1 files changed, 19 insertions, 10 deletions
diff --git a/test/CodeGen/ARM/vshll.ll b/test/CodeGen/ARM/vshll.ll
index f81c09a7b9d31..8e85b98f49b10 100644
--- a/test/CodeGen/ARM/vshll.ll
+++ b/test/CodeGen/ARM/vshll.ll
@@ -1,45 +1,48 @@
-; RUN: llvm-as < %s | llc -march=arm -mattr=+neon > %t
-; RUN: grep {vshll\\.s8} %t | count 1
-; RUN: grep {vshll\\.s16} %t | count 1
-; RUN: grep {vshll\\.s32} %t | count 1
-; RUN: grep {vshll\\.u8} %t | count 1
-; RUN: grep {vshll\\.u16} %t | count 1
-; RUN: grep {vshll\\.u32} %t | count 1
-; RUN: grep {vshll\\.i8} %t | count 1
-; RUN: grep {vshll\\.i16} %t | count 1
-; RUN: grep {vshll\\.i32} %t | count 1
+; RUN: llc < %s -march=arm -mattr=+neon | FileCheck %s
 
 define <8 x i16> @vshlls8(<8 x i8>* %A) nounwind {
+;CHECK: vshlls8:
+;CHECK: vshll.s8
 	%tmp1 = load <8 x i8>* %A
 	%tmp2 = call <8 x i16> @llvm.arm.neon.vshiftls.v8i16(<8 x i8> %tmp1, <8 x i8> < i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7 >)
 	ret <8 x i16> %tmp2
 }
 
 define <4 x i32> @vshlls16(<4 x i16>* %A) nounwind {
+;CHECK: vshlls16:
+;CHECK: vshll.s16
 	%tmp1 = load <4 x i16>* %A
 	%tmp2 = call <4 x i32> @llvm.arm.neon.vshiftls.v4i32(<4 x i16> %tmp1, <4 x i16> < i16 15, i16 15, i16 15, i16 15 >)
 	ret <4 x i32> %tmp2
 }
 
 define <2 x i64> @vshlls32(<2 x i32>* %A) nounwind {
+;CHECK: vshlls32:
+;CHECK: vshll.s32
 	%tmp1 = load <2 x i32>* %A
 	%tmp2 = call <2 x i64> @llvm.arm.neon.vshiftls.v2i64(<2 x i32> %tmp1, <2 x i32> < i32 31, i32 31 >)
 	ret <2 x i64> %tmp2
 }
 
 define <8 x i16> @vshllu8(<8 x i8>* %A) nounwind {
+;CHECK: vshllu8:
+;CHECK: vshll.u8
 	%tmp1 = load <8 x i8>* %A
 	%tmp2 = call <8 x i16> @llvm.arm.neon.vshiftlu.v8i16(<8 x i8> %tmp1, <8 x i8> < i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7 >)
 	ret <8 x i16> %tmp2
 }
 
 define <4 x i32> @vshllu16(<4 x i16>* %A) nounwind {
+;CHECK: vshllu16:
+;CHECK: vshll.u16
 	%tmp1 = load <4 x i16>* %A
 	%tmp2 = call <4 x i32> @llvm.arm.neon.vshiftlu.v4i32(<4 x i16> %tmp1, <4 x i16> < i16 15, i16 15, i16 15, i16 15 >)
 	ret <4 x i32> %tmp2
 }
 
 define <2 x i64> @vshllu32(<2 x i32>* %A) nounwind {
+;CHECK: vshllu32:
+;CHECK: vshll.u32
 	%tmp1 = load <2 x i32>* %A
 	%tmp2 = call <2 x i64> @llvm.arm.neon.vshiftlu.v2i64(<2 x i32> %tmp1, <2 x i32> < i32 31, i32 31 >)
 	ret <2 x i64> %tmp2
@@ -48,18 +51,24 @@ define <2 x i64> @vshllu32(<2 x i32>* %A) nounwind {
 ; The following tests use the maximum shift count, so the signedness is
 ; irrelevant.  Test both signed and unsigned versions.
 define <8 x i16> @vshlli8(<8 x i8>* %A) nounwind {
+;CHECK: vshlli8:
+;CHECK: vshll.i8
 	%tmp1 = load <8 x i8>* %A
 	%tmp2 = call <8 x i16> @llvm.arm.neon.vshiftls.v8i16(<8 x i8> %tmp1, <8 x i8> < i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8, i8 8 >)
 	ret <8 x i16> %tmp2
 }
 
 define <4 x i32> @vshlli16(<4 x i16>* %A) nounwind {
+;CHECK: vshlli16:
+;CHECK: vshll.i16
 	%tmp1 = load <4 x i16>* %A
 	%tmp2 = call <4 x i32> @llvm.arm.neon.vshiftlu.v4i32(<4 x i16> %tmp1, <4 x i16> < i16 16, i16 16, i16 16, i16 16 >)
 	ret <4 x i32> %tmp2
 }
 
 define <2 x i64> @vshlli32(<2 x i32>* %A) nounwind {
+;CHECK: vshlli32:
+;CHECK: vshll.i32
 	%tmp1 = load <2 x i32>* %A
 	%tmp2 = call <2 x i64> @llvm.arm.neon.vshiftls.v2i64(<2 x i32> %tmp1, <2 x i32> < i32 32, i32 32 >)
 	ret <2 x i64> %tmp2