Fix PR14161

- Check index being extracted to be constant 0 before simplfiying.
  Otherwise, retain the original sequence.

llvm-svn: 166504

1 files changed, 38 insertions, 0 deletions

diff --git a/llvm/test/CodeGen/X86/pr14161.ll b/llvm/test/CodeGen/X86/pr14161.ll
new file mode 100644
index 00000000000..ff4532eac3a
--- /dev/null
+++ b/llvm/test/CodeGen/X86/pr14161.ll
@@ -0,0 +1,38 @@
+; RUN: llc < %s -mtriple=x86_64-linux-pc -mcpu=corei7 | FileCheck %s
+
+declare <4 x i32> @llvm.x86.sse41.pminud(<4 x i32>, <4 x i32>)
+
+define <2 x i16> @good(<4 x i32>*, <4 x i8>*) {
+entry:
+  %2 = load <4 x i32>* %0, align 16
+  %3 = call <4 x i32> @llvm.x86.sse41.pminud(<4 x i32> %2, <4 x i32> <i32 127, i32 127, i32 127, i32 127>)
+  %4 = extractelement <4 x i32> %3, i32 0
+  %5 = extractelement <4 x i32> %3, i32 1
+  %6 = extractelement <4 x i32> %3, i32 2
+  %7 = extractelement <4 x i32> %3, i32 3
+  %8 = bitcast i32 %4 to <2 x i16>
+  %9 = bitcast i32 %5 to <2 x i16>
+  ret <2 x i16> %8
+; CHECK: good
+; CHECK: pminud
+; CHECK-NEXT: pmovzxwq
+; CHECK: ret
+}
+
+define <2 x i16> @bad(<4 x i32>*, <4 x i8>*) {
+entry:
+  %2 = load <4 x i32>* %0, align 16
+  %3 = call <4 x i32> @llvm.x86.sse41.pminud(<4 x i32> %2, <4 x i32> <i32 127, i32 127, i32 127, i32 127>)
+  %4 = extractelement <4 x i32> %3, i32 0
+  %5 = extractelement <4 x i32> %3, i32 1
+  %6 = extractelement <4 x i32> %3, i32 2
+  %7 = extractelement <4 x i32> %3, i32 3
+  %8 = bitcast i32 %4 to <2 x i16>
+  %9 = bitcast i32 %5 to <2 x i16>
+  ret <2 x i16> %9
+; CHECK: bad
+; CHECK: pminud
+; CHECK: pextrd
+; CHECK: pmovzxwq
+; CHECK: ret
+}