Relax an invariant that block placement was trying to assert a bit
[oota-llvm.git] / lib / CodeGen /
2011-11-23 Chandler CarruthRelax an invariant that block placement was trying...
2011-11-23 Chandler CarruthHandle the case of a no-return invoke correctly. It...
2011-11-23 Bob WilsonEnable stack protectors for all arrays, not just char...
2011-11-23 Jakob Stoklund OlesenFix PR11422.
2011-11-23 Chandler CarruthFix a crash in block placement due to an inner loop...
2011-11-22 Chandler CarruthFix a devilish miscompile exposed by block placement...
2011-11-22 Chandler CarruthFix an obvious omission in the SelectionDAGBuilder...
2011-11-22 Rafael EspindolaIf a register is both an early clobber and part of...
2011-11-20 Chandler CarruthThe logic for breaking the CFG in the presence of hot...
2011-11-19 Chandler CarruthMove the handling of unanalyzable branches out of the...
2011-11-17 Devang PatelDISubrange supports unsigned lower/upper array bounds...
2011-11-17 Chad RosierWhen fast iseling a GEP, accumulate the offset rather...
2011-11-16 Eli FriedmanMake sure to replace the chain properly when DAGCombini...
2011-11-16 Chad RosierAdd fast-isel stats to determine who's doing all the...
2011-11-16 Chad RosierFix the stats collection for fast-isel. The failed...
2011-11-16 Evan ChengDisable expensive two-address optimizations at -O0...
2011-11-16 Evan ChengDisable the assertion again. Looks like fastisel is...
2011-11-16 Evan ChengSink codegen optimization level into MCCodeGenInfo...
2011-11-16 Bob WilsonRecord landing pads with a SmallSetVector to avoid...
2011-11-16 Bob WilsonUpdate the SP in the SjLj jmpbuf whenever it changes...
2011-11-16 Evan ChengRevert r144568 now that r144730 has fixed the fast...
2011-11-16 Evan ChengIf the 2addr instruction has other kills, don't move...
2011-11-16 Evan ChengRescheduleKillAboveMI() must backtrack to before the...
2011-11-16 Evan ChengProcess all uses first before defs to accurately captur...
2011-11-16 Eli FriedmanCONCAT_VECTORS can have more than two operands. PR11389.
2011-11-16 Eli FriedmanAdd a couple asserts so it will be easier to debug...
2011-11-16 Owen AndersonRename MVT::untyped to MVT::Untyped to match similar...
2011-11-15 Eric ChristopherStabilize the output of the dwarf accelerator tables...
2011-11-15 Chad RosierGEPs with all zero indices are trivially coalesced...
2011-11-15 Pete CooperAdded custom lowering for load->dec->store sequence...
2011-11-15 Devang PatelInsert modified DBG_VALUE into LiveDbgValueMap.
2011-11-15 Rafael EspindolaWe currently use a callback to handle an IL pass deleti...
2011-11-15 Benjamin KramerRemove all remaining uses of Value::getNameStr().
2011-11-15 Benjamin KramerTwinify GraphWriter a little bit.
2011-11-15 Jakob Stoklund OlesenCheck all overlaps when looking for used registers.
2011-11-15 Jay FoadMake use of MachinePointerInfo::getFixedStack.
2011-11-15 Jay FoadRemove some unnecessary includes of PseudoSourceValue.h.
2011-11-15 Evan ChengSet SeenStore to true to prevent loads from being moved...
2011-11-15 Chandler CarruthRather than trying to use the loop block sequence ...
2011-11-15 Jakob Stoklund OlesenBreak false dependencies before partial register updates.
2011-11-15 Jakob Stoklund OlesenTrack register ages more accurately.
2011-11-14 Evan ChengAvoid dereferencing off the beginning of lists.
2011-11-14 Evan ChengAt -O0, multiple uses of a virtual registers in the...
2011-11-14 Evan ChengTeach two-address pass to re-schedule two-address instr...
2011-11-14 Jakob Stoklund OlesenFix early-clobber handling in shrinkToUses.
2011-11-14 Chandler CarruthIt helps to deallocate memory as well as allocate it...
2011-11-14 Chandler CarruthRemove an over-eager assert that was firing on one...
2011-11-14 Chandler CarruthBegin chipping away at one of the biggest quadratic...
2011-11-14 Chandler CarruthUnder the hood, MBPI is doing a linear scan of every...
2011-11-14 Chandler CarruthReuse the logic in getEdgeProbability within getHotSucc...
2011-11-14 Chandler CarruthFix an overflow bug in MachineBranchProbabilityInfo...
2011-11-14 Jakob Stoklund OlesenUse getVNInfoBefore() when it makes sense.
2011-11-14 Chandler CarruthTeach machine block placement to cope with unnatural...
2011-11-13 Jakob Stoklund OlesenUse kill slots instead of the previous slot in shrinkTo...
2011-11-13 Chandler CarruthCleanup some 80-columns violations and poor formatting...
2011-11-13 Jakob Stoklund OlesenTerminate all dead defs at the dead slot instead of...
2011-11-13 Jakob Stoklund OlesenSimplify early clobber slots a bit.
2011-11-13 Chandler CarruthEnhance the assertion mechanisms in place to make it...
2011-11-13 Jakob Stoklund OlesenRename SlotIndexes to match how they are used.
2011-11-13 Chandler CarruthTeach MBP to force-merge layout successors for blocks...
2011-11-13 Chandler CarruthHoist another gross nested loop into a helper method.
2011-11-13 Chandler CarruthAdd a missing doxygen comment for a helper method.
2011-11-13 Chandler CarruthHoist a nested loop into its own method.
2011-11-13 Chandler CarruthRewrite #3 of machine block placement. This is based...
2011-11-13 NAKAMURA TakumiPrune more RALinScan. RALinScan was also here!
2011-11-13 Jakob Stoklund OlesenMore dead code elimination in VirtRegMap.
2011-11-13 Jakob Stoklund OlesenStop tracking spill slot uses in VirtRegMap.
2011-11-13 Jakob Stoklund OlesenRemove dead code and data from VirtRegMap.
2011-11-13 Jakob Stoklund OlesenStop tracking unused registers in VirtRegMap.
2011-11-13 Jakob Stoklund OlesenRemove the -color-ss-with-regs option.
2011-11-13 Jakob Stoklund OlesenDelete VirtRegRewriter.
2011-11-13 Jakob Stoklund OlesenSwitch PBQP to VRM's trivial rewriter.
2011-11-12 Jakob Stoklund OlesenDelete the old spilling framework from LiveIntervalAnal...
2011-11-12 Jakob Stoklund OlesenDelete the 'standard' spiller with used the old spillin...
2011-11-12 Jakob Stoklund OlesenSwitch PBQP to the modern InlineSpiller framework.
2011-11-12 Jakob Stoklund OlesenDelete the linear scan register allocator.
2011-11-12 Rafael EspindolaThe dwarf standard says that the only differences betwe...
2011-11-12 Eli FriedmanDon't try to form pre/post-indexed loads/stores until...
2011-11-11 Eli FriedmanSome cleanup and bulletproofing for node replacement...
2011-11-11 Nicolas GeoffrayAdd a custom safepoint method, in order for language...
2011-11-11 Eric ChristopherInitialize variable.
2011-11-11 Eric ChristopherIf we have a DIE with an AT_specification use that...
2011-11-10 Rafael EspindolaCheck in getOrCreateSubprogramDIE if a declaration...
2011-11-10 Eric ChristopherMake types and namespaces take multiple DIEs for the...
2011-11-10 Eric ChristopherMove type handling to make sure we get all created...
2011-11-10 Eric ChristopherRework adding function names to the dwarf accelerator...
2011-11-10 Evan ChengUse a bigger hammer to fix PR11314 by disabling the...
2011-11-10 Jakob Stoklund OlesenStrip old implicit operands after foldMemoryOperand.
2011-11-09 Eli FriedmanAdd check so we don't try to perform an impossible...
2011-11-09 Benjamin KramerAdd comments.
2011-11-09 Duncan SandsSpeculatively revert commit 144124 (djg) in the hope...
2011-11-09 Benjamin KramerTake advantage of the zero byte in StringMap when emitt...
2011-11-09 Devang PatelRemove extra ';'
2011-11-09 Eric ChristopherRemove the pubnames section, no one consumes it.
2011-11-09 Jakob Stoklund OlesenCollapse DomainValues across loop back-edges.
2011-11-09 Jakob Stoklund OlesenLink to the live DomainValue after merging.
2011-11-08 Jakob Stoklund OlesenTrack reference count independently from clear().
2011-11-08 Jakob Stoklund OlesenCall release() directly when cleaning up the remaining...
2011-11-08 Jakob Stoklund OlesenRename all methods to follow style guide.
2011-11-08 Jakob Stoklund OlesenHandle reference counts in one function: release().
next