oota-llvm.git
2005-04-09 Chris LattnerSwitch this instruction selector over to using liveins...
2005-04-09 Chris Lattnerthere is no need to remove this instruction, linscan...
2005-04-09 Chris LattnerAdjust live intervals to support a livein set
2005-04-09 Chris LattnerUse live out sets for return values instead of imp_defs...
2005-04-09 Chris LattnerConsider the livein/out set for a function, allowing...
2005-04-09 Chris Lattneradd routines to track the livein/out set for a function
2005-04-09 Duraid Madinaok, the "ia64 has a boatload of registers" joke stopped...
2005-04-09 Nate BegemanOptimize FSEL a bit for fneg arguments. This fixes...
2005-04-09 Nate BegemanAdd a testcase to make sure that we don't emit two...
2005-04-09 Nate BegemanAdd cases to cover the rest of the patterns we should...
2005-04-09 Chris LattnerFix CodeGen/SparcV9/2005-05-09-GEP-Crash.ll a crash...
2005-04-09 Chris LattnerNew testcase that the sparc backend crashes on
2005-04-09 Chris Lattnerrecognize some patterns as fabs operations, so that...
2005-04-09 Chris Lattnermake this test more interesting
2005-04-09 Chris Lattneradd a test for fnabs
2005-04-09 Chris Lattneradd a partial test for the fma operations that ppc...
2005-04-09 Chris LattnerEmit BRCONDTWOWAY when possible.
2005-04-09 Chris LattnerLegalize BRCONDTWOWAY into a BRCOND/BR pair if a target...
2005-04-09 Chris Lattnerprint and fold BRCONDTWOWAY correctly
2005-04-09 Chris LattnerThis target does not support/want ISD::BRCONDTWOWAY
2005-04-09 Chris LattnerThis target does not yet support ISD::BRCONDTWOWAY
2005-04-09 Chris LattnerAdd a new node
2005-04-09 Nate Begeman64b: Expand S/UREM
2005-04-09 Chris Lattnercanonicalize a bunch of operations involving fneg
2005-04-08 Nate BegemanFix 64b shifts
2005-04-08 Chris Lattnerfix this method for 64-bit constants
2005-04-08 Nate BegemanMatch Mac OS X 64 bit calling conventions
2005-04-08 Andrew Lenharthcollect a few statistics, factor constants (constant...
2005-04-08 Andrew Lenharthoops
2005-04-08 Andrew Lenharthadded some tests to check stupid pattern matching mistakes
2005-04-08 Duraid Madinafix bogus division-by-power-of-2 (was wrong for negativ...
2005-04-08 Chris LattnerFix bug: InstCombine/2005-05-07-UDivSelectCrash.ll
2005-04-08 Chris Lattnernew testcase that crashes the instcombiner.
2005-04-07 Nate BegemanOptimized code sequences for setcc reg, 0
2005-04-07 Andrew LenharthAlpha zero extends setcc results
2005-04-07 Chris LattnerIf a target zero or sign extends the result of its...
2005-04-07 Chris LattnerPowerPC zero extends setcc results
2005-04-07 Chris LattnerX86 zero extends setcc results
2005-04-07 Chris LattnerAllow targets which produce setcc results in non-MVT...
2005-04-07 Chris LattnerRemove somethign I had for testing
2005-04-07 Andrew Lenharthfix a small optimization opertunity and make gcc happy
2005-04-07 Chris LattnerThis patch does two things. First, it canonicalizes...
2005-04-07 Andrew Lenharthmatch case change in codegen
2005-04-07 Andrew Lenharthfixup magic constant making code. tested by thousands...
2005-04-07 Andrew Lenharthlowercase instructions, makes diff happier
2005-04-07 Chris LattnerImplement the following xforms:
2005-04-07 Chris Lattnernew test
2005-04-07 Chris LattnerImplement InstCombine/add.ll:test28, transforming C1...
2005-04-07 Chris Lattnernew testcase
2005-04-07 Chris LattnerTransform X-(X+Y) == -Y and X-(Y+X) == -Y
2005-04-07 Andrew LenharthIt wasn't happy about this either
2005-04-07 Andrew LenharthYea, it wasn't happy
2005-04-07 Duraid Madinateach asmprinter to print s8/s14 operands
2005-04-07 Duraid Madinacodegen immediate forms of add/sub/shift
2005-04-07 Duraid Madinaadd immediate forms of add, sub, shift
2005-04-07 Chris LattnerFix a really scary bug that Nate found where we weren...
2005-04-06 Nate BegemanPattern match bitfield insert, which helps shift long...
2005-04-06 Nate BegemanFix some shift bugs
2005-04-06 Alkis EvlogimenosMake these 64 bit constants so that this compiles on...
2005-04-06 Andrew Lenharthadded sdiv by 2^k and works for neg divisors also
2005-04-06 Chris LattnerDon't make this require loopsimplify. It works BETTER...
2005-04-06 Nate BegemanTeach ExpandShift how to handle shifts by a constant...
2005-04-06 Andrew Lenharthfix copy/paste errors, and add imm support to SxADDQ...
2005-04-06 Chris LattnerFix SingleSource/Regression/C/2005-05-06-LongLongSigned...
2005-04-06 Andrew Lenharthsimplified
2005-04-06 Andrew Lenharthadded first alpha codegen regression test
2005-04-06 Andrew LenharthAdded Nate's div by constant stuff, also scaled operations!
2005-04-06 Chris LattnerFix a namespace issue, reported by Vladimir Merzliakov!
2005-04-06 Duraid Madinasteal sampo's div-by-constant-power-of-2 stuff
2005-04-06 Duraid Madinaadd fms instruction
2005-04-06 Nate BegemanFixed version of optimized integer divide is now fixed...
2005-04-06 Duraid Madinalie a bit and say that r1/r12 (GP/SP) _aren't_ callee...
2005-04-06 Duraid Madinamake sure 'special' registers don't get allocated
2005-04-06 Chris Lattnerdocument these nodes, as they are nonobvious
2005-04-06 Chris LattnerAdd (untested) support for MULHS and MULHU.
2005-04-06 Chris Lattneradd signed versions of the extra precision multiplies
2005-04-06 Nate BegemanTurn off the div -> mul optimization until it works...
2005-04-06 Nate BegemanAdd support for MULHS and MULHU nodes
2005-04-06 Nate BegemanExpand SREM and UREM for targets that claim not to...
2005-04-05 Nate BegemanAdd MULHU and MULHS nodes for the high part of an ...
2005-04-05 Andrew Lenharthadded lowerargs support for varargs
2005-04-05 Nate BegemanBehold, rlwinm with certain immediate arguments is...
2005-04-05 Nate BegemanFix cut & paste errors (32->64), and codegen float...
2005-04-05 Tanya LattnerUpdated to use dep analyzer.
2005-04-05 Nate BegemanRemove 64 bit simple ISel, it never worked correctly
2005-04-05 Nate BegemanBack out the previous change to SelectBranchCC, since...
2005-04-05 Nate BegemanRename canUseAsImmediateForOpcode to getImmediateForOpc...
2005-04-05 Chris Lattnerdo not crash when using -debug
2005-04-05 Nate BegemanImplement SDIV by power of 2 as srawi/addze rather...
2005-04-04 Nate BegemanPattern match fp mul-add, mul-sub, neg-mul-add, and...
2005-04-04 Nate BegemanAdd support for multiply-add, multiply-sub, and their...
2005-04-04 Chris Lattnerdo not dereference an extra layer of pointers to determ...
2005-04-04 Chris Lattnernew testcase
2005-04-04 Nate BegemanMake sure that arg regs used by the call instruction...
2005-04-04 Nate BegemanRemove unnecessary register copy now that regalloc...
2005-04-04 Chris LattnerMake sure to notice that explicit physregs are used...
2005-04-04 Nate Begemani1 loads should also be from the low byte of the argume...
2005-04-04 Nate BegemanFix i64 return, fix CopyFromReg
2005-04-04 Duraid Madinafix SREM/UREM, which gave incorrect results for x%y...
2005-04-04 Duraid Madinaadd implicit use op
next