diff --git a/llvm/lib/Target/ARM/ARMISelDAGToDAG.cpp b/llvm/lib/Target/ARM/ARMISelDAGToDAG.cpp
--- a/llvm/lib/Target/ARM/ARMISelDAGToDAG.cpp
+++ b/llvm/lib/Target/ARM/ARMISelDAGToDAG.cpp
@@ -1791,8 +1791,8 @@
   SDValue Ops[] = {Base, NewOffset,
                    CurDAG->getTargetConstant(Pred, SDLoc(N), MVT::i32), PredReg,
                    Chain};
-  SDNode *New = CurDAG->getMachineNode(Opcode, SDLoc(N), N->getValueType(0),
-                                       MVT::i32, MVT::Other, Ops);
+  SDNode *New = CurDAG->getMachineNode(Opcode, SDLoc(N), MVT::i32,
+                                       N->getValueType(0), MVT::Other, Ops);
   transferMemOperands(N, New);
   ReplaceUses(SDValue(N, 0), SDValue(New, 1));
   ReplaceUses(SDValue(N, 1), SDValue(New, 0));
@@ -2514,7 +2514,16 @@
 
   Ops.push_back(N->getOperand(0)); // chain
 
-  CurDAG->SelectNodeTo(N, Opcode, N->getVTList(), makeArrayRef(Ops));
+  SmallVector<EVT, 8> VTs;
+  VTs.push_back(N->getValueType(1));
+  VTs.push_back(N->getValueType(0));
+  VTs.push_back(N->getValueType(2));
+
+  SDNode *New = CurDAG->getMachineNode(Opcode, SDLoc(N), VTs, Ops);
+  ReplaceUses(SDValue(N, 0), SDValue(New, 1));
+  ReplaceUses(SDValue(N, 1), SDValue(New, 0));
+  ReplaceUses(SDValue(N, 2), SDValue(New, 2));
+  CurDAG->RemoveDeadNode(N);
 }
 
 void ARMDAGToDAGISel::SelectMVE_LongShift(SDNode *N, uint16_t Opcode,
diff --git a/llvm/test/CodeGen/Thumb2/mve-intrinsics/scatter-gather.ll b/llvm/test/CodeGen/Thumb2/mve-intrinsics/scatter-gather.ll
--- a/llvm/test/CodeGen/Thumb2/mve-intrinsics/scatter-gather.ll
+++ b/llvm/test/CodeGen/Thumb2/mve-intrinsics/scatter-gather.ll
@@ -202,8 +202,8 @@
 define arm_aapcs_vfpcc <2 x i64> @test_vldrdq_gather_base_wb_s64(<2 x i64>* %addr) {
 ; CHECK-LABEL: test_vldrdq_gather_base_wb_s64:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vldrw.u32 q0, [r0]
-; CHECK-NEXT:    vldrd.u64 q1, [q0, #576]!
+; CHECK-NEXT:    vldrw.u32 q1, [r0]
+; CHECK-NEXT:    vldrd.u64 q0, [q1, #576]!
 ; CHECK-NEXT:    vstrw.32 q1, [r0]
 ; CHECK-NEXT:    bx lr
 entry:
@@ -220,8 +220,8 @@
 define arm_aapcs_vfpcc <2 x i64> @test_vldrdq_gather_base_wb_u64(<2 x i64>* %addr) {
 ; CHECK-LABEL: test_vldrdq_gather_base_wb_u64:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vldrw.u32 q0, [r0]
-; CHECK-NEXT:    vldrd.u64 q1, [q0, #-328]!
+; CHECK-NEXT:    vldrw.u32 q1, [r0]
+; CHECK-NEXT:    vldrd.u64 q0, [q1, #-328]!
 ; CHECK-NEXT:    vstrw.32 q1, [r0]
 ; CHECK-NEXT:    bx lr
 entry:
@@ -237,9 +237,9 @@
 ; CHECK-LABEL: test_vldrdq_gather_base_wb_z_s64:
 ; CHECK:       @ %bb.0: @ %entry
 ; CHECK-NEXT:    vmsr p0, r1
-; CHECK-NEXT:    vldrw.u32 q0, [r0]
+; CHECK-NEXT:    vldrw.u32 q1, [r0]
 ; CHECK-NEXT:    vpst
-; CHECK-NEXT:    vldrdt.u64 q1, [q0, #664]!
+; CHECK-NEXT:    vldrdt.u64 q0, [q1, #664]!
 ; CHECK-NEXT:    vstrw.32 q1, [r0]
 ; CHECK-NEXT:    bx lr
 entry:
@@ -259,9 +259,9 @@
 ; CHECK-LABEL: test_vldrdq_gather_base_wb_z_u64:
 ; CHECK:       @ %bb.0: @ %entry
 ; CHECK-NEXT:    vmsr p0, r1
-; CHECK-NEXT:    vldrw.u32 q0, [r0]
+; CHECK-NEXT:    vldrw.u32 q1, [r0]
 ; CHECK-NEXT:    vpst
-; CHECK-NEXT:    vldrdt.u64 q1, [q0, #656]!
+; CHECK-NEXT:    vldrdt.u64 q0, [q1, #656]!
 ; CHECK-NEXT:    vstrw.32 q1, [r0]
 ; CHECK-NEXT:    bx lr
 entry:
@@ -727,8 +727,8 @@
 define arm_aapcs_vfpcc <4 x float> @test_vldrwq_gather_base_wb_f32(<4 x i32>* %addr) {
 ; CHECK-LABEL: test_vldrwq_gather_base_wb_f32:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vldrw.u32 q0, [r0]
-; CHECK-NEXT:    vldrw.u32 q1, [q0, #-64]!
+; CHECK-NEXT:    vldrw.u32 q1, [r0]
+; CHECK-NEXT:    vldrw.u32 q0, [q1, #-64]!
 ; CHECK-NEXT:    vstrw.32 q1, [r0]
 ; CHECK-NEXT:    bx lr
 entry:
@@ -745,8 +745,8 @@
 define arm_aapcs_vfpcc <4 x i32> @test_vldrwq_gather_base_wb_s32(<4 x i32>* %addr) {
 ; CHECK-LABEL: test_vldrwq_gather_base_wb_s32:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vldrw.u32 q0, [r0]
-; CHECK-NEXT:    vldrw.u32 q1, [q0, #80]!
+; CHECK-NEXT:    vldrw.u32 q1, [r0]
+; CHECK-NEXT:    vldrw.u32 q0, [q1, #80]!
 ; CHECK-NEXT:    vstrw.32 q1, [r0]
 ; CHECK-NEXT:    bx lr
 entry:
@@ -763,8 +763,8 @@
 define arm_aapcs_vfpcc <4 x i32> @test_vldrwq_gather_base_wb_u32(<4 x i32>* %addr) {
 ; CHECK-LABEL: test_vldrwq_gather_base_wb_u32:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vldrw.u32 q0, [r0]
-; CHECK-NEXT:    vldrw.u32 q1, [q0, #480]!
+; CHECK-NEXT:    vldrw.u32 q1, [r0]
+; CHECK-NEXT:    vldrw.u32 q0, [q1, #480]!
 ; CHECK-NEXT:    vstrw.32 q1, [r0]
 ; CHECK-NEXT:    bx lr
 entry:
@@ -780,9 +780,9 @@
 ; CHECK-LABEL: test_vldrwq_gather_base_wb_z_f32:
 ; CHECK:       @ %bb.0: @ %entry
 ; CHECK-NEXT:    vmsr p0, r1
-; CHECK-NEXT:    vldrw.u32 q0, [r0]
+; CHECK-NEXT:    vldrw.u32 q1, [r0]
 ; CHECK-NEXT:    vpst
-; CHECK-NEXT:    vldrwt.u32 q1, [q0, #-352]!
+; CHECK-NEXT:    vldrwt.u32 q0, [q1, #-352]!
 ; CHECK-NEXT:    vstrw.32 q1, [r0]
 ; CHECK-NEXT:    bx lr
 entry:
@@ -802,9 +802,9 @@
 ; CHECK-LABEL: test_vldrwq_gather_base_wb_z_s32:
 ; CHECK:       @ %bb.0: @ %entry
 ; CHECK-NEXT:    vmsr p0, r1
-; CHECK-NEXT:    vldrw.u32 q0, [r0]
+; CHECK-NEXT:    vldrw.u32 q1, [r0]
 ; CHECK-NEXT:    vpst
-; CHECK-NEXT:    vldrwt.u32 q1, [q0, #276]!
+; CHECK-NEXT:    vldrwt.u32 q0, [q1, #276]!
 ; CHECK-NEXT:    vstrw.32 q1, [r0]
 ; CHECK-NEXT:    bx lr
 entry:
@@ -824,9 +824,9 @@
 ; CHECK-LABEL: test_vldrwq_gather_base_wb_z_u32:
 ; CHECK:       @ %bb.0: @ %entry
 ; CHECK-NEXT:    vmsr p0, r1
-; CHECK-NEXT:    vldrw.u32 q0, [r0]
+; CHECK-NEXT:    vldrw.u32 q1, [r0]
 ; CHECK-NEXT:    vpst
-; CHECK-NEXT:    vldrwt.u32 q1, [q0, #88]!
+; CHECK-NEXT:    vldrwt.u32 q0, [q1, #88]!
 ; CHECK-NEXT:    vstrw.32 q1, [r0]
 ; CHECK-NEXT:    bx lr
 entry:
diff --git a/llvm/test/CodeGen/Thumb2/mve-intrinsics/vldr.ll b/llvm/test/CodeGen/Thumb2/mve-intrinsics/vldr.ll
--- a/llvm/test/CodeGen/Thumb2/mve-intrinsics/vldr.ll
+++ b/llvm/test/CodeGen/Thumb2/mve-intrinsics/vldr.ll
@@ -4,8 +4,8 @@
 define arm_aapcs_vfpcc <4 x i32> @test_vldrwq_gather_base_wb_s32(<4 x i32>* %addr) {
 ; CHECK-LABEL: test_vldrwq_gather_base_wb_s32:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vldrw.u32 q0, [r0]
-; CHECK-NEXT:    vldrw.u32 q1, [q0, #80]!
+; CHECK-NEXT:    vldrw.u32 q1, [r0]
+; CHECK-NEXT:    vldrw.u32 q0, [q1, #80]!
 ; CHECK-NEXT:    vstrw.32 q1, [r0]
 ; CHECK-NEXT:    bx lr
 entry:
@@ -22,8 +22,8 @@
 define arm_aapcs_vfpcc <4 x float> @test_vldrwq_gather_base_wb_f32(<4 x i32>* %addr) {
 ; CHECK-LABEL: test_vldrwq_gather_base_wb_f32:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vldrw.u32 q0, [r0]
-; CHECK-NEXT:    vldrw.u32 q1, [q0, #64]!
+; CHECK-NEXT:    vldrw.u32 q1, [r0]
+; CHECK-NEXT:    vldrw.u32 q0, [q1, #64]!
 ; CHECK-NEXT:    vstrw.32 q1, [r0]
 ; CHECK-NEXT:    bx lr
 entry:
@@ -41,9 +41,9 @@
 ; CHECK-LABEL: test_vldrdq_gather_base_wb_z_u64:
 ; CHECK:       @ %bb.0: @ %entry
 ; CHECK-NEXT:    vmsr p0, r1
-; CHECK-NEXT:    vldrw.u32 q0, [r0]
+; CHECK-NEXT:    vldrw.u32 q1, [r0]
 ; CHECK-NEXT:    vpst
-; CHECK-NEXT:    vldrdt.u64 q1, [q0, #656]!
+; CHECK-NEXT:    vldrdt.u64 q0, [q1, #656]!
 ; CHECK-NEXT:    vstrw.32 q1, [r0]
 ; CHECK-NEXT:    bx lr
 entry: