Commits · db5a0a5818750ebcd075ee2b3f56c87447c0c411 · Roger Ferrer / llvm-epi · GitLab

Jul 24, 2013
- Update testing cases to pass debug info verifier. · db5a0a58
  Manman Ren authored Jul 24, 2013
  db5a0a58
- Add -disable-debug-info-verifier to a RUN line. · f0ae066e
  Rafael Espindola authored Jul 24, 2013
  f0ae066e
- Fix a problem I introduced in r187029 where we would over-eagerly · 58e25d39
  Chandler Carruth authored Jul 24, 2013
  58e25d39
- Fix PR16687 where we were incorrectly promoting an alloca that had · 83ea195d
  Chandler Carruth authored Jul 24, 2013
  83ea195d
- Update testing cases to make them pass debug info verification. · 3f93d3d6
  Manman Ren authored Jul 24, 2013
  3f93d3d6
- DAGCombiner: Pass the correct type to TargetLowering::isF(Abs|Neg)Free · c54731aa
  Tom Stellard authored Jul 23, 2013
  c54731aa
- R600: Treat CONSTANT_ADDRESS loads like GLOBAL_ADDRESS loads when necessary · 8cb0e47c
  Tom Stellard authored Jul 23, 2013
  8cb0e47c
- Debug Info: improve the Finder. · 8f1a3cf4
  Manman Ren authored Jul 23, 2013
  8f1a3cf4
- [ARM][ISel] Improve the lowering of vector loads. · 0f2fe74a
  Quentin Colombet authored Jul 23, 2013
  0f2fe74a
Jul 23, 2013
- Add not so that these tests pass with pipefail enabled. · 91997c0f
  Rafael Espindola authored Jul 23, 2013
  91997c0f
- Add -disable-debug-info-verifier. · 2dfb1cfd
  Rafael Espindola authored Jul 23, 2013
  2dfb1cfd
- Add an initial implementation of archive symbol table generation. · ce7f52db
  Rafael Espindola authored Jul 23, 2013
  ce7f52db
- llvm/test/DebugInfo/X86/dbg-value-range.ll: Add -disable-debug-info-verifier, too, since r186902. · 2e13d10c
  NAKAMURA Takumi authored Jul 23, 2013
  2e13d10c
- Don't let x86 asm printer use the no operand movsd alias. It should use the normal movsl instead. · db90f65b
  Craig Topper authored Jul 23, 2013
  db90f65b
- R600: Add support for 24-bit MAD instructions · 5263948a
  Tom Stellard authored Jul 23, 2013
  5263948a
- R600: Add support for 24-bit MUL instructions · 41fc7853
  Tom Stellard authored Jul 23, 2013
  41fc7853
- R600: Improve support for < 32-bit loads · 9f95033d
  Tom Stellard authored Jul 23, 2013
  9f95033d
- R600: Move CONST_ADDRESS folding into AMDGPUDAGToDAGISel::Select() · 84021443
  Tom Stellard authored Jul 23, 2013
  84021443
- R600: Use KCache for kernel arguments · 1e80309e
  Tom Stellard authored Jul 23, 2013
  1e80309e
- R600: Use the same compute kernel calling convention for all GPUs · acfeebf8
  Tom Stellard authored Jul 23, 2013
  acfeebf8
- R600: Use correct LoadExtType when lowering kernel arguments · 78e01296
  Tom Stellard authored Jul 23, 2013
  78e01296
- R600: Clean up extended load patterns · 33dd04bf
  Tom Stellard authored Jul 23, 2013
  33dd04bf
- R600: Expand vector FNEG · beed74af
  Tom Stellard authored Jul 23, 2013
  beed74af
- Debug Info Finder: use processDeclare and processValue to list debug info · 9974c88f
  Manman Ren authored Jul 23, 2013
  9974c88f
- When we vectorize across multiple basic blocks we may vectorize PHINodes that... · cf0dcdc7
  Nadav Rotem authored Jul 22, 2013
  cf0dcdc7
Jul 22, 2013
- Fix the move to/from accumulator register instructions that use a full 64-bit · 285da020
  Kevin Enderby authored Jul 22, 2013
  285da020
- Fix typo in test. · d07dd7f2
  Jakub Staszak authored Jul 22, 2013
  d07dd7f2
- Recommit r186813: More Intel syntax alias fixes. With the addition of... · 998bcf95
  Craig Topper authored Jul 22, 2013
  998bcf95
- Recommit r186217 with testcase fix: · c02a0aab
  Bill Wendling authored Jul 22, 2013
  c02a0aab
- [mips] Fix MipsAsmParser::parseCCRRegs. · 4d2ea3c6
  Akira Hatanaka authored Jul 22, 2013
  4d2ea3c6
- Mark ELF's first symbol as format specific so that llvm-nm ignores it. · 0d6b347f
  Rafael Espindola authored Jul 22, 2013
  0d6b347f
- [mips] Use ADDu instead of OR to copy general purpose registers. Also, delete · 44ff81d4
  Akira Hatanaka authored Jul 22, 2013
  44ff81d4
- This adds range checking for "ldr Rn, [pc, #imm]" Thumb · 8a9da5b0
  Mihai Popa authored Jul 22, 2013
  8a9da5b0
- Replace archive members in the old position. · 623c3d84
  Rafael Espindola authored Jul 22, 2013
  623c3d84
- Don't depend on a llvm-ar specific behavior. · 9cea44a5
  Rafael Espindola authored Jul 22, 2013
  9cea44a5
- [NVPTX] Use approximate FP ops when unsafe-fp-math is used, and append · cd069e6d
  Justin Holewinski authored Jul 22, 2013
  cd069e6d
- Treat nothrow forms of ::operator delete and ::operator delete[] as · 70523c79
  Richard Smith authored Jul 21, 2013
  70523c79
Jul 21, 2013
- Handle replacement into a position past the original member. · 9cd2435b
  Rafael Espindola authored Jul 21, 2013
  9cd2435b
- Don't crash when llvm.compiler.used becomes empty. · c2bb73fc
  Rafael Espindola authored Jul 20, 2013
  c2bb73fc
Jul 20, 2013
- Disallow global aliases to bitcast between address spaces · 828b565c
  Matt Arsenault authored Jul 20, 2013
  828b565c