| 2009-03-12 |
Evan Cheng | Add this test back.
|
commit | commitdiff | tree |
| 2009-03-12 |
Evan Cheng | Re-apply 66024 with fixes: 1. Fixed indirect call to...
|
commit | commitdiff | tree |
| 2009-03-12 |
Evan Cheng | Typo.
|
commit | commitdiff | tree |
| 2009-03-12 |
Evan Cheng | Fix test after Chris' select changes.
|
commit | commitdiff | tree |
| 2009-03-12 |
Evan Cheng | Enable Chris' value propagation change. It make available...
|
commit | commitdiff | tree |
| 2009-03-12 |
Evan Cheng | On x86, if the only use of a i64 load is a i64 store...
|
commit | commitdiff | tree |
| 2009-03-12 |
Evan Cheng | Also pass -gcc-tool-args when building a shared object.
|
commit | commitdiff | tree |
| 2009-03-11 |
Evan Cheng | My last coalescer fix introduced a subtler one. It...
|
commit | commitdiff | tree |
| 2009-03-11 |
Evan Cheng | Two coalescer fixes in one.
|
commit | commitdiff | tree |
| 2009-03-10 |
Evan Cheng | Revert 66358 for now. It's breaking povray, 450.soplex...
|
commit | commitdiff | tree |
| 2009-03-10 |
Evan Cheng | If a function is marked alwaysinline, it must be inlined...
|
commit | commitdiff | tree |
| 2009-03-09 |
Evan Cheng | ARM target now also recognize triplets like thumbv6...
|
commit | commitdiff | tree |
| 2009-03-09 |
Evan Cheng | ARM isLegalAddressImmediate should check if type is...
|
commit | commitdiff | tree |
| 2009-03-09 |
Evan Cheng | Yet another case where the spiller marked two uses...
|
commit | commitdiff | tree |
| 2009-03-09 |
Evan Cheng | Re-apply 66315, but restrict it to Darwin only.
|
commit | commitdiff | tree |
| 2009-03-08 |
Evan Cheng | Recognize triplets starting with armv5-, armv6- etc...
|
commit | commitdiff | tree |
| 2009-03-08 |
Evan Cheng | If a MI uses the same register more than once, only...
|
commit | commitdiff | tree |
| 2009-03-07 |
Evan Cheng | If ARCH is x86_64, pass -m64 to the host compiler....
|
commit | commitdiff | tree |
| 2009-03-06 |
Evan Cheng | SRThreshold is meant to be inclusive.
|
commit | commitdiff | tree |
| 2009-03-05 |
Evan Cheng | Do not split edges to EH landing pads. It will cause...
|
commit | commitdiff | tree |
| 2009-03-05 |
Evan Cheng | Fix how livein live intervals are handled. Previously...
|
commit | commitdiff | tree |
| 2009-03-04 |
Evan Cheng | Fix PR3666: isel calls to constant addresses.
|
commit | commitdiff | tree |
| 2009-03-04 |
Evan Cheng | Rename test.
|
commit | commitdiff | tree |
| 2009-03-04 |
Evan Cheng | Fix PR3701. 1. X86 target renamed eflags register to...
|
commit | commitdiff | tree |
| 2009-03-01 |
Evan Cheng | Minor optimization:
|
commit | commitdiff | tree |
| 2009-02-28 |
Evan Cheng | Last commit accidentially deleted this code.
|
commit | commitdiff | tree |
| 2009-02-28 |
Evan Cheng | Avoid unused parameter warning.
|
commit | commitdiff | tree |
| 2009-02-27 |
Evan Cheng | Make sure this test passes on linux-ppc.
|
commit | commitdiff | tree |
| 2009-02-27 |
Evan Cheng | MachineLICM CSE should match destination register classes...
|
commit | commitdiff | tree |
| 2009-02-26 |
Evan Cheng | ADDS{D|S}rr_Int and MULS{D|S}rr_Int are not commutable...
|
commit | commitdiff | tree |
| 2009-02-26 |
Evan Cheng | The last commit was overly conservative. It's ok to...
|
commit | commitdiff | tree |
| 2009-02-26 |
Evan Cheng | If an available register falls through to a succ block...
|
commit | commitdiff | tree |
| 2009-02-25 |
Evan Cheng | Revert BuildVectorSDNode related patches: 65426, 65427...
|
commit | commitdiff | tree |
| 2009-02-25 |
Evan Cheng | Clean up dwarf writer, part 1. This eliminated the...
|
commit | commitdiff | tree |
| 2009-02-23 |
Evan Cheng | Only v1i16 (i.e. _m64) is returned via RAX / RDX.
|
commit | commitdiff | tree |
| 2009-02-22 |
Evan Cheng | If a use operand is marked isKill, don't forget to...
|
commit | commitdiff | tree |
| 2009-02-22 |
Evan Cheng | Add a note.
|
commit | commitdiff | tree |
| 2009-02-22 |
Evan Cheng | Be bug compatible with gcc by returning MMX values...
|
commit | commitdiff | tree |
| 2009-02-22 |
Evan Cheng | Do not consider MMX_MOVD64rr a move instructions. The...
|
commit | commitdiff | tree |
| 2009-02-22 |
Evan Cheng | Only try to sink immediate when TLI is not null. It...
|
commit | commitdiff | tree |
| 2009-02-21 |
Evan Cheng | Add AddrModeMatcher.cpp
|
commit | commitdiff | tree |
| 2009-02-21 |
Evan Cheng | If two-address def is dead and the instruction does...
|
commit | commitdiff | tree |
| 2009-02-21 |
Evan Cheng | Teach LSR sink to sink the immediate portion of the...
|
commit | commitdiff | tree |
| 2009-02-20 |
Evan Cheng | Fix strange logic in CollectIVUsers used to determine...
|
commit | commitdiff | tree |
| 2009-02-20 |
Evan Cheng | Support return of MMX values in 64-bit mode.
|
commit | commitdiff | tree |
| 2009-02-20 |
Evan Cheng | Factor address mode matcher out of codegen prepare...
|
commit | commitdiff | tree |
| 2009-02-18 |
Evan Cheng | GV with null value initializer shouldn't go to BSS...
|
commit | commitdiff | tree |
| 2009-02-17 |
Evan Cheng | A couple of places where reused use operands should...
|
commit | commitdiff | tree |
| 2009-02-17 |
Evan Cheng | Strengthen the "non-constant stride must dominate loop...
|
commit | commitdiff | tree |
| 2009-02-15 |
Evan Cheng | Fix PR3522. It's not safe to sink into landing pad...
|
commit | commitdiff | tree |
| 2009-02-15 |
Evan Cheng | Fix pr3571: If stride is a value defined by an instruction...
|
commit | commitdiff | tree |
| 2009-02-15 |
Evan Cheng | ifdef out unneeded if statement.
|
commit | commitdiff | tree |
| 2009-02-13 |
Evan Cheng | Teach x86 target -soft-float.
|
commit | commitdiff | tree |
| 2009-02-13 |
Evan Cheng | Switch from new[] + delete[] to malloc + free since...
|
commit | commitdiff | tree |
| 2009-02-13 |
Evan Cheng | If new[] fails, return 0 rather then trying to dereference...
|
commit | commitdiff | tree |
| 2009-02-12 |
Evan Cheng | It's (currently) not safe to keep certain physical...
|
commit | commitdiff | tree |
| 2009-02-12 |
Evan Cheng | Oops. Last second clean up messed things up.
|
commit | commitdiff | tree |
| 2009-02-12 |
Evan Cheng | If availability info is kept when fallthrough into...
|
commit | commitdiff | tree |
| 2009-02-12 |
Evan Cheng | Replace one of burr scheduling heuristic with something...
|
commit | commitdiff | tree |
| 2009-02-11 |
Evan Cheng | Apparently some MachineBasicBlock's don't have corresponding...
|
commit | commitdiff | tree |
| 2009-02-11 |
Evan Cheng | Remove a bogus assertion. It's possible a live-in available...
|
commit | commitdiff | tree |
| 2009-02-11 |
Evan Cheng | Implement PR3495: local spiller optimization. The local...
|
commit | commitdiff | tree |
| 2009-02-10 |
Evan Cheng | Handle llvm.x86.sse2.maskmov.dqu in 64-bit.
|
commit | commitdiff | tree |
| 2009-02-10 |
Evan Cheng | 80 col violations.
|
commit | commitdiff | tree |
| 2009-02-10 |
Evan Cheng | Fix PR3457: Ignore control successors when looking...
|
commit | commitdiff | tree |
| 2009-02-09 |
Evan Cheng | Implement FpSET_ST1_*.
|
commit | commitdiff | tree |
| 2009-02-09 |
Evan Cheng | If the target cannot issue a copy for the given source...
|
commit | commitdiff | tree |
| 2009-02-09 |
Evan Cheng | Simplify code.
|
commit | commitdiff | tree |
| 2009-02-09 |
Evan Cheng | Make sure constant subscript is truncated to ptr size...
|
commit | commitdiff | tree |
| 2009-02-09 |
Evan Cheng | Re-enable machine sinking pass now that the coalescer...
|
commit | commitdiff | tree |
| 2009-02-09 |
Evan Cheng | Fix another case ShortenDeadCopySrcLiveRange is shortening...
|
commit | commitdiff | tree |
| 2009-02-09 |
Evan Cheng | Turns out AnalyzeBranch can modify the mbb being analyzed...
|
commit | commitdiff | tree |
| 2009-02-09 |
Evan Cheng | Eliminate a 'control reaches end of non-void function...
|
commit | commitdiff | tree |
| 2009-02-08 |
Evan Cheng | Fix PR3486. Fix a bug in code that manually patch physical...
|
commit | commitdiff | tree |
| 2009-02-08 |
Evan Cheng | Strengthen the previous check.
|
commit | commitdiff | tree |
| 2009-02-08 |
Evan Cheng | r64073 commit message is lost. Here it is:
|
commit | commitdiff | tree |
| 2009-02-08 |
Evan Cheng | git-svn-id: https://llvm.org/svn/llvm-project/llvm...
|
commit | commitdiff | tree |
| 2009-02-07 |
Evan Cheng | Revert 64023. make prefers GNUmakefile over makefile.
|
commit | commitdiff | tree |
| 2009-02-07 |
Evan Cheng | Move Apple style build makefiles to the top level....
|
commit | commitdiff | tree |
| 2009-02-07 |
Evan Cheng | Enable machine sinking pass in non-fast mode.
|
commit | commitdiff | tree |
| 2009-02-07 |
Evan Cheng | Don't sink the instruction if TargetRegisterInfo::isSafeToMo...
|
commit | commitdiff | tree |
| 2009-02-06 |
Evan Cheng | We don't really need the abort here. This unbreak x86_64...
|
commit | commitdiff | tree |
| 2009-02-06 |
Evan Cheng | Move getPointerRegClass from TargetInstrInfo to TargetRegist...
|
commit | commitdiff | tree |
| 2009-02-06 |
Evan Cheng | Add TargetInstrInfo::isSafeToMoveRegisterClassDefs...
|
commit | commitdiff | tree |
| 2009-02-06 |
Evan Cheng | Change -1 => negative number.
|
commit | commitdiff | tree |
| 2009-02-06 |
Evan Cheng | Document the meaning of -1 for getCopyCost.
|
commit | commitdiff | tree |
| 2009-02-06 |
Evan Cheng | Fix test. It produces unexpected code if sse4.1 is on.
|
commit | commitdiff | tree |
| 2009-02-05 |
Evan Cheng | isAsCheapAsMove instructions can have register src...
|
commit | commitdiff | tree |
| 2009-02-05 |
Evan Cheng | Machine LICM increases register pressure and it almost...
|
commit | commitdiff | tree |
| 2009-02-05 |
Evan Cheng | Turn on machine LICM in non-fast mode.
|
commit | commitdiff | tree |
| 2009-02-05 |
Evan Cheng | Teach machine licm to CSE hoisted instructions.
|
commit | commitdiff | tree |
| 2009-02-05 |
Evan Cheng | ReMaterializeTrivialDef need to trim the live interval...
|
commit | commitdiff | tree |
| 2009-02-05 |
Evan Cheng | A few more isAsCheapAsAMove.
|
commit | commitdiff | tree |
| 2009-02-05 |
Evan Cheng | Clarify comments.
|
commit | commitdiff | tree |
| 2009-02-04 |
Evan Cheng | Skip over zero registers.
|
commit | commitdiff | tree |
| 2009-02-04 |
Evan Cheng | Don't call isInvariantLoad twice.
|
commit | commitdiff | tree |
| 2009-02-04 |
Evan Cheng | Don't bother hoisting out a "cheap" instruction if...
|
commit | commitdiff | tree |
| 2009-02-04 |
Evan Cheng | For now, only hoist re-materilizable instructions....
|
commit | commitdiff | tree |
| 2009-02-03 |
Evan Cheng | APInt'fy SimplifyDemandedVectorElts so it can analyze...
|
commit | commitdiff | tree |
| 2009-02-02 |
Evan Cheng | Refactor PerformPHIConstruction, no functionality changes.
|
commit | commitdiff | tree |
| next |
projects / oota-llvm.git / search