| 2012-07-18 |
Chandler Carruth | Fix a somewhat nasty crasher in PR13378. This crashes... |
blob | commitdiff | raw |
| 2012-07-17 |
Jakob Stoklund Olesen | Add some trace output to TwoAddressInstructionPass. |
blob | commitdiff | raw | diff to current |
| 2012-07-15 |
Chandler Carruth | Reapply r160194, switching to use LV information for... |
blob | commitdiff | raw | diff to current |
| 2012-07-13 |
Chandler Carruth | Revert r160194, which switched to use LV information... |
blob | commitdiff | raw | diff to current |
| 2012-07-13 |
Chandler Carruth | Use the LiveVariables information to efficiently get... |
blob | commitdiff | raw | diff to current |
| 2012-07-02 |
Manman Ren | Added assertion in getVRegDef of MachineRegisterInfo... |
blob | commitdiff | raw | diff to current |
| 2012-06-25 |
Jakob Stoklund Olesen | Handle <undef> operands in TwoAddressInstructionPass. |
blob | commitdiff | raw | diff to current |
| 2012-06-05 |
Andrew Trick | misched: API for minimum vs. expected latency. |
blob | commitdiff | raw | diff to current |
| 2012-05-20 |
Jakob Stoklund Olesen | Properly constrain register classes in 2-addr. |
blob | commitdiff | raw | diff to current |
| 2012-05-18 |
Evan Cheng | Teach two-address pass to update the "source" map so... |
blob | commitdiff | raw | diff to current |
| 2012-05-07 |
Jakob Stoklund Olesen | Add an MF argument to TRI::getPointerRegClass() and... |
blob | commitdiff | raw | diff to current |
| 2012-05-03 |
Evan Cheng | Fix two-address pass's aggressive instruction commuting... |
blob | commitdiff | raw | diff to current |
| 2012-05-03 |
Andrew Trick | Added TargetRegisterInfo::getAllocatableClass. |
blob | commitdiff | raw | diff to current |
| 2012-04-09 |
Lang Hames | Patch r153892 for PR11861 apparently broke an external... |
blob | commitdiff | raw | diff to current |
| 2012-04-04 |
Pete Cooper | REG_SEQUENCE expansion to COPY instructions wasn't... |
blob | commitdiff | raw | diff to current |
| 2012-04-04 |
Jakob Stoklund Olesen | Remove spurious debug output. |
blob | commitdiff | raw | diff to current |
| 2012-04-02 |
Lang Hames | During two-address lowering, rescheduling an instructio... |
blob | commitdiff | raw | diff to current |
| 2012-02-10 |
Andrew Trick | RegAlloc superpass: includes phi elimination, coalescin... |
blob | commitdiff | raw | diff to current |
| 2012-02-03 |
Andrew Trick | whitespace |
blob | commitdiff | raw | diff to current |
| 2012-01-24 |
Jakob Stoklund Olesen | Set correct <def,undef> flags when lowering REG_SEQUENCE. |
blob | commitdiff | raw | diff to current |
| 2012-01-24 |
Jakob Stoklund Olesen | Preserve <def,undef> flags in CoalesceExtSubRegs. |
blob | commitdiff | raw | diff to current |
| 2012-01-18 |
Pete Cooper | Fix ISD::REG_SEQUENCE to accept physical registers... |
blob | commitdiff | raw | diff to current |
| 2011-12-07 |
Evan Cheng | Add bundle aware API for querying instruction propertie... |
blob | commitdiff | raw | diff to current |
| 2011-12-03 |
Benjamin Kramer | Simplify code. No functionality change. |
blob | commitdiff | raw | diff to current |
| 2011-11-16 |
Evan Cheng | Disable expensive two-address optimizations at -O0... |
blob | commitdiff | raw | diff to current |
| 2011-11-16 |
Evan Cheng | Disable the assertion again. Looks like fastisel is... |
blob | commitdiff | raw | diff to current |
| 2011-11-16 |
Evan Cheng | Revert r144568 now that r144730 has fixed the fast... |
blob | commitdiff | raw | diff to current |
| 2011-11-16 |
Evan Cheng | If the 2addr instruction has other kills, don't move... |
blob | commitdiff | raw | diff to current |
| 2011-11-16 |
Evan Cheng | RescheduleKillAboveMI() must backtrack to before the... |
blob | commitdiff | raw | diff to current |
| 2011-11-16 |
Evan Cheng | Process all uses first before defs to accurately captur... |
blob | commitdiff | raw | diff to current |
| 2011-11-15 |
Evan Cheng | Set SeenStore to true to prevent loads from being moved... |
blob | commitdiff | raw | diff to current |
| 2011-11-14 |
Evan Cheng | Avoid dereferencing off the beginning of lists. |
blob | commitdiff | raw | diff to current |
| 2011-11-14 |
Evan Cheng | At -O0, multiple uses of a virtual registers in the... |
blob | commitdiff | raw | diff to current |
| 2011-11-14 |
Evan Cheng | Teach two-address pass to re-schedule two-address instr... |
blob | commitdiff | raw | diff to current |
| 2011-09-23 |
Eli Friedman | PR10998: It is not legal to sink an instruction past... |
blob | commitdiff | raw | diff to current |
| 2011-07-29 |
Jakob Stoklund Olesen | Add an isSSA() flag to MachineRegisterInfo. |
blob | commitdiff | raw | diff to current |
| 2011-06-28 |
Evan Cheng | - Rename TargetInstrDesc, TargetOperandInfo to MCInstrD... |
blob | commitdiff | raw | diff to current |
| 2011-06-27 |
Evan Cheng | More refactoring. Move getRegClass from TargetOperandIn... |
blob | commitdiff | raw | diff to current |
| 2011-06-18 |
Benjamin Kramer | Simplify code. No change in functionality. |
blob | commitdiff | raw | diff to current |
| 2011-06-07 |
Cameron Zwarich | Fix an issue where the two-address conversion pass... |
blob | commitdiff | raw | diff to current |
| 2011-03-02 |
Evan Cheng | Catch more cases where 2-address pass should 3-addressi... |
blob | commitdiff | raw | diff to current |
| 2011-02-10 |
Evan Cheng | After 3-addressifying a two-address instruction, update... |
blob | commitdiff | raw | diff to current |
| 2011-01-10 |
Jakob Stoklund Olesen | Simplify a bunch of isVirtualRegister() and isPhysicalR... |
blob | commitdiff | raw | diff to current |
| 2011-01-09 |
Jakob Stoklund Olesen | Shrink a BitVector that didn't mean to store bits for... |
blob | commitdiff | raw | diff to current |
| 2011-01-07 |
Evan Cheng | Do not model all INLINEASM instructions as having unmod... |
blob | commitdiff | raw | diff to current |
| 2010-12-19 |
Cameron Zwarich | StrongPHIElimination will never run before TwoAddressIn... |
blob | commitdiff | raw | diff to current |
| 2010-12-19 |
Cameron Zwarich | Remove some checks for StrongPHIElim. These checks... |
blob | commitdiff | raw | diff to current |
| 2010-12-17 |
Bob Wilson | Fix crash compiling a QQQQ REG_SEQUENCE for a Neon... |
blob | commitdiff | raw | diff to current |
| 2010-12-14 |
Evan Cheng | Fix a minor bug in two-address pass. It was missing... |
blob | commitdiff | raw | diff to current |
| 2010-10-21 |
Duncan Sands | Remove some variables that are never really used |
blob | commitdiff | raw | diff to current |
| 2010-10-19 |
Owen Anderson | Get rid of static constructors for pass registration... |
blob | commitdiff | raw | diff to current |
| 2010-10-12 |
Owen Anderson | Begin adding static dependence information to passes... |
blob | commitdiff | raw | diff to current |
| 2010-10-07 |
Owen Anderson | Now with fewer extraneous semicolons! |
blob | commitdiff | raw | diff to current |
| 2010-08-23 |
Owen Anderson | Now that PassInfo and Pass::ID have been separated... |
blob | commitdiff | raw | diff to current |
| 2010-08-09 |
Jakob Stoklund Olesen | A REG_SEQUENCE instruction may use the same register... |
blob | commitdiff | raw | diff to current |
| 2010-08-06 |
Owen Anderson | Reapply r110396, with fixes to appease the Linux buildb... |
blob | commitdiff | raw | diff to current |
| 2010-08-06 |
Owen Anderson | Revert r110396 to fix buildbots. |
blob | commitdiff | raw | diff to current |
| 2010-08-05 |
Owen Anderson | Don't use PassInfo* as a type identifier for passes... |
blob | commitdiff | raw | diff to current |
| 2010-07-16 |
Jakob Stoklund Olesen | Remove many calls to TII::isMoveInstr. Targets should... |
blob | commitdiff | raw | diff to current |
| 2010-07-10 |
Jakob Stoklund Olesen | Replace copyRegToReg with COPY everywhere in lib/CodeGe... |
blob | commitdiff | raw | diff to current |
| 2010-07-10 |
Jakob Stoklund Olesen | Only collect subreg extracting copies for later coalescing. |
blob | commitdiff | raw | diff to current |
| 2010-07-10 |
Jakob Stoklund Olesen | Emit COPY instructions instead of using copyRegToReg... |
blob | commitdiff | raw | diff to current |
| 2010-07-08 |
Jakob Stoklund Olesen | Convert EXTRACT_SUBREG to COPY when emitting machine... |
blob | commitdiff | raw | diff to current |
| 2010-07-08 |
Jakob Stoklund Olesen | Remove references to INSERT_SUBREG after de-SSA. |
blob | commitdiff | raw | diff to current |
| 2010-07-07 |
Jakob Stoklund Olesen | Revert "Remove references to INSERT_SUBREG after de... |
blob | commitdiff | raw | diff to current |
| 2010-07-06 |
Jakob Stoklund Olesen | Remove references to INSERT_SUBREG after de-SSA |
blob | commitdiff | raw | diff to current |
| 2010-07-06 |
Jakob Stoklund Olesen | Convert INSERT_SUBREG to COPY in TwoAddressInstructionPass. |
blob | commitdiff | raw | diff to current |
| 2010-07-03 |
Jakob Stoklund Olesen | Detect and handle COPY in many places. |
blob | commitdiff | raw | diff to current |
| 2010-07-02 |
Evan Cheng | - Two-address pass should not assume unfolding is alway... |
blob | commitdiff | raw | diff to current |
| 2010-06-22 |
Dan Gohman | When unfolding a load, avoid assuming which instruction... |
blob | commitdiff | raw | diff to current |
| 2010-06-22 |
Dan Gohman | Fix the new load-unfolding code to update LiveVariable... |
blob | commitdiff | raw | diff to current |
| 2010-06-21 |
Dan Gohman | Teach two-address lowering how to unfold a load to... |
blob | commitdiff | raw | diff to current |
| 2010-06-18 |
Jakob Stoklund Olesen | Only run CoalesceExtSubRegs when we can expect LiveInte... |
blob | commitdiff | raw | diff to current |
| 2010-06-15 |
Bob Wilson | Add some missing checks for the case where the extract_... |
blob | commitdiff | raw | diff to current |
| 2010-06-15 |
Bob Wilson | Generalize the pre-coalescing of extract_subregs feedin... |
blob | commitdiff | raw | diff to current |
| 2010-06-09 |
Evan Cheng | Allow target to place 2-address pass inserted copies... |
blob | commitdiff | raw | diff to current |
| 2010-06-07 |
Bob Wilson | Fix a mistake in my previous change r105437: don't... |
blob | commitdiff | raw | diff to current |
| 2010-06-03 |
Bob Wilson | Add some missing checks in TwoAddressInstructionPass... |
blob | commitdiff | raw | diff to current |
| 2010-06-02 |
Jakob Stoklund Olesen | Slightly change the meaning of the reMaterialize target... |
blob | commitdiff | raw | diff to current |
| 2010-06-02 |
Bob Wilson | Rename canCombinedSubRegIndex method to something more... |
blob | commitdiff | raw | diff to current |
| 2010-06-02 |
Bob Wilson | Fix an obvious mistake: don't change the operands until... |
blob | commitdiff | raw | diff to current |
| 2010-05-29 |
Jakob Stoklund Olesen | Handle composed subreg indices when processing REQ_SEQU... |
blob | commitdiff | raw | diff to current |
| 2010-05-21 |
Evan Cheng | - Change MachineInstr::findRegisterDefOperandIdx so... |
blob | commitdiff | raw | diff to current |
| 2010-05-19 |
Jakob Stoklund Olesen | TwoAddressInstructionPass doesn't really know how to... |
blob | commitdiff | raw | diff to current |
| 2010-05-17 |
Evan Cheng | Fix PR7175. Insert copies of a REG_SEQUENCE source... |
blob | commitdiff | raw | diff to current |
| 2010-05-17 |
Evan Cheng | Fix PR7156. If the sources of a REG_SEQUENCE are all... |
blob | commitdiff | raw | diff to current |
| 2010-05-17 |
Evan Cheng | Careful with reg_sequence coalescing to not to overwrit... |
blob | commitdiff | raw | diff to current |
| 2010-05-14 |
Evan Cheng | Teach two-address pass to do some coalescing while... |
blob | commitdiff | raw | diff to current |
| 2010-05-13 |
Evan Cheng | If REG_SEQUENCE source is livein, copy it first. Also... |
blob | commitdiff | raw | diff to current |
| 2010-05-12 |
Evan Cheng | Code clean up. |
blob | commitdiff | raw | diff to current |
| 2010-05-11 |
Evan Cheng | Ensure REG_SEQUENCE source operands are unique. |
blob | commitdiff | raw | diff to current |
| 2010-05-10 |
Evan Cheng | Clear RegSequences vector after eliminating REG_SEQUENC... |
blob | commitdiff | raw | diff to current |
| 2010-05-06 |
Dan Gohman | Add a DebugLoc argument to TargetInstrInfo::copyRegToRe... |
blob | commitdiff | raw | diff to current |
| 2010-05-05 |
Evan Cheng | Move REG_SEQUENCE removal to 2addr pass. |
blob | commitdiff | raw | diff to current |
| 2010-03-23 |
Evan Cheng | Ignore dbg_value's. |
blob | commitdiff | raw | diff to current |
| 2010-03-03 |
Evan Cheng | Add MachineRegisterInfo::hasOneUse and hasOneNonDBGUse. |
blob | commitdiff | raw | diff to current |
| 2010-03-02 |
Evan Cheng | Swap parameters of isSafeToMove and isSafeToReMat for... |
blob | commitdiff | raw | diff to current |
| 2010-02-11 |
Dale Johannesen | Don't allow DBG_VALUE to affect codegen. |
blob | commitdiff | raw | diff to current |
| 2010-02-10 |
Dale Johannesen | Skip debug info in a couple of places. |
blob | commitdiff | raw | diff to current |
| 2010-02-09 |
Chris Lattner | move target-independent opcodes out of TargetInstrInfo |
blob | commitdiff | raw | diff to current |
| next |
projects / oota-llvm.git / history