[llvm] r357062 - [X86][SSE] Add shuffle test case for PR41249

[Simon Pilgrim via llvm-commits]

Author: rksimon
Date: Wed Mar 27 04:21:09 2019
New Revision: 357062

URL: http://llvm.org/viewvc/llvm-project?rev=357062&view=rev
Log:
[X86][SSE] Add shuffle test case for PR41249

Modified:
    llvm/trunk/test/CodeGen/X86/vector-shuffle-128-v4.ll

Modified: llvm/trunk/test/CodeGen/X86/vector-shuffle-128-v4.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/vector-shuffle-128-v4.ll?rev=357062&r1=357061&r2=357062&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/vector-shuffle-128-v4.ll (original)
+++ llvm/trunk/test/CodeGen/X86/vector-shuffle-128-v4.ll Wed Mar 27 04:21:09 2019
@@ -2339,6 +2339,43 @@ define <4 x i32> @insert_dup_mem_v4i32(i
   ret <4 x i32> %tmp2
 }
 
+; PR41249
+define <4 x float> @shuffle_mem_pmovzx_v4f32(<2 x float>* %p0, <4 x float>* %p1) {
+; SSE-LABEL: shuffle_mem_pmovzx_v4f32:
+; SSE:       # %bb.0:
+; SSE-NEXT:    movsd {{.*#+}} xmm0 = mem[0],zero
+; SSE-NEXT:    xorps %xmm1, %xmm1
+; SSE-NEXT:    movaps %xmm0, %xmm2
+; SSE-NEXT:    unpcklps {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1]
+; SSE-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,0,0,0]
+; SSE-NEXT:    movaps %xmm2, (%rsi)
+; SSE-NEXT:    retq
+;
+; AVX1-LABEL: shuffle_mem_pmovzx_v4f32:
+; AVX1:       # %bb.0:
+; AVX1-NEXT:    vmovsd {{.*#+}} xmm0 = mem[0],zero
+; AVX1-NEXT:    vxorps %xmm1, %xmm1, %xmm1
+; AVX1-NEXT:    vunpcklps {{.*#+}} xmm1 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; AVX1-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[0,0,0,0]
+; AVX1-NEXT:    vmovaps %xmm1, (%rsi)
+; AVX1-NEXT:    retq
+;
+; AVX2OR512VL-LABEL: shuffle_mem_pmovzx_v4f32:
+; AVX2OR512VL:       # %bb.0:
+; AVX2OR512VL-NEXT:    vmovsd {{.*#+}} xmm0 = mem[0],zero
+; AVX2OR512VL-NEXT:    vxorps %xmm1, %xmm1, %xmm1
+; AVX2OR512VL-NEXT:    vunpcklps {{.*#+}} xmm1 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; AVX2OR512VL-NEXT:    vbroadcastss %xmm0, %xmm0
+; AVX2OR512VL-NEXT:    vmovaps %xmm1, (%rsi)
+; AVX2OR512VL-NEXT:    retq
+  %1 = load <2 x float>, <2 x float>* %p0
+  %2 = shufflevector <2 x float> %1, <2 x float> undef, <4 x i32> <i32 undef, i32 undef, i32 0, i32 1>
+  %3 = shufflevector <4 x float> %2, <4 x float> <float undef, float undef, float 0.000000e+00, float 0.000000e+00>, <4 x i32> <i32 2, i32 6, i32 3, i32 7>
+  %4 = shufflevector <2 x float> %1, <2 x float> undef, <4 x i32> zeroinitializer
+  store <4 x float> %3, <4 x float>* %p1
+  ret <4 x float> %4
+}
+
 ;
 ; Shuffle to logical bit shifts
 ;