Commits · e6c6a945ca93432c930cf255dbf55069377d4426 · Roger Ferrer / llvm-epi · GitLab

Feb 17, 2017
- Fix windows bots by locking down the target triple on this testcase. · e6c6a945
  Adrian Prantl authored Feb 17, 2017
  e6c6a945
- AMDGPU: Fix crashes on invalid icmp/fcmp intrinsics · f6cf1032
  Matt Arsenault authored Feb 17, 2017
  f6cf1032
- WholeProgramDevirt: For VCP use a 32-bit ConstantInt for the byte offset. · 184773d8
  Peter Collingbourne authored Feb 17, 2017
  184773d8
- Debug Info: Sort frame index expressions before emitting them. · 67c24422
  Adrian Prantl authored Feb 17, 2017
  67c24422
- [X86][SHA] Add SHA stack folding tests · a3f28039
  Simon Pilgrim authored Feb 17, 2017
  a3f28039
- In Thumb1 mode, the custom lowering for ARMISD::CMPZ could never emit tADDi3 · 4592f620
  Artyom Skrobov authored Feb 17, 2017
  4592f620
- [X86][TBM] Add TBM stack folding tests · f4f5cd5d
  Simon Pilgrim authored Feb 17, 2017
  f4f5cd5d
- GlobalISel: verify that generic loads & stores have a mem operand. · 88634996
  Tim Northover authored Feb 17, 2017
  88634996
- [AArch64] Add Cavium ThunderX support · ab0f3b43
  Joel Jones authored Feb 17, 2017
  ab0f3b43
- WholeProgramDevirt: Examine the function body when deciding whether functions are readnone. · 37317f12
  Peter Collingbourne authored Feb 17, 2017
  37317f12
- [X86][BMI] Add BMI2 stack folding tests · 99193de8
  Simon Pilgrim authored Feb 17, 2017
  99193de8
- opt: Rename -default-data-layout flag to -data-layout and make it always override the layout. · 10c500dd
  Peter Collingbourne authored Feb 17, 2017
  10c500dd
- MC/COFF: Do not emit forward associative section referenceds. · ac20c179
  Rui Ueyama authored Feb 17, 2017
  ac20c179
- [X86][BMI] Add BMI stack folding tests · 09dde435
  Simon Pilgrim authored Feb 17, 2017
  09dde435
- [PowerPC] add tests for select-of-constants; NFC · f2a345c8
  Sanjay Patel authored Feb 17, 2017
  f2a345c8
- [ARM] add tests for select-of-constants; NFC · 9b6cfaa7
  Sanjay Patel authored Feb 17, 2017
  9b6cfaa7
- [LV] Remove constant restriction for vector phi creation · f68e183f
  Matthew Simpson authored Feb 17, 2017
  f68e183f
- [ARM] Replace HasT2ExtractPack with HasDSP · 58af0c55
  Sam Parker authored Feb 17, 2017
  58af0c55
- [DAGCombine] Recognise any_extend_vector_inreg and truncation style shuffle masks · 511d788a
  Simon Pilgrim authored Feb 17, 2017
  511d788a
- [ARM] GlobalISel: Use Subtarget in Legalizer · 7cab0786
  Diana Picus authored Feb 17, 2017
  7cab0786
- [ARM] GlobalISel: Add end-to-end tests for double · d2f3ba71
  Diana Picus authored Feb 17, 2017
  d2f3ba71
- InstCombine: fix extraction when performing vector/array punning · 958fcd75
  Eugene Leviant authored Feb 17, 2017
  958fcd75
- [JumpThreading] Re-enable JumpThreading for guards · 8b859c26
  Sanjoy Das authored Feb 17, 2017
  8b859c26
- Revert "[Hexagon] Start using regmasks on calls" · 6eab4044
  Rafael Espindola authored Feb 17, 2017
  6eab4044
- [XRAY] [x86_64] Adding a Flight Data filetype reader to the llvm-xray Trace implementation. · 4f83c4d1
  Dean Michael Berris authored Feb 17, 2017
  4f83c4d1
- Move test to X86 subdirectory for bot failures · 95ed51dc
  Teresa Johnson authored Feb 17, 2017
  95ed51dc
- [x86] Give this test a triple so that we don't have to cope with two · 96d86a7f
  Chandler Carruth authored Feb 17, 2017
  96d86a7f
- Attempt to fix bot failures by adding -mtriple to llc invocation · 1ede03d5
  Teresa Johnson authored Feb 17, 2017
  1ede03d5
- Bug 31948: Fix assertion when bitcasting constantexpr pointers · c18b6774
  Matt Arsenault authored Feb 17, 2017
  c18b6774
- FileCheck-ize some tests in test/CodeGen/X86/ · 89606869
  Chandler Carruth authored Feb 17, 2017
  89606869
- Handle link of NoDebug CU with a CU that has debug emission enabled · 76b5b749
  Teresa Johnson authored Feb 17, 2017
  76b5b749
Feb 16, 2017
- [LSR] Prevent formula with SCEVAddRecExpr type of Reg from Sibling loops · 493fb266
  Wei Mi authored Feb 16, 2017
  493fb266
- [Hexagon] Start using regmasks on calls · fb9503c0
  Krzysztof Parzyszek authored Feb 16, 2017
  fb9503c0
- [X86][SSE] Add PR31309 test case (load-extend i32 to i128). · e5215751
  Simon Pilgrim authored Feb 16, 2017
  e5215751
- AMDGPU: Remove llvm.AMDGPU.cube intrinsic · b95ddd7c
  Matt Arsenault authored Feb 16, 2017
  b95ddd7c
- AMDGPU: Remove llvm.AMDGPU.rsq intrinsic · eb65cda9
  Matt Arsenault authored Feb 16, 2017
  eb65cda9
- Re-apply r282920 "X86: Allow conditional tail calls in Win64 "leaf" functions (PR26302)" · 35905d6a
  Hans Wennborg authored Feb 16, 2017
  35905d6a
- Refactor DebugHandlerBase a bit to common non-debug-having-function filtering · b2fbb4b2
  David Blaikie authored Feb 16, 2017
  b2fbb4b2
- InstCombine: Canonicalize fast fmuladd to fmul + fadd · 92057604
  Matt Arsenault authored Feb 16, 2017
  92057604
- x86 interrupt calling convention: only save xmm registers if the target supports SSE · 42f7712e
  Andrea Di Biagio authored Feb 16, 2017
  42f7712e