llvm-6502/lib/CodeGen
Dan Gohman 87bad707ed Remove old DBG_LABEL code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@90669 91177308-0d34-0410-b5e6-96231b3b80d8
2009-12-05 17:56:26 +00:00
..
AsmPrinter Don't print a space before the : between the file name and line number. 2009-12-05 02:00:34 +00:00
PBQP
SelectionDAG Remove old DBG_LABEL code. 2009-12-05 17:56:26 +00:00
AggressiveAntiDepBreaker.cpp <rdar://problem/6721894>. Allow multiple registers to be renamed together (super and sub) if necessary to break an anti-dependence. 2009-11-24 00:59:08 +00:00
AggressiveAntiDepBreaker.h Remove some old experimental code that is no longer needed. Remove additional, speculative scheduling pass as its cost did not translate into significant performance improvement. Minor tweaks. 2009-11-20 19:32:48 +00:00
AntiDepBreaker.h Remove some old experimental code that is no longer needed. Remove additional, speculative scheduling pass as its cost did not translate into significant performance improvement. Minor tweaks. 2009-11-20 19:32:48 +00:00
BranchFolding.cpp Remove the target hook TargetInstrInfo::BlockHasNoFallThrough in favor of 2009-12-05 00:44:40 +00:00
BranchFolding.h Split tail duplication into a separate pass. This is needed to avoid 2009-11-26 00:32:21 +00:00
CMakeLists.txt Fix CMake makefiles 2009-12-02 22:19:31 +00:00
CodePlacementOpt.cpp improve portability to avoid conflicting with std::next in c++'0x. 2009-12-03 00:50:42 +00:00
CriticalAntiDepBreaker.cpp Remove some old experimental code that is no longer needed. Remove additional, speculative scheduling pass as its cost did not translate into significant performance improvement. Minor tweaks. 2009-11-20 19:32:48 +00:00
CriticalAntiDepBreaker.h Remove some old experimental code that is no longer needed. Remove additional, speculative scheduling pass as its cost did not translate into significant performance improvement. Minor tweaks. 2009-11-20 19:32:48 +00:00
DeadMachineInstructionElim.cpp
DwarfEHPrepare.cpp Pull LLVMContext out of PromoteMemToReg. 2009-11-23 03:50:44 +00:00
ELF.h
ELFCodeEmitter.cpp
ELFCodeEmitter.h
ELFWriter.cpp
ELFWriter.h
ExactHazardRecognizer.cpp
ExactHazardRecognizer.h
GCMetadata.cpp
GCMetadataPrinter.cpp
GCStrategy.cpp
IfConversion.cpp Allow target to disable if-converting predicable instructions. e.g. NEON instructions under ARM mode. 2009-11-21 06:20:26 +00:00
IntrinsicLowering.cpp
LatencyPriorityQueue.cpp Remove some old experimental code that is no longer needed. Remove additional, speculative scheduling pass as its cost did not translate into significant performance improvement. Minor tweaks. 2009-11-20 19:32:48 +00:00
LiveInterval.cpp
LiveIntervalAnalysis.cpp Temporarily revert r90502. It was causing the llvm-gcc bootstrap on PPC to fail. 2009-12-05 07:30:23 +00:00
LiveStackAnalysis.cpp
LiveVariables.cpp Move PHIElimination::isLiveOut method to LiveVariables. 2009-12-01 17:13:31 +00:00
LLVMTargetMachine.cpp Add a pre-regalloc tail duplication pass. 2009-12-04 09:42:45 +00:00
LowerSubregs.cpp improve portability to avoid conflicting with std::next in c++'0x. 2009-12-03 00:50:42 +00:00
MachineBasicBlock.cpp Simplify this code: don't call AnalyzeBranch before doing simpler checks. 2009-12-05 00:32:59 +00:00
MachineDominators.cpp
MachineFunction.cpp improve portability to avoid conflicting with std::next in c++'0x. 2009-12-03 00:50:42 +00:00
MachineFunctionAnalysis.cpp
MachineFunctionPass.cpp
MachineInstr.cpp Don't print the debug directory; it's often long and uninteresting. Omit 2009-12-05 00:20:51 +00:00
MachineLICM.cpp Enable hoisting load from constant memories. 2009-11-20 23:31:34 +00:00
MachineLoopInfo.cpp improve portability to avoid conflicting with std::next in c++'0x. 2009-12-03 00:50:42 +00:00
MachineModuleInfo.cpp Remove DebugLabelFolder pass. It is not used by dwarf writer anymore. 2009-11-24 19:37:07 +00:00
MachineModuleInfoImpls.cpp
MachinePassRegistry.cpp
MachineRegisterInfo.cpp
MachineSink.cpp
MachineSSAUpdater.cpp Don't try to be cute with undef optimization here. Let ProcessImplicitDefs handle it. 2009-12-04 09:23:37 +00:00
MachineVerifier.cpp Remove the target hook TargetInstrInfo::BlockHasNoFallThrough in favor of 2009-12-05 00:44:40 +00:00
MachO.h
MachOCodeEmitter.cpp
MachOCodeEmitter.h
MachOWriter.cpp
MachOWriter.h
Makefile
MaxStackAlignment.cpp Factor the stack alignment calculations out into a target independent pass. 2009-12-02 19:30:24 +00:00
ObjectCodeEmitter.cpp
OcamlGC.cpp
Passes.cpp
PHIElimination.cpp improve portability to avoid conflicting with std::next in c++'0x. 2009-12-03 00:50:42 +00:00
PHIElimination.h Move PHIElimination::isLiveOut method to LiveVariables. 2009-12-01 17:13:31 +00:00
PostRASchedulerList.cpp Don't hang on to pointers or references after vector::push_back. 2009-12-03 01:49:56 +00:00
PreAllocSplitting.cpp improve portability to avoid conflicting with std::next in c++'0x. 2009-12-03 00:50:42 +00:00
ProcessImplicitDefs.cpp When all defs of a vr are implicit_def, delete all of the defs. 2009-11-26 00:32:36 +00:00
PrologEpilogInserter.cpp improve portability to avoid conflicting with std::next in c++'0x. 2009-12-03 00:50:42 +00:00
PrologEpilogInserter.h
PseudoSourceValue.cpp
README.txt
RegAllocLinearScan.cpp Temporarily revert r90502. It was causing the llvm-gcc bootstrap on PPC to fail. 2009-12-05 07:30:23 +00:00
RegAllocLocal.cpp
RegAllocPBQP.cpp
RegisterCoalescer.cpp
RegisterScavenging.cpp improve portability to avoid conflicting with std::next in c++'0x. 2009-12-03 00:50:42 +00:00
ScheduleDAG.cpp Remove some old experimental code that is no longer needed. Remove additional, speculative scheduling pass as its cost did not translate into significant performance improvement. Minor tweaks. 2009-11-20 19:32:48 +00:00
ScheduleDAGEmit.cpp
ScheduleDAGInstrs.cpp
ScheduleDAGInstrs.h
ScheduleDAGPrinter.cpp Remove ShortNames from getNodeLabel in DOTGraphTraits 2009-11-30 12:38:47 +00:00
ShadowStackGC.cpp
ShrinkWrapping.cpp
SimpleHazardRecognizer.h
SimpleRegisterCoalescing.cpp improve portability to avoid conflicting with std::next in c++'0x. 2009-12-03 00:50:42 +00:00
SimpleRegisterCoalescing.h
SjLjEHPrepare.cpp
SlotIndexes.cpp
Spiller.cpp improve portability to avoid conflicting with std::next in c++'0x. 2009-12-03 00:50:42 +00:00
Spiller.h Removed references to LiveStacks from Spiller.* . They're no longer needed. 2009-11-20 00:53:30 +00:00
StackProtector.cpp
StackSlotColoring.cpp improve portability to avoid conflicting with std::next in c++'0x. 2009-12-03 00:50:42 +00:00
StrongPHIElimination.cpp
TailDuplication.cpp Handle recursive PHI's. 2009-12-04 19:09:10 +00:00
TargetInstrInfoImpl.cpp improve portability to avoid conflicting with std::next in c++'0x. 2009-12-03 00:50:42 +00:00
TwoAddressInstructionPass.cpp improve portability to avoid conflicting with std::next in c++'0x. 2009-12-03 00:50:42 +00:00
UnreachableBlockElim.cpp
VirtRegMap.cpp
VirtRegMap.h
VirtRegRewriter.cpp improve portability to avoid conflicting with std::next in c++'0x. 2009-12-03 00:50:42 +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.