- 3b008a3 Fix a use-after-free in post-ra-scheduling. by Benjamin Kramer · 16 years ago
- d725159 All callee-saved registers are live-out of a return block. by David Goodwin · 16 years ago
- 1cc6dd9 Remove neonfp attribute and instead set default based on CPU string. Add -arm-use-neon-fp to override the default. by David Goodwin · 16 years ago
- 9a051a5 Restore the -post-RA-scheduler flag as an override for the target specification. Remove -mattr for setting PostRAScheduler enable and instead use CPU string. by David Goodwin · 16 years ago
- be3039e Use MachineFrameInfo.getPristineRegs() to determine which callee-saved registers are available for anti-dependency breaking. Some cleanup. by David Goodwin · 16 years ago
- 4fb2891 Observe hasExtraSrcRegAllocReq and hasExtraDefRegAllocReq. Do not change by Evan Cheng · 16 years ago
- 8ff5c19 Fix integer overflow in instruction scheduling. This can happen if we have by Reid Kleckner · 16 years ago
- 17199b5 Remove -post-RA-schedule flag and add a TargetSubtarget method to enable post-register-allocation scheduling. By default it is off. For ARM, enable/disable with -mattr=+/-postrasched. Enable by default for cortex-a8. by David Goodwin · 16 years ago
- 0bb5af3 Use KILL instead of IMPLICIT_DEF in LowerSubregs pass. by Jakob Stoklund Olesen · 16 years ago
- a4c98a3 Fix bug in kill flag updating for post-register-allocation scheduling. When the kill flag of a superreg needs to be cleared because there are one or more subregs live, we instead add implicit-defs of those subregs and leave the kill flag on the superreg. This allows us to end the live-range of the superreg without ending the live-ranges of the subregs. by David Goodwin · 16 years ago
- 270d0f9 Enhance EmitInstrWithCustomInserter() so target can specify CFG changes that sdisel will use to properly complete phi nodes. by Evan Cheng · 16 years ago
- e3c9d23 It's a bool, so treat it like one. Fixes a MSVC warning. by Benjamin Kramer · 16 years ago
- 6c08cfc Create our own block initializer for kill fixups as the scheduling one wasn't doing the right thing. by David Goodwin · 16 years ago
- 7f65169 Add hidden flags to allow binary search of post-RA scheduling errors. by David Goodwin · 16 years ago
- c898520 Don't mark a register live at an undef use. by David Goodwin · 16 years ago
- 7cb103d Another stab at fixing up register kill flags after post-RA scheduling. by David Goodwin · 16 years ago
- ae6bc82 Fixup register kills after scheduling. by David Goodwin · 16 years ago
- 4883d90 convert LoopInfo.h and GraphWriter.h to use raw_ostream by Chris Lattner · 16 years ago
- 1f8c7a7 Fix counting of Post-RA scheduling stalls. Improve debug output. by David Goodwin · 16 years ago
- 2a767c0 This logic was accidentally inverted in r78767. by Dan Gohman · 16 years ago
- f7912ed Factor out the code for finding an available register for use by Dan Gohman · 16 years ago
- 92492f2 Use DEBUG macro for debug output. by David Goodwin · 16 years ago
- ebd694b Add some debug output. by David Goodwin · 16 years ago
- f20236a Replace DOUT. by David Goodwin · 16 years ago
- 6021b4d Post RA scheduler changes. Introduce a hazard recognizer that uses the target schedule information to accurately model the pipeline. Update the scheduler to correctly handle multi-issue targets. by David Goodwin · 16 years ago
- 0402315 Use setPreservesAll and setPreservesCFG in CodeGen passes. by Dan Gohman · 16 years ago
- 7667332 inline the global 'getInstrOperandRegClass' function into its callers by Chris Lattner · 16 years ago
- fbcc663 llvm_unreachable->llvm_unreachable(0), LLVM_UNREACHABLE->llvm_unreachable. by Torok Edwin · 16 years ago
- 56d0659 assert(0) -> LLVM_UNREACHABLE. by Torok Edwin · 16 years ago
- 1ff2727 Move getInstrOperandRegClass from the scheduler to TargetInstrInfo. by Evan Cheng · 17 years ago
- 5185617 Fix pr3954. The register scavenger asserts for inline assembly with by Bob Wilson · 17 years ago
- be69d60 Add parentheses to pacify gcc-4.3. by Duncan Sands · 17 years ago
- 64613ac Fix a post-RA scheduling liveness bug. When a basic block is being by Dan Gohman · 17 years ago
- dfaf646 When scheduling a block in parts, keep track of the overall by Dan Gohman · 17 years ago
- b3dbb21 Consider any instruction that modifies the stack pointer to be by Dan Gohman · 17 years ago
- b954343 Factor out more code for computing register live-range informationfor by Dan Gohman · 17 years ago
- 066757e Move getPointerRegClass from TargetInstrInfo to TargetRegisterInfo. by Evan Cheng · 17 years ago
- f4b08b4 Move ScheduleDAGInstrs.h to be a private header. Front-ends by Dan Gohman · 17 years ago
- d564353 Change the post-RA scheduler to iterate through the by Dan Gohman · 17 years ago
- 5f8a259 Instead of adding dependence edges between terminator instructions by Dan Gohman · 17 years ago
- 157e008 If an anti-dependence uses a non-allocatable register, set AntiDepReg by Dan Gohman · 17 years ago
- 014142f Fix the check for an empty basic block to check for an empty SUnits by Dan Gohman · 17 years ago
- b903071 Fix a "comparison between signed and unsigned integer expressions" warning. by Dan Gohman · 17 years ago
- ceac7c3 Initial hazard recognizer support in post-pass scheduling. This includes by Dan Gohman · 17 years ago
- 619ef48 Move a few containers out of ScheduleDAGInstrs::BuildSchedGraph by Dan Gohman · 17 years ago
- 906152a Tidy up #includes, deleting a bunch of unnecessary #includes. by Dan Gohman · 17 years ago
- 04543e7 Rename BuildSchedUnits to BuildSchedGraph, and refactor the by Dan Gohman · 17 years ago
- 650b1e7 Use ~0u instead of -1u as the special value, to hopefully avoid by Dan Gohman · 17 years ago
- b4d41e8 Eliminate the loop that walks the critical path. Instead, just track the by Dan Gohman · 17 years ago
- 5155918 Enable anti-dependence breaking by default when post-RA scheduling is enabled. by Dan Gohman · 17 years ago
- 4302b4a When breaking an anti-dependency, don't use a register which has seen by Dan Gohman · 17 years ago
- dddc1ac Fix some register-alias-related bugs in the post-RA scheduler liveness by Dan Gohman · 17 years ago
- 2d17089 Rewrite the SDep class, and simplify some of the related code. by Dan Gohman · 17 years ago
- d24be45 Drop the reg argument to isRegReDefinedByTwoAddr, which was redundant. by Dan Gohman · 17 years ago
- c1dee22 Ignore IMPLICIT_DEF instructions when computing physreg liveness. by Dan Gohman · 17 years ago
- 44f57df Use register names instead of numbers in debug output. by Dan Gohman · 17 years ago
- 84efaf6 Rewrite the liveness bookkeeping code to fix a bunch of by Dan Gohman · 17 years ago
- 0c91a5f Fix an inconsistency in a comment. by Dan Gohman · 17 years ago
- 434a3ca Don't charge the full latency for anti and output dependencies. This is by Dan Gohman · 17 years ago
- 444baea When looking for anti-dependences on the critical path, don't bother by Dan Gohman · 17 years ago
- 1a32dda Add a comment about callee-saved registers. by Dan Gohman · 17 years ago
- 2c96bdd Silence a warning. by Nick Lewycky · 17 years ago
- bb1298e Suppress warnings. by Dan Gohman · 17 years ago
- ad2134d Initial support for anti-dependence breaking. Currently this code does not by Dan Gohman · 17 years ago
- 4ce15e1 Factor out the code for verifying the work of the scheduler, by Dan Gohman · 17 years ago
- 60cb69e Experimental post-pass scheduling support. Post-pass scheduling by Dan Gohman · 17 years ago
- a79db30 Tidy up several unbeseeming casts from pointer to intptr_t. by Dan Gohman · 17 years ago
- 9947184 don't create the post-ra scheduler unless it is enabled. by Chris Lattner · 18 years ago
- f3ebc3f Remove attribution from file headers, per discussion on llvmdev. by Chris Lattner · 18 years ago
- 4dc35db Modify previous patch per review comments. by Dale Johannesen · 18 years ago
- 2182f06 Skeleton of post-RA scheduler; doesn't do anything yet. by Dale Johannesen · 18 years ago