Commits · 4b40a42653f7e292c2257c9cdde7228bf6fb5ae8 · Roger Ferrer / llvm-epi-0.8

Feb 14, 2006
- Rename maxStoresPerMemSet to maxStoresPerMemset, etc. · 4b40a426
  Evan Cheng authored Feb 14, 2006
```
llvm-svn: 26174
```
  4b40a426
- Set maxStoresPerMemSet to 16. Ditto for maxStoresPerMemCpy and · 6a37456d
  Evan Cheng authored Feb 14, 2006
```
maxStoresPerMemMove. Although the last one is not used.

llvm-svn: 26172
```
  6a37456d
Feb 13, 2006
- Switch targets over to using SelectionDAG::getCALLSEQ_START to create · 62c3484e
  Chris Lattner authored Feb 13, 2006
```
CALLSEQ_START nodes.

llvm-svn: 26143
```
  62c3484e
Feb 07, 2006
- Darwin ABI issues: weak, linkonce, etc. dynamic-no-pic support is complete. · 5a76680d
  Evan Cheng authored Feb 07, 2006
```
Also fixed a function stub bug. Added weak and linkonce support for
x86 Linux.

llvm-svn: 26038
```
  5a76680d
Feb 04, 2006
- Separate FILD and FILD_FLAG, the later is only used for SSE2. It produces a · 11613a52
  Evan Cheng authored Feb 04, 2006
```
flag so it can be flagged to a FST.

llvm-svn: 25953
```
  11613a52
Feb 02, 2006
- Fix a erroneous comment. · d8fba3a1
  Evan Cheng authored Feb 02, 2006
```
llvm-svn: 25894
```
  d8fba3a1
Feb 01, 2006
- Fix some of the stuff in the PPC README file, and clean up legalization · 7e7f439f
  Nate Begeman authored Feb 01, 2006
```
of the SELECT_CC, BR_CC, and BRTWOWAY_CC nodes.

llvm-svn: 25875
```
  7e7f439f
- Return's chain should be matching either the chain produced by the · a24617f5
  Evan Cheng authored Feb 01, 2006
```
value or the chain going into the load.

llvm-svn: 25863
```
  a24617f5
- When folding a load into a return of SSE value, check the chain to · e1ce4d71
  Evan Cheng authored Feb 01, 2006
```
ensure the memory location has not been clobbered.

llvm-svn: 25861
```
  e1ce4d71
- Be smarter about whether to store the SSE return value in memory. If · 5659ca8f
  Evan Cheng authored Jan 31, 2006
```
it is already available in memory, do a fld directly from there.

llvm-svn: 25859
```
  5659ca8f
Jan 31, 2006
- - Allow XMM load (for scalar use) to be folded into ANDP* and XORP*. · 72d5c256
  Evan Cheng authored Jan 31, 2006
```
- Use XORP* to implement fneg.

llvm-svn: 25857
```
  72d5c256
- * Fix 80-column violations · c642aa5e
  Chris Lattner authored Jan 31, 2006
```
* Rename hasSSE -> hasSSE1 to avoid my continual confusion with 'has any SSE'.
* Add inline asm constraint specification.

llvm-svn: 25854
```
  c642aa5e
- Added custom lowering of fabs · 2dd217b8
  Evan Cheng authored Jan 31, 2006
```
llvm-svn: 25831
```
  2dd217b8
- Don't generate complex sequence for SETOLE, SETOLT, SETULT, and SETUGT. Flip · 45df7f84
  Evan Cheng authored Jan 30, 2006
```
the order of the compare operands and generate SETOGT, SETOGE, SETUGE, and
SETULE instead.

llvm-svn: 25824
```
  45df7f84
Jan 30, 2006
- i64 -> f32, f32 -> i64 and some clean up. · 08390f6a
  Evan Cheng authored Jan 30, 2006
```
llvm-svn: 25818
```
  08390f6a
- Always use FP stack instructions to perform i64 to f64 as well as f64 to i64 · 5b97fcf0
  Evan Cheng authored Jan 30, 2006
```
conversions. SSE does not have instructions to handle these tasks.

llvm-svn: 25817
```
  5b97fcf0
- Move MaskedValueIsZero from the DAGCombiner to the TargetLowering... · f0b24d2d
  Chris Lattner authored Jan 30, 2006
```
Move MaskedValueIsZero from the DAGCombiner to the TargetLowering interface,making isMaskedValueZeroForTargetNode simpler, and useable from other partsof the compiler.

llvm-svn: 25803
```
  f0b24d2d
- adjust prototype · c6fa0282
  Chris Lattner authored Jan 30, 2006
```
llvm-svn: 25798
```
  c6fa0282
Jan 29, 2006
- The FP stack doesn't support UNDEF, ask the legalizer to legalize it · 132177e1
  Chris Lattner authored Jan 29, 2006
```
instead of lying and saying we have it.

llvm-svn: 25775
```
  132177e1
- Targets all now request ConstantFP to be legalized into TargetConstantFP. · 61c9a8e9
  Chris Lattner authored Jan 29, 2006
```
'fpimm' in .td files is now TargetConstantFP.

llvm-svn: 25771
```
  61c9a8e9
Jan 28, 2006
- silence a warning · ccd2a20c
  Chris Lattner authored Jan 28, 2006
```
llvm-svn: 25745
```
  ccd2a20c
Jan 27, 2006
- Bye bye Pattern ISel, hello DAG ISel. · a814f0b3
  Evan Cheng authored Jan 27, 2006
```
llvm-svn: 25700
```
  a814f0b3
- Remove TLI.LowerReturnTo, and just let targets custom lower ISD::RET for · 8c47c3a3
  Nate Begeman authored Jan 27, 2006
```
the same functionality.  This addresses another piece of bug 680.  Next,
on to fixing Alpha VAARG, which I broke last time.

llvm-svn: 25696
```
  8c47c3a3
- x86 CPU detection and proper subtarget support · cde9e30b
  Evan Cheng authored Jan 27, 2006
```
llvm-svn: 25679
```
  cde9e30b
Jan 26, 2006
- 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
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
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
Jan 20, 2006
- LowerReturn now doesn't have to handle f32 returns. · 335b46dd
  Chris Lattner authored Jan 20, 2006
```
llvm-svn: 25484
```
  335b46dd
Jan 19, 2006
- Avoid generating a redundant setcc. · 3d2cc7e2
  Evan Cheng authored Jan 19, 2006
```
llvm-svn: 25457
```
  3d2cc7e2
- 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
- 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
Jan 16, 2006
- Fix FP_TO_INT**_IN_MEM lowering. · 911c68d7
  Evan Cheng authored Jan 16, 2006
```
llvm-svn: 25368
```
  911c68d7