Commits · 26070c5cffd4aa1d00c838d0327686b9bed8f230 · Roger Ferrer / llvm-epi-0.8

Oct 09, 2009
- when previous scratch register is killed, flag the value as no longer tracking · 26070c5c
  Jim Grosbach authored Oct 09, 2009
```
llvm-svn: 83653
```
  26070c5c
- Convert some ARM tests with lots of greps to use FileCheck. · 6dd3b9ad
  Bob Wilson authored Oct 09, 2009
```
llvm-svn: 83651
```
  6dd3b9ad
- Revert r83606 and add comments explaining why it isn't safe. · dc4893ab
  Dan Gohman authored Oct 09, 2009
```
llvm-svn: 83649
```
  dc4893ab
- As it turns out, the bug fixes in GC codegen did not make it · 31c4d8b0
  Nicolas Geoffray authored Oct 09, 2009
```
to llvm-2.6. Remove the precise garbage collection feature.

llvm-svn: 83638
```
  31c4d8b0
- 80-columns! · 7d33d21d
  Nicolas Geoffray authored Oct 09, 2009
```
llvm-svn: 83628
```
  7d33d21d
- Add initial information on VMKit. · 4b7479ab
  Nicolas Geoffray authored Oct 09, 2009
```
llvm-svn: 83627
```
  4b7479ab
- more random updates. · 0dd2acae
  Chris Lattner authored Oct 09, 2009
```
llvm-svn: 83625
```
  0dd2acae
- Give Dan and my recent changes, machine LICM is now code size neutral. · 7ddb844a
  Evan Cheng authored Oct 09, 2009
```
llvm-svn: 83624
```
  7ddb844a
- checkpoint. · 7c442bdd
  Chris Lattner authored Oct 09, 2009
```
llvm-svn: 83623
```
  7c442bdd
- Fix a logic error that caused non-rematable loop invariants loads to be licm'ed out of loop. · ccd4545e
  Evan Cheng authored Oct 09, 2009
```
llvm-svn: 83622
```
  ccd4545e
- checkpoint. · c9bf49df
  Chris Lattner authored Oct 09, 2009
```
llvm-svn: 83621
```
  c9bf49df
- Slight rewording. · 28bca602
  Mikhail Glushenkov authored Oct 09, 2009
```
llvm-svn: 83620
```
  28bca602
- Omit the 'out_file_index != -1' check when possible. · 0bd47063
  Mikhail Glushenkov authored Oct 09, 2009
```
llvm-svn: 83619
```
  0bd47063
- Use llvm-as only for compiling .ll -> .bc. · 8f301485
  Mikhail Glushenkov authored Oct 09, 2009
```
llc can compile .ll files directly these days.

llvm-svn: 83618
```
  8f301485
- Commit one last NEON test to use FileCheck. That's all of them now! · e9b19f76
  Bob Wilson authored Oct 09, 2009
```
llvm-svn: 83617
```
  e9b19f76
- Convert more NEON tests to use FileCheck. · 24b84fec
  Bob Wilson authored Oct 09, 2009
```
llvm-svn: 83616
```
  24b84fec
- update clang section. · 303d2325
  Chris Lattner authored Oct 09, 2009
```
llvm-svn: 83615
```
  303d2325
- Raise the limit on built-in plugins in llvmc to 10. · fc8809d0
  Mikhail Glushenkov authored Oct 09, 2009
```
llvm-svn: 83614
```
  fc8809d0
- Reconfigure automatically when Base.td.in is changed. · 9fdea9e4
  Mikhail Glushenkov authored Oct 09, 2009
```
Thanks to Chris for heads-up!

llvm-svn: 83613
```
  9fdea9e4
- Reset kill markers after live interval is reconstructed. · be27d61c
  Evan Cheng authored Oct 09, 2009
```
llvm-svn: 83608
```
  be27d61c
- Indentation. · f5970a01
  Evan Cheng authored Oct 09, 2009
```
llvm-svn: 83607
```
  f5970a01
- Preserve HasNSW and HasNUW when constructing SCEVs for Add and Mul · d65387a4
  Dan Gohman authored Oct 09, 2009
```
instructions.

llvm-svn: 83606
```
  d65387a4
- When considering whether to inline Callee into Caller, · 3059924b
  Dale Johannesen authored Oct 09, 2009
```
and that will make Caller too big to inline, see if it
might be better to inline Caller into its callers instead.
This situation is described in PR 2973, although I haven't
tried the specific case in SPASS.

llvm-svn: 83602
```
  3059924b
- Add the ability to track HasNSW and HasNUW on more kinds of SCEV expressions. · 816fe0a4
  Dan Gohman authored Oct 09, 2009
```
llvm-svn: 83601
```
  816fe0a4
- Add codegen support for NEON vst4lane intrinsics with 128-bit vectors. · 84e7967f
  Bob Wilson authored Oct 09, 2009
```
llvm-svn: 83600
```
  84e7967f
- Add codegen support for NEON vst3lane intrinsics with 128-bit vectors. · c4090308
  Bob Wilson authored Oct 08, 2009
```
llvm-svn: 83598
```
  c4090308
- Add codegen support for NEON vst2lane intrinsics with 128-bit vectors. · b851eb35
  Bob Wilson authored Oct 08, 2009
```
llvm-svn: 83596
```
  b851eb35
- Convert more NEON tests to use FileCheck. · 1fd98d67
  Bob Wilson authored Oct 08, 2009
```
llvm-svn: 83595
```
  1fd98d67
- Add codegen support for NEON vld4lane intrinsics with 128-bit vectors. · 38ba4722
  Bob Wilson authored Oct 08, 2009
```
Also fix some copy-and-paste errors in previous changes.

llvm-svn: 83590
```
  38ba4722
- Remove code that makes no sense. · e23984fb
  Evan Cheng authored Oct 08, 2009
```
llvm-svn: 83589
```
  e23984fb
- Convert more NEON tests to use FileCheck. · f4482550
  Bob Wilson authored Oct 08, 2009
```
llvm-svn: 83587
```
  f4482550
- Add codegen support for NEON vld3lane intrinsics with 128-bit vectors. · cf54e934
  Bob Wilson authored Oct 08, 2009
```
llvm-svn: 83585
```
  cf54e934
Oct 08, 2009
- Update CMake build yet again after a source file was removed · da3c7969
  Douglas Gregor authored Oct 08, 2009
```
llvm-svn: 83575
```
  da3c7969
- It's possible for a global variable to be optimized out of a metadata object. So · 3dc625cd
  Bill Wendling authored Oct 08, 2009
```
we should allow a "null" with this dyn_cast.

llvm-svn: 83573
```
  3dc625cd
- Use lower16 / upper16 imm modifiers to asmprint 32-bit imms splitted via movt/movw pair. · 222b86cd
  Anton Korobeynikov authored Oct 08, 2009
```
llvm-svn: 83572
```
  222b86cd
- Clear variable debug info map at the end of the function. · e1969dcd
  Devang Patel authored Oct 08, 2009
```
llvm-svn: 83571
```
  e1969dcd
- Add codegen support for NEON vld2lane intrinsics with 128-bit vectors. · c2728f44
  Bob Wilson authored Oct 08, 2009
```
llvm-svn: 83568
```
  c2728f44
- Clean up some unnecessary initializations. · fac94765
  Bob Wilson authored Oct 08, 2009
```
llvm-svn: 83566
```
  fac94765
- Clean up a comment (indentation was wrong). · 4facd965
  Bob Wilson authored Oct 08, 2009
```
llvm-svn: 83565
```
  4facd965
- Add a SelectionDAG getTargetInsertSubreg convenience function, · 2a45a655
  Bob Wilson authored Oct 08, 2009
```
similar to getTargetExtractSubreg.

llvm-svn: 83564
```
  2a45a655