Commits · 374a3ac74490eba5be81f2bcccab83085161abde · Lorenzo Albano / LLVM bpEVL · GitLab

Mar 05, 2010
- inline a small function with one call site. · 374a3ac7
  Chris Lattner authored Mar 05, 2010
  374a3ac7
- Revert r97778 and r97779. They're somehow breaking llvm-gcc builds. · 998c7c26
  Dan Gohman authored Mar 05, 2010
  998c7c26
- Fix these constants to be more portable. · ba9eb0bf
  Dan Gohman authored Mar 05, 2010
  ba9eb0bf
- Rewrite i64-to-f64 conversion using an algorithm which handles · 7fbeeeba
  Dan Gohman authored Mar 05, 2010
  7fbeeeba
- Fix some more places where dbg_value affected codegen. · 2061c841
  Dale Johannesen authored Mar 05, 2010
  2061c841
Mar 04, 2010
- Fix memcheck-found leaks: one false positive from using new[], and one true · 735b0ae2
  Jeffrey Yasskin authored Mar 04, 2010
  735b0ae2
- For SJLJ exception handling, make sure that all calls that are not marked · 1201f293
  Jim Grosbach authored Mar 04, 2010
  1201f293
- Run machine licm before machine cse to avoid messing up licm opportunities. · a325e562
  Evan Cheng authored Mar 04, 2010
  a325e562
- Avoid cse load instructions unless they are known to be invariant loads. · 1abd1a9f
  Evan Cheng authored Mar 04, 2010
  1abd1a9f
- add a statistic for # times fastisel fails. · c1cb75eb
  Chris Lattner authored Mar 04, 2010
  c1cb75eb
- Fix a typo Duncan noticed. · 9cc886b9
  Dan Gohman authored Mar 04, 2010
  9cc886b9
- Rename -machine-cse to -enable-machine-cse. · c58aea00
  Evan Cheng authored Mar 04, 2010
  c58aea00
- Look ahead a bit to determine if a physical register def that is not marked... · 36f8aabb
  Evan Cheng authored Mar 04, 2010
  36f8aabb
- change the new isel matcher to emit ComplexPattern matches · 0acbb71b
  Chris Lattner authored Mar 04, 2010
  0acbb71b
- Fix more code to work properly with vector operands. Based on · e14c4087
  Dan Gohman authored Mar 04, 2010
  e14c4087
- inline CannotYetSelectIntrinsic into CannotYetSelect and simplify. · 878b3e46
  Chris Lattner authored Mar 04, 2010
  878b3e46
- Fix a logic error. An instruction that has a live physical register def cannot... · 2922641a
  Evan Cheng authored Mar 03, 2010
  2922641a
- Remove PHINodeTraits and use MachineInstrExpressionTrait instead. · 2d23779e
  Evan Cheng authored Mar 03, 2010
  2d23779e
- Move MachineInstrExpressionTrait::getHashValue() out of line so it can skip... · 59d27fe5
  Evan Cheng authored Mar 03, 2010
  59d27fe5
- Re-apply r97667 but with a little bit of thought put into the patch. This... · 0abbb399
  Evan Cheng authored Mar 03, 2010
  0abbb399
Mar 03, 2010
- Revert 97667. It broke a bunch of tests. · a791914e
  Dan Gohman authored Mar 03, 2010
  a791914e
- Fix funky indentation and add comments. · 0f260e17
  Evan Cheng authored Mar 03, 2010
  0f260e17
- Move DenseMapInfo for MachineInstr* to MachineInstr.h · 0dd0c47b
  Evan Cheng authored Mar 03, 2010
  0dd0c47b
- Fix a bug in SelectionDAG's ReplaceAllUsesWith in the case where · 7d099f7e
  Dan Gohman authored Mar 03, 2010
  7d099f7e
- Machine CSE work in progress. It's doing some CSE now. But implicit def of... · b386cd38
  Evan Cheng authored Mar 03, 2010
  b386cd38
- Add MachineRegisterInfo::hasOneUse and hasOneNonDBGUse. · f94d6839
  Evan Cheng authored Mar 03, 2010
  f94d6839
- Revert... · 182eea8f
  Bill Wendling authored Mar 03, 2010
  182eea8f
- add some of the more obscure predicate types to the · dc1b6f79
  Chris Lattner authored Mar 03, 2010
  dc1b6f79
- speed up scope node processing: if the first element of a scope · 796f1da4
  Chris Lattner authored Mar 03, 2010
  796f1da4
- introduce a new SwitchTypeMatcher node (which is analogous to · 3e1ffd06
  Chris Lattner authored Mar 03, 2010
  3e1ffd06
- Work in progress. Finding some cse now. · 4eab008b
  Evan Cheng authored Mar 03, 2010
  4eab008b
- Use APInt instead of zext value. · c8d3add0
  Bill Wendling authored Mar 03, 2010
  c8d3add0
- - Change MachineInstr::isIdenticalTo to take a new option that determines... · e9c46c25
  Evan Cheng authored Mar 03, 2010
  e9c46c25
- Add an option to enable machine cse (it's not doing anything yet. · 68dabc70
  Evan Cheng authored Mar 03, 2010
  68dabc70
- This test case: · af13d829
  Bill Wendling authored Mar 03, 2010
  af13d829
Mar 02, 2010
- Fix some issues in WalkChainUsers dealing with · dd030701
  Chris Lattner authored Mar 02, 2010
  dd030701
- Okay. One last attempt: · ce153aff
  Bill Wendling authored Mar 02, 2010
  ce153aff
- run HandleMergeInputChains even if we only have one input chain. · 27a184b8
  Chris Lattner authored Mar 02, 2010
  27a184b8
- Swap parameters of isSafeToMove and isSafeToReMat for consistency. · 62e795ab
  Evan Cheng authored Mar 02, 2010
  62e795ab
- Fix typo. · 10194a42
  Evan Cheng authored Mar 02, 2010
  10194a42