aboutsummaryrefslogtreecommitdiff
path: root/lib
AgeCommit message (Expand)Author
2010-06-07Make bugpoint dead-argument-hacking actually work, and actually test it.Dan Gohman
2010-06-07The FoldingSet hash data includes pointer values, so it isn'tDan Gohman
2010-06-07Optimize this code somewhat by taking advantage of the factDan Gohman
2010-06-07Another place where the code wanted to access the argument list and not all ofBill Wendling
2010-06-07Micro-optimize this, to speed up this hotspot in debug builds a little.Dan Gohman
2010-06-07Micro-optimize this.Dan Gohman
2010-06-07Move exit check where it really belongs.Jim Grosbach
2010-06-07Optimize ScalarEvolution's SCEVComplexityCompare predicate: don't goDan Gohman
2010-06-07Create new accessors to get arguments for call/invoke instructions. It breaksBill Wendling
2010-06-05Partial specialization was not checking the callsite to make sure it was usin...Kenneth Uildriks
2010-06-05This bug is also present in MSVC10. Requested by Elrood on IRC.Duncan Sands
2010-06-05revert r105521, which is breaking the buildbots with stuff like this:Chris Lattner
2010-06-05Initial AVX support for some instructions. No patterns matchedBruno Cardoso Lopes
2010-06-05Improvements to tail call code. No functional effectDale Johannesen
2010-06-05getFoldedOffsetOf no longer does anything special with vector types.Dan Gohman
2010-06-05Revert 105492 & 105493 due to a testcase regression. Radar 7424645.Stuart Hastings
2010-06-05LSR needs to remember inserted instructions even in postinc mode, becauseDan Gohman
2010-06-05Fix some liveout handling related to tail calls, see comments.Dale Johannesen
2010-06-04Re-apply 105308 with fix.Evan Cheng
2010-06-04Don't track users of undef values; they aren't interesting forDan Gohman
2010-06-04Make if-conversion ignore dbg_value instructions in its analysis. rdar://7797940Jim Grosbach
2010-06-04Support for nested functions/classes in debug output. Radar 7424645.Stuart Hastings
2010-06-04Copy location info for current function argument from dbg.declare if respecti...Devang Patel
2010-06-04Skip dbg_value instructions when scanning instructions in register scavenging.Jim Grosbach
2010-06-04Fix normalization and de-normalization of non-affine SCEVs.Dan Gohman
2010-06-04Keep track of the call instructions whose clobber lists were skipped during fastJakob Stoklund Olesen
2010-06-04More thoroughly disable tails calls by default.Dale Johannesen
2010-06-04Another fix to prevent debug info from affecting codegen. rdar://7797940Jim Grosbach
2010-06-04more dbg_value adjustments so debug info doesn't affect codegenJim Grosbach
2010-06-04Fixed a bug during widening where we would avoid legalizing a node. When weMon P Wang
2010-06-04fix typoJim Grosbach
2010-06-04For NEON vectors with 32- or 64-bit elements, select BUILD_VECTORs andBob Wilson
2010-06-03Add some missing checks in TwoAddressInstructionPass::CoalesceExtSubRegs.Bob Wilson
2010-06-03Cleanup 80-column and trim trailing whitespaceJim Grosbach
2010-06-03Teach the ARM load-store optimizer to deal with dbg_value instructions.Jim Grosbach
2010-06-03Early implementation of tail call for ARM.Dale Johannesen
2010-06-03Fix SimplifyDemandedBits' AssertZext logic to demand all the bits. ItDan Gohman
2010-06-03Revert 105308.Bob Wilson
2010-06-03Machine sink could potentially sink instructions into a block where the physicalBill Wendling
2010-06-03Add first pass at darwin tls compiler support.Eric Christopher
2010-06-03One underscore, not two.Eric Christopher
2010-06-03Implement expansion in type legalization for add/sub with overflow. TheEli Friedman
2010-06-03Remove some already-fixed README entries.Eli Friedman
2010-06-03Remove README entry which no longer compiles to something sane.Eli Friedman
2010-06-03Remove a fixed item, update a couple partially-fixed items.Eli Friedman
2010-06-03Use the fast register allocator by default for -O0 builds.Jakob Stoklund Olesen
2010-06-03Use readsWritesVirtualRegister instead of counting uses and defs when insertingJakob Stoklund Olesen
2010-06-02Add full register <imp-def> operands when the coalescer is creating partialJakob Stoklund Olesen
2010-06-02Speedup bitcode writer. Do not walk all values for all functions to emit func...Devang Patel
2010-06-02Compulsive reformating. No functionalitical changes.Bill Wendling