- 28252b6 fix PR8932, a case where arg promotion could infinitely promote. by Chris Lattner · 15 years ago
- 54cfe7e simplify a little by Chris Lattner · 15 years ago
- b0daffc add some commentary by Chris Lattner · 15 years ago
- 7e9b427 if an alloca is only ever accessed as a unit, and is accessed with load/store instructions, by Chris Lattner · 15 years ago
- 7072853 Use an irbuilder to get some trivial constant folding when doing a store by Chris Lattner · 15 years ago
- 6eb6116 remove a dead check, this was needed before we had an explicit veto on uses of phis. by Chris Lattner · 15 years ago
- 192228e enhance FoldOpIntoPhi in instcombine to try harder when a phi has by Chris Lattner · 15 years ago
- df55fea Spill R4 if it's going to be used to restore SP from FP. by Evan Cheng · 15 years ago
- 9922ccf remove the AllowAggressive argument to FoldOpIntoPhi. It is forced to false in the by Chris Lattner · 15 years ago
- 7dfe8fd more cleanups: use the IR builder. by Chris Lattner · 15 years ago
- 5aac832 tidy up code. by Chris Lattner · 15 years ago
- 66f708f Improve the safety of my globalopt enhancement by ensuring that the bitcast by Owen Anderson · 15 years ago
- 156eb0a fix PR8983, a broken assertion. by Chris Lattner · 15 years ago
- c1a6283 Implement AnalyzeBranch in Sparc Backend. by Venkatraman Govindaraju · 15 years ago
- 9cd3da4 fix PR8981, a crash trying to form a conditional inc with a floating point compare. by Chris Lattner · 15 years ago
- b99fdee reapply my fix for PR8961 with a tweak to properly handle by Chris Lattner · 15 years ago
- d5f656f simplify this code, it is still broken but will follow up on llvm-commits. by Chris Lattner · 15 years ago
- c850965 Revert "Archive: Replace all internal uses of PathV1 with PathV2. The external API still uses PathV1." by Michael J. Spencer · 15 years ago
- cad33c6 Simplify a README.txt entry significantly to expose the core issue. by Chandler Carruth · 15 years ago
- 02efda4 one of michael's recent patches broke this, temporarily disable by Chris Lattner · 15 years ago
- 0092b11 remove the partial specialization pass. It is unmaintained and has bugs. by Chris Lattner · 15 years ago
- a7a71a3 Archive: Fix spelling. by Michael J. Spencer · 15 years ago
- 770772e Archive: Replace all internal uses of PathV1 with PathV2. The external API still uses PathV1. by Michael J. Spencer · 15 years ago
- 36852d3 Support/GraphWriter: Replace all internal uses of PathV1 with PathV2. The external API still uses PathV1. by Michael J. Spencer · 15 years ago
- 5df5a22 Add an assert so we don't silently miscompile ctpop for bit widths > 128. by Benjamin Kramer · 15 years ago
- 28f0ed5 Support/PathV2: Add identify_magic. by Michael J. Spencer · 15 years ago
- b6516ae Reimplement CTPOP legalization with the "best" algorithm from by Benjamin Kramer · 15 years ago
- 248f9f2 Unittests/Support/Path: Tweak test. by Michael J. Spencer · 15 years ago
- b33594b Support/PathV2: Implement has_magic in terms of get_magic. by Michael J. Spencer · 15 years ago
- d6cdf1d Support/PathV2: Implement get_magic. by Michael J. Spencer · 15 years ago
- cd7f0a1 Add missing whitespace. by Nick Lewycky · 15 years ago
- 2820c25 Make constmerge a two-pass algorithm so that it won't miss merging by Nick Lewycky · 15 years ago
- 111fd9c Make config.h.cmake similar to config.h.in by Oscar Fuentes · 15 years ago
- bfa3b90 Try to unbreak selfhost. by Benjamin Kramer · 15 years ago
- e8f8139 Add a cache that protects mergefunc's internals from more surprises in DenseSet. by Nick Lewycky · 15 years ago
- 786c7cd Teach LazyValueInfo that allocas aren't NULL. Over all of llvm-test, this saves by Nick Lewycky · 15 years ago
- a5eaa86 Add a clarification about merging constants with and without unnamed_addr. by Rafael Espindola · 15 years ago
- ba7c38c Allow unnamed_addr on declarations. by Rafael Espindola · 15 years ago
- 6ccb5ef temporarily revert r123526. While working on a follow-on patch I by Chris Lattner · 15 years ago
- eeba3f5 fix rdar://8785296 - -fcatch-undefined-behavior generates inefficient code by Chris Lattner · 15 years ago
- 1a8943a simplify code, no functionality change. by Chris Lattner · 15 years ago
- 94e8e0c Now that instruction optzns can update the iterator as they go, we can by Chris Lattner · 15 years ago
- 7579609 make the current instruction iterator an ivar, allowing xforms that by Chris Lattner · 15 years ago
- 62fe406 implement an instcombine xform that canonicalizes casts outside of and-with-constant operations. by Chris Lattner · 15 years ago
- bdf466c fix typo by Chris Lattner · 15 years ago
- 4de2c76 Fix m_Not and m_Neg to not match random ConstantInt's. Before by Chris Lattner · 15 years ago
- 6792032 one more instcombine variant that is needed to work with future changes, by Chris Lattner · 15 years ago
- 27a9848 fix typo by Chris Lattner · 15 years ago
- fdb5b01 Catch ~x < cst just like ~x < ~y, we currently handle this through by Chris Lattner · 15 years ago
- 28621cb reduce indentation by Chris Lattner · 15 years ago
- a0f720f 80-col. by Eric Christopher · 15 years ago
- deaf55f Generalize LoadAndStorePromoter a bit and switch LICM to use it. by Chris Lattner · 15 years ago
- ca3f069 Fix a comment. by Bob Wilson · 15 years ago
- 41262da Fix 80-cols. by Eric Christopher · 15 years ago
- 439ea27 Update CMake build. by Ted Kremenek · 15 years ago
- dc1a2bd Fix some tablegen issues to allow using zero_reg for InstAlias definitions. by Bob Wilson · 15 years ago
- 7cefd0e Fix memory leak found by clang static analyzer. by Ted Kremenek · 15 years ago
- d7f696e 'HiReg' is written but never read. Nuke its declaration and its assignments. by Ted Kremenek · 15 years ago
- e95a32c Fix a false-positive warning. by Owen Anderson · 15 years ago
- ca5f616 Delete an assignment to ThisBB which isn't needed, and tidy up some comments. by Dan Gohman · 15 years ago
- cff6b37 Enhance GlobalOpt to be able evaluate initializers that involve stores through by Owen Anderson · 15 years ago
- 3965b5e Add a possibility to switch between CFI directives- and table-based frame description emission. Currently all the backends use table-based stuff. by Anton Korobeynikov · 15 years ago
- e749911 Cleanup by Anton Korobeynikov · 15 years ago
- 9a1ef4e Add CFI directives-based frame information emission. Not hooked yet. by Anton Korobeynikov · 15 years ago
- d7e8ddc Split stuff as a preparation for CFI directives-based frame information emission by Anton Korobeynikov · 15 years ago
- c658520 Use common style for .cfi directives by Anton Korobeynikov · 15 years ago
- f697c8a Support for precise scheduling of the instruction selection DAG, by Andrew Trick · 15 years ago
- d0f5613 switch SRoA to use LoadAndStorePromoter instead of its own copy of the code. by Chris Lattner · 15 years ago
- a2d845a Add a new LoadAndStorePromoter class, which implements the general by Chris Lattner · 15 years ago
- f18e4c3 OperandTraits<>::Layout isn't used for anything. Remove it. by Jay Foad · 15 years ago
- 04b9a43 Update llvm-gcc's tests. by Rafael Espindola · 15 years ago
- da7e287 Reorder macros on config.h.cmake to easily compare it against config.h.in. by Oscar Fuentes · 15 years ago
- 3f84a45 Disable debug mode. by Devang Patel · 15 years ago
- c087e20 Turn X-(X-Y) into Y. According to my auto-simplifier this is the most common by Duncan Sands · 15 years ago
- cf80bc1 Factorize common code out of the InstructionSimplify shift logic. Add in by Duncan Sands · 15 years ago
- 62becca Rename this test. by Duncan Sands · 15 years ago
- 998ef4a switch the second scalarrepl pass to use SSAUpdater. We run two scalarrepl passes: one by Chris Lattner · 15 years ago
- b352d6e split SROA into two passes: one that uses DomFrontiers (-scalarrepl) by Chris Lattner · 15 years ago
- 25052d8 Remove casts between Value** and Constant**, which won't work if a by Jay Foad · 15 years ago
- e0a1a5b Implement full support for promoting allocas to registers using SSAUpdater by Chris Lattner · 15 years ago
- bdb6b7f relax testcase a bit. by Chris Lattner · 15 years ago
- 0422390 Try for the third time to teach getFirstTerminator() about debug values. by Jakob Stoklund Olesen · 15 years ago
- d357e88 revert my fastisel patch again which apparently still gives the by Chris Lattner · 15 years ago
- 9e27cc8 reapply r123414 now that the botz are calmed down and the fix is already in. by Chris Lattner · 15 years ago
- 70e5e22 indentation by Chris Lattner · 15 years ago
- f3eb3bb Completed :lower16: / :upper16: support for movw / movt pairs on Darwin. by Evan Cheng · 15 years ago
- b6436e5 Revert r123419. It still breaks llvm-gcc-i386-linux-selfhost. by Jakob Stoklund Olesen · 15 years ago
- a899d1c r123414 broke llvm-gcc bootstrap apparently, revert by Chris Lattner · 15 years ago
- 67d9891 Set the insertion point correctly for instructions generated by load folding: by Chris Lattner · 15 years ago
- 09befe9 Try again to teach getFirstTerminator() about debug values. by Jakob Stoklund Olesen · 15 years ago
- c5f608b Rather than doing early instcombine, try doing early CSE instead. This should still handle by Owen Anderson · 15 years ago
- c43cee3 Move some shift transforms out of instcombine and into InstructionSimplify. by Duncan Sands · 15 years ago
- 68a659d Don't bother conditionalizing the use of SROA in -O1 mode. We're already running it unconditionally by Owen Anderson · 15 years ago
- d754041 fix PR8961 - a fast isel miscompilation where we'd insert a new instruction by Chris Lattner · 15 years ago
- eea666f Better terminator avoidance. by Jakob Stoklund Olesen · 15 years ago
- 8fbbd1c Add comment about Thumb2 fixup comments being completely bogus. by Evan Cheng · 15 years ago
- 21d842c Add single entry / single exit accessors. by Tobias Grosser · 15 years ago
- 0c9f250 Recognize alternative register names like ip -> r12. by Owen Anderson · 15 years ago
- f7ca976 Fix a few more places that should use MBB::getLastNonDebugInstr(). by Jakob Stoklund Olesen · 15 years ago
- 0fba714 As far as I can tell, unified syntax uses c0-c15 instead of cr0-cr15 for mcr and friends. by Owen Anderson · 15 years ago