summaryrefslogtreecommitdiff
path: root/lib
AgeCommit message (Expand)AuthorFilesLines
2009-09-16Big change #1 for personality function references:Chris Lattner9-36/+149
2009-09-16remove a dead variable.Chris Lattner2-3/+0
2009-09-16add a helper method for creating MCSymbol and MCSymbolRefExpr atChris Lattner1-0/+6
2009-09-16Added a variety of floating-point and SSE instructions.Sean Callanan2-0/+51
2009-09-16inline AsmPrinter::getCurrentFunctionEHName into its only caller.Chris Lattner2-11/+6
2009-09-16Expand some more vector operations not supported by Neon.Bob Wilson1-0/+6
2009-09-16Eliminate AsmPrinter::EmitExternalGlobal, inlining its (now)Chris Lattner2-8/+1
2009-09-16Neon does not support vector divide or remainder. Expand them.Bob Wilson1-0/+8
2009-09-16eliminate the PPC backend's implementation of EmitExternalGlobalChris Lattner2-15/+3
2009-09-16eliminate the horrid AsmPrinter::getGlobalLinkName method, inliningChris Lattner3-15/+4
2009-09-16simplify some codeChris Lattner1-6/+5
2009-09-15Expand all v2f64 arithmetic operations for Neon.Bob Wilson1-0/+27
2009-09-15Added far return instructions (that is, returns to Sean Callanan1-0/+4
2009-09-15remove some horrible MAI hooks which fortunately turn out to be always empty.Chris Lattner2-26/+3
2009-09-15strength reduce a call to PrintRelDirective(true).Chris Lattner1-1/+3
2009-09-15add hooks to hang target-specific goop off MachineModuleInfo,Chris Lattner2-7/+10
2009-09-15Do not add the SVOffset to the Node CSE ID. The same pointer argument cannot...Nate Begeman1-5/+0
2009-09-15Expand on comment.Eric Christopher1-1/+4
2009-09-15Updated comments per Eli's suggestion.Sean Callanan2-2/+4
2009-09-15Added register-to-register ADD instructions to theSean Callanan2-0/+13
2009-09-15Drop the raw_ostream required buffer size to 1.Daniel Dunbar1-2/+2
2009-09-15Fix -Asserts warning.Daniel Dunbar1-1/+2
2009-09-15Better solution for tracking both the original alignment of the access, and t...Nate Begeman3-44/+33
2009-09-15Added a new register class for segment registersSean Callanan3-0/+25
2009-09-15Change the marker byte for stubs from 0xcd to 0xce (another form ofDale Johannesen1-3/+6
2009-09-15fix PR4984 by ensuring that fastisel adds properly sign extended GEP displace...Chris Lattner1-1/+1
2009-09-15add missing fileChris Lattner1-0/+32
2009-09-15Handle AddrMode4 for Thumb2 in rewriteT2FrameIndex. This occurs forBob Wilson1-0/+5
2009-09-15Fix superreg use in ARMAsmPrinter. Approved by Anton Korobeynikov.Sandeep Patel1-2/+2
2009-09-15several major improvements to the sparc backend: support for weak linkageChris Lattner8-48/+197
2009-09-15Teach ValueTracking how to look through GlobalAliases. GlobalAliases areDan Gohman1-0/+12
2009-09-15Fix an accidental inversion of the inbounds flag.Dan Gohman1-3/+3
2009-09-15When a constant's type is refined, update the constant in placeDan Gohman3-215/+157
2009-09-15Add more newlines to make up for the ones removed from the end of instructions.Nick Lewycky1-9/+12
2009-09-15Another try at early partial coalescing. Identity phi source copies (their so...Evan Cheng1-58/+56
2009-09-15Forbid arrays of function-type and structures with function-typed fields.Nick Lewycky1-17/+13
2009-09-15fix PR4963: folding insertvalue would sometimes turn a packed struct intoChris Lattner1-20/+24
2009-09-15add a new CallGraphNode::replaceCallEdge method and use it fromChris Lattner2-4/+21
2009-09-15add newline to debug dumpChris Lattner1-1/+1
2009-09-15make -debug-pass=Executions show information about what call graph nodesChris Lattner2-5/+27
2009-09-15add some missing quotes in debug outputChris Lattner1-2/+2
2009-09-15Remove invalid add_dependencies line to unbreak the CMake build.Ted Kremenek1-1/+1
2009-09-15Get rid of GetProcessId in Win32/Program.inc.Mikhail Glushenkov2-43/+22
2009-09-15Don't pull a load through a callseq_start if the load's chainDan Gohman1-1/+2
2009-09-15Remove incorrect CSE code from r81813.Nate Begeman1-2/+0
2009-09-15Modified the Intel instruction tables to includeSean Callanan2-4/+37
2009-09-15Added the first bits of the ARM target assembler to llvm-mc. For now it onlyKevin Enderby4-1/+115
2009-09-15Substantially speed up combiner-aa in the following ways:Nate Begeman1-33/+69
2009-09-15Teach the legalizer to propagate the original alignment of loads and store whenNate Begeman1-6/+11
2009-09-15On x86-64, the 32-bit cmov doesn't actually clear the high 32-bit ofDan Gohman1-3/+5