Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | Properly pseudo-ize the ARM LDMIA_RET instruction. This has the nice side- | Jim Grosbach | 2011-03-11 | 1 | -1/+1 |
| | | | | | | effect that we get proper instruction printing using the "pop" mnemonic for it. llvm-svn: 127502 | ||||
* | Eliminate the unused CodeGenPrepare option to split critical edges. | Cameron Zwarich | 2011-03-02 | 1 | -1/+1 |
| | | | | llvm-svn: 126825 | ||||
* | Explicitly disable CGP critical edge splitting for this test so it won't break | Jakob Stoklund Olesen | 2010-09-27 | 1 | -1/+1 |
| | | | | | | by reenabling it temporarily. llvm-svn: 114858 | ||||
* | Don't depend on basic block numbering. | Jakob Stoklund Olesen | 2010-09-27 | 1 | -7/+8 |
| | | | | llvm-svn: 114857 | ||||
* | Enable code placement optimization pass for ARM. | Evan Cheng | 2010-09-24 | 1 | -1/+50 |
| | | | | llvm-svn: 114746 | ||||
* | PHI elimination should not break back edge. It can cause some significant ↵ | Evan Cheng | 2010-08-17 | 1 | -0/+29 |
code placement issues. rdar://8263994 good: LBB0_2: mov r2, r0 . . . mov r1, r2 bne LBB0_2 bad: LBB0_2: mov r2, r0 . . . @ BB#3: mov r1, r2 b LBB0_2 llvm-svn: 111221 |