Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / be8b084d8ab51ee7f402cad3b56347c8a260d42c / lib / Target / X86 / X86ISelLowering.cpp
  1. be8b084 Reapply piecies of r112942 and r112934 which don't do functional changes by Bruno Cardoso Lopes · 15 years ago
  2. b3e0669 Reapply Fix comment by Bruno Cardoso Lopes · 15 years ago
  3. 3139422 Revert r112934, "- Use specific nodes to match unpckl masks.", which introduced by Daniel Dunbar · 15 years ago
  4. 78541f2 Revert r112938 "Fix comment", which depends on r112934, which introduced some by Daniel Dunbar · 15 years ago
  5. a87ccce Revert r112942, "Use punpckh and unpckh family of nodes instead of using unpckh by Daniel Dunbar · 15 years ago
  6. 4b0c9f3 Use punpckh and unpckh family of nodes instead of using unpckh mask pattern fragment by Bruno Cardoso Lopes · 15 years ago
  7. 01f0847 Fix comment by Bruno Cardoso Lopes · 15 years ago
  8. 5e5342b - Use specific nodes to match unpckl masks. by Bruno Cardoso Lopes · 15 years ago
  9. c7c62bb Revert win64 changes. They seem to be incomplete by Anton Korobeynikov · 15 years ago
  10. 2f4fad9 Properly allocate win64 shadow reg area. Patch by Jan Sjodin! by Anton Korobeynikov · 15 years ago
  11. 3722f00 Replace unpckl_undef and unpckh_undef matching with target specific opcodes by Bruno Cardoso Lopes · 15 years ago
  12. dd69db8 Move condition out to prepare for more matching by Bruno Cardoso Lopes · 15 years ago
  13. ad10fb2 Remove checking for isUNPCKL_v_undef_Mask, the specific node is already emitted for it by Bruno Cardoso Lopes · 15 years ago
  14. d00bfe1 become more strict about when it's safe to use X86ISD::MOVLPS by Bruno Cardoso Lopes · 15 years ago
  15. 4783a3e Revert r112689, avoid those kind of checks cause they mess up with mmx by Bruno Cardoso Lopes · 15 years ago
  16. 56098f5 Use movlps, movlpd, movss and movsd specific nodes instead of pattern matching with movlp pattern fragment by Bruno Cardoso Lopes · 15 years ago
  17. 9cfad89 minor change, simplify some logic by Bruno Cardoso Lopes · 15 years ago
  18. e654b56 Move some functions around so they can be used for some other to come function by Bruno Cardoso Lopes · 15 years ago
  19. 013bb3d Use x86 specific MOVSLDUP node, add more patterns to match it and remove useless load nodes by Bruno Cardoso Lopes · 15 years ago
  20. 5023ef2 Use x86 specific MOVSHDUP node and add more patterns to match it by Bruno Cardoso Lopes · 15 years ago
  21. 7ff30bb Use MOVHLPS node instead of matching using movhlps and movhlps_undef pattern fragments by Bruno Cardoso Lopes · 15 years ago
  22. f2db5b4 Use MOVLHPS and MOVHLPS x86 nodes whenever possible. Also remove some useless nodes by Bruno Cardoso Lopes · 15 years ago
  23. 20a07f4 Use X86ISD::MOVSS and MOVSD to represent the movl mask pattern, also fix the handling of those nodes when seeking for scalars inside vector shuffles by Bruno Cardoso Lopes · 15 years ago
  24. 24faf61 fix the buildvector->insertp[sd] logic to not always create a redundant by Chris Lattner · 15 years ago
  25. 3ddcc43 fix the BuildVector -> unpcklps logic to not do pointless shuffles by Chris Lattner · 15 years ago
  26. 6e80e44 improve comments in the unpcklps generating logic, introduce by Chris Lattner · 15 years ago
  27. 27f1279 Clean up the logic of vector shuffles -> vector shifts. by Bruno Cardoso Lopes · 15 years ago
  28. c52bedb Properly handle passing of FP stuff to varargs function on Win64: by Anton Korobeynikov · 15 years ago
  29. af57738 zap the now unused MVT::getIntVectorWithNumElements by Bruno Cardoso Lopes · 15 years ago
  30. 8306968 implement SplitVecOp_CONCAT_VECTORS, fixing the included testcase with SSE1. by Chris Lattner · 15 years ago
  31. 97a2a56 fix sse1 only codegen in x86-64 mode, which is something we by Chris Lattner · 15 years ago
  32. 3e60a23 Revert this for now, PUNPCKLDQ dont operate on v4f32 by Bruno Cardoso Lopes · 15 years ago
  33. 9f7f83b Fix nasty mingw32 bug, which e.g. prevented llvm-gcc bootstrap there. by Anton Korobeynikov · 15 years ago
  34. f76c55a PUNPCKLDQ should also be used for v4f32 by Bruno Cardoso Lopes · 15 years ago
  35. 7338bbd teach lowering to get target specific nodes for pshufd, emulating the same isel behavior for now, so we can pass all vector shuffle tests by Bruno Cardoso Lopes · 15 years ago
  36. 92b651f Fix X86's isLegalAddressingMode to recognize that static addresses by Dan Gohman · 15 years ago
  37. 8878e21 Use pshufhw and pshuflw in more cases and fix getTargetShuffleNode number of arguments by Bruno Cardoso Lopes · 15 years ago
  38. 3efc077 Start using target speficic nodes for shuffles: pshufhw and pshuflw by Bruno Cardoso Lopes · 15 years ago
  39. 4654a07 Revert invalid r111792. Jump tables are not broken on x86-64 / coff, by Anton Korobeynikov · 15 years ago
  40. 3464cec Workaround broken jump tables on x86-64 COFF. by Michael J. Spencer · 15 years ago
  41. bf8154a Prepare LowerVECTOR_SHUFFLEv8i16 to use x86 target specific nodes directly by Bruno Cardoso Lopes · 15 years ago
  42. 3157ef1 This is the first step towards refactoring the x86 vector shuffle code. The by Bruno Cardoso Lopes · 15 years ago
  43. 3a1e54a More fixes for win64: by Anton Korobeynikov · 15 years ago
  44. c0b2a20 Rework how the non-sse2 memory barrier is lowered so that the by Eric Christopher · 15 years ago
  45. 132929a improve indentation by Chris Lattner · 15 years ago
  46. bb0a948 Fix comment to reflect code, and remove an unused argument by Bruno Cardoso Lopes · 15 years ago
  47. 8c05a85 Begin to support some vector operations for AVX 256-bit intructions. The long by Bruno Cardoso Lopes · 15 years ago
  48. d881627 Use ISD::ADD instead of ISD::SUB with a negated constant. This by Dan Gohman · 15 years ago
  49. 045573c Add AVX matching patterns to Packed Bit Test intrinsics. by Bruno Cardoso Lopes · 15 years ago
  50. 405f11b Support AVX 256-bit load and store intrinsics by Bruno Cardoso Lopes · 15 years ago
  51. ac09835 Support very basic (doesn't include ABI support in the front-end, varags, ...) 256-bit argument passing and return for AVX by Bruno Cardoso Lopes · 15 years ago
  52. b6729dc Make x86-64 membarriers work without sse and clean up some of the uses. by Eric Christopher · 15 years ago
  53. 98f9856 Support all 128-bit AVX vector intrinsics. Most part of them I already by Bruno Cardoso Lopes · 15 years ago
  54. b2eeed7 Revert r109652, and remove the offending assert in loadRegFromStackSlot instead. by Jakob Stoklund Olesen · 15 years ago
  55. 4c010ec Create a fixed stack object for varargs that is as large as any register. by Jakob Stoklund Olesen · 15 years ago
  56. 5140921 Implement a vectorized algorithm for <16 x i8> << <16 x i8> by Nate Begeman · 15 years ago
  57. bdcb5af ~40% faster vector shl <4 x i32> on SSE 4.1 Larger improvements for smaller types coming in future patches. by Nate Begeman · 15 years ago
  58. dee8101 On x86, f32 / f64 nodes share the same registers as 128-bit vector values. by Evan Cheng · 15 years ago
  59. 70017e4 Add an ILP scheduler. This is a register pressure aware scheduler that's by Evan Cheng · 15 years ago
  60. c76d23f The only supported calling convention for X86-64 uses by Dale Johannesen · 15 years ago
  61. 9a9d275 Custom lower the memory barrier instructions and add support by Eric Christopher · 15 years ago
  62. 90eb402 80-columns. by Eric Christopher · 15 years ago
  63. c8ea673 Fix a couple issues with Win64 ABI by Nate Begeman · 15 years ago
  64. dab4dac Pulling out previous patch, must've run the tests in the wrong directory. by Eric Christopher · 15 years ago
  65. 87f4137 Lower MEMBARRIER on x86 and support processors without SSE2. by Eric Christopher · 15 years ago
  66. 60108e9 Split -enable-finite-only-fp-math to two options: by Evan Cheng · 15 years ago
  67. b5378ea Use TargetOpcode::COPY instead of X86-native register copy instructions when by Jakob Stoklund Olesen · 15 years ago
  68. dedd974 Fix for PR7193 was overly conservative. The only case where sibcall callee by Evan Cheng · 15 years ago
  69. 84023e0 Reapply bottom-up fast-isel, with several fixes for x86-32: by Dan Gohman · 15 years ago
  70. d737fca An x86 function returns a floating point value in st(0), and we must make sure by Jakob Stoklund Olesen · 15 years ago
  71. 02266e2 --- Reverse-merging r107947 into '.': by Bob Wilson · 15 years ago
  72. 01dcb18 Fix the memoperand offsets in code generated for va_start. by Dan Gohman · 15 years ago
  73. bf87e24 Re-apply bottom-up fast-isel, with fixes. Be very careful to avoid emitting by Dan Gohman · 15 years ago
  74. 599b531 Change LEA to have 5 operands for its memory operand, just by Chris Lattner · 15 years ago
  75. ac0ed5d add some long-overdue enums to refer to the parts of the 5-operand by Chris Lattner · 15 years ago
  76. f595141 Revert 107840 107839 107813 107804 107800 107797 107791. by Dan Gohman · 15 years ago
  77. bcc8017 Move getExtLoad() and (some) getLoad() DebugLoc argument after EVT argument for consistency sake. by Evan Cheng · 15 years ago
  78. f423a69 Add X86FastISel support for return statements. This entails refactoring by Dan Gohman · 15 years ago
  79. a4160c3 Simplify FastISel's constructor by giving it a FunctionLoweringInfo by Dan Gohman · 15 years ago
  80. c940365 Split the SDValue out of OutputArg so that SelectionDAG-independent by Dan Gohman · 15 years ago
  81. e2b448c Accept RIP-relative symbols with 'i' constraint, and by Dale Johannesen · 15 years ago
  82. c9af33c CanLowerReturn doesn't need a SelectionDAG; it just needs an LLVMContext. by Dan Gohman · 15 years ago
  83. 0d881da Propagate debug loc. by Devang Patel · 15 years ago
  84. 14152b4 Reapply r107655 with fixes; insert the pseudo instruction into by Dan Gohman · 15 years ago
  85. 258c58c Revert r107655. by Dan Gohman · 15 years ago
  86. b81c771 Fix a bunch of custom-inserter functions to handle the case where by Dan Gohman · 15 years ago
  87. f7a0c7b Fix up -fstack-protector on linux to use the segment by Eric Christopher · 15 years ago
  88. 62f35a2 Have the X86 backend use Triple instead of a string and some enums. by Eric Christopher · 15 years ago
  89. 32b4b5a more tidying. by Chris Lattner · 15 years ago
  90. f172ecd Just rip v2f32 support completely out of the X86 backend. In by Chris Lattner · 15 years ago
  91. e35d984 fix PR7518 - terrible codegen of <2 x float>, by only marking by Chris Lattner · 15 years ago
  92. ed2ae13 Remove isSS argument from CreateFixedObject. Fixed objects cannot be spill slots so it's always false. by Evan Cheng · 15 years ago
  93. e1c2b9c use ArgOperand API by Gabor Greif · 15 years ago
  94. 78337b4 Remove pointless and unused variables. by Duncan Sands · 15 years ago
  95. c25ccf8 Reduce indentation via early exit. NFC. by Bill Wendling · 15 years ago
  96. 1cfe44a use ArgOperand API by Gabor Greif · 15 years ago
  97. 1784d16 The hasMemory argument is irrelevant to how the argument by Dale Johannesen · 15 years ago
  98. 730c07e - Reapply r106066 now that the bzip2 build regression has been fixed. by Bill Wendling · 15 years ago
  99. e5ff9ef Disallow matching "i" constraint to symbol addresses when by Dale Johannesen · 15 years ago
  100. 4e39e9d Reapply r106634, now that the bug it exposed is fixed. by Dan Gohman · 15 years ago