| 2008-11-04 |
Evan Cheng | hasDisassembler should return false if disassembler...
|
commit | commitdiff | tree |
| 2008-11-04 |
Evan Cheng | Handle ARM machine constantpool entries.
|
commit | commitdiff | tree |
| 2008-11-03 |
Evan Cheng | Remove a dead switch statement.
|
commit | commitdiff | tree |
| 2008-11-03 |
Evan Cheng | Minor code restructuring. No functionality change.
|
commit | commitdiff | tree |
| 2008-11-03 |
Evan Cheng | Silence a compiler warning.
|
commit | commitdiff | tree |
| 2008-10-31 |
Evan Cheng | Add comment.
|
commit | commitdiff | tree |
| 2008-10-31 |
Evan Cheng | Use better data structure for ConstPoolId2AddrMap.
|
commit | commitdiff | tree |
| 2008-10-31 |
Evan Cheng | Actually make debug output understandable.
|
commit | commitdiff | tree |
| 2008-10-31 |
Evan Cheng | Forgot this in last commit.
|
commit | commitdiff | tree |
| 2008-10-31 |
Evan Cheng | Encode PICADD; some code clean up.
|
commit | commitdiff | tree |
| 2008-10-31 |
Evan Cheng | Change x86 register allocation ordering to match that...
|
commit | commitdiff | tree |
| 2008-10-31 |
Evan Cheng | Add a fixme.
|
commit | commitdiff | tree |
| 2008-10-30 |
Evan Cheng | I think we got non-machine specific constpool entries...
|
commit | commitdiff | tree |
| 2008-10-30 |
Evan Cheng | ARM JIT should observe -relocation-model command line...
|
commit | commitdiff | tree |
| 2008-10-29 |
Evan Cheng | Correct way to handle CONSTPOOL_ENTRY instructions.
|
commit | commitdiff | tree |
| 2008-10-29 |
Evan Cheng | Add debugging support.
|
commit | commitdiff | tree |
| 2008-10-29 |
Evan Cheng | Let target resolve some relocation results.
|
commit | commitdiff | tree |
| 2008-10-29 |
Evan Cheng | This is not needed anymore.
|
commit | commitdiff | tree |
| 2008-10-29 |
Evan Cheng | Add a bit to MachineRelocation that tells JIT that...
|
commit | commitdiff | tree |
| 2008-10-29 |
Evan Cheng | - More pre-split fixes: spill slot live interval computation...
|
commit | commitdiff | tree |
| 2008-10-29 |
Evan Cheng | - Rewrite code that update register live interval that...
|
commit | commitdiff | tree |
| 2008-10-28 |
Evan Cheng | Avoid calls to setSubgraphColor in release mode. They...
|
commit | commitdiff | tree |
| 2008-10-28 |
Evan Cheng | If def is in the same mbb as the barrier, spilt the...
|
commit | commitdiff | tree |
| 2008-10-28 |
Evan Cheng | Add command line option to limit the number splits...
|
commit | commitdiff | tree |
| 2008-10-28 |
Evan Cheng | Avoid putting a split past the end of the live range...
|
commit | commitdiff | tree |
| 2008-10-27 |
Evan Cheng | Silence a bogus compile time warning.
|
commit | commitdiff | tree |
| 2008-10-27 |
Evan Cheng | Remove val# defined by a remat'ed def that is now dead.
|
commit | commitdiff | tree |
| 2008-10-27 |
Evan Cheng | For now, don't split live intervals around x87 stack...
|
commit | commitdiff | tree |
| 2008-10-26 |
Evan Cheng | Do not shrink wrap live interval in a mbb if it's livein...
|
commit | commitdiff | tree |
| 2008-10-25 |
Evan Cheng | Handle cases where there aren't uses in the barrier...
|
commit | commitdiff | tree |
| 2008-10-25 |
Evan Cheng | Add storeRegTo{StackSlot|Addr} and loadRegFrom{StackSlot...
|
commit | commitdiff | tree |
| 2008-10-25 |
Evan Cheng | If val# def is ~0U, meaning it's defined by a PHI,...
|
commit | commitdiff | tree |
| 2008-10-24 |
Evan Cheng | Fix a pasto.
|
commit | commitdiff | tree |
| 2008-10-24 |
Evan Cheng | Fix a end() dereference; remove an abort() that wasn...
|
commit | commitdiff | tree |
| 2008-10-24 |
Evan Cheng | Avoid splitting an interval multiple times; avoid splitting...
|
commit | commitdiff | tree |
| 2008-10-23 |
Evan Cheng | Committing a good chunk of the pre-register allocation...
|
commit | commitdiff | tree |
| 2008-10-20 |
Evan Cheng | Add skeleton for the pre-register allocation live interval...
|
commit | commitdiff | tree |
| 2008-10-20 |
Evan Cheng | Add a register class -> virtual registers map.
|
commit | commitdiff | tree |
| 2008-10-20 |
Evan Cheng | This forward declaration is unnecessary.
|
commit | commitdiff | tree |
| 2008-10-18 |
Evan Cheng | By min, I mean max.
|
commit | commitdiff | tree |
| 2008-10-18 |
Evan Cheng | When creating intervals, leave min(1, numdefs) holes...
|
commit | commitdiff | tree |
| 2008-10-17 |
Evan Cheng | Add implicit defs of XMM8 to XMM15 on 32-bit call instructio...
|
commit | commitdiff | tree |
| 2008-10-17 |
Evan Cheng | Add RCBarriers to TargetInstrDesc. It's a list of register...
|
commit | commitdiff | tree |
| 2008-10-17 |
Evan Cheng | Fix PR2898. Spiller delete a store for reuse before...
|
commit | commitdiff | tree |
| 2008-10-17 |
Evan Cheng | Fix lfence and mfence encoding. These look like MRM5r...
|
commit | commitdiff | tree |
| 2008-10-17 |
Evan Cheng | getX86RegNum has long been moved to X86RegisterInfo.
|
commit | commitdiff | tree |
| 2008-10-17 |
Evan Cheng | Fix a very subtle spiller bug: UpdateKills should not...
|
commit | commitdiff | tree |
| 2008-10-15 |
Evan Cheng | - Add target lowering hooks that specify which setcc...
|
commit | commitdiff | tree |
| 2008-10-14 |
Evan Cheng | Rename LoadX to LoadExt.
|
commit | commitdiff | tree |
| 2008-10-14 |
Evan Cheng | Combine (fcmp cc0 x, y) | (fcmp cc1 x, y) into a single...
|
commit | commitdiff | tree |
| 2008-10-14 |
Evan Cheng | - Somehow I forgot about one / une.
|
commit | commitdiff | tree |
| 2008-10-14 |
Evan Cheng | Fix indentation.
|
commit | commitdiff | tree |
| 2008-10-14 |
Evan Cheng | Optimize anding of two fcmp into a single fcmp if the...
|
commit | commitdiff | tree |
| 2008-10-13 |
Evan Cheng | FIX PR2794. Make sure SIGN_EXTEND_INREG nodes introduced...
|
commit | commitdiff | tree |
| 2008-10-13 |
Evan Cheng | Also update sub-register intervals after a trivial...
|
commit | commitdiff | tree |
| 2008-10-13 |
Evan Cheng | Add a test case for _Complex passed as a FCA.
|
commit | commitdiff | tree |
| 2008-10-13 |
Evan Cheng | Clarify meaning of copyRegToReg's return value.
|
commit | commitdiff | tree |
| 2008-10-06 |
Evan Cheng | Fix PR2850 and PR2863. Only generate movddup for 128...
|
commit | commitdiff | tree |
| 2008-10-06 |
Evan Cheng | Cosmetic.
|
commit | commitdiff | tree |
| 2008-10-06 |
Evan Cheng | This is an objective-c test, not an objective-c++ one.
|
commit | commitdiff | tree |
| 2008-10-03 |
Evan Cheng | New test case.
|
commit | commitdiff | tree |
| 2008-10-03 |
Evan Cheng | Fix typos pointed out by Duncan. Also untabify these...
|
commit | commitdiff | tree |
| 2008-10-02 |
Evan Cheng | A Partitioned Boolean Quadratic Programming (PBQP)...
|
commit | commitdiff | tree |
| 2008-09-30 |
Evan Cheng | Add runStaticConstructorsDestructors which runs ctors...
|
commit | commitdiff | tree |
| 2008-09-30 |
Evan Cheng | Re-apply 56835 along with header file changes.
|
commit | commitdiff | tree |
| 2008-09-30 |
Evan Cheng | If a re-materializable instruction has a register operand...
|
commit | commitdiff | tree |
| 2008-09-29 |
Evan Cheng | Fix PR2835. Do not change the width of a volatile load.
|
commit | commitdiff | tree |
| 2008-09-27 |
Evan Cheng | Re-apply 56683 with fixes.
|
commit | commitdiff | tree |
| 2008-09-26 |
Evan Cheng | Implement "punpckldq %xmm0, $xmm0" as "pshufd $0x50...
|
commit | commitdiff | tree |
| 2008-09-26 |
Evan Cheng | unpckhps requires sse1, punpckhdq requires sse2.
|
commit | commitdiff | tree |
| 2008-09-26 |
Evan Cheng | Fix @llvm.frameaddress codegen. FP elimination optimization...
|
commit | commitdiff | tree |
| 2008-09-26 |
Evan Cheng | Avoid spilling EBP / RBP twice in the prologue.
|
commit | commitdiff | tree |
| 2008-09-26 |
Evan Cheng | X86 address displacement field must be interpreted...
|
commit | commitdiff | tree |
| 2008-09-25 |
Evan Cheng | Prefer movlhps over punpcklqdq, etc. in more cases.
|
commit | commitdiff | tree |
| 2008-09-25 |
Evan Cheng | With sse3 and when the source is a load or has multiple...
|
commit | commitdiff | tree |
| 2008-09-25 |
Evan Cheng | Temporarily backing out 56585:56589 to unbreak the...
|
commit | commitdiff | tree |
| 2008-09-25 |
Evan Cheng | pmovsxbq etc. requires sse4.1.
|
commit | commitdiff | tree |
| 2008-09-25 |
Evan Cheng | <rdar://problem/6234798> Assertion failed: (!OpInfo...
|
commit | commitdiff | tree |
| 2008-09-24 |
Evan Cheng | Fix patterns for SSE4.1 move and sign extend instructions...
|
commit | commitdiff | tree |
| 2008-09-24 |
Evan Cheng | X86ISD::VZEXT_LOAD should produce and fold a chain.
|
commit | commitdiff | tree |
| 2008-09-24 |
Evan Cheng | Add DisableGVCompilation which forces the JIT to assert...
|
commit | commitdiff | tree |
| 2008-09-24 |
Evan Cheng | Commit CodeGenPrepare.cpp changes which was accidentially...
|
commit | commitdiff | tree |
| 2008-09-24 |
Evan Cheng | Test fp not.
|
commit | commitdiff | tree |
| 2008-09-24 |
Evan Cheng | Properly handle 'm' inline asm constraints. If a GV...
|
commit | commitdiff | tree |
| 2008-09-22 |
Evan Cheng | Support x86 specific inline asm modifier 'J'.
|
commit | commitdiff | tree |
| 2008-09-22 |
Evan Cheng | Livestacks really does preserve everything.
|
commit | commitdiff | tree |
| 2008-09-22 |
Evan Cheng | Instead of setPreservesAll, just mark them preseving...
|
commit | commitdiff | tree |
| 2008-09-22 |
Evan Cheng | Test case for fp logical instructions.
|
commit | commitdiff | tree |
| 2008-09-22 |
Evan Cheng | Mark several codegen passes as preserving all analysis.
|
commit | commitdiff | tree |
| 2008-09-22 |
Evan Cheng | Per review feedback: Only perform
|
commit | commitdiff | tree |
| 2008-09-20 |
Evan Cheng | Fix PR2808. When regalloc runs out of register, it...
|
commit | commitdiff | tree |
| 2008-09-20 |
Evan Cheng | Clean up the test.
|
commit | commitdiff | tree |
| 2008-09-20 |
Evan Cheng | No need to print function stubs for Mac OS X 10.5 and...
|
commit | commitdiff | tree |
| 2008-09-20 |
Evan Cheng | 80 column violation.
|
commit | commitdiff | tree |
| 2008-09-19 |
Evan Cheng | Continue after removing the current MI.
|
commit | commitdiff | tree |
| 2008-09-19 |
Evan Cheng | Re-materalized definition instructions may be dead...
|
commit | commitdiff | tree |
| 2008-09-18 |
Evan Cheng | Somehow RegAllocLinearScan is keeping two pointers...
|
commit | commitdiff | tree |
| 2008-09-18 |
Evan Cheng | Preliminary support for systems which require changing...
|
commit | commitdiff | tree |
| 2008-09-18 |
Evan Cheng | Duh. Default to ARMCC::AL (always).
|
commit | commitdiff | tree |
| 2008-09-18 |
Evan Cheng | Clean up.
|
commit | commitdiff | tree |
| next |
projects / oota-llvm.git / search