llvm-6502/lib/CodeGen
Eli Bendersky 2ad047e04d Optimize MachineBasicBlock::getSymbol by caching the symbol. Since the symbol
name computation is expensive, this helps save about 25% of the time spent in
this function.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180049 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-22 21:21:08 +00:00
..
AsmPrinter Clarify that llvm.used can contain aliases. 2013-04-22 14:58:02 +00:00
SelectionDAG Legalize vector truncates by parts rather than just splitting. 2013-04-21 23:47:41 +00:00
AggressiveAntiDepBreaker.cpp Remove special-casing of return blocks for liveness. 2013-02-05 18:21:52 +00:00
AggressiveAntiDepBreaker.h
AllocationOrder.cpp Assert that the target provided hints are in the allocation order. 2013-02-19 18:41:01 +00:00
AllocationOrder.h Limit the search space in RAGreedy::tryEvict(). 2013-01-12 00:57:44 +00:00
Analysis.cpp Add CodeGen support for functions that always return arguments via a new parameter attribute 'returned', which is taken advantage of in target-independent tail call opportunity detection and in ARM call lowering (when placed on an integral first parameter). 2013-04-20 05:14:40 +00:00
AntiDepBreaker.h
BasicTargetTransformInfo.cpp Document the decision to assume that the cost of floats is twice as much as integers. 2013-04-14 05:55:18 +00:00
BranchFolding.cpp Move all of the header files which are involved in modelling the LLVM IR 2013-01-02 11:36:10 +00:00
BranchFolding.h
CalcSpillWeights.cpp typo 2013-04-06 04:24:12 +00:00
CallingConvLower.cpp Move all of the header files which are involved in modelling the LLVM IR 2013-01-02 11:36:10 +00:00
CMakeLists.txt Remove the old CodePlacementOpt pass. 2013-03-29 17:14:24 +00:00
CodeGen.cpp Remove the old CodePlacementOpt pass. 2013-03-29 17:14:24 +00:00
CriticalAntiDepBreaker.cpp Remove special-casing of return blocks for liveness. 2013-02-05 18:21:52 +00:00
CriticalAntiDepBreaker.h This patch addresses bug 15031. 2013-01-28 18:36:58 +00:00
DeadMachineInstructionElim.cpp Remove special-casing of return blocks for liveness. 2013-02-05 18:21:52 +00:00
DFAPacketizer.cpp
DwarfEHPrepare.cpp Split TargetLowering into a CodeGen and a SelectionDAG part. 2013-01-11 20:05:37 +00:00
EarlyIfConversion.cpp Allow MachineTraceMetrics to be used when the model has no resources. 2013-04-02 22:27:45 +00:00
EdgeBundles.cpp
ErlangGC.cpp Add a GC plugin for Erlang 2013-03-25 13:47:46 +00:00
ExecutionDepsFix.cpp
ExpandISelPseudos.cpp
ExpandPostRAPseudos.cpp Re-apply r175688, with the changes suggested by Jakob in PR15320. 2013-02-21 22:16:43 +00:00
GCMetadata.cpp Fix GCMetadaPrinter::finishAssembly not executed, patch by Yiannis Tsiouris. 2013-02-19 16:51:44 +00:00
GCMetadataPrinter.cpp
GCStrategy.cpp Move all of the header files which are involved in modelling the LLVM IR 2013-01-02 11:36:10 +00:00
IfConversion.cpp Manually remove successors in if conversion when CopyAndPredicateBlock is used 2013-04-10 22:05:25 +00:00
InlineSpiller.cpp
InterferenceCache.cpp
InterferenceCache.h
IntrinsicLowering.cpp Revert "Add LLVMContext::emitWarning methods and use them. <rdar://problem/12867368>" 2013-02-08 21:48:29 +00:00
JITCodeEmitter.cpp
LatencyPriorityQueue.cpp
LexicalScopes.cpp Correct indentation for dumping LexicalScope. 2013-02-02 00:02:03 +00:00
LiveDebugVariables.cpp Clean up LDV, no functionality change. 2013-02-13 20:23:48 +00:00
LiveDebugVariables.h Clean up LDV, no functionality change. 2013-02-13 20:23:48 +00:00
LiveInterval.cpp Don't allocate memory in LiveInterval::join(). 2013-02-20 23:51:10 +00:00
LiveIntervalAnalysis.cpp Rewrite the physreg part of findLastUseBefore(). 2013-03-08 18:08:57 +00:00
LiveIntervalUnion.cpp
LiveRangeCalc.cpp Copy single reaching defs directly into the LiveInterval. 2013-02-20 23:08:26 +00:00
LiveRangeCalc.h Copy single reaching defs directly into the LiveInterval. 2013-02-20 23:08:26 +00:00
LiveRangeEdit.cpp Add some constantness. 2013-03-18 23:40:46 +00:00
LiveRegMatrix.cpp
LiveStackAnalysis.cpp
LiveVariables.cpp Remove special-casing of return blocks for liveness. 2013-02-05 18:21:52 +00:00
LLVMBuild.txt Extracted ObjCARC.cpp into its own library libLLVMObjCARCOpts in preparation for refactoring the ARC Optimizer. 2013-01-28 01:35:51 +00:00
LLVMTargetMachine.cpp Fix GCMetadaPrinter::finishAssembly not executed, patch by Yiannis Tsiouris. 2013-02-19 16:51:44 +00:00
LocalStackSlotAllocation.cpp Move all of the header files which are involved in modelling the LLVM IR 2013-01-02 11:36:10 +00:00
MachineBasicBlock.cpp Optimize MachineBasicBlock::getSymbol by caching the symbol. Since the symbol 2013-04-22 21:21:08 +00:00
MachineBlockFrequencyInfo.cpp
MachineBlockPlacement.cpp Don't disable block layout when forcing block alignment. 2013-04-12 01:24:16 +00:00
MachineBranchProbabilityInfo.cpp Move all of the header files which are involved in modelling the LLVM IR 2013-01-02 11:36:10 +00:00
MachineCodeEmitter.cpp
MachineCopyPropagation.cpp
MachineCSE.cpp
MachineDominators.cpp
MachineFunction.cpp Move estimateStackSize from ARM into MachineFrameInfo 2013-03-14 21:15:20 +00:00
MachineFunctionAnalysis.cpp
MachineFunctionPass.cpp Move all of the header files which are involved in modelling the LLVM IR 2013-01-02 11:36:10 +00:00
MachineFunctionPrinterPass.cpp
MachineInstr.cpp Remove liveout lists from MachineRegisterInfo. 2013-02-05 18:21:56 +00:00
MachineInstrBundle.cpp Move an assertion so it doesn't dereference end(). 2013-01-04 22:17:31 +00:00
MachineLICM.cpp Split TargetLowering into a CodeGen and a SelectionDAG part. 2013-01-11 20:05:37 +00:00
MachineLoopInfo.cpp
MachineModuleInfo.cpp Clarify that llvm.used can contain aliases. 2013-04-22 14:58:02 +00:00
MachineModuleInfoImpls.cpp
MachinePassRegistry.cpp
MachinePostDominators.cpp
MachineRegisterInfo.cpp Add an MRI::verifyUseLists() function. 2013-04-19 21:40:57 +00:00
MachineScheduler.cpp MI-Sched: DEBUG formatting. 2013-04-13 06:07:49 +00:00
MachineSink.cpp
MachineSSAUpdater.cpp Use MachineInstrBuilder in a few CodeGen passes. 2012-12-20 18:08:06 +00:00
MachineTraceMetrics.cpp Allow MachineTraceMetrics to be used when the model has no resources. 2013-04-02 22:27:45 +00:00
MachineVerifier.cpp Add an MRI::verifyUseLists() function. 2013-04-19 21:40:57 +00:00
Makefile
OcamlGC.cpp
OptimizePHIs.cpp Move all of the header files which are involved in modelling the LLVM IR 2013-01-02 11:36:10 +00:00
Passes.cpp Add braces around || in && to pacify GCC. 2013-04-11 11:57:01 +00:00
PeepholeOptimizer.cpp Add debug prints for when optimizeLoadInstr folds a load. 2012-12-17 03:56:00 +00:00
PHIElimination.cpp Don't rely on the isDead() MachineOperand flag when updating LiveIntervals. 2013-02-21 08:51:58 +00:00
PHIEliminationUtils.cpp
PHIEliminationUtils.h
PostRASchedulerList.cpp Remove special-casing of return blocks for liveness. 2013-02-05 18:21:52 +00:00
ProcessImplicitDefs.cpp
PrologEpilogInserter.cpp Reapply r178845 with fix - Fix bug in PEI's virtual-register scavenging 2013-04-05 22:31:56 +00:00
PrologEpilogInserter.h
PseudoSourceValue.cpp Move all of the header files which are involved in modelling the LLVM IR 2013-01-02 11:36:10 +00:00
README.txt
RegAllocBase.cpp
RegAllocBase.h
RegAllocBasic.cpp Replace uses of the deprecated std::auto_ptr with OwningPtr. 2013-04-12 10:56:28 +00:00
RegAllocFast.cpp Make RAFast::UsedInInstr indexed by register units. 2013-02-21 19:35:21 +00:00
RegAllocGreedy.cpp Replace uses of the deprecated std::auto_ptr with OwningPtr. 2013-04-12 10:56:28 +00:00
RegAllocPBQP.cpp Replace uses of the deprecated std::auto_ptr with OwningPtr. 2013-04-15 12:06:32 +00:00
RegisterClassInfo.cpp Precompute some information about register costs. 2013-01-12 00:54:59 +00:00
RegisterCoalescer.cpp Remove unused #includes. 2013-03-05 01:00:45 +00:00
RegisterCoalescer.h
RegisterPressure.cpp Replace erase loop with std::remove_if. 2013-02-16 17:06:38 +00:00
RegisterScavenging.cpp Reapply r178845 with fix - Fix bug in PEI's virtual-register scavenging 2013-04-05 22:31:56 +00:00
ScheduleDAG.cpp Scheduler diagnostics. Print the register name. 2013-03-01 00:19:14 +00:00
ScheduleDAGInstrs.cpp MI-Sched: schedule physreg copies. 2013-04-13 06:07:40 +00:00
ScheduleDAGPrinter.cpp ScheduleDAG: colorize the DOT graph and improve formatting. 2013-01-25 07:45:25 +00:00
ScoreboardHazardRecognizer.cpp
ShadowStackGC.cpp Move all of the header files which are involved in modelling the LLVM IR 2013-01-02 11:36:10 +00:00
ShrinkWrapping.cpp Remove duplicate includes. 2012-12-21 17:06:44 +00:00
SjLjEHPrepare.cpp Revert r176154 in favor of a better approach. 2013-03-08 02:21:08 +00:00
SlotIndexes.cpp Make some fixes for LiveInterval repair with debug info. Debug value 2013-02-23 10:25:25 +00:00
Spiller.cpp
Spiller.h
SpillPlacement.cpp Move #include of BitVector from .h to .cpp file. 2013-03-18 23:45:45 +00:00
SpillPlacement.h
SplitKit.cpp
SplitKit.h
StackColoring.cpp Couple more sets of tidying. 2013-03-25 21:26:36 +00:00
StackProtector.cpp Remove unused #includes. 2013-03-05 01:00:45 +00:00
StackSlotColoring.cpp Move all of the header files which are involved in modelling the LLVM IR 2013-01-02 11:36:10 +00:00
StrongPHIElimination.cpp
TailDuplication.cpp Move all of the header files which are involved in modelling the LLVM IR 2013-01-02 11:36:10 +00:00
TargetFrameLoweringImpl.cpp
TargetInstrInfo.cpp Add static cast to unsigned char whenever a character classification function is called with a signed char argument, in order to avoid assertions in Windows Debug configuration. 2013-02-12 21:21:59 +00:00
TargetLoweringBase.cpp Remove unused ShouldFoldAtomicFences flag. 2013-04-20 12:32:43 +00:00
TargetLoweringObjectFileImpl.cpp Windows TLS: Section name prefix to ensure correct order 2013-03-18 08:10:10 +00:00
TargetOptionsImpl.cpp Use the target options specified on a function to reset the back-end. 2013-04-05 21:52:40 +00:00
TargetRegisterInfo.cpp Remove unneeded "TargetMachine.h" #includes. 2013-02-09 20:54:05 +00:00
TargetSchedule.cpp MI-Sched cleanup. If an instruction has no valid sched class, do not attempt to check for a variant. 2013-04-13 06:07:45 +00:00
TwoAddressInstructionPass.cpp Remove unused #includes. 2013-03-05 01:00:45 +00:00
UnreachableBlockElim.cpp Move all of the header files which are involved in modelling the LLVM IR 2013-01-02 11:36:10 +00:00
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 lr, [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.