- c4b1abd Actually source file has already been uniquified into an id during isel. Eliminate the StringMap. by Evan Cheng · 17 years ago
- d0adbb5 Add data structure to define and track debug location during codegen. by Evan Cheng · 17 years ago
- fe6d2cd Fixed optimization of combining two shuffles where the first shuffle inputs by Mon P Wang · 17 years ago
- e7852d0 Take the next steps in making SDUse more consistent with LLVM Use, and by Dan Gohman · 17 years ago
- 19c10e6 Untabify code. by Scott Michel · 17 years ago
- c9c8b2a CellSPU: by Scott Michel · 17 years ago
- 5bf4b75 Fix a typo by Nate Begeman · 17 years ago
- c2fd67f De-identifying per sabre review by Nate Begeman · 17 years ago
- 18f0231 Handle single-entry phi nodes gracefully in condprop. by Chris Lattner · 17 years ago
- 3796a26 Fix PR3408 by making a non-obvious assumption very obvious, and by Chris Lattner · 17 years ago
- f6784a3 More cleanups and simplifications, no functionality change. by Chris Lattner · 17 years ago
- 2ba2543 tidy asserts by Chris Lattner · 17 years ago
- 51a0437 Map address space 256 to gs; similar mappings could be supported for the by Nate Begeman · 17 years ago
- 740ab03 Support pattern matching various x86 sse shifts. by Nate Begeman · 17 years ago
- 4469c53 silence a warning when assertions are disabled. by Chris Lattner · 17 years ago
- afb564c should have removed the + when manually applying a patch! by Torok Edwin · 17 years ago
- c8ad2d2 revert this patch for now, because Codegen does still want to generate SSE code, by Torok Edwin · 17 years ago
- 7f1d525 If user explicitly asks not to use SSE, don't force it. This fixes LLVM part of PR3402. by Torok Edwin · 17 years ago
- 3994610 Eliminate the loop that searches through each of the operands by Dan Gohman · 17 years ago
- 74692c0 Whitespace tidiments. by Dan Gohman · 17 years ago
- e77f89d Move the N->use_empty() assert from DeleteNode to by Dan Gohman · 17 years ago
- 237d873 Start generating arbitrary precision integer SCEVs. This removes the temporary by Nick Lewycky · 17 years ago
- 9d49791 The function that does nothing but call malloc is noalias return. by Nick Lewycky · 17 years ago
- 1902a12 Private linkage support for PPC / Darwin. by Evan Cheng · 17 years ago
- d498c8f Teach 2addr pass to be do more commuting. If both uses of a two-address instruction are killed, but the first operand has a use before and after the def, commute if the second operand does not suffer from the same issue. by Evan Cheng · 17 years ago
- 9b99485 Fix an indent and a typo. by Nate Begeman · 17 years ago
- 381e6f6 Revert previous change; even this mild and clearly by Dale Johannesen · 17 years ago
- e46a686 add note about possible GEP improvement with fields of size 0. by Torok Edwin · 17 years ago
- 08ffee5 testcase for PR3381. Also it was an empty struct, not a void after all. by Torok Edwin · 17 years ago
- 629e92b void* is represented as pointer to empty struct {}. by Torok Edwin · 17 years ago
- 32ca865 Some cleanups. No functional changes. by Owen Anderson · 17 years ago
- e4e094b Improve the inlining cost function a bit. Little practical effect. by Dale Johannesen · 17 years ago
- 3914f72 Make InstCombineStoreToCast handle aggregates more aggressively, by Chris Lattner · 17 years ago
- 8191371 Refactor code. No functionality change. by Evan Cheng · 17 years ago
- 3b64c6b Introduce two DWARF attribute extentions DW_AT_APPLE_optimized, DW_AT_APPLE_flags. by Devang Patel · 17 years ago
- 91387de hopefully address PR3379 by making the P modifier work in x86 inline asm. by Chris Lattner · 17 years ago
- 40b65c8 use CallSite::isCalle instead of slow getOperandNo by Gabor Greif · 17 years ago
- a36791d Simplify the logic of getting hold of a PHI predecessor block. by Gabor Greif · 17 years ago
- 60789e4 Add support for deleting a module provider from a JIT in such a way that it does not cause the owned module to be fully materialized. by Nate Begeman · 17 years ago
- 80303aa Empty DIType represents void. In this case no need to construct any type DIE. by Devang Patel · 17 years ago
- a90c8e6 Fold x-0 to x in unsafe-fp-math mode. This comes up in the by Dan Gohman · 17 years ago
- 95900f2 fix two more cases where we could let the NLPDI cache get unsorted. by Chris Lattner · 17 years ago
- 4433a09 Unconditionally reset 'cache' to zero, even if we don't need to resort it. by Chris Lattner · 17 years ago
- c16d37e Only check if coalescing is worthwhile when the result is targeting a more restrictive register class. by Evan Cheng · 17 years ago
- 9ce499a Stage two of fixing pre-alloc-splitting's code size issues: filter out restores that are just by Owen Anderson · 17 years ago
- 8c08d8c Cross register class coalescing. Not yet enabled. by Evan Cheng · 17 years ago
- a608687 Code did not follow associated comment. not a good idea. by Devang Patel · 17 years ago
- 49f38cb Set appropriate tag for the composite type. by Devang Patel · 17 years ago
- b54bfc2 a minor tweak to my previous patch, handle the invalidation case by Chris Lattner · 17 years ago
- 956ec27 Add an initial pass at dead spill/restore removal for pre alloc splitting. by Owen Anderson · 17 years ago
- bbfa244 do not sign extend characters input to isprint. This improves by Chris Lattner · 17 years ago
- 67ba223 Fix a minor bug in DAGCombiner's folding of SELECT. Folding "select C, 0, 1" by Bob Wilson · 17 years ago
- 760f86f Don't create ISD::FNEG nodes after legalize if they aren't legal. by Dan Gohman · 17 years ago
- edc4d69 introduce a useful abstraction to find out if a Use is in the call position of an instruction by Gabor Greif · 17 years ago
- 32aaee6 Revert r62553 and r62616 due to issues with portability. by Tanya Lattner · 17 years ago
- b67c958 rename methods in System/Host to be more consistent. by Chris Lattner · 17 years ago
- 4c24546 Add SelectionDAG::getNOT method to construct bitwise NOT operations, by Bob Wilson · 17 years ago
- fcc6f15 Few targets like the tiny little PIC16 have only 16-bit pointers. by Sanjiv Gupta · 17 years ago
- 536ab13 Eliminate a couple of fields from TargetRegisterClass: SubRegClasses and SuperRegClasses. These are not necessary. Also eliminate getSubRegisterRegClass and getSuperRegisterRegClass. These are slow and their results can change if register file names change. Just use TargetLowering::getRegClassFor() to get the right TargetRegisterClass instead. by Evan Cheng · 17 years ago
- e70909b fix a typo by Chris Lattner · 17 years ago
- 256baa4 add a note by Chris Lattner · 17 years ago
- 12a7db3 Fix PR3358, a really nasty bug where recursive phi translated by Chris Lattner · 17 years ago
- dbf5f25 Recognize inline asm for bswap on x86-64 GLIBC. This allows it by Dan Gohman · 17 years ago
- 541ed9f Do not use host floating point types when emitting by Dale Johannesen · 17 years ago
- a095c97 Also favors NOT64r. by Evan Cheng · 17 years ago
- 0f2d995 fix warning in release-asserts mode and spelling of assert. by Chris Lattner · 17 years ago
- 1e42c5b add getPointerToGlobal to the C bindings, patch by Lennart Augustsson! PR3364 by Chris Lattner · 17 years ago
- 455e9ab Remove uses of uint32_t in favor of 'unsigned' for better by Chris Lattner · 17 years ago
- 764fd0c Simplify ReduceLoadWidth's logic: it doesn't need several different by Dan Gohman · 17 years ago
- 1ef4d8f Fix a recent regression. ClrOpcode is not set for i8; for i8, if by Dan Gohman · 17 years ago
- 1421b7b Mimic gcc behaviour with regard to response files. by Mikhail Glushenkov · 17 years ago
- 2530566 Fixed build warnings. Restoring changes done in 62600, they were lost in 62655. by Sanjiv Gupta · 17 years ago
- 9fbc7e2 Cleanup whitespace and comments, and tweak some by Duncan Sands · 17 years ago
- d3be462 I accidentally removed this check in an earlier commit, which cause breakage in the pre alloc splitter. by Owen Anderson · 17 years ago
- 08b9b05 Implement LowerOperationWrapper for legalizer. by Sanjiv Gupta · 17 years ago
- d1e8d9c CellSPU: by Scott Michel · 17 years ago
- bb326bb Allow targets to legalize operations (with illegal operands) that produces multiple values. For example, a load with an illegal operand (a load produces two values, a value and chain). by Sanjiv Gupta · 17 years ago
- aaf414c Favors generating "not" over "xor -1". For example. by Evan Cheng · 17 years ago
- ed6af24 Make special cases (0 inf nan) work for frem. by Dale Johannesen · 17 years ago
- 6cf7c39 Be more aggressive about renumbering vregs after splitting them. by Owen Anderson · 17 years ago
- 4766159 Encode member accessibility. by Devang Patel · 17 years ago
- 2a57466 Appropriately mark fowrad decls. by Devang Patel · 17 years ago
- a894ae1 Fix PR3243: a LiveVariables bug. When HandlePhysRegKill is checking whether the last reference is also the last def (i.e. dead def), it should also check if last reference is the current machine instruction being processed. This can happen when it is processing a physical register use and setting the current machine instruction as sub-register's last ref. by Evan Cheng · 17 years ago
- 9c70819 Fix typo. Patch by Alexei Svitkine. by Duncan Sands · 17 years ago
- e9a7286 Use "SINT_TO_FP" instead of "UINT_TO_FP" when getting the exponent. This was by Bill Wendling · 17 years ago
- 2be5893 Fix struct member's debug info. by Devang Patel · 17 years ago
- 6906ba5 Need only one set of debug info versions enum. by Devang Patel · 17 years ago
- 04ee5a1 Change TargetInstrInfo::isMoveInstr to return source and destination sub-register indices as well. by Evan Cheng · 17 years ago
- 9ec50ab Fix global variable's address in a DIE. by Devang Patel · 17 years ago
- eab4a2e Enable debug info for enums. by Devang Patel · 17 years ago
- 24b66a8 Add an IEEE remainder function, which is not by Dale Johannesen · 17 years ago
- ccb3ab4 improve compatibility with various versions of graphviz, patch by by Chris Lattner · 17 years ago
- 79abedb Eliminate use of uint32_t to improve compatibility with cygwin by Chris Lattner · 17 years ago
- ce31b02 Enable debug info for composite types. by Devang Patel · 17 years ago
- e08eb9c Refactor code. No functionality change. by Evan Cheng · 17 years ago
- 6c53334 Shift types need to match. by Bill Wendling · 17 years ago
- 63bf29b another fix for PR3354 by Chris Lattner · 17 years ago
- 8289b05 Add a README entry noticed while investigating PR3216. by Dan Gohman · 17 years ago
- f5add58 Fix a dagcombine to not generate loads of non-round integer types, by Dan Gohman · 17 years ago
- c2997f4 Do not use DenseMap because the iterator is invalidated while constructing types. After all there was a reason why std::map was used initially! by Devang Patel · 17 years ago