summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
...
* AMDGPU: Fix breaking SMEM clausesMatt Arsenault2017-11-177-39/+415
| | | | | | | | This was completely ignoring subregisters, so was not very useful. Also only break them if xnack is actually enabled. llvm-svn: 318505
* Fix LLDB build.Eugene Zemtsov2017-11-171-0/+1
| | | | | | It was broken by r318489. llvm-svn: 318504
* [PCH+Modules] Improve diagnosticts to help out users pass an extra header ↵Bruno Cardoso Lopes2017-11-173-2/+28
| | | | | | | | | | | | | | | search path When mixing PCH and Implicit Modules, missing a header search path can lead to the implicit built PCM to complaint about not finding its matching module map. Instead of adding more magic to implicit modules engine, add a note to help the user add the appropriate path. rdar://problem/33388847 llvm-svn: 318503
* [llvm-profdata] Fix a dangling reference to an error stringVedant Kumar2017-11-176-3/+43
| | | | llvm-svn: 318502
* [GISel]: DCE copy instructions during legalizationAditya Nandakumar2017-11-172-9/+71
| | | | | | | | We might have instructions such as ext(copy(trunc)), and while cleaning up legalization artifacts, we can also dce the copies that are in between legalization artifacts. llvm-svn: 318501
* Update xcode project file to build new ppc64le files.Jason Molenda2017-11-173-3/+27
| | | | llvm-svn: 318500
* [SelectionDAG] Consolidate (t|T)ransferDbgValues methods, NFC (reapply)Vedant Kumar2017-11-173-45/+35
| | | | | | | | | | | | | | | | | | TransferDbgValues (capital 'T') is wired into ReplaceAllUsesWith, and transferDbgValues (lowercase 't') is used elsewhere (e.g in Legalize). Both functions should be doing the exact same thing. This patch consolidates the logic into one place. This was reverted in r318455 because some newly introduced asserts, which I thought were NFC, were firing. I filed PR35338. For now I've weakened the asserts. Testing: check-llvm, check-clang, and a stage2 Rel+Deb build of clang Differential Revision: https://reviews.llvm.org/D40104 llvm-svn: 318498
* Remove IsLocal.Rafael Espindola2017-11-178-70/+61
| | | | | | | Since we always have Binding in the current symbol design IsLocal is redundant. llvm-svn: 318497
* Recognize another magic token sent in the LaunchInfo's EventData.Jim Ingham2017-11-171-29/+71
| | | | | | <rdar://problem/32505028> llvm-svn: 318496
* "source list -a" was calling DisplaySourceLinesWithNumbers incorrectly.Jim Ingham2017-11-171-4/+5
| | | | | | | | The parameters were just passed in the wrong order. <rdar://problem/35150446> llvm-svn: 318495
* [CUDA] Remove implementations of nexttoward.Justin Lebar2017-11-172-23/+8
| | | | | | | | | | | | | | | | | | Summary: __builtin_nexttoward lowers to a libcall, e.g. nexttowardf(), that CUDA does not have. Rather than try to implement it, we simply remove these functions -- nvcc doesn't support them either, and nextafter, which does work, does essentially the same thing on GPUs, because GPUs don't have long double. Reviewers: tra Subscribers: cfe-commits, sanjoy Differential Revision: https://reviews.llvm.org/D40152 llvm-svn: 318494
* [asan] Test ASan instrumentation for shadow scale value of 5Walter Lee2017-11-1710-14/+34
| | | | | | | | | Add additional RUN clauses to test for -asan-mapping-scale=5 in selective tests, with special CHECK statements where needed. Differential Revision: https://reviews.llvm.org/D39775 llvm-svn: 318493
* Update for layering fix in LLVM CodeGen<>TargetDavid Blaikie2017-11-171-1/+1
| | | | llvm-svn: 318491
* Fix a bunch more layering of CodeGen headers that are in TargetDavid Blaikie2017-11-17410-702/+698
| | | | | | | | All these headers already depend on CodeGen headers so moving them into CodeGen fixes the layering (since CodeGen depends on Target, not the other way around). llvm-svn: 318490
* Don't #include MemoryBuffer.h from Host.h.Zachary Turner2017-11-1711-1/+12
| | | | | | | | | | It turns out this #include isn't used from Host.h anyway, but by having it it causes circular include dependencies. This issues only surfaced while I was working on a separate patch, so I'm submitting this first so that it's independent of the other, unrelated patch. llvm-svn: 318489
* [fuzzer] Initialize PcDescr buffer before calling __sanitizer_symbolize_pcVitaly Buka2017-11-171-1/+1
| | | | | | | __sanitizer_symbolize_pc is not instrumented so msan assumes that PcDescr was not initialized. llvm-svn: 318488
* [PredicateInfo] Add comment about why we require stable sortMandeep Singh Grang2017-11-171-0/+5
| | | | llvm-svn: 318487
* Re-revert "Refactor debuginfo-tests"Zachary Turner2017-11-1728-163/+29
| | | | | | | | | | | | | | | | | | | | | | This is still broken because it causes certain tests to be run twice with slightly different configurations, which is wrong in some cases. You can observe this by running: ninja -nv check-all | grep debuginfo-tests And seeing that it passes clang/test and clang/test/debuginfo-tests to lit, which causes it to run debuginfo-tests twice. The fix is going to involve either: a) figuring out that we're running in this "deprecated" configuration, and then deleting the clang/test/debuginfo-tests path, which should cause it to behave identically to before, or: b) make lit smart enough that it doesn't descend into a sub-suite if that sub-suite already has a lit.cfg file. llvm-svn: 318486
* Current implementation of Value::replaceUsesExceptBlockAddr() uses UseListDmitry Mikulin2017-11-172-1/+35
| | | | | | | | | | | | | iterator to walk the list which keeps changing inside the loop. When the UseList contains several uses with the same user, we end processing the same user more than once, which leads to an assert. With this fix, unique users are saved and processed later to avoid processing duplicates. Differential Revision: https://reviews.llvm.org/D39864 llvm-svn: 318477
* [libFuzzer] Fix flaky merge.test.Matt Morehouse2017-11-171-1/+1
| | | | llvm-svn: 318475
* PR22763: if a defaulted (non-user-provided) special member function isRichard Smith2017-11-164-2/+33
| | | | | | | | | | | | | | | | explicitly instantiated, still emit it with each use. We don't emit a definition of the member with an explicit instantiation definition (and indeed it appears that we're not allowed to, since an explicit instantiation definition does not constitute an odr-use and only odr-use permits definition for defaulted special members). So we still need to emit a weak definition with each use. This also makes defaulted-in-class declarations behave more like implicitly-declared special members, which matches their design intent. And it matches the way this problem was solved in GCC. llvm-svn: 318474
* [ARM] 't' asm constraint should accept i32Yi Kong2017-11-162-2/+9
| | | | | | | | | | | | | | | | | | | | 't' constraint normally only accepts f32 operands, but for VCVT the operands can be i32. LLVM is overly restrictive and rejects asm like: float foo() { float result; __asm__ __volatile__( "vcvt.f32.s32 %[result], %[arg1]\n" : [result]"=t"(result) : [arg1]"t"(0x01020304) ); return result; } Relax the value type for 't' constraint to either f32 or i32. Differential Revision: https://reviews.llvm.org/D40137 llvm-svn: 318472
* [asan] Properly mark or disable tests that only work with shadow scale of 3Walter Lee2017-11-1613-12/+19
| | | | | | Differential Revision: https://reviews.llvm.org/D39774 llvm-svn: 318471
* [asan] Port tests to shadow scale of 5Walter Lee2017-11-166-8/+26
| | | | | | | | | | | | | | | | | | | | | | | | | | The tests are ported as follows: contiguous_container_crash.cc use-after-delete.cc use-after-free.cc Replace hardwired shadow granularity in CHECK statements with regex. max_redzone.cc Bump max_redzone parameter to 32. memset_test.cc Bump size parameter of __asan_poison_memory_region to 32. scariness_score_test.cc For "far-from-bounds" heap overflow, make sure overflow is more than one shadow granularity away. At large shadow granularity, there is not enough redzone between stack elements to detect far-from-bounds, so fake out that test. Differential Revision: https://reviews.llvm.org/D39773 llvm-svn: 318470
* [asan] Add lit feature to indicate compiler_rt's shadow scale valueWalter Lee2017-11-161-0/+5
| | | | | | | | This will be used to mark tests that require a specific shadow scale. Differential Revision: https://reviews.llvm.org/D39772 llvm-svn: 318469
* [X86] Add DAG combine to remove sext i32->i64 from gather/scatter instructions.Craig Topper2017-11-162-256/+125
| | | | | | | | | | Only do this pre-legalize in case we're using the sign extend to legalize for KNL. This recovers all of the tests that changed when I stopped SelectionDAGBuilder from deleting sign extends. There's more work that could be done here particularly to fix the i8->i64 test case that experienced split. llvm-svn: 318468
* [X86] Add gather test with index sign extended from i8 type.Craig Topper2017-11-161-0/+63
| | | | | | | | Previously SelectionDAGBuilder would remove this sign extend leading to a failure during isel. The codegen here isn't very nice as we ended up triggering a split. llvm-svn: 318467
* [X86] Don't remove sign extend of gather/scatter indices during ↵Craig Topper2017-11-162-110/+255
| | | | | | | | | | | | | | SelectionDAGBuilder. The sign extend might be from an i16 or i8 type and was inserted by InstCombine to match the pointer width. X86 gather legalization isn't currently detecting this to reinsert a sign extend to make things legal. It's a bit weird for the SelectionDAGBuilder to do this kind of optimization in the first place. With this removed we can at least lean on InstCombine somewhat to ensure the index is i32 or i64. I'll work on trying to recover some of the test cases by removing sign extends in the backend when its safe to do so with an understanding of the current legalizer capabilities. This should fix PR30690. llvm-svn: 318466
* [libFuzzer] Make merge-sigusr.test work on Mac.Matt Morehouse2017-11-161-2/+1
| | | | llvm-svn: 318465
* [Support] Support NetBSD PaX MPROTECT in sys::Memory.Lang Hames2017-11-164-256/+30
| | | | | | | | | Removes AllocateRWX, setWritable and setExecutable from sys::Memory and standardizes on allocateMappedMemory / protectMappedMemory. The allocateMappedMemory method is updated to request full permissions for memory blocks so that they can be marked executable later. llvm-svn: 318464
* Attempt to fix inscrutible build break...David Blaikie2017-11-161-1/+1
| | | | llvm-svn: 318463
* Fix some undefined beahvior in FileMapping.Zachary Turner2017-11-161-2/+2
| | | | | | | This was broken when building a 32-bit native toolchain, as shifting a size_t right by 32 is UB when sizeof(size_t) == 8. llvm-svn: 318462
* llvm-readobj/ARMEHABIPrinter.h: Make this a real/modular headerDavid Blaikie2017-11-161-62/+79
| | | | | | | Had several non-inline/strong function definitions that needed to be marked inline, etc. llvm-svn: 318461
* [asan] Avoid assert failure for non-default shadow scaleWalter Lee2017-11-161-2/+2
| | | | | | | | | Rather than assertion failing, we can fall back to the non-optimized version which works for any shadow scale. Differential Revision: https://reviews.llvm.org/D39474 llvm-svn: 318460
* Convert the last use of sys::fs::createUniqueFile in bugpoint.Rafael Espindola2017-11-161-14/+8
| | | | llvm-svn: 318459
* Convert another use of createUniqueFile to TempFile::create.Rafael Espindola2017-11-166-12/+44
| | | | | | | This one requires a new small feature in TempFile: the ability to keep the temporary file with the temporary name. llvm-svn: 318458
* [CMake][runtimes] Use cmake_parse_arguments in runtimes functionsPetr Hosek2017-11-161-9/+18
| | | | | | | | | | | | Passing lists to functions in CMake is tricky, any list argument has to be quoted otherwise it'll be expanded. To avoid this issue, use cmake_parse_arguments in runtime functions and pass lists using a keyword argument which eliminates any ambiguity when dealing with lists. Differential Revision: https://reviews.llvm.org/D40087 llvm-svn: 318457
* Issue -Wempty-body warnings for else blocksReid Kleckner2017-11-163-1/+25
| | | | | | | | This looks like it was just an oversight. Fixes http://llvm.org/pr35319 llvm-svn: 318456
* Revert "[SelectionDAG] Consolidate (t|T)ransferDbgValues methods, NFC."Vedant Kumar2017-11-163-33/+45
| | | | | | | | | This reverts commit r318448. It looks like some of the asserts need to be weakened. http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-autoconf/builds/16296 llvm-svn: 318455
* [RISCV] Fix 64-bit data layout mismatch between backend and target descriptionMandeep Singh Grang2017-11-161-1/+1
| | | | | | | | | | | | Reviewers: asb Reviewed By: asb Subscribers: rbar, johnrusso, simoncook, jordy.potman.lists, llvm-commits Differential Revision: https://reviews.llvm.org/D40145 llvm-svn: 318454
* [libFuzzer] add more tests for mergeKostya Serebryany2017-11-161-0/+10
| | | | llvm-svn: 318453
* [X86] Pre-truncate gather/scatter indices that have element sizes larger ↵Craig Topper2017-11-162-2/+82
| | | | | | | | | | than 64-bits before Legalize. The wider element type will normally cause legalize to try to split and scalarize the gather/scatter, but we can't handle that. Instead, truncate the index early so the gather/scatter node is insulated from the legalization. This really shouldn't happen in practice since InstCombine will normalize index types to the same size as pointers. llvm-svn: 318452
* [X86] DAGCombinerInfo is in TargetLowering not X86TargetLowering.Craig Topper2017-11-161-1/+1
| | | | llvm-svn: 318451
* [DAGCombiner] Use cast instead of an unchecked dyn_cast.Craig Topper2017-11-161-1/+1
| | | | llvm-svn: 318450
* llvm-objdump: Add missing requires for the test in r313921Justin Bogner2017-11-161-0/+1
| | | | | | | This test needs an arm backend in addition to the arm64 one that's already checked in the test directory. llvm-svn: 318449
* [SelectionDAG] Consolidate (t|T)ransferDbgValues methods, NFC.Vedant Kumar2017-11-163-45/+33
| | | | | | | | | | | | TransferDbgValues (capital 'T') is wired into ReplaceAllUsesWith, and transferDbgValues (lowercase 't') is used elsewhere (e.g in Legalize). Both functions should be doing the exact same thing. This patch consolidates the logic into one place. Differential Revision: https://reviews.llvm.org/D40104 llvm-svn: 318448
* Fix my typo of PDB_TableTypeReid Kleckner2017-11-161-1/+1
| | | | llvm-svn: 318447
* Remove a FIXME about truncated section namesMartin Storsjo2017-11-161-2/+0
| | | | | | | | | | | | | | | | | If the linker chose to store the full section name instead of truncating it, this field doesn't contain a truncated name, but an offset into the string table of the binary. The string table isn't loaded/mapped into memory during runtime though, so it's not possible to read the full section name, unless we try to locate the DLL/EXE on disk that the HMODULE corresponds to and load that manually. Due to this, lld now always prefers writing a truncated section name for sections that will be mapped at runtime, even when debug info is enabled. Differential Revision: https://reviews.llvm.org/D39918 llvm-svn: 318446
* [VirtualFileSystem] Support creating directories then adding files insideBen Hamilton2017-11-163-10/+33
| | | | | | | | | | | | | | | | | | | | | | | | | | | Summary: In https://reviews.llvm.org/D39572 , I added support for specifying `Type` when invoking `InMemoryFileSystem::addFile()`. However, I didn't account for the fact that when `Type` is `directory_file`, we need to construct an `InMemoryDirectory`, not an `InMemoryFile`, or else clients cannot create files inside that directory. This diff fixes the bug and adds a test. Test Plan: New test added. Ran test with: % make -j12 check-clang-tools Reviewers: bkramer, hokein Reviewed By: bkramer Subscribers: cfe-commits Differential Revision: https://reviews.llvm.org/D40140 llvm-svn: 318445
* Fix -Wreturn-type falling off the end of a function in new DIA codeReid Kleckner2017-11-161-0/+1
| | | | llvm-svn: 318444
OpenPOWER on IntegriCloud