summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Don't implicitly construct an Archive::child_iterator.Rafael Espindola2015-11-051-6/+5
| | | | llvm-svn: 252166
* [tsan] Fix build warnings on OS XKuba Brecka2015-11-052-2/+12
| | | | | | | | Fixing `tsan_interceptors.cc`, which on OS X produces a bunch of warnings about unused constants and functions. Differential Revision: http://reviews.llvm.org/D14381 llvm-svn: 252165
* Fix a signed/unsigned mismatch warning; NFC.Aaron Ballman2015-11-051-1/+1
| | | | llvm-svn: 252164
* Lint warning fixup for r252160 ("[tsan] Fix pthread_once interceptor for OS X").Kuba Brecka2015-11-051-1/+1
| | | | llvm-svn: 252163
* [tsan] Fix the memcpy interceptor to be memmove compatible on OS XKuba Brecka2015-11-051-1/+4
| | | | | | | | On OS X, memcpy and memmove are actually aliases of the same implementation, which means the interceptor of memcpy is also invoked when memmove is called. The current implementation of the interceptor uses `internal_memcpy` to perform the actual memory operation, which can produce an incorrect result when memmove semantics are expected. Let's call `internal_memmove` instead. Differential Revision: http://reviews.llvm.org/D14336 llvm-svn: 252162
* [tsan] Allow memmove interceptor to be used when TSan is not initializedKuba Brecka2015-11-051-3/+5
| | | | | | | | A call to memmove is used early during new thread initialization on OS X. This patch uses the `COMMON_INTERCEPTOR_NOTHING_IS_INITIALIZED` check, similarly to how we deal with other early-used interceptors. Differential Revision: http://reviews.llvm.org/D14377 llvm-svn: 252161
* [tsan] Fix pthread_once interceptor for OS XKuba Brecka2015-11-051-1/+5
| | | | | | | | TSan has a re-implementation of `pthread_once` in its interceptor, which assumes that the `pthread_once_t *once_control` pointer is actually pointing to a "storage" which is zero-initialized and used for the atomic operations. However, that's not true on OS X, where pthread_once_t is a structure, that contains a header (with a magic value) and the actual storage follows after that. This patch skips the header to make the interceptor work on OS X. Differential Revision: http://reviews.llvm.org/D14379 llvm-svn: 252160
* [tsan] Alternative ThreadState storage for OS XKuba Brecka2015-11-054-2/+70
| | | | | | | | This implements a "poor man's TLV" to be used for TSan's ThreadState on OS X. Based on the fact that `pthread_self()` is always available and reliable and returns a valid pointer to memory, we'll use the shadow memory of this pointer as a thread-local storage. No user code should ever read/write to this internal libpthread structure, so it's safe to use it for this purpose. We lazily allocate the ThreadState object and store the pointer here. Differential Revision: http://reviews.llvm.org/D14288 llvm-svn: 252159
* Fix escaping value to subregion entry node phiMichael Kruse2015-11-052-21/+75
| | | | | | | | | An incoming value from a block the is not inside the scop is an external use, even if the phi is inside the scop. A previous fix in r251208 did not apply if the phi is inside a non-affine subregion. We move the check for this phi case before the non-affine subregion check. llvm-svn: 252157
* [x86] Additional small fix for MCU psABI supportAndrey Bokhanko2015-11-052-3/+7
| | | | | | | | This patch fixes one more thing in MCU psABI support: LongDoubleWidth should be set to 64. Differential Revision: http://reviews.llvm.org/D14285 llvm-svn: 252156
* [tsan] Use malloc zone interceptors on OS X, part 2Kuba Brecka2015-11-056-8/+72
| | | | | | | | TSan needs to use a custom malloc zone on OS X, which is already implemented in ASan. This patch uses the sanitizer_common implementation in `sanitizer_malloc_mac.inc` for TSan as well. Reviewed at http://reviews.llvm.org/D14330 llvm-svn: 252155
* revert rev. 252153 due to build failure on ubuntuAsaf Badouh2015-11-057-309/+1
| | | | | | [X86][AVX512] add comi with Sae llvm-svn: 252154
* [X86][AVX512] add comi with SaeAsaf Badouh2015-11-057-1/+309
| | | | | | | | add builtin_ia32_vcomisd and builtin_ia32_vcomisd Differential Revision: http://reviews.llvm.org/D14331 llvm-svn: 252153
* [SimplifyCFG] Tweak heuristic for merging conditional storesJames Molloy2015-11-051-7/+13
| | | | | | We were correctly skipping dbginfo intrinsics and terminators, but the initial bailout wasn't, causing it to bail out on almost any block. llvm-svn: 252152
* [X86][AVX512] small bugfix in VPBROADCASTMAsaf Badouh2015-11-052-2/+17
| | | | | | | | VPBROADCASTMW2D and VPBROADCASTMB2Q Differential Revision: http://reviews.llvm.org/D14335 llvm-svn: 252151
* RuntimeDyld: fix -Wtype-limitsSaleem Abdulrasool2015-11-051-2/+2
| | | | | | | Adjust the casted type. By casting to the same size rather than just the signed-ness, we were asserting tautological statements. NFC. llvm-svn: 252150
* Fix LoopAccessAnalysis when potentially nullptr check are involvedMehdi Amini2015-11-052-1/+44
| | | | | | | | | | | | | | Summary: GetUnderlyingObjects() can return "null" among its list of objects, we don't want to deduce that two pointers can point to the same memory in this case, so filter it out. Reviewers: anemet Subscribers: dexonsmith, llvm-commits From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 252149
* MCStreamer.h: Prune \return, corresponding to r252102. [-Wdocumentation]NAKAMURA Takumi2015-11-051-1/+0
| | | | llvm-svn: 252148
* Use profile data template file for covmap func record (NFC)Xinliang David Li2015-11-051-11/+10
| | | | llvm-svn: 252147
* Fix a bug exposed by uses in CFEXinliang David Li2015-11-051-2/+4
| | | | llvm-svn: 252146
* AMDGPU: Also track whether SGPRs were spilledMatt Arsenault2015-11-053-2/+20
| | | | llvm-svn: 252145
* AMDGPU: Print number user SGPRsMatt Arsenault2015-11-051-0/+6
| | | | | | | This doesn't quite match how SC prints it, which doesn't put it in a comment. llvm-svn: 252144
* Fix build for platforms without ObjC++.Chaoren Lin2015-11-052-6/+9
| | | | | | | | | | Reviewers: jingham, jasonmolenda Subscribers: lldb-commits Differential Revision: http://reviews.llvm.org/D14365 llvm-svn: 252143
* AMDGPU: Disallow s[102:103] on VI in assemblerMatt Arsenault2015-11-053-18/+74
| | | | llvm-svn: 252142
* [FunctionAttrs] Remove a loop, NFC refactorSanjoy Das2015-11-051-16/+14
| | | | | | | | | | | | | | | | | | | | | | Summary: Remove the loop over the uses of the CallSite in ArgumentUsesTracker. Since we have the `Use *` for actual argument operand, we can just use pointer subtraction. The time complexity remains the same though (except for a vararg argument) -- `std::advance` is O(UseIndex) for the ArgumentList iterator. The real motivation is to make a later change adding support for operand bundles simpler. Reviewers: reames, chandlerc, nlewycky Subscribers: llvm-commits Differential Revision: http://reviews.llvm.org/D14363 llvm-svn: 252141
* AMDGPU: Fix assert when legalizing atomic operandsMatt Arsenault2015-11-054-15/+111
| | | | | | | | | | The operand layout is slightly different for the atomic opcodes from the usual MUBUF loads and stores. This should only fix it on SI/CI. VI is still broken because it still emits the addr64 replacement. llvm-svn: 252140
* AMDGPU: Make addr64 atomic operand order consistentMatt Arsenault2015-11-051-2/+2
| | | | | | | vaddr comes before srsrc in every other MUBUF instruction, and is the order it is printed. llvm-svn: 252139
* Accommodate interface change in r252134.Alexander Kornienko2015-11-051-1/+2
| | | | llvm-svn: 252138
* Fix OSX build after r252118 (missing parameter for findModulesAndOffsets())Mehdi Amini2015-11-051-1/+2
| | | | | From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 252137
* Remove empty linesMehdi Amini2015-11-051-2/+2
| | | | | From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 252136
* [WinEH] Fix establisher param reg in CLR funcletsJoseph Tremoulet2015-11-052-9/+28
| | | | | | | | | | | | | | Summary: The CLR's personality routine passes the pointer to the establisher frame in RCX, not RDX. Reviewers: pgavlin, majnemer, rnk Subscribers: llvm-commits Differential Revision: http://reviews.llvm.org/D14343 llvm-svn: 252135
* Make ArgumentAdjuster aware of the current file being processed.Alexander Kornienko2015-11-055-12/+12
| | | | | | | | | | | | | | Summary: This is needed to handle per-project configurations when adding extra arguments in clang-tidy for example. Reviewers: klimek, djasper Subscribers: djasper, cfe-commits, klimek Differential Revision: http://reviews.llvm.org/D14191 llvm-svn: 252134
* Fix TestGoFormatters.py.Chaoren Lin2015-11-051-3/+2
| | | | llvm-svn: 252133
* [elf2] Fix R_X86_64_TPOFF32 handling.Michael J. Spencer2015-11-054-6/+11
| | | | | | | | For x86-64 the initial executable TLS block is placed directly before the thread specific data register so compilers can directly access it via R_X86_64_TPOFF32. Generate the correct (negative) offset for this case. llvm-svn: 252131
* [IR] Add bounds checking to dataOperandHasImpliedAttrSanjoy Das2015-11-051-0/+8
| | | | | | This is similar to the bounds check added to paramHasAttr in r252073. llvm-svn: 252130
* Add IRExecutionUnitSP to lldb-forward and use it instead of hand-rolling in ↵Jim Ingham2015-11-053-2/+4
| | | | | | place. llvm-svn: 252129
* TestModuleFileExtension.cpp: Include <cstdio> explicitly for fprintf(3).NAKAMURA Takumi2015-11-051-0/+1
| | | | llvm-svn: 252128
* Python 3 - Don't use `os.path.walk`, it's removed in Py3.Zachary Turner2015-11-051-4/+2
| | | | | | | | It was deprecated even in 2.7, but not removed until 3.x. os.walk provides all of the same functionality and is the correct way to do this now. llvm-svn: 252127
* Python 3 - Use universal_newlines=True in subprocess.Popen.Zachary Turner2015-11-051-0/+2
| | | | | | | | | | | This follows the spirit of a previous patch which did essentially the same thing. In Python 3, when you use Popen.communicate(), you get back a bytes object which cannot normally be treated as a string. We could decode this manually, but universal_newlines=True does this automatically, and there's no disadvantage to doing so even on Python 2. So just enable it always. llvm-svn: 252126
* [modules] Don't merge an anonymous enum definition into a named enum definition.Richard Smith2015-11-052-4/+21
| | | | llvm-svn: 252125
* Fix build on Linux.Chaoren Lin2015-11-051-0/+4
| | | | llvm-svn: 252124
* [libFuzzer] print a bit fewer linesKostya Serebryany2015-11-052-2/+3
| | | | llvm-svn: 252123
* CommandObjectProxy needs to proxy GenerateHelpText as well asJim Ingham2015-11-052-0/+11
| | | | | | | | everything else. <rdar://problem/18383591> llvm-svn: 252122
* [Sanitizer] Use ReportMmapFailureAndDie() in all applicable mmap variants on ↵Alexey Samsonov2015-11-054-16/+12
| | | | | | Posix. llvm-svn: 252121
* Fix build breakDavid Majnemer2015-11-051-2/+1
| | | | llvm-svn: 252120
* Go back to producing relocations for out of range symbols.Rafael Espindola2015-11-053-13/+9
| | | | | | | | This brings back the behavior from before r252090 for out of range symbols. Should bring some arm bots back. llvm-svn: 252119
* [Windows] Symbolize with llvm-symbolizer instead of dbghelp in a self-hostReid Kleckner2015-11-053-97/+227
| | | | | | | | | | | | | | | | Summary: llvm-symbolizer understands both PDBs and DWARF, so it is more likely to succeed at symbolization. If llvm-symbolizer is unavailable, we will fall back to dbghelp. This also makes our crash traces more similar between Windows and Linux. Reviewers: Bigcheese, zturner, chapuni Subscribers: llvm-commits Differential Revision: http://reviews.llvm.org/D12884 llvm-svn: 252118
* AMDGPU: Add missing v2f64 fadd testsMatt Arsenault2015-11-051-10/+42
| | | | llvm-svn: 252117
* AMDGPU: Fix typoMatt Arsenault2015-11-051-2/+2
| | | | llvm-svn: 252116
* [Lex] Add __has_builtin support for __make_integer_seqDavid Majnemer2015-11-052-1/+11
| | | | | | Differential Revision: http://reviews.llvm.org/D14349 llvm-svn: 252115
OpenPOWER on IntegriCloud