Commits · e3abb14503d2ea1375002adafc75e615d691384d · Lorenzo Albano / LLVM bpEVL

Jun 14, 2006
- Use the PotDoms map to memoize 'dominating value' lookup. With this patch, · e3abb145
  Chris Lattner authored Jun 14, 2006
```
LCSSA is still the slowest pass when gccas'ing 252.eon, but now it only takes
39s instead of 289s. :)

llvm-svn: 28776
```
  e3abb145
Jun 13, 2006
- getOperandNum(): error if specified operand number is out of range. · 632ee8de
  Evan Cheng authored Jun 13, 2006
```
llvm-svn: 28775
```
  632ee8de
- Fix another instance where PHI nodes need special treatment. · e714a5c5
  Owen Anderson authored Jun 13, 2006
```
llvm-svn: 28774
```
  e714a5c5
- Update isLCSSAForm to handle PHI nodes specially for live-out detection. This · 7e428625
  Owen Anderson authored Jun 13, 2006
```
is the same as the recent patch to LCSSA.cpp.

llvm-svn: 28773
```
  7e428625
- I am sure I had commited this workaround before. Perhaps soon I should sort it all out · 7c69df96
  Andrew Lenharth authored Jun 13, 2006
```
llvm-svn: 28772
```
  7c69df96
- Fix a bug that was causing major slowdowns in povray. This was due to LCSSA · 3f8ff044
  Owen Anderson authored Jun 13, 2006
```
not handling PHI nodes correctly when determining if a value was live-out.

This patch reduces the number of detected live-out variables in the testcase
from 6565 to 485.

llvm-svn: 28771
```
  3f8ff044
- It really helps to be returning to the correct place · f570feea
  Andrew Lenharth authored Jun 13, 2006
```
llvm-svn: 28769
```
  f570feea
- Port some bugfixes in shift handling from SimplifyDemandedBits over to · c5bb8ab1
  Chris Lattner authored Jun 13, 2006
```
ComputeMaskedBits.  DemandedMasks and KnownZero/One masks should never have
bits set out of the range of the base datatype.

llvm-svn: 28768
```
  c5bb8ab1
- My original test case was bogus. Reverting to crasher case. · a48a1763
  Jim Laskey authored Jun 13, 2006
```
llvm-svn: 28767
```
  a48a1763
- Tests to make sure that ComputeMaskedBits discards bits for shifts. · fe25b30c
  Jim Laskey authored Jun 13, 2006
```
llvm-svn: 28766
```
  fe25b30c
- TargetLowering::ComputeMaskedBits was not clearing reciprocal bits on shifts. · 8cac9cd5
  Jim Laskey authored Jun 13, 2006
```
llvm-svn: 28765
```
  8cac9cd5
- Cygwin support: use _alloca to allocate stack if > 4k. Patch by Anton Korobeynikov. · 17ca732b
  Evan Cheng authored Jun 13, 2006
```
llvm-svn: 28764
```
  17ca732b
- Teach bugpoint to kill optimization passes that run over the timeout limit, · 20ec1654
  Chris Lattner authored Jun 13, 2006
```
which allows it to debug optimizer infinite loops.  This patch is contributed
by Nick Lewycky, thanks!

llvm-svn: 28763
```
  20ec1654
- Decribe the "implementation" keyword. · d2d29a08
  Chris Lattner authored Jun 13, 2006
```
llvm-svn: 28762
```
  d2d29a08
- Gaar! Don't use r11 for CR save/restore, use R0. R11 can be register · ac59ab51
  Chris Lattner authored Jun 12, 2006
```
allocated, thus live across the save/reload.  This fixes

llc-beta /MultiSource/Applications/spiff/spiff
llc-beta /MultiSource/Benchmarks/sim/sim:
llc-beta /MultiSource/Benchmarks/Ptrdist/bc/bc
llc-beta /MultiSource/Benchmarks/McCat/12-IOtest/iotest:
llc-beta /MultiSource/Benchmarks/FreeBench/fourinarow/fourinarow
llc-beta /MultiSource/Benchmarks/Fhourstones-3.1/fhourstones3.1
llc-beta /MultiSource/Benchmarks/mediabench/adpcm/rawdaudio/rawdaudio
llc-beta /MultiSource/Benchmarks/mediabench/adpcm/rawcaudio/rawcaudio
llc-beta /MultiSource/Benchmarks/mediabench/g721/g721encode/encode
llc-beta /MultiSource/Benchmarks/mediabench/jpeg/jpeg-6a/cjpeg

and probably others, with -regalloc=local.

llvm-svn: 28761
```
  ac59ab51
Jun 12, 2006
- Fix spilling and reloading of CR regs to reload the right values. This fixes · 6b043a24
  Chris Lattner authored Jun 12, 2006
```
Olden/power (and probably others) with -regalloc=local.

llvm-svn: 28760
```
  6b043a24
- Reapply my 6/9 changes. The bug Evan saw no longer occurs. · fd0a3d6e
  Owen Anderson authored Jun 12, 2006
```
llvm-svn: 28759
```
  fd0a3d6e
- Fix an infinite loop on Transforms/SimplifyCFG/2006-06-12-InfLoop.ll · b5c9d7a0
  Chris Lattner authored Jun 12, 2006
```
llvm-svn: 28758
```
  b5c9d7a0
- TEstcase that makes simplifycfg infloop. · 9c4ddf83
  Chris Lattner authored Jun 12, 2006
```
llvm-svn: 28757
```
  9c4ddf83
- Add a new -compileflags options, remove old externals location specifiers. · b3c62e43
  Chris Lattner authored Jun 12, 2006
```
People should just use -with-externals.

llvm-svn: 28756
```
  b3c62e43
- Make sure to update the CFG correctly if a switch only has a default dest. · 32d92e00
  Chris Lattner authored Jun 12, 2006
```
This fixes CodeGen/Generic/2006-06-12-LowerSwitchCrash.ll

llvm-svn: 28755
```
  32d92e00
- New testcase that crashes llc · eecff6c2
  Chris Lattner authored Jun 12, 2006
```
llvm-svn: 28754
```
  eecff6c2
- Let the alpha breakage begin. First Formals and RET. next Calls · 80528499
  Andrew Lenharth authored Jun 12, 2006
```
llvm-svn: 28753
```
  80528499
- Start on my todo list · 0e57b2cb
  Andrew Lenharth authored Jun 12, 2006
```
llvm-svn: 28752
```
  0e57b2cb
- lower more then 4 formal arguments. The offset is currently hard coded. · 4e76015e
  Rafael Espindola authored Jun 12, 2006
```
implement SelectFrameIndex

llvm-svn: 28751
```
  4e76015e
- Fix for 2006-06-26-MultipleExitsSingleBlock. · 0ac33696
  Owen Anderson authored Jun 12, 2006
```
If a single exit block has multiple predecessors within the loop, it will
appear in the exit blocks list more than once.  LCSSA needs to take that into
account so that it doesn't double process that exit block.

llvm-svn: 28750
```
  0ac33696
- Make sure that LCSSA works properly when it encounters an exit block that has · ecb8036e
  Owen Anderson authored Jun 12, 2006
```
more than one predecessor inside the loop.

llvm-svn: 28749
```
  ecb8036e
Jun 11, 2006
- Re-commit the safe parts of my 6/9 patch. Still working on fixing the unsafe parts. · b538f14d
  Owen Anderson authored Jun 11, 2006
```
llvm-svn: 28748
```
  b538f14d
- Back out Owen's 6/9 changes. They broke... · 1b6e310e
  Evan Cheng authored Jun 11, 2006
```
Back out Owen's 6/9 changes. They broke MultiSource/Benchmarks/Prolangs-C/bison (and perhaps others).

llvm-svn: 28747
```
  1b6e310e
Jun 10, 2006
- Add a missing assertion that would have helped out Reid · bfd0b6d8
  Chris Lattner authored Jun 10, 2006
```
llvm-svn: 28746
```
  bfd0b6d8
- Work around a nasty tblgen bug where it doesn't add operands for varargs · b055c873
  Chris Lattner authored Jun 10, 2006
```
nodes correctly.

llvm-svn: 28745
```
  b055c873
- Fix a problem exposed by the local allocator. CALL instructions are not marked · 006b2c6a
  Chris Lattner authored Jun 10, 2006
```
as using incoming argument registers, so the local allocator would clobber them
between their set and use.  To fix this, we give the call instructions a variable
number of uses in the CALL MachineInstr itself, so live variables understands
the live ranges of these register arguments.

llvm-svn: 28744
```
  006b2c6a
- Wrap to 80 cols · deafba03
  Chris Lattner authored Jun 09, 2006
```
llvm-svn: 28743
```
  deafba03
Jun 09, 2006
- After telling GCC to type of the input file with -x asm/-x c, switch back to · 572e78c2
  Chris Lattner authored Jun 09, 2006
```
-x none, to not foul up autodetection of .a file or .dylibs.

llvm-svn: 28741
```
  572e78c2
- Don't pull in environ, not always safe. Global variables are bad anyway. · 09cf8299
  Evan Cheng authored Jun 09, 2006
```
Use execve when explicit environment variables ptr is available. Otherwise
just use execv.

llvm-svn: 28740
```
  09cf8299
- Add LCSSA as a requirement for LoopUnswitch, and assert that LoopUnswitch preserves · b1dc1d44
  Owen Anderson authored Jun 09, 2006
```
LCSSA.

llvm-svn: 28739
```
  b1dc1d44
- Make Loop able to verify that it is in LCSSA-form, and have the LCSSA pass assert · 505adff3
  Owen Anderson authored Jun 09, 2006
```
on this.

llvm-svn: 28738
```
  505adff3
- Comments to appease sabre. · beedf824
  Evan Cheng authored Jun 09, 2006
```
llvm-svn: 28737
```
  beedf824
- Minor compilation speed improvement. · 0e14a56d
  Evan Cheng authored Jun 09, 2006
```
llvm-svn: 28736
```
  0e14a56d
- RewriteExpr, either the new PHI node of induction variable or the · 398f7029
  Evan Cheng authored Jun 09, 2006
```
post-increment value, should be first cast to the appropriated type (to the
type of the common expr). Otherwise, the rewrite of a use based on (common +
iv) may end up with an incorrect type.

llvm-svn: 28735
```
  398f7029