prepare for a change I'm about to make
[oota-llvm.git] / lib / CodeGen / SelectionDAG /
2006-10-27 Jim LaskeyClean up.
2006-10-27 Chris LattnerFix a bug in merged condition handling (CodeGen/Generic...
2006-10-27 Jim LaskeySwitch over from SelectionNodeCSEMap to FoldingSet.
2006-10-27 Chris Lattnerremove debug code
2006-10-27 Chris LattnerCodegen cond&cond with two branches. This compiles...
2006-10-27 Chris LattnerTurn conditions like x<Y|z==q into multiple blocks.
2006-10-26 Evan ChenggetPreIndexedLoad -> getIndexedLoad.
2006-10-26 Reid SpencerFor PR950:
2006-10-24 Chris LattnervisitSwitchCase knows how to insert conditional branche...
2006-10-24 Chris LattnerGeneralize CaseBlock a bit more:
2006-10-24 Chris Lattnergeneralize 'CaseBlock'. It really allows any compariso...
2006-10-23 Chris LattnerMinor tweak. Instead of generating:
2006-10-22 Chris LattnerFix phi node updating for switches lowered to linear...
2006-10-22 Chris Lattnerdisable this code for now, it's not yet safely updating...
2006-10-22 Chris LattnerImplement PR964 and Regression/CodeGen/Generic/SwitchLo...
2006-10-20 Chris LattnerMake flag and chain edges visually distinguishable...
2006-10-20 Reid SpencerFor PR950:
2006-10-19 Bill WendlingPartially in response to PR926: insert the newly create...
2006-10-18 Jim LaskeyAdd option for controlling inclusion of global AA.
2006-10-18 Jim LaskeyUse global info for alias analysis.
2006-10-17 Chris LattnerTrivial patch to speed up legalizing common i64 constants.
2006-10-17 Chris LattnerFix CodeGen/PowerPC/2006-10-17-brcc-miscompile.ll
2006-10-17 Evan ChengFix printer for StoreSDNode.
2006-10-17 Evan ChengReflect MemOpAddrMode change; added a helper to create...
2006-10-17 Jim LaskeyMake it simplier to dump DAGs while in DAGCombiner...
2006-10-17 Evan ChengMake sure operand does have size and element type operands.
2006-10-16 Evan ChengBe careful when looking through a vbit_convert. Optimiz...
2006-10-16 Jim LaskeyPass AliasAnalysis thru to DAGCombiner.
2006-10-14 Jim LaskeyTidy up after truncstore changes.
2006-10-14 Evan ChengDebug tweak.
2006-10-14 Chris LattnerMake sure that the node returned by SimplifySetCC is...
2006-10-14 Chris Lattnerfold setcc of a setcc.
2006-10-14 Chris LattnerWhen SimplifySetCC was moved to the DAGCombiner, it...
2006-10-13 Jim LaskeyReduce the workload by not adding chain users to work...
2006-10-13 Chris LattnerFix a bug where we incorrectly turned '(X & 0) == 0...
2006-10-13 Evan ChengMerge ISD::TRUNCSTORE to ISD::STORE. Switch to using...
2006-10-12 Chris LattnerLower X%C into X/C+stuff. This allows the 'division...
2006-10-12 Evan ChengAdd RemoveDeadNode to remove a dead node and its (poten...
2006-10-12 Chris Lattneradd a minor dag combine noticed when looking at PR945
2006-10-12 Jim LaskeyD'oh - need to use the rigth kind of store.
2006-10-11 Jim LaskeyAlias analysis of TRUNCSTORE.
2006-10-11 Jim LaskeyTypo
2006-10-11 Jim LaskeyHandle aliasing of loadext.
2006-10-11 Jim LaskeyFix regression in combiner alias analysis.
2006-10-11 Evan ChengNaming consistency.
2006-10-11 Andrew LenharthJimptables working again on alpha.
2006-10-11 Chris Lattneradd two helper methods.
2006-10-11 Evan ChengFindModifiedNodeSlot needs to add LoadSDNode ivars...
2006-10-10 Evan ChengAlso update getNodeLabel for LoadSDNode.
2006-10-10 Evan ChengSDNode::dump should also print out extension type and VT.
2006-10-10 Chris LattnerFix another bug in extload promotion.
2006-10-10 Evan ChengFix a bug introduced by my LOAD/LOADX changes.
2006-10-09 Evan ChengReflects ISD::LOAD / ISD::LOADX / LoadSDNode changes.
2006-10-08 Chris LattnerEliminate more token factors by taking advantage of...
2006-10-07 Jim LaskeyCombiner alias analysis passes Multisource (release...
2006-10-07 Chris LattnerFix a bug legalizing zero-extending i64 loads into...
2006-10-06 Chris Lattnerinitialize ivar
2006-10-06 Chris Lattnerjump tables handle pic
2006-10-06 Chris LattnerFix a miscompilation of:
2006-10-05 Evan ChengMake use of getStore().
2006-10-05 Evan ChengAdd getStore() helper function to create ISD::STORE...
2006-10-05 Jim LaskeyAlias analysis code clean ups.
2006-10-04 Evan ChengFix some typos that can cause a flag value to have...
2006-10-04 Jim LaskeyMore extensive alias analysis.
2006-10-04 Evan ChengCombine ISD::EXTLOAD, ISD::SEXTLOAD, ISD::ZEXTLOAD...
2006-10-03 Evan ChengFix an obvious typo.
2006-10-02 Jim LaskeyDebugging kruft
2006-10-02 Jim LaskeyAdd ability to annotate (color) nodes in a viewGraph.
2006-09-28 Chris Lattnerrefactor critical edge breaking out into the SplitCritE...
2006-09-26 Andrew LenharthComments on JumpTableness
2006-09-26 Jim LaskeyLoad chain check is not needed
2006-09-26 Jim LaskeyChain can be any operand
2006-09-26 Jim LaskeyWrong size for load
2006-09-26 Jim LaskeyCan't move a load node if it's chain is not used.
2006-09-25 Jim LaskeyAccidental enable of bad code
2006-09-25 Jim LaskeyFix chain dropping in load and drop unused stores in...
2006-09-25 Jim LaskeyCore antialiasing for load and store.
2006-09-24 Andrew LenharthAdd support for other relocation bases to jump tables...
2006-09-24 Evan ChengPIC jump table entries are always 32-bit. This fixes...
2006-09-21 Evan ChengMake it work for DAG combine of multi-value nodes.
2006-09-21 Jim Laskeycore corrections
2006-09-21 Jim LaskeyBasic "in frame" alias analysis.
2006-09-21 Chris Lattnerfold (aext (and (trunc x), cst)) -> (and x, cst).
2006-09-21 Chris LattnerCheck the right value type. This fixes 186.crafty...
2006-09-21 Chris LattnerCompile:
2006-09-21 Chris LattnerGeneralize (zext (truncate x)) and (sext (truncate...
2006-09-20 Chris LattnerCompile:
2006-09-20 Chris LattnerFold the full generality of (any_extend (truncate x))
2006-09-20 Chris LattnerTwo things:
2006-09-20 Chris LattnerExpand 64-bit shifts more optimally if we know that...
2006-09-19 Chris LattnerFold extract_element(cst) to cst
2006-09-19 Chris LattnerMinor speedup for legalize by avoiding some malloc...
2006-09-18 Evan ChengFix a typo.
2006-09-18 Evan ChengAllow i32 UDIV, SDIV, UREM, SREM to be expanded into...
2006-09-18 Andrew Lenharthabsolute addresses must match pointer size
2006-09-16 Chris LattnerOh yeah, this is needed too
2006-09-16 Chris Lattnersimplify control flow, no functionality change
2006-09-16 Chris LattnerAllow custom expand of mul
2006-09-14 Chris LattnerFold (X & C1) | (Y & C2) -> (X|Y) & C3 when possible.
2006-09-14 Chris LattnerSplit rotate matching code out to its own function...
next