Commits · 3b2b1e794242b96121cc696dc72d35d911bc82f6 · Roger Ferrer / llvm-epi · GitLab

Oct 01, 2010
- FileCheck-ize this test. · 3b2b1e79
  Bill Wendling authored Oct 01, 2010
  3b2b1e79
- FileCheck-ize this test. · 9b6853c6
  Bill Wendling authored Oct 01, 2010
  9b6853c6
- fix rdar://8494845 + PR8244 - a miscompile exposed by my patch in r101350 · a2050558
  Chris Lattner authored Oct 01, 2010
  a2050558
- One more +sse2. · f419de08
  Dale Johannesen authored Oct 01, 2010
  f419de08
- Mark all these as needing SSE2. Should fix PPC and · bb6b9618
  Dale Johannesen authored Oct 01, 2010
  bb6b9618
- Disable these tests for now; it's not obvious why they fail on Linux. · ab60ae3c
  Dale Johannesen authored Oct 01, 2010
  ab60ae3c
- Make test not sensitive to register choice. · c6f17f74
  Dale Johannesen authored Oct 01, 2010
  c6f17f74
- Massive rewrite of MMX: · dd224d23
  Dale Johannesen authored Sep 30, 2010
  dd224d23
- test/CodeGen/X86/sibcall.ll: Add explicit triplets and remove XFAIL: apple-darwin8. · bb995ae2
  NAKAMURA Takumi authored Sep 30, 2010
  bb995ae2
Sep 30, 2010
- Try again to disable critical edge splitting in CodeGenPrepare. · eb12f49f
  Jakob Stoklund Olesen authored Sep 30, 2010
  eb12f49f
- When isel is emitting instructions for an x86 target without CMOV, the CFG is · 665aa6ef
  Jakob Stoklund Olesen authored Sep 30, 2010
  665aa6ef
Sep 29, 2010
- And remove r114997's test. · cc916012
  Bill Wendling authored Sep 28, 2010
  cc916012
- Revert r114997. It was causing a failure on darwin10-selfhost. · b0b2c571
  Bill Wendling authored Sep 28, 2010
  b0b2c571
- Fix a FIXME. _foo.eh symbols are currently always exported so that the linker · d848beb1
  Bill Wendling authored Sep 28, 2010
  d848beb1
Sep 27, 2010
- Revert "Disable codegen prepare critical edge splitting. Machine instruction passes now" · 415a7a6f
  Jakob Stoklund Olesen authored Sep 27, 2010
  415a7a6f
Sep 23, 2010
- Disable codegen prepare critical edge splitting. Machine instruction passes now · 794aaa79
  Evan Cheng authored Sep 23, 2010
  794aaa79
- A select between a constant and zero, when fed by a bit test, can be efficiently · 3231d13d
  Owen Anderson authored Sep 22, 2010
  3231d13d
- Fix PR8201: Update the code to call via X86::CALL64pcrel32 in the 64-bit case. · bbb92870
  Cameron Esfahani authored Sep 22, 2010
  bbb92870
Sep 22, 2010
- Fix an inconsistency in the x86 backend that led it to reject "calll foo" on · bd857253
  Chris Lattner authored Sep 22, 2010
  bd857253
- reimplement elf TLS support in terms of addressing modes, eliminating SegmentBaseAddress. · 8a236b63
  Chris Lattner authored Sep 22, 2010
  8a236b63
- linux has a different stack alignment than the mac, relax this a bit. · 505af598
  Chris Lattner authored Sep 22, 2010
  505af598
- give VZEXT_LOAD a memory operand, it now works with segment registers. · 54e53295
  Chris Lattner authored Sep 22, 2010
  54e53295
- revert r114386 now that address modes work correctly, we get a nice · 07827ba9
  Chris Lattner authored Sep 22, 2010
  07827ba9
- give LCMPXCHG_DAG[8] a memory operand, allowing it to work with addrspace 256/257 · e479e964
  Chris Lattner authored Sep 21, 2010
  e479e964
- filecheckize · 0cefa511
  Chris Lattner authored Sep 21, 2010
  0cefa511
Sep 21, 2010
- Use FileCheck · d92f42d1
  Devang Patel authored Sep 21, 2010
  d92f42d1
- When adding the carry bit to another value on X86, exploit the fact that the carry-materialization · f4b1a5bd
  Owen Anderson authored Sep 21, 2010
  f4b1a5bd
- fix rdar://8453210 , a crash handling a call through a GS relative load. · bb0a1c44
  Chris Lattner authored Sep 21, 2010
  bb0a1c44
- Enable machine sinking critical edge splitting. e.g. · f3e9a485
  Evan Cheng authored Sep 20, 2010
  f3e9a485
Sep 20, 2010
- When TCO is turned on, it is possible to end up with aliasing FrameIndex's. Therefore, · 272ff949
  Owen Anderson authored Sep 20, 2010
  272ff949
Sep 19, 2010
- test/CodeGen/X86: Add explicit triplet -mtriple=i686-linux to 3 tests... · b912c27f
  NAKAMURA Takumi authored Sep 19, 2010
  b912c27f
Sep 18, 2010
- Invert the logic of reachesChainWithoutSideEffects(). What we want to check is that there is · b92b13d8
  Owen Anderson authored Sep 18, 2010
  b92b13d8
- Teach machine sink to · e53ab6df
  Evan Cheng authored Sep 17, 2010
  e53ab6df
Sep 17, 2010
- Avoid emitting a PIC base register if no PIC addresses are needed. · 534db8a5
  Dan Gohman authored Sep 17, 2010
  534db8a5
Sep 16, 2010
- When substituting sunkaddrs into indirect arguments an asm, we were · f95f59a0
  Dale Johannesen authored Sep 16, 2010
  f95f59a0
Sep 09, 2010
- Add one more pattern to fallback movddup · e8501a46
  Bruno Cardoso Lopes authored Sep 09, 2010
  e8501a46
Sep 08, 2010
- remove these tests for now. · 3f4abf39
  Devang Patel authored Sep 07, 2010
  3f4abf39
Sep 07, 2010
- There is no need to force target if the test is going to run on other x86 platforms. · b0af23a1
  Devang Patel authored Sep 07, 2010
  b0af23a1
- Fix command line used to link these test cases. · e50b23e2
  Devang Patel authored Sep 07, 2010
  e50b23e2
- Reintroduce dbg-declare tests. · 9dc0e5be
  Devang Patel authored Sep 07, 2010
  9dc0e5be