- Sep 03, 2011
-
-
Bruno Cardoso Lopes authored
llvm-svn: 139069
-
Bruno Cardoso Lopes authored
duplicate it for AVX mode. llvm-svn: 139068
-
Bruno Cardoso Lopes authored
llvm-svn: 139067
-
Bruno Cardoso Lopes authored
that the AVX versions (even the 128-bit ones) all clear the upper part of the destination register. llvm-svn: 139066
-
Bruno Cardoso Lopes authored
pattern should be matched llvm-svn: 139065
-
Bruno Cardoso Lopes authored
llvm-svn: 139064
-
Bruno Cardoso Lopes authored
it! llvm-svn: 139063
-
Bruno Cardoso Lopes authored
llvm-svn: 139062
-
Bruno Cardoso Lopes authored
llvm-svn: 139061
-
Bruno Cardoso Lopes authored
OptForSize pattern llvm-svn: 139060
-
Owen Anderson authored
If we have a chain of zext -> assert_zext -> zext -> use, the first zext would get simplified away because of the later zext, and then the later zext would get simplified away because of the assert. The solution is to teach SimplifyDemandedBits that assert_zext demands all of the high bits of its input, rather than only those demanded by its users. No testcase because the only example I have manifests as llvm-gcc miscompiling LLVM, and I haven't found a smaller case that reproduces this problem. Fixes <rdar://problem/10063365>. llvm-svn: 139059
-
Dan Gohman authored
llvm-svn: 139058
-
Jakob Stoklund Olesen authored
The explanation about a 0 argument being materialized as xor is no longer valid. Rematerialization will check if EFLAGS is live before clobbering it. The code produced by X86TargetLowering::EmitLoweredSelect does not clobber EFLAGS. This causes one less testb instruction to be generated in the cmov.ll test case. llvm-svn: 139057
-
Jakob Stoklund Olesen authored
It is only allowed to clobber EFLAGS at the end of a block if it isn't live-in to any successor. llvm-svn: 139056
-
Jakob Stoklund Olesen authored
llvm-svn: 139055
-
Jim Grosbach authored
llvm-svn: 139054
-
Jim Grosbach authored
llvm-svn: 139053
-
Jim Grosbach authored
llvm-svn: 139052
-
Jim Grosbach authored
llvm-svn: 139051
-
Jim Grosbach authored
llvm-svn: 139050
-
Jim Grosbach authored
Tweak handling of IT blocks a bit to enable this. The differentiation between B and Bcc needs special sauce. llvm-svn: 139049
-
Andrew Trick authored
llvm-svn: 139048
-
Jakob Stoklund Olesen authored
llvm-svn: 139047
-
Bill Wendling authored
llvm-svn: 139046
-
Eli Friedman authored
llvm-svn: 139044
-
- Sep 02, 2011
-
-
Fariborz Jahanian authored
llvm-svn: 139043
-
Richard Trieu authored
llvm-svn: 139042
-
Fariborz Jahanian authored
// rdar://10033896 llvm-svn: 139041
-
Jim Grosbach authored
For other shift and rotate instructions, too. Tests for those forthcoming as I work my way through the ISA. llvm-svn: 139040
-
Bill Wendling authored
llvm-svn: 139039
-
Bill Wendling authored
llvm-svn: 139038
-
Andrew Trick authored
llvm-svn: 139037
-
Andrew Trick authored
llvm-svn: 139036
-
Bill Wendling authored
slots. This fixes a bug where the number of nodes coming into the PHI node may not equal the number of predecessors. E.g., two or more landingpad instructions may require a PHI before reaching the eh.exception and eh.selector instructions. llvm-svn: 139035
-
Johnny Chen authored
Add a virtual method GetHardwareWatchpointHit() to the DNBArchProtocol base class which consults the architecture to return the watchpoint hit; otherwise return an invalid index. Add impl. of the method to X86_64 and I386 subclasses, plus reset the debug status register before we resume execution of the inferior thread. llvm-svn: 139034
-
Richard Trieu authored
llvm-svn: 139033
-
Howard Hinnant authored
llvm-svn: 139032
-
David Greene authored
Store a RecordVal's name as an Init to allow class-qualified Record members to reference Records that have Init names. We'll use this to provide more programmability in how we name defs and their associated members. llvm-svn: 139031
-
Fariborz Jahanian authored
llvm-svn: 139029
-
Kevin Enderby authored
case those instructions that the immediate is not sign-extend. radr://8795217 llvm-svn: 139028
-