[llvm] r345626 - [X86] In lowerVectorShuffleAsBroadcast, make peeking through CONCAT_VECTORS work correctly if we already walked through a bitcast that changed the element size.

[Craig Topper via llvm-commits]

Author: ctopper
Date: Tue Oct 30 11:48:42 2018
New Revision: 345626

URL: http://llvm.org/viewvc/llvm-project?rev=345626&view=rev
Log:
[X86] In lowerVectorShuffleAsBroadcast, make peeking through CONCAT_VECTORS work correctly if we already walked through a bitcast that changed the element size.

The CONCAT_VECTORS case was using the original mask element count to determine how to adjust the broadcast index. But if we looked through a bitcast the original mask size doesn't tell us anything about the concat_vectors.

This patch switchs to using the concat_vectors input element count directly instead.

Differential Revision: https://reviews.llvm.org/D53823

Modified:
    llvm/trunk/lib/Target/X86/X86ISelLowering.cpp
    llvm/trunk/test/CodeGen/X86/vector-shuffle-256-v8.ll

Modified: llvm/trunk/lib/Target/X86/X86ISelLowering.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86ISelLowering.cpp?rev=345626&r1=345625&r2=345626&view=diff
==============================================================================
--- llvm/trunk/lib/Target/X86/X86ISelLowering.cpp (original)
+++ llvm/trunk/lib/Target/X86/X86ISelLowering.cpp Tue Oct 30 11:48:42 2018
@@ -11238,7 +11238,8 @@ static SDValue lowerVectorShuffleAsBroad
       continue;
     }
     case ISD::CONCAT_VECTORS: {
-      int OperandSize = Mask.size() / V.getNumOperands();
+      int OperandSize =
+          V.getOperand(0).getSimpleValueType().getVectorNumElements();
       V = V.getOperand(BroadcastIdx / OperandSize);
       BroadcastIdx %= OperandSize;
       continue;

Modified: llvm/trunk/test/CodeGen/X86/vector-shuffle-256-v8.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/vector-shuffle-256-v8.ll?rev=345626&r1=345625&r2=345626&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/vector-shuffle-256-v8.ll (original)
+++ llvm/trunk/test/CodeGen/X86/vector-shuffle-256-v8.ll Tue Oct 30 11:48:42 2018
@@ -2848,3 +2848,32 @@ entry:
   %add = add <8 x i32> %shuffle, %shuffle1
   ret <8 x i32> %add
 }
+
+; This test used to crash due to bad handling of concat_vectors after a bitcast
+; in lowerVectorShuffleAsBroadcast.
+define <8 x float> @broadcast_concat_crash(<4 x float> %x, <4 x float> %y, float %z) {
+; AVX1-LABEL: broadcast_concat_crash:
+; AVX1:       # %bb.0: # %entry
+; AVX1-NEXT:    vpermilps {{.*#+}} xmm0 = xmm1[3,3,1,1]
+; AVX1-NEXT:    vinsertps {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[2,3]
+; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
+; AVX1-NEXT:    retq
+;
+; AVX2OR512VL-LABEL: broadcast_concat_crash:
+; AVX2OR512VL:       # %bb.0: # %entry
+; AVX2OR512VL-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; AVX2OR512VL-NEXT:    vpermpd {{.*#+}} ymm0 = ymm0[3,1,2,3]
+; AVX2OR512VL-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[1,1,3,3]
+; AVX2OR512VL-NEXT:    vinsertps {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[2,3]
+; AVX2OR512VL-NEXT:    vinsertf128 $1, %xmm0, %ymm0, %ymm0
+; AVX2OR512VL-NEXT:    retq
+entry:
+  %tmp = shufflevector <4 x float> %x, <4 x float> %y, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+  %bc = bitcast <8 x float> %tmp to <4 x i64>
+  %tmp1 = extractelement <4 x i64> %bc, i32 3
+  %tmp2 = bitcast i64 %tmp1 to <2 x float>
+  %tmp4 = extractelement <2 x float> %tmp2, i32 1
+  %tmp5 = insertelement <8 x float> undef, float %tmp4, i32 4
+  %tmp6 = insertelement <8 x float> %tmp5, float %z, i32 5
+  ret <8 x float> %tmp6
+}