- 61425c0 MCJIT support for non-function sections. by Jim Grosbach · 13 years ago
- 27bf560 Fixed comment in loop-unswitch. by Stepan Dyatkovskiy · 13 years ago
- 63c6672 Give better scavenger errors by invoking the verifier. by Jakob Stoklund Olesen · 13 years ago
- 7739cad Add a new kind of MachineOperand: MO_RegisterMask. by Jakob Stoklund Olesen · 13 years ago
- 1857b51 Make sure the non-SSE lowering for fences correctly clobbers EFLAGS. PR11768. by Eli Friedman · 13 years ago
- a20b715 Get rid of unused codegen-only instruction. by Eli Friedman · 13 years ago
- a0c1fc3 Adding a Hexagon cell for segmented stacks, as they have been implemented for X86 and not Sparc... by Joe Abbey · 13 years ago
- b7ab7fe Give priority to AVX over SSE for 128-bit floating point unpck instructions. by Craig Topper · 13 years ago
- 9223822 A fix for the previous commit: "integer constant is too large for ‘long’ type" error on some 32-bit bots by Eli Bendersky · 13 years ago
- a66a185 Adding a basic ELF dynamic loader and MC-JIT for ELF. Functionality is currently basic and will be enhanced with future patches. by Eli Bendersky · 13 years ago
- 810d6d3 Refactor variables unused under non-assert builds (& remove two entirely unused variables). by David Blaikie · 13 years ago
- bf42139 Changed intrinsic ID operand to a target constant as its not used in any arithmetic so should not be checked in legalisation by Pete Cooper · 13 years ago
- cc61656 [AVX] Optimize x86 VSELECT instructions using SimplifyDemandedBits. by Nadav Rotem · 13 years ago
- ed4c8c6 Return an ArrayRef from ShuffleVectorSDNode::getMask and push it through CodeGen. by Benjamin Kramer · 13 years ago
- ae4746b DAGCombiner: Deduplicate code. by Benjamin Kramer · 13 years ago
- 0499c1f Cosmetic patch for r148215. by Stepan Dyatkovskiy · 13 years ago
- 209287d Fixup for r148132. Type replacement for LoopsProperties: from DenseMap to std::map, since we need to keep a valid pointer to properties of current loop. by Stepan Dyatkovskiy · 13 years ago
- 980bce2 Relax the FileCheck assertion a bit -- all we really care about is that by Chandler Carruth · 13 years ago
- 482e4a8 FileCheck-ize a test, make it more specific to directly test the shift by Chandler Carruth · 13 years ago
- 4aab4ab Remove SetWorkingDirectory from the Process interface. Nothing in LLVM by Chandler Carruth · 13 years ago
- 49c0a9a Remove dead code. by David Blaikie · 13 years ago
- 799ea5c Truncate of undef is just undef of smaller size. by Craig Topper · 13 years ago
- d07ef50 Fix the memop type on a couple 256-bit AVX instructions that were using f128mem instead of f256mem. by Craig Topper · 13 years ago
- 446626d Add a bunch of AVX instructions to the folding tables. Also fixed the alignment on 256-bit AVX2 instructions. by Craig Topper · 13 years ago
- c464e99 Speculatively revert commit 148175 (rafael), to see if this fixes by Duncan Sands · 13 years ago
- dd1f22f Fix a corner case hit by redundant phi elimination running after LSR. by Andrew Trick · 13 years ago
- da22344 Only emit the Leh_func_endN symbol when needed. by Rafael Espindola · 13 years ago
- 3c58ba8 misched: Initial code for building an MI level scheduling DAG by Andrew Trick · 13 years ago
- 7ebcaf4 Move physreg dependency generation into aptly named addPhysRegDeps. by Andrew Trick · 13 years ago
- 5e920d7 misched: Added ScheduleDAGInstrs::IsPostRA by Andrew Trick · 13 years ago
- e9ef4ed misched: Invoke the DAG builder on each sequence of schedulable instructions. by Andrew Trick · 13 years ago
- 5edf2f0 Move things around to make the file navigable, even though it will probably be split up later. by Andrew Trick · 13 years ago
- bfe8afa After r147827 and r147902, it's now possible for unallocatable registers to be by Evan Cheng · 13 years ago
- d32d3b7 Fix pasto from r146196. by Chad Rosier · 13 years ago
- 4f00c08 Cleanup test case by adding checks for test names. by Chad Rosier · 13 years ago
- bd6dc3b Add TRI::getCallPreservedMask() hook. by Jakob Stoklund Olesen · 13 years ago
- 9218098 Fix an unused variable warning that Chad noticed. by Dan Gohman · 13 years ago
- 20fb487 Add a test showing how the Leh_func_endN symbol is used. by Rafael Espindola · 13 years ago
- 54f7533 Remove previous commit while I debug the bot failures. by Rafael Espindola · 13 years ago
- ece8b73 Use RegisterTuples to generate pseudo-registers. by Jakob Stoklund Olesen · 13 years ago
- cf05d03 Remove label that is not used anymore. by Rafael Espindola · 13 years ago
- 09a536a Speculatively revert r148132+r148133 to try and fix a buildbot failure. by Eli Friedman · 13 years ago
- 794481d Skip the NAME field when forming tuples. by Jakob Stoklund Olesen · 13 years ago
- 23d1c5c Remove pointless mode line in .cpp file. by Andrew Trick · 13 years ago
- 8d6796b Try to clarify a little how exception handling works. by Duncan Sands · 13 years ago
- 85d5aae by Devang Patel · 13 years ago
- 5dab2bf Cosmetic patch for r148132. by Stepan Dyatkovskiy · 13 years ago
- a5a1e46 LoopUnswitch: All helper data that is collected during loop-unswitch iterations was moved to separated class (LUAnalysisCache). by Stepan Dyatkovskiy · 13 years ago
- c798cc4 Refactor. by Devang Patel · 13 years ago
- 4a5c0fd Add new test. by Devang Patel · 13 years ago
- 0f3023c Fixed comment. Thanks Duncan! by Pete Cooper · 13 years ago
- 0518970 Convert SHUFPD with the same register for both sources to PSHUFD if it would prevent a register copy. Similar to SHUFPS, but requires the mask to be converted. by Craig Topper · 13 years ago
- 562659f use v8i32 as optimal mem type over v8f32 if AVX2 is enabled. Similar to SSE2 vs SSE1. by Craig Topper · 13 years ago
- 1221617 Make X86 instruction selection use 256-bit VPXOR for build_vector of all ones if AVX2 is enabled. This gives the ExeDepsFix pass a chance to choose FP vs int as appropriate. Also use v8i32 as the type for getZeroVector if AVX2 is enabled. This is consistent with SSE2 using prefering v4i32. by Craig Topper · 13 years ago
- db080e8 test/CodeGen/ARM/test-sharedidx.ll: Fix for -Asserts. by NAKAMURA Takumi · 13 years ago
- c30432a Add patterns for v16i16 and v32i8 immAllZerosV to select VPXOR to match v4i64 and v8i32. by Craig Topper · 13 years ago
- 96f678f Added the MachineSchedulerPass skeleton. by Andrew Trick · 13 years ago
- b3fe88f whitespace by Andrew Trick · 13 years ago
- 524ff02 wrong filename by Andrew Trick · 13 years ago
- 07680ec 80-col violation by Andrew Trick · 13 years ago
- b9c7f65 Use 8i32 constant pool entry for converting AVX2_SETALLONES. Possibly fixes PR11750. by Craig Topper · 13 years ago
- e6cf4a0 Fix typo in PerformAddCombine that caused any vector type to be checked for horizontal add/sub if AVX2 is enabled. This caused an assert to fail for non 128/256-bit vectors when done before type legalizing. Fixes PR11749. by Craig Topper · 13 years ago
- 8dd6f0c Delete CodeInit and CodeRecTy from TableGen. by Jakob Stoklund Olesen · 13 years ago
- ebaf92c Use uniqued StringInit pointers for lookups. by Jakob Stoklund Olesen · 13 years ago
- c4b527a DAGCombine's logic for forming pre- and post- indexed loads / stores were being by Evan Cheng · 13 years ago
- b4ee516 Fix off-by-one error. by Bill Wendling · 13 years ago
- 79522dc Implement proper ObjC ARC objc_retainBlock "escape" analysis, so that by Dan Gohman · 13 years ago
- ddfda5c Added MVT::v2f16 by Pete Cooper · 13 years ago
- bf5b13d Revert accidental commit. by Bill Wendling · 13 years ago
- 86b1a7d Fix the code that was WRONG. by Bill Wendling · 13 years ago
- d578b90 Added FPOW, FEXP, FLOG to PromoteNode so that custom actions can be set to Promote for those operations. by Pete Cooper · 13 years ago
- 16db710 Fixed a bug in LowerVECTOR_SHUFFLE caused assertion failure by Elena Demikhovsky · 13 years ago
- c16c25f When hoisting common code, watch out for uses which are marked "kill". If the by Evan Cheng · 13 years ago
- 0577c59 Add error-reporting tests for platforms that don't support segmented stacks. by Rafael Espindola · 13 years ago
- 85b9d43 Support segmented stacks on 64-bit FreeBSD. by Rafael Espindola · 13 years ago
- e4d18de Support segmented stacks on win32. by Rafael Espindola · 13 years ago
- 989a681 Remove test case, as Chris suggested. by Devang Patel · 13 years ago
- 21d3c40 Add test case to check intel syntax parsing. by Devang Patel · 13 years ago
- f0a9535 Allow targets to select source order pre-RA scheduler. by Evan Cheng · 13 years ago
- dd929fc Rename X86ATTAsmParser -> X86AsmParser by Devang Patel · 13 years ago
- b20b518 Make SplitAnalysis::UseSlots private. by Jakob Stoklund Olesen · 13 years ago
- e2eb925 After Jakob's r147938 exception handling on i386 was completely broken. by Benjamin Kramer · 13 years ago
- d2070b0 Fix a bug in the AVX 256-bit shuffle code in cases where the splat element is on the boundary of two 128-bit vectors. by Nadav Rotem · 13 years ago
- fb418ba X86: Generalize the x << (y & const) optimization to also catch masks with more set bits set than 31 or 63. by Benjamin Kramer · 13 years ago
- e5dafc3 Move Sched::Preference out of TargetMachine.h where it is not referenced. by Evan Cheng · 13 years ago
- c59d9df Add predicate method check match memory operand size, if available. by Devang Patel · 13 years ago
- 805e224 A DenseMap of a std::map isn't a very good idea because the "grow()" method will by Bill Wendling · 13 years ago
- 0a33886 Add intel style operand parser skeleton. This is a work in progress. by Devang Patel · 13 years ago
- 0fe9a92 Switch all of the uses of my InsertDAGNode helper to follow the exact by Chandler Carruth · 13 years ago
- aa8b994 Revert r147978. A DenseMap's iterators may become invalidated here. by Bill Wendling · 13 years ago
- 9384111 Make data structures private. by Jakob Stoklund Olesen · 13 years ago
- 337a269 Use a DenseMap. by Bill Wendling · 13 years ago
- a8bd9a6 Sink spillInterferences into RABasic. by Jakob Stoklund Olesen · 13 years ago
- 078203f Cleanup. by Jakob Stoklund Olesen · 13 years ago
- ccc9581 Move RegAllocBase into its own cpp file separate from RABasic. by Jakob Stoklund Olesen · 13 years ago
- e15f421 Re-fix the issue Bill fixed in r147899 in a slightly different way, which doesn't abuse the semantics of linker_private. We don't really want to merge any string constant with a weak_odr global. by Eli Friedman · 13 years ago
- 4992881 80 columns. by Jim Grosbach · 13 years ago
- f6c22c6 lli should only create memmgr when JITing. by Jim Grosbach · 13 years ago
- bf9ab93 lli should create a JIT memory manager. by Jim Grosbach · 13 years ago
- fae699a Fix assert. by Eric Christopher · 13 years ago