Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / 205d92589bc8c59d4bba9ddf89e0eb3c4d548cda / lib / CodeGen / SelectionDAG / ScheduleDAG.cpp
  1. 3a09d89 Optimize ScheduleDAG's ComputeDepths and ComputeHeights to not need by Dan Gohman · 17 years ago
  2. 99a6cb9 disallow direct access to SDValue::ResNo, provide a getter instead by Gabor Greif · 17 years ago
  3. 475871a Rename SDOperand to SDValue. by Dan Gohman · 17 years ago
  4. 8968450 Tidy SDNode::use_iterator, and complete the transition to have it by Dan Gohman · 17 years ago
  5. e8be6c6 Add a new function, ReplaceAllUsesOfValuesWith, which handles bulk by Dan Gohman · 17 years ago
  6. 5e84368 Reapply 53476 and 53480, with a fix so that it properly updates by Dan Gohman · 17 years ago
  7. b5eec33 Back out 53476 and 53480 for now. Somehow they cause llc to miscompile 179.art. by Evan Cheng · 17 years ago
  8. 5e76c3b Fix an obsolete top-level comment. by Dan Gohman · 17 years ago
  9. 0f4012f Factor out debugging code into the common base class. by Dan Gohman · 17 years ago
  10. 8e5f2c6 Pool-allocation for MachineInstrs, MachineBasicBlocks, and by Dan Gohman · 17 years ago
  11. 6b345ee Make DenseMap's insert return a pair, to more closely resemble std::map. by Dan Gohman · 17 years ago
  12. cd920d9 Replace a few uses of SelectionDAG::getTargetNode with by Dan Gohman · 17 years ago
  13. c6be777 - Use a faster priority comparison function if -fast. - Code clean up. by Evan Cheng · 17 years ago
  14. 4406604 Split ISD::LABEL into ISD::DBG_LABEL and ISD::EH_LABEL, eliminating by Dan Gohman · 17 years ago
  15. 8d96144 Remove a redundant return. by Dan Gohman · 17 years ago
  16. 94d7a5f Remove ScheduleDAG's SUnitMap altogether. Instead, use SDNode's NodeId by Dan Gohman · 17 years ago
  17. 3627e34 Add a priority queue class, which is a wrapper around std::priority_queue by Dan Gohman · 17 years ago
  18. 4c8c830 Change ScheduleDAG's SUnitMap from DenseMap<SDNode*, vector<SUnit*> > by Dan Gohman · 17 years ago
  19. 3461cc9 Clean up some uses of std::distance, now that we have allnodes_size. by Dan Gohman · 17 years ago
  20. 83ec4b6 Wrap MVT::ValueType in a struct to get type safety by Duncan Sands · 17 years ago
  21. 97e60d9 Use a better idiom to silence compiler warnings. by Evan Cheng · 17 years ago
  22. b203114 Really silence compiler warnings. by Evan Cheng · 17 years ago
  23. 5087124 Silence some compiler warnings. by Evan Cheng · 17 years ago
  24. 02a260a Switch to using Simplified ConstantFP::get API. by Chris Lattner · 17 years ago
  25. 9cac525 Ongoing work on improving the instruction selection infrastructure: by Roman Levenstein · 17 years ago
  26. 80792f3 Treat EntryToken nodes as "passive" so that they aren't added to the by Dan Gohman · 17 years ago
  27. 36b5c13 Rename MemOperand to MachineMemOperand. This was suggested by by Dan Gohman · 17 years ago
  28. dc1adac Re-commit of the r48822, where the infinite looping problem discovered by Roman Levenstein · 17 years ago
  29. 8a50f1f Start of a series of patches related to implicit_def. by Evan Cheng · 17 years ago
  30. 6397c64 Backing out 48222 temporarily. by Evan Cheng · 17 years ago
  31. e326332 Use a linked data structure for the uses lists of an SDNode, just like by Roman Levenstein · 17 years ago
  32. cfbb2f0 A quick nm audit turned up several fixed tables and objects that were by Dan Gohman · 17 years ago
  33. c929823 Make insert_subreg a two-address instruction, vastly simplifying LowerSubregs pass. Add a new TII, subreg_to_reg, which is like insert_subreg except that it takes an immediate implicit value to insert into rather than a register. by Christopher Lamb · 17 years ago
  34. da47e6e Replace all target specific implicit def instructions with a target independent one: TargetInstrInfo::IMPLICIT_DEF. by Evan Cheng · 17 years ago
  35. db2d773 Livein copy scheduling fixes: do not coalesce physical register copies, correctly determine the safe location to insert the copies. by Evan Cheng · 17 years ago
  36. 9e23336 Experimental scheduler change to schedule / coalesce the copies added for function livein's. Take 2008-03-10-RegAllocInfLoop.ll, the schedule looks like this after these copies are inserted: by Evan Cheng · 17 years ago
  37. 1fab4a6 Recommitting parts of r48130. These do not appear to cause the observed failures. by Christopher Lamb · 17 years ago
  38. 676dd7c When the register allocator runs out of registers, spill a physical register around the def's and use's of the interval being allocated to make it possible for the interval to target a register and spill it right away and restore a register for uses. This likely generates terrible code but is before than aborting. by Evan Cheng · 17 years ago
  39. c5733ac variadic instructions don't have operand info for variadic arguments. by Chris Lattner · 17 years ago
  40. b779580 If a register operand comes from the variadic part of a node, don't by Chris Lattner · 17 years ago
  41. 45f36ea Correctly clone FlaggedNodes. by Dan Gohman · 17 years ago
  42. 4499e49 Revert 48125, 48126, and 48130 for now to unbreak some x86-64 tests. by Evan Cheng · 17 years ago
  43. 3feb017 Allow insert_subreg into implicit, target-specific values. by Christopher Lamb · 17 years ago
  44. f30e1cf teach X86InstrInfo::copyRegToReg how to copy into ST(0) from by Chris Lattner · 17 years ago
  45. 02b6d25 Add ScheduleDAG support for copytoreg where the src/dst register are by Chris Lattner · 17 years ago
  46. d9c4c45 fix 80 col violation by Chris Lattner · 17 years ago
  47. d86449e Some improvements related to the computation of heights, depths of SUnits. by Roman Levenstein · 17 years ago
  48. 917be68 Rename isOperand() to isOperandOf() (and other similar methods). It always confuses me. by Evan Cheng · 17 years ago
  49. 643afa5 Keep track how many commutes are performed by the scheduler. by Evan Cheng · 17 years ago
  50. 42a7788 Rename CountMemOperands to ComputeMemOperandsEnd to reflect what by Dan Gohman · 17 years ago
  51. e179584 Change how FP immediates are handled. by Nate Begeman · 18 years ago
  52. cc20cd5 From Chris' review: use isa instead of explicitly using classof. by Dan Gohman · 18 years ago
  53. 027ee7e From Chris' review: minor corrections in comments. by Dan Gohman · 18 years ago
  54. 6f0d024 Rename MRegisterInfo to TargetRegisterInfo. by Dan Gohman · 18 years ago
  55. 69de193 Re-apply the memory operand changes, with a fix for the static by Dan Gohman · 18 years ago
  56. a844bde SDIsel processes llvm.dbg.declare by recording the variable debug information descriptor and its corresponding stack frame index in MachineModuleInfo. This only works if the local variable is "homed" in the stack frame. It does not work for byval parameter, etc. by Evan Cheng · 18 years ago
  57. 1b08bbc Remove the nasty LABEL hack with a much less evil one. Now llvm.dbg.func.start implies a stoppoint is set. SelectionDAGISel records a new source line but does not create a ISD::LABEL node for this special stoppoint. Asm printer will magically print this label. This ensures nothing is emitted before. by Evan Cheng · 18 years ago
  58. 334dc1f Revert 46556 and 46585. Dan please fix the PseudoSourceValue problem and re-commit. by Evan Cheng · 18 years ago
  59. 7bdc5f7 Add a comment for a nasty short term hack. by Evan Cheng · 18 years ago
  60. bb81d97 Add an extra operand to LABEL nodes which distinguishes between debug, EH, or misc labels. This fixes the EH breakage. However I am not convinced this is *the* solution. by Evan Cheng · 18 years ago
  61. 175e815 Add more thorough error checking for NULL register classes. by Christopher Lamb · 18 years ago
  62. c6c391d Create a new class, MemOperand, for describing memory references by Dan Gohman · 18 years ago
  63. 965d1b4 Treat the label for the first @llvm.dbg.stoppoint the same way as the dbg_func_start label. Make sure nothing else is inserted before them. by Evan Cheng · 18 years ago
  64. 6b2cf28 A semi-gross fix for a debug info issue. When inserting the "function start" label (i.e. first label in the entry block) take care to insert it at the beginning of the block. by Evan Cheng · 18 years ago
  65. ff9b373 Even though InsertAtEndOfBasicBlock is an ugly hack it still deserves a proper name. Rename it to EmitInstrWithCustomInserter since it does not necessarily insert by Evan Cheng · 18 years ago
  66. 3035959 Use empty() instead of comparing size() with zero. by Dan Gohman · 18 years ago
  67. 749c6f6 rename TargetInstrDescriptor -> TargetInstrDesc. by Chris Lattner · 18 years ago
  68. 0ff2396 Rename all the M_* flags to be namespace qualified enums, and switch by Chris Lattner · 18 years ago
  69. e32d765 add more and significantly better comments to the rest of the machineinstr by Chris Lattner · 18 years ago
  70. 8f707e1 rename hasVariableOperands() -> isVariadic(). Add some comments. by Chris Lattner · 18 years ago
  71. 349c495 Move a bunch more accessors from TargetInstrInfo to TargetInstrDescriptor by Chris Lattner · 18 years ago
  72. ba6da5d remove some uses of MachineOpCode, move getSchedClass by Chris Lattner · 18 years ago
  73. 8ca5c67 Add predicates methods to TargetOperandInfo, and switch all clients by Chris Lattner · 18 years ago
  74. 6924430 Rename MachineInstr::getInstrDescriptor -> getDesc(), which reflects by Chris Lattner · 18 years ago
  75. d10fd97 Move copyRegToReg from MRegisterInfo to TargetInstrInfo. This is part of the by Owen Anderson · 18 years ago
  76. 534bcfb update a couple of references to SSARegMap. by Chris Lattner · 18 years ago
  77. 84bc542 Rename SSARegMap -> MachineRegisterInfo in keeping with the idea by Chris Lattner · 18 years ago
  78. fec65d5 use simplified operand addition methods. by Chris Lattner · 18 years ago
  79. 8019f41 Start using the simplified methods for adding operands. by Chris Lattner · 18 years ago
  80. 4ee451d Remove attribution from file headers, per discussion on llvmdev. by Chris Lattner · 18 years ago
  81. 7c07aeb Bug fix. Must also match ResNo when matching an operand with a user. by Evan Cheng · 18 years ago
  82. eb57ea7 Make labels work in asm blocks; allow labels as by Dale Johannesen · 18 years ago
  83. 514ab34 Executive summary: getTypeSize -> getTypeStoreSize / getABITypeSize. by Duncan Sands · 18 years ago
  84. 32dfbea EXTRACT_SUBREG coalescing support. The coalescer now treats EXTRACT_SUBREG like by Evan Cheng · 18 years ago
  85. f10c973 If a node that defines a physical register that is expensive to copy. The by Evan Cheng · 18 years ago
  86. cb406c2 Use empty() member functions when that's what's being tested for instead by Dan Gohman · 18 years ago
  87. 22a5299 If two instructions are both two-address code, favors (schedule closer to by Evan Cheng · 18 years ago
  88. 74d2fd8 Trim some unneeded fields. by Evan Cheng · 18 years ago
  89. 42d6027 - Move getPhysicalRegisterRegClass() from ScheduleDAG to MRegisterInfo. by Evan Cheng · 18 years ago
  90. 9efce63 Allow copyRegToReg to emit cross register classes copies. by Evan Cheng · 18 years ago
  91. a6fb1b6 Added major new capabilities to scheduler (only BURR for now) to support physical register dependency. The BURR scheduler can now backtrace and duplicate instructions in order to avoid "expensive / impossible to copy" values (e.g. status flag EFLAGS for x86) from being clobbered. by Evan Cheng · 18 years ago
  92. 713a98d Use struct SDep instead of std::pair for SUnit pred and succ lists. First step by Evan Cheng · 18 years ago
  93. 6900132 Remove dead code. by Evan Cheng · 18 years ago
  94. 7df31dc Teach the dag scheduler to handle inline asm nodes with multi-value immediate operands. by Chris Lattner · 18 years ago
  95. 33d5595 Do not emit copies for physical register output if it's not used. by Evan Cheng · 18 years ago
  96. 8409747 Instead of adding copyfromreg's to handle physical definitions. Now isel can by Evan Cheng · 18 years ago
  97. e24f8f1 Teach DAG scheduling how to properly emit subreg insert/extract machine instructions. PR1350 by Christopher Lamb · 18 years ago
  98. e7e7d0d Skeleton of post-RA scheduler; doesn't do anything yet. by Dale Johannesen · 18 years ago
  99. 5e2456c If the operand is marked M_OPTIONAL_DEF_OPERAND, then it's a def. by Evan Cheng · 18 years ago
  100. af825c8 When a node value is only used by a CopyToReg, use the user's dest. This should not be restricted to nodes that produce only a single value. by Evan Cheng · 18 years ago