llvm-6502/lib/Target/Sparc
Duncan Sands f9516208e5 Revert the SelectionDAG optimization that makes
it impossible to create a MERGE_VALUES node with
only one result: sometimes it is useful to be able
to create a node with only one result out of one of
the results of a node with more than one result, for
example because the new node will eventually be used
to replace a one-result node using ReplaceAllUsesWith,
cf X86TargetLowering::ExpandFP_TO_SINT.  On the other
hand, most users of MERGE_VALUES don't need this and
for them the optimization was valuable.  So add a new
utility method getMergeValues for creating MERGE_VALUES
nodes which by default performs the optimization.
Change almost everywhere to use getMergeValues (and
tidy some stuff up at the same time).


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@52893 91177308-0d34-0410-b5e6-96231b3b80d8
2008-06-30 10:19:09 +00:00
..
DelaySlotFiller.cpp rename TargetInstrDescriptor -> TargetInstrDesc. 2008-01-07 07:27:27 +00:00
FPMover.cpp rename MachineInstr::setInstrDescriptor -> setDesc 2008-01-11 18:10:50 +00:00
Makefile Start moving sparc to use SparcCallingConv.td, switching over 2008-03-17 05:41:48 +00:00
README.txt fix CodeGen/Generic/2008-01-25-dag-combine-mul.ll on sparc, PR2105 2008-02-28 05:44:20 +00:00
Sparc.h Remove attribution from file headers, per discussion on llvmdev. 2007-12-29 20:36:04 +00:00
Sparc.td Start moving sparc to use SparcCallingConv.td, switching over 2008-03-17 05:41:48 +00:00
SparcAsmPrinter.cpp Add CommonLinkage; currently tentative definitions 2008-05-14 20:12:51 +00:00
SparcCallingConv.td Check in some #ifdef'd out code switching call argument 2008-03-17 06:58:37 +00:00
SparcInstrFormats.td Remove attribution from file headers, per discussion on llvmdev. 2007-12-29 20:36:04 +00:00
SparcInstrInfo.cpp It's not always safe to fold movsd into xorpd, etc. Check the alignment of the load address first to make sure it's 16 byte aligned. 2008-02-08 21:20:40 +00:00
SparcInstrInfo.h Change target-specific classes to use more precise static types. 2008-05-14 01:58:56 +00:00
SparcInstrInfo.td Replace all target specific implicit def instructions with a target independent one: TargetInstrInfo::IMPLICIT_DEF. 2008-03-15 00:03:38 +00:00
SparcISelDAGToDAG.cpp split sparc lowering out into SparcISelLowering.{cpp|h} to follow 2008-03-17 03:21:36 +00:00
SparcISelLowering.cpp Revert the SelectionDAG optimization that makes 2008-06-30 10:19:09 +00:00
SparcISelLowering.h Switch sparc from using LowerCallTo to using LowerOperation(CALL) like 2008-03-17 06:01:07 +00:00
SparcRegisterInfo.cpp Move reMaterialize() from TargetRegisterInfo to TargetInstrInfo. 2008-03-31 20:40:39 +00:00
SparcRegisterInfo.h Move reMaterialize() from TargetRegisterInfo to TargetInstrInfo. 2008-03-31 20:40:39 +00:00
SparcRegisterInfo.td Remove attribution from file headers, per discussion on llvmdev. 2007-12-29 20:36:04 +00:00
SparcSubtarget.cpp Remove attribution from file headers, per discussion on llvmdev. 2007-12-29 20:36:04 +00:00
SparcSubtarget.h Remove attribution from file headers, per discussion on llvmdev. 2007-12-29 20:36:04 +00:00
SparcTargetAsmInfo.cpp Remove attribution from file headers, per discussion on llvmdev. 2007-12-29 20:36:04 +00:00
SparcTargetAsmInfo.h Remove attribution from file headers, per discussion on llvmdev. 2007-12-29 20:36:04 +00:00
SparcTargetMachine.cpp Clean up the use of static and anonymous namespaces. This turned up 2008-05-13 00:00:25 +00:00
SparcTargetMachine.h Change target-specific classes to use more precise static types. 2008-05-14 01:58:56 +00:00

To-do
-----

* Keep the address of the constant pool in a register instead of forming its
  address all of the time.
* We can fold small constant offsets into the %hi/%lo references to constant
  pool addresses as well.
* When in V9 mode, register allocate %icc[0-3].
* Add support for isel'ing UMUL_LOHI instead of marking it as Expand.
* Emit the 'Branch on Integer Register with Prediction' instructions.  It's
  not clear how to write a pattern for this though:

float %t1(int %a, int* %p) {
        %C = seteq int %a, 0
        br bool %C, label %T, label %F
T:
        store int 123, int* %p
        br label %F
F:
        ret float undef
}

codegens to this:

t1:
        save -96, %o6, %o6
1)      subcc %i0, 0, %l0
1)      bne .LBBt1_2    ! F
        nop
.LBBt1_1:       ! T
        or %g0, 123, %l0
        st %l0, [%i1]
.LBBt1_2:       ! F
        restore %g0, %g0, %g0
        retl
        nop

1) should be replaced with a brz in V9 mode.

* Same as above, but emit conditional move on register zero (p192) in V9 
  mode.  Testcase:

int %t1(int %a, int %b) {
        %C = seteq int %a, 0
        %D = select bool %C, int %a, int %b
        ret int %D
}

* Emit MULX/[SU]DIVX instructions in V9 mode instead of fiddling 
  with the Y register, if they are faster.

* Codegen bswap(load)/store(bswap) -> load/store ASI

* Implement frame pointer elimination, e.g. eliminate save/restore for 
  leaf fns.
* Fill delay slots