[PATCH] D23002: DAGCombiner: check isZExtFree before doing combine

[Justin Bogner via llvm-commits]

escha via llvm-commits <llvm-commits at lists.llvm.org> writes:
> escha created this revision.
> escha added a reviewer: arsenm.
> escha added a subscriber: llvm-commits.
> escha set the repository for this revision to rL LLVM.
>
> No in-tree effect as of now (as far as I know), but is logical, and
> allows me to do some certain combines that would otherwise infinite
> loop.
>
> What the change does: doesn't convert zext(op(zext(X))) to op(zext(X))

LGTM. If zext is free this is pointless.

> Why: the larger op may be more expensive, and if the zext is free,
> there is no purpose to such a transform.
>
> Really why: I want a combine in our GPU tree that shrinks (and zero
> extends the result of) large arithmetic operations, because smaller
> operations take fewer registers and are often faster as well. This
> will infinite loop if any combines occur on (zext (OP (x)) that result
> in making OP bigger again.
>
> Repository:
>   rL LLVM
>
> https://reviews.llvm.org/D23002
>
> Files:
>   lib/CodeGen/SelectionDAG/DAGCombiner.cpp
>
> Index: lib/CodeGen/SelectionDAG/DAGCombiner.cpp
> ===================================================================
> --- lib/CodeGen/SelectionDAG/DAGCombiner.cpp
> +++ lib/CodeGen/SelectionDAG/DAGCombiner.cpp
> @@ -6309,7 +6309,8 @@
>    if ((N0.getOpcode() == ISD::SHL || N0.getOpcode() == ISD::SRL) &&
>        isa<ConstantSDNode>(N0.getOperand(1)) &&
>        N0.getOperand(0).getOpcode() == ISD::ZERO_EXTEND &&
> -      N0.hasOneUse()) {
> +      N0.hasOneUse() &&
> +      !TLI.isZExtFree(N0.getValueType(), VT)) {
>      SDValue ShAmt = N0.getOperand(1);
>      unsigned ShAmtVal = cast<ConstantSDNode>(ShAmt)->getZExtValue();
>      if (N0.getOpcode() == ISD::SHL) {