NAKAMURA Takumi [Wed, 21 Sep 2016 04:43:11 +0000 (04:43 +0000)]
llvm/test/CodeGen/NVPTX/zero-cs.ll: Relax an expression to match in -Asserts.
LLVM ERROR: Cannot select: 0x3607bf0: i32 = ExternalSymbol'__powidf2'
llvm-svn: 282053
Craig Topper [Wed, 21 Sep 2016 03:58:44 +0000 (03:58 +0000)]
[AVX-512] Don't add an additional rounding mode operand to the avx512 vcvtps2ph intrinsic lowering.
There was no way to control its value so it was always FROUND_CURRENT making it unnecessary. The true rounding mode is encoded in the immediate operand of the instruction.
This also removes the pattern from the rb form of the instructions since there is no way to specify the FROUND_NO_EXC rounding mode it required.
llvm-svn: 282052
Craig Topper [Wed, 21 Sep 2016 03:58:41 +0000 (03:58 +0000)]
[AVX-512] Simplify handling of INTR_TYPE_1OP_MASK_RM to remove support for the second opcode since its never used. This makes it consistent with INTR_TYPE_2OP_MASK_RM and INTR_TYPE_3OP_MASK_RM.
And even if it was used we were passing the same operands to both so it wouldn't make sense to have two opcodes.
llvm-svn: 282051
Kostya Serebryany [Wed, 21 Sep 2016 03:50:37 +0000 (03:50 +0000)]
[libFuzzer] fix libc++ build
llvm-svn: 282050
Rui Ueyama [Wed, 21 Sep 2016 03:22:18 +0000 (03:22 +0000)]
Accept sh_entsize = 0.
This surfaced again with Rust. As per bug 30435, rustc creates a
mergeable section with a sh_entsize zero. It bit us before, too.
I think we should relax the input check rather than being too picky.
Differential Revision: https://reviews.llvm.org/D24789
llvm-svn: 282049
Adam Nemet [Wed, 21 Sep 2016 03:14:20 +0000 (03:14 +0000)]
[LV] When reporting about a specific instruction without debug location use loop's
This can occur for example if some optimization drops the debug location.
llvm-svn: 282048
Kostya Serebryany [Wed, 21 Sep 2016 02:05:39 +0000 (02:05 +0000)]
[libFuzzer] more refactoring; NFC
llvm-svn: 282047
Craig Topper [Wed, 21 Sep 2016 02:05:22 +0000 (02:05 +0000)]
[AVX-512] Don't lower avx512 vcvtps2ph/vcvtph2ps nodes to ISD::FP16_TO_FP/ISD::FP_TO_FP16 with an extra x86 specific rounding mode operand. We should use a target specific ISD opcode.
llvm-svn: 282046
Jacques Pienaar [Wed, 21 Sep 2016 01:57:57 +0000 (01:57 +0000)]
[NVPTX] Check if callsite is defined when computing argument allignment
Summary: In getArgumentAlignment check if the ImmutableCallSite pointer CS is non-null before dereferencing. If CS is 0x0 fall back to the ABI type alignment else compute the alignment as before.
Reviewers: eliben, jpienaar
Subscribers: jlebar, vchuravy, cfe-commits, jholewinski
Differential Revision: https://reviews.llvm.org/D9168
llvm-svn: 282045
Kostya Serebryany [Wed, 21 Sep 2016 01:50:50 +0000 (01:50 +0000)]
[libFuzzer] refactoring: split the large header into many; NFC
llvm-svn: 282044
Jim Ingham [Wed, 21 Sep 2016 01:21:19 +0000 (01:21 +0000)]
Adds tests for breakpoint names, and a FindBreakpointsByName.
Also if you set a breakpoint with an invalid name, we'll
refuse to set the breakpoint rather than silently ignoring
the name.
llvm-svn: 282043
Kostya Serebryany [Wed, 21 Sep 2016 01:04:43 +0000 (01:04 +0000)]
[libFuzzer] refactoring: move the Corpus into a separate class; delete two unused experimental features
llvm-svn: 282042
Eugene Zelenko [Wed, 21 Sep 2016 00:59:22 +0000 (00:59 +0000)]
Fix Clang initialization and Clang-tidy modernize-use-nullptr warnings in source/Plugins/Process/Utility.
Differential revision: https://reviews.llvm.org/D24694
llvm-svn: 282041
Justin Bogner [Wed, 21 Sep 2016 00:25:45 +0000 (00:25 +0000)]
Revert "TableGen: Switch from a std::map to a DenseMap in CodeGenSubRegIndex. NFC"
It turns out we iterate over this map a fair amount and the order
matters for clang to be deterministic. See:
http://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-
20160919/391315.html
This reverts r279875.
llvm-svn: 282040
Michael Kuperstein [Tue, 20 Sep 2016 23:10:31 +0000 (23:10 +0000)]
[InferAttributes] Don't access parameters that don't exist.
Check for the correct number of parameters before querying their type.
This fixes PR30455.
llvm-svn: 282038
Teresa Johnson [Tue, 20 Sep 2016 23:07:17 +0000 (23:07 +0000)]
[ThinLTO] Always emit a summary when compiling in ThinLTO mode
Summary:
Emit an empty summary section, instead of no summary section, when
there are no global variables in the index. This ensures that LTO
will treat these files as ThinLTO inputs, instead of as regular
LTO inputs.
In addition to not being what the user likely intended when
compiling with -flto=thin, the current behavior is problematic for
distributed build systems that expect to get ThinLTO index and imports
files back for each input compiled with -flto=thin. Combining into
a single regular LTO module also reduces the backend parallelism.
And in the case where the index was suppressed due to uses in
inline assembly, combining into a single LTO module could provoke
renaming of duplicates that we were trying to prevent by suppressing
the index.
This change required a couple of fixes to handle the empty summary
section.
Reviewers: mehdi_amini
Subscribers: mehdi_amini, llvm-commits, pcc
Differential Revision: https://reviews.llvm.org/D24779
llvm-svn: 282037
Jim Ingham [Tue, 20 Sep 2016 22:54:49 +0000 (22:54 +0000)]
Add some more tests for breakpoint serialization.
Serialize breakpoint names & the hardware_requested attributes.
Also added a few missing affordances to SBBreakpoint whose absence
writing the tests pointed out.
<rdar://problem/
12611863>
llvm-svn: 282036
Rafael Espindola [Tue, 20 Sep 2016 22:43:15 +0000 (22:43 +0000)]
Revert "Revert "Only restrict order if both sections are in the script.""
This reverts commit r282021, bringing back r282015.
The problem was that the comparison function was not a strict weak
ordering anymore, which this patch fixes.
Original message:
Only restrict order if both sections are in the script.
This matches gold and bfd behavior and is required to handle some scripts.
The script has to assume where PT_LOADs start in order to align that
spot. If we don't allow section it doesn't know about to move to the
middle, we can need more PT_LOADs and those will not be aligned.
llvm-svn: 282035
Xinliang David Li [Tue, 20 Sep 2016 22:39:47 +0000 (22:39 +0000)]
code cleanup -- commoning IR travsersals
llvm-svn: 282034
Chris Bieneman [Tue, 20 Sep 2016 22:29:22 +0000 (22:29 +0000)]
Revert "[CMake] Explicitly add --target option to compiler flags"
This reverts commit r282024.
This broke some bots, and I'm going to revert while I figure it out.
See: http://lab.llvm.org:8011/builders/clang-x86_64-linux-abi-test/builds/21120
llvm-svn: 282033
Kamil Rytarowski [Tue, 20 Sep 2016 22:26:29 +0000 (22:26 +0000)]
Unbreak the NetBSD build after recent changes
Similar to r281922 "Try to fix freebsd and android builds."
Replace .AppendArgument(cstr) with .AppendArgument(llvm::StringRef(cstr))
llvm-svn: 282032
Eric Christopher [Tue, 20 Sep 2016 22:19:33 +0000 (22:19 +0000)]
Remove the default subtarget from the x86 port as it isn't necessary (or
correct) anymore.
llvm-svn: 282031
Kostya Kortchinsky [Tue, 20 Sep 2016 22:17:59 +0000 (22:17 +0000)]
[scudo] Fix a bug in the new Secondary Allocator
Summary:
GetActuallyAllocatedSize() was not accounting for the last page of the mapping
being a guard page, and was returning the wrong number of actually allocated
bytes, which in turn would mess up with the realloc logic. Current tests didn't
find this as the size exercised was only serviced by the Primary.
Correct the issue by subtracting PageSize, and update the realloc test to
exercise paths in both the Primary and the Secondary.
Reviewers: kcc
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D24787
llvm-svn: 282030
Zachary Turner [Tue, 20 Sep 2016 22:15:07 +0000 (22:15 +0000)]
Enable clang attributes when using clang-cl.
llvm-svn: 282029
Eric Christopher [Tue, 20 Sep 2016 22:03:28 +0000 (22:03 +0000)]
Revert "Remove extra argument used once on
TargetMachine::getNameWithPrefix and inline the result into the singular
caller." and "Remove more guts of TargetMachine::getNameWithPrefix and
migrate one check to the TLOF mach-o version." temporarily until I can
get the whole call migrated out of the TargetMachine as we could hit
places where TLOF isn't valid.
This reverts commits r281981 and r281983.
llvm-svn: 282028
Chris Bieneman [Tue, 20 Sep 2016 21:36:33 +0000 (21:36 +0000)]
[CMake] Fix error preventing simulator sanitizers from building
This should be checking for sim archs not target archs. This bug has
been around for a long time.
llvm-svn: 282025
Chris Bieneman [Tue, 20 Sep 2016 21:36:30 +0000 (21:36 +0000)]
[CMake] Explicitly add --target option to compiler flags
Summary:
Much of the non-Darwin build system assumes that COMPILER_RT_DEFAULT_TARGET_TRIPLE is the default target triple for the compiler being used. With clang as your compiler this isn't necessarily true.
To ensure that the rest of the build system behaves as expected this patch adds "--target=${COMPILER_RT_DEFAULT_TARGET_TRIPLE}" to the compiler flags for C, CXX and ASM sources.
Reviewers: compnerd, rengolin, fjricci
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D24156
llvm-svn: 282024
Anna Thomas [Tue, 20 Sep 2016 21:36:02 +0000 (21:36 +0000)]
[RS4GC] Refactor code for Rematerializing in presence of phi. NFC
Summary:
This is an NFC refactoring change as a precursor to the actual fix for rematerializing in
presence of phi.
https://reviews.llvm.org/D24399
Pasted from review:
findRematerializableChainToBasePointer changed to return the root of the
chain. instead of true or false.
move the PHI matching logic into the caller by inspecting the root return value.
This includes an assertion that the alternate root is in the liveset for the
call.
Tested with current RS4GC tests.
Reviewers: reames, sanjoy
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D24780
llvm-svn: 282023
George Burgess IV [Tue, 20 Sep 2016 21:30:01 +0000 (21:30 +0000)]
[CodeGen] stop short-circuiting the SSP code for sspstrong.
This check caused us to skip adding layout information for calls to
alloca in sspreq/sspstrong mode. We check properly for sspstrong later
on (and add the correct layout info when doing so), so removing this
shouldn't hurt.
No test is included, since testing this using lit seems to require
checking for exact offsets in asm, which is something that the lit tests
for this avoid. If someone cares deeply, I'm happy to write a unittest
or something to cover this, but that feels like overkill.
Patch by Daniel Micay.
Differential Revision: https://reviews.llvm.org/D22714
llvm-svn: 282022
Rafael Espindola [Tue, 20 Sep 2016 21:28:19 +0000 (21:28 +0000)]
Revert "Only restrict order if both sections are in the script."
This reverts commit r282015. It broke some bots.
llvm-svn: 282021
Vedant Kumar [Tue, 20 Sep 2016 21:27:48 +0000 (21:27 +0000)]
[llvm-cov] Demangle names for hidden instantiation views
llvm-svn: 282020
Kostya Serebryany [Tue, 20 Sep 2016 21:27:10 +0000 (21:27 +0000)]
[asan] Add C++17 aligned new/delete entrypoints. Patch by Jakub Jelinek, see https://reviews.llvm.org/D24771
llvm-svn: 282019
Reid Kleckner [Tue, 20 Sep 2016 21:09:31 +0000 (21:09 +0000)]
XFAIL cfi/stats.cpp on Windows until we fix LLD
llvm-svn: 282018
Xinliang David Li [Tue, 20 Sep 2016 21:04:22 +0000 (21:04 +0000)]
[Profile] dump ic value profile value/site-count histogram
Differential Revision: http://reviews.google.com/D24783
llvm-svn: 282017
Rafael Espindola [Tue, 20 Sep 2016 20:54:39 +0000 (20:54 +0000)]
Only restrict order if both sections are in the script.
This matches gold and bfd behavior and is required to handle some scripts.
The script has to assume where PT_LOADs start in order to align that
spot. If we don't allow section it doesn't know about to move to the
middle, we can need more PT_LOADs and those will not be aligned.
llvm-svn: 282015
Rafael Espindola [Tue, 20 Sep 2016 20:48:54 +0000 (20:48 +0000)]
Make test less dependent on the section order.
This means either relaxing CHECKs or listing more sections and
addresses in linker scripts.
llvm-svn: 282014
Zachary Turner [Tue, 20 Sep 2016 20:44:50 +0000 (20:44 +0000)]
[PECOFF] Use the COFF subsystem field when determining the Triple.
Patch by walter erquinigo
Differential revision: https://reviews.llvm.org/D24284
llvm-svn: 282013
Filipe Cabecinhas [Tue, 20 Sep 2016 20:33:18 +0000 (20:33 +0000)]
[asan] Fix GlobalAddressDescription::Print()
Summary: Check bug_type for nullptr before calling internal_strcmp
Reviewers: kcc, vitalybuka, eugenis
Subscribers: kubabrecka, llvm-commits
Differential Revision: https://reviews.llvm.org/D24773
llvm-svn: 282012
Anna Zaks [Tue, 20 Sep 2016 20:28:50 +0000 (20:28 +0000)]
[analyzer] Add a checker that detects blocks in critical sections
This checker should find the calls to blocking functions (for example: sleep, getc, fgets,read,recv etc.) inside a critical section. When sleep(x) is called while a mutex is held, other threads cannot lock the same mutex. This might take some time, leading to bad performance or even deadlock.
Example:
mutex_t m;
void f() {
sleep(1000); // Error: sleep() while m is locked! [f() is called from foobar() while m is locked]
// do some work
}
void foobar() {
lock(m);
f();
unlock(m);
}
A patch by zdtorok (Zoltán Dániel Török)!
Differential Revision: https://reviews.llvm.org/D21506
llvm-svn: 282011
Petr Hosek [Tue, 20 Sep 2016 20:21:13 +0000 (20:21 +0000)]
Mark ELF sections whose name start with .note as note
Previously, such section would be marked as SHT_PROGBITS which
makes it impossible to use an initialized C variable declaration
to emit an (allocated) ELF note. The new behavior is also consistent
with ELF assembly parser.
Differential Revision: https://reviews.llvm.org/D24692
llvm-svn: 282010
Xinliang David Li [Tue, 20 Sep 2016 20:20:01 +0000 (20:20 +0000)]
[Profile] Do not annotate select insts not covered in profile.
Fixed PR/30466
llvm-svn: 282009
Kevin Enderby [Tue, 20 Sep 2016 20:14:14 +0000 (20:14 +0000)]
Next set of additional error checks for invalid Mach-O files for bad load commands
that use the Mach::dylib_command type for the load commands that are
currently used in the MachOObjectFile constructor.
This contains the missing checks for LC_ID_DYLIB, LC_ID_DYLIB, etc.
load commands and the fields for the Mach::dylib_command type.
Also checks that only an MH_DYLIB or MH_STUB_DYLIB has an
LC_ID_DYLIB load command (and others filetype don’t) and there
is not more than one of these load commands.
llvm-svn: 282008
Eric Christopher [Tue, 20 Sep 2016 19:45:06 +0000 (19:45 +0000)]
Remove some boilerplate comments that don't explain anything.
llvm-svn: 282007
Rui Ueyama [Tue, 20 Sep 2016 19:42:41 +0000 (19:42 +0000)]
Simplify SORT and --sort-section command line option handling.
Differential Revision: https://reviews.llvm.org/D24685
llvm-svn: 282006
Sanjay Patel [Tue, 20 Sep 2016 19:31:30 +0000 (19:31 +0000)]
[x86] split up tests, regenerate checks
Note that we fail to eliminate 'or' with 0!
llvm-svn: 282005
Zachary Turner [Tue, 20 Sep 2016 19:10:56 +0000 (19:10 +0000)]
Fix a regex error breaking tests.
llvm-svn: 282004
Chris Bieneman [Tue, 20 Sep 2016 19:09:21 +0000 (19:09 +0000)]
[CMake] Support overriding CLANG_VERSION_*
As with how we handle LLVM_VERSION_* variables we should support
overriding clang version variables.
llvm-svn: 282003
Xinliang David Li [Tue, 20 Sep 2016 19:07:22 +0000 (19:07 +0000)]
[Profile] code refactoring: make getStep a method in base class
llvm-svn: 282002
Evandro Menezes [Tue, 20 Sep 2016 19:02:09 +0000 (19:02 +0000)]
Revert part of "AArch64: Do not test for CPUs, use SubtargetFeatures"
This reverts part of commit
119e358d9635c8d1f3e7aee67e3ea3b8a62f8db6 by
removing FeatureUseRSqrt et al per request by Eric Christopher
<echristo@gmail.com> (v. http://bit.ly/2cmz6kW).
llvm-svn: 282001
Evandro Menezes [Tue, 20 Sep 2016 19:02:06 +0000 (19:02 +0000)]
Revert "[AArch64] Use the reciprocal estimation machinery"
This reverts commit
b7d42b0048f65346e9fa37fb65defeea7ce8c337 per request by
Eric Christopher <echristo@gmail.com> (v. http://bit.ly/2cmz6kW).
llvm-svn: 282000
Evandro Menezes [Tue, 20 Sep 2016 19:02:02 +0000 (19:02 +0000)]
Revert "[AArch64] Properly validate the reciprocal estimation."
This reverts commit
ad8ca1528242e2a4cb363e3779309e70eb7a430e per request by
Eric Christopher <echristo@gmail.com> (v. http://bit.ly/2cmz6kW).
llvm-svn: 281999
Zachary Turner [Tue, 20 Sep 2016 18:41:19 +0000 (18:41 +0000)]
Add some entropy to the folder name in Driver/warning-options.cpp.
It was trying to check that things behave correctly when a
non-existant folder was specified for -isysroot. Incidentally,
I have a folder named FOO in the root of my drive, so this test
was failing. Make this impossible by using %T to refer to a
definitely non-existant folder.:
llvm-svn: 281998
Saleem Abdulrasool [Tue, 20 Sep 2016 18:38:54 +0000 (18:38 +0000)]
CodeGen: further merge cstring literal construction
Use the new CreateCStringLiteral in an additional site. Now all the C string
literals are created in one function. Furthermore, mark the additional literal
as an `unnamed_addr constant`.
llvm-svn: 281997
Nick Lewycky [Tue, 20 Sep 2016 18:37:25 +0000 (18:37 +0000)]
Fix typo in documentation.
Since this is a header it will break links to this section.
llvm-svn: 281996
Alexander Shaposhnikov [Tue, 20 Sep 2016 18:32:48 +0000 (18:32 +0000)]
[cleanup] Remove excessive padding from TextTokenRetokenizer::Position
Reorder the fields of the struct TextTokenRetokenizer::Position to remove excessive padding.
Test plan: make -j8 check-clang
Differential revision: https://reviews.llvm.org/D24751
llvm-svn: 281995
Adrian Prantl [Tue, 20 Sep 2016 18:28:42 +0000 (18:28 +0000)]
ASAN: Don't drop debug info attachements for global variables.
This is a follow-up to r281284. Global Variables now can have
!dbg attachements, so ASAN should clone these when generating a
sanitized copy of a global variable.
<rdar://problem/
24899262>
llvm-svn: 281994
Enrico Granata [Tue, 20 Sep 2016 18:26:30 +0000 (18:26 +0000)]
Make it so that one can register prefix matches as well as identical matches as extra cases for NSDictionary data formatting
llvm-svn: 281993
Dmitry Vyukov [Tue, 20 Sep 2016 18:05:06 +0000 (18:05 +0000)]
tsan: revert r281970
r281970 extended the check in a useful way,
but caused (true) failures on aarch64.
Revert it for now.
llvm-svn: 281992
Adrian McCarthy [Tue, 20 Sep 2016 17:42:13 +0000 (17:42 +0000)]
Fix syntactical nit from r281990.
llvm-svn: 281991
Adrian McCarthy [Tue, 20 Sep 2016 17:20:51 +0000 (17:20 +0000)]
Emit S_COMPILE3 CodeView record
CodeView has an S_COMPILE3 record to identify the compiler and source language of the compiland. This record comes first in the debug$S section for the compiland. The debuggers rely on this record to know the source language of the code.
There was a little test fallout from introducing a new record into the symbols subsection.
Differential Revision: https://reviews.llvm.org/D24317
llvm-svn: 281990
Rafael Espindola [Tue, 20 Sep 2016 17:14:16 +0000 (17:14 +0000)]
Don't produce an error for undefined entry symbol.
This is particularly important when the symbol comes from a linker
script. It is common to use the same linker script for shared
libraries and executables. Without this we would always fail to link
shared libraries with -z,defs and a linker script with an ENTRY
directive.
llvm-svn: 281989
Vedant Kumar [Tue, 20 Sep 2016 17:11:18 +0000 (17:11 +0000)]
[docs] Extend the code coverage docs some more
Flesh out the section on interpreting coverage reports, mention the
coverage export feature, and add notes on how to build llvm with
coverage turned on.
llvm-svn: 281988
Tobias Grosser [Tue, 20 Sep 2016 17:05:22 +0000 (17:05 +0000)]
[ScopDetection] Remove redundant checks for endless loops
Summary:
Both `canUseISLTripCount()` and `addOverApproximatedRegion()` contained checks
to reject endless loops which are now removed and replaced by a single check
in `isValidLoop()`.
For reporting such loops the `ReportLoopOverlapWithNonAffineSubRegion` is
renamed to `ReportLoopHasNoExit`. The test case
`ReportLoopOverlapWithNonAffineSubRegion.ll` is adapted and renamed as well.
The schedule generation in `buildSchedule()` is based on the following
assumption:
Given some block B that is contained in a loop L and a SESE region R,
we assume that L is contained in R or the other way around.
However, this assumption is broken in the presence of endless loops that are
nested inside other loops. Therefore, in order to prevent erroneous behavior
in `buildSchedule()`, r265280 introduced a corresponding check in
`canUseISLTripCount()` to reject endless loops. Unfortunately, it was possible
to bypass this check with -polly-allow-nonaffine-loops which was fixed by adding
another check to reject endless loops in `allowOverApproximatedRegion()` in
r273905. Hence there existed two separate locations that handled this case.
Thank you Johannes Doerfert for helping to provide the above background
information.
Reviewers: Meinersbur, grosser
Subscribers: _jdoerfert, pollydev
Differential Revision: https://reviews.llvm.org/D24560
Contributed-by: Matthias Reisinger <d412vv1n@gmail.com>
llvm-svn: 281987
Saleem Abdulrasool [Tue, 20 Sep 2016 17:05:04 +0000 (17:05 +0000)]
X86: loosen an overly aggressive MachO assertion
We would assert that the FP setup CFI used esp/rsp always. This held up in
practice when the code was generated from IR. However, with the integrated
assembler, it is possible to have the input be user specified assembly. In such
a case, we cannot assume that the function implementation has a compact unwind
representation. Loosen the assertion into a check and bail if we cannot
represent the frame pointer in the compact unwinding.
Addresses PR30453!
llvm-svn: 281986
Rafael Espindola [Tue, 20 Sep 2016 17:02:38 +0000 (17:02 +0000)]
Redirect stderr to /dev/null. NFC.
Makes the error output less confusing when this test fails.
llvm-svn: 281985
Davide Italiano [Tue, 20 Sep 2016 16:57:02 +0000 (16:57 +0000)]
[Writer] Fix a typo. NFC.
llvm-svn: 281984
Eric Christopher [Tue, 20 Sep 2016 16:05:02 +0000 (16:05 +0000)]
Remove more guts of TargetMachine::getNameWithPrefix and migrate one check to the TLOF mach-o version.
NFC intended.
llvm-svn: 281983
Eric Christopher [Tue, 20 Sep 2016 16:04:59 +0000 (16:04 +0000)]
Remove a use of subtarget initialization in the X86 backend so we can get rid of the default subtarget.
NFC intended.
llvm-svn: 281982
Eric Christopher [Tue, 20 Sep 2016 16:04:50 +0000 (16:04 +0000)]
Remove extra argument used once on TargetMachine::getNameWithPrefix and inline the result into the singular caller.
llvm-svn: 281981
Keith Walker [Tue, 20 Sep 2016 16:04:31 +0000 (16:04 +0000)]
Improve the -debug output for Debug Range Extension (NFC)
Include header messages and remove unnecessary blank lines.
llvm-svn: 281980
Nick Lewycky [Tue, 20 Sep 2016 15:49:58 +0000 (15:49 +0000)]
Replace 'isProvablyNonNull' with existing utility llvm::IsKnownNonNull which handles more cases. Noticed by inspection.
Because of how the IR generation works, this isn't expected to cause an observable difference.
llvm-svn: 281979
Rafael Espindola [Tue, 20 Sep 2016 15:22:27 +0000 (15:22 +0000)]
Disable --rosegment when we have linker scripts.
Linker scripts are responsible for aliging '.'. Since they are
designed for bfd which has no --rosegment, they don't align the RO to
RX transition.
llvm-svn: 281978
Tim Northover [Tue, 20 Sep 2016 15:20:36 +0000 (15:20 +0000)]
GlobalISel: split aggregates for PCS lowering
This should match the existing behaviour for passing complicated struct and
array types, in particular HFAs come through like that from Clang.
For C & C++ we still need to somehow support all the weird ABI flags, or at
least those that are present in the IR (signext, byval, ...), and stack-based
parameter passing.
llvm-svn: 281977
Rafael Espindola [Tue, 20 Sep 2016 15:08:24 +0000 (15:08 +0000)]
Make tests less dependent on the exact layout.
In most cases that means just not checking the address when we don't
need it.
For some tests it is easier to just set . to a known value.
llvm-svn: 281976
Simon Dardis [Tue, 20 Sep 2016 15:07:36 +0000 (15:07 +0000)]
[mips] MSA intrinsics header file
This patch adds the msa.h header file containing the shorter names for the
MSA instrinsics, e.g. msa_sll_b for builtin_msa_sll_b.
Reviewers: vkalintiris, zoran.jovanovic
Differential Review: https://reviews.llvm.org/D24674
llvm-svn: 281975
Rafael Espindola [Tue, 20 Sep 2016 14:49:18 +0000 (14:49 +0000)]
Move test to the correct directory. NFC.
llvm-svn: 281974
Simon Pilgrim [Tue, 20 Sep 2016 14:42:45 +0000 (14:42 +0000)]
[X86][SSE] Regenerate multiple combine tests
llvm-svn: 281973
Sanjay Patel [Tue, 20 Sep 2016 14:36:14 +0000 (14:36 +0000)]
move variables closer to their uses; add FIXMEs; NFC
llvm-svn: 281972
Dean Michael Berris [Tue, 20 Sep 2016 14:35:57 +0000 (14:35 +0000)]
[XRay] ARM 32-bit no-Thumb support in compiler-rt
This is a port of XRay to ARM 32-bit, without Thumb support yet.
This is one of 3 commits to different repositories of XRay ARM port. The other 2 are:
https://reviews.llvm.org/D23931 (LLVM)
https://reviews.llvm.org/D23932 (Clang test)
Differential Revision: https://reviews.llvm.org/D23933
llvm-svn: 281971
Dmitry Vyukov [Tue, 20 Sep 2016 13:30:01 +0000 (13:30 +0000)]
tsan: check more addresses in CheckShadowMapping
There is still a handful of them, so should not slow down
tsan apps. But gives assurance if we change/complicate
shadow mappings.
llvm-svn: 281970
Dmitry Vyukov [Tue, 20 Sep 2016 13:28:20 +0000 (13:28 +0000)]
tsan: make CHECK more robust
Enable more ignores when we start crashing.
Unwind in CHECK SIGSEGVs if happens early:
FATAL: ThreadSanitizer CHECK failed: ../projects/compiler-rt/lib/tsan/rtl/tsan_platform_posix.cc:105 "((beg)) <= ((end))" (0x8000000000, 0x4000000000)
Program received signal SIGSEGV, Segmentation fault.
__tsan::MetaMap::GetAndLock (this=0x1337c88 <__tsan::ctx_placeholder+8>, thr=thr@entry=0x7ffff7f91800, pc=pc@entry=4639488, addr=addr@entry=
140737339086992, write_lock=write_lock@entry=true,
create=create@entry=true) at ../projects/compiler-rt/lib/tsan/rtl/tsan_sync.cc:208
208 u32 idx0 = *meta;
(gdb) bt
#0 __tsan::MetaMap::GetAndLock (this=0x1337c88 <__tsan::ctx_placeholder+8>, thr=thr@entry=0x7ffff7f91800, pc=pc@entry=4639488, addr=addr@entry=
140737339086992, write_lock=write_lock@entry=true,
create=create@entry=true) at ../projects/compiler-rt/lib/tsan/rtl/tsan_sync.cc:208
#1 0x00000000004a965f in __tsan::MetaMap::GetOrCreateAndLock (this=<optimized out>, thr=thr@entry=0x7ffff7f91800, pc=pc@entry=4639488, addr=addr@entry=
140737339086992, write_lock=write_lock@entry=true)
at ../projects/compiler-rt/lib/tsan/rtl/tsan_sync.cc:198
#2 0x00000000004a162a in __tsan::Release (thr=0x7ffff7f91800, pc=pc@entry=4639488, addr=addr@entry=
140737339086992) at ../projects/compiler-rt/lib/tsan/rtl/tsan_rtl_mutex.cc:395
#3 0x000000000046cc40 in __interceptor_pthread_once (o=0x7ffff71a5890 <once_regsizes>, f=0x7ffff6f9d9c0 <init_dwarf_reg_size_table>) at ../projects/compiler-rt/lib/tsan/rtl/tsan_interceptors.cc:1334
#4 0x00007ffff6f9fe86 in __gthread_once (__func=0x7ffff6f9d9c0 <init_dwarf_reg_size_table>, __once=0x7ffff71a5890 <once_regsizes>) at ./gthr-default.h:699
#5 uw_init_context_1 (context=context@entry=0x7fffffffd6d0, outer_cfa=outer_cfa@entry=0x7fffffffd980, outer_ra=0x437d13 <__sanitizer::BufferedStackTrace::SlowUnwindStack(unsigned long, unsigned int)+67>)
at ../../../src/libgcc/unwind-dw2.c:1572
#6 0x00007ffff6fa06a8 in _Unwind_Backtrace (trace=0x437c30 <__sanitizer::Unwind_Trace(_Unwind_Context*, void*)>, trace_argument=0x7fffffffd980) at ../../../src/libgcc/unwind.inc:283
#7 0x0000000000437d13 in __sanitizer::BufferedStackTrace::SlowUnwindStack (this=0x7ffff6103208, pc=pc@entry=4863574, max_depth=max_depth@entry=256)
at ../projects/compiler-rt/lib/sanitizer_common/sanitizer_unwind_linux_libcdep.cc:125
#8 0x0000000000434f4a in __sanitizer::BufferedStackTrace::Unwind (this=this@entry=0x7ffff6103208, max_depth=max_depth@entry=256, pc=pc@entry=4863574, bp=bp@entry=0, context=context@entry=0x0,
stack_top=stack_top@entry=0, stack_bottom=stack_bottom@entry=0, request_fast_unwind=request_fast_unwind@entry=false) at ../projects/compiler-rt/lib/sanitizer_common/sanitizer_stacktrace_libcdep.cc:76
#9 0x00000000004a36b3 in PrintCurrentStackSlow (pc=4863574) at ../projects/compiler-rt/lib/tsan/rtl/tsan_rtl_report.cc:696
#10 __tsan::TsanCheckFailed (file=<optimized out>, line=<optimized out>, cond=<optimized out>, v1=<optimized out>, v2=<optimized out>) at ../projects/compiler-rt/lib/tsan/rtl/tsan_rtl_report.cc:44
#11 0x000000000042dfd6 in __sanitizer::CheckFailed (file=file@entry=0x4b9fd0 "../projects/compiler-rt/lib/tsan/rtl/tsan_platform_posix.cc", line=line@entry=105,
cond=cond@entry=0x4ba049 "((beg)) <= ((end))", v1=v1@entry=
549755813888, v2=v2@entry=
274877906944) at ../projects/compiler-rt/lib/sanitizer_common/sanitizer_termination.cc:79
#12 0x00000000004aa36c in ProtectRange (end=
274877906944, beg=
549755813888) at ../projects/compiler-rt/lib/tsan/rtl/tsan_platform_posix.cc:105
#13 __tsan::CheckAndProtect () at ../projects/compiler-rt/lib/tsan/rtl/tsan_platform_posix.cc:133
#14 0x00000000004a9e95 in __tsan::InitializePlatform () at ../projects/compiler-rt/lib/tsan/rtl/tsan_platform_linux.cc:280
#15 0x0000000000497e73 in __tsan::Initialize (thr=0x7ffff7f91800) at ../projects/compiler-rt/lib/tsan/rtl/tsan_rtl.cc:343
#16 0x00007ffff7dea25a in _dl_init (main_map=0x7ffff7ffe1c8, argc=1, argv=0x7fffffffdb88, env=0x7fffffffdb98) at dl-init.c:111
#17 0x00007ffff7ddb30a in _dl_start_user () at rtld.c:871
llvm-svn: 281969
Rafael Espindola [Tue, 20 Sep 2016 13:12:07 +0000 (13:12 +0000)]
Remove empty section commands.
We were already not creating them, and with this other parts of the
code don't have to worry about them.
llvm-svn: 281968
Artem Tamazov [Tue, 20 Sep 2016 11:58:40 +0000 (11:58 +0000)]
[AMDGPU][mc] Add regression tests for Bug 28168
llvm-svn: 281967
Elena Demikhovsky [Tue, 20 Sep 2016 11:02:26 +0000 (11:02 +0000)]
AVX-512: Fixed a bug in lowering saturated operations on KNL.
The generated code is still not optimal.
Differential Revision: https://reviews.llvm.org/D24723
llvm-svn: 281966
Valery Pykhtin [Tue, 20 Sep 2016 10:41:16 +0000 (10:41 +0000)]
[AMDGPU] Refactor VOP3 instruction TD definitions
Differential revision: https://reviews.llvm.org/D24664
llvm-svn: 281965
Keith Walker [Tue, 20 Sep 2016 10:36:17 +0000 (10:36 +0000)]
Make llvm::ConvertDebugDeclareToDebugValue() be a void function (NFC)
The routines llvm::ConvertDebugDeclareToDebugValue() always returned
a true value which was never checked at the call site; change the
function return type to void.
This NFC cleanup was approved in the review https://reviews.llvm.org/D23715
llvm-svn: 281964
Nikolay Haustov [Tue, 20 Sep 2016 09:04:51 +0000 (09:04 +0000)]
AMDGPU: Improve documentation.
Summary:
Add links to ISA manuals and ABI.
Add text about assembler syntax.
Add info about instructions operands.
Add instruction examples for each encoding.
Update directives section, add missing .amdgpu_hsa_kernel.
Reviewers: tstellarAMD, SamWot, vpykhtin
Subscribers: kzhuravl, wdng, nhaehnle, yaxunl, tony-tye, artem.tamazov, llvm-commits
Differential Revision: https://reviews.llvm.org/D24724
llvm-svn: 281962
Dorit Nuzman [Tue, 20 Sep 2016 08:27:48 +0000 (08:27 +0000)]
Reverting revision 281960 due to test failures.
llvm-svn: 281961
Dorit Nuzman [Tue, 20 Sep 2016 07:50:49 +0000 (07:50 +0000)]
[SROA] Preserve llvm.mem.parallel_loop_access metadata.
SROA doesn't preserve the llvm.mem.parallel_loop_access metadata when it
transforms loads/stores. This patch fixes a couple occurences of this
issue.
(Partially addresses PR28981).
Differential Revision: https://reviews.llvm.org/D23549
llvm-svn: 281960
Craig Topper [Tue, 20 Sep 2016 06:49:17 +0000 (06:49 +0000)]
[AVX-512] Teach X86InstrInfo::copyPhysReg to use a 512-bit move if XMM16-XMM31 or YMM16-YMM31 are the source or dest of the copy and VLX is not supported.
This can happen with SUBREG_TO_REG of ZMM16-ZMM31. Fixes PR30430.
llvm-svn: 281959
Craig Topper [Tue, 20 Sep 2016 05:44:47 +0000 (05:44 +0000)]
[AVX-512] Use 512-bit vcvtps2ph/vcvtph2ps to implement fp_to_f16/f16_to_fp when F16C and VLX are not supported.
Fixes PR23941.
llvm-svn: 281958
Matthias Braun [Tue, 20 Sep 2016 01:14:42 +0000 (01:14 +0000)]
BranchFolder: Fix invalid undef flags after merge.
It is legal to merge instructions with different undef flags; However we
must drop the undef flag from the merged instruction if it isn't present
everywhere.
This fixes http://llvm.org/PR30199
llvm-svn: 281957
Matthias Braun [Tue, 20 Sep 2016 01:14:39 +0000 (01:14 +0000)]
Machine{Instr|Operand}: Clarify some isIdenticalTo() subtleties.
llvm-svn: 281956
Quentin Colombet [Tue, 20 Sep 2016 00:48:44 +0000 (00:48 +0000)]
[RegisterBankInfo] Avoid heap allocation in InstructionMapping.
Use SmallVector instead of dynamically allocated arrays for the mapping of the
operands in the InstructionMapping. That way we avoid heap allocation for most
of the cases. Ultimately, we should not have to rely on such tricky, the
instances of InstructionMapping would be TableGen'ed.
This improves the compilation time of the RegBankSelect pass.
llvm-svn: 281955
NAKAMURA Takumi [Tue, 20 Sep 2016 00:44:45 +0000 (00:44 +0000)]
clang-change-namespace: Update libdeps.
FIXME: It seems clangFormat is not referred.
llvm-svn: 281954
Sanjay Patel [Tue, 20 Sep 2016 00:27:22 +0000 (00:27 +0000)]
[x86] fix variable names; NFC
llvm-svn: 281953
Kostya Serebryany [Tue, 20 Sep 2016 00:16:54 +0000 (00:16 +0000)]
[sanitizer-coverage] add comdat to coverage guards if needed
llvm-svn: 281952
Rui Ueyama [Tue, 20 Sep 2016 00:02:06 +0000 (00:02 +0000)]
Remove a use of std::list.
llvm-svn: 281951
Sanjay Patel [Mon, 19 Sep 2016 23:44:50 +0000 (23:44 +0000)]
[x86] auto-generate checks
llvm-svn: 281950
Philip Reames [Mon, 19 Sep 2016 23:30:23 +0000 (23:30 +0000)]
[LCSSA] Cache LoopExits to avoid wasted work
When looking at the scribus_1.3 example from https://llvm.org/bugs/show_bug.cgi?id=10584, I noticed that we were spending a large amount of time computing loop exits in LCSSA. This code appears to be written with the assumption that LoopExits are stored in the Loop and thus cheap to query. This is not true, so we should cache the result across the potentially long running loop which tends to visit a small handful of Loops.
On the particular example from 10584, this change drops the time spent in LCSSA computation by about 80%.
Differential Revision: https://reviews.llvm.org/D24509
llvm-svn: 281949