in workspaceSymbols: it just presents the
results in the index order, which is not sorted in practice, it needs to rank
them!
Also, index implementations care about scoring (both query-dependent and
independent) in order to truncate result lists appropriately.
The main yak shaved here is the build() function that had 3 variants across
unit tests is unified in TestTU.h (rather than adding a 4th variant).
Reviewers: ilya-biryukov
Subscribers: klimek, mgorny, ioeric, MaskRay, jkorous, mgrang, cfe-commits
Differential Revision: https://reviews.llvm.org/D46524
llvm-svn: 332378
Konstantin Zhuravlyov [Tue, 15 May 2018 17:39:13 +0000 (17:39 +0000)]
AMDGPU: Add disasm tests for deep learning instructions + fix v_fmac_f32 disasm
Differential Revision: https://reviews.llvm.org/D46853
llvm-svn: 332377
Simon Pilgrim [Tue, 15 May 2018 17:36:49 +0000 (17:36 +0000)]
[X86] Split WriteCvtF2F into F32->F64 and F64->F32 scheduler classes
BtVer2 - Fixes schedules for (V)CVTPS2PD instructions
A lot of the Intel models still have too many InstRW overrides for these new classes - this needs cleaning up but I wanted to get the classes in first
llvm-svn: 332376
Sanjay Patel [Tue, 15 May 2018 17:14:23 +0000 (17:14 +0000)]
[InstCombine] fix binop-of-shuffles to check uses
llvm-svn: 332375
Han Shen [Tue, 15 May 2018 17:02:35 +0000 (17:02 +0000)]
[lld] Mitigate relocation overflow [part 1 of 2].
This CL places .dynsym and .dynstr at the beginning of SHF_ALLOC
sections. We do this to mitigate the possibility that huge .dynsym and
.dynstr sections placed between ro-data and text sections cause
relocation overflow.
Differential Revision: https://reviews.llvm.org/D45788
llvm-svn: 332374
Sanjay Patel [Tue, 15 May 2018 16:47:47 +0000 (16:47 +0000)]
[InstCombine] add multi-use shuffle tests and regenerate checks; NFC
llvm-svn: 332373
Fangrui Song [Tue, 15 May 2018 16:40:54 +0000 (16:40 +0000)]
Remove \brief commands from doxygen comments.
Summary: This is similar to D46290 D46320.
Reviewers: ruiu, grimar
Subscribers: mehdi_amini, llvm-commits
Differential Revision: https://reviews.llvm.org/D46861
llvm-svn: 332372
Nicola Zaghen [Tue, 15 May 2018 16:37:45 +0000 (16:37 +0000)]
[clang-tools-extra] Update uses of DEBUG macro to LLVM_DEBUG.
The DEBUG() macro is very generic so it might clash with other projects.
The renaming was done as follows:
- git grep -l 'DEBUG' | xargs sed -i 's/\bDEBUG\s\?(/LLVM_DEBUG(/g'
- git diff -U0 master | ../clang/tools/clang-format/clang-format-diff.py -i -p1 -style LLVM
Differential Revision: https://reviews.llvm.org/D44976
llvm-svn: 332371
Nico Weber [Tue, 15 May 2018 16:37:00 +0000 (16:37 +0000)]
update two comments as suggested on https://reviews.llvm.org/D46843
llvm-svn: 332370
Krzysztof Parzyszek [Tue, 15 May 2018 16:32:24 +0000 (16:32 +0000)]
[Hexagon] Remove unused function from subtarget
llvm-svn: 332369
Nico Weber [Tue, 15 May 2018 16:30:30 +0000 (16:30 +0000)]
Rename three cxx files in unittests to cpp.
LLVM uses cpp as its C++ file extension, these are the only three cxx file in
the monorepo. These files apparently were called to escape a CMake check -- use
the LLVM_OPTIONAL_SOURCES mechanism that's meant as an escape for this case
instead.
No intended behavior change.
https://reviews.llvm.org/D46843
llvm-svn: 332368
Geoff Berry [Tue, 15 May 2018 16:27:34 +0000 (16:27 +0000)]
[AArch64] Fix mir test case liveins info.
The test case added in r332265 had incomplete livein information which
was caught by the EXPENSIVE_CHECKS bot. Fix the livein information and
add -verify-machineinstrs to the test case.
llvm-svn: 332367
Eric Liu [Tue, 15 May 2018 16:22:43 +0000 (16:22 +0000)]
[clangd] Log error message instead write to errs(). NFC
llvm-svn: 332366
Krzysztof Parzyszek [Tue, 15 May 2018 16:13:52 +0000 (16:13 +0000)]
[Hexagon] Remove unused flag from subtarget and (non)corresponding test
llvm-svn: 332365
Simon Dardis [Tue, 15 May 2018 16:05:04 +0000 (16:05 +0000)]
[mips] Mark select instructions correctly
Reviewers: atanasyan, abeserminji, smaksimovic
Differential Revision: https://reviews.llvm.org/D46702
llvm-svn: 332364
Eric Liu [Tue, 15 May 2018 15:29:32 +0000 (15:29 +0000)]
[clangd] Populate #include insertions as additional edits in completion items.
Summary:
o Remove IncludeInsertion LSP command.
o Populate include insertion edits synchromously in completion items.
o Share the code completion compiler instance and precompiled preamble to get existing inclusions in main file.
o Include insertion logic lives only in CodeComplete now.
o Use tooling::HeaderIncludes for inserting new includes.
o Refactored tests.
Reviewers: sammccall
Reviewed By: sammccall
Subscribers: klimek, ilya-biryukov, MaskRay, jkorous, cfe-commits
Differential Revision: https://reviews.llvm.org/D46497
llvm-svn: 332363
Eric Liu [Tue, 15 May 2018 15:23:53 +0000 (15:23 +0000)]
[clangd] Remove LSP command-based #include insertion.
Summary:
clangd will populate #include insertions as addtionalEdits in completion items.
The code completion tests in ClangdServerTest will be added back in D46497.
Reviewers: ilya-biryukov, sammccall
Reviewed By: ilya-biryukov
Subscribers: klimek, MaskRay, jkorous, cfe-commits
Differential Revision: https://reviews.llvm.org/D46676
llvm-svn: 332362
Andrea Di Biagio [Tue, 15 May 2018 15:18:05 +0000 (15:18 +0000)]
[llvm-mca] Strip leading tabs and spaces from instruction strings before printing. NFC
llvm-svn: 332361
Gheorghe-Teodor Bercea [Tue, 15 May 2018 15:16:43 +0000 (15:16 +0000)]
[OpenMP][libomptarget] Add function for checking SPMD mode
Summary: Add function to the NVPTX libomptarget library that will return true if the current target region is being executed in SPMD mode.
Reviewers: ABataev, grokos, carlo.bertolli, caomhin
Reviewed By: grokos
Subscribers: guansong, openmp-commits
Differential Revision: https://reviews.llvm.org/D46840
llvm-svn: 332360
Philip Pfaffe [Tue, 15 May 2018 14:53:25 +0000 (14:53 +0000)]
[SI] Create Scop Name lazily
Summary: Creating the Scop name is expensive, because creating the
Region name it's derived from is expensive. So create the name lazily,
because getName() is actually called rarely.
This is a reiteration of r328666, which introduced a use-after-free and
got reverted in r331363.
Differential Revision: https://reviews.llvm.org/D46868
llvm-svn: 332359
Sanjay Patel [Tue, 15 May 2018 14:16:24 +0000 (14:16 +0000)]
[DAG] propagate FMF for all FPMathOperators
This is a simple hack based on what's proposed in D37686, but we can extend it if needed in follow-ups.
It gets us most of the FMF functionality that we want without adding any state bits to the flags. It
also intentionally leaves out non-FMF flags (nsw, etc) to minimize the patch.
It should provide a superset of the functionality from D46563 - the extra tests show propagation and
codegen diffs for fcmp, vecreduce, and FP libcalls.
The PPC log2() test shows the limits of this most basic approach - we only applied 'afn' to the last
node created for the call. AFAIK, there aren't any libcall optimizations based on the flags currently,
so that shouldn't make any difference.
Differential Revision: https://reviews.llvm.org/D46854
llvm-svn: 332358
Simon Pilgrim [Tue, 15 May 2018 14:12:32 +0000 (14:12 +0000)]
[X86] Split off F16C WriteCvtPH2PS/WriteCvtPS2PH scheduler classes
Btver2 - VCVTPH2PSYrm needs to double pump the AGU
Broadwell - missing VCVTPS2PH*mr stores extra latency
Allows us to remove the WriteCvtF2FSt conversion store class
llvm-svn: 332357
Paul Semel [Tue, 15 May 2018 14:09:37 +0000 (14:09 +0000)]
[llvm-objcopy] Add --keep-symbol (-K) option
This option permits to explicitly keep the specified
symbol so that it doesn't get removed.
Differential Revision: https://reviews.llvm.org/D46819
llvm-svn: 332356
Philip Pfaffe [Tue, 15 May 2018 13:43:42 +0000 (13:43 +0000)]
[arc] Remove unittesting from arcconfig
Summary:
Unittests aren't working, and I don't think they ever were. Just remove
them, so that we don't have to write `arc --nounit` all the time.
Reviewers: grosser, Meinersbur, bollu
Subscribers: bollu, pollydev, llvm-commits
Differential Revision: https://reviews.llvm.org/D46869
llvm-svn: 332355
Eric Liu [Tue, 15 May 2018 13:43:20 +0000 (13:43 +0000)]
Fix broken asan Support tests
The asan failures were caught in google internal asan tests after r332311
o Make StackOption support cl::list
o Rememeber to removeArguments for cl::alias in tests.
llvm-svn: 332354
Pavel Labath [Tue, 15 May 2018 13:42:26 +0000 (13:42 +0000)]
Reapply "Remove Process references from the Host module"
This re-lands r332250/D46395, after fixing Mac build errors.
llvm-svn: 332353
Nicola Zaghen [Tue, 15 May 2018 13:37:17 +0000 (13:37 +0000)]
[polly] Update uses of DEBUG macro to LLVM_DEBUG.
The DEBUG() macro is very generic so it might clash with other projects.
The renaming was done as follows:
- git grep -l 'DEBUG' | xargs sed -i 's/\bDEBUG\s\?(/LLVM_DEBUG(/g'
- git diff -U0 master | ../clang/tools/clang-format/clang-format-diff.py -i -p1 -style LLVM
Differential Revision: https://reviews.llvm.org/D44978
llvm-svn: 332352
Nicola Zaghen [Tue, 15 May 2018 13:36:20 +0000 (13:36 +0000)]
[lld] Update uses of DEBUG macro to LLVM_DEBUG.
The DEBUG() macro is very generic so it might clash with other projects.
The renaming was done as follows:
- git grep -l 'DEBUG' | xargs sed -i 's/\bDEBUG\s\?(/LLVM_DEBUG(/g'
- git diff -U0 master | ../clang/tools/clang-format/clang-format-diff.py -i -p1 -style LLVM
Differential Revision: https://reviews.llvm.org/D44977
llvm-svn: 332351
Nicola Zaghen [Tue, 15 May 2018 13:30:56 +0000 (13:30 +0000)]
[clang] Update uses of DEBUG macro to LLVM_DEBUG.
The DEBUG() macro is very generic so it might clash with other projects.
The renaming was done as follows:
- git grep -l 'DEBUG' | xargs sed -i 's/\bDEBUG\s\?(/LLVM_DEBUG(/g'
- git diff -U0 master | ../clang/tools/clang-format/clang-format-diff.py -i -p1 -style LLVM
Explicitly avoided changing the strings in the clang-format tests.
Differential Revision: https://reviews.llvm.org/D44975
llvm-svn: 332350
Pavel Labath [Tue, 15 May 2018 13:24:10 +0000 (13:24 +0000)]
Reapply "DWARFVerifier: Check "completeness" of .debug_names section"
This is a resubmit of r331868 (D46583), which was reverted due to
failures on the PS4 bot.
These have been resolved with r332246/D46748.
llvm-svn: 332349
Clement Courbet [Tue, 15 May 2018 13:07:05 +0000 (13:07 +0000)]
[llvm-exegesis] Split AsmTemplate.Name into components.
Summary:
AsmTemplate becomes IntructionBenchmarkKey, which has three components.
This allows retreiving the opcode for analysis.
Reviewers: gchatelet
Subscribers: tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D46873
llvm-svn: 332348
Simon Pilgrim [Tue, 15 May 2018 12:50:06 +0000 (12:50 +0000)]
[llvm-mca][x86] Add F16C instruction tests
llvm-svn: 332347
Clement Courbet [Tue, 15 May 2018 12:38:06 +0000 (12:38 +0000)]
Fix compilation under pre-c++14 gccs.
llvm-svn: 332346
Clement Courbet [Tue, 15 May 2018 12:27:36 +0000 (12:27 +0000)]
Fix r332344: only the native target is linked.
llvm-svn: 332345
Clement Courbet [Tue, 15 May 2018 12:08:00 +0000 (12:08 +0000)]
[llvm-exegesis] Add an analysis mode.
Summary:
The analysis mode gives the user a clustered view of the measurement results.
Next steps are (requires the split ok AsmTemplate.Name into {mnemonic, mode}):
- Show the sched class.
- Highlight any inconsistencies with the checked-in data.
Reviewers: gchatelet
Subscribers: mgorny, llvm-commits, tschuett
Differential Revision: https://reviews.llvm.org/D46865
llvm-svn: 332344
Rainer Orth [Tue, 15 May 2018 11:36:00 +0000 (11:36 +0000)]
[Solaris] Only define _REENTRANT if -pthread
When looking at lib/Basic/Targets/OSTargets.h, I noticed that _REENTRANT is defined
unconditionally on Solaris, unlike all other targets and what either Studio cc (only define
it with -mt) or gcc (only define it with -pthread) do.
This patch follows that lead.
Differential Revision: https://reviews.llvm.org/D41241
llvm-svn: 332343
whitequark [Tue, 15 May 2018 11:31:07 +0000 (11:31 +0000)]
[MergeFunctions] Fix merging of small weak functions
When two interposable functions are merged, we cannot replace
uses and have to emit calls to a common internal function. However,
writeThunk() will not actually emit a thunk if the function is too
small. This leaves us in a broken state where mergeTwoFunctions
already rewired the functions, but writeThunk doesn't do anything.
This patch changes the implementation so that:
* writeThunk() does just that.
* The direct replacement of calls is moved into mergeTwoFunctions()
into the non-interposable case only.
* isThunkProfitable() is extracted and will be called for
the non-iterposable case always, and in the interposable case
only if uses are still left after replacement.
This issue has been introduced in https://reviews.llvm.org/D34806,
where the code for checking thunk profitability has been moved.
Differential Revision: https://reviews.llvm.org/D46804
Reviewed By: whitequark
llvm-svn: 332342
Simon Dardis [Tue, 15 May 2018 11:21:07 +0000 (11:21 +0000)]
[mips] Fix formatting of floating point conversion patterns
llvm-svn: 332341
Simon Dardis [Tue, 15 May 2018 11:18:24 +0000 (11:18 +0000)]
[mips] Add disassembly support for comparison instructions
llvm-svn: 332340
Simon Dardis [Tue, 15 May 2018 11:10:30 +0000 (11:10 +0000)]
[mips] Fix predicates of mfc1, mtc1 instructions
Reviewers: atanasyan, abeserminji, smaksimovic
Differential Revision: https://reviews.llvm.org/D46692
llvm-svn: 332339
Aleksei Sidorin [Tue, 15 May 2018 11:09:07 +0000 (11:09 +0000)]
[ASTImporter] Extend lookup logic in class templates
During import of a class template, lookup may find a forward
declaration and structural match falsely reports equivalency
between a forward decl and a definition. The result is that
some definitions are not imported if we had imported a forward
decl previously. This patch gives a fix.
Patch by Gabor Marton!
Differential Revision: https://reviews.llvm.org/D46353
llvm-svn: 332338
Andrea Di Biagio [Tue, 15 May 2018 10:30:39 +0000 (10:30 +0000)]
[llvm-mca] Remove unused include header files. NFC
Also, run clang-format on RetireControlUnit.cpp.
llvm-svn: 332337
Artur Gainullin [Tue, 15 May 2018 10:24:12 +0000 (10:24 +0000)]
[X86] Improve unsigned saturation downconvert detection.
Summary:
New unsigned saturation downconvert patterns detection was implemented in
X86 Codegen:
(truncate (smin (smax (x, C1), C2)) to dest_type),
where C1 >= 0 and C2 is unsigned max of destination type.
(truncate (smax (smin (x, C2), C1)) to dest_type)
where C1 >= 0, C2 is unsigned max of destination type and C1 <= C2.
These two patterns are equivalent to:
(truncate (umin (smax(x, C1), unsigned_max_of_dest_type)) to dest_type)
Reviewers: RKSimon
Subscribers: llvm-commits, a.elovikov
Differential Revision: https://reviews.llvm.org/D45315
llvm-svn: 332336
Hans Wennborg [Tue, 15 May 2018 10:19:24 +0000 (10:19 +0000)]
Make ast-print-record-decl.c pass on Windows after r332314
It was failing because on Windows, -ast-print prints
__single_inheritance(1) before T1.
Adding a triple is a stop-gap fix until it can be fixed properly.
llvm-svn: 332335
Andrea Di Biagio [Tue, 15 May 2018 09:31:32 +0000 (09:31 +0000)]
[llvm-mca] Add file header to RetireControlUnit.cpp.
Strictly speaking, this is not necessary for .cpp files. However, other .cpp
files from this same tool have it. This also matches what we do in other tools.
llvm-svn: 332334
Peter Smith [Tue, 15 May 2018 08:57:21 +0000 (08:57 +0000)]
[ELF] Implement --keep-unique option
The --keep-unique <symbol> option is taken from gold. The intention is that
<symbol> will be prevented from being folded by ICF. Although not
specifically mentioned in the documentation <symbol> only matches
global symbols, with a warning if the symbol is not found.
The implementation finds the Section defining <symbol> and removes it from
the set of sections considered for ICF.
Differential Revision: https://reviews.llvm.org/D46755
llvm-svn: 332332
Clement Courbet [Tue, 15 May 2018 07:40:21 +0000 (07:40 +0000)]
[llvm-exegesis] InMemoryAssembler: handle return-less targets (e.g. arm).
Summary: Arm does not have a ret code per se.
Reviewers: gchatelet
Subscribers: mgorny, javed.absar, kristof.beyls, tschuett, llvm-commits
Differential Revision: https://reviews.llvm.org/D45672
llvm-svn: 332331
Clement Courbet [Tue, 15 May 2018 07:35:21 +0000 (07:35 +0000)]
[llvm-exegesis] Check perf event validity.
This was part of https://reviews.llvm.org/D46821.
Authored by Guillaume Chatelet
llvm-svn: 332330
Martin Storsjo [Tue, 15 May 2018 06:35:29 +0000 (06:35 +0000)]
[llvm-rc] Add support for parsing memory flags
Most of the handling is pretty straightforward; fetch the default
memory flags for the specific resource type before parsing the flags
and apply them on top of that, except that some flags imply others
and some flags clear more than one flag.
For icons and cursors, the flags set get passed on to all individual
single icon/cursor resources, while only some flags affect the icon/cursor
group resource.
For stringtables, the behaviour is pretty simple; the first stringtable
resource of a bundle sets the flags for the whole bundle.
The output of these tests match rc.exe byte for byte.
The actual use of these memory flags is deprecated and they have no
effect since Win16, but some resource script files may still happen
to have them in place.
Differential Revision: https://reviews.llvm.org/D46818
llvm-svn: 332329
Martin Storsjo [Tue, 15 May 2018 06:35:20 +0000 (06:35 +0000)]
[llvm-rc] Read the Planes/BitCount fields from BITMAPINFOHEADER for icons
Previously these fields were only read from this header for cursors,
while Planes was hardcoded to 1 for icons (with a comment that it was
unknown why this was needed) and BitCount was left at the value
read originally in the RESDIRENTRY.
This fixes the single byte that was differing for the icon/cursor test
compared to rc.exe.
This is based on research/testing by Nico Weber.
Differential Revision: https://reviews.llvm.org/D46816
llvm-svn: 332328
Martin Storsjo [Tue, 15 May 2018 06:34:18 +0000 (06:34 +0000)]
[MinGW] Add a LLD specific option -pdb
This allows producing pdb debug info. This is an LLD specific option
since GCC and GNU binutils doesn't support the PDB file format.
Differential Revision: https://reviews.llvm.org/D46796
llvm-svn: 332327
Craig Topper [Tue, 15 May 2018 04:26:27 +0000 (04:26 +0000)]
[X86] Add fast isel tests for some of the avx512 truncate intrinsics to match current clang codegen.
llvm-svn: 332326
JF Bastien [Tue, 15 May 2018 04:23:48 +0000 (04:23 +0000)]
[NFC] pull a function into its own lambda
As requested in D46858, pulling this function into its own lambda makes it
easier to read that part of the code and reason as to what's going on because
the scope it can be called from is extremely limited. We want to keep it as a
function because it's called from the two subsequent lines.
llvm-svn: 332325
Jan Vesely [Tue, 15 May 2018 04:22:43 +0000 (04:22 +0000)]
rootn: Use denormal path only
It's OK to either flush to 0 or return denormal result if the device
does not support denormals. See sec 7.2 and 7.5.3 of OCL specs
Use 0.0f explicitly intead of relying on GPU to flush it.
Fixes CTS on carrizo and turks
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
Acked-by: Aaron Watry <awatry@gmail.com>
Tested-by: Aaron Watry <awatry@gmail.com>
llvm-svn: 332324
JF Bastien [Tue, 15 May 2018 04:06:28 +0000 (04:06 +0000)]
[NFC] Update comments
Don't prepend function or data name before each comment. Split into its own NFC patch as requested in D46858.
llvm-svn: 332323
Craig Topper [Tue, 15 May 2018 03:17:52 +0000 (03:17 +0000)]
[X86] Revert part of r332266: Use __builtin_convertvector to replace some of the avx512 truncate builtins.
The masking doesn't work right in the backend for the ones that produce byte or word elements without avx512bw.
llvm-svn: 332322
Craig Topper [Tue, 15 May 2018 03:17:05 +0000 (03:17 +0000)]
[X86] Revert part of r332267: Remove GCCBuiltin from the intrinsics that clang stopped using in r332266.
The masking of instructions that produce byte or word elements doesn't work right without avx512bw since the generic i8/i16 select won't be legal in the SelectionDAG.
llvm-svn: 332321
Vitaly Buka [Tue, 15 May 2018 01:39:25 +0000 (01:39 +0000)]
[sanitizer] Intercept __pthread_mutex_lock and __pthread_mutex_unlock
Reviewers: eugenis, dvyukov
Subscribers: srhines, kubamracek, llvm-commits
Differential Revision: https://reviews.llvm.org/D46793
llvm-svn: 332320
Vitaly Buka [Tue, 15 May 2018 01:39:13 +0000 (01:39 +0000)]
[asan] Workaround to avoid hangs in Chromium tests
Summary:
For some reasons on Chromium when we start leak checking we get own pid as 1.
After that we see threads with PPID:0 assuming that thread is dead in infinite
loop.
To resolve particularly this case and possible issues like this, when IsAlive check failed to detect thread status, we need to limit the number of SuspendAllThreads
iterations.
Reviewers: eugenis
Subscribers: kubamracek, llvm-commits
Differential Revision: https://reviews.llvm.org/D46852
llvm-svn: 332319
Shiva Chen [Tue, 15 May 2018 01:28:50 +0000 (01:28 +0000)]
[RISCV] Define FeatureRelax and shouldForceRelocation for RISCV linker relaxation
1. Deine FeatureRelax to enable/disable linker relaxation.
2. Define shouldForceRelocation to preserve relocation types even if the fixup
can be resolved when linker relaxation enabled. This is necessary for
correctness as offsets may change during relaxation.
Differential Revision: https://reviews.llvm.org/D46674
llvm-svn: 332318
Max Kazantsev [Tue, 15 May 2018 01:21:56 +0000 (01:21 +0000)]
[NFC] Add const to method signature
llvm-svn: 332317
Kostya Serebryany [Tue, 15 May 2018 01:15:47 +0000 (01:15 +0000)]
[libFuzzer] deprecate equivalence_server
llvm-svn: 332316
Sam Clegg [Tue, 15 May 2018 00:46:43 +0000 (00:46 +0000)]
Fix debug build by adding missing dependencies on libBinaryFormat
Debug BUILD_SHARED_LIBS build was broken by rL332305
Differential Revision: https://reviews.llvm.org/D46859
llvm-svn: 332315
Joel E. Denny [Tue, 15 May 2018 00:44:14 +0000 (00:44 +0000)]
[AST] Fix printing tag decl groups in decl contexts
For example, given:
struct T1 {
struct T2 *p0;
};
-ast-print produced:
struct T1 {
struct T2;
struct T2 *p0;
};
Compiling that produces a warning that the first struct T2 declaration
does not declare anything.
Details:
A tag decl group is one or more decls that share a type specifier that
is a tag decl (that is, a struct/union/class/enum decl). Within
functions, the parser builds such a tag decl group as part of a
DeclStmt. However, in decl contexts, such as file scope or a member
list, the parser does not group together the members of a tag decl
group. Previously, detection of tag decl groups during printing was
implemented but only if the tag decl was unnamed. Otherwise, as in
the above example, the members of the group did not print together and
so sometimes introduced warnings.
This patch extends detection of tag decl groups in decl contexts to
any tag decl that is recorded in the AST as not free-standing.
Reviewed by: rsmith
Differential Revision: https://reviews.llvm.org/D45465
llvm-svn: 332314
Dean Michael Berris [Tue, 15 May 2018 00:42:36 +0000 (00:42 +0000)]
[XRay][profiler] Part 2: XRay Function Call Trie
Summary:
This is part of the larger XRay Profiling Mode effort.
This patch implements a central data structure for capturing statistics
about XRay instrumented function call stacks. The `FunctionCallTrie`
type does the following things:
* It keeps track of a shadow function call stack of XRay instrumented
functions as they are entered (function enter event) and as they are
exited (function exit event).
* When a function is entered, the shadow stack contains information
about the entry TSC, and updates the trie (or prefix tree)
representing the current function call stack. If we haven't
encountered this function call before, this creates a unique node for
the function in this position on the stack. We update the list of
callees of the parent function as well to reflect this newly found
path.
* When a function is exited, we compute statistics (TSC deltas,
function call count frequency) for the associated function(s) up the
stack as we unwind to find the matching entry event.
This builds upon the XRay `Allocator` and `Array` types in Part 1 of
this series of patches.
Depends on D45756.
Reviewers: echristo, pelikan, kpw
Reviewed By: kpw
Subscribers: llvm-commits, mgorny
Differential Revision: https://reviews.llvm.org/D45757
llvm-svn: 332313
Vedant Kumar [Tue, 15 May 2018 00:29:27 +0000 (00:29 +0000)]
[Debugify] Add -debugify-each for testing each pass in a pipeline
This adds a -debugify-each mode to opt which, when enabled, wraps each
{Module,Function}Pass in a pipeline with logic to add, check, and strip
synthetic debug info for testing purposes.
This mode can be used to test complex pipelines for debug info bugs, or
to collect statistics about the number of debug values & locations lost
throughout various stages of a pipeline.
Patch by Son Tuan Vu!
Differential Revision: https://reviews.llvm.org/D46525
llvm-svn: 332312
Keno Fischer [Mon, 14 May 2018 23:26:06 +0000 (23:26 +0000)]
[CommandLine] Error message for incorrect PositionalEatArgs usage
Summary:
bugpoint has several options specified as `PositionalEatArgs` to pass
options through to the underlying tool, e.g. `-tool-args`. The `-help`
message suggests the usage is: `-tool-args=<string>`. However, this is
misleading, because that's not how these arguments work. Rather than taking
a value, the option consumes all positional arguments until the next
recognized option (or all arguments if `--` is specified at some point).
To make this slightly clearer, instead print the help as:
```
-tool-args <string>... - <tool arguments>...
```
Additionally, add an error if the user attempts to use a `PositionalEatArgs`
argument with a value, instead of silently ignoring it. Example:
```
./bin/bugpoint -tool-args=-mpcu=skylake-avx512
bugpoint: for the -tool-args option: This argument does not take a value.
Instead, it consumes any positional arguments until the next recognized option.
```
Reviewed By: aprantl
Differential Revision: https://reviews.llvm.org/D46787
llvm-svn: 332311
Richard Trieu [Mon, 14 May 2018 23:21:48 +0000 (23:21 +0000)]
Enable control flow pruning of float overflow warnings.
Like other conversion warnings, allow float overflow warnings to be disabled
in known dead paths of template instantiation. This often occurs when a
template template type is a numeric type and the template will check the
range of the numeric type before performing the conversion.
llvm-svn: 332310
Eli Friedman [Mon, 14 May 2018 23:05:43 +0000 (23:05 +0000)]
[SCEVAffinator] Fix handling of pwaff complexity limit.
nullptr is not a valid affine expression, and none of the callers check
for null, so we eventually hit an isl error and crash.
Instead, invalidate the scop and return a constant zero.
Differential Revision: https://reviews.llvm.org/D46445
llvm-svn: 332309
Sam Clegg [Mon, 14 May 2018 23:01:16 +0000 (23:01 +0000)]
[WebAssembly] Allow signautre of entry function to be flexible
Since we a no longer using this function for the wasm start
section we don't actually care what its signature is.
Differential Revision: https://reviews.llvm.org/D46594
llvm-svn: 332308
Volodymyr Sapsai [Mon, 14 May 2018 22:49:44 +0000 (22:49 +0000)]
[c++17] Fix assertion on synthesizing deduction guides after a fatal error.
After a fatal error Sema::InstantiatingTemplate doesn't allow further
instantiation and doesn't push a CodeSynthesisContext. When we tried to
synthesize implicit deduction guides from constructors we hit the
assertion
> Assertion failed: (!CodeSynthesisContexts.empty() && "Cannot perform an instantiation without some context on the " "instantiation stack"), function SubstType, file clang/lib/Sema/SemaTemplateInstantiate.cpp, line 1580.
Fix by avoiding deduction guide synthesis if InstantiatingTemplate is invalid.
rdar://problem/
39051732
Reviewers: rsmith
Reviewed By: rsmith
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D46446
llvm-svn: 332307
Sam Clegg [Mon, 14 May 2018 22:42:33 +0000 (22:42 +0000)]
[WebAssembly] Update to match llvm changes
Subscribers: dschuff, jgravelle-google, aheejin, sunfish, llvm-commits
Differential Revision: https://reviews.llvm.org/D46849
llvm-svn: 332306
Sam Clegg [Mon, 14 May 2018 22:42:07 +0000 (22:42 +0000)]
[WebAssembly] Move toString helpers to BinaryFormat
Subscribers: dschuff, mgorny, jgravelle-google, aheejin, sunfish, llvm-commits
Differential Revision: https://reviews.llvm.org/D46847
llvm-svn: 332305
Brian Gesiak [Mon, 14 May 2018 22:36:47 +0000 (22:36 +0000)]
Revert "[Option] Fix PR37006 prefix choice in findNearest"
Summary:
This revision causes build failures in PS4 and ppc64le buildbots (for example,
http://lab.llvm.org:8011/builders/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/builds/29988).
I'll revert for now and try to diagnose the issue.
Test Plan: check-llvm check-clang
llvm-svn: 332304
Artem Dergachev [Mon, 14 May 2018 22:32:24 +0000 (22:32 +0000)]
[analyzer] Re-apply r331096 "CStringChecker: Add support for BSD strlcpy()...".
Fixed after revert in r331401.
Patch by David Carlier!
Differential Revision: https://reviews.llvm.org/D45177
llvm-svn: 332303
Keno Fischer [Mon, 14 May 2018 22:05:01 +0000 (22:05 +0000)]
[InstCombine] fix crash due to ignored addrspacecast
Summary:
Part of the InstCombine code for simplifying GEPs looks through
addrspacecasts. However, this was done by updating a variable
also used by the next transformation, for marking GEPs as
inbounds. This led to replacing a GEP with a similar instruction
in a different addrspace, which caused an assertion failure in RAUW.
This caused julia issue https://github.com/JuliaLang/julia/issues/27055
Patch by Jeff Bezanson <jeff@juliacomputing.com>
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D46722
llvm-svn: 332302
Sanjay Patel [Mon, 14 May 2018 21:53:21 +0000 (21:53 +0000)]
[AArch64] enhance test to show FMF loss; NFC
llvm-svn: 332301
George Karpenkov [Mon, 14 May 2018 21:39:54 +0000 (21:39 +0000)]
[analyzer] Extend the ObjCAutoreleaseWriteChecker to warn on captures as well
A common pattern is that the code in the block does not write into the
variable explicitly, but instead passes it to a helper function which
performs the write.
Differential Revision: https://reviews.llvm.org/D46772
llvm-svn: 332300
Brian Gesiak [Mon, 14 May 2018 21:35:00 +0000 (21:35 +0000)]
[Option] Fix PR37006 prefix choice in findNearest
Summary:
In https://bugs.llvm.org/show_bug.cgi?id=37006 Nico Weber points out a
flaw in `OptTable::findNearest`: if an option "foo"'s prefixes are "--"
and "-", then the nearest option for "--fob" will be "-foo". This is
incorrect, however, since the function is expected to return "--foo".
The bug is due to a naive loop that attempts to predetermines which
prefix is best. Instead, compute the edit distance for each prefix/name
pair.
Test Plan: `check-llvm`
Reviewers: thakis
Reviewed By: thakis
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D46776
llvm-svn: 332299
Martin Storsjo [Mon, 14 May 2018 21:32:52 +0000 (21:32 +0000)]
[ARM] Back up R4 and LR if calling the stack probe function
Differential Revision: https://reviews.llvm.org/D46777
llvm-svn: 332298
Martin Storsjo [Mon, 14 May 2018 21:32:47 +0000 (21:32 +0000)]
[llvm-rc] Add missing inputs for tag-icon-cursor.test.
This adds the missing input files used for this test, except for
the separate input files for specific error cases; matching
test input files were provided by Nico Weber.
The extra copying of files into the %t directory doesn't seem to
be necessary since that directory only ever is used for output here,
not for inputs.
Differential Revision: https://reviews.llvm.org/D46813
llvm-svn: 332297
Brock Wyma [Mon, 14 May 2018 21:21:22 +0000 (21:21 +0000)]
[CodeView] Improve debugging of virtual base class member variables
Initial support for passing the virtual base pointer offset to CodeViewDebug.
https://reviews.llvm.org/D46271
llvm-svn: 332296
Sanjay Patel [Mon, 14 May 2018 21:17:49 +0000 (21:17 +0000)]
[PowerPC] add more tests for FMF propagation; NFC
llvm-svn: 332295