llvm/lib/CodeGen
Tim Northover 5393254646 DAG: move sret demotion into most basic LowerCallTo implementation.
It looks like there are two versions of LowerCallTo here: the
SelectionDAGBuilder one is designed to operate on LLVM IR, and the
TargetLowering one in the case where everything is at DAG level.

Previously, only the SelectionDAGBuilder variant could handle demoting
an impossible return to sret semantics (before delegating to the
TargetLowering version), but this functionality is also useful for
certain libcalls (e.g. 128-bit operations on 32-bit x86).  So this
commit moves the sret handling down a level.

rdar://problem/17242889

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211155 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-18 11:52:44 +00:00
..
AsmPrinter PR20038: DebugInfo missing DIEs for some concrete variables. 2014-06-15 19:34:26 +00:00
SelectionDAG DAG: move sret demotion into most basic LowerCallTo implementation. 2014-06-18 11:52:44 +00:00
AggressiveAntiDepBreaker.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
AggressiveAntiDepBreaker.h [C++11] More 'nullptr' conversion. In some cases just using a boolean check instead of comparing to nullptr. 2014-04-16 04:21:27 +00:00
AllocationOrder.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
AllocationOrder.h
Analysis.cpp Have isInTailCallPosition take the DAG so that we can use the 2014-06-10 20:39:38 +00:00
AntiDepBreaker.h
AtomicExpandLoadLinkedPass.cpp Atomics: make use of the "cmpxchg weak" instruction. 2014-06-13 16:45:52 +00:00
BasicTargetTransformInfo.cpp Fix a spelling error 2014-05-08 13:42:57 +00:00
BranchFolding.cpp Prevent hoisting the instruction whose def might be clobbered by the terminator. 2014-06-05 13:42:48 +00:00
BranchFolding.h
CalcSpillWeights.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
CallingConvLower.cpp [C++11] More 'nullptr' conversion. In some cases just using a boolean check instead of comparing to nullptr. 2014-04-14 00:51:57 +00:00
CMakeLists.txt Move GlobalMerge from Transform to CodeGen. 2014-06-13 22:57:59 +00:00
CodeGen.cpp Atomics: promote ARM's IR-based atomics pass to CodeGen. 2014-04-17 18:22:47 +00:00
CodeGenPrepare.cpp Create macro INITIALIZE_TM_PASS. 2014-06-11 07:04:37 +00:00
CriticalAntiDepBreaker.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
CriticalAntiDepBreaker.h [C++11] Add 'override' keyword to virtual methods that override their base class. 2014-03-07 09:26:03 +00:00
DeadMachineInstructionElim.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
DFAPacketizer.cpp [C++11] More 'nullptr' conversion. In some cases just using a boolean check instead of comparing to nullptr. 2014-04-14 00:51:57 +00:00
DwarfEHPrepare.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
EarlyIfConversion.cpp Return false if we're not going to do anything. 2014-05-22 17:49:33 +00:00
EdgeBundles.cpp Convert several loops over MachineFunction basic blocks to range-based loops 2014-04-30 18:29:51 +00:00
ErlangGC.cpp [C++11] Add 'override' keyword to virtual methods that override their base class. 2014-03-07 09:26:03 +00:00
ExecutionDepsFix.cpp Clean up language and grammar. 2014-05-20 17:11:11 +00:00
ExpandISelPseudos.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
ExpandPostRAPseudos.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
GCMetadata.cpp Use unique_ptr for the result of Registry entries. 2014-04-15 05:53:26 +00:00
GCMetadataPrinter.cpp
GCStrategy.cpp Use unique_ptr to manage ownership of GCFunctionInfos in GCStrategy 2014-04-15 06:07:26 +00:00
GlobalMerge.cpp Move GlobalMerge from Transform to CodeGen. 2014-06-13 22:57:59 +00:00
IfConversion.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
InlineSpiller.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
InterferenceCache.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
InterferenceCache.h [C++11] More 'nullptr' conversion. In some cases just using a boolean check instead of comparing to nullptr. 2014-04-16 04:21:27 +00:00
IntrinsicLowering.cpp [C++11] More 'nullptr' conversion. In some cases just using a boolean check instead of comparing to nullptr. 2014-04-14 00:51:57 +00:00
JITCodeEmitter.cpp
JumpInstrTables.cpp [C++11] Use 'nullptr'. 2014-06-08 22:29:17 +00:00
LatencyPriorityQueue.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
LexicalScopes.cpp DebugInfo: Fix inlining with #file directives a little harder 2014-05-25 18:11:35 +00:00
LiveDebugVariables.cpp This reapplies r207235 with an additional bugfixes caught by the msan 2014-04-25 20:49:25 +00:00
LiveDebugVariables.h [C++11] Add 'override' keyword to virtual methods that override their base class. 2014-03-07 09:26:03 +00:00
LiveInterval.cpp [C++11] More 'nullptr' conversion. In some cases just using a boolean check instead of comparing to nullptr. 2014-04-14 00:51:57 +00:00
LiveIntervalAnalysis.cpp Calculate dead instructions when a live interval is created. 2014-06-03 22:42:10 +00:00
LiveIntervalUnion.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
LivePhysRegs.cpp
LiveRangeCalc.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
LiveRangeCalc.h [C++11] More 'nullptr' conversion. In some cases just using a boolean check instead of comparing to nullptr. 2014-04-16 04:21:27 +00:00
LiveRangeEdit.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
LiveRegMatrix.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
LiveStackAnalysis.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
LiveVariables.cpp Convert more loops to range-based equivalents 2014-04-30 22:17:38 +00:00
LLVMBuild.txt
LLVMTargetMachine.cpp Add a new attribute called 'jumptable' that creates jump-instruction tables for functions marked with this attribute. 2014-06-05 19:29:43 +00:00
LocalStackSlotAllocation.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
MachineBasicBlock.cpp [Modules] Make Support/Debug.h modular. This requires it to not change 2014-04-21 22:55:11 +00:00
MachineBlockFrequencyInfo.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
MachineBlockPlacement.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
MachineBranchProbabilityInfo.cpp [C++11] More 'nullptr' conversion. In some cases just using a boolean check instead of comparing to nullptr. 2014-04-14 00:51:57 +00:00
MachineCodeEmitter.cpp
MachineCopyPropagation.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
MachineCSE.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
MachineDominators.cpp
MachineFunction.cpp Convert several loops over MachineFunction basic blocks to range-based loops 2014-04-30 18:29:51 +00:00
MachineFunctionAnalysis.cpp [C++11] More 'nullptr' conversion. In some cases just using a boolean check instead of comparing to nullptr. 2014-04-14 00:51:57 +00:00
MachineFunctionPass.cpp
MachineFunctionPrinterPass.cpp [C++11] Add 'override' keyword to virtual methods that override their base class. 2014-03-07 09:26:03 +00:00
MachineInstr.cpp [BUG][REFACTOR] 2014-05-07 09:51:22 +00:00
MachineInstrBundle.cpp [C++11] Add 'override' keyword to virtual methods that override their base class. 2014-03-07 09:26:03 +00:00
MachineLICM.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
MachineLoopInfo.cpp
MachineModuleInfo.cpp [C++11] More 'nullptr' conversion. In some cases just using a boolean check instead of comparing to nullptr. 2014-04-14 00:51:57 +00:00
MachineModuleInfoImpls.cpp
MachinePassRegistry.cpp [C++11] More 'nullptr' conversion. In some cases just using a boolean check instead of comparing to nullptr. 2014-04-14 00:51:57 +00:00
MachinePostDominators.cpp
MachineRegisterInfo.cpp [C++11] More 'nullptr' conversion. In some cases just using a boolean check instead of comparing to nullptr. 2014-04-14 00:51:57 +00:00
MachineScheduler.cpp Fix the scheduler's MaxObservedStall computation. 2014-06-12 22:36:28 +00:00
MachineSink.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
MachineSSAUpdater.cpp [Modules] Make Support/Debug.h modular. This requires it to not change 2014-04-21 22:55:11 +00:00
MachineTraceMetrics.cpp Convert more loops to range-based equivalents 2014-04-30 22:17:38 +00:00
MachineVerifier.cpp MachineVerifier: Clean up some syntactic weirdness left behind by find&replace. 2014-05-24 13:31:10 +00:00
Makefile
module.modulemap [modules] Add module maps for LLVM. These are not quite ready for prime-time 2014-05-21 02:46:14 +00:00
OcamlGC.cpp
OptimizePHIs.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
Passes.cpp Add a subtarget hook: enablePostMachineScheduler. 2014-06-04 07:06:27 +00:00
PeepholeOptimizer.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
PHIElimination.cpp Convert more loops to range-based equivalents 2014-04-30 22:17:38 +00:00
PHIEliminationUtils.cpp Switch a number of loops in lib/CodeGen over to range-based for-loops, now that 2014-03-17 19:36:09 +00:00
PHIEliminationUtils.h
PostRASchedulerList.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
ProcessImplicitDefs.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
PrologEpilogInserter.cpp Fix typos 2014-06-07 21:23:09 +00:00
PrologEpilogInserter.h [C++11] Add 'override' keyword to virtual methods that override their base class. 2014-03-07 09:26:03 +00:00
PseudoSourceValue.cpp Break PseudoSourceValue out of the Value hierarchy. It is now the root of its own tree containing FixedStackPseudoSourceValue (which you can use isa/dyn_cast on) and MipsCallEntry (which you can't). Anything that needs to use either a PseudoSourceValue* and Value* is strongly encouraged to use a MachinePointerInfo instead. 2014-04-15 07:22:52 +00:00
README.txt
RegAllocBase.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
RegAllocBase.h [C++11] More 'nullptr' conversion. In some cases just using a boolean check instead of comparing to nullptr. 2014-04-16 04:21:27 +00:00
RegAllocBasic.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
RegAllocFast.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
RegAllocGreedy.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
RegAllocPBQP.cpp Convert more loops to range-based equivalents 2014-04-30 22:17:38 +00:00
RegisterClassInfo.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
RegisterCoalescer.cpp Convert more loops to range-based equivalents 2014-04-30 22:17:38 +00:00
RegisterCoalescer.h [C++11] More 'nullptr' conversion. In some cases just using a boolean check instead of comparing to nullptr. 2014-04-16 04:21:27 +00:00
RegisterPressure.cpp [C++11] More 'nullptr' conversion. In some cases just using a boolean check instead of comparing to nullptr. 2014-04-14 00:51:57 +00:00
RegisterScavenging.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
ScheduleDAG.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
ScheduleDAGInstrs.cpp Fix typos 2014-05-15 01:52:21 +00:00
ScheduleDAGPrinter.cpp
ScoreboardHazardRecognizer.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
ShadowStackGC.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
SjLjEHPrepare.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
SlotIndexes.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
Spiller.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
Spiller.h
SpillPlacement.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
SpillPlacement.h [C++11] More 'nullptr' conversion. In some cases just using a boolean check instead of comparing to nullptr. 2014-04-16 04:21:27 +00:00
SplitKit.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
SplitKit.h Clean up language and grammar. 2014-05-20 17:11:11 +00:00
StackColoring.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
StackMapLivenessAnalysis.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
StackMaps.cpp [Stackmaps] Pacify windows buildbot. 2014-05-01 22:39:26 +00:00
StackProtector.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
StackSlotColoring.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
TailDuplication.cpp Convert more loops to range-based equivalents 2014-04-30 22:17:38 +00:00
TargetFrameLoweringImpl.cpp
TargetInstrInfo.cpp The hazard recognizer only needs a subtarget, not a target machine 2014-06-13 22:38:52 +00:00
TargetLoweringBase.cpp IR: add "cmpxchg weak" variant to support permitted failure. 2014-06-13 14:24:07 +00:00
TargetLoweringObjectFileImpl.cpp Fix a few issues with comdat handling on COFF. 2014-06-06 19:26:12 +00:00
TargetOptionsImpl.cpp
TargetRegisterInfo.cpp [C++11] More 'nullptr' conversion. In some cases just using a boolean check instead of comparing to nullptr. 2014-04-14 00:51:57 +00:00
TargetSchedule.cpp
TwoAddressInstructionPass.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +00:00
UnreachableBlockElim.cpp [C++11] Add 'override' keyword to virtual methods that override their base class. 2014-03-07 09:26:03 +00:00
VirtRegMap.cpp [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE 2014-04-22 02:02:50 +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 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.