| Commit message (Expand) | Author | Age | Files | Lines |
* | Disable rematerialization in TwoAddressInstructionPass. | Manman Ren | 2012-07-25 | 1 | -78/+6 |
* | Fix a somewhat nasty crasher in PR13378. This crashes inside of | Chandler Carruth | 2012-07-18 | 1 | -22/+32 |
* | Add some trace output to TwoAddressInstructionPass. | Jakob Stoklund Olesen | 2012-07-17 | 1 | -1/+4 |
* | Reapply r160194, switching to use LV information for finding local kills. | Chandler Carruth | 2012-07-15 | 1 | -56/+32 |
* | Revert r160194, which switched to use LV information for finding local | Chandler Carruth | 2012-07-13 | 1 | -19/+54 |
* | Use the LiveVariables information to efficiently get local kills. This | Chandler Carruth | 2012-07-13 | 1 | -54/+19 |
* | Added assertion in getVRegDef of MachineRegisterInfo to make sure the virtual | Manman Ren | 2012-07-02 | 1 | -5/+6 |
* | Handle <undef> operands in TwoAddressInstructionPass. | Jakob Stoklund Olesen | 2012-06-25 | 1 | -12/+31 |
* | misched: API for minimum vs. expected latency. | Andrew Trick | 2012-06-05 | 1 | -1/+1 |
* | Properly constrain register classes in 2-addr. | Jakob Stoklund Olesen | 2012-05-20 | 1 | -0/+6 |
* | Teach two-address pass to update the "source" map so it doesn't perform a | Evan Cheng | 2012-05-18 | 1 | -1/+7 |
* | Add an MF argument to TRI::getPointerRegClass() and TII::getRegClass(). | Jakob Stoklund Olesen | 2012-05-07 | 1 | -1/+1 |
* | Fix two-address pass's aggressive instruction commuting heuristics. It's meant | Evan Cheng | 2012-05-03 | 1 | -15/+16 |
* | Added TargetRegisterInfo::getAllocatableClass. | Andrew Trick | 2012-05-03 | 1 | -1/+2 |
* | Patch r153892 for PR11861 apparently broke an external project (see PR12493). | Lang Hames | 2012-04-09 | 1 | -16/+17 |
* | REG_SEQUENCE expansion to COPY instructions wasn't taking account of sub regi... | Pete Cooper | 2012-04-04 | 1 | -1/+2 |
* | Remove spurious debug output. | Jakob Stoklund Olesen | 2012-04-04 | 1 | -1/+0 |
* | During two-address lowering, rescheduling an instruction does not untie | Lang Hames | 2012-04-02 | 1 | -2/+2 |
* | RegAlloc superpass: includes phi elimination, coalescing, and scheduling. | Andrew Trick | 2012-02-10 | 1 | -1/+0 |
* | whitespace | Andrew Trick | 2012-02-03 | 1 | -8/+8 |
* | Set correct <def,undef> flags when lowering REG_SEQUENCE. | Jakob Stoklund Olesen | 2012-01-24 | 1 | -0/+46 |
* | Preserve <def,undef> flags in CoalesceExtSubRegs. | Jakob Stoklund Olesen | 2012-01-24 | 1 | -3/+7 |
* | Fix ISD::REG_SEQUENCE to accept physical registers and change TwoAddressInstr... | Pete Cooper | 2012-01-18 | 1 | -9/+12 |
* | Add bundle aware API for querying instruction properties and switch the code | Evan Cheng | 2011-12-07 | 1 | -18/+13 |
* | Simplify code. No functionality change. | Benjamin Kramer | 2011-12-03 | 1 | -2/+1 |
* | Disable expensive two-address optimizations at -O0. rdar://10453055 | Evan Cheng | 2011-11-16 | 1 | -0/+8 |
* | Disable the assertion again. Looks like fastisel is still generating bad kill... | Evan Cheng | 2011-11-16 | 1 | -1/+2 |
* | Revert r144568 now that r144730 has fixed the fast-isel kill marker bug. | Evan Cheng | 2011-11-16 | 1 | -2/+1 |
* | If the 2addr instruction has other kills, don't move it below any other uses ... | Evan Cheng | 2011-11-16 | 1 | -2/+7 |
* | RescheduleKillAboveMI() must backtrack to before the rescheduled DBG_VALUE in... | Evan Cheng | 2011-11-16 | 1 | -1/+1 |
* | Process all uses first before defs to accurately capture register liveness. r... | Evan Cheng | 2011-11-16 | 1 | -7/+13 |
* | Set SeenStore to true to prevent loads from being moved; also eliminates a no... | Evan Cheng | 2011-11-15 | 1 | -2/+2 |
* | Avoid dereferencing off the beginning of lists. | Evan Cheng | 2011-11-14 | 1 | -7/+4 |
* | At -O0, multiple uses of a virtual registers in the same BB are being marked | Evan Cheng | 2011-11-14 | 1 | -1/+2 |
* | Teach two-address pass to re-schedule two-address instructions (or the kill | Evan Cheng | 2011-11-14 | 1 | -19/+356 |
* | PR10998: It is not legal to sink an instruction past the terminator of a bloc... | Eli Friedman | 2011-09-23 | 1 | -1/+9 |
* | Add an isSSA() flag to MachineRegisterInfo. | Jakob Stoklund Olesen | 2011-07-29 | 1 | -0/+3 |
* | - Rename TargetInstrDesc, TargetOperandInfo to MCInstrDesc and MCOperandInfo and | Evan Cheng | 2011-06-28 | 1 | -16/+17 |
* | More refactoring. Move getRegClass from TargetOperandInfo to TargetInstrInfo. | Evan Cheng | 2011-06-27 | 1 | -1/+1 |
* | Simplify code. No change in functionality. | Benjamin Kramer | 2011-06-18 | 1 | -6/+1 |
* | Fix an issue where the two-address conversion pass incorrectly rewrites untied | Cameron Zwarich | 2011-06-07 | 1 | -9/+16 |
* | Catch more cases where 2-address pass should 3-addressify instructions. rdar:... | Evan Cheng | 2011-03-02 | 1 | -49/+70 |
* | After 3-addressifying a two-address instruction, update the register maps; ad... | Evan Cheng | 2011-02-10 | 1 | -4/+9 |
* | Simplify a bunch of isVirtualRegister() and isPhysicalRegister() logic. | Jakob Stoklund Olesen | 2011-01-10 | 1 | -1/+1 |
* | Shrink a BitVector that didn't mean to store bits for all physical registers. | Jakob Stoklund Olesen | 2011-01-09 | 1 | -6/+4 |
* | Do not model all INLINEASM instructions as having unmodelled side effects. | Evan Cheng | 2011-01-07 | 1 | -1/+1 |
* | StrongPHIElimination will never run before TwoAddressInstructionPass. | Cameron Zwarich | 2010-12-19 | 1 | -1/+0 |
* | Remove some checks for StrongPHIElim. These checks make it impossible to use an | Cameron Zwarich | 2010-12-19 | 1 | -4/+2 |
* | Fix crash compiling a QQQQ REG_SEQUENCE for a Neon vld3_lane operation. | Bob Wilson | 2010-12-17 | 1 | -2/+5 |
* | Fix a minor bug in two-address pass. It was missing a commute opportunity. | Evan Cheng | 2010-12-14 | 1 | -1/+2 |