- Oct 06, 2010
-
-
Jakob Stoklund Olesen authored
never kept after splitting. Keeping the original interval made sense when the split region doesn't modify the register, and the original is spilled. We can get the same effect by detecting reloaded values when spilling around copies. llvm-svn: 115695
-
Jakob Stoklund Olesen authored
llvm-svn: 115694
-
Chris Lattner authored
not getting marked as mayStore. This fixes llvm-gcc bootstrap. llvm-svn: 115693
-
Chris Lattner authored
llvm-svn: 115692
-
- Oct 05, 2010
-
-
Chris Lattner authored
X86InstrCMovSetCC.td | 200 ++++++--------------------------------------------- 1 file changed, 27 insertions(+), 173 deletions(-) llvm-svn: 115689
-
Rafael Espindola authored
llvm-svn: 115687
-
Chris Lattner authored
llvm-svn: 115686
-
Chris Lattner authored
gross hack (having the asmmatcher handle the alias). llvm-svn: 115685
-
Rafael Espindola authored
llvm-svn: 115684
-
Chris Lattner authored
the right places. X86Instr64bit.td now dies, long live x86-64! llvm-svn: 115669
-
Jakob Stoklund Olesen authored
erasing it from the visited set. That ensures we create the right phi defs. llvm-svn: 115666
-
Jakob Stoklund Olesen authored
Insert copy after defining instruction. Fix LiveIntervalMap::extendTo to properly handle live segments starting before the current basic block. Make sure the open live range is extended to the inserted copy's use slot. llvm-svn: 115665
-
Chris Lattner authored
llvm-svn: 115663
-
Chris Lattner authored
llvm-svn: 115660
-
Michael J. Spencer authored
llvm-svn: 115657
-
Michael J. Spencer authored
llvm-svn: 115656
-
Rafael Espindola authored
llvm-svn: 115655
-
Jakob Stoklund Olesen authored
llvm-svn: 115650
-
Jakob Stoklund Olesen authored
llvm-svn: 115649
-
Jim Grosbach authored
addressing mode from four to five. llvm-svn: 115645
-
Rafael Espindola authored
isInSymtab and isLocal logic in the two loops don't get easily out of sync. llvm-svn: 115643
-
Owen Anderson authored
Use a more efficient lowering of uint64_t --> float that can take advantage of hardware signed integer conversion without having to do a double cast (uint64_t --> double --> float). This is based on the algorithm from compiler_rt's __floatundisf for X86-64. llvm-svn: 115634
-
Chris Lattner authored
llvm-svn: 115632
-
Chris Lattner authored
llvm-svn: 115631
-
Chris Lattner authored
llvm-svn: 115627
-
Rafael Espindola authored
llvm-svn: 115623
-
Dan Gohman authored
Graphviz program, print something with a newline, to avoid leaving the line unfinished. llvm-svn: 115620
-
Rafael Espindola authored
they should be in the symbol table or not. Instead of "guessing", just compute the symbol table after the relocations are known. llvm-svn: 115619
-
Chris Lattner authored
llvm-svn: 115608
-
Chris Lattner authored
llvm-svn: 115607
-
Chris Lattner authored
llvm-svn: 115606
-
Chris Lattner authored
llvm-svn: 115605
-
Chris Lattner authored
bswap32 doesn't read eflags. llvm-svn: 115604
-
Chris Lattner authored
pseudo instructions. Move POPCNT to InstrSSE since they are SSE4 instructions. llvm-svn: 115603
-
Chris Lattner authored
with the 32-bit stuff. llvm-svn: 115602
-
Chris Lattner authored
llvm-svn: 115601
-
Chris Lattner authored
llvm-svn: 115600
-
Chris Lattner authored
llvm-svn: 115599
-
Chris Lattner authored
llvm-svn: 115598
-
Chris Lattner authored
llvm-svn: 115597
-