- 2e3f44e Disable pthread support in googletest if llvm was configured without threads. by Benjamin Kramer · 14 years ago
- c72f8e9 Don't preserve all symbols in a .so and instead trust gold to know what is by Rafael Espindola · 14 years ago
- 869d60d Machine sink could potentially sink instructions into a block where the physical by Bill Wendling · 14 years ago
- 8ef297e Turns out gtest still prefers the system <tr1/tuple> over it's own by Benjamin Kramer · 14 years ago
- 30ef0e5 Add first pass at darwin tls compiler support. by Eric Christopher · 14 years ago
- 162d3ba arm_neon.h now makes it through clang and generates appropriate code for those functions which can use by Nate Begeman · 14 years ago
- 7310064 One underscore, not two. by Eric Christopher · 14 years ago
- 268637e Implement expansion in type legalization for add/sub with overflow. The by Eli Friedman · 14 years ago
- db1bf34 Remove some already-fixed README entries. by Eli Friedman · 14 years ago
- a04a2c0 Remove README entry which no longer compiles to something sane. by Eli Friedman · 14 years ago
- 2ad7e43 Remove a fixed item, update a couple partially-fixed items. by Eli Friedman · 14 years ago
- 8b89c64 Use the fast register allocator by default for -O0 builds. by Jakob Stoklund Olesen · 14 years ago
- ead06be Use readsWritesVirtualRegister instead of counting uses and defs when inserting by Jakob Stoklund Olesen · 14 years ago
- 02571a3 Add full register <imp-def> operands when the coalescer is creating partial by Jakob Stoklund Olesen · 14 years ago
- 6209869 Speedup bitcode writer. Do not walk all values for all functions to emit function local metadata. In one testcase, probably worst case scenario, the 70x speed up is seen. by Devang Patel · 14 years ago
- 05c6837 Compulsive reformating. No functionalitical changes. by Bill Wendling · 14 years ago
- 9edf7de Slightly change the meaning of the reMaterialize target hook when the original by Jakob Stoklund Olesen · 14 years ago
- 4839d87 Update Readme and Makefiles for the new gtest. by Benjamin Kramer · 14 years ago
- 57240ff Merge gtest-1.5.0. by Benjamin Kramer · 14 years ago
- 190f8ee Merge gtest-1.4.0. by Benjamin Kramer · 14 years ago
- e4b9c93 Merge gtest-1.3.0. by Benjamin Kramer · 14 years ago
- 18f30e6 Clean up 80 column violations. No functional change. by Jim Grosbach · 14 years ago
- 7c8c883 arm_neon.h emitter now mostly complete for the purposes of initial testing. by Nate Begeman · 14 years ago
- 42d075c Remove the TargetRegisterClass member from CalleeSavedInfo by Rafael Espindola · 14 years ago
- 962f549 Fix comment so it doesn't include comments which are irrelevant to the x86 by Eli Friedman · 14 years ago
- 71c62a2 Use comments to document non-obvious code rather than mailing list archives. by Dan Gohman · 14 years ago
- 354eb7e Skip identical instruction while calculating DBG_VALUE range. by Devang Patel · 14 years ago
- 91a74da Rename canCombinedSubRegIndex method to something more grammatically correct by Bob Wilson · 14 years ago
- 20fae65 Replace ARM's getCalleeSavedRegClasses with a simpler solution by Rafael Espindola · 14 years ago
- 63ef367 remove trailing whitespace by Jim Grosbach · 14 years ago
- b9abe9f by Devang Patel · 14 years ago
- 691820a Remove unused function. by Rafael Espindola · 14 years ago
- 086723d Not all entries in the range will have an SUnit. Check for that when looking by Jim Grosbach · 14 years ago
- b26f279 Update the documentation, getCalleeSavedRegClasses is not required anymore. by Rafael Espindola · 14 years ago
- 6f07bd6 cleanup by Rafael Espindola · 14 years ago
- ce48c1d Remove uses of getCalleeSavedRegClasses from outside the by Rafael Espindola · 14 years ago
- 8dbbace Pacify recent gcc: remove a pointless const qualifier. by Duncan Sands · 14 years ago
- e66aab5 Checkpoint; handle 'int' and 'void' correctly by Nate Begeman · 14 years ago
- af905ef Emit full function prototypes. Definitions & typedefs to come. by Nate Begeman · 14 years ago
- 9d709a8 Enable machine cse of instructions which define physical registers. by Evan Cheng · 14 years ago
- 2223777 Checkpoint arm_neon.h generation with tablegen by Nate Begeman · 14 years ago
- a993f0a Don't try to custom-lower 64-bit add-with-overflow and friends on x86-32; the by Eli Friedman · 14 years ago
- b2c2a97 Fix an obvious mistake: don't change the operands until all of them have been by Bob Wilson · 14 years ago
- b242563 Remove outdated README entries. by Eli Friedman · 14 years ago
- 533934e Update debug information when breaking anti-dependencies. rdar://7759363 by Jim Grosbach · 14 years ago
- 891f273 Fix comment by Sean Hunt · 14 years ago
- 735419d DwarfWrite is now smart enough to drop debug value pointing to undefined register. Update this test to avoid this. by Devang Patel · 14 years ago
- a5135f6 Properly compose subregister indices when coalescing. by Jakob Stoklund Olesen · 14 years ago
- 23d3d45 Remove unused code by Jim Grosbach · 14 years ago
- f2e7767 Ignore line number of debug value in undefined register. by Devang Patel · 14 years ago
- 69081e9 fix think-o by Jim Grosbach · 14 years ago
- d8acddd Fix the allocation of shadow space for the Win64 calling convention by Dan Gohman · 14 years ago
- aeafb37 Simplify things a bit more. Fix prototype to use SmallVectorImpl and by Jim Grosbach · 14 years ago
- 10e7726 Fix the noalias checking so that it doesn't worry about by Dan Gohman · 14 years ago
- 9126c0d Keep track of incoming debug value of unused argument. Radar 7927666. by Devang Patel · 14 years ago
- 956b349 Fill in missing support for ISD::FEXP, ISD::FPOWI, and friends. by Dan Gohman · 14 years ago
- 2f3257e Add a FIXME by Jim Grosbach · 14 years ago
- 0fd8b62 mirror of r105280 changes for LowerInvoke, which uses the same basic logic here by Jim Grosbach · 14 years ago
- dc58b25 When processing function arguments when splitting live ranges across invokes, by Jim Grosbach · 14 years ago
- 601f9d6 Use SmallVector instead of std::vector. by Jim Grosbach · 14 years ago
- 426b06d Delete unused variables (that weren't even getting expanded). by Dan Gohman · 14 years ago
- f4ef72e Freebsd's sed doesn't support \< in regexps, and GNU sed doesn't by Dan Gohman · 14 years ago
- 7f5b0d8 Refactor some SSE 2 unpack instructions by Bruno Cardoso Lopes · 14 years ago
- 4b1000d Don't call flush() at a library level which isn't checking for errors by Dan Gohman · 14 years ago
- 943005e Fix a spello affecting platforms where the configure check fails. by Dan Gohman · 14 years ago
- d86e6ac Fix a typo in a comment. by Dan Gohman · 14 years ago
- da612d6 -retain-symbols-file is not what it seems. Update the makefiles by Dan Gohman · 14 years ago
- c6e1f67 Fix extra fread after EOF, non-wires-crossed version. by Dan Gohman · 14 years ago
- 11fe246 Fix handling of 'load' nodes. by Kalle Raiskila · 14 years ago
- 2133054 Debreak test for non-Darwin. by Bill Wendling · 14 years ago
- 2914ba6 Fix PR7272: when inlining through a callsite with byval arguments, by Duncan Sands · 14 years ago
- 89e7356 Add a test for the llvm-gcc commit in r90200. by Eric Christopher · 14 years ago
- 6c41ad8 fix PR6623: when optimizing for size, don't inline memcpy/memsets by Chris Lattner · 14 years ago
- 0ece9a1 upgrade and filecheckize this test. by Chris Lattner · 14 years ago
- 6b37861 the 'limit' argument to FindOptimalMemOpLowering is unsigned, not uint64_t. by Chris Lattner · 14 years ago
- caa3dcf revert r105223 which broke all my testing. by Chris Lattner · 14 years ago
- b6bd72e Fix count so that it doesn't make an extra fread call after EOF is detected. by Dan Gohman · 14 years ago
- c29df3c Avoid swap when a copy suffices. by Benjamin Kramer · 14 years ago
- a9d6680 The memcpy intrinsic only takes i8* for %src and %dst, so cast them to that by Nick Lewycky · 14 years ago
- 10e02a0 Use `llvm::next' instead of `next' to make VC++ 2010 happy. by Oscar Fuentes · 14 years ago
- c10a62b Allow for creation of clang DeclNodes tables. by Sean Hunt · 14 years ago
- 4ed81ec Some A9 load/store cleanups by Anton Korobeynikov · 14 years ago
- 8207fce Some rough approximations for load/stores on A9 by Anton Korobeynikov · 14 years ago
- 1098ef5 NEON/VFP stuff can be issued only via Pipe1 on A9 by Anton Korobeynikov · 14 years ago
- 1845a38 Add some integer instruction itineraries for A9 by Anton Korobeynikov · 14 years ago
- 462f6b5 Reorder some code in SelectionDAGBuilder. by Dan Gohman · 14 years ago
- 7451d3e SelectionDAG shouldn't have a FunctionLoweringInfo member. RegsForValue by Dan Gohman · 14 years ago
- 0f9c658 Remove unused function. by Benjamin Kramer · 14 years ago
- 9e130ce Document aggregate operation constant expressions. by Nick Lewycky · 14 years ago
- 6ed9d40 Fix typo. by Nick Lewycky · 14 years ago
- 701d4d3 Remove schedule-livein-copies. It's not being used. by Evan Cheng · 14 years ago
- 9c04467 Fix PR7193: if sibling call address can take a register, make sure there are enough registers available by counting inreg arguments. by Evan Cheng · 14 years ago
- 5a0d4fc Handle composed subreg indices when processing REQ_SEQUENCE instructions. by Jakob Stoklund Olesen · 14 years ago
- 82e0a1a Doh. Machine LICM is re-initializing the CSE map over and over. Patch by Anna Zaks. rdar://8037934. by Evan Cheng · 14 years ago
- 7c9a6e3 Emit TargetRegisterInfo::composeSubRegIndices(). by Jakob Stoklund Olesen · 14 years ago
- 05ce489 Merge the SystemZ subreg_even32 SubRegIndex into subreg_32bit. The SubRegIndices by Jakob Stoklund Olesen · 14 years ago
- 046fa3f Fix some latency computation bugs: if the use is not a machine opcode do not just return zero. by Evan Cheng · 14 years ago
- c10f543 Schedule high latency instructions for latency reduction even if they are not vfp / NEON instructions. by Evan Cheng · 14 years ago
- 2f05cc0 Fix comment typos. by Dale Johannesen · 14 years ago
- f39e0ce More SSE 1 & 2 merge, this time with logical instructions by Bruno Cardoso Lopes · 14 years ago