Add NEON 'laned' operations. This fixes another bunch of gcc testsuite fails and

makes the code faster.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@81220 91177308-0d34-0410-b5e6-96231b3b80d8

9 files changed, 367 insertions, 0 deletions

diff --git a/test/CodeGen/ARM/vmlal_lane.ll b/test/CodeGen/ARM/vmlal_lane.ll
new file mode 100644
index 0000000..5bb0621
--- /dev/null
+++ b/test/CodeGen/ARM/vmlal_lane.ll
@@ -0,0 +1,47 @@
+; RUN: llc -mattr=+neon < %s | FileCheck %s
+target datalayout = "e-p:32:32:32-i1:8:32-i8:8:32-i16:16:32-i32:32:32-i64:32:32-f32:32:32-f64:32:32-v64:64:64-v128:128:128-a0:0:32"
+target triple = "thumbv7-elf"
+
+define arm_aapcs_vfpcc <4 x i32> @test_vmlal_lanes16(<4 x i32> %arg0_int32x4_t, <4 x i16> %arg1_int16x4_t, <4 x i16> %arg2_int16x4_t) nounwind readnone {
+entry:
+; CHECK: test_vmlal_lanes16
+; CHECK: vmlal.s16 q0, d2, d3[1]
+  %0 = shufflevector <4 x i16> %arg2_int16x4_t, <4 x i16> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1> ; <<4 x i16>> [#uses=1]
+  %1 = tail call <4 x i32> @llvm.arm.neon.vmlals.v4i32(<4 x i32> %arg0_int32x4_t, <4 x i16> %arg1_int16x4_t, <4 x i16> %0) ; <<4 x i32>> [#uses=1]
+  ret <4 x i32> %1
+}
+
+declare <4 x i32> @llvm.arm.neon.vmlals.v4i32(<4 x i32>, <4 x i16>, <4 x i16>) nounwind readnone
+
+define arm_aapcs_vfpcc <2 x i64> @test_vmlal_lanes32(<2 x i64> %arg0_int64x2_t, <2 x i32> %arg1_int32x2_t, <2 x i32> %arg2_int32x2_t) nounwind readnone {
+entry:
+; CHECK: test_vmlal_lanes32
+; CHECK: vmlal.s32 q0, d2, d3[1]
+  %0 = shufflevector <2 x i32> %arg2_int32x2_t, <2 x i32> undef, <2 x i32> <i32 1, i32 1> ; <<2 x i32>> [#uses=1]
+  %1 = tail call <2 x i64> @llvm.arm.neon.vmlals.v2i64(<2 x i64> %arg0_int64x2_t, <2 x i32> %arg1_int32x2_t, <2 x i32> %0) ; <<2 x i64>> [#uses=1]
+  ret <2 x i64> %1
+}
+
+declare <2 x i64> @llvm.arm.neon.vmlals.v2i64(<2 x i64>, <2 x i32>, <2 x i32>) nounwind readnone
+
+define arm_aapcs_vfpcc <4 x i32> @test_vmlal_laneu16(<4 x i32> %arg0_uint32x4_t, <4 x i16> %arg1_uint16x4_t, <4 x i16> %arg2_uint16x4_t) nounwind readnone {
+entry:
+; CHECK: test_vmlal_laneu16
+; CHECK: vmlal.u16 q0, d2, d3[1]
+  %0 = shufflevector <4 x i16> %arg2_uint16x4_t, <4 x i16> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1> ; <<4 x i16>> [#uses=1]
+  %1 = tail call <4 x i32> @llvm.arm.neon.vmlalu.v4i32(<4 x i32> %arg0_uint32x4_t, <4 x i16> %arg1_uint16x4_t, <4 x i16> %0) ; <<4 x i32>> [#uses=1]
+  ret <4 x i32> %1
+}
+
+declare <4 x i32> @llvm.arm.neon.vmlalu.v4i32(<4 x i32>, <4 x i16>, <4 x i16>) nounwind readnone
+
+define arm_aapcs_vfpcc <2 x i64> @test_vmlal_laneu32(<2 x i64> %arg0_uint64x2_t, <2 x i32> %arg1_uint32x2_t, <2 x i32> %arg2_uint32x2_t) nounwind readnone {
+entry:
+; CHECK: test_vmlal_laneu32
+; CHECK: vmlal.u32 q0, d2, d3[1]
+  %0 = shufflevector <2 x i32> %arg2_uint32x2_t, <2 x i32> undef, <2 x i32> <i32 1, i32 1> ; <<2 x i32>> [#uses=1]
+  %1 = tail call <2 x i64> @llvm.arm.neon.vmlalu.v2i64(<2 x i64> %arg0_uint64x2_t, <2 x i32> %arg1_uint32x2_t, <2 x i32> %0) ; <<2 x i64>> [#uses=1]
+  ret <2 x i64> %1
+}
+
+declare <2 x i64> @llvm.arm.neon.vmlalu.v2i64(<2 x i64>, <2 x i32>, <2 x i32>) nounwind readnone
diff --git a/test/CodeGen/ARM/vmlsl_lane.ll b/test/CodeGen/ARM/vmlsl_lane.ll
new file mode 100644
index 0000000..1effbd6
--- /dev/null
+++ b/test/CodeGen/ARM/vmlsl_lane.ll
@@ -0,0 +1,47 @@
+; RUN: llc -mattr=+neon < %s | FileCheck %s
+target datalayout = "e-p:32:32:32-i1:8:32-i8:8:32-i16:16:32-i32:32:32-i64:32:32-f32:32:32-f64:32:32-v64:64:64-v128:128:128-a0:0:32"
+target triple = "thumbv7-elf"
+
+define arm_aapcs_vfpcc <4 x i32> @test_vmlsl_lanes16(<4 x i32> %arg0_int32x4_t, <4 x i16> %arg1_int16x4_t, <4 x i16> %arg2_int16x4_t) nounwind readnone {
+entry:
+; CHECK: test_vmlsl_lanes16
+; CHECK: vmlsl.s16 q0, d2, d3[1]
+  %0 = shufflevector <4 x i16> %arg2_int16x4_t, <4 x i16> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1> ; <<4 x i16>> [#uses=1]
+  %1 = tail call <4 x i32> @llvm.arm.neon.vmlsls.v4i32(<4 x i32> %arg0_int32x4_t, <4 x i16> %arg1_int16x4_t, <4 x i16> %0) ; <<4 x i32>> [#uses=1]
+  ret <4 x i32> %1
+}
+
+declare <4 x i32> @llvm.arm.neon.vmlsls.v4i32(<4 x i32>, <4 x i16>, <4 x i16>) nounwind readnone
+
+define arm_aapcs_vfpcc <2 x i64> @test_vmlsl_lanes32(<2 x i64> %arg0_int64x2_t, <2 x i32> %arg1_int32x2_t, <2 x i32> %arg2_int32x2_t) nounwind readnone {
+entry:
+; CHECK: test_vmlsl_lanes32
+; CHECK: vmlsl.s32 q0, d2, d3[1]
+  %0 = shufflevector <2 x i32> %arg2_int32x2_t, <2 x i32> undef, <2 x i32> <i32 1, i32 1> ; <<2 x i32>> [#uses=1]
+  %1 = tail call <2 x i64> @llvm.arm.neon.vmlsls.v2i64(<2 x i64> %arg0_int64x2_t, <2 x i32> %arg1_int32x2_t, <2 x i32> %0) ; <<2 x i64>> [#uses=1]
+  ret <2 x i64> %1
+}
+
+declare <2 x i64> @llvm.arm.neon.vmlsls.v2i64(<2 x i64>, <2 x i32>, <2 x i32>) nounwind readnone
+
+define arm_aapcs_vfpcc <4 x i32> @test_vmlsl_laneu16(<4 x i32> %arg0_uint32x4_t, <4 x i16> %arg1_uint16x4_t, <4 x i16> %arg2_uint16x4_t) nounwind readnone {
+entry:
+; CHECK: test_vmlsl_laneu16
+; CHECK: vmlsl.u16 q0, d2, d3[1]
+  %0 = shufflevector <4 x i16> %arg2_uint16x4_t, <4 x i16> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1> ; <<4 x i16>> [#uses=1]
+  %1 = tail call <4 x i32> @llvm.arm.neon.vmlslu.v4i32(<4 x i32> %arg0_uint32x4_t, <4 x i16> %arg1_uint16x4_t, <4 x i16> %0) ; <<4 x i32>> [#uses=1]
+  ret <4 x i32> %1
+}
+
+declare <4 x i32> @llvm.arm.neon.vmlslu.v4i32(<4 x i32>, <4 x i16>, <4 x i16>) nounwind readnone
+
+define arm_aapcs_vfpcc <2 x i64> @test_vmlsl_laneu32(<2 x i64> %arg0_uint64x2_t, <2 x i32> %arg1_uint32x2_t, <2 x i32> %arg2_uint32x2_t) nounwind readnone {
+entry:
+; CHECK: test_vmlsl_laneu32
+; CHECK: vmlsl.u32 q0, d2, d3[1]
+  %0 = shufflevector <2 x i32> %arg2_uint32x2_t, <2 x i32> undef, <2 x i32> <i32 1, i32 1> ; <<2 x i32>> [#uses=1]
+  %1 = tail call <2 x i64> @llvm.arm.neon.vmlslu.v2i64(<2 x i64> %arg0_uint64x2_t, <2 x i32> %arg1_uint32x2_t, <2 x i32> %0) ; <<2 x i64>> [#uses=1]
+  ret <2 x i64> %1
+}
+
+declare <2 x i64> @llvm.arm.neon.vmlslu.v2i64(<2 x i64>, <2 x i32>, <2 x i32>) nounwind readnone
diff --git a/test/CodeGen/ARM/vmul_lane.ll b/test/CodeGen/ARM/vmul_lane.ll
new file mode 100644
index 0000000..7edd873
--- /dev/null
+++ b/test/CodeGen/ARM/vmul_lane.ll
@@ -0,0 +1,57 @@
+; RUN: llc -mattr=+neon < %s | FileCheck %s
+target datalayout = "e-p:32:32:32-i1:8:32-i8:8:32-i16:16:32-i32:32:32-i64:32:32-f32:32:32-f64:32:32-v64:64:64-v128:128:128-a0:0:32"
+target triple = "thumbv7-elf"
+
+define arm_aapcs_vfpcc <2 x float> @test_vmul_lanef32(<2 x float> %arg0_float32x2_t, <2 x float> %arg1_float32x2_t) nounwind readnone {
+entry:
+; CHECK: test_vmul_lanef32:
+; CHECK: vmul.f32 d0, d0, d1[0]
+  %0 = shufflevector <2 x float> %arg1_float32x2_t, <2 x float> undef, <2 x i32> zeroinitializer ; <<2 x float>> [#uses=1]
+  %1 = fmul <2 x float> %0, %arg0_float32x2_t     ; <<2 x float>> [#uses=1]
+  ret <2 x float> %1
+}
+
+define arm_aapcs_vfpcc <4 x i16> @test_vmul_lanes16(<4 x i16> %arg0_int16x4_t, <4 x i16> %arg1_int16x4_t) nounwind readnone {
+entry:
+; CHECK: test_vmul_lanes16:
+; CHECK: vmul.i16 d0, d0, d1[1]
+  %0 = shufflevector <4 x i16> %arg1_int16x4_t, <4 x i16> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1> ; <<4 x i16>> [#uses$
+  %1 = mul <4 x i16> %0, %arg0_int16x4_t          ; <<4 x i16>> [#uses=1]
+  ret <4 x i16> %1
+}
+
+define arm_aapcs_vfpcc <2 x i32> @test_vmul_lanes32(<2 x i32> %arg0_int32x2_t, <2 x i32> %arg1_int32x2_t) nounwind readnone {
+entry:
+; CHECK: test_vmul_lanes32:
+; CHECK: vmul.i32 d0, d0, d1[1]
+  %0 = shufflevector <2 x i32> %arg1_int32x2_t, <2 x i32> undef, <2 x i32> <i32 1, i32 1> ; <<2 x i32>> [#uses=1]
+  %1 = mul <2 x i32> %0, %arg0_int32x2_t          ; <<2 x i32>> [#uses=1]
+  ret <2 x i32> %1
+}
+
+define arm_aapcs_vfpcc <4 x float> @test_vmulQ_lanef32(<4 x float> %arg0_float32x4_t, <2 x float> %arg1_float32x2_t) nounwind readnone {
+entry:
+; CHECK: test_vmulQ_lanef32:
+; CHECK: vmul.f32 q0, q0, d2[1]
+  %0 = shufflevector <2 x float> %arg1_float32x2_t, <2 x float> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1> ; <<4 x float>$
+  %1 = fmul <4 x float> %0, %arg0_float32x4_t     ; <<4 x float>> [#uses=1]
+  ret <4 x float> %1
+}
+
+define arm_aapcs_vfpcc <8 x i16> @test_vmulQ_lanes16(<8 x i16> %arg0_int16x8_t, <4 x i16> %arg1_int16x4_t) nounwind readnone {
+entry:
+; CHECK: test_vmulQ_lanes16:
+; CHECK: vmul.i16 q0, q0, d2[1]
+  %0 = shufflevector <4 x i16> %arg1_int16x4_t, <4 x i16> undef, <8 x i32> <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
+  %1 = mul <8 x i16> %0, %arg0_int16x8_t          ; <<8 x i16>> [#uses=1]
+  ret <8 x i16> %1
+}
+
+define arm_aapcs_vfpcc <4 x i32> @test_vmulQ_lanes32(<4 x i32> %arg0_int32x4_t, <2 x i32> %arg1_int32x2_t) nounwind readnone {
+entry:
+; CHECK: test_vmulQ_lanes32:
+; CHECK: vmul.i32 q0, q0, d2[1]
+  %0 = shufflevector <2 x i32> %arg1_int32x2_t, <2 x i32> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1> ; <<4 x i32>> [#uses$
+  %1 = mul <4 x i32> %0, %arg0_int32x4_t          ; <<4 x i32>> [#uses=1]
+  ret <4 x i32> %1
+}
diff --git a/test/CodeGen/ARM/vmull_lane.ll b/test/CodeGen/ARM/vmull_lane.ll
new file mode 100644
index 0000000..72cb3b1
--- /dev/null
+++ b/test/CodeGen/ARM/vmull_lane.ll
@@ -0,0 +1,47 @@
+; RUN: llc -mattr=+neon < %s | FileCheck %s
+target datalayout = "e-p:32:32:32-i1:8:32-i8:8:32-i16:16:32-i32:32:32-i64:32:32-f32:32:32-f64:32:32-v64:64:64-v128:128:128-a0:0:32"
+target triple = "thumbv7-elf"
+
+define arm_aapcs_vfpcc <4 x i32> @test_vmull_lanes16(<4 x i16> %arg0_int16x4_t, <4 x i16> %arg1_int16x4_t) nounwind readnone {
+entry:
+; CHECK: test_vmull_lanes16
+; CHECK: vmull.s16 q0, d0, d1[1]
+  %0 = shufflevector <4 x i16> %arg1_int16x4_t, <4 x i16> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1> ; <<4 x i16>> [#uses=1]
+  %1 = tail call <4 x i32> @llvm.arm.neon.vmulls.v4i32(<4 x i16> %arg0_int16x4_t, <4 x i16> %0) ; <<4 x i32>> [#uses=1]
+  ret <4 x i32> %1
+}
+
+declare <4 x i32> @llvm.arm.neon.vmulls.v4i32(<4 x i16>, <4 x i16>) nounwind readnone
+
+define arm_aapcs_vfpcc <2 x i64> @test_vmull_lanes32(<2 x i32> %arg0_int32x2_t, <2 x i32> %arg1_int32x2_t) nounwind readnone {
+entry:
+; CHECK: test_vmull_lanes32
+; CHECK: vmull.s32 q0, d0, d1[1]
+  %0 = shufflevector <2 x i32> %arg1_int32x2_t, <2 x i32> undef, <2 x i32> <i32 1, i32 1> ; <<2 x i32>> [#uses=1]
+  %1 = tail call <2 x i64> @llvm.arm.neon.vmulls.v2i64(<2 x i32> %arg0_int32x2_t, <2 x i32> %0) ; <<2 x i64>> [#uses=1]
+  ret <2 x i64> %1
+}
+
+declare <2 x i64> @llvm.arm.neon.vmulls.v2i64(<2 x i32>, <2 x i32>) nounwind readnone
+
+define arm_aapcs_vfpcc <4 x i32> @test_vmull_laneu16(<4 x i16> %arg0_uint16x4_t, <4 x i16> %arg1_uint16x4_t) nounwind readnone {
+entry:
+; CHECK: test_vmull_laneu16
+; CHECK: vmull.u16 q0, d0, d1[1]
+  %0 = shufflevector <4 x i16> %arg1_uint16x4_t, <4 x i16> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1> ; <<4 x i16>> [#uses=1]
+  %1 = tail call <4 x i32> @llvm.arm.neon.vmullu.v4i32(<4 x i16> %arg0_uint16x4_t, <4 x i16> %0) ; <<4 x i32>> [#uses=1]
+  ret <4 x i32> %1
+}
+
+declare <4 x i32> @llvm.arm.neon.vmullu.v4i32(<4 x i16>, <4 x i16>) nounwind readnone
+
+define arm_aapcs_vfpcc <2 x i64> @test_vmull_laneu32(<2 x i32> %arg0_uint32x2_t, <2 x i32> %arg1_uint32x2_t) nounwind readnone {
+entry:
+; CHECK: test_vmull_laneu32
+; CHECK: vmull.u32 q0, d0, d1[1]
+  %0 = shufflevector <2 x i32> %arg1_uint32x2_t, <2 x i32> undef, <2 x i32> <i32 1, i32 1> ; <<2 x i32>> [#uses=1]
+  %1 = tail call <2 x i64> @llvm.arm.neon.vmullu.v2i64(<2 x i32> %arg0_uint32x2_t, <2 x i32> %0) ; <<2 x i64>> [#uses=1]
+  ret <2 x i64> %1
+}
+
+declare <2 x i64> @llvm.arm.neon.vmullu.v2i64(<2 x i32>, <2 x i32>) nounwind readnone
diff --git a/test/CodeGen/ARM/vqRdmulh_lane.ll b/test/CodeGen/ARM/vqRdmulh_lane.ll
new file mode 100644
index 0000000..f308c52
--- /dev/null
+++ b/test/CodeGen/ARM/vqRdmulh_lane.ll
@@ -0,0 +1,47 @@
+; RUN: llc -mattr=+neon < %s | FileCheck %s
+target datalayout = "e-p:32:32:32-i1:8:32-i8:8:32-i16:16:32-i32:32:32-i64:32:32-f32:32:32-f64:32:32-v64:64:64-v128:128:128-a0:0:32"
+target triple = "thumbv7-elf"
+
+define arm_aapcs_vfpcc <8 x i16> @test_vqRdmulhQ_lanes16(<8 x i16> %arg0_int16x8_t, <4 x i16> %arg1_int16x4_t) nounwind readnone {
+entry:
+; CHECK: test_vqRdmulhQ_lanes16
+; CHECK: vqrdmulh.s16 q0, q0, d2[1]
+  %0 = shufflevector <4 x i16> %arg1_int16x4_t, <4 x i16> undef, <8 x i32> <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1> ; <<8 x i16>> [#uses=1]
+  %1 = tail call <8 x i16> @llvm.arm.neon.vqrdmulh.v8i16(<8 x i16> %arg0_int16x8_t, <8 x i16> %0) ; <<8 x i16>> [#uses=1]
+  ret <8 x i16> %1
+}
+
+declare <8 x i16> @llvm.arm.neon.vqrdmulh.v8i16(<8 x i16>, <8 x i16>) nounwind readnone
+
+define arm_aapcs_vfpcc <4 x i32> @test_vqRdmulhQ_lanes32(<4 x i32> %arg0_int32x4_t, <2 x i32> %arg1_int32x2_t) nounwind readnone {
+entry:
+; CHECK: test_vqRdmulhQ_lanes32
+; CHECK: vqrdmulh.s32 q0, q0, d2[1]
+  %0 = shufflevector <2 x i32> %arg1_int32x2_t, <2 x i32> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1> ; <<4 x i32>> [#uses=1]
+  %1 = tail call <4 x i32> @llvm.arm.neon.vqrdmulh.v4i32(<4 x i32> %arg0_int32x4_t, <4 x i32> %0) ; <<4 x i32>> [#uses=1]
+  ret <4 x i32> %1
+}
+
+declare <4 x i32> @llvm.arm.neon.vqrdmulh.v4i32(<4 x i32>, <4 x i32>) nounwind readnone
+
+define arm_aapcs_vfpcc <4 x i16> @test_vqRdmulh_lanes16(<4 x i16> %arg0_int16x4_t, <4 x i16> %arg1_int16x4_t) nounwind readnone {
+entry:
+; CHECK: test_vqRdmulh_lanes16
+; CHECK: vqrdmulh.s16 d0, d0, d1[1]
+  %0 = shufflevector <4 x i16> %arg1_int16x4_t, <4 x i16> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1> ; <<4 x i16>> [#uses=1]
+  %1 = tail call <4 x i16> @llvm.arm.neon.vqrdmulh.v4i16(<4 x i16> %arg0_int16x4_t, <4 x i16> %0) ; <<4 x i16>> [#uses=1]
+  ret <4 x i16> %1
+}
+
+declare <4 x i16> @llvm.arm.neon.vqrdmulh.v4i16(<4 x i16>, <4 x i16>) nounwind readnone
+
+define arm_aapcs_vfpcc <2 x i32> @test_vqRdmulh_lanes32(<2 x i32> %arg0_int32x2_t, <2 x i32> %arg1_int32x2_t) nounwind readnone {
+entry:
+; CHECK: test_vqRdmulh_lanes32
+; CHECK: vqrdmulh.s32 d0, d0, d1[1]
+  %0 = shufflevector <2 x i32> %arg1_int32x2_t, <2 x i32> undef, <2 x i32> <i32 1, i32 1> ; <<2 x i32>> [#uses=1]
+  %1 = tail call <2 x i32> @llvm.arm.neon.vqrdmulh.v2i32(<2 x i32> %arg0_int32x2_t, <2 x i32> %0) ; <<2 x i32>> [#uses=1]
+  ret <2 x i32> %1
+}
+
+declare <2 x i32> @llvm.arm.neon.vqrdmulh.v2i32(<2 x i32>, <2 x i32>) nounwind readnone
diff --git a/test/CodeGen/ARM/vqdmlal_lanes.ll b/test/CodeGen/ARM/vqdmlal_lanes.ll
new file mode 100644
index 0000000..ff532f3
--- /dev/null
+++ b/test/CodeGen/ARM/vqdmlal_lanes.ll
@@ -0,0 +1,25 @@
+; RUN: llc -mattr=+neon < %s | FileCheck %s
+target datalayout = "e-p:32:32:32-i1:8:32-i8:8:32-i16:16:32-i32:32:32-i64:32:32-f32:32:32-f64:32:32-v64:64:64-v128:128:128-a0:0:32"
+target triple = "thumbv7-elf"
+
+define arm_aapcs_vfpcc <4 x i32> @test_vqdmlal_lanes16(<4 x i32> %arg0_int32x4_t, <4 x i16> %arg1_int16x4_t, <4 x i16> %arg2_int16x4_t) nounwind readnone {
+entry:
+; CHECK: test_vqdmlal_lanes16
+; CHECK: vqdmlal.s16 q0, d2, d3[1]
+  %0 = shufflevector <4 x i16> %arg2_int16x4_t, <4 x i16> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1> ; <<4 x i16>> [#uses=1]
+  %1 = tail call <4 x i32> @llvm.arm.neon.vqdmlal.v4i32(<4 x i32> %arg0_int32x4_t, <4 x i16> %arg1_int16x4_t, <4 x i16> %0) ; <<4 x i32>> [#uses=1]
+  ret <4 x i32> %1
+}
+
+declare <4 x i32> @llvm.arm.neon.vqdmlal.v4i32(<4 x i32>, <4 x i16>, <4 x i16>) nounwind readnone
+
+define arm_aapcs_vfpcc <2 x i64> @test_vqdmlal_lanes32(<2 x i64> %arg0_int64x2_t, <2 x i32> %arg1_int32x2_t, <2 x i32> %arg2_int32x2_t) nounwind readnone {
+entry:
+; CHECK: test_vqdmlal_lanes32
+; CHECK: vqdmlal.s32 q0, d2, d3[1]
+  %0 = shufflevector <2 x i32> %arg2_int32x2_t, <2 x i32> undef, <2 x i32> <i32 1, i32 1> ; <<2 x i32>> [#uses=1]
+  %1 = tail call <2 x i64> @llvm.arm.neon.vqdmlal.v2i64(<2 x i64> %arg0_int64x2_t, <2 x i32> %arg1_int32x2_t, <2 x i32> %0) ; <<2 x i64>> [#uses=1]
+  ret <2 x i64> %1
+}
+
+declare <2 x i64> @llvm.arm.neon.vqdmlal.v2i64(<2 x i64>, <2 x i32>, <2 x i32>) nounwind readnone
diff --git a/test/CodeGen/ARM/vqdmlsl_lanes.ll b/test/CodeGen/ARM/vqdmlsl_lanes.ll
new file mode 100644
index 0000000..1a834ff
--- /dev/null
+++ b/test/CodeGen/ARM/vqdmlsl_lanes.ll
@@ -0,0 +1,25 @@
+; RUN: llc -mattr=+neon < %s | FileCheck %s
+target datalayout = "e-p:32:32:32-i1:8:32-i8:8:32-i16:16:32-i32:32:32-i64:32:32-f32:32:32-f64:32:32-v64:64:64-v128:128:128-a0:0:32"
+target triple = "thumbv7-elf"
+
+define arm_aapcs_vfpcc <4 x i32> @test_vqdmlsl_lanes16(<4 x i32> %arg0_int32x4_t, <4 x i16> %arg1_int16x4_t, <4 x i16> %arg2_int16x4_t) nounwind readnone {
+entry:
+; CHECK: test_vqdmlsl_lanes16
+; CHECK: vqdmlsl.s16 q0, d2, d3[1]
+  %0 = shufflevector <4 x i16> %arg2_int16x4_t, <4 x i16> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1> ; <<4 x i16>> [#uses=1]
+  %1 = tail call <4 x i32> @llvm.arm.neon.vqdmlsl.v4i32(<4 x i32> %arg0_int32x4_t, <4 x i16> %arg1_int16x4_t, <4 x i16> %0) ; <<4 x i32>> [#uses=1]
+  ret <4 x i32> %1
+}
+
+declare <4 x i32> @llvm.arm.neon.vqdmlsl.v4i32(<4 x i32>, <4 x i16>, <4 x i16>) nounwind readnone
+
+define arm_aapcs_vfpcc <2 x i64> @test_vqdmlsl_lanes32(<2 x i64> %arg0_int64x2_t, <2 x i32> %arg1_int32x2_t, <2 x i32> %arg2_int32x2_t) nounwind readnone {
+entry:
+; CHECK: test_vqdmlsl_lanes32
+; CHECK: vqdmlsl.s32 q0, d2, d3[1]
+  %0 = shufflevector <2 x i32> %arg2_int32x2_t, <2 x i32> undef, <2 x i32> <i32 1, i32 1> ; <<2 x i32>> [#uses=1]
+  %1 = tail call <2 x i64> @llvm.arm.neon.vqdmlsl.v2i64(<2 x i64> %arg0_int64x2_t, <2 x i32> %arg1_int32x2_t, <2 x i32> %0) ; <<2 x i64>> [#uses=1]
+  ret <2 x i64> %1
+}
+
+declare <2 x i64> @llvm.arm.neon.vqdmlsl.v2i64(<2 x i64>, <2 x i32>, <2 x i32>) nounwind readnone
diff --git a/test/CodeGen/ARM/vqdmulh_lane.ll b/test/CodeGen/ARM/vqdmulh_lane.ll
new file mode 100644
index 0000000..874f5f3
--- /dev/null
+++ b/test/CodeGen/ARM/vqdmulh_lane.ll
@@ -0,0 +1,47 @@
+; RUN: llc -mattr=+neon < %s | FileCheck %s
+target datalayout = "e-p:32:32:32-i1:8:32-i8:8:32-i16:16:32-i32:32:32-i64:32:32-f32:32:32-f64:32:32-v64:64:64-v128:128:128-a0:0:32"
+target triple = "thumbv7-elf"
+
+define arm_aapcs_vfpcc <8 x i16> @test_vqdmulhQ_lanes16(<8 x i16> %arg0_int16x8_t, <4 x i16> %arg1_int16x4_t) nounwind readnone {
+entry:
+; CHECK: test_vqdmulhQ_lanes16
+; CHECK: vqdmulh.s16 q0, q0, d2[1]
+  %0 = shufflevector <4 x i16> %arg1_int16x4_t, <4 x i16> undef, <8 x i32> <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1> ; <<8 x i16>> [#uses=1]
+  %1 = tail call <8 x i16> @llvm.arm.neon.vqdmulh.v8i16(<8 x i16> %arg0_int16x8_t, <8 x i16> %0) ; <<8 x i16>> [#uses=1]
+  ret <8 x i16> %1
+}
+
+declare <8 x i16> @llvm.arm.neon.vqdmulh.v8i16(<8 x i16>, <8 x i16>) nounwind readnone
+
+define arm_aapcs_vfpcc <4 x i32> @test_vqdmulhQ_lanes32(<4 x i32> %arg0_int32x4_t, <2 x i32> %arg1_int32x2_t) nounwind readnone {
+entry:
+; CHECK: test_vqdmulhQ_lanes32
+; CHECK: vqdmulh.s32 q0, q0, d2[1]
+  %0 = shufflevector <2 x i32> %arg1_int32x2_t, <2 x i32> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1> ; <<4 x i32>> [#uses=1]
+  %1 = tail call <4 x i32> @llvm.arm.neon.vqdmulh.v4i32(<4 x i32> %arg0_int32x4_t, <4 x i32> %0) ; <<4 x i32>> [#uses=1]
+  ret <4 x i32> %1
+}
+
+declare <4 x i32> @llvm.arm.neon.vqdmulh.v4i32(<4 x i32>, <4 x i32>) nounwind readnone
+
+define arm_aapcs_vfpcc <4 x i16> @test_vqdmulh_lanes16(<4 x i16> %arg0_int16x4_t, <4 x i16> %arg1_int16x4_t) nounwind readnone {
+entry:
+; CHECK: test_vqdmulh_lanes16
+; CHECK: vqdmulh.s16 d0, d0, d1[1]
+  %0 = shufflevector <4 x i16> %arg1_int16x4_t, <4 x i16> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1> ; <<4 x i16>> [#uses=1]
+  %1 = tail call <4 x i16> @llvm.arm.neon.vqdmulh.v4i16(<4 x i16> %arg0_int16x4_t, <4 x i16> %0) ; <<4 x i16>> [#uses=1]
+  ret <4 x i16> %1
+}
+
+declare <4 x i16> @llvm.arm.neon.vqdmulh.v4i16(<4 x i16>, <4 x i16>) nounwind readnone
+
+define arm_aapcs_vfpcc <2 x i32> @test_vqdmulh_lanes32(<2 x i32> %arg0_int32x2_t, <2 x i32> %arg1_int32x2_t) nounwind readnone {
+entry:
+; CHECK: test_vqdmulh_lanes32
+; CHECK: vqdmulh.s32 d0, d0, d1[1]
+  %0 = shufflevector <2 x i32> %arg1_int32x2_t, <2 x i32> undef, <2 x i32> <i32 1, i32 1> ; <<2 x i32>> [#uses=1]
+  %1 = tail call <2 x i32> @llvm.arm.neon.vqdmulh.v2i32(<2 x i32> %arg0_int32x2_t, <2 x i32> %0) ; <<2 x i32>> [#uses=1]
+  ret <2 x i32> %1
+}
+
+declare <2 x i32> @llvm.arm.neon.vqdmulh.v2i32(<2 x i32>, <2 x i32>) nounwind readnone
diff --git a/test/CodeGen/ARM/vqdmull_lane.ll b/test/CodeGen/ARM/vqdmull_lane.ll
new file mode 100644
index 0000000..21f4e94
--- /dev/null
+++ b/test/CodeGen/ARM/vqdmull_lane.ll
@@ -0,0 +1,25 @@
+; RUN: llc -mattr=+neon < %s | FileCheck %s
+target datalayout = "e-p:32:32:32-i1:8:32-i8:8:32-i16:16:32-i32:32:32-i64:32:32-f32:32:32-f64:32:32-v64:64:64-v128:128:128-a0:0:32"
+target triple = "thumbv7-elf"
+
+define arm_aapcs_vfpcc <4 x i32> @test_vqdmull_lanes16(<4 x i16> %arg0_int16x4_t, <4 x i16> %arg1_int16x4_t) nounwind readnone {
+entry:
+; CHECK: test_vqdmull_lanes16
+; CHECK: vqdmull.s16 q0, d0, d1[1]
+  %0 = shufflevector <4 x i16> %arg1_int16x4_t, <4 x i16> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1> ; <<4 x i16>> [#uses=1]
+  %1 = tail call <4 x i32> @llvm.arm.neon.vqdmull.v4i32(<4 x i16> %arg0_int16x4_t, <4 x i16> %0) ; <<4 x i32>> [#uses=1]
+  ret <4 x i32> %1
+}
+
+declare <4 x i32> @llvm.arm.neon.vqdmull.v4i32(<4 x i16>, <4 x i16>) nounwind readnone
+
+define arm_aapcs_vfpcc <2 x i64> @test_vqdmull_lanes32(<2 x i32> %arg0_int32x2_t, <2 x i32> %arg1_int32x2_t) nounwind readnone {
+entry:
+; CHECK: test_vqdmull_lanes32
+; CHECK: vqdmull.s32 q0, d0, d1[1]
+  %0 = shufflevector <2 x i32> %arg1_int32x2_t, <2 x i32> undef, <2 x i32> <i32 1, i32 1> ; <<2 x i32>> [#uses=1]
+  %1 = tail call <2 x i64> @llvm.arm.neon.vqdmull.v2i64(<2 x i32> %arg0_int32x2_t, <2 x i32> %0) ; <<2 x i64>> [#uses=1]
+  ret <2 x i64> %1
+}
+
+declare <2 x i64> @llvm.arm.neon.vqdmull.v2i64(<2 x i32>, <2 x i32>) nounwind readnone