[llvm] r235079 - TRUNCATE constant folding - minor fix for rL233224

[Mehdi Amini]

Hi Simon,

I have an out-of-tree target broken by this commits, and I have a feeling that the problem is in this change.

I have this node:  v2i32  BUILD_VECTOR { i32 = Constant<0>,  i32 = undef  }
That I try to truncate:
 getNode(ISD::TRUNCATE, dl, v2i16, N)

After your change, getNode() returns this node:

 N = v2i16  BUILD_VECTOR { i32 = Constant<0>,  i32 = undef  }

Note that the operands are not truncated, which breaks the DAG type system.
I assume you are relying on FoldConstantArithmetic() to magically convert the constant to the truncated type, however it does not work with undef.

I have the feeling that the proper way requires to wrap every operand with a TRUNCATE before calling getNode, but you may see an alternative?

Thanks,

Mehdi



> On Apr 16, 2015, at 1:21 AM, Simon Pilgrim <llvm-dev at redking.me.uk> wrote:
> 
> Author: rksimon
> Date: Thu Apr 16 03:21:09 2015
> New Revision: 235079
> 
> URL: http://llvm.org/viewvc/llvm-project?rev=235079&view=rev
> Log:
> TRUNCATE constant folding - minor fix for rL233224
> 
> Fix for test case found by James Molloy - TRUNCATE of constant build vectors can be more simply achieved by simply replacing with a new build vector node with the truncated value type - no need to touch the scalar operands at all.
> 
> Added:
>    llvm/trunk/test/CodeGen/AArch64/fold-constants.ll
> Modified:
>    llvm/trunk/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
> 
> Modified: llvm/trunk/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/CodeGen/SelectionDAG/SelectionDAG.cpp?rev=235079&r1=235078&r2=235079&view=diff
> ==============================================================================
> --- llvm/trunk/lib/CodeGen/SelectionDAG/SelectionDAG.cpp (original)
> +++ llvm/trunk/lib/CodeGen/SelectionDAG/SelectionDAG.cpp Thu Apr 16 03:21:09 2015
> @@ -2851,13 +2851,16 @@ SDValue SelectionDAG::getNode(unsigned O
>         // FIXME: Entirely reasonable to perform folding of other unary
>         // operations here as the need arises.
>         break;
> +      case ISD::TRUNCATE:
> +        // Constant build vector truncation can be done with the original scalar
> +        // operands but with a new build vector with the truncated value type.
> +        return getNode(ISD::BUILD_VECTOR, DL, VT, BV->ops());
>       case ISD::FNEG:
>       case ISD::FABS:
>       case ISD::FCEIL:
>       case ISD::FTRUNC:
>       case ISD::FFLOOR:
>       case ISD::FP_EXTEND:
> -      case ISD::TRUNCATE:
>       case ISD::UINT_TO_FP:
>       case ISD::SINT_TO_FP: {
>         // Let the above scalar folding handle the folding of each element.
> 
> Added: llvm/trunk/test/CodeGen/AArch64/fold-constants.ll
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/AArch64/fold-constants.ll?rev=235079&view=auto
> ==============================================================================
> --- llvm/trunk/test/CodeGen/AArch64/fold-constants.ll (added)
> +++ llvm/trunk/test/CodeGen/AArch64/fold-constants.ll Thu Apr 16 03:21:09 2015
> @@ -0,0 +1,21 @@
> +; RUN: llc -mtriple=aarch64-linux-gnu -o - %s | FileCheck %s
> +
> +define i64 @dotests_616() {
> +; CHECK-LABEL: dotests_616
> +; CHECK:       movi d0, #0000000000000000
> +; CHECK-NEXT:  umov w8, v0.b[2]
> +; CHECK-NEXT:  sbfx w8, w8, #0, #1
> +; CHECK-NEXT:  fmov s0, w8
> +; CHECK-NEXT:  fmov x0, d0
> +; CHECK-NEXT:  ret
> +entry:
> +  %0 = bitcast <2 x i64> zeroinitializer to <8 x i16>
> +  %1 = and <8 x i16> zeroinitializer, %0
> +  %2 = icmp ne <8 x i16> %1, zeroinitializer
> +  %3 = extractelement <8 x i1> %2, i32 2
> +  %vgetq_lane285 = sext i1 %3 to i16
> +  %vset_lane = insertelement <4 x i16> undef, i16 %vgetq_lane285, i32 0
> +  %4 = bitcast <4 x i16> %vset_lane to <1 x i64>
> +  %vget_lane = extractelement <1 x i64> %4, i32 0
> +  ret i64 %vget_lane
> +}
> 
> 
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits