Duncan P. N. Exon Smith [Sun, 7 Feb 2016 05:06:35 +0000 (05:06 +0000)]
LangRef: Fix example code for cmpxchg
Patch by Daniel Robertson!
llvm-svn: 260018
Saleem Abdulrasool [Sun, 7 Feb 2016 02:30:59 +0000 (02:30 +0000)]
Sema: handle typo correction on ARC'ed ivar
The ivar ref would be transformed by the Typo Correction TreeTransform, but not
be owned, resulting in the source location being invalid. This would eventually
lead to an assertion in findCapturingExpr. Prevent this assertion from
triggering.
Resolves PR25113.
llvm-svn: 260017
Saleem Abdulrasool [Sun, 7 Feb 2016 02:30:55 +0000 (02:30 +0000)]
Sema: handle typo correction with ARC'ed objc properties
We would previously assert in findCapturingExpr when performing a typo
correction resulting in an assignment of an ObjC property with a strong lifetype
specifier due to the expression not being rooted in the file (invalid SLoc)
during the retain cycle check on the typo-corrected expression. Handle the
expression type appropriately during the TreeTransform to ensure that we have a
source location associated with the expression.
Fixes PR26486.
llvm-svn: 260016
Daniel Berlin [Sun, 7 Feb 2016 02:03:39 +0000 (02:03 +0000)]
Don't use module context here. It's unnecessary and makes it harder to write unittests
llvm-svn: 260015
Daniel Berlin [Sun, 7 Feb 2016 01:52:19 +0000 (01:52 +0000)]
Compute live-in for MemorySSA
llvm-svn: 260014
Daniel Berlin [Sun, 7 Feb 2016 01:52:15 +0000 (01:52 +0000)]
Only insert into definingblocks once per block
llvm-svn: 260013
Eric Fiselier [Sun, 7 Feb 2016 00:36:33 +0000 (00:36 +0000)]
Cleanup node-type handling in the unordered containers
This patch is the first in a series of patches that's meant to better
support unordered_map. unordered_map has a special "value_type" that
differs from pair<const Key, Value>. In order to meet the EmplaceConstructible
and CopyInsertable requirements we need to teach __hash_table about this
special value_type.
This patch creates a "__hash_node_types" traits class that contains
all of the typedefs needed by the unordered containers and it's iterators.
These typedefs include ones for each node type and node pointer type,
as well as special typedefs for "unordered_map"'s value type.
As a result of this change all of the unordered containers now all support
incomplete types.
As a drive-by fix I changed the difference_type in __hash_table to always
be ptrdiff_t. There is a corresponding change to size_type but it cannot
take affect until an ABI break.
This patch will be followed up shortly with fixes for various unordered_map
fixes.
llvm-svn: 260012
Saleem Abdulrasool [Sat, 6 Feb 2016 22:36:34 +0000 (22:36 +0000)]
Index: provide adjustment thunk information for C++ manglings
Add support for exposing the adjustment thunk for virtual methods as
appropriate.
llvm-svn: 260011
Simon Pilgrim [Sat, 6 Feb 2016 22:09:25 +0000 (22:09 +0000)]
[X86][AVX2] Regenerated broadcast domain tests
llvm-svn: 260010
Tobias Grosser [Sat, 6 Feb 2016 21:23:39 +0000 (21:23 +0000)]
IslNodeBuilder: Invariant load hoisting of elements with differing sizes
Always use access-instruction pointer type to load the invariant values.
Otherwise mismatches between ScopArrayInfo element type and memory access
element type will result in invalid casts. These type mismatches are after
r259784 a lot more common and also arise with types of different size, which
have not been handled before.
Interestingly, this change actually simplifies the code, as we now have only
one code path that is always taken, rather then a standard code path for the
common case and a "fixup" code path that replaces the standard code path in
case of mismatching types.
llvm-svn: 260009
Simon Pilgrim [Sat, 6 Feb 2016 20:11:52 +0000 (20:11 +0000)]
[X86][SSE] Add tests for MOVHLPS/MOVLHPS shuffle lowering.
As raised in PR26491, we don't make use of these instructions at the moment.
llvm-svn: 260008
Simon Pilgrim [Sat, 6 Feb 2016 19:51:21 +0000 (19:51 +0000)]
[X86][AVX512] Added support for VPMOVZX shuffle decoding.
llvm-svn: 260007
Philip Reames [Sat, 6 Feb 2016 19:43:40 +0000 (19:43 +0000)]
[docs] Warn against slow serial builds
llvm-svn: 260006
Justin Lebar [Sat, 6 Feb 2016 19:32:44 +0000 (19:32 +0000)]
[NVPTX] Mark nvvm synchronizing intrinsics as convergent.
Summary:
This is the attribute purpose-made for e.g. __syncthreads. It appears
that NoDuplicate may not be sufficient to prevent Sink from touching a
call to __syncthreads.
Reviewers: jingyue, hfinkel
Subscribers: llvm-commits, jholewinski, jhen, rnk, tra, majnemer
Differential Revision: http://reviews.llvm.org/D16941
llvm-svn: 260005
Philip Reames [Sat, 6 Feb 2016 19:29:23 +0000 (19:29 +0000)]
[docs] Redirect new contributors to the right starting point
llvm-svn: 260004
Philip Reames [Sat, 6 Feb 2016 19:20:26 +0000 (19:20 +0000)]
[docs] Clarify a couple of getting started issues identified during Sprint
llvm-svn: 260003
Adrian Prantl [Sat, 6 Feb 2016 18:39:34 +0000 (18:39 +0000)]
Add a missing call to MDNode::deleteTemporary().
Follow-up to r259975. Kudos to the ASAN bots!
<rdar://problem/
24493203>
llvm-svn: 260002
Kamil Rytarowski [Sat, 6 Feb 2016 18:19:29 +0000 (18:19 +0000)]
Introduce NetBSD support
Current FreeBSD and NetBSD code is compatible.
llvm-svn: 260001
Simon Pilgrim [Sat, 6 Feb 2016 18:07:19 +0000 (18:07 +0000)]
[X86][AVX512] Fixed prefix ordering for lzcnt tests.
Let AVX512 targets share the same CHECKs.
llvm-svn: 260000
Simon Pilgrim [Sat, 6 Feb 2016 17:57:15 +0000 (17:57 +0000)]
[X86][SSE] Regenerate vector shift tests
llvm-svn: 259999
Devin Coughlin [Sat, 6 Feb 2016 17:17:32 +0000 (17:17 +0000)]
[analyzer] DeallocChecker: Don't warn on release of readonly assign property in dealloc.
It is common for the ivars for read-only assign properties to always be stored retained,
so don't warn for a release in dealloc for the ivar backing these properties.
llvm-svn: 259998
Simon Pilgrim [Sat, 6 Feb 2016 17:02:15 +0000 (17:02 +0000)]
[X86][SSE] Moved shuffle decode CASE macros earlier. NFC.
To allow the helper functions to make use of them.
llvm-svn: 259997
Simon Pilgrim [Sat, 6 Feb 2016 16:33:42 +0000 (16:33 +0000)]
[X86][SSE] Refactored PMOVZX shuffle decoding to use scalar input types
First step towards being able to decode AVX512 PMOVZX instructions without a massive bloat in the shuffle decode switch statement.
This should also make it easier to decode X86ISD::VZEXT target shuffles in the future.
llvm-svn: 259995
Kamil Rytarowski [Sat, 6 Feb 2016 16:23:18 +0000 (16:23 +0000)]
Adapt LLVM_CMAKE_PATH for recent cmake path changes
Current LLVM installs CMake files under lib/cmake/llvm.
llvm-svn: 259994
Teresa Johnson [Sat, 6 Feb 2016 16:07:35 +0000 (16:07 +0000)]
[ThinLTO] Include linkage type in function summary
Summary:
Adds the linkage type to both the per-module and combined function
summaries, which subsumes the current islocal bit. This will eventually
be used to optimized linkage types based on global summary-based
analysis.
Reviewers: joker.eph
Subscribers: joker.eph, davidxl, llvm-commits
Differential Revision: http://reviews.llvm.org/D16943
llvm-svn: 259993
Simon Pilgrim [Sat, 6 Feb 2016 15:38:25 +0000 (15:38 +0000)]
line endings fix
llvm-svn: 259992
Simon Pilgrim [Sat, 6 Feb 2016 15:37:09 +0000 (15:37 +0000)]
[X86][SSE] Don't replace an existing 32-bit load with its duplicate
If we are already loading a single 32-bit float/integer then just reuse it.
Fix for regression in D16729
llvm-svn: 259991
Simon Pilgrim [Sat, 6 Feb 2016 14:21:49 +0000 (14:21 +0000)]
Comment fix
llvm-svn: 259990
Ashutosh Nema [Sat, 6 Feb 2016 12:34:41 +0000 (12:34 +0000)]
Corrected tests for Loop Versioning LICM, by adding “REQUIRES: asserts”.
Earlier they were failing under no-assert build.
llvm-svn: 259989
Ashutosh Nema [Sat, 6 Feb 2016 09:24:37 +0000 (09:24 +0000)]
Fixed short underline error in LangRef.rst for recently added
metadata 'llvm.loop.licm_versioning.disable' description.
llvm-svn: 259988
Michael Kruse [Sat, 6 Feb 2016 09:19:40 +0000 (09:19 +0000)]
Follow uses to create value MemoryAccesses
The previously implemented approach is to follow value definitions and
create write accesses ("push defs") while searching for uses. This
requires the same relatively validity- and requirement conditions to be
replicated at multiple locations (PHI instructions, other instructions,
uses by PHIs).
We replace this by iterating over the uses in a SCoP ("pull in
requirements"), and add writes only when at least one read has been
added. It turns out to be simpler code because each use is only iterated
over once and writes are added for the first access that reads it. We
need another iteration to identify escaping values (uses not in the
SCoP), which also makes the difference between such accesses more
obvious. As a side-effect, the order of scalar MemoryAccess can change.
Differential Revision: http://reviews.llvm.org/D15706
llvm-svn: 259987
Ashutosh Nema [Sat, 6 Feb 2016 07:47:48 +0000 (07:47 +0000)]
New Loop Versioning LICM Pass
Summary:
When alias analysis is uncertain about the aliasing between any two accesses,
it will return MayAlias. This uncertainty from alias analysis restricts LICM
from proceeding further. In cases where alias analysis is uncertain we might
use loop versioning as an alternative.
Loop Versioning will create a version of the loop with aggressive aliasing
assumptions in addition to the original with conservative (default) aliasing
assumptions. The version of the loop making aggressive aliasing assumptions
will have all the memory accesses marked as no-alias. These two versions of
loop will be preceded by a memory runtime check. This runtime check consists
of bound checks for all unique memory accessed in loop, and it ensures the
lack of memory aliasing. The result of the runtime check determines which of
the loop versions is executed: If the runtime check detects any memory
aliasing, then the original loop is executed. Otherwise, the version with
aggressive aliasing assumptions is used.
The pass is off by default and can be enabled with command line option
-enable-loop-versioning-licm.
Reviewers: hfinkel, anemet, chatur01, reames
Subscribers: MatzeB, grosser, joker.eph, sanjoy, javed.absar, sbaranga,
llvm-commits
Differential Revision: http://reviews.llvm.org/D9151
llvm-svn: 259986
Samuel Antao [Sat, 6 Feb 2016 06:52:48 +0000 (06:52 +0000)]
Re-apply r259977 - [OpenMP] Reorganize code to allow specialized code generation for different devices.
This was reverted due to a failure in a buildbot, but it turned out the failure was unrelated.
llvm-svn: 259985
Samuel Antao [Sat, 6 Feb 2016 06:22:46 +0000 (06:22 +0000)]
Revert r259977 - [OpenMP] Reorganize code to allow specialized code generation for different devices.
It triggered some problem in the configuration related with zlib and exposed in the driver.
llvm-svn: 259984
Jason Molenda [Sat, 6 Feb 2016 04:55:26 +0000 (04:55 +0000)]
ProcessMachCore scans through the core file pages looking for a
user process dyld binary and/or a mach kernel binary image. By
default, it prefers the kernel if it finds both.
But if it finds two kernel binary images (which can happen when
random things are mapped into memory), it may pick the wrong
kernel image.
DynamicLoaderDarwinKernel has heuristics to find a kernel in memory;
once we've established that there is a kernel binary in memory,
call over to that class to see if it can find a kernel address via
its search methods. If it does, use that.
Some minor cleanups to DynamicLoaderDarwinKernel while I was at it.
<rdar://problem/
24446112>
llvm-svn: 259983
Kostya Serebryany [Sat, 6 Feb 2016 04:50:30 +0000 (04:50 +0000)]
[asan] trying to fix the non-x86 bots
llvm-svn: 259981
Kostya Serebryany [Sat, 6 Feb 2016 03:22:24 +0000 (03:22 +0000)]
[asan] properly report an un-aligned global variable instead of just crashing
llvm-svn: 259979
Kostya Serebryany [Sat, 6 Feb 2016 02:37:39 +0000 (02:37 +0000)]
[asan] disabled one subtest in scariness_score_test.cc that may fail when running under GNU make. Add more subtests
llvm-svn: 259978
Samuel Antao [Sat, 6 Feb 2016 02:12:34 +0000 (02:12 +0000)]
[OpenMP] Reorganize code to allow specialized code generation for different devices.
Summary:
Different devices may in some cases require different code generation schemes in order to implement OpenMP. This is required not only for performance reasons, but also because it may not be possible to have the current (default) implementation working for these devices. E.g. GPU's cannot implement the same scheme a target such as powerpc or x86b would use, in the sense that it does not have the ability to fork threads, instead all the threads are always executing and need to be managed by the implementation.
This patch proposes a reorganization of the code in the OpenMP code generation to pave the way to have specialized implementation of OpenMP support. More than a "real" patch this is more a request for comments in order to understand if what is proposed is acceptable or if there are better/easier ways to do it.
In this patch part of the common OpenMP codegen infrastructure is moved to a new file under a new namespace (CGOpenMPCommon) so it can be shared between the default implementation and the specialized one. When CGOpenMPRuntime is created, an attempt to select a specialized implementation is done.
In the patch a specialization for nvptx targets is done which currently checks if the target is an OpenMP device and trap if it is not.
Let me know comments suggestions you may have.
Reviewers: hfinkel, carlo.bertolli, arpith-jacob, kkwli0, ABataev
Subscribers: Hahnfeld, cfe-commits, fraggamuffin, caomhin, jholewinski
Differential Revision: http://reviews.llvm.org/D16784
llvm-svn: 259977
Richard Smith [Sat, 6 Feb 2016 02:06:43 +0000 (02:06 +0000)]
[modules] Compress files embedded into a .pcm file, to reduce the disk usage of -fembed-all-files mode.
llvm-svn: 259976
Adrian Prantl [Sat, 6 Feb 2016 01:59:09 +0000 (01:59 +0000)]
Fix a crash when emitting dbeug info for forward-declared scoped enums.
It is possible for enums to be created as part of their own
declcontext. We need to cache a placeholder to avoid the type being
created twice before hitting the cache.
<rdar://problem/
24493203>
llvm-svn: 259975
Adrian Prantl [Sat, 6 Feb 2016 01:56:55 +0000 (01:56 +0000)]
Relax assertion in ReplaceableMetadataImpl::replaceAllUsesWith().
There is a legitimate use-case in clang where we need to replace a
temporary placeholder node with the temporary node that may be a
forward declaration.
<rdar://problem/
24493203>
llvm-svn: 259973
Enrico Granata [Sat, 6 Feb 2016 01:36:07 +0000 (01:36 +0000)]
Per Jim's suggestion, move checks that we're not mixing and matching Debuggers and Commands deeper in the bowels of LLDB
NFC
llvm-svn: 259972
David Blaikie [Sat, 6 Feb 2016 01:15:26 +0000 (01:15 +0000)]
[llvm-dwp] Merge cu_index from DWPs
This is almost feature complete - just missing tu_index merging now.
llvm-svn: 259971
Devin Coughlin [Sat, 6 Feb 2016 00:59:14 +0000 (00:59 +0000)]
[www] Update analyzer release notes to correct the checker-278 build date.
This is not the future.
llvm-svn: 259969
Lang Hames [Sat, 6 Feb 2016 00:55:08 +0000 (00:55 +0000)]
[Orc] Slightly improve the x86-64 resolver block machine code.
Replace leaq + movq of a pointer with a single movabsq.
llvm-svn: 259968
Devin Coughlin [Sat, 6 Feb 2016 00:53:33 +0000 (00:53 +0000)]
[www] Update analyzer website for checker-278.
llvm-svn: 259967
Pete Cooper [Sat, 6 Feb 2016 00:51:16 +0000 (00:51 +0000)]
Set max segment protection level.
The initial segment protection was also being used to set the maximum
segment protection level. Instead, the maximum should be set according
to the architecture we are linking. For example on Mac OS it should be
RWX on most pages, but on iOS is often on R_X.
rdar://problem/
24515136
llvm-svn: 259966
Richard Smith [Sat, 6 Feb 2016 00:46:09 +0000 (00:46 +0000)]
llvm-bcanalyzer: Produce summary information for the BLOCKINFO block, it can be
a significant fraction of the file size (for files that otherwise have few
records). Also include an average size per record in the summary information.
llvm-svn: 259965
Enrico Granata [Sat, 6 Feb 2016 00:43:07 +0000 (00:43 +0000)]
Fix an issue where certain CommandObjects (or Options thereof) were being created once, bound to a specific CommandInterpreter (and hence a specific Debugger), and then cached for reuse across different Debugger instances
Obviously, if the original Debugger goes away, those commands are holding on to now stale memory, which has the potential to cause crashes
Fixes rdar://
24460882
llvm-svn: 259964
George Burgess IV [Sat, 6 Feb 2016 00:42:52 +0000 (00:42 +0000)]
Add note of suboptimal behavior in MemorySSA. NFC.
llvm-svn: 259963
Jim Ingham [Sat, 6 Feb 2016 00:31:23 +0000 (00:31 +0000)]
Fix "thread backtrace -s": option was misparsed because of a missing break.
<rdar://problem/
24525106>
llvm-svn: 259962
Kostya Serebryany [Sat, 6 Feb 2016 00:29:44 +0000 (00:29 +0000)]
[asan] add an experimental feature that prints the scariness score of the error message. To enable it use ASAN_OPTIONS=print_scariness=1
llvm-svn: 259961
Pete Cooper [Sat, 6 Feb 2016 00:14:15 +0000 (00:14 +0000)]
Move creation of "__LINKEDIT" segment. NFC.
We currently tag on a "__LINKEDIT" when we are emitting the segments.
However, an upcoming patch aims to set the initprot and maxprot segment members
to their correct values, and in order to share code, its better to create this
segment for real and handle it in buildFileOffsets the same way ld64 does.
The commit for segment protections will add a test for this all being correct so
no test here until that code is committed.
llvm-svn: 259960
Rafael Espindola [Sat, 6 Feb 2016 00:06:26 +0000 (00:06 +0000)]
Sort headers. NFC.
llvm-svn: 259959
Evandro Menezes [Sat, 6 Feb 2016 00:01:41 +0000 (00:01 +0000)]
[AArch64] Add the scheduling model for Exynos-M1
Summary:
Add the core scheduling model for the Samsung Exynos-M1 (ARMv8-A).
Reviewers: jmolloy, rengolin, christof, MinSeongKIM, t.p.northover
Subscribers: aemerson, rengolin, MatzeB
Differential Revision: http://reviews.llvm.org/D16644
llvm-svn: 259958
Sanjoy Das [Fri, 5 Feb 2016 23:40:04 +0000 (23:40 +0000)]
[StatepointLower] Use None instead of Optional<int>()
llvm-svn: 259956
Xinliang David Li [Fri, 5 Feb 2016 23:36:08 +0000 (23:36 +0000)]
[PGO] Test case update
Temporarily relax check in test to avoid
breakage for format change in LLVM side. Once that is
done, the test case will be retightened.
llvm-svn: 259955
Eric Christopher [Fri, 5 Feb 2016 23:28:03 +0000 (23:28 +0000)]
Make the OCaml tests temporarily unsupported until they can be updated.
llvm-svn: 259954
Lang Hames [Fri, 5 Feb 2016 23:27:48 +0000 (23:27 +0000)]
[Orc] Fix a typo in the comments for the x86_64 resolver block.
llvm-svn: 259953
Xinliang David Li [Fri, 5 Feb 2016 23:24:42 +0000 (23:24 +0000)]
Variable naming style fix /nfc
llvm-svn: 259952
Rui Ueyama [Fri, 5 Feb 2016 23:24:05 +0000 (23:24 +0000)]
ELF: Simplify readEntryLength.
I removed "CIE/FIE size is too large" error because that was not
checking for correct error conditions. [UINT_MAX - 4, UINT_MAX) is
a correct range as a size of a CIE/FDE record. It's just that the
size cannot be larger than the section size.
llvm-svn: 259951
Paul Robinson [Fri, 5 Feb 2016 23:23:25 +0000 (23:23 +0000)]
Eliminate an unnecessary enum, use the LLVM version. NFC
llvm-svn: 259950
Paul Robinson [Fri, 5 Feb 2016 23:20:02 +0000 (23:20 +0000)]
Add qualifiers that are about to be necessary. NFC
llvm-svn: 259949
Richard Smith [Fri, 5 Feb 2016 23:05:09 +0000 (23:05 +0000)]
Attempt#2 to work around MSVC rejects-valid.
llvm-svn: 259948
Richard Trieu [Fri, 5 Feb 2016 23:02:38 +0000 (23:02 +0000)]
Exempt char array initializers from -Wconstant-converion.
Sometimes, char arrays are used as bit storage, with no difference made between
signed and unsigned char. Thus, it is reasonable to use 0 to 255 instead of
-128 to 127 and not trigger this warning.
llvm-svn: 259947
Rui Ueyama [Fri, 5 Feb 2016 22:56:03 +0000 (22:56 +0000)]
ELF: Make EHOutputSection::readEntryLength a non-member function.
This function did not use any fields of the class.
llvm-svn: 259946
Rafael Espindola [Fri, 5 Feb 2016 22:55:06 +0000 (22:55 +0000)]
Fix PT_GNU_RELRO computation.
We were not including TLS sections.
llvm-svn: 259945
Artem Belevich [Fri, 5 Feb 2016 22:54:05 +0000 (22:54 +0000)]
[CUDA] Bug 26497 : Remove wrappers for variants provided by CUDA headers.
... and pull global-scope ones into std namespace with using-declaration.
Differential Revision: http://reviews.llvm.org/D16932
llvm-svn: 259944
Richard Smith [Fri, 5 Feb 2016 22:48:19 +0000 (22:48 +0000)]
Attempt to work around an MSVC rejects-valid. Apparently it gets the access
check wrong when inheriting a member through two levels of private inheritance,
where the middle one is a class template specialization.
llvm-svn: 259943
Richard Smith [Fri, 5 Feb 2016 22:32:52 +0000 (22:32 +0000)]
More workarounds for undefined behavior exposed when compiling in C++14 with
-fsized-deallocation. Disable sized deallocation for all objects derived from
TrailingObjects, as we expect the storage allocated for these objects to be
larger than the size of their dynamic type.
llvm-svn: 259942
Xinliang David Li [Fri, 5 Feb 2016 22:32:01 +0000 (22:32 +0000)]
[PGO] Speed up name tab reading
The change allows skipping duplicate strings
early to avoid redundant md5 computation and
string copying/swapping.
llvm-svn: 259941
Davide Italiano [Fri, 5 Feb 2016 22:10:42 +0000 (22:10 +0000)]
[llvm-nm] Prefer empty() over size() == 0.
Thanks to David Blaikie for pointing this out!
llvm-svn: 259938
Davide Italiano [Fri, 5 Feb 2016 22:07:09 +0000 (22:07 +0000)]
[llvm-nm] Transform a switch() statement in a pair of if(s).
This is more uniform wrt what other tools do and makes the code
a little bit more readable.
llvm-svn: 259937
Paul Robinson [Fri, 5 Feb 2016 21:54:42 +0000 (21:54 +0000)]
Move DebugInfoKind enum from Driver to Basic. NFC
llvm-svn: 259935
John McCall [Fri, 5 Feb 2016 21:37:38 +0000 (21:37 +0000)]
Add an ARC autoreleased-return-value caller marker on i386.
rdar://
24531556
llvm-svn: 259932
Ulrich Weigand [Fri, 5 Feb 2016 21:34:28 +0000 (21:34 +0000)]
[SystemZ] Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP macros
Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_[1248] macros on SystemZ.
This fixes a miscompile of GCC C++11 standard library headers
due to use of those macros in an ABI-changing manner.
See e.g. /usr/include/c++/4.8.5/ext/concurrence.h:
// Compile time constant that indicates prefered locking policy in
// the current configuration.
static const _Lock_policy __default_lock_policy =
#ifdef __GTHREADS
#if (defined(__GCC_HAVE_SYNC_COMPARE_AND_SWAP_2) \
&& defined(__GCC_HAVE_SYNC_COMPARE_AND_SWAP_4))
_S_atomic;
#else
_S_mutex;
#endif
#else
_S_single;
#endif
A different choice of __default_lock_policy causes different
sizes of several of the C++11 data structures, which are then
incompatible when inlined in clang-compiled code with what the
(GCC-compiled) external library expects.
This in turn leads to various crashes when using std::thread
in code compiled with clang, as see e.g. via the ThreadPool
unit tests. See PR 26473 for an example.
llvm-svn: 259931
Davide Italiano [Fri, 5 Feb 2016 21:10:48 +0000 (21:10 +0000)]
[llvm-nm] Simplify code logic. NFCI.
llvm-svn: 259917
Sunil Srivastava [Fri, 5 Feb 2016 20:50:02 +0000 (20:50 +0000)]
Do not honor explicit alignment attribute on fields for PS4.
This change reverts r257462 for PS4 triple.
Differential Revision: http://reviews.llvm.org/D16788
llvm-svn: 259916
Hans Wennborg [Fri, 5 Feb 2016 20:32:42 +0000 (20:32 +0000)]
CallAnalyzer::analyzeCall: change the condition back to "Cost < Threshold"
In r252595, I inadvertently changed the condition to "Cost <= Threshold",
which caused a significant size regression in Chrome. This commit rectifies
that.
llvm-svn: 259915
Jun Bum Lim [Fri, 5 Feb 2016 20:02:03 +0000 (20:02 +0000)]
[AArch64] Refactoring aarch64-ldst-opt. NCF.
Remove narrow load / store instructions from getMatchingPairOpcode(),
and add getMatchingWideOpcode().
llvm-svn: 259914
Tom Stellard [Fri, 5 Feb 2016 19:59:33 +0000 (19:59 +0000)]
TableGen: Add IsOptional field to AsmOperandClass
Summary:
This makes it possible to specify some operands as optional to the AsmMatcher.
Setting this field to true will prevent the AsmMatcher from emitting
'too few operands' errors when there are missing optional operands.
Reviewers: olista01, ab
Subscribers: nhaustov, arsenm, llvm-commits
Differential Revision: http://reviews.llvm.org/D15755
llvm-svn: 259913
Matt Arsenault [Fri, 5 Feb 2016 19:47:29 +0000 (19:47 +0000)]
AMDGPU: Account for LDS alignment
The current situation isn't great, because the amount of padding
requires is determined by the inverse order of the first encountered
use. We should eventually somehow sort these to minimize wasted space.
Another problem is the alignment of kernel arguments isn't
respected. The group_segment_alignment is always emitted as
the default 16, and typed arguments with higher alignments
or an explicitly set alignment are also ignored.
llvm-svn: 259912
Matt Arsenault [Fri, 5 Feb 2016 19:47:23 +0000 (19:47 +0000)]
AMDGPU: Preserve alignments on new created globals
Also switch to internal linkage, and include the name of the function in
the name.
llvm-svn: 259911
Bruno Cardoso Lopes [Fri, 5 Feb 2016 19:36:39 +0000 (19:36 +0000)]
[Parser] Perform CachedTokens update dependent on token consumption
In the context where we break one tok::greatergreater into two
tok::greater in order to correctly update the cached tokens; update the
CachedTokens with two tok::greater only if ParseGreaterThanInTemplateList
clients asks to consume the last token. Otherwise we only need to add
one because the second is already added later on, as a not yet cached token.
Differential Revision: http://reviews.llvm.org/D16906
rdar://problem/
24488367
llvm-svn: 259910
Reid Kleckner [Fri, 5 Feb 2016 19:15:45 +0000 (19:15 +0000)]
[codeview] Dump a missing field and change its signedness
llvm-svn: 259904
Rui Ueyama [Fri, 5 Feb 2016 19:13:18 +0000 (19:13 +0000)]
Simplify. NFC.
llvm-svn: 259903
Siva Chandra [Fri, 5 Feb 2016 19:10:04 +0000 (19:10 +0000)]
Take 2: Use an artifical namespace so that member vars do not hide local vars.
Summary: This relands r259810 with fix for failures on Mac.
Reviewers: spyffe, tfiala
Subscribers: tfiala, lldb-commits
Differential Revision: http://reviews.llvm.org/D16900
llvm-svn: 259902
Richard Smith [Fri, 5 Feb 2016 19:03:40 +0000 (19:03 +0000)]
[modules] Separately track whether an identifier's preprocessor information and
name lookup information have changed since deserialization. For a C++ modules
build, we do not need to re-emit the identifier into the serialized identifier
table if only the name lookup information has changed (and in all cases, we
don't need to re-emit the macro information if only the name lookup information
has changed).
llvm-svn: 259901
Tom Stellard [Fri, 5 Feb 2016 18:44:57 +0000 (18:44 +0000)]
AMDGPU: Remove some purely R600 functions from AMDGPUInstrInfo
Reviewers: arsenm
Subscribers: arsenm, llvm-commits
Differential Revision: http://reviews.llvm.org/D16862
llvm-svn: 259900
Rui Ueyama [Fri, 5 Feb 2016 18:41:40 +0000 (18:41 +0000)]
ELF: Make Out<ELFT> initialization less error-prone.
Previously, it was easy to leave some Out<ELFT> fields uninitialized
because assignments to the fields are mixed with output section
instantiations. In this patch, I separate initializations from assignments
to improve readability.
http://reviews.llvm.org/D16864
llvm-svn: 259899
Samuel Benzaquen [Fri, 5 Feb 2016 18:29:24 +0000 (18:29 +0000)]
[ASTMatchers] Allow hasName() to look through inline namespaces
Summary:
Allow hasName() to look through inline namespaces.
This will fix the interaction between some clang-tidy checks and libc++.
libc++ defines names in an inline namespace named std::<version_#>.
When we try to match a name using hasName("std::xxx") it fails to match and the clang-tidy check does not work.
Reviewers: klimek
Subscribers: klimek, cfe-commits
Differential Revision: http://reviews.llvm.org/D15506
llvm-svn: 259898
Tom Stellard [Fri, 5 Feb 2016 18:29:17 +0000 (18:29 +0000)]
AMDGPU: Fix ordering of CPU and FS parameters in TargetMachine constructors
Reviewers: arsenm
Subscribers: arsenm, llvm-commits
Differential Revision: http://reviews.llvm.org/D16863
llvm-svn: 259897
Reid Kleckner [Fri, 5 Feb 2016 18:21:28 +0000 (18:21 +0000)]
Fix echo.ll test failing due to DOS line endings
llvm-svn: 259896
Wei Mi [Fri, 5 Feb 2016 18:14:24 +0000 (18:14 +0000)]
Some stackslots are allocated to vregs which have no real reference.
LiveRangeEdit::eliminateDeadDef is used to remove dead define instructions
after rematerialization. To remove a VNI for a vreg from its LiveInterval,
LiveIntervals::removeVRegDefAt is used. However, after non-PHI VNIs are all
removed, PHI VNI are still left in the LiveInterval. Such unused vregs will
be kept in RegsToSpill[] at the end of InlineSpiller::reMaterializeAll and
spiller will allocate stackslot for them.
The fix is to get rid of unused reg by checking whether it has non-dbg
reference instead of whether it has non-empty interval.
llvm-svn: 259895
Tom Stellard [Fri, 5 Feb 2016 17:42:38 +0000 (17:42 +0000)]
AMDGPU/SI: Correctly initialize SIInsertWaits pass
Reviewers: arsenm
Subscribers: arsenm, llvm-commits
Differential Revision: http://reviews.llvm.org/D16724
llvm-svn: 259894
Dan Gohman [Fri, 5 Feb 2016 17:14:59 +0000 (17:14 +0000)]
[WebAssembly] Update the select instructions' operand orders to match the spec.
llvm-svn: 259893
Rafael Espindola [Fri, 5 Feb 2016 15:27:15 +0000 (15:27 +0000)]
Rename IsUsedInDynamicReloc to MustBeInDynSym.
The variable was marking various cases where a symbol must be included
in the dynamic symbol table. Being used by a dynamic relocation was only
one of them.
llvm-svn: 259889
Nemanja Ivanovic [Fri, 5 Feb 2016 15:03:17 +0000 (15:03 +0000)]
Add the missing test case for PR26193
llvm-svn: 259888
Rafael Espindola [Fri, 5 Feb 2016 15:03:10 +0000 (15:03 +0000)]
Centralize most calls to setUsedInDynamicReloc.
llvm-svn: 259887