mirror of
https://github.com/c64scene-ar/llvm-6502.git
synced 2024-08-21 03:29:40 +00:00
As the AggressiveAntiDepBreaker iterated backward through a scheduling region, we must leave super registers live through subregister definitions so that all relevant subregister definitions are renamed together. The problem was that we were also discarding sub-register use locations as the sub-registers are redefined. The result is that we'd rename the super register along with some, but not all, subregister definitions. R0_D = {R0_L, R1_L} R0_L = {R0_S, R1_S} %R0_L<def> = TRLi9 16, pred:8, pred:%noreg %R1_L<def> = LSRLrr %R1_L<kill>, %R0_S, pred:8, pred:%noreg %R0_L<def> = LSRLrr %R2_L, %R0_S, pred:8, pred:%noreg, %R0_L<imp-use,kill> %R1_L<def> = ANDLri %R1_L<kill>, 2047, pred:8, pred:%noreg %R0_L<def> = ANDLri %R0_L<kill>, 2047, pred:8, pred:%noreg %R4_D<def> = ASRDrr %R0_D<kill>, %R6_S Anti: %R4_D<def> = ASRDrr %R0_D<kill>, %R6_S Def Groups: R4_D=g213->g215(via R4_S)->g214(via R4_L)->g216(via R5_S)->g216(via R4_L)->g217(via R5_L) Use Groups: R0_D=g0->g218(last-use) R1_L->g219(last-use) R6_S=g204->g220(last-use) Anti: %R0_L<def> = ANDLri %R0_L<kill>, 2047, pred:8, pred:%noreg Def Groups: R0_L=g208->g209(via R0_S)->g218(via R0_D)->g210(via R1_S)->g210(via R0_D) Antidep reg: R0_L (real dependency) Use Groups: R0_L=g210->g224(last-use) R0_S->g225(last-use) R1_S->g226(last-use) Anti: %R1_L<def> = ANDLri %R1_L<kill>, 2047, pred:8, pred:%noreg Def Groups: R1_L=g219->g210(via R0_D) Antidep reg: R1_L (real dependency) Use Groups: R1_L=g210->g229(last-use) Anti: %R0_L<def> = LSRLrr %R2_L, %R0_S, pred:8, pred:%noreg, %R0_L<imp-use,kill> Def Groups: R0_L=g224->g225(via R0_S)->g210(via R0_D)->g226(via R1_S)->g226(via R0_D) Antidep reg: R0_L Use Groups: R2_L=g192 R0_S=g226->g230(last-use) R0_L=g226->g231(last-use) R1_S->g232(last-use) Anti: %R1_L<def> = LSRLrr %R1_L<kill>, %R0_S, pred:8, pred:%noreg Def Groups: R1_L=g229->g226(via R0_D) Antidep reg: R1_L Use Groups: R1_L=g226->g233(last-use) R0_S=g230 Anti: %R0_L<def> = TRLi9 16, pred:8, pred:%noreg Def Groups: R0_L=g231->g230(via R0_S)->g226(via R0_D)->g232(via R1_S)->g232(via R0_D) Antidep reg: R0_L Rename Candidates for Group g232: R0_D: elcInt64Regs :: R0_D R1_D R2_D R3_D R4_D R5_D R8_D R9_D R10_D R11_D R12_D R13_D R14_D R15_D R16_D R17_D R18_D R19_D R20_D R21_D R22_D R23_D R24_D R25_D R0_L: elcIntRegs :: R0_L R1_L R2_L R3_L R4_L R5_L R8_L R9_L R10_L R11_L R12_L R13_L R14_L R15_L R16_L R17_L R18_L R19_L R20_L R21_L R22_L R23_L R24_L R25_L R0_S: elcShrtRegs elcShrtRegs :: R0_S R1_S R2_S R3_S R4_S R5_S R8_S R9_S R10_S R11_S R12_S R13_S R14_S R15_S R16_S R17_S R18_S R19_S R20_S R21_S R22_S R23_S R24_S R25_S Find Registers: [R12_D: R12_D R12_L R12_S] Breaking anti-dependence edge on R0_L: R0_D->R12_D(1 refs) R0_L->R12_L(2 refs) R0_S->R12_S(2 refs) Use Groups: ... %R12_L<def> = TRLi9 16, pred:8, pred:%noreg %R1_L<def> = LSRLrr %R1_L<kill>, %R12_S, pred:8, pred:%noreg %R0_L<def> = LSRLrr %R2_L<kill>, %R12_S, pred:8, pred:%noreg, %R12_L<imp-use> %R1_L<def> = ANDLri %R1_L<kill>, 2047, pred:8, pred:%noreg %R0_L<def> = ANDLri %R0_L<kill>, 2047, pred:8, pred:%noreg %R4_D<def> = ASRDrr %R12_D<kill>, %R6_S With this change, we now produce: Anti: %R4_D<def> = ASRDrr %R0_D<kill>, %R6_S Def Groups: R4_D=g213->g215(via R4_S)->g214(via R4_L)->g216(via R5_S)->g216(via R4_L)->g217(via R5_L) Use Groups: R0_D=g0->g218(last-use) R1_L->g219(last-use) R6_S=g204->g220(last-use) Anti: %R0_L<def> = ANDLri %R0_L<kill>, 2047, pred:8, pred:%noreg Def Groups: R0_L=g208->g209(via R0_S)->g218(via R0_D)->g210(via R1_S)->g210(via R0_D) Antidep reg: R0_L (real dependency) Use Groups: R0_L=g210 Anti: %R1_L<def> = ANDLri %R1_L<kill>, 2047, pred:8, pred:%noreg Def Groups: R1_L=g219->g210(via R0_D) Antidep reg: R1_L (real dependency) Use Groups: R1_L=g210 Anti: %R0_L<def> = LSRLrr %R2_L, %R0_S, pred:8, pred:%noreg, %R0_L<imp-use,kill> Def Groups: R0_L=g210->g210(via R0_D)->g210(via R0_D) Antidep reg: R0_L Use Groups: R2_L=g192 R0_S=g210 R0_L=g210 Anti: %R1_L<def> = LSRLrr %R1_L<kill>, %R0_S, pred:8, pred:%noreg Def Groups: R1_L=g210->g210(via R0_D) Antidep reg: R1_L Use Groups: R1_L=g210 R0_S=g210 Anti: %R0_L<def> = TRLi9 16, pred:8, pred:%noreg Def Groups: R0_L=g210->g210(via R0_D)->g210(via R0_D) Antidep reg: R0_L Rename Candidates for Group g210: R0_D: elcInt64Regs :: R0_D R1_D R2_D R3_D R4_D R5_D R8_D R9_D R10_D R11_D R12_D R13_D R14_D R15_D R16_D R17_D R18_D R19_D R20_D R21_D R22_D R23_D R24_D R25_D R0_L: elcIntRegs elcIntAIRegs elcIntRegs elcIntRegs elcIntRegs elcIntRegs :: R0_L R1_L R2_L R3_L R4_L R5_L R8_L R9_L R10_L R11_L R12_L R13_L R14_L R15_L R16_L R17_L R18_L R19_L R20_L R21_L R22_L R23_L R24_L R25_L R1_L: elcIntRegs elcIntRegs elcIntRegs elcIntRegs elcIntRegs :: R0_L R1_L R2_L R3_L R4_L R5_L R8_L R9_L R10_L R11_L R12_L R13_L R14_L R15_L R16_L R17_L R18_L R19_L R20_L R21_L R22_L R23_L R24_L R25_L R0_S: elcShrtRegs elcShrtRegs :: R0_S R1_S R2_S R3_S R4_S R5_S R8_S R9_S R10_S R11_S R12_S R13_S R14_S R15_S R16_S R17_S R18_S R19_S R20_S R21_S R22_S R23_S R24_S R25_S Find Registers: [R12_D: R12_D R12_L R13_L R12_S] Breaking anti-dependence edge on R0_L: R0_D->R12_D(1 refs) R0_L->R12_L(7 refs) R1_L->R13_L(5 refs) R0_S->R12_S(2 refs) Use Groups: ... %R12_L<def> = TRLi9 16, pred:8, pred:%noreg %R13_L<def> = LSRLrr %R13_L<kill>, %R12_S, pred:8, pred:%noreg %R12_L<def> = LSRLrr %R2_L<kill>, %R12_S<kill>, pred:8, pred:%noreg, %R12_L<imp-use,kill> %R13_L<def> = ANDLri %R13_L<kill>, 2047, pred:8, pred:%noreg %R12_L<def> = ANDLri %R12_L<kill>, 2047, pred:8, pred:%noreg %R4_D<def> = ASRDrr %R12_D, %R6_S, %R12_L<imp-def>, %R12_S<imp-def>, %R13_S<imp-def> As demonstrated by this example, this is also somewhat unfortunate, because there is actually no need to rename the super register in this case (it is fully covered by later subregister definitions), but we don't seem to track enough information here to exploit that either. Thanks to Daniil Troshkov for reporting the issue. The debug outputs in this commit message are from Daniil. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227311 91177308-0d34-0410-b5e6-96231b3b80d8 |
||
---|---|---|
.. | ||
AsmPrinter | ||
SelectionDAG | ||
AggressiveAntiDepBreaker.cpp | ||
AggressiveAntiDepBreaker.h | ||
AllocationOrder.cpp | ||
AllocationOrder.h | ||
Analysis.cpp | ||
AntiDepBreaker.h | ||
AtomicExpandPass.cpp | ||
BasicTargetTransformInfo.cpp | ||
BranchFolding.cpp | ||
BranchFolding.h | ||
CalcSpillWeights.cpp | ||
CallingConvLower.cpp | ||
CMakeLists.txt | ||
CodeGen.cpp | ||
CodeGenPrepare.cpp | ||
CriticalAntiDepBreaker.cpp | ||
CriticalAntiDepBreaker.h | ||
DeadMachineInstructionElim.cpp | ||
DFAPacketizer.cpp | ||
DwarfEHPrepare.cpp | ||
EarlyIfConversion.cpp | ||
EdgeBundles.cpp | ||
ErlangGC.cpp | ||
ExecutionDepsFix.cpp | ||
ExpandISelPseudos.cpp | ||
ExpandPostRAPseudos.cpp | ||
ForwardControlFlowIntegrity.cpp | ||
GCMetadata.cpp | ||
GCMetadataPrinter.cpp | ||
GCRootLowering.cpp | ||
GCStrategy.cpp | ||
GlobalMerge.cpp | ||
IfConversion.cpp | ||
InlineSpiller.cpp | ||
InterferenceCache.cpp | ||
InterferenceCache.h | ||
IntrinsicLowering.cpp | ||
JumpInstrTables.cpp | ||
LatencyPriorityQueue.cpp | ||
LexicalScopes.cpp | ||
LiveDebugVariables.cpp | ||
LiveDebugVariables.h | ||
LiveInterval.cpp | ||
LiveIntervalAnalysis.cpp | ||
LiveIntervalUnion.cpp | ||
LivePhysRegs.cpp | ||
LiveRangeCalc.cpp | ||
LiveRangeCalc.h | ||
LiveRangeEdit.cpp | ||
LiveRegMatrix.cpp | ||
LiveStackAnalysis.cpp | ||
LiveVariables.cpp | ||
LLVMBuild.txt | ||
LLVMTargetMachine.cpp | ||
LocalStackSlotAllocation.cpp | ||
MachineBasicBlock.cpp | ||
MachineBlockFrequencyInfo.cpp | ||
MachineBlockPlacement.cpp | ||
MachineBranchProbabilityInfo.cpp | ||
MachineCombiner.cpp | ||
MachineCopyPropagation.cpp | ||
MachineCSE.cpp | ||
MachineDominanceFrontier.cpp | ||
MachineDominators.cpp | ||
MachineFunction.cpp | ||
MachineFunctionAnalysis.cpp | ||
MachineFunctionPass.cpp | ||
MachineFunctionPrinterPass.cpp | ||
MachineInstr.cpp | ||
MachineInstrBundle.cpp | ||
MachineLICM.cpp | ||
MachineLoopInfo.cpp | ||
MachineModuleInfo.cpp | ||
MachineModuleInfoImpls.cpp | ||
MachinePassRegistry.cpp | ||
MachinePostDominators.cpp | ||
MachineRegionInfo.cpp | ||
MachineRegisterInfo.cpp | ||
MachineScheduler.cpp | ||
MachineSink.cpp | ||
MachineSSAUpdater.cpp | ||
MachineTraceMetrics.cpp | ||
MachineVerifier.cpp | ||
Makefile | ||
module.modulemap | ||
OcamlGC.cpp | ||
OptimizePHIs.cpp | ||
Passes.cpp | ||
PeepholeOptimizer.cpp | ||
PHIElimination.cpp | ||
PHIEliminationUtils.cpp | ||
PHIEliminationUtils.h | ||
PostRASchedulerList.cpp | ||
ProcessImplicitDefs.cpp | ||
PrologEpilogInserter.cpp | ||
PrologEpilogInserter.h | ||
PseudoSourceValue.cpp | ||
README.txt | ||
RegAllocBase.cpp | ||
RegAllocBase.h | ||
RegAllocBasic.cpp | ||
RegAllocFast.cpp | ||
RegAllocGreedy.cpp | ||
RegAllocPBQP.cpp | ||
RegisterClassInfo.cpp | ||
RegisterCoalescer.cpp | ||
RegisterCoalescer.h | ||
RegisterPressure.cpp | ||
RegisterScavenging.cpp | ||
ScheduleDAG.cpp | ||
ScheduleDAGInstrs.cpp | ||
ScheduleDAGPrinter.cpp | ||
ScoreboardHazardRecognizer.cpp | ||
ShadowStackGC.cpp | ||
SjLjEHPrepare.cpp | ||
SlotIndexes.cpp | ||
Spiller.h | ||
SpillPlacement.cpp | ||
SpillPlacement.h | ||
SplitKit.cpp | ||
SplitKit.h | ||
StackColoring.cpp | ||
StackMapLivenessAnalysis.cpp | ||
StackMaps.cpp | ||
StackProtector.cpp | ||
StackSlotColoring.cpp | ||
StatepointExampleGC.cpp | ||
TailDuplication.cpp | ||
TargetFrameLoweringImpl.cpp | ||
TargetInstrInfo.cpp | ||
TargetLoweringBase.cpp | ||
TargetLoweringObjectFileImpl.cpp | ||
TargetOptionsImpl.cpp | ||
TargetRegisterInfo.cpp | ||
TargetSchedule.cpp | ||
TwoAddressInstructionPass.cpp | ||
UnreachableBlockElim.cpp | ||
VirtRegMap.cpp |
//===---------------------------------------------------------------------===// Common register allocation / spilling problem: mul lr, r4, lr str lr, [sp, #+52] ldr lr, [r1, #+32] sxth r3, r3 ldr r4, [sp, #+52] mla r4, r3, lr, r4 can be: mul lr, r4, lr mov r4, lr str lr, [sp, #+52] ldr lr, [r1, #+32] sxth r3, r3 mla r4, r3, lr, r4 and then "merge" mul and mov: mul r4, r4, lr str r4, [sp, #+52] ldr lr, [r1, #+32] sxth r3, r3 mla r4, r3, lr, r4 It also increase the likelihood the store may become dead. //===---------------------------------------------------------------------===// bb27 ... ... %reg1037 = ADDri %reg1039, 1 %reg1038 = ADDrs %reg1032, %reg1039, %NOREG, 10 Successors according to CFG: 0x8b03bf0 (#5) bb76 (0x8b03bf0, LLVM BB @0x8b032d0, ID#5): Predecessors according to CFG: 0x8b0c5f0 (#3) 0x8b0a7c0 (#4) %reg1039 = PHI %reg1070, mbb<bb76.outer,0x8b0c5f0>, %reg1037, mbb<bb27,0x8b0a7c0> Note ADDri is not a two-address instruction. However, its result %reg1037 is an operand of the PHI node in bb76 and its operand %reg1039 is the result of the PHI node. We should treat it as a two-address code and make sure the ADDri is scheduled after any node that reads %reg1039. //===---------------------------------------------------------------------===// Use local info (i.e. register scavenger) to assign it a free register to allow reuse: ldr r3, [sp, #+4] add r3, r3, #3 ldr r2, [sp, #+8] add r2, r2, #2 ldr r1, [sp, #+4] <== add r1, r1, #1 ldr r0, [sp, #+4] add r0, r0, #2 //===---------------------------------------------------------------------===// LLVM aggressively lift CSE out of loop. Sometimes this can be negative side- effects: R1 = X + 4 R2 = X + 7 R3 = X + 15 loop: load [i + R1] ... load [i + R2] ... load [i + R3] Suppose there is high register pressure, R1, R2, R3, can be spilled. We need to implement proper re-materialization to handle this: R1 = X + 4 R2 = X + 7 R3 = X + 15 loop: R1 = X + 4 @ re-materialized load [i + R1] ... R2 = X + 7 @ re-materialized load [i + R2] ... R3 = X + 15 @ re-materialized load [i + R3] Furthermore, with re-association, we can enable sharing: R1 = X + 4 R2 = X + 7 R3 = X + 15 loop: T = i + X load [T + 4] ... load [T + 7] ... load [T + 15] //===---------------------------------------------------------------------===// It's not always a good idea to choose rematerialization over spilling. If all the load / store instructions would be folded then spilling is cheaper because it won't require new live intervals / registers. See 2003-05-31-LongShifts for an example. //===---------------------------------------------------------------------===// With a copying garbage collector, derived pointers must not be retained across collector safe points; the collector could move the objects and invalidate the derived pointer. This is bad enough in the first place, but safe points can crop up unpredictably. Consider: %array = load { i32, [0 x %obj] }** %array_addr %nth_el = getelementptr { i32, [0 x %obj] }* %array, i32 0, i32 %n %old = load %obj** %nth_el %z = div i64 %x, %y store %obj* %new, %obj** %nth_el If the i64 division is lowered to a libcall, then a safe point will (must) appear for the call site. If a collection occurs, %array and %nth_el no longer point into the correct object. The fix for this is to copy address calculations so that dependent pointers are never live across safe point boundaries. But the loads cannot be copied like this if there was an intervening store, so may be hard to get right. Only a concurrent mutator can trigger a collection at the libcall safe point. So single-threaded programs do not have this requirement, even with a copying collector. Still, LLVM optimizations would probably undo a front-end's careful work. //===---------------------------------------------------------------------===// The ocaml frametable structure supports liveness information. It would be good to support it. //===---------------------------------------------------------------------===// The FIXME in ComputeCommonTailLength in BranchFolding.cpp needs to be revisited. The check is there to work around a misuse of directives in inline assembly. //===---------------------------------------------------------------------===// It would be good to detect collector/target compatibility instead of silently doing the wrong thing. //===---------------------------------------------------------------------===// It would be really nice to be able to write patterns in .td files for copies, which would eliminate a bunch of explicit predicates on them (e.g. no side effects). Once this is in place, it would be even better to have tblgen synthesize the various copy insertion/inspection methods in TargetInstrInfo. //===---------------------------------------------------------------------===// Stack coloring improvements: 1. Do proper LiveStackAnalysis on all stack objects including those which are not spill slots. 2. Reorder objects to fill in gaps between objects. e.g. 4, 1, <gap>, 4, 1, 1, 1, <gap>, 4 => 4, 1, 1, 1, 1, 4, 4 //===---------------------------------------------------------------------===// The scheduler should be able to sort nearby instructions by their address. For example, in an expanded memset sequence it's not uncommon to see code like this: movl $0, 4(%rdi) movl $0, 8(%rdi) movl $0, 12(%rdi) movl $0, 0(%rdi) Each of the stores is independent, and the scheduler is currently making an arbitrary decision about the order. //===---------------------------------------------------------------------===// Another opportunitiy in this code is that the $0 could be moved to a register: movl $0, 4(%rdi) movl $0, 8(%rdi) movl $0, 12(%rdi) movl $0, 0(%rdi) This would save substantial code size, especially for longer sequences like this. It would be easy to have a rule telling isel to avoid matching MOV32mi if the immediate has more than some fixed number of uses. It's more involved to teach the register allocator how to do late folding to recover from excessive register pressure.