Commits · cde9e30bc6acdb22b21e1553f3460a9ab5363e62 · Roger Ferrer / llvm-epi-0.8

Jan 27, 2006
- x86 CPU detection and proper subtarget support · cde9e30b
  Evan Cheng authored Jan 27, 2006
```
llvm-svn: 25679
```
  cde9e30b
- PHI and INLINEASM are now built-in instructions provided by Target.td · 12405746
  Chris Lattner authored Jan 27, 2006
```
llvm-svn: 25674
```
  12405746
Jan 26, 2006
- Improve compatibility with VC2005, patch by Morten Ofstad! · 15a8c15a
  Jeff Cohen authored Jan 26, 2006
```
llvm-svn: 25661
```
  15a8c15a
- Improve compatibility with VC2005, patch by Morten Ofstad! · ebbfb386
  Chris Lattner authored Jan 26, 2006
```
llvm-svn: 25653
```
  ebbfb386
- Added preliminary x86 subtarget support. · 54c13da2
  Evan Cheng authored Jan 26, 2006
```
llvm-svn: 25645
```
  54c13da2
- Work around some x86 Darwin assembler bugs · fcdce6d2
  Evan Cheng authored Jan 26, 2006
```
llvm-svn: 25638
```
  fcdce6d2
- When trying to fold X86::SETCC into a Select, make a copy if it has more than · 944d1e91
  Evan Cheng authored Jan 26, 2006
```
one use. This allows more CMOV instructions.

llvm-svn: 25634
```
  944d1e91
- Remove the uses of STATUS flag register. Rely on node property SDNPInFlag, · 97c68f0f
  Evan Cheng authored Jan 26, 2006
```
SDNPOutFlag, and SDNPOptInFlag instead.

llvm-svn: 25629
```
  97c68f0f
Jan 25, 2006
- First part of bug 680: · e74795cd
  Nate Begeman authored Jan 25, 2006
```
Remove TLI.LowerVA* and replace it with SDNodes that are lowered the same
way as everything else.

llvm-svn: 25606
```
  e74795cd
- X86 prefer scheduling for reduced register pressure. · 83eeefbb
  Evan Cheng authored Jan 25, 2006
```
llvm-svn: 25602
```
  83eeefbb
- Fix a selectcc lowering bug. Make a copy of X86ISD::CMP when folding it. · aff0800f
  Evan Cheng authored Jan 25, 2006
```
llvm-svn: 25596
```
  aff0800f
- Loosen up these checks to allow direct uses of ESP · bc7226a7
  Chris Lattner authored Jan 25, 2006
```
llvm-svn: 25595
```
  bc7226a7
Jan 24, 2006
- use ESP directly, not a copy of ESP into some other register for fastcc calls · 27d30a5f
  Chris Lattner authored Jan 24, 2006
```
llvm-svn: 25584
```
  27d30a5f
- Emit the copies out of call return registers *after* the ISD::CALLSEQ_END · 6f33eaeb
  Chris Lattner authored Jan 24, 2006
```
node, fixing fastcc and the case where a function has a frame pointer due
to dynamic allocas.

llvm-svn: 25580
```
  6f33eaeb
- Allow jit-beta to work · 68e62a51
  Chris Lattner authored Jan 24, 2006
```
llvm-svn: 25578
```
  68e62a51
Jan 23, 2006
- Add explicit #includes of <iostream> · de02d772
  Chris Lattner authored Jan 22, 2006
```
llvm-svn: 25515
```
  de02d772
Jan 21, 2006
- Rename fcmovae to fcmovnb and fcmova to fcmovnbe (following Intel manual). · 468fecdc
  Evan Cheng authored Jan 21, 2006
```
Some assemblers can't recognize the aliases.

llvm-svn: 25494
```
  468fecdc
Jan 20, 2006
- LowerReturn now doesn't have to handle f32 returns. · 335b46dd
  Chris Lattner authored Jan 20, 2006
```
llvm-svn: 25484
```
  335b46dd
- Stop doing that accidental commit. · 0c5de286
  Evan Cheng authored Jan 20, 2006
```
llvm-svn: 25474
```
  0c5de286
- A few more SH{L|R}D peepholes. · cce748d3
  Evan Cheng authored Jan 20, 2006
```
llvm-svn: 25473
```
  cce748d3
- Didn't mean to commit the last one. · 9c30bd5e
  Evan Cheng authored Jan 19, 2006
```
llvm-svn: 25469
```
  9c30bd5e
- Added i16 SH{L|R}D patterns. · 8591b9f2
  Evan Cheng authored Jan 19, 2006
```
llvm-svn: 25468
```
  8591b9f2
Jan 19, 2006
- Avoid generating a redundant setcc. · 3d2cc7e2
  Evan Cheng authored Jan 19, 2006
```
llvm-svn: 25457
```
  3d2cc7e2
- adc and sbb need an incoming flag to ensure it reads the carry flag · 91007126
  Evan Cheng authored Jan 19, 2006
```
from add / sub.

llvm-svn: 25444
```
  91007126
- Two peepholes: · a7bfbe99
  Evan Cheng authored Jan 19, 2006
```
(or (x >> c) | (y << (32 - c))) ==> (shrd x, y, c)
(or (x << c) | (y >> (32 - c))) ==> (shld x, y, c)

llvm-svn: 25438
```
  a7bfbe99
- Didn't mean to check that in. · 6135a7a5
  Evan Cheng authored Jan 19, 2006
```
llvm-svn: 25436
```
  6135a7a5
- A obvious typo · 267ba596
  Evan Cheng authored Jan 19, 2006
```
llvm-svn: 25435
```
  267ba596
Jan 18, 2006
- SRA shift amount must be in i8 · 621674a1
  Evan Cheng authored Jan 18, 2006
```
llvm-svn: 25416
```
  621674a1
- If a call return type is i1, insert a truncate from X86::AL to i1. · 4b3774e0
  Evan Cheng authored Jan 18, 2006
```
llvm-svn: 25415
```
  4b3774e0
Jan 17, 2006
- Fix lowering of calls which return f32 values. · feaed4d1
  Evan Cheng authored Jan 17, 2006
```
llvm-svn: 25413
```
  feaed4d1
- Zero extending load from i1 to i8. · 14417ed9
  Evan Cheng authored Jan 17, 2006
```
llvm-svn: 25391
```
  14417ed9
- SSE does not support i64 SINT_TO_FP (FP stack doesn't either, but we custom · 0d5b69f7
  Evan Cheng authored Jan 17, 2006
```
expand it), so ask legalizer to expand i32 UINT_TO_FP.

llvm-svn: 25386
```
  0d5b69f7
- Added a FIXME comment about why FST is currently flagged to fpGETRESULT. · 561881f3
  Evan Cheng authored Jan 17, 2006
```
llvm-svn: 25381
```
  561881f3
- Bug fixes: fpGETRESULT should produces a flag result and X86ISD::FST should · bec9d720
  Evan Cheng authored Jan 17, 2006
```
read a flag.

llvm-svn: 25378
```
  bec9d720
- More typo's · c14bb102
  Evan Cheng authored Jan 16, 2006
```
llvm-svn: 25375
```
  c14bb102
Jan 16, 2006
- Some typo's · 64eeed27
  Evan Cheng authored Jan 16, 2006
```
llvm-svn: 25374
```
  64eeed27
- Fix FP_TO_INT**_IN_MEM lowering. · 911c68d7
  Evan Cheng authored Jan 16, 2006
```
llvm-svn: 25368
```
  911c68d7
- transfer some notes from my email to somewhere useful. · b2eacf48
  Chris Lattner authored Jan 16, 2006
```
llvm-svn: 25361
```
  b2eacf48
Jan 15, 2006
- Added patterns for 8-bit multiply · 2494ce49
  Evan Cheng authored Jan 15, 2006
```
llvm-svn: 25338
```
  2494ce49
- Use the default lowering of ISD::DYNAMIC_STACKALLOC, delete now dead code. · 78c358d1
  Chris Lattner authored Jan 15, 2006
```
llvm-svn: 25333
```
  78c358d1