Amy Huang [Thu, 10 Oct 2019 23:35:53 +0000 (23:35 +0000)]
Change test case so that it accepts backslashes in file path, in the case that the test runs on Windows
llvm-svn: 374473
Reid Kleckner [Thu, 10 Oct 2019 23:30:54 +0000 (23:30 +0000)]
Fix check-interception link error in compiler-rt debug mode
llvm-svn: 374472
Alina Sbirlea [Thu, 10 Oct 2019 23:27:21 +0000 (23:27 +0000)]
[MemorySSA] Update Phi simplification.
When simplifying a Phi to the unique value found incoming, check that
there wasn't a Phi already created to break a cycle. If so, remove it.
Resolves PR43541.
Some additional nits included.
llvm-svn: 374471
Michael Liao [Thu, 10 Oct 2019 23:05:55 +0000 (23:05 +0000)]
[tooling] Teach Tooling to understand compilation with offloading.
Summary:
- So far, we only recognize the host compilation with offloading and
skip the offloading part.
Reviewers: tra, yaxunl
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D68660
llvm-svn: 374470
Alex Langford [Thu, 10 Oct 2019 22:43:03 +0000 (22:43 +0000)]
[lldb-test] Remove unused header
llvm-svn: 374468
Yaxun Liu [Thu, 10 Oct 2019 22:43:00 +0000 (22:43 +0000)]
Fix help message for -ffp-contract
Differential Revision: https://reviews.llvm.org/D68823
llvm-svn: 374467
Richard Smith [Thu, 10 Oct 2019 22:31:17 +0000 (22:31 +0000)]
PR43629: Fix crash evaluating constexpr placement new on a subobject of
an out-of-lifetime object.
llvm-svn: 374465
Marcello Maggioni [Thu, 10 Oct 2019 21:51:30 +0000 (21:51 +0000)]
[GISel] Simplifying return from else in function. NFC
Forgot to integrate this little change in previous commit
llvm-svn: 374463
Jonas Devlieghere [Thu, 10 Oct 2019 21:49:25 +0000 (21:49 +0000)]
[Make] Remove spurious line break
This test is disabled upstream and therefore this went unnoticed.
llvm-svn: 374462
Jonas Devlieghere [Thu, 10 Oct 2019 21:49:22 +0000 (21:49 +0000)]
[test] Reduce inconsistency between lit configuration files.
Add the Python extension to the configuration files in the API directory
to match the other test suites.
llvm-svn: 374461
Craig Topper [Thu, 10 Oct 2019 21:46:52 +0000 (21:46 +0000)]
[X86] Guard against leaving a dangling node in combineTruncateWithSat.
When handling the packus pattern for i32->i8 we do a two step
process using a packss to i16 followed by a packus to i8. If the
final i8 step is a type with less than 64-bits the packus step
will return SDValue(), but the i32->i16 step might have succeeded.
This leaves the nodes from the middle step dangling.
Guard against this by pre-checking that the number of elements is
at least 8 before doing the middle step.
With that check in place this should mean the only other
case the middle step itself can fail is when SSE2 is disabled. So
add an early SSE2 check then just assert that neither the middle
or final step ever fail.
llvm-svn: 374460
Craig Topper [Thu, 10 Oct 2019 21:46:44 +0000 (21:46 +0000)]
[X86] Add test cases for packus/ssat/usat 32i32->v32i8 test cases. NFC
llvm-svn: 374459
Marcello Maggioni [Thu, 10 Oct 2019 21:46:26 +0000 (21:46 +0000)]
[GISel] Allow getConstantVRegVal() to return G_FCONSTANT values.
In GISel we have both G_CONSTANT and G_FCONSTANT, but because
in GISel we don't really have a concept of Float vs Int value
the only difference between the two is where the data originates
from.
What both G_CONSTANT and G_FCONSTANT return is just a bag of bits
with the constant representation in it.
By making getConstantVRegVal() return G_FCONSTANTs bit representation
as well we allow ConstantFold and other things to operate with
G_FCONSTANT.
Adding tests that show ConstantFolding to work on mixed G_CONSTANT
and G_FCONSTANT sources.
Differential Revision: https://reviews.llvm.org/D68739
llvm-svn: 374458
Richard Smith [Thu, 10 Oct 2019 21:40:56 +0000 (21:40 +0000)]
Fix some errors in <arm_neon.h> tests that cause them to fail with lax
vector conversions disabled.
llvm-svn: 374457
Nico Weber [Thu, 10 Oct 2019 21:34:32 +0000 (21:34 +0000)]
Revert 374450 "Fix __builtin_assume_aligned with too large values."
The test fails on Windows, with
error: 'warning' diagnostics expected but not seen:
File builtin-assume-aligned.c Line 62: requested alignment
must be
268435456 bytes or smaller; assumption ignored
error: 'warning' diagnostics seen but not expected:
File builtin-assume-aligned.c Line 62: requested alignment
must be 8192 bytes or smaller; assumption ignored
llvm-svn: 374456
Stanislav Mekhanoshin [Thu, 10 Oct 2019 21:32:41 +0000 (21:32 +0000)]
[AMDGPU] Handle undef old operand in DPP combine
It was missing an undef flag.
Differential Revision: https://reviews.llvm.org/D68813
llvm-svn: 374455
Rong Xu [Thu, 10 Oct 2019 21:30:43 +0000 (21:30 +0000)]
[ValueTracking] Improve pointer offset computation for cases of same base
This patch improves the handling of pointer offset in GEP expressions where
one argument is the base pointer. isPointerOffset() is being used by memcpyopt
where current code synthesizes consecutive 32 bytes stores to one store and
two memset intrinsic calls. With this patch, we convert the stores to one
memset intrinsic.
Differential Revision: https://reviews.llvm.org/D67989
llvm-svn: 374454
Evandro Menezes [Thu, 10 Oct 2019 21:29:10 +0000 (21:29 +0000)]
[InstCombine] Add test case for PR43617 (NFC)
Also, refactor check in `LibCallSimplifier::optimizeLog()`.
llvm-svn: 374453
Julian Lettner [Thu, 10 Oct 2019 21:24:41 +0000 (21:24 +0000)]
[lit] Break main into smaller functions
This change is purely mechanical. I will do further cleanups of
parameter usages.
Reviewed By: rnk
Differential Revision: https://reviews.llvm.org/D68830
llvm-svn: 374452
Frederic Riss [Thu, 10 Oct 2019 21:21:16 +0000 (21:21 +0000)]
TestMTCSimple: Make Makefile portable.
r374262 left out the Makefile changes needed to cross compile this test.
llvm-svn: 374451
Erich Keane [Thu, 10 Oct 2019 21:08:28 +0000 (21:08 +0000)]
Fix __builtin_assume_aligned with too large values.
Code to handle __builtin_assume_aligned was allowing larger values, but
would convert this to unsigned along the way. This patch removes the
EmitAssumeAligned overloads that take unsigned to do away with this
problem.
Additionally, it adds a warning that values greater than 1 <<29 are
ignored by LLVM.
Differential Revision: https://reviews.llvm.org/D68824
llvm-svn: 374450
Reid Kleckner [Thu, 10 Oct 2019 21:04:25 +0000 (21:04 +0000)]
Add -fgnuc-version= to control __GNUC__ and other GCC macros
I noticed that compiling on Windows with -fno-ms-compatibility had the
side effect of defining __GNUC__, along with __GNUG__, __GXX_RTTI__, and
a number of other macros for GCC compatibility. This is undesirable and
causes Chromium to do things like mix __attribute__ and __declspec,
which doesn't work. We should have a positive language option to enable
GCC compatibility features so that we can experiment with
-fno-ms-compatibility on Windows. This change adds -fgnuc-version= to be
that option.
My issue aside, users have, for a long time, reported that __GNUC__
doesn't match their expectations in one way or another. We have
encouraged users to migrate code away from this macro, but new code
continues to be written assuming a GCC-only environment. There's really
nothing we can do to stop that. By adding this flag, we can allow them
to choose their own adventure with __GNUC__.
This overlaps a bit with the "GNUMode" language option from -std=gnu*.
The gnu language mode tends to enable non-conforming behaviors that we'd
rather not enable by default, but the we want to set things like
__GXX_RTTI__ by default, so I've kept these separate.
Helps address PR42817
Reviewed By: hans, nickdesaulniers, MaskRay
Differential Revision: https://reviews.llvm.org/D68055
llvm-svn: 374449
Evgeniy Stepanov [Thu, 10 Oct 2019 20:47:22 +0000 (20:47 +0000)]
Add a missing include in test.
A fix for r373993.
llvm-svn: 374448
Alina Sbirlea [Thu, 10 Oct 2019 20:43:06 +0000 (20:43 +0000)]
[MemorySSA] Additional handling of unreachable blocks.
Summary:
Whenever we get the previous definition, the assumption is that the
recursion starts ina reachable block.
If the recursion starts in an unreachable block, we may recurse
indefinitely. Handle this case by returning LoE if the block is
unreachable.
Resolves PR43426.
Reviewers: george.burgess.iv
Subscribers: Prazek, sanjoy.google, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D68809
llvm-svn: 374447
David Greene [Thu, 10 Oct 2019 20:39:27 +0000 (20:39 +0000)]
[System Model] [TTI] Move default cache/prefetch implementations
Move the default implementations of cache and prefetch queries to
TargetTransformInfoImplBase and delete them from NoTIIImpl. This brings these
interfaces in line with how other TTI interfaces work.
Differential Revision: https://reviews.llvm.org/D68804
llvm-svn: 374446
Greg Clayton [Thu, 10 Oct 2019 20:29:11 +0000 (20:29 +0000)]
Fix a documentation warning from GSYM commit.
llvm-svn: 374445
Dimitry Andric [Thu, 10 Oct 2019 20:26:59 +0000 (20:26 +0000)]
Fix process launch failure on FreeBSD after r365761
Summary:
After rLLDB365761, and with `LLVM_ENABLE_ABI_BREAKING_CHECKS` enabled,
launching any process on FreeBSD crashes lldb with:
```
Expected<T> must be checked before access or destruction.
Expected<T> value was in success state. (Note: Expected<T> values in success mode must still be checked prior to being destroyed).
```
This is because `m_operation_thread` and `m_monitor_thread` were wrapped
in `llvm::Expected<>`, but this requires the objects to be correctly
initialized before accessing them.
To fix the crashes, use `llvm::Optional<>` for the members (as indicated
by labath), and use local variables to store the return values of
`LaunchThread` and `StartMonitoringChildProcess`. Then, only assign to
the member variables after checking if the return values indicated
success.
Reviewers: devnexen, emaste, MaskRay, mgorny
Reviewed By: devnexen
Subscribers: jfb, labath, krytarowski, lldb-commits
Differential Revision: https://reviews.llvm.org/D68723
llvm-svn: 374444
Zachary Turner [Thu, 10 Oct 2019 20:25:54 +0000 (20:25 +0000)]
[MSVC] Automatically add atlmfc folder to include and libpath.
Differential Revision: https://reviews.llvm.org/D68736
llvm-svn: 374443
Zachary Turner [Thu, 10 Oct 2019 20:25:51 +0000 (20:25 +0000)]
[PDB] Fix bug when using multiple PCH header objects with the same name.
A common pattern in Windows is to have all your precompiled headers
use an object named stdafx.obj. If you've got a project with many
different static libs, you might use a separate PCH for each one of
these.
During the final link step, a file from A might reference the PCH
object from A, but it will have the same name (stdafx.obj) as any
other PCH from another project. The only difference will be the
path. For example, A might be A/stdafx.obj while B is B/stdafx.obj.
The existing algorithm checks only the filename that was passed on
the command line (or stored in archive), but this is insufficient in
the case where relative paths are used, because depending on the
command line object file / library order, it might find the wrong
PCH object first resulting in a signature mismatch.
The fix here is to simply check whether the absolute path of the
PCH object (which is stored in the input obj file for the file that
references the PCH) *ends with* the full relative path of whatever
is specified on the command line (or is in the archive).
Differential Revision: https://reviews.llvm.org/D66431
llvm-svn: 374442
Julian Lettner [Thu, 10 Oct 2019 20:23:28 +0000 (20:23 +0000)]
[lit] Add comment explaining the LIT_OPTS env var overrides command line options
Normally, command line options override environment variables. Add
comment to state that we are doing the reverse on purpose.
llvm-svn: 374441
Jordan Rose [Thu, 10 Oct 2019 20:22:53 +0000 (20:22 +0000)]
ADT: Save a word in every StringSet entry
Add a specialization to StringMap (actually StringMapEntry) for a
value type of NoneType (the type of llvm::None), and use it for
StringSet. This'll save us a word from every entry in a StringSet,
used for alignment with the size_t that stores the string length.
I could have gone all the way to some kind of empty base class
optimization, but that seemed like overkill. Someone can consider
adding that in the future, though.
https://reviews.llvm.org/D68586
llvm-svn: 374440
Michael J. Spencer [Thu, 10 Oct 2019 20:19:02 +0000 (20:19 +0000)]
[ScanDeps] clang-format, 80 cols.
llvm-svn: 374439
Alexey Bataev [Thu, 10 Oct 2019 20:15:54 +0000 (20:15 +0000)]
[OPENMP]Update doc for supported constructs, NFC.
llvm-svn: 374438
Alexey Bataev [Thu, 10 Oct 2019 20:13:02 +0000 (20:13 +0000)]
[OPENMP50]Support for 'master taskloop' directive.
Added full support for master taskloop directive.
llvm-svn: 374437
Sanjay Patel [Thu, 10 Oct 2019 19:52:27 +0000 (19:52 +0000)]
[x86] reduce duplicate test assertions; NFC
llvm-svn: 374436
Jonas Devlieghere [Thu, 10 Oct 2019 19:51:50 +0000 (19:51 +0000)]
[test] Add timeout to API tests.
Before the reorganiziation, the API tests were inheriting the timeout
from the top-level lit file. Now that this is no longer the case, the
lldb-api test suite needs to set its own timeout.
llvm-svn: 374435
Jonas Devlieghere [Thu, 10 Oct 2019 19:51:47 +0000 (19:51 +0000)]
[test] Cleanup top-level lit.cfg.py
llvm-svn: 374434
Julian Lettner [Thu, 10 Oct 2019 19:43:57 +0000 (19:43 +0000)]
[lit] Bring back `--threads` option alias
Bring back `--threads` option which was lost in the move of the
command line argument parsing code to cl_arguments.py. Update docs
since `--workers` is preferred.
llvm-svn: 374432
Craig Topper [Thu, 10 Oct 2019 19:40:44 +0000 (19:40 +0000)]
[X86] Use packusdw+vpmovuswb to implement v16i32->V16i8 that clamps signed inputs to be between 0 and 255 when zmm registers are disabled on SKX.
If we've disable zmm registers, the v16i32 will need to be split. This split will propagate through min/max the truncate. This creates two sequences that need to be concatenated back to v16i8. We can instead use packusdw to do part of the clamping, truncating, and concatenating all at once. Then we can use a vpmovuswb to finish off the clamp.
Differential Revision: https://reviews.llvm.org/D68763
llvm-svn: 374431
Jonas Devlieghere [Thu, 10 Oct 2019 19:35:20 +0000 (19:35 +0000)]
[Docs] Update testing documentaton
Update the test documentation after the directory reorganization.
llvm-svn: 374430
Joel E. Denny [Thu, 10 Oct 2019 19:25:39 +0000 (19:25 +0000)]
Revert r374388: "[lit] Make internal diff work in pipelines"
This breaks a Windows bot.
llvm-svn: 374429
Jinsong Ji [Thu, 10 Oct 2019 19:25:30 +0000 (19:25 +0000)]
[PowerPC][docs] Update IBM official docs in Compiler Writers Info page
Summary:
Just realized that most of the links in this page are deprecated.
So update some important reference here:
* adding PowerISA 3.0B/2.7B
* adding P8/P9 User Manual
* ELFv2 ABI and errata
Move deprecated ones into "Other documents..".
Reviewers: #powerpc, hfinkel, nemanjai
Reviewed By: hfinkel
Subscribers: shchenz, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D68817
llvm-svn: 374428
Joel E. Denny [Thu, 10 Oct 2019 19:25:24 +0000 (19:25 +0000)]
Revert r374389: "[lit] Clean up internal diff's encoding handling"
This breaks a Windows bot.
llvm-svn: 374427
Joel E. Denny [Thu, 10 Oct 2019 19:25:11 +0000 (19:25 +0000)]
Revert r374390: "[lit] Extend internal diff to support `-` argument"
This breaks a Windows bot.
llvm-svn: 374426
Joel E. Denny [Thu, 10 Oct 2019 19:24:57 +0000 (19:24 +0000)]
Revert r374392: "[lit] Extend internal diff to support -U"
This breaks a Windows bot.
llvm-svn: 374425
Lawrence D'Anna [Thu, 10 Oct 2019 19:24:07 +0000 (19:24 +0000)]
update TestRunCommandInterpreterAPI to use SBFile
Summary:
If you look at what this test is doing, it's actually quite
mysterious why it works at all. It sets the input file
inside a "with open". As soon as the with block ends,
that file will be closed. And yet somehow LLDB reads
commands from it anyway. What's actually happening is that
the file descriptor gets dup'd when something inside LLDB
calls File::GetStream(). I think it's fair to say that
what this test is doing is illegal and it has no right
to expect it to work.
This patch updates the test with two cases. One uses
the SBFile api, and actually transfers ownership of
the original file descriptor to the debugger. The other
just uses the old FILE* API, but in a sane way.
I also set NO_DEBUG_INFO_TESTCASE, because this test doesn't
use any debug info and doesn't need to run three times.
Reviewers: JDevlieghere, jasonmolenda, labath
Reviewed By: labath
Subscribers: aprantl, lldb-commits
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D68738
llvm-svn: 374424
Lawrence D'Anna [Thu, 10 Oct 2019 19:10:59 +0000 (19:10 +0000)]
update SBDebugger::SetInputFile() etc to work on native Files
Summary:
This patch adds FileSP versions of SetInputFile(),
SetOutputFile, and SetErrorFile(). SWIG will convert native
python file objects into FileSP.
Reviewers: JDevlieghere, jasonmolenda, labath
Reviewed By: labath
Subscribers: clayborg, lldb-commits
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D68734
llvm-svn: 374422
Nico Weber [Thu, 10 Oct 2019 18:57:23 +0000 (18:57 +0000)]
win: Move Parallel.h off concrt to cross-platform code
r179397 added Parallel.h and implemented it terms of concrt in 2013.
In 2015, a cross-platform implementation of the functions has appeared
and is in use everywhere but on Windows (r232419). r246219 hints that
<thread> had issues in MSVC2013, but r296906 suggests they've been fixed
now that we require 2015+.
So remove the concrt code. It's less code, and it sounds like concrt has
conceptual and performance issues, see PR41198.
I built blink_core.dll in a debug component build with full symbols and
in a release component build without any symbols. I couldn't measure a
performance difference for linking blink_core.dll before and after this
patch.
Differential Revision: https://reviews.llvm.org/D68820
llvm-svn: 374421
Xiangling Liao [Thu, 10 Oct 2019 18:56:42 +0000 (18:56 +0000)]
[NFC][PowerPC]Clean up PPCAsmPrinter for TOC related pseudo opcode
Add a helper function getMCSymbolForTOCPseudoMO to clean up PPCAsmPrinter
a little bit.
Differential Revision: https://reviews.llvm.org/D68721
llvm-svn: 374420
Eli Friedman [Thu, 10 Oct 2019 18:45:34 +0000 (18:45 +0000)]
[ARM] Fix arm_neon.h with -flax-vector-conversions=none, part 2.
Just running -fsyntax-only over arm_neon.h doesn't cover some intrinsics
which are defined using macros. Add more test coverage for that.
arm-neon-header.c wasn't checking the full set of available NEON target
features; change the target architecture of the test to account for
that.
Fix the generator for arm_neon.h to generate casts in more cases where
they are necessary.
Fix VFMLAL_LOW etc. to express their signatures differently, so the
builtins have the expected type. Maybe the TableGen backend should
detect intrinsics that are defined the wrong way, and produce an error.
The rules here are sort of strange.
Differential Revision: https://reviews.llvm.org/D68743
llvm-svn: 374419
Reid Kleckner [Thu, 10 Oct 2019 18:42:06 +0000 (18:42 +0000)]
Fix one more clang test which didn't have \5C in it
llvm-svn: 374418
Lawrence D'Anna [Thu, 10 Oct 2019 18:38:23 +0000 (18:38 +0000)]
TestFileHandle.py: fix for Python 3.6
Summary:
Python 3.6 stringifies exceptions as `ExceptionClass("foo",)` instead
of `ExceptionClass("foo")`. This patch makes the test assertions a
little more flexible so the test passes anyway.
Reviewers: JDevlieghere, jasonmolenda, labath
Reviewed By: labath
Subscribers: lldb-commits
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D68745
llvm-svn: 374417
Reid Kleckner [Thu, 10 Oct 2019 18:36:41 +0000 (18:36 +0000)]
Update clang tests for new LLVM IR backslash printing in r374415
llvm-svn: 374416
Reid Kleckner [Thu, 10 Oct 2019 18:31:57 +0000 (18:31 +0000)]
Print quoted backslashes in LLVM IR as \\ instead of \5C
This improves readability of Windows path string literals in LLVM IR.
The LLVM assembler has supported \\ in IR strings for a long time, but
the lexer doesn't tolerate escaped quotes, so they have to be printed as
\22 for now.
llvm-svn: 374415
Adrian Prantl [Thu, 10 Oct 2019 18:30:47 +0000 (18:30 +0000)]
Remove merge marker :-)
llvm-svn: 374414
Nico Weber [Thu, 10 Oct 2019 18:20:16 +0000 (18:20 +0000)]
Fix Windows build after r374381
llvm-svn: 374413
Jim Ingham [Thu, 10 Oct 2019 18:19:27 +0000 (18:19 +0000)]
Die, TABS, die, die, die, die...
Not sure how tabs got into this file, but they don't belong there.
llvm-svn: 374412
Reid Kleckner [Thu, 10 Oct 2019 18:17:24 +0000 (18:17 +0000)]
Remove strings.h include to fix GSYM Windows build
Fifth time's the charm.
llvm-svn: 374411
Greg Clayton [Thu, 10 Oct 2019 18:13:13 +0000 (18:13 +0000)]
Unbreak buildbots.
llvm-svn: 374410
Greg Clayton [Thu, 10 Oct 2019 18:11:49 +0000 (18:11 +0000)]
Fix buildbots by using memset instead of bzero.
llvm-svn: 374409
Casey Carter [Thu, 10 Oct 2019 18:07:12 +0000 (18:07 +0000)]
[libc++][test] Change IsSmallObject's calculation for std::any's small object buffer
`sizeof(std::any) - sizeof(void*)` is correct for both libc++ and the MSVC standard library.
Differential Revision: https://reviews.llvm.org/D68756
llvm-svn: 374407
Michael Liao [Thu, 10 Oct 2019 18:04:52 +0000 (18:04 +0000)]
Fix build by adding the missing dependency.
llvm-svn: 374406
Julian Lettner [Thu, 10 Oct 2019 18:03:37 +0000 (18:03 +0000)]
[lit] Leverage argparse features to remove some code
Reviewed By: rnk, serge-sans-paille
Differential Revision: https://reviews.llvm.org/D68589
llvm-svn: 374405
Reid Kleckner [Thu, 10 Oct 2019 18:01:27 +0000 (18:01 +0000)]
Fix test to avoid check-not matching the temp file absolute path
Fix for PR43636
llvm-svn: 374404
Reid Kleckner [Thu, 10 Oct 2019 18:01:20 +0000 (18:01 +0000)]
Re-land "Use -fdebug-compilation-dir to form absolute paths in coverage mappings"
This reverts r374324 (git commit
62808631acceaa8b78f8ab9b407eb6b943ff5f77)
I changed the test to not rely on finding the sequence "clang, test,
CoverageMapping" in the CWD used to run the test. Instead it makes its
own internal directory hierarchy of foo/bar/baz and looks for that.
llvm-svn: 374403
Adrian Prantl [Thu, 10 Oct 2019 17:59:18 +0000 (17:59 +0000)]
Remove CC autodetection from Makefile.rules
Auto-detecting CC in Makefile.rules is no longer useful. Ever since
out-of-tree builds we are better off just running lldb-dotest which
sets it directly. This also makes it harder to accidentally unset CC
in a Makefile.
Differential Revision: https://reviews.llvm.org/D68731
llvm-svn: 374402
Adrian Prantl [Thu, 10 Oct 2019 17:59:15 +0000 (17:59 +0000)]
Speed up accelerator table lookups
When debugging a large program like clang and doing "frame variable
*this", the ValueObject pretty printer is doing hundreds of scoped
FindTypes lookups. The ones that take longest are the ones where the
DWARFDeclContext ends in something like ::Iterator which produces many
false positives that need to be filtered out *after* extracting the
DIEs. This patch demonstrates a way to filter out false positives at
the accerator table lookup step.
With this patch
lldb clang-10 -o "b EmitFunctionStart" -o r -o "f 2" -o "fr v *this" -b -- ...
goes (in user time) from 5.6s -> 4.8s
or (in wall clock) from 6.9s -> 6.0s.
Differential Revision: https://reviews.llvm.org/D68678
llvm-svn: 374401
Julian Lettner [Thu, 10 Oct 2019 17:58:38 +0000 (17:58 +0000)]
[lit] Move argument parsing/validation to separate file
Reviewed By: serge-sans-paille
Differential Revision: https://reviews.llvm.org/D68529
llvm-svn: 374400
Paul Hoad [Thu, 10 Oct 2019 17:54:47 +0000 (17:54 +0000)]
[clang-format] throws an incorrect assertion in consumeToken() formatting the MSVC stl
Summary:
An incorrect assertion is thrown when clang-formatting MSVC's STL library
```
Assertion failed: !Line.startsWith(tok::hash), file C:/llvm/llvm-project/clang/lib/Format/TokenAnnotator.cpp, line 847
Stack dump:
0. Program arguments: C:\llvm\build\bin\clang-format.exe -i -n ./stl/inc/xkeycheck.h
```
```
Enable warning C4005 to find the forbidden define.
```
Reviewers: mitchell-stellar, STL_MSFT, klimek, krasimir
Reviewed By: mitchell-stellar
Subscribers: cfe-commits
Tags: #clang-format, #clang-tools-extra, #clang
Differential Revision: https://reviews.llvm.org/D68707
llvm-svn: 374399
Greg Clayton [Thu, 10 Oct 2019 17:52:33 +0000 (17:52 +0000)]
Unbreak llvm-clang-lld-x86_64-scei-ps4-windows10pro-fast buildbot.
llvm-svn: 374398
Sanjay Patel [Thu, 10 Oct 2019 17:52:02 +0000 (17:52 +0000)]
[DAGCombiner] fold select-of-constants to shift
This reverses the scalar canonicalization proposed in D63382.
Pre: isPowerOf2(C1)
%r = select i1 %cond, i32 C1, i32 0
=>
%z = zext i1 %cond to i32
%r = shl i32 %z, log2(C1)
https://rise4fun.com/Alive/Z50
x86 already tries to fold this pattern, but it isn't done
uniformly, so we still see a diff. AArch64 probably should
enable the TLI hook to benefit too, but that's a follow-on.
llvm-svn: 374397
Greg Clayton [Thu, 10 Oct 2019 17:49:33 +0000 (17:49 +0000)]
Unbreak windows buildbots.
llvm-svn: 374396
Nico Weber [Thu, 10 Oct 2019 17:47:18 +0000 (17:47 +0000)]
gn build: restore tablegen restat optimization after r373664
llvm-svn: 374395
Jim Ingham [Thu, 10 Oct 2019 17:44:50 +0000 (17:44 +0000)]
Implement serializing scripted breakpoints and their extra args.
Differential Revision: https://reviews.llvm.org/D68750
llvm-svn: 374394
Joel E. Denny [Thu, 10 Oct 2019 17:40:12 +0000 (17:40 +0000)]
[lit] Extend internal diff to support -U
When using lit's internal shell, RUN lines like the following
accidentally execute an external `diff` instead of lit's internal
`diff`:
```
# RUN: program | diff -U1 file -
```
Such cases exist now, in `clang/test/Analysis` for example. We are
preparing patches to ensure lit's internal `diff` is called in such
cases, which will then fail because lit's internal `diff` doesn't
recognize `-U` as a command-line option. This patch adds `-U`
support.
Reviewed By: rnk
Differential Revision: https://reviews.llvm.org/D68668
llvm-svn: 374392
Nico Weber [Thu, 10 Oct 2019 17:40:00 +0000 (17:40 +0000)]
gn build: merge r374381 more (effectively a no-op)
llvm-svn: 374391
Joel E. Denny [Thu, 10 Oct 2019 17:39:57 +0000 (17:39 +0000)]
[lit] Extend internal diff to support `-` argument
When using lit's internal shell, RUN lines like the following
accidentally execute an external `diff` instead of lit's internal
`diff`:
```
# RUN: program | diff file -
```
Such cases exist now, in `clang/test/Analysis` for example. We are
preparing patches to ensure lit's internal `diff` is called in such
cases, which will then fail because lit's internal `diff` doesn't
recognize `-` as a command-line option. This patch adds support for
`-` to mean stdin.
Reviewed By: probinson, rnk
Differential Revision: https://reviews.llvm.org/D67643
llvm-svn: 374390
Joel E. Denny [Thu, 10 Oct 2019 17:39:41 +0000 (17:39 +0000)]
[lit] Clean up internal diff's encoding handling
As suggested by rnk at D67643#1673043, instead of reading files
multiple times until an appropriate encoding is found, read them once
as binary, and then try to decode what was read.
For python >= 3.5, don't fail when attempting to decode the
`diff_bytes` output in order to print it.
Finally, add some tests for encoding handling.
Reviewed By: rnk
Differential Revision: https://reviews.llvm.org/D68664
llvm-svn: 374389
Joel E. Denny [Thu, 10 Oct 2019 17:39:24 +0000 (17:39 +0000)]
[lit] Make internal diff work in pipelines
When using lit's internal shell, RUN lines like the following
accidentally execute an external `diff` instead of lit's internal
`diff`:
```
# RUN: program | diff file -
# RUN: not diff file1 file2 | FileCheck %s
```
Such cases exist now, in `clang/test/Analysis` for example. We are
preparing patches to ensure lit's internal `diff` is called in such
cases, which will then fail because lit's internal `diff` cannot
currently be used in pipelines and doesn't recognize `-` as a
command-line option.
To enable pipelines, this patch moves lit's `diff` implementation into
an out-of-process script, similar to lit's `cat` implementation. A
follow-up patch will implement `-` to mean stdin.
Reviewed By: probinson, stella.stamenova
Differential Revision: https://reviews.llvm.org/D66574
llvm-svn: 374388
Alexey Bataev [Thu, 10 Oct 2019 17:28:10 +0000 (17:28 +0000)]
[OPENMP50]Support for declare variant directive for NVPTX target.
NVPTX does not support global aliases. Instead, we have to copy the full
body of the variant function for the original function.
llvm-svn: 374387
Jonas Devlieghere [Thu, 10 Oct 2019 17:27:09 +0000 (17:27 +0000)]
[test] Use a different module cache for Shell and API tests.
Before the test reorganization, everything was part of a single test
suite with a single module cache. Now that things are properly separated
this is no longer the case. Only the shell tests inherited the logic to
properly configure and wipe the module caches. This patch adds that
logic back for the API tests. While doing so, I noticed that we were
configuring a Clang module cache in CMake, but weren't actually using it
from dotest.py. I included a fix for that in this patch as well.
Differential revision: https://reviews.llvm.org/D68755
llvm-svn: 374386
Jonas Devlieghere [Thu, 10 Oct 2019 17:27:06 +0000 (17:27 +0000)]
[test] Pass DSYMUTIL and SDKROOT as part of the Make invocation.
Pass the DSYMUTIL and SDKROOT variables on the command line instead of
the environment. This makes it easier to reproduce the make invocation
during development.
Differential revision: https://reviews.llvm.org/D68812
llvm-svn: 374385
Julian Lettner [Thu, 10 Oct 2019 17:19:58 +0000 (17:19 +0000)]
Reland "[ASan] Do not misrepresent high value address dereferences as null dereferences"
Updated: Removed offending TODO comment.
Dereferences with addresses above the 48-bit hardware addressable range
produce "invalid instruction" (instead of "invalid access") hardware
exceptions (there is no hardware address decoding logic for those bits),
and the address provided by this exception is the address of the
instruction (not the faulting address). The kernel maps the "invalid
instruction" to SEGV, but fails to provide the real fault address.
Because of this ASan lies and says that those cases are null
dereferences. This downgrades the severity of a found bug in terms of
security. In the ASan signal handler, we can not provide the real
faulting address, but at least we can try not to lie.
rdar://
50366151
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D68676
> llvm-svn: 374265
llvm-svn: 374384
GN Sync Bot [Thu, 10 Oct 2019 17:14:20 +0000 (17:14 +0000)]
gn build: Merge r374381
llvm-svn: 374383
Jonas Devlieghere [Thu, 10 Oct 2019 17:12:57 +0000 (17:12 +0000)]
[test] Trim the symbtab test case (NFC)
We don't actually need the section content for this tests.
llvm-svn: 374382
Greg Clayton [Thu, 10 Oct 2019 17:10:11 +0000 (17:10 +0000)]
Add GsymCreator and GsymReader.
This patch adds the ability to create GSYM files with GsymCreator, and read them with GsymReader. Full testing has been added for both new classes.
This patch differs from the original patch https://reviews.llvm.org/D53379 in that is uses a StringTableBuilder class from llvm instead of a custom version. Support for big and little endian files has been added. If the endianness matches the current host, we use efficient extraction for the header, address table and address info offset tables.
Differential Revision: https://reviews.llvm.org/D68744
llvm-svn: 374381
Jonas Devlieghere [Thu, 10 Oct 2019 16:37:38 +0000 (16:37 +0000)]
[dsymutil] Move CommonSymbols.clear() in resetParserState.
This seems like a more natural place to clear the vector, especially
since this method is clearing other data structures as well.
llvm-svn: 374378
David Green [Thu, 10 Oct 2019 16:34:30 +0000 (16:34 +0000)]
[ARM] VQSUB instruction
Same as VQADD, VQSUB can be selected from llvm.ssub.sat intrinsics.
Differential Revision: https://reviews.llvm.org/D68567
llvm-svn: 374377
Adrian Prantl [Thu, 10 Oct 2019 16:16:49 +0000 (16:16 +0000)]
Increase timeout in pexpect to lower chances of tests failing under ASAN.
If this doesn't actually work, I'll revert the change and just disable
the remaining thee pexpect tests under asan.
llvm-svn: 374375
Siva Chandra [Thu, 10 Oct 2019 16:06:21 +0000 (16:06 +0000)]
Use arrays on stack and avoid use of new and delete operators.
Summary: Also fix an error found with LLVM_USE_SANITIZER=Address.
Reviewers: nelhage
Subscribers: libc-commits
Tags: #libc-project
Differential Revision: https://reviews.llvm.org/D68761
llvm-svn: 374374
David Green [Thu, 10 Oct 2019 16:04:49 +0000 (16:04 +0000)]
[Codegen] Alter the default promotion for saturating adds and subs
The default promotion for the add_sat/sub_sat nodes currently does:
1. ANY_EXTEND iN to iM
2. SHL by M-N
3. [US][ADD|SUB]SAT
4. L/ASHR by M-N
If the promoted add_sat or sub_sat node is not legal, this can produce code
that effectively does a lot of shifting (and requiring large constants to be
materialised) just to use the overflow flag. It is simpler to just do the
saturation manually, using the higher bitwidth addition and a min/max against
the saturating bounds. That is what this patch attempts to do.
Differential Revision: https://reviews.llvm.org/D68643
llvm-svn: 374373
Kadir Cetinkaya [Thu, 10 Oct 2019 16:04:45 +0000 (16:04 +0000)]
Fix assertions disabled builds after rL374367
llvm-svn: 374372
Adrian Prantl [Thu, 10 Oct 2019 15:59:17 +0000 (15:59 +0000)]
Increase timeout in gdbclientutils.py to decrease chance of test failing under ASAN.
llvm-svn: 374371
Sanjay Patel [Thu, 10 Oct 2019 15:38:29 +0000 (15:38 +0000)]
[DAGCombiner] reduce code duplication; NFC
llvm-svn: 374370
Guillaume Chatelet [Thu, 10 Oct 2019 15:38:22 +0000 (15:38 +0000)]
[Alignment][NFC] Use llv::Align in GISelKnownBits
Summary:
This is patch is part of a series to introduce an Alignment type.
See this thread for context: http://lists.llvm.org/pipermail/llvm-dev/2019-July/133851.html
See this patch for the introduction of the type: https://reviews.llvm.org/D64790
Reviewers: courbet
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D68786
llvm-svn: 374369
Yonghong Song [Thu, 10 Oct 2019 15:33:09 +0000 (15:33 +0000)]
[BPF] Remove relocation for patchable externs
Previously, patchable extern relocations are introduced to patch
external variables used for multi versioning in
compile once, run everywhere use case. The load instruction
will be converted into a move with an patchable immediate
which can be changed by bpf loader on the host.
The kernel verifier has evolved and is able to load
and propagate constant values, so compiler relocation
becomes unnecessary. This patch removed codes related to this.
Differential Revision: https://reviews.llvm.org/D68760
llvm-svn: 374367
Kousik Kumar [Thu, 10 Oct 2019 15:29:01 +0000 (15:29 +0000)]
In openFileForRead don't cache erroneous entries if the error relates to them being directories. Add tests.
Summary:
It seems that when the CachingFileSystem is first given a file to open that is actually a directory, it incorrectly
caches that path to be errenous and throws an error when subsequently a directory open call is made for the same
path.
This change makes it so that we do NOT cache a path if it turns out we asked for a file when its a directory.
Reviewers: arphaman
Subscribers: dexonsmith, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D68193
llvm-svn: 374366
Stanislav Mekhanoshin [Thu, 10 Oct 2019 15:28:52 +0000 (15:28 +0000)]
[AMDGPU] Fixed dpp_combine.mir with expensive checks. NFC.
llvm-svn: 374365
Simon Pilgrim [Thu, 10 Oct 2019 15:25:16 +0000 (15:25 +0000)]
Fix Wdocumentation warnings. NFCI.
llvm-svn: 374364
Alexey Bataev [Thu, 10 Oct 2019 15:15:26 +0000 (15:15 +0000)]
[OPENMP50]Register vendor name only once in vendor context selector.
No need to store multiple copies of the same vendor names in the context
selector, keep only single copy.
llvm-svn: 374363