- May 18, 2009
-
-
Dale Johannesen authored
Some comment cleanup. llvm-svn: 72022
-
- May 15, 2009
-
-
Mike Stump authored
llvm-svn: 71813
-
Mike Stump authored
llvm-svn: 71812
-
- May 13, 2009
-
-
Evan Cheng authored
llvm-svn: 71726
-
Bill Wendling authored
booleans. This gives a better indication of what the "addReg()" is doing. Remembering what all of those booleans mean isn't easy, especially if you aren't spending all of your time in that code. I took Jakob's suggestion and made it illegal to pass in "true" for the flag. This should hopefully prevent any unintended misuse of this (by reverting to the old way of using addReg()). llvm-svn: 71722
-
- May 12, 2009
-
-
Bill Wendling authored
llvm-svn: 71601
-
Evan Cheng authored
llvm-svn: 71582
-
Evan Cheng authored
llvm-svn: 71520
-
- May 11, 2009
-
-
Dan Gohman authored
address folding. llvm-svn: 71446
-
- May 09, 2009
-
-
Duncan Sands authored
will make it more obvious what it represents, and stop it being confused with the StoreSize. llvm-svn: 71349
-
- May 08, 2009
-
-
Anton Korobeynikov authored
llvm-svn: 71241
-
Chris Lattner authored
need to work a bit to combine things like (x+c1+c2) into x+c3. llvm-svn: 71232
-
Evan Cheng authored
Optimize code placement in loop to eliminate unconditional branches or move unconditional branch to the outside of the loop. e.g. /// A: /// ... /// <fallthrough to B> /// /// B: --> loop header /// ... /// jcc <cond> C, [exit] /// /// C: /// ... /// jmp B /// /// ==> /// /// A: /// ... /// jmp B /// /// C: --> new loop header /// ... /// <fallthough to B> /// /// B: /// ... /// jcc <cond> C, [exit] llvm-svn: 71209
-
- May 06, 2009
-
-
Dale Johannesen authored
more place. This fixes a bunch of x86-64 JIT regressions. (Introduced when the value of the magic constant changed in 68645. At the time apparently nobody noticed; failures were hidden in 70343-70439 by an unrelated bug, so showed up again as "new" failures in 70440.) llvm-svn: 71106
-
- May 05, 2009
-
-
Chris Lattner authored
addrspace(257) -> FS relative on x86. Patch by Zoltan Varga! llvm-svn: 70992
-
Evan Cheng authored
Revert part of 70929 that has to do with determining whether a SIB byte is needed. It causes a lot of x86_64 JIT failures. llvm-svn: 70986
-
Evan Cheng authored
- Synchronize instruction length computation code in X86InstrInfo with code in X86CodeEmitter.cpp Patch by Zoltan Varga. llvm-svn: 70929
-
- May 04, 2009
-
-
Dan Gohman authored
llvm-svn: 70902
-
- May 01, 2009
-
-
Argyrios Kyrtzidis authored
llvm-svn: 70522
-
Argyrios Kyrtzidis authored
-Replace DebugLocTuple's Source ID with CompileUnit's GlobalVariable* -Remove DwarfWriter::getOrCreateSourceID -Make necessary changes for the above (fix callsites, etc.) llvm-svn: 70520
-
- Apr 30, 2009
-
-
Dan Gohman authored
llvm-svn: 70466
-
Evan Cheng authored
llvm-svn: 70461
-
Bill Wendling authored
which better identifies what the optimization is doing. And is more flexible for future uses. llvm-svn: 70440
-
Nate Begeman authored
llvm-svn: 70425
-
- Apr 29, 2009
-
-
Nate Begeman authored
llvm-svn: 70372
-
Bill Wendling authored
Massive check in. This changes the "-fast" flag to "-O#" in llc. If you want to use the old behavior, the flag is -O0. This change allows for finer-grained control over which optimizations are run at different -O levels. Most of this work was pretty mechanical. The majority of the fixes came from verifying that a "fast" variable wasn't used anymore. The JIT still uses a "Fast" flag. I'll change the JIT with a follow-up patch. llvm-svn: 70343
-
- Apr 28, 2009
-
-
Anton Korobeynikov authored
This should fix PR3379 and PR4064. Patch inspired by Edwin Török! llvm-svn: 70328
-
Bill Wendling authored
llvm-svn: 70275
-
Bill Wendling authored
use the old behavior, the flag is -O0. This change allows for finer-grained control over which optimizations are run at different -O levels. Most of this work was pretty mechanical. The majority of the fixes came from verifying that a "fast" variable wasn't used anymore. The JIT still uses a "Fast" flag. I'm not 100% sure if it's necessary to change it there... llvm-svn: 70270
-
- Apr 27, 2009
-
-
Nate Begeman authored
PR2957 ISD::VECTOR_SHUFFLE now stores an array of integers representing the shuffle mask internal to the node, rather than taking a BUILD_VECTOR of ConstantSDNodes as the shuffle mask. A value of -1 represents UNDEF. In addition to eliminating the creation of illegal BUILD_VECTORS just to represent shuffle masks, we are better about canonicalizing the shuffle mask, resulting in substantially better code for some classes of shuffles. llvm-svn: 70225
-
Dan Gohman authored
to precisely describe the h-register subreg register classes. Thanks to Jakob Stoklund Olesen for spotting this and for the initial patch! Also, make getStoreRegOpcode and getLoadRegOpcode aware of the needs of h registers. llvm-svn: 70211
-
Dan Gohman authored
GR32_ABCD, and GR64_ABCD, respectively, to help describe them. llvm-svn: 70210
-
Dan Gohman authored
llvm-svn: 70209
-
Mon P Wang authored
llvm-svn: 70197
-
- Apr 25, 2009
-
-
Chris Lattner authored
Nicolas Capens! llvm-svn: 70057
-
- Apr 24, 2009
-
-
Rafael Espindola authored
very elegant, but neither is the tls specification :-( llvm-svn: 69968
-
Rafael Espindola authored
llvm-svn: 69967
-
Nate Begeman authored
ISD::VECTOR_SHUFFLE now stores an array of integers representing the shuffle mask internal to the node, rather than taking a BUILD_VECTOR of ConstantSDNodes as the shuffle mask. A value of -1 represents UNDEF. In addition to eliminating the creation of illegal BUILD_VECTORS just to represent shuffle masks, we are better about canonicalizing the shuffle mask, resulting in substantially better code for some classes of shuffles. A clean up of x86 shuffle code, and some canonicalizing in DAGCombiner is next. llvm-svn: 69952
-
- Apr 23, 2009
-
-
Dan Gohman authored
memory operand tuples. This doesn't ever come up in normal code however. llvm-svn: 69848
-
- Apr 21, 2009
-
-
Duncan Sands authored
Spotted by gcc-4.5. llvm-svn: 69673
-
