- Sep 05, 2008
-
-
Duncan Sands authored
edges one by one by hand. llvm-svn: 55836
-
Duncan Sands authored
llvm-svn: 55835
-
Duncan Sands authored
llvm-svn: 55834
-
Duncan Sands authored
it modifies a functions attributes. llvm-svn: 55831
-
Duncan Sands authored
elsewhere due to a missing pattern for v2f64 = sint_to_fp v2i32. That is PR2687. llvm-svn: 55828
-
Dan Gohman authored
llvm-svn: 55824
-
Dale Johannesen authored
but less accurate (non-IEEE) code sequences for certain math library functions. Add the first of several such expansions. Don't worry, if you don't turn it on it won't affect you. llvm-svn: 55823
-
Dan Gohman authored
expanding its operands. llvm-svn: 55820
-
Dan Gohman authored
to check the result of getRegForValue before using it, and to check for illegal operand types. llvm-svn: 55819
-
Dan Gohman authored
llvm-svn: 55818
-
Dan Gohman authored
in the same block. Fix the entry-block handling to only run at at the beginning of the entry block, and not any other times. llvm-svn: 55817
-
Dan Gohman authored
llvm-svn: 55816
-
Owen Anderson authored
pool loads on X86 in fast isel. This isn't actually used yet. llvm-svn: 55814
-
Dan Gohman authored
llvm-svn: 55811
-
Devang Patel authored
llvm-svn: 55809
-
Devang Patel authored
llvm-svn: 55808
-
Evan Cheng authored
For whatever the reason, x86 CallingConv::Fast (i.e. fastcc) was not passing scalar arguments in registers. This patch defines a new fastcc CC which is slightly different from the FastCall CC. In addition to passing integer arguments in ECX and EDX, it also specify doubles are passed in 8-byte slots which are 8-byte aligned (instead of 4-byte aligned). This avoids a potential performance hazard where doubles span cacheline boundaries. llvm-svn: 55807
-
Devang Patel authored
llvm-svn: 55806
-
- Sep 04, 2008
-
-
Dale Johannesen authored
emitted only when errno is not in use. llvm-svn: 55797
-
Devang Patel authored
llvm-svn: 55794
-
Dan Gohman authored
llvm-svn: 55793
-
Devang Patel authored
llvm-svn: 55792
-
Duncan Sands authored
llvm-svn: 55789
-
Devang Patel authored
llvm-svn: 55786
-
Andrew Lenharth authored
llvm-svn: 55785
-
Dale Johannesen authored
llvm-svn: 55784
-
Dale Johannesen authored
forms remain to handle older IR files, but will go away soon. llvm-svn: 55781
-
Dan Gohman authored
llvm-svn: 55779
-
Owen Anderson authored
llvm-svn: 55777
-
Dan Gohman authored
llvm-svn: 55769
-
Andrew Lenharth authored
llvm-svn: 55766
-
Owen Anderson authored
Dan or Evan, please review. llvm-svn: 55764
-
Evan Cheng authored
llvm-svn: 55763
-
Evan Cheng authored
Fix an overly strict assertion. Source register of a copy may not be killed, it may be killed by an implicit super-register use. llvm-svn: 55762
-
Evan Cheng authored
Remove code that pad number of bytes to pop for X86_FastCall CC. The code doesn't do the "aligning" for Cygwin, Mingw, and Windows. But aligning it on Darwin and Linux breaks gcc compatibility. That ruled out all the platforms we support! llvm-svn: 55756
-
Dale Johannesen authored
No functional change (and no FE change to generate them). llvm-svn: 55753
-
Dan Gohman authored
in FastISel. llvm-svn: 55748
-
Dan Gohman authored
classes in the llvm namespace having members with types from anonymous namespaces. llvm-svn: 55747
-
Dan Gohman authored
HandlePHINodesInSuccessorBlocks that works FastISel-style. This allows PHI nodes to be updated correctly while using FastISel. This also involves some code reorganization; ValueMap and MBBMap are now members of the FastISel class, so they needn't be passed around explicitly anymore. Also, SelectInstructions is changed to SelectInstruction, and only does one instruction at a time. llvm-svn: 55746
-
Devang Patel authored
llvm-svn: 55745
-