llvm/lib/Target/Sparc
Evan Cheng 9efce638d3 Allow copyRegToReg to emit cross register classes copies.
Tested with "make check"!


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@42346 91177308-0d34-0410-b5e6-96231b3b80d8
2007-09-26 06:25:56 +00:00
..
DelaySlotFiller.cpp Drop 'const' 2007-05-03 01:11:54 +00:00
FPMover.cpp Drop 'const' 2007-05-03 01:11:54 +00:00
Makefile
README.txt
Sparc.h
Sparc.td
SparcAsmPrinter.cpp Don't ignore the return value of AsmPrinter::doInitialization and 2007-07-25 19:33:14 +00:00
SparcInstrFormats.td Change instruction description to split OperandList into OutOperandList and 2007-07-19 01:14:50 +00:00
SparcInstrInfo.cpp Add lengthof and endof templates that hide a lot of sizeof computations. 2007-09-07 04:06:50 +00:00
SparcInstrInfo.h RemoveBranch() and InsertBranch() now returns number of instructions deleted / inserted. 2007-05-18 00:18:17 +00:00
SparcInstrInfo.td Remove (somewhat confusing) Imp<> helper, use let Defs = [], Uses = [] instead. 2007-09-11 19:55:27 +00:00
SparcISelDAGToDAG.cpp Fold the adjust_trampoline intrinsic into 2007-09-11 14:10:23 +00:00
SparcRegisterInfo.cpp Allow copyRegToReg to emit cross register classes copies. 2007-09-26 06:25:56 +00:00
SparcRegisterInfo.h Allow copyRegToReg to emit cross register classes copies. 2007-09-26 06:25:56 +00:00
SparcRegisterInfo.td Fix for PR1540: Specify F0, F1 are sub-registers of D0, etc. 2007-07-13 23:55:50 +00:00
SparcSubtarget.cpp
SparcSubtarget.h
SparcTargetAsmInfo.cpp
SparcTargetAsmInfo.h More explicit keywords. 2007-09-25 20:27:06 +00:00
SparcTargetMachine.cpp long double patch 2 of N. Handle it in TargetData. 2007-08-03 20:20:50 +00:00
SparcTargetMachine.h

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].
* 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