Commits · 5a30d8367afa94689aeb14ce07251ab613f00114 · Roger Ferrer / llvm-epi-0.8

Feb 13, 2006
- Add a method · 5a30d836
  Chris Lattner authored Feb 13, 2006
```
llvm-svn: 26141
```
  5a30d836
- this passes now, due to Nate's recent efforts · d062f090
  Chris Lattner authored Feb 13, 2006
```
llvm-svn: 26140
```
  d062f090
- Reduce this testcase a bit more, with the help of llvm-extract and some hand tweaks · 4c388281
  Chris Lattner authored Feb 13, 2006
```
llvm-svn: 26139
```
  4c388281
- Be careful not to request or look at bits shifted in from outside the size · 68e74757
  Chris Lattner authored Feb 13, 2006
```
of the input.  This fixes the mediabench/gsm/toast failure last night.

llvm-svn: 26138
```
  68e74757
Feb 12, 2006
- Added a test case for a libcall insertion bug. · 0684e9ae
  Evan Cheng authored Feb 12, 2006
```
llvm-svn: 26137
```
  0684e9ae
- Add missing patterns for andi. and andis., fixing test/Regression/CodeGen/ · bc3ec1d3
  Nate Begeman authored Feb 12, 2006
```
PowerPC/and-imm.ll

llvm-svn: 26136
```
  bc3ec1d3
- remove some more dead special case code · f5b4ef7f
  Chris Lattner authored Feb 12, 2006
```
llvm-svn: 26135
```
  f5b4ef7f
- Eliminate special case hacks that are superceded by general purpose hacks · 5b2edb1f
  Chris Lattner authored Feb 12, 2006
```
llvm-svn: 26134
```
  5b2edb1f
- tweaks · e6ebe1af
  Chris Lattner authored Feb 12, 2006
```
llvm-svn: 26133
```
  e6ebe1af
- Three changes: · ee0f2807
  Chris Lattner authored Feb 12, 2006
```
1. Teach GetConstantInType to handle boolean constants.
2. Teach instcombine to fold (compare X, CST) when X has known 0/1 bits.
   Testcase here: set.ll:test22
3. Improve the "(X >> c1) & C2 == 0" folding code to allow a noop cast
   between the shift and and.  More aggressive bitfolding for other reasons
   was turning signed shr's into unsigned shr's, leaving the noop cast in
   the way.

llvm-svn: 26131
```
  ee0f2807
- new testcase · 2234a136
  Chris Lattner authored Feb 12, 2006
```
llvm-svn: 26130
```
  2234a136
- move a failing testcase from bit-tracking.ll to narrow.ll, and move the · 5e488c9c
  Chris Lattner authored Feb 12, 2006
```
xfail marker with it

llvm-svn: 26129
```
  5e488c9c
- Revert my last patch. It too breaks stuff · 02f53ad3
  Chris Lattner authored Feb 12, 2006
```
llvm-svn: 26128
```
  02f53ad3
- Make these tests fail if opt crashes. · 8e99b0ba
  Chris Lattner authored Feb 12, 2006
```
llvm-svn: 26127
```
  8e99b0ba
Feb 11, 2006
- Fix for my previously reverted patch · 35248e06
  Chris Lattner authored Feb 11, 2006
```
llvm-svn: 26126
```
  35248e06
- Update comments to be actually accurate · f3ed0348
  Chris Lattner authored Feb 11, 2006
```
llvm-svn: 26124
```
  f3ed0348
- This is implemented by the simplify-libcalls pass, not instcombine · 00601c0c
  Chris Lattner authored Feb 11, 2006
```
llvm-svn: 26123
```
  00601c0c
- Port the recent innovations in ComputeMaskedBits to SimplifyDemandedBits. · 0157e7f5
  Chris Lattner authored Feb 11, 2006
```
This allows us to simplify on conditions where bits are not known, but they
are not demanded either!  This also fixes a couple of bugs in
ComputeMaskedBits that were exposed during this work.

In the future, swaths of instcombine should be removed, as this code
subsumes a bunch of ad-hockery.

llvm-svn: 26122
```
  0157e7f5
- revert my previous change, it exposed other problems. · b24ce3a2
  Chris Lattner authored Feb 11, 2006
```
llvm-svn: 26121
```
  b24ce3a2
- · 4698e4f5
  Duraid Madina authored Feb 11, 2006
```
fix storing booleans (grawp missed this one)

llvm-svn: 26120
```
  4698e4f5
- · 0010a923
  Duraid Madina authored Feb 11, 2006
```
now short immediates will get matched (previously constants were all
triggering movl 64bit imm fat instructions)

llvm-svn: 26119
```
  0010a923
- Make this check stricter. Disallow loop exit blocks from being shared by · 05bf90dd
  Chris Lattner authored Feb 11, 2006
```
loops and their subloops.

llvm-svn: 26118
```
  05bf90dd
- Prevent certain nodes that have already been selected from being folded into · a86ba85d
  Evan Cheng authored Feb 11, 2006
```
X86 addressing mode. Currently we do not allow any node whose target node
produces a chain as well as any node that is at the root of the addressing
mode expression tree.

llvm-svn: 26117
```
  a86ba85d
- remove dead expr · a6ae101a
  Chris Lattner authored Feb 11, 2006
```
llvm-svn: 26116
```
  a6ae101a
- Reorg for integration with gcc4. Old style debug info will not be passed though · 5995d016
  Jim Laskey authored Feb 11, 2006
```
to SelIDAG.

llvm-svn: 26115
```
  5995d016
- implement unswitching of loops with switch stmts and selects in them · fbadd7e1
  Chris Lattner authored Feb 11, 2006
```
llvm-svn: 26114
```
  fbadd7e1
- Update PHI nodes in successors of exit blocks. · f1b15168
  Chris Lattner authored Feb 10, 2006
```
llvm-svn: 26113
```
  f1b15168
- Reform the unswitching code in terms of edge splitting, not block splitting. · fe4151ef
  Chris Lattner authored Feb 10, 2006
```
llvm-svn: 26112
```
  fe4151ef
Feb 10, 2006
- Nicer code. :-) · 2b6f78b6
  Evan Cheng authored Feb 10, 2006
```
llvm-svn: 26111
```
  2b6f78b6
- Added X86 isel debugging stuff. · d49cc363
  Evan Cheng authored Feb 10, 2006
```
llvm-svn: 26110
```
  d49cc363
- Remove a level of indirection. · 975486db
  Chris Lattner authored Feb 10, 2006
```
llvm-svn: 26109
```
  975486db
- Fix a case where UnswitchTrivialCondition broke critical edges with · ec6b40a0
  Chris Lattner authored Feb 10, 2006
```
phi's in the successors

llvm-svn: 26108
```
  ec6b40a0
- Use the auto-generated call matcher. Remove a broken impl of the frameaddr/returnaddr · fcb8a3aa
  Chris Lattner authored Feb 10, 2006
```
intrinsics.

Autogen frameindex matcher

llvm-svn: 26107
```
  fcb8a3aa
- Update to new-style flags usage, simplifying the .td file · 0c4dea4c
  Chris Lattner authored Feb 10, 2006
```
llvm-svn: 26106
```
  0c4dea4c
- Remove a completed entry; add a new entry about fisttp op · 907be3e2
  Evan Cheng authored Feb 10, 2006
```
llvm-svn: 26105
```
  907be3e2
- add some notes, move some code around. Implement unswitching of loops · 6e263155
  Chris Lattner authored Feb 10, 2006
```
with branches on partially invariant computations.

llvm-svn: 26104
```
  6e263155
- Move code around to be more logical, no functionality change. · 4935417a
  Chris Lattner authored Feb 10, 2006
```
llvm-svn: 26103
```
  4935417a
- When unswitching a trivial loop, do admit we are doing it! :) · 3fc3148b
  Chris Lattner authored Feb 10, 2006
```
llvm-svn: 26102
```
  3fc3148b
- Implement unconditional unswitching of 'trivial' loops, those loops that contain · ed7a67b0
  Chris Lattner authored Feb 10, 2006
```
branches in their entry block that control whether or not the loop is a noop or not.

llvm-svn: 26101
```
  ed7a67b0
Feb 09, 2006
- Simplify control flow a bit, note that unswitch preserves canonical loop form · 4f0e66df
  Chris Lattner authored Feb 09, 2006
```
llvm-svn: 26098
```
  4f0e66df