aboutsummaryrefslogtreecommitdiff
path: root/lib/CodeGen/SelectionDAG
AgeCommit message (Expand)Author
2005-08-31Sigh, not my day. Fix typo.Nate Begeman
2005-08-31Fix a mistake in my previous patch pointed out by sabre; the AssertZextNate Begeman
2005-08-31Remove some unnecessary casts, and add the AssertZext case toNate Begeman
2005-08-30Allow physregs to occur in the dag with multiple types. Though I don't liket...Chris Lattner
2005-08-30Fix FreeBench/fourinarow with the dag isel, by not adding a bogus resultChris Lattner
2005-08-30Fix a miscompile of PtrDist/bc. Sign extending bools is not the right thing,Chris Lattner
2005-08-30Remove a bogus piece of my AssertSext/AssertZext patch. oops.Nate Begeman
2005-08-30Add support for AssertSext and AssertZext, folding other extensions withNate Begeman
2005-08-30Name this variable to be what it really is!Chris Lattner
2005-08-30Handle CopyToReg nodes with flag operands correctlyChris Lattner
2005-08-29Add a hack to avoid some horrible code in some cases by always emittingChris Lattner
2005-08-29Add a new API for NateChris Lattner
2005-08-29Some of us cared about the the promote pathAndrew Lenharth
2005-08-29Fix an infinite loop on x86Chris Lattner
2005-08-28Fix a bug in ReplaceAllUsesWithChris Lattner
2005-08-27Change the names of member variables per Chris' instructions, and documentReid Spencer
2005-08-27Disable this code, which broke many tests last nightChris Lattner
2005-08-27fix PHI node emission for basic blocks that have select_cc's in them on ppc32Chris Lattner
2005-08-26Nate noticed that Andrew never did this. This fixes PR600Chris Lattner
2005-08-26Don't copy regs that are only used in the entry block into a vreg. ThisChris Lattner
2005-08-26Checking types here is not safe, because multiple types can map to the sameChris Lattner
2005-08-26Call the InsertAtEndOfBasicBlock hook if the usesCustomDAGSchedInserterChris Lattner
2005-08-26Revampt ReplaceAllUsesWith to be more efficient and easier to use.Chris Lattner
2005-08-26Change ConstantPoolSDNode to actually hold the Constant itself instead ofChris Lattner
2005-08-26Fix a huge annoyance: SelectNodeTo took types before the opcode unlikeChris Lattner
2005-08-26the 5th operand is the 4th numberChris Lattner
2005-08-26Add support for targets that want to custom expand select_cc in some cases.Chris Lattner
2005-08-26Allow LowerOperation to return a null SDOperand in case it wants to lowerChris Lattner
2005-08-26Fix a nasty bug from a previous patch of mineChris Lattner
2005-08-25New fold for SELECT_CCNate Begeman
2005-08-25Don't auto-cse nodes that return flagsChris Lattner
2005-08-25simplify the code a bit using isOperationLegalChris Lattner
2005-08-25Add support for flag operandsChris Lattner
2005-08-25ADd support for TargetConstantPool nodesChris Lattner
2005-08-25add a new TargetFrameIndex nodeChris Lattner
2005-08-24add a methodChris Lattner
2005-08-24Add ReplaceAllUsesWith that can take a vector of replacement values.Chris Lattner
2005-08-24Add support for external symbols, and support for variable arity instructionsChris Lattner
2005-08-24Fix pasto that prevented VT ndoes from showing up in -view-isel-dags correctlyChris Lattner
2005-08-24teach selection dag mask tracking about the fact that select_cc operates likeChris Lattner
2005-08-24Start using isOperationLegal and isTypeLegal to simplify the codeChris Lattner
2005-08-24Adjust to new interfaceChris Lattner
2005-08-24Teach SelectionDAG how to simplify a few more setcc-equivalent select_ccNate Begeman
2005-08-24Make -view-isel-dags show the dag before instruction selecting, in caseChris Lattner
2005-08-24Fix optimization of select_cc seteq X, 0, 1, 0 -> srl (ctlz X), log2 X sizeNate Begeman
2005-08-23Teach the SelectionDAG how to transform select_cc eq, X, 0, 1, 0 intoNate Begeman
2005-08-23Teach Legalize how to turn setcc into select_ccNate Begeman
2005-08-22Fix a problem where constant expr shifts would not have their shift amountChris Lattner
2005-08-22Add a fast-path for register values. Add support for constant pool entries,Chris Lattner
2005-08-21add anew methodChris Lattner