64-bit (MMX) vectors do not need restrictive alignment.
[oota-llvm.git] / lib / Target / X86 / X86ISelLowering.cpp
2008-02-08 Dale Johannesen64-bit (MMX) vectors do not need restrictive alignment.
2008-02-08 Dan GohmanAvoid needlessly casting away const qualifiers.
2008-02-07 Dan GohmanFollow Chris' suggestion; change the PseudoSourceValue...
2008-02-07 Dan GohmanAdd SourceValue information for outgoing argument store...
2008-02-06 Dan GohmanRe-apply the memory operand changes, with a fix for...
2008-02-05 Dale JohannesenImplement sseregparm.
2008-02-02 Nick LewyckyDon't use uninitialized values. Fixes vec_align.ll...
2008-02-02 Evan ChengSDIsel processes llvm.dbg.declare by recording the...
2008-01-31 Evan ChengRevert 46556 and 46585. Dan please fix the PseudoSource...
2008-01-31 Dan GohmanAvoid unnecessarily casting away const.
2008-01-31 Dan GohmanRename ISD::FLT_ROUNDS to ISD::FLT_ROUNDS_ to avoid...
2008-01-31 Dan GohmanCreate a new class, MemOperand, for describing memory...
2008-01-30 Evan ChengEven though InsertAtEndOfBasicBlock is an ugly hack...
2008-01-29 Evan ChengWork in progress. This patch *fixes* x86-64 calls which...
2008-01-29 Dale JohannesenHandle 'X' constraint in asm's better.
2008-01-27 Chris LattnerUse fldz and fld1 for long double constants instead...
2008-01-26 Chris LattnerRemove some code for inferring alignment info from...
2008-01-25 Chris Lattneroptimize fxor like for
2008-01-25 Chris LattnerAdd target-specific dag combines for FAND(x,0) and...
2008-01-24 Chris LattnerSignificantly simplify and improve handling of FP funct...
2008-01-23 Evan ChengLet each target decide byval alignment. For X86, it...
2008-01-23 Duncan SandsThe last pieces needed for loading arbitrary
2008-01-17 Chris LattnerThis commit changes:
2008-01-17 Chris Lattner* Introduce a new SelectionDAG::getIntPtrConstant method
2008-01-16 Duncan SandsTrampoline support for x86-64. This looks like
2008-01-16 Chris Lattnermake it more clear that this predicate only applies...
2008-01-16 Chris Lattnerintroduce a isTypeInSSEReg predicate, which allows...
2008-01-16 Chris LattnerMy previous commit had an incomplete message, it should...
2008-01-16 Chris Lattnermake the 'fp return in ST(0)' optimization smart enough to
2008-01-16 Chris Lattnervarious whitespace cleanups, no functionality change.
2008-01-15 Chris Lattnerno need to expand ISD::TRAP to X86ISD::TRAP, just match...
2008-01-15 Anton KorobeynikovFor PR1839: add initial support for __builtin_trap...
2008-01-13 Duncan SandsWhitespace tweak.
2008-01-12 Evan ChengCode clean up.
2008-01-11 Arnold Schwaighoferhrm - correct spelling.
2008-01-11 Arnold SchwaighoferImprove tail call optimized call's argument lowering...
2008-01-11 Arnold SchwaighoferCorrect a copy and paste error.
2008-01-10 Evan ChengMark byval parameter stack objects mutable for now.
2008-01-10 Evan ChengDo not use the stack pointer directly, issue a copyfrom...
2008-01-10 Evan ChengRemove comments that do not correspond to anything...
2008-01-07 Evan ChengUnbreak x86-64.
2008-01-05 Nate BegemanRemove an incorrect optimization that is performed...
2008-01-05 Gordon HenriksenRefactoring the x86 and x86-64 calling convention imple...
2008-01-03 Gordon HenriksenFirst steps in in X86 calling convention cleanup.
2007-12-31 Chris LattnerRename SSARegMap -> MachineRegisterInfo in keeping...
2007-12-30 Chris LattnerAdd new shorter predicates for testing machine operands...
2007-12-29 Chris LattnerRemove attribution from file headers, per discussion...
2007-12-29 Chris LattnerCodegen:
2007-12-29 Chris Lattneravoid going through a stack slot to convert from fpstac...
2007-12-16 Chris Lattnerfix a questionable cast, thanks to Mike Stump for point...
2007-12-15 Evan ChengMake better use of instructions that clear high bits...
2007-12-14 Evan ChengFix ctlz and cttz. llvm definition requires them to...
2007-12-14 Evan ChengImplement ctlz and cttz with bsr and bsf.
2007-12-12 Dan GohmanAllow vector integer constants to be created with
2007-12-12 Evan ChengUse shuffles to implement insert_vector_elt for i32...
2007-12-12 Evan ChengLower a build_vector with all constants into a constpoo...
2007-12-11 Evan Cheng- Improved v8i16 shuffle lowering. It now uses pshuflw...
2007-12-11 Nate Begemanx86 doesn't actually want to custom lower v3i32
2007-12-07 Evan ChengAdd comment.
2007-12-07 Evan ChengMuch improved v8i16 shuffles. (Step 1).
2007-12-06 Evan ChengRemove a bogus optimization. It's not possible to do...
2007-11-27 Duncan SandsFix PR1146: parameter attributes are longer part of
2007-11-25 Chris LattnerFix a long standing deficiency in the X86 backend:...
2007-11-24 Chris Lattnerremove bogus assertion that broke CodeGen/Generic/cast...
2007-11-24 Chris LattnerSeveral changes:
2007-11-16 Anton KorobeynikovImplement codegen for flt_rounds on x86
2007-11-13 Bill WendlingUnify CALLSEQ_{START,END}. They take 4 parameters:...
2007-11-10 Arnold SchwaighoferUpdate tailcall code to include inline attribute operan...
2007-11-09 Evan ChengMuch improved pic jumptable codegen:
2007-11-05 Rafael EspindolaMove the LowerMEMCPY and LowerMEMCPYCall to a common...
2007-11-04 Chris LattnerFix PR1763 by allowing the 'q' constraint to work with...
2007-11-02 Evan ChengUnbreak tailcall opt.
2007-11-02 Evan ChengMissing a getNumOperands check.
2007-10-31 Rafael EspindolaMake ARM and X86 LowerMEMCPY identical by moving the...
2007-10-31 Rafael EspindolaMake ARM an X86 memcpy expansion more similar to each...
2007-10-31 Dale JohannesenMake i64=expand_vector_elt(v2i64) work in 32-bit mode.
2007-10-30 Dale JohannesenAdd missing MMX PSUBQ.
2007-10-29 Evan ChengEnable more fold (sext (load x)) -> (sext (truncate...
2007-10-29 Evan ChengAvoid doing something dumb like rewriting using a 64...
2007-10-26 Evan ChengLoosen up iv reuse to allow reuse of the same stride...
2007-10-21 Dale JohannesenAllow for copysign having f80 second argument.
2007-10-19 Rafael EspindolaAdd support for byval function whose argument is not...
2007-10-17 Chris LattnerChange fp to sint legalization on x86-32 to do 2 x i32
2007-10-17 Chris Lattnerfix some funny indentation, add comments.
2007-10-16 Dale JohannesenCheck for invalid cc's in f80 select.
2007-10-16 Arnold SchwaighoferCorrection to tail call optimization code. The new...
2007-10-15 Evan ChengLowerFP_TO_SINT must not create a stack object if it...
2007-10-14 Evan ChengUnbreak x86-64.
2007-10-12 Arnold SchwaighoferCorrecting the corrections. Bad bad baaad emacs!
2007-10-12 Arnold SchwaighoferCorrected many typing errors. And removed 'nest' parame...
2007-10-12 Duncan SandsDue to the new tail call optimization, trampolines...
2007-10-12 Dan GohmanMark vector ctpop, cttz, and ctlz as Expand on x86.
2007-10-11 Dan GohmanSet ISD::FPOW to Expand.
2007-10-11 Arnold SchwaighoferAdded tail call optimization to the x86 back end. It...
2007-10-08 Evan ChengBug fix. X86 was emitting redundant setcc and test...
2007-10-08 Dan GohmanMigrate X86 and ARM from using X86ISD::{,I}DIV and...
2007-10-05 Evan ChengNot needed any more.
2007-09-29 Evan ChengEnabling new condition code modeling scheme.
2007-09-28 Rafael EspindolaRefactor the memcpy lowering for the x86 target.
2007-09-26 Dale JohannesenEnable codegen for long double abs, sin, cos
next