Commits · ce7196b9033feed933127544f1b8af560d58cd53 · Roger Ferrer / llvm-epi-0.8 · GitLab

This project is archived. Its data is read-only.

May 04, 2009
- -Remove the DwarfWriter::RecordSourceLine calls from the instruction selectors. · ce7196b9
  Argyrios Kyrtzidis authored May 03, 2009
  ce7196b9
May 03, 2009
- Fix typo · 2745bc92
  Anton Korobeynikov authored May 03, 2009
  2745bc92
- Properly handle sdiv / udiv / srem / urem libcalls · 05b7a7c8
  Anton Korobeynikov authored May 03, 2009
  05b7a7c8
- Proper name 16 bit libcalls · 399ad444
  Anton Korobeynikov authored May 03, 2009
  399ad444
- Add libcall expansion for 16 and 128 bit muls · f3fc92d6
  Anton Korobeynikov authored May 03, 2009
  f3fc92d6
- -Move the DwarfWriter::ValidDebugInfo check to a static DIDescriptor::ValidDebugInfo · 97324cec
  Argyrios Kyrtzidis authored May 03, 2009
  97324cec
May 01, 2009
- Allow CONCAT_VECTORS nodes to be legal or have custom lowering for some targets. · 62a3124f
  Bob Wilson authored May 01, 2009
  62a3124f
- Make DebugLoc independent of DwarfWriter. · a5037484
  Argyrios Kyrtzidis authored Apr 30, 2009
  a5037484
Apr 30, 2009
- Move helper functions for optimizing division by constant into the APInt · fe0c648f
  Jay Foad authored Apr 30, 2009
  fe0c648f
- fix a regression handling indirect results: these need to be considered · 5ab42e93
  Chris Lattner authored Apr 30, 2009
  5ab42e93
- Instead of passing in an unsigned value for the optimization level, use an enum, · 026e5d76
  Bill Wendling authored Apr 29, 2009
  026e5d76
- Fix infinite recursion in the C++ code which handles movddup by making it unnecessary. · 7e6e3527
  Nate Begeman authored Apr 29, 2009
  7e6e3527
Apr 29, 2009
- Update comment, replace theoretically impossible check with an assert. · 39b59db2
  Nate Begeman authored Apr 29, 2009
  39b59db2
- Implement review feedback for vector shuffle work. · 5f829d89
  Nate Begeman authored Apr 29, 2009
  5f829d89
- Add a public method called getAddressSpace() to the GlobalAddressSDNode. · ccd30945
  Sanjiv Gupta authored Apr 29, 2009
  ccd30945
- Disable the load-shrinking optimization from looking at · 7d103861
  Chris Lattner authored Apr 29, 2009
  7d103861
- Second attempt: · 084669a1
  Bill Wendling authored Apr 29, 2009
  084669a1
Apr 28, 2009
- Move getSubRegisterRegClass from ScheduleDagSDNodesEmit.cpp to a TargetRegisterClass method. · 604248e8
  Jakob Stoklund Olesen authored Apr 28, 2009
  604248e8
- r70270 isn't ready yet. Back this out. Sorry for the noise. · 56f2987a
  Bill Wendling authored Apr 28, 2009
  56f2987a
- Massive check in. This changes the "-fast" flag to "-O#" in llc. If you want to · d0ae1594
  Bill Wendling authored Apr 28, 2009
  d0ae1594
Apr 27, 2009
- Now that PR2957 is resolved, remove a bunch of · bfa03770
  Duncan Sands authored Apr 27, 2009
  bfa03770
- 2nd attempt, fixing SSE4.1 issues and implementing feedback from duncan. · 8d6d4b92
  Nate Begeman authored Apr 27, 2009
  8d6d4b92
- When transforming sext(trunc(load(x))) into sext(smaller load(x)), · be36f5cc
  Dan Gohman authored Apr 27, 2009
  be36f5cc
Apr 25, 2009
- Refactor the code to grab the low and high parts of a value · fe9e1d5b
  Dan Gohman authored Apr 25, 2009
  fe9e1d5b
- Add a top-level comment about DAGCombiner's role in the compiler. · 45399879
  Dan Gohman authored Apr 25, 2009
  45399879
Apr 24, 2009
- Fix PR 4057, a crash doing float->char const folding. · 56cb14c8
  Dale Johannesen authored Apr 24, 2009
  56cb14c8
- Revert 69952. Causes testsuite failures on linux x86-64. · b93db668
  Rafael Espindola authored Apr 24, 2009
  b93db668
- PR2957 · bb881d66
  Nate Begeman authored Apr 24, 2009
  bb881d66
- Instead of requiring TLI.LowerCallTo to return an ISD::BUILD_PAIR, · 640a161c
  Dan Gohman authored Apr 24, 2009
  640a161c
- Factor out a bit of code that appears in several places into a · 9478c3f8
  Dan Gohman authored Apr 23, 2009
  9478c3f8
- Handle Void types in ComputeValueVTs. This doesn't currently occur, · a290ab44
  Dan Gohman authored Apr 23, 2009
  a290ab44
Apr 21, 2009
- Make X86's copyRegToReg able to handle copies to and from subclasses. · 1addf647
  Dan Gohman authored Apr 20, 2009
  1addf647
- Simplify this code. getConstant knows how to make · e014b699
  Dan Gohman authored Apr 20, 2009
  e014b699
Apr 20, 2009
- Revise my previous change 68996 as suggested by Duncan. · da188ebb
  Bob Wilson authored Apr 20, 2009
  da188ebb
Apr 19, 2009
- Now that BUILD_VECTOR operands are allowed to be · f2e7133d
  Duncan Sands authored Apr 19, 2009
  f2e7133d
Apr 18, 2009
- Fix PR3898, which manifests as failures on are an Xcore, · 7b01e664
  Chris Lattner authored Apr 18, 2009
  7b01e664
- Don't try to make BUILD_VECTOR operands have the same · e4ff21ba
  Duncan Sands authored Apr 18, 2009
  e4ff21ba
- Inline asm's were still introducing bogus dependencies; · ad968ee2
  Dale Johannesen authored Apr 18, 2009
  ad968ee2
Apr 16, 2009
- In the list-burr's pseudo two-addr dependency heuristics, don't · eefba6bb
  Dan Gohman authored Apr 16, 2009
  eefba6bb
- Handle SUBREG_TO_REG instructions with the same heuristics · 3027bb69
  Dan Gohman authored Apr 16, 2009
  3027bb69