llvm-6502/lib/CodeGen
Evan Cheng 6edb0eac87 Teach machine sink to
1) Do forward copy propagation. This makes it easier to estimate the cost of the
   instruction being sunk.
2) Break critical edges on demand, including cases where the value is used by
   PHI nodes.
Critical edge splitting is not yet enabled by default.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114227 91177308-0d34-0410-b5e6-96231b3b80d8
2010-09-17 22:28:18 +00:00
..
AsmPrinter If FE forgot to provide a file name (usually it uses "stdin" as name in such situation) then make one up to ensure that debug info is not malformed. 2010-09-16 20:57:49 +00:00
PBQP Added initialisers for reduction rule counters. 2010-09-05 13:42:32 +00:00
SelectionDAG Check bb to ensure that alloca is in separate basic block. 2010-09-15 18:13:55 +00:00
AggressiveAntiDepBreaker.cpp Anti-dependency breaking needs to be careful not to use reserved regs 2010-09-02 17:12:55 +00:00
AggressiveAntiDepBreaker.h
Analysis.cpp
AntiDepBreaker.h
BranchFolding.cpp
BranchFolding.h
CalcSpillWeights.cpp
CallingConvLower.cpp
CMakeLists.txt Revert "CMake: Get rid of LLVMLibDeps.cmake and export the libraries normally." 2010-09-13 23:59:48 +00:00
CodePlacementOpt.cpp
CriticalAntiDepBreaker.cpp Fix a comment typo. 2010-09-10 22:42:21 +00:00
CriticalAntiDepBreaker.h
DeadMachineInstructionElim.cpp Track liveness of unallocatable, unreserved registers in machine DCE. 2010-08-31 21:51:05 +00:00
DwarfEHPrepare.cpp Reapply commit 112702 which was speculatively reverted by echristo. 2010-09-03 08:31:48 +00:00
ELF.h
ELFCodeEmitter.cpp
ELFCodeEmitter.h
ELFWriter.cpp
ELFWriter.h
GCMetadata.cpp zap dead code. 2010-09-04 18:12:00 +00:00
GCMetadataPrinter.cpp
GCStrategy.cpp
IfConversion.cpp Teach if-converter to be more careful with predicating instructions that would 2010-09-10 01:29:16 +00:00
InlineSpiller.cpp
IntrinsicLowering.cpp
LatencyPriorityQueue.cpp
LiveInterval.cpp Remove dead code. 2010-09-08 18:50:24 +00:00
LiveIntervalAnalysis.cpp
LiveStackAnalysis.cpp
LiveVariables.cpp
LLVMTargetMachine.cpp Stop using the dom frontier in DwarfEHPrepare by not promoting alloca's 2010-08-31 09:05:06 +00:00
LocalStackSlotAllocation.cpp Improve virtual frame base register allocation heuristics. 2010-08-31 17:58:19 +00:00
LowerSubregs.cpp
MachineBasicBlock.cpp
MachineCSE.cpp Machine CSE was forgetting to clear some data structures. 2010-09-17 21:59:42 +00:00
MachineDominators.cpp
MachineFunction.cpp
MachineFunctionAnalysis.cpp
MachineFunctionPass.cpp
MachineFunctionPrinterPass.cpp
MachineInstr.cpp
MachineLICM.cpp
MachineLoopInfo.cpp
MachineModuleInfo.cpp zap dead code. 2010-09-04 18:12:00 +00:00
MachineModuleInfoImpls.cpp
MachinePassRegistry.cpp
MachineRegisterInfo.cpp
MachineSink.cpp Teach machine sink to 2010-09-17 22:28:18 +00:00
MachineSSAUpdater.cpp
MachineVerifier.cpp
Makefile
ObjectCodeEmitter.cpp
OcamlGC.cpp
OptimizePHIs.cpp
Passes.cpp
PeepholeOptimizer.cpp must not peephole away side effects 2010-09-14 20:46:08 +00:00
PHIElimination.cpp
PHIElimination.h
PostRAHazardRecognizer.cpp Teach if-converter to be more careful with predicating instructions that would 2010-09-10 01:29:16 +00:00
PostRASchedulerList.cpp Teach if-converter to be more careful with predicating instructions that would 2010-09-10 01:29:16 +00:00
PreAllocSplitting.cpp
ProcessImplicitDefs.cpp
PrologEpilogInserter.cpp Simplify eliminateFrameIndex() interface back down now that PEI doesn't need 2010-08-26 23:32:16 +00:00
PrologEpilogInserter.h Simplify eliminateFrameIndex() interface back down now that PEI doesn't need 2010-08-26 23:32:16 +00:00
PseudoSourceValue.cpp
README.txt
RegAllocFast.cpp Add DEBUG message. 2010-09-10 20:32:09 +00:00
RegAllocLinearScan.cpp Tweak to ignoring reserved regs. The allocator was occasionally still looking 2010-09-01 22:48:34 +00:00
RegAllocPBQP.cpp Added support for register allocators to record which intervals are spill intervals, and where the uses and defs of the original intervals were in the original code. 2010-09-02 08:27:00 +00:00
RegisterCoalescer.cpp
RegisterScavenging.cpp The scavenger should just use getAllocatableSet() rather than reinventing it 2010-09-02 18:29:04 +00:00
RenderMachineFunction.cpp Added support for register allocators to record which intervals are spill intervals, and where the uses and defs of the original intervals were in the original code. 2010-09-02 08:27:00 +00:00
RenderMachineFunction.h Added support for register allocators to record which intervals are spill intervals, and where the uses and defs of the original intervals were in the original code. 2010-09-02 08:27:00 +00:00
ScheduleDAG.cpp
ScheduleDAGEmit.cpp
ScheduleDAGInstrs.cpp Teach if-converter to be more careful with predicating instructions that would 2010-09-10 01:29:16 +00:00
ScheduleDAGInstrs.h Teach if-converter to be more careful with predicating instructions that would 2010-09-10 01:29:16 +00:00
ScheduleDAGPrinter.cpp
ShadowStackGC.cpp
ShrinkWrapping.cpp
SimpleRegisterCoalescing.cpp Teach RemoveCopyByCommutingDef to check all aliases, not just subregisters. 2010-09-01 22:15:35 +00:00
SimpleRegisterCoalescing.h
SjLjEHPrepare.cpp
SlotIndexes.cpp
Spiller.cpp
Spiller.h
SplitKit.cpp Use the value mapping provided by LiveIntervalMap. This simplifies the code a 2010-09-16 00:01:36 +00:00
SplitKit.h Use the value mapping provided by LiveIntervalMap. This simplifies the code a 2010-09-16 00:01:36 +00:00
Splitter.cpp
Splitter.h
StackProtector.cpp
StackSlotColoring.cpp remove dead proto 2010-08-28 03:45:03 +00:00
StrongPHIElimination.cpp
TailDuplication.cpp
TargetInstrInfoImpl.cpp Teach if-converter to be more careful with predicating instructions that would 2010-09-10 01:29:16 +00:00
TargetLoweringObjectFileImpl.cpp two changes: 2010-08-30 18:12:35 +00:00
TwoAddressInstructionPass.cpp
UnreachableBlockElim.cpp
VirtRegMap.cpp
VirtRegMap.h
VirtRegRewriter.cpp Don't add <imp-def> operands during register rewriting. 2010-09-07 22:38:45 +00:00
VirtRegRewriter.h

//===---------------------------------------------------------------------===//

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 likelyhood 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 improvments:

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.