llvm-6502/lib/CodeGen
Evan Cheng 7f56625447 Local spiller optimization:
Turn this:
movswl  %ax, %eax
movl    %eax, -36(%ebp)
xorl    %edi, -36(%ebp)
into
movswl  %ax, %eax
xorl    %edi, %eax
movl    %eax, -36(%ebp)
by unfolding the load / store xorl into an xorl and a store when we know the
value in the spill slot is available in a register. This doesn't change the
number of instructions but reduce the number of times memory is accessed.

Also unfold some load folding instructions and reuse the value when similar
situation presents itself.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42947 91177308-0d34-0410-b5e6-96231b3b80d8
2007-10-13 02:50:24 +00:00
..
SelectionDAG Corrected many typing errors. And removed 'nest' parameter handling 2007-10-12 21:30:57 +00:00
AsmPrinter.cpp Change the names used for internal labels to use the current 2007-10-12 14:53:36 +00:00
BranchFolding.cpp More explicit keywords. 2007-08-02 21:21:54 +00:00
Collector.cpp Collector is the base class for garbage collection code generators. 2007-09-29 02:13:43 +00:00
CollectorMetadata.cpp CollectorMetadata abstractly describes stack maps for a function. 2007-09-27 22:18:46 +00:00
Collectors.cpp My previous Registry.h header, as well as Collectors.h, which is the 2007-09-27 19:34:27 +00:00
DwarfWriter.cpp Move the code that emits the .file directives so that it runs after the 2007-10-01 22:40:20 +00:00
ELFWriter.cpp Drop 'const' 2007-05-03 01:11:54 +00:00
ELFWriter.h Here is the bulk of the sanitizing. 2007-07-05 17:07:56 +00:00
IfConversion.cpp Somehow this wasn't committed last time. M_CLOBBERS_PRED is gone. 2007-07-10 17:50:43 +00:00
IntrinsicLowering.cpp Rewrite sqrt and powi to use anyfloat. By popular demand. 2007-10-02 17:43:59 +00:00
LiveInterval.cpp EXTRACT_SUBREG coalescing support. The coalescer now treats EXTRACT_SUBREG like 2007-10-12 08:50:34 +00:00
LiveIntervalAnalysis.cpp Did mean to leave this in. INSERT_SUBREG isn't being coalesced yet. 2007-10-12 17:16:50 +00:00
LiveVariables.cpp Remove isReg, isImm, and isMBB, and change all their users to use 2007-09-14 20:33:02 +00:00
LLVMTargetMachine.cpp Move subreg lowering pass to be right after regalloc, per feedback. 2007-07-27 07:36:14 +00:00
LowerSubregs.cpp Allow copyRegToReg to emit cross register classes copies. 2007-09-26 06:25:56 +00:00
MachineBasicBlock.cpp Use empty() member functions when that's what's being tested for instead 2007-10-03 19:26:29 +00:00
MachineFunction.cpp Use empty() member functions when that's what's being tested for instead 2007-10-03 19:26:29 +00:00
MachineInstr.cpp Optionally create a MachineInstr without default implicit operands. 2007-10-13 02:23:01 +00:00
MachineModuleInfo.cpp Fix PR1628. When exception handling is turned on, 2007-09-05 11:27:52 +00:00
MachinePassRegistry.cpp Final polish on machine pass registries. 2006-08-02 12:30:23 +00:00
MachOWriter.cpp Revise previous patch per review comments. 2007-09-12 03:30:33 +00:00
MachOWriter.h Drop 'const' 2007-05-03 01:11:54 +00:00
Makefile this will work better 2006-11-03 19:15:55 +00:00
Passes.cpp *** empty log message *** 2006-11-16 20:11:33 +00:00
PHIElimination.cpp Allow copyRegToReg to emit cross register classes copies. 2007-09-26 06:25:56 +00:00
PhysRegTracker.h Add explicit keywords and remove spurious trailing semicolons. 2007-08-27 14:50:10 +00:00
PostRASchedulerList.cpp Modify previous patch per review comments. 2007-07-13 17:31:29 +00:00
PrologEpilogInserter.cpp Long live the exception handling! 2007-07-14 14:06:15 +00:00
README.txt This is done already. 2007-09-29 02:23:08 +00:00
RegAllocBigBlock.cpp Merge DenseMapKeyInfo & DenseMapValueInfo into DenseMapInfo 2007-09-17 18:34:04 +00:00
RegAllocLinearScan.cpp Some clean up. 2007-10-12 08:45:27 +00:00
RegAllocLocal.cpp Correctly handle implcit def / use operands. 2007-06-26 21:05:13 +00:00
RegAllocSimple.cpp Drop 'const' 2007-05-03 01:11:54 +00:00
RegisterCoalescer.cpp Pluggable coalescers inplementation. 2007-09-06 16:18:45 +00:00
RegisterScavenging.cpp Remove isReg, isImm, and isMBB, and change all their users to use 2007-09-14 20:33:02 +00:00
SimpleRegisterCoalescing.cpp Restrict EXTRACT_SUBREG coalescing to avoid negative performance impact. 2007-10-12 09:15:53 +00:00
TwoAddressInstructionPass.cpp Allow copyRegToReg to emit cross register classes copies. 2007-09-26 06:25:56 +00:00
UnreachableBlockElim.cpp Fix typo in comment. 2007-05-06 13:37:16 +00:00
VirtRegMap.cpp Local spiller optimization: 2007-10-13 02:50:24 +00:00
VirtRegMap.h Local spiller optimization: 2007-10-13 02:50:24 +00:00

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

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.

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

I think we should have a "hasSideEffects" flag (which is automatically set for
stuff that "isLoad" "isCall" etc), and the remat pass should eventually be able
to remat any instruction that has no side effects, if it can handle it and if
profitable.

For now, I'd suggest having the remat stuff work like this:

1. I need to spill/reload this thing.
2. Check to see if it has side effects.
3. Check to see if it is simple enough: e.g. it only has one register
destination and no register input.
4. If so, clone the instruction, do the xform, etc.

Advantages of this are:

1. the .td file describes the behavior of the instructions, not the way the
   algorithm should work.
2. as remat gets smarter in the future, we shouldn't have to be changing the .td
   files.
3. it is easier to explain what the flag means in the .td file, because you
   don't have to pull in the explanation of how the current remat algo works.

Some potential added complexities:

1. Some instructions have to be glued to it's predecessor or successor. All of
   the PC relative instructions and condition code setting instruction. We could
   mark them as hasSideEffects, but that's not quite right. PC relative loads
   from constantpools can be remat'ed, for example. But it requires more than
   just cloning the instruction. Some instructions can be remat'ed but it
   expands to more than one instruction. But allocator will have to make a
   decision.

4. As stated in 3, not as simple as cloning in some cases. The target will have
   to decide how to remat it. For example, an ARM 2-piece constant generation
   instruction is remat'ed as a load from constantpool.

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

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.