[PATCH] Add a target legalize hook for SplitVectorOperand

[Eli Friedman]

Missing testcase for the x86 change.

On Tue, Jul 23, 2013 at 5:10 AM, Justin Holewinski
<justin.holewinski at gmail.com> wrote:
>   Updated with X86 legalization patch by Eli Friedman
>
> http://llvm-reviews.chandlerc.com/D1195
>
> CHANGE SINCE LAST DIFF
>   http://llvm-reviews.chandlerc.com/D1195?vs=2942&id=2965#toc
>
> Files:
>   lib/CodeGen/SelectionDAG/LegalizeVectorTypes.cpp
>   lib/Target/X86/X86ISelLowering.cpp
>   test/CodeGen/NVPTX/vector-stores.ll
>
> Index: lib/CodeGen/SelectionDAG/LegalizeVectorTypes.cpp
> ===================================================================
> --- lib/CodeGen/SelectionDAG/LegalizeVectorTypes.cpp
> +++ lib/CodeGen/SelectionDAG/LegalizeVectorTypes.cpp
> @@ -1031,6 +1031,10 @@
>          dbgs() << "\n");
>    SDValue Res = SDValue();
>
> +  // See if the target wants to custom split this node.
> +  if (CustomLowerNode(N, N->getOperand(OpNo).getValueType(), false))
> +    return false;
> +
>    if (Res.getNode() == 0) {
>      switch (N->getOpcode()) {
>      default:
> Index: lib/Target/X86/X86ISelLowering.cpp
> ===================================================================
> --- lib/Target/X86/X86ISelLowering.cpp
> +++ lib/Target/X86/X86ISelLowering.cpp
> @@ -996,7 +996,7 @@
>      setLoadExtAction(ISD::EXTLOAD,              MVT::v2f32, Legal);
>    }
>
> -  if (Subtarget->hasSSE41()) {
> +  if (!TM.Options.UseSoftFloat && Subtarget->hasSSE41()) {
>      setOperationAction(ISD::FFLOOR,             MVT::f32,   Legal);
>      setOperationAction(ISD::FCEIL,              MVT::f32,   Legal);
>      setOperationAction(ISD::FTRUNC,             MVT::f32,   Legal);
> Index: test/CodeGen/NVPTX/vector-stores.ll
> ===================================================================
> --- /dev/null
> +++ test/CodeGen/NVPTX/vector-stores.ll
> @@ -0,0 +1,30 @@
> +; RUN: llc < %s -march=nvptx -mcpu=sm_20 | FileCheck %s
> +
> +; CHECK: .visible .func foo1
> +; CHECK: st.v2.f32
> +define void @foo1(<2 x float> %val, <2 x float>* %ptr) {
> +  store <2 x float> %val, <2 x float>* %ptr
> +  ret void
> +}
> +
> +; CHECK: .visible .func foo2
> +; CHECK: st.v4.f32
> +define void @foo2(<4 x float> %val, <4 x float>* %ptr) {
> +  store <4 x float> %val, <4 x float>* %ptr
> +  ret void
> +}
> +
> +; CHECK: .visible .func foo3
> +; CHECK: st.v2.u32
> +define void @foo3(<2 x i32> %val, <2 x i32>* %ptr) {
> +  store <2 x i32> %val, <2 x i32>* %ptr
> +  ret void
> +}
> +
> +; CHECK: .visible .func foo4
> +; CHECK: st.v4.u32
> +define void @foo4(<4 x i32> %val, <4 x i32>* %ptr) {
> +  store <4 x i32> %val, <4 x i32>* %ptr
> +  ret void
> +}
> +
>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
>