2009-11-01 |
Chris Lattner | only run GlobalDCE at -O3 and run it late instead of...
|
commit | commitdiff | tree |
2009-11-01 |
Chris Lattner | cleanups, switch GlobalDCE to SmallPtrSet instead of...
|
commit | commitdiff | tree |
2009-11-01 |
Chris Lattner | We currently only run ipsccp at LTO time, which is...
|
commit | commitdiff | tree |
2009-11-01 |
Chris Lattner | remove a bunch of locking from LLVMContextImpl. Since...
|
commit | commitdiff | tree |
2009-11-01 |
Chris Lattner | improve comment.
|
commit | commitdiff | tree |
2009-11-01 |
Chris Lattner | add a comment about why we don't allow inlining indbr.
|
commit | commitdiff | tree |
2009-11-01 |
Chris Lattner | the verifier shouldn't modify the IR.
|
commit | commitdiff | tree |
2009-11-01 |
Chris Lattner | teach ipsccp and ipconstprop that a blockaddress doesn...
|
commit | commitdiff | tree |
2009-11-01 |
Chris Lattner | change llvm::MergeBlockIntoPredecessor to not merge...
|
commit | commitdiff | tree |
2009-11-01 |
Chris Lattner | strengthen an assumption: RevectorBlockTo knows that...
|
commit | commitdiff | tree |
2009-11-01 |
Chris Lattner | fix an issue where the verifier would reject a function...
|
commit | commitdiff | tree |
2009-11-01 |
Chris Lattner | if CostMetrics says to never duplicate some code, don...
|
commit | commitdiff | tree |
2009-11-01 |
Chris Lattner | constant fold indirectbr(blockaddress(%bb)) -> br label...
|
commit | commitdiff | tree |
2009-11-01 |
Chris Lattner | improve x86 codegen support for blockaddress. We now...
|
commit | commitdiff | tree |
2009-11-01 |
Chris Lattner | pull check for return inst out of loop, never inline...
|
commit | commitdiff | tree |
2009-11-01 |
Chris Lattner | Fix BlockAddress::replaceUsesOfWithOnConstant to correctly
|
commit | commitdiff | tree |
2009-11-01 |
Chris Lattner | implement linker support for BlockAddress.
|
commit | commitdiff | tree |
2009-11-01 |
Chris Lattner | Revert 85678/85680. The decision is to stay with the...
|
commit | commitdiff | tree |
2009-10-31 |
Chris Lattner | Make sure PRE doesn't split crit edges from indirectbr.
|
commit | commitdiff | tree |
2009-10-31 |
Chris Lattner | llvm::SplitEdge should refuse to split an edge from...
|
commit | commitdiff | tree |
2009-10-31 |
Chris Lattner | update the comment above llvm::SplitCriticalEdge, and...
|
commit | commitdiff | tree |
2009-10-31 |
Chris Lattner | adjust a couple xforms to work with null bb's in BlockAddress.
|
commit | commitdiff | tree |
2009-10-31 |
Chris Lattner | Make blockaddress(@func, null) be valid, and make ...
|
commit | commitdiff | tree |
2009-10-31 |
Chris Lattner | add a comment.
|
commit | commitdiff | tree |
2009-10-30 |
Chris Lattner | if basic blocks are destroyed while there are *just...
|
commit | commitdiff | tree |
2009-10-30 |
Chris Lattner | make hasAddressTaken() constant time by storing a refcount...
|
commit | commitdiff | tree |
2009-10-30 |
Chris Lattner | it isn't valid to take the address of the entry block.
|
commit | commitdiff | tree |
2009-10-29 |
Chris Lattner | add sanity check for indbr.
|
commit | commitdiff | tree |
2009-10-29 |
Chris Lattner | just for the hell of it, allow globalopt to statically...
|
commit | commitdiff | tree |
2009-10-29 |
Chris Lattner | add interpreter support for indirect goto / blockaddress...
|
commit | commitdiff | tree |
2009-10-29 |
Chris Lattner | greatly improve the LLVM IR bitcode encoding documentation,
|
commit | commitdiff | tree |
2009-10-29 |
Chris Lattner | teach various passes about blockaddress. We no longer
|
commit | commitdiff | tree |
2009-10-29 |
Chris Lattner | teach ValueMapper about BlockAddress', making bugpoint...
|
commit | commitdiff | tree |
2009-10-29 |
Chris Lattner | unindent massive blocks, no functionality change.
|
commit | commitdiff | tree |
2009-10-28 |
Chris Lattner | add IRBuilder support for IndirectBr
|
commit | commitdiff | tree |
2009-10-28 |
Chris Lattner | add a new 'SetCurrentDebugType' API (requested by Andrew...
|
commit | commitdiff | tree |
2009-10-28 |
Chris Lattner | add bitcode reader support for blockaddress. We can...
|
commit | commitdiff | tree |
2009-10-28 |
Chris Lattner | bitcode writer support for blockaddress.
|
commit | commitdiff | tree |
2009-10-28 |
Chris Lattner | Previously, all operands to Constant were themselves...
|
commit | commitdiff | tree |
2009-10-28 |
Chris Lattner | doc bug spotted by apinski
|
commit | commitdiff | tree |
2009-10-28 |
Chris Lattner | 'static const void *X = &&y' can only be put in the
|
commit | commitdiff | tree |
2009-10-28 |
Chris Lattner | full asmparser support for blockaddress. We can now do:
|
commit | commitdiff | tree |
2009-10-28 |
Chris Lattner | asmprinter support for BlockAddress.
|
commit | commitdiff | tree |
2009-10-28 |
Chris Lattner | when we tear down a module, we need to be careful to
|
commit | commitdiff | tree |
2009-10-28 |
Chris Lattner | basic blocks can now have non-instruction users.
|
commit | commitdiff | tree |
2009-10-28 |
Chris Lattner | rename indbr -> indirectbr to appease the residents...
|
commit | commitdiff | tree |
2009-10-28 |
Chris Lattner | IR support for the new BlockAddress constant kind....
|
commit | commitdiff | tree |
2009-10-27 |
Chris Lattner | typo
|
commit | commitdiff | tree |
2009-10-27 |
Chris Lattner | you can't take the address of the entry block of a...
|
commit | commitdiff | tree |
2009-10-27 |
Chris Lattner | improvements from gabor.
|
commit | commitdiff | tree |
2009-10-27 |
Chris Lattner | make the build build.
|
commit | commitdiff | tree |
2009-10-27 |
Chris Lattner | Random updates to passes for indbr, I need blockaddress...
|
commit | commitdiff | tree |
2009-10-27 |
Chris Lattner | cppbackend support for indbr
|
commit | commitdiff | tree |
2009-10-27 |
Chris Lattner | CBE support for indbr.
|
commit | commitdiff | tree |
2009-10-27 |
Chris Lattner | fix things pointed out by Dan!
|
commit | commitdiff | tree |
2009-10-27 |
Chris Lattner | document the forthcoming blockaddress constant.
|
commit | commitdiff | tree |
2009-10-27 |
Chris Lattner | don't use stdio
|
commit | commitdiff | tree |
2009-10-27 |
Chris Lattner | fix pasto pointed out by Rafael
|
commit | commitdiff | tree |
2009-10-27 |
Chris Lattner | add enough support for indirect branch for the feature...
|
commit | commitdiff | tree |
2009-10-27 |
Chris Lattner | change of mind :)
|
commit | commitdiff | tree |
2009-10-27 |
Chris Lattner | rename test.
|
commit | commitdiff | tree |
2009-10-27 |
Chris Lattner | Type.h doesn't need to #include LLVMContext.h
|
commit | commitdiff | tree |
2009-10-27 |
Chris Lattner | pseudosourcevalue is also still using getGlobalContext...
|
commit | commitdiff | tree |
2009-10-27 |
Chris Lattner | apparently the X86 JIT isn't fully contextized, it...
|
commit | commitdiff | tree |
2009-10-27 |
Chris Lattner | trim another #include
|
commit | commitdiff | tree |
2009-10-27 |
Chris Lattner | remove an unneeded #include.
|
commit | commitdiff | tree |
2009-10-27 |
Chris Lattner | Fix a pretty serious misfeature of the inliner: if...
|
commit | commitdiff | tree |
2009-10-27 |
Chris Lattner | convert to filecheck.
|
commit | commitdiff | tree |
2009-10-27 |
Chris Lattner | lang points out that the comment is out of date with...
|
commit | commitdiff | tree |
2009-10-26 |
Chris Lattner | reapply r85085 with a bugfix to avoid infinite looping.
|
commit | commitdiff | tree |
2009-10-26 |
Chris Lattner | Implement PR3266 & PR5276, folding:
|
commit | commitdiff | tree |
2009-10-25 |
Chris Lattner | convert or.ll to filecheck and merge or2 into it.
|
commit | commitdiff | tree |
2009-10-25 |
Chris Lattner | fix PR5295 where the .ll parser didn't reject a function...
|
commit | commitdiff | tree |
2009-10-25 |
Chris Lattner | fix PR5186: the JIT shouldn't try to codegen available_exter...
|
commit | commitdiff | tree |
2009-10-25 |
Chris Lattner | this is done.
|
commit | commitdiff | tree |
2009-10-25 |
Chris Lattner | Teach FoldBitCast to be able to handle bitcasts from...
|
commit | commitdiff | tree |
2009-10-25 |
Chris Lattner | move FoldBitCast earlier in the file, and use it instead of
|
commit | commitdiff | tree |
2009-10-25 |
Chris Lattner | refactor FoldBitCast to reduce nesting and to always...
|
commit | commitdiff | tree |
2009-10-24 |
Chris Lattner | various cleanups suggested by Duncan
|
commit | commitdiff | tree |
2009-10-24 |
Chris Lattner | fix PR5287, a serious regression from my previous patches...
|
commit | commitdiff | tree |
2009-10-23 |
Chris Lattner | some stuff is done, we still have constantexpr simplificatio...
|
commit | commitdiff | tree |
2009-10-23 |
Chris Lattner | teach libanalysis to simplify vector loads with bitcast...
|
commit | commitdiff | tree |
2009-10-23 |
Chris Lattner | enhance FoldReinterpretLoadFromConstPtr to handle loads...
|
commit | commitdiff | tree |
2009-10-23 |
Chris Lattner | teach libanalysis to fold int and fp loads from almost...
|
commit | commitdiff | tree |
2009-10-22 |
Chris Lattner | move another load optimization from instcombine ->...
|
commit | commitdiff | tree |
2009-10-22 |
Chris Lattner | move 'loading i32 from string' optimization from instcombine
|
commit | commitdiff | tree |
2009-10-22 |
Chris Lattner | Move some constant folding logic for loads out of instcombin...
|
commit | commitdiff | tree |
2009-10-22 |
Chris Lattner | testcase for PR4678 & rdar://7309675
|
commit | commitdiff | tree |
2009-10-22 |
Chris Lattner | fix warning.
|
commit | commitdiff | tree |
2009-10-22 |
Chris Lattner | llvm-ld doesn't throw.
|
commit | commitdiff | tree |
2009-10-22 |
Chris Lattner | this doesn't use EH either.
|
commit | commitdiff | tree |
2009-10-22 |
Chris Lattner | nothing opt uses can throw, remove the try block and...
|
commit | commitdiff | tree |
2009-10-22 |
Chris Lattner | Add some command line options for twiddling the default...
|
commit | commitdiff | tree |
2009-10-22 |
Chris Lattner | fix PR5262.
|
commit | commitdiff | tree |
2009-10-21 |
Chris Lattner | revert r84754, it isn't the right approach. Edwin...
|
commit | commitdiff | tree |
2009-10-21 |
Chris Lattner | Set comment string, patch by Johnny Chen!
|
commit | commitdiff | tree |
2009-10-21 |
Chris Lattner | make GVN work better when TD is not around:
|
commit | commitdiff | tree |
2009-10-21 |
Chris Lattner | tidy
|
commit | commitdiff | tree |
2009-10-21 |
Chris Lattner | IPSCCP is missing stuff.
|
commit | commitdiff | tree |
2009-10-20 |
Chris Lattner | add a real testcase for PR4313
|
commit | commitdiff | tree |
next |