From: David Green <david.green@arm.com>
Date: Sun, 28 May 2023 18:12:45 +0000 (+0100)
Subject: [ARM] Allow D-reg copies to use VMOVD with fpregs64
X-Git-Tag: upstream/17.0.6~6944
X-Git-Url: http://review.tizen.org/git/?a=commitdiff_plain;h=0a762ec1b09d96734a3462f8792a5574d089b24d;p=platform%2Fupstream%2Fllvm.git

[ARM] Allow D-reg copies to use VMOVD with fpregs64

This instruction should be available with MVE, where we have D regs, not
requiring the full FP64 target feature.
---

diff --git a/llvm/lib/Target/ARM/ARMBaseInstrInfo.cpp b/llvm/lib/Target/ARM/ARMBaseInstrInfo.cpp
index 2ffa540..792d486 100644
--- a/llvm/lib/Target/ARM/ARMBaseInstrInfo.cpp
+++ b/llvm/lib/Target/ARM/ARMBaseInstrInfo.cpp
@@ -913,7 +913,7 @@ void ARMBaseInstrInfo::copyPhysReg(MachineBasicBlock &MBB,
     Opc = ARM::VMOVRS;
   else if (SPRDest && GPRSrc)
     Opc = ARM::VMOVSR;
-  else if (ARM::DPRRegClass.contains(DestReg, SrcReg) && Subtarget.hasFP64())
+  else if (ARM::DPRRegClass.contains(DestReg, SrcReg) && Subtarget.hasFPRegs64())
     Opc = ARM::VMOVD;
   else if (ARM::QPRRegClass.contains(DestReg, SrcReg))
     Opc = Subtarget.hasNEON() ? ARM::VORRq : ARM::MQPRCopy;
diff --git a/llvm/test/CodeGen/Thumb2/aapcs.ll b/llvm/test/CodeGen/Thumb2/aapcs.ll
index 651b994..b8a93d4 100644
--- a/llvm/test/CodeGen/Thumb2/aapcs.ll
+++ b/llvm/test/CodeGen/Thumb2/aapcs.ll
@@ -1,6 +1,6 @@
-; RUN: llc < %s -mtriple=thumbv7-none-eabi   -mcpu=cortex-m4 -mattr=-vfp2             | FileCheck %s -check-prefix=CHECK -check-prefix=SOFT
-; RUN: llc < %s -mtriple=thumbv7-none-eabihf -mcpu=cortex-m4 -mattr=+vfp4,-fp64 | FileCheck %s -check-prefix=CHECK -check-prefix=HARD -check-prefix=SP
-; RUN: llc < %s -mtriple=thumbv7-none-eabihf -mcpu=cortex-a8 -mattr=+vfp3             | FileCheck %s -check-prefix=CHECK -check-prefix=HARD -check-prefix=DP
+; RUN: llc < %s -mtriple=thumbv7-none-eabi   -mcpu=cortex-m4 -mattr=-vfp2 | FileCheck %s -check-prefix=CHECK -check-prefix=SOFT
+; RUN: llc < %s -mtriple=thumbv7-none-eabihf -mcpu=cortex-m4 -mattr=+vfp4d16sp | FileCheck %s -check-prefix=CHECK -check-prefix=HARD -check-prefix=SP
+; RUN: llc < %s -mtriple=thumbv7-none-eabihf -mcpu=cortex-a8 -mattr=+vfp3 | FileCheck %s -check-prefix=CHECK -check-prefix=HARD -check-prefix=DP
 
 define float @float_in_reg(float %a, float %b) {
 entry:
diff --git a/llvm/test/CodeGen/Thumb2/mve-complex-deinterleaving-i64-add.ll b/llvm/test/CodeGen/Thumb2/mve-complex-deinterleaving-i64-add.ll
index dea6990..157ca2a 100644
--- a/llvm/test/CodeGen/Thumb2/mve-complex-deinterleaving-i64-add.ll
+++ b/llvm/test/CodeGen/Thumb2/mve-complex-deinterleaving-i64-add.ll
@@ -57,16 +57,13 @@ define arm_aapcs_vfpcc <4 x i64> @complex_add_v4i64(<4 x i64> %a, <4 x i64> %b)
 ; CHECK-NEXT:    sbc.w r12, r3, r1
 ; CHECK-NEXT:    vmov r2, r3, d1
 ; CHECK-NEXT:    vmov r1, r0, d4
-; CHECK-NEXT:    vmov.f32 s2, s4
-; CHECK-NEXT:    vmov.f32 s3, s5
+; CHECK-NEXT:    vmov.f64 d1, d2
 ; CHECK-NEXT:    subs r1, r1, r2
 ; CHECK-NEXT:    vmov q2[2], q2[0], r1, lr
 ; CHECK-NEXT:    sbcs r0, r3
 ; CHECK-NEXT:    vmov q2[3], q2[1], r0, r12
-; CHECK-NEXT:    vmov.f32 s0, s8
-; CHECK-NEXT:    vmov.f32 s4, s10
-; CHECK-NEXT:    vmov.f32 s1, s9
-; CHECK-NEXT:    vmov.f32 s5, s11
+; CHECK-NEXT:    vmov.f64 d0, d4
+; CHECK-NEXT:    vmov.f64 d2, d5
 ; CHECK-NEXT:    vpop {d8, d9}
 ; CHECK-NEXT:    pop {r7, pc}
 entry:
@@ -109,8 +106,7 @@ define arm_aapcs_vfpcc <8 x i64> @complex_add_v8i64(<8 x i64> %a, <8 x i64> %b)
 ; CHECK-NEXT:    sbc.w r12, r3, r1
 ; CHECK-NEXT:    vmov r2, r3, d1
 ; CHECK-NEXT:    vmov r1, r0, d12
-; CHECK-NEXT:    vmov.f32 s2, s4
-; CHECK-NEXT:    vmov.f32 s3, s5
+; CHECK-NEXT:    vmov.f64 d1, d2
 ; CHECK-NEXT:    subs r1, r1, r2
 ; CHECK-NEXT:    add r2, sp, #104
 ; CHECK-NEXT:    vldrw.u32 q5, [r2]
@@ -119,10 +115,8 @@ define arm_aapcs_vfpcc <8 x i64> @complex_add_v8i64(<8 x i64> %a, <8 x i64> %b)
 ; CHECK-NEXT:    vmov q4[3], q4[1], r0, r12
 ; CHECK-NEXT:    vmov r0, r1, d6
 ; CHECK-NEXT:    vmov r2, r3, d11
-; CHECK-NEXT:    vmov.f32 s0, s16
-; CHECK-NEXT:    vmov.f32 s4, s18
-; CHECK-NEXT:    vmov.f32 s1, s17
-; CHECK-NEXT:    vmov.f32 s5, s19
+; CHECK-NEXT:    vmov.f64 d0, d8
+; CHECK-NEXT:    vmov.f64 d2, d9
 ; CHECK-NEXT:    adds.w lr, r2, r0
 ; CHECK-NEXT:    adc.w r12, r3, r1
 ; CHECK-NEXT:    add r1, sp, #88
@@ -139,16 +133,13 @@ define arm_aapcs_vfpcc <8 x i64> @complex_add_v8i64(<8 x i64> %a, <8 x i64> %b)
 ; CHECK-NEXT:    sbc.w r12, r3, r1
 ; CHECK-NEXT:    vmov r2, r3, d5
 ; CHECK-NEXT:    vmov r1, r0, d12
-; CHECK-NEXT:    vmov.f32 s10, s16
-; CHECK-NEXT:    vmov.f32 s11, s17
+; CHECK-NEXT:    vmov.f64 d5, d8
 ; CHECK-NEXT:    subs r1, r1, r2
 ; CHECK-NEXT:    vmov q3[2], q3[0], r1, lr
 ; CHECK-NEXT:    sbcs r0, r3
 ; CHECK-NEXT:    vmov q3[3], q3[1], r0, r12
-; CHECK-NEXT:    vmov.f32 s16, s14
-; CHECK-NEXT:    vmov.f32 s8, s12
-; CHECK-NEXT:    vmov.f32 s17, s15
-; CHECK-NEXT:    vmov.f32 s9, s13
+; CHECK-NEXT:    vmov.f64 d8, d7
+; CHECK-NEXT:    vmov.f64 d4, d6
 ; CHECK-NEXT:    vmov q3, q4
 ; CHECK-NEXT:    vpop {d8, d9, d10, d11, d12, d13}
 ; CHECK-NEXT:    pop {r7, pc}
diff --git a/llvm/test/CodeGen/Thumb2/mve-fptosi-sat-vector.ll b/llvm/test/CodeGen/Thumb2/mve-fptosi-sat-vector.ll
index 3ca01cf..bd66c90 100644
--- a/llvm/test/CodeGen/Thumb2/mve-fptosi-sat-vector.ll
+++ b/llvm/test/CodeGen/Thumb2/mve-fptosi-sat-vector.ll
@@ -490,18 +490,16 @@ define arm_aapcs_vfpcc <3 x i32> @test_signed_v3f64_v3i32(<3 x double> %f) {
 ; CHECK-NEXT:    vpush {d8, d9}
 ; CHECK-NEXT:    .pad #24
 ; CHECK-NEXT:    sub sp, #24
-; CHECK-NEXT:    vmov.f32 s16, s0
-; CHECK-NEXT:    vmov.f32 s17, s1
+; CHECK-NEXT:    vmov.f64 d8, d0
 ; CHECK-NEXT:    vldr d0, .LCPI10_0
 ; CHECK-NEXT:    vmov r4, r6, d1
+; CHECK-NEXT:    vmov.f64 d9, d2
 ; CHECK-NEXT:    vmov r2, r11, d0
-; CHECK-NEXT:    vmov.f32 s18, s4
-; CHECK-NEXT:    vmov.f32 s19, s5
-; CHECK-NEXT:    str r2, [sp, #20] @ 4-byte Spill
 ; CHECK-NEXT:    mov r0, r4
 ; CHECK-NEXT:    mov r1, r6
-; CHECK-NEXT:    mov r3, r11
+; CHECK-NEXT:    str r2, [sp, #20] @ 4-byte Spill
 ; CHECK-NEXT:    str.w r11, [sp, #12] @ 4-byte Spill
+; CHECK-NEXT:    mov r3, r11
 ; CHECK-NEXT:    bl __aeabi_dcmpgt
 ; CHECK-NEXT:    vldr d0, .LCPI10_1
 ; CHECK-NEXT:    mov r1, r6
@@ -791,19 +789,15 @@ define arm_aapcs_vfpcc <5 x i32> @test_signed_v5f64_v5i32(<5 x double> %f) {
 ; CHECK-NEXT:    vpush {d8, d9, d10, d11}
 ; CHECK-NEXT:    .pad #32
 ; CHECK-NEXT:    sub sp, #32
-; CHECK-NEXT:    vmov.f32 s16, s0
+; CHECK-NEXT:    vmov.f64 d8, d0
 ; CHECK-NEXT:    mov r7, r0
-; CHECK-NEXT:    vmov.f32 s17, s1
+; CHECK-NEXT:    str r0, [sp, #16] @ 4-byte Spill
 ; CHECK-NEXT:    vldr d0, .LCPI12_0
 ; CHECK-NEXT:    vmov r5, r4, d4
-; CHECK-NEXT:    str r0, [sp, #16] @ 4-byte Spill
+; CHECK-NEXT:    vmov.f64 d10, d3
 ; CHECK-NEXT:    vmov r2, r3, d0
-; CHECK-NEXT:    vmov.f32 s20, s6
-; CHECK-NEXT:    vmov.f32 s18, s4
-; CHECK-NEXT:    vmov.f32 s22, s2
-; CHECK-NEXT:    vmov.f32 s21, s7
-; CHECK-NEXT:    vmov.f32 s19, s5
-; CHECK-NEXT:    vmov.f32 s23, s3
+; CHECK-NEXT:    vmov.f64 d9, d2
+; CHECK-NEXT:    vmov.f64 d11, d1
 ; CHECK-NEXT:    mov r0, r5
 ; CHECK-NEXT:    mov r1, r4
 ; CHECK-NEXT:    strd r2, r3, [sp, #20] @ 8-byte Folded Spill
@@ -1000,25 +994,20 @@ define arm_aapcs_vfpcc <6 x i32> @test_signed_v6f64_v6i32(<6 x double> %f) {
 ; CHECK-NEXT:    vpush {d8, d9, d10, d11, d12}
 ; CHECK-NEXT:    .pad #40
 ; CHECK-NEXT:    sub sp, #40
-; CHECK-NEXT:    vmov.f32 s16, s0
+; CHECK-NEXT:    vmov.f64 d8, d0
 ; CHECK-NEXT:    str r0, [sp, #20] @ 4-byte Spill
-; CHECK-NEXT:    vmov.f32 s17, s1
 ; CHECK-NEXT:    vldr d0, .LCPI13_0
 ; CHECK-NEXT:    vmov r9, r4, d5
+; CHECK-NEXT:    vmov.f64 d11, d4
 ; CHECK-NEXT:    vmov r2, r6, d0
-; CHECK-NEXT:    vmov.f32 s22, s8
-; CHECK-NEXT:    vmov.f32 s20, s6
-; CHECK-NEXT:    vmov.f32 s18, s4
-; CHECK-NEXT:    vmov.f32 s24, s2
-; CHECK-NEXT:    vmov.f32 s23, s9
-; CHECK-NEXT:    vmov.f32 s21, s7
-; CHECK-NEXT:    vmov.f32 s19, s5
-; CHECK-NEXT:    vmov.f32 s25, s3
-; CHECK-NEXT:    str r2, [sp, #24] @ 4-byte Spill
+; CHECK-NEXT:    vmov.f64 d10, d3
+; CHECK-NEXT:    vmov.f64 d9, d2
+; CHECK-NEXT:    vmov.f64 d12, d1
 ; CHECK-NEXT:    mov r0, r9
 ; CHECK-NEXT:    mov r1, r4
-; CHECK-NEXT:    mov r3, r6
+; CHECK-NEXT:    str r2, [sp, #24] @ 4-byte Spill
 ; CHECK-NEXT:    str r6, [sp, #28] @ 4-byte Spill
+; CHECK-NEXT:    mov r3, r6
 ; CHECK-NEXT:    bl __aeabi_dcmpgt
 ; CHECK-NEXT:    vldr d0, .LCPI13_1
 ; CHECK-NEXT:    mov r1, r4
diff --git a/llvm/test/CodeGen/Thumb2/mve-fptoui-sat-vector.ll b/llvm/test/CodeGen/Thumb2/mve-fptoui-sat-vector.ll
index 8ea12bd..0fdd4a2 100644
--- a/llvm/test/CodeGen/Thumb2/mve-fptoui-sat-vector.ll
+++ b/llvm/test/CodeGen/Thumb2/mve-fptoui-sat-vector.ll
@@ -432,19 +432,17 @@ define arm_aapcs_vfpcc <3 x i32> @test_unsigned_v3f64_v3i32(<3 x double> %f) {
 ; CHECK-NEXT:    vpush {d8, d9}
 ; CHECK-NEXT:    .pad #24
 ; CHECK-NEXT:    sub sp, #24
-; CHECK-NEXT:    vmov.f32 s18, s0
-; CHECK-NEXT:    vmov.f32 s19, s1
+; CHECK-NEXT:    vmov.f64 d9, d0
 ; CHECK-NEXT:    vldr d0, .LCPI10_0
 ; CHECK-NEXT:    vmov r4, r5, d1
+; CHECK-NEXT:    vmov.f64 d8, d2
 ; CHECK-NEXT:    vmov r9, r7, d0
-; CHECK-NEXT:    vmov.f32 s16, s4
-; CHECK-NEXT:    vmov.f32 s17, s5
-; CHECK-NEXT:    str.w r9, [sp, #8] @ 4-byte Spill
 ; CHECK-NEXT:    mov r0, r4
 ; CHECK-NEXT:    mov r1, r5
+; CHECK-NEXT:    str.w r9, [sp, #8] @ 4-byte Spill
+; CHECK-NEXT:    str r7, [sp, #12] @ 4-byte Spill
 ; CHECK-NEXT:    mov r2, r9
 ; CHECK-NEXT:    mov r3, r7
-; CHECK-NEXT:    str r7, [sp, #12] @ 4-byte Spill
 ; CHECK-NEXT:    bl __aeabi_dcmpgt
 ; CHECK-NEXT:    vldr d0, .LCPI10_1
 ; CHECK-NEXT:    mov r1, r5
@@ -665,19 +663,15 @@ define arm_aapcs_vfpcc <5 x i32> @test_unsigned_v5f64_v5i32(<5 x double> %f) {
 ; CHECK-NEXT:    vpush {d8, d9, d10, d11}
 ; CHECK-NEXT:    .pad #40
 ; CHECK-NEXT:    sub sp, #40
-; CHECK-NEXT:    vmov.f32 s16, s0
+; CHECK-NEXT:    vmov.f64 d8, d0
 ; CHECK-NEXT:    mov r4, r0
-; CHECK-NEXT:    vmov.f32 s17, s1
+; CHECK-NEXT:    str r0, [sp, #28] @ 4-byte Spill
 ; CHECK-NEXT:    vldr d0, .LCPI12_0
 ; CHECK-NEXT:    vmov r5, r6, d4
-; CHECK-NEXT:    str r0, [sp, #28] @ 4-byte Spill
+; CHECK-NEXT:    vmov.f64 d10, d3
 ; CHECK-NEXT:    vmov r2, r3, d0
-; CHECK-NEXT:    vmov.f32 s20, s6
-; CHECK-NEXT:    vmov.f32 s18, s4
-; CHECK-NEXT:    vmov.f32 s22, s2
-; CHECK-NEXT:    vmov.f32 s21, s7
-; CHECK-NEXT:    vmov.f32 s19, s5
-; CHECK-NEXT:    vmov.f32 s23, s3
+; CHECK-NEXT:    vmov.f64 d9, d2
+; CHECK-NEXT:    vmov.f64 d11, d1
 ; CHECK-NEXT:    mov r0, r5
 ; CHECK-NEXT:    mov r1, r6
 ; CHECK-NEXT:    strd r2, r3, [sp, #32] @ 8-byte Folded Spill
@@ -834,25 +828,20 @@ define arm_aapcs_vfpcc <6 x i32> @test_unsigned_v6f64_v6i32(<6 x double> %f) {
 ; CHECK-NEXT:    vpush {d8, d9, d10, d11, d12}
 ; CHECK-NEXT:    .pad #40
 ; CHECK-NEXT:    sub sp, #40
-; CHECK-NEXT:    vmov.f32 s16, s0
+; CHECK-NEXT:    vmov.f64 d8, d0
 ; CHECK-NEXT:    str r0, [sp, #32] @ 4-byte Spill
-; CHECK-NEXT:    vmov.f32 s17, s1
 ; CHECK-NEXT:    vldr d0, .LCPI13_0
 ; CHECK-NEXT:    vmov r5, r6, d5
+; CHECK-NEXT:    vmov.f64 d11, d4
 ; CHECK-NEXT:    vmov r11, r3, d0
-; CHECK-NEXT:    vmov.f32 s22, s8
-; CHECK-NEXT:    vmov.f32 s20, s6
-; CHECK-NEXT:    vmov.f32 s18, s4
-; CHECK-NEXT:    vmov.f32 s24, s2
-; CHECK-NEXT:    vmov.f32 s23, s9
-; CHECK-NEXT:    vmov.f32 s21, s7
-; CHECK-NEXT:    vmov.f32 s19, s5
-; CHECK-NEXT:    vmov.f32 s25, s3
-; CHECK-NEXT:    str r3, [sp, #36] @ 4-byte Spill
+; CHECK-NEXT:    vmov.f64 d10, d3
+; CHECK-NEXT:    vmov.f64 d9, d2
+; CHECK-NEXT:    vmov.f64 d12, d1
 ; CHECK-NEXT:    mov r0, r5
 ; CHECK-NEXT:    mov r1, r6
-; CHECK-NEXT:    mov r2, r11
+; CHECK-NEXT:    str r3, [sp, #36] @ 4-byte Spill
 ; CHECK-NEXT:    str.w r11, [sp, #28] @ 4-byte Spill
+; CHECK-NEXT:    mov r2, r11
 ; CHECK-NEXT:    bl __aeabi_dcmpgt
 ; CHECK-NEXT:    vldr d0, .LCPI13_1
 ; CHECK-NEXT:    mov r7, r0
diff --git a/llvm/test/CodeGen/Thumb2/mve-shuffle.ll b/llvm/test/CodeGen/Thumb2/mve-shuffle.ll
index 93a0588..3f2310b 100644
--- a/llvm/test/CodeGen/Thumb2/mve-shuffle.ll
+++ b/llvm/test/CodeGen/Thumb2/mve-shuffle.ll
@@ -921,10 +921,8 @@ entry:
 define arm_aapcs_vfpcc <2 x i64> @shuffle2_i64(<2 x i64> %src) {
 ; CHECK-LABEL: shuffle2_i64:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s4, s2
-; CHECK-NEXT:    vmov.f32 s6, s0
-; CHECK-NEXT:    vmov.f32 s5, s3
-; CHECK-NEXT:    vmov.f32 s7, s1
+; CHECK-NEXT:    vmov.f64 d2, d1
+; CHECK-NEXT:    vmov.f64 d3, d0
 ; CHECK-NEXT:    vmov q0, q1
 ; CHECK-NEXT:    bx lr
 entry:
@@ -1381,10 +1379,8 @@ entry:
 define arm_aapcs_vfpcc <2 x double> @shuffle2_f64(<2 x double> %src) {
 ; CHECK-LABEL: shuffle2_f64:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s4, s2
-; CHECK-NEXT:    vmov.f32 s6, s0
-; CHECK-NEXT:    vmov.f32 s5, s3
-; CHECK-NEXT:    vmov.f32 s7, s1
+; CHECK-NEXT:    vmov.f64 d2, d1
+; CHECK-NEXT:    vmov.f64 d3, d0
 ; CHECK-NEXT:    vmov q0, q1
 ; CHECK-NEXT:    bx lr
 entry:
@@ -1404,12 +1400,9 @@ entry:
 define arm_aapcs_vfpcc <4 x double> @shuffle4_f64(<2 x double> %src1, <2 x double> %src2) {
 ; CHECK-LABEL: shuffle4_f64:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s8, s6
-; CHECK-NEXT:    vmov.f32 s6, s0
-; CHECK-NEXT:    vmov.f32 s9, s7
-; CHECK-NEXT:    vmov.f32 s7, s1
-; CHECK-NEXT:    vmov.f32 s10, s2
-; CHECK-NEXT:    vmov.f32 s11, s3
+; CHECK-NEXT:    vmov.f64 d4, d3
+; CHECK-NEXT:    vmov.f64 d5, d1
+; CHECK-NEXT:    vmov.f64 d3, d0
 ; CHECK-NEXT:    vmov q0, q2
 ; CHECK-NEXT:    bx lr
 entry:
@@ -1419,14 +1412,10 @@ entry:
 define arm_aapcs_vfpcc <4 x double> @shuffle5_f64(<2 x double> %src1, <2 x double> %src2) {
 ; CHECK-LABEL: shuffle5_f64:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s8, s6
-; CHECK-NEXT:    vmov.f32 s10, s4
-; CHECK-NEXT:    vmov.f32 s4, s2
-; CHECK-NEXT:    vmov.f32 s6, s0
-; CHECK-NEXT:    vmov.f32 s9, s7
-; CHECK-NEXT:    vmov.f32 s11, s5
-; CHECK-NEXT:    vmov.f32 s5, s3
-; CHECK-NEXT:    vmov.f32 s7, s1
+; CHECK-NEXT:    vmov.f64 d4, d3
+; CHECK-NEXT:    vmov.f64 d5, d2
+; CHECK-NEXT:    vmov.f64 d2, d1
+; CHECK-NEXT:    vmov.f64 d3, d0
 ; CHECK-NEXT:    vmov q0, q2
 ; CHECK-NEXT:    bx lr
 entry:
@@ -1436,8 +1425,7 @@ entry:
 define arm_aapcs_vfpcc <2 x double> @shuffle6_f64(<2 x double> %src1, <2 x double> %src2) {
 ; CHECK-LABEL: shuffle6_f64:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s2, s6
-; CHECK-NEXT:    vmov.f32 s3, s7
+; CHECK-NEXT:    vmov.f64 d1, d3
 ; CHECK-NEXT:    bx lr
 entry:
   %out = shufflevector <2 x double> %src1, <2 x double> %src2, <2 x i32> <i32 0, i32 3>
@@ -1446,8 +1434,7 @@ entry:
 define arm_aapcs_vfpcc <2 x double> @shuffle7_f64(<2 x double> %src1, <2 x double> %src2) {
 ; CHECK-LABEL: shuffle7_f64:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s0, s6
-; CHECK-NEXT:    vmov.f32 s1, s7
+; CHECK-NEXT:    vmov.f64 d0, d3
 ; CHECK-NEXT:    bx lr
 entry:
   %out = shufflevector <2 x double> %src1, <2 x double> %src2, <2 x i32> <i32 3, i32 1>
@@ -1456,8 +1443,7 @@ entry:
 define arm_aapcs_vfpcc <2 x double> @shuffle8_f64(<2 x double> %src1, <2 x double> %src2) {
 ; CHECK-LABEL: shuffle8_f64:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s6, s2
-; CHECK-NEXT:    vmov.f32 s7, s3
+; CHECK-NEXT:    vmov.f64 d3, d1
 ; CHECK-NEXT:    vmov q0, q1
 ; CHECK-NEXT:    bx lr
 entry:
@@ -1470,19 +1456,13 @@ define arm_aapcs_vfpcc <8 x double> @shuffle9_f64(<4 x double> %src1, <4 x doubl
 ; CHECK-NEXT:    .vsave {d8, d9, d10, d11}
 ; CHECK-NEXT:    vpush {d8, d9, d10, d11}
 ; CHECK-NEXT:    vmov q5, q2
-; CHECK-NEXT:    vmov.f32 s16, s0
-; CHECK-NEXT:    vmov.f32 s18, s20
-; CHECK-NEXT:    vmov.f32 s20, s2
-; CHECK-NEXT:    vmov.f32 s10, s12
-; CHECK-NEXT:    vmov.f32 s19, s21
-; CHECK-NEXT:    vmov.f32 s8, s4
-; CHECK-NEXT:    vmov.f32 s17, s1
-; CHECK-NEXT:    vmov.f32 s21, s3
+; CHECK-NEXT:    vmov.f64 d8, d0
+; CHECK-NEXT:    vmov.f64 d9, d10
+; CHECK-NEXT:    vmov.f64 d10, d1
 ; CHECK-NEXT:    vmov q0, q4
-; CHECK-NEXT:    vmov.f32 s12, s6
-; CHECK-NEXT:    vmov.f32 s11, s13
-; CHECK-NEXT:    vmov.f32 s9, s5
-; CHECK-NEXT:    vmov.f32 s13, s7
+; CHECK-NEXT:    vmov.f64 d5, d6
+; CHECK-NEXT:    vmov.f64 d4, d2
+; CHECK-NEXT:    vmov.f64 d6, d3
 ; CHECK-NEXT:    vmov q1, q5
 ; CHECK-NEXT:    vpop {d8, d9, d10, d11}
 ; CHECK-NEXT:    bx lr
@@ -1497,12 +1477,9 @@ entry:
 define arm_aapcs_vfpcc <4 x i64> @shuffle4_i64(<2 x i64> %src1, <2 x i64> %src2) {
 ; CHECK-LABEL: shuffle4_i64:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s8, s6
-; CHECK-NEXT:    vmov.f32 s6, s0
-; CHECK-NEXT:    vmov.f32 s9, s7
-; CHECK-NEXT:    vmov.f32 s7, s1
-; CHECK-NEXT:    vmov.f32 s10, s2
-; CHECK-NEXT:    vmov.f32 s11, s3
+; CHECK-NEXT:    vmov.f64 d4, d3
+; CHECK-NEXT:    vmov.f64 d5, d1
+; CHECK-NEXT:    vmov.f64 d3, d0
 ; CHECK-NEXT:    vmov q0, q2
 ; CHECK-NEXT:    bx lr
 entry:
@@ -1512,14 +1489,10 @@ entry:
 define arm_aapcs_vfpcc <4 x i64> @shuffle5_i64(<2 x i64> %src1, <2 x i64> %src2) {
 ; CHECK-LABEL: shuffle5_i64:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s8, s6
-; CHECK-NEXT:    vmov.f32 s10, s4
-; CHECK-NEXT:    vmov.f32 s4, s2
-; CHECK-NEXT:    vmov.f32 s6, s0
-; CHECK-NEXT:    vmov.f32 s9, s7
-; CHECK-NEXT:    vmov.f32 s11, s5
-; CHECK-NEXT:    vmov.f32 s5, s3
-; CHECK-NEXT:    vmov.f32 s7, s1
+; CHECK-NEXT:    vmov.f64 d4, d3
+; CHECK-NEXT:    vmov.f64 d5, d2
+; CHECK-NEXT:    vmov.f64 d2, d1
+; CHECK-NEXT:    vmov.f64 d3, d0
 ; CHECK-NEXT:    vmov q0, q2
 ; CHECK-NEXT:    bx lr
 entry:
@@ -1529,8 +1502,7 @@ entry:
 define arm_aapcs_vfpcc <2 x i64> @shuffle6_i64(<2 x i64> %src1, <2 x i64> %src2) {
 ; CHECK-LABEL: shuffle6_i64:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s2, s6
-; CHECK-NEXT:    vmov.f32 s3, s7
+; CHECK-NEXT:    vmov.f64 d1, d3
 ; CHECK-NEXT:    bx lr
 entry:
   %out = shufflevector <2 x i64> %src1, <2 x i64> %src2, <2 x i32> <i32 0, i32 3>
@@ -1539,8 +1511,7 @@ entry:
 define arm_aapcs_vfpcc <2 x i64> @shuffle7_i64(<2 x i64> %src1, <2 x i64> %src2) {
 ; CHECK-LABEL: shuffle7_i64:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s0, s6
-; CHECK-NEXT:    vmov.f32 s1, s7
+; CHECK-NEXT:    vmov.f64 d0, d3
 ; CHECK-NEXT:    bx lr
 entry:
   %out = shufflevector <2 x i64> %src1, <2 x i64> %src2, <2 x i32> <i32 3, i32 1>
@@ -1549,8 +1520,7 @@ entry:
 define arm_aapcs_vfpcc <2 x i64> @shuffle8_i64(<2 x i64> %src1, <2 x i64> %src2) {
 ; CHECK-LABEL: shuffle8_i64:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s6, s2
-; CHECK-NEXT:    vmov.f32 s7, s3
+; CHECK-NEXT:    vmov.f64 d3, d1
 ; CHECK-NEXT:    vmov q0, q1
 ; CHECK-NEXT:    bx lr
 entry:
@@ -1563,19 +1533,13 @@ define arm_aapcs_vfpcc <8 x i64> @shuffle9_i64(<4 x i64> %src1, <4 x i64> %src2)
 ; CHECK-NEXT:    .vsave {d8, d9, d10, d11}
 ; CHECK-NEXT:    vpush {d8, d9, d10, d11}
 ; CHECK-NEXT:    vmov q5, q2
-; CHECK-NEXT:    vmov.f32 s16, s0
-; CHECK-NEXT:    vmov.f32 s18, s20
-; CHECK-NEXT:    vmov.f32 s20, s2
-; CHECK-NEXT:    vmov.f32 s10, s12
-; CHECK-NEXT:    vmov.f32 s19, s21
-; CHECK-NEXT:    vmov.f32 s8, s4
-; CHECK-NEXT:    vmov.f32 s17, s1
-; CHECK-NEXT:    vmov.f32 s21, s3
+; CHECK-NEXT:    vmov.f64 d8, d0
+; CHECK-NEXT:    vmov.f64 d9, d10
+; CHECK-NEXT:    vmov.f64 d10, d1
 ; CHECK-NEXT:    vmov q0, q4
-; CHECK-NEXT:    vmov.f32 s12, s6
-; CHECK-NEXT:    vmov.f32 s11, s13
-; CHECK-NEXT:    vmov.f32 s9, s5
-; CHECK-NEXT:    vmov.f32 s13, s7
+; CHECK-NEXT:    vmov.f64 d5, d6
+; CHECK-NEXT:    vmov.f64 d4, d2
+; CHECK-NEXT:    vmov.f64 d6, d3
 ; CHECK-NEXT:    vmov q1, q5
 ; CHECK-NEXT:    vpop {d8, d9, d10, d11}
 ; CHECK-NEXT:    bx lr
@@ -1810,8 +1774,7 @@ entry:
 define arm_aapcs_vfpcc double @extract_f64_1(<2 x double> %a) {
 ; CHECK-LABEL: extract_f64_1:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s0, s2
-; CHECK-NEXT:    vmov.f32 s1, s3
+; CHECK-NEXT:    vmov.f64 d0, d1
 ; CHECK-NEXT:    bx lr
 entry:
   %res = extractelement <2 x double> %a, i32 1
diff --git a/llvm/test/CodeGen/Thumb2/mve-shufflemov.ll b/llvm/test/CodeGen/Thumb2/mve-shufflemov.ll
index 6ce7550..4812ae5 100644
--- a/llvm/test/CodeGen/Thumb2/mve-shufflemov.ll
+++ b/llvm/test/CodeGen/Thumb2/mve-shufflemov.ll
@@ -7,10 +7,8 @@
 define arm_aapcs_vfpcc <8 x i16> @shuffle_i16_45670123(<8 x i16> %s1, <8 x i16> %s2) {
 ; CHECK-LABEL: shuffle_i16_45670123:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s4, s2
-; CHECK-NEXT:    vmov.f32 s6, s0
-; CHECK-NEXT:    vmov.f32 s5, s3
-; CHECK-NEXT:    vmov.f32 s7, s1
+; CHECK-NEXT:    vmov.f64 d2, d1
+; CHECK-NEXT:    vmov.f64 d3, d0
 ; CHECK-NEXT:    vmov q0, q1
 ; CHECK-NEXT:    bx lr
 entry:
@@ -58,8 +56,7 @@ entry:
 define arm_aapcs_vfpcc <8 x i16> @shuffle_i16_0123cdef(<8 x i16> %s1, <8 x i16> %s2) {
 ; CHECK-LABEL: shuffle_i16_0123cdef:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s2, s6
-; CHECK-NEXT:    vmov.f32 s3, s7
+; CHECK-NEXT:    vmov.f64 d1, d3
 ; CHECK-NEXT:    bx lr
 entry:
   %out = shufflevector <8 x i16> %s1, <8 x i16> %s2, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 12, i32 13, i32 14, i32 15>
@@ -278,10 +275,8 @@ entry:
 define arm_aapcs_vfpcc <8 x half> @shuffle_f16_45670123(<8 x half> %s1, <8 x half> %s2) {
 ; CHECK-LABEL: shuffle_f16_45670123:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s4, s2
-; CHECK-NEXT:    vmov.f32 s6, s0
-; CHECK-NEXT:    vmov.f32 s5, s3
-; CHECK-NEXT:    vmov.f32 s7, s1
+; CHECK-NEXT:    vmov.f64 d2, d1
+; CHECK-NEXT:    vmov.f64 d3, d0
 ; CHECK-NEXT:    vmov q0, q1
 ; CHECK-NEXT:    bx lr
 entry:
@@ -329,8 +324,7 @@ entry:
 define arm_aapcs_vfpcc <8 x half> @shuffle_f16_0123cdef(<8 x half> %s1, <8 x half> %s2) {
 ; CHECK-LABEL: shuffle_f16_0123cdef:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s2, s6
-; CHECK-NEXT:    vmov.f32 s3, s7
+; CHECK-NEXT:    vmov.f64 d1, d3
 ; CHECK-NEXT:    bx lr
 entry:
   %out = shufflevector <8 x half> %s1, <8 x half> %s2, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 12, i32 13, i32 14, i32 15>
diff --git a/llvm/test/CodeGen/Thumb2/mve-vdup.ll b/llvm/test/CodeGen/Thumb2/mve-vdup.ll
index 9ba3866..b87cc83 100644
--- a/llvm/test/CodeGen/Thumb2/mve-vdup.ll
+++ b/llvm/test/CodeGen/Thumb2/mve-vdup.ll
@@ -208,8 +208,7 @@ entry:
 define arm_aapcs_vfpcc <2 x double> @vdup_f64(double %src) {
 ; CHECK-LABEL: vdup_f64:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s2, s0
-; CHECK-NEXT:    vmov.f32 s3, s1
+; CHECK-NEXT:    vmov.f64 d1, d0
 ; CHECK-NEXT:    bx lr
 entry:
   %0 = insertelement <2 x double> undef, double %src, i32 0
@@ -279,8 +278,7 @@ entry:
 define arm_aapcs_vfpcc <2 x i64> @vduplane_i64(<2 x i64> %src) {
 ; CHECK-LABEL: vduplane_i64:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s0, s2
-; CHECK-NEXT:    vmov.f32 s1, s3
+; CHECK-NEXT:    vmov.f64 d0, d1
 ; CHECK-NEXT:    bx lr
 entry:
   %out = shufflevector <2 x i64> %src, <2 x i64> undef, <2 x i32> <i32 1, i32 1>
@@ -328,8 +326,7 @@ entry:
 define arm_aapcs_vfpcc <2 x double> @vduplane_f64(<2 x double> %src) {
 ; CHECK-LABEL: vduplane_f64:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s0, s2
-; CHECK-NEXT:    vmov.f32 s1, s3
+; CHECK-NEXT:    vmov.f64 d0, d1
 ; CHECK-NEXT:    bx lr
 entry:
   %out = shufflevector <2 x double> %src, <2 x double> undef, <2 x i32> <i32 1, i32 1>
@@ -506,8 +503,7 @@ define arm_aapcs_vfpcc <2 x i64> @bitcast_i64_v2f64(i64 %a) {
 define arm_aapcs_vfpcc <2 x i64> @bitcast_v2f64_v2i64(<2 x double> %a) {
 ; CHECK-LABEL: bitcast_v2f64_v2i64:
 ; CHECK:       @ %bb.0:
-; CHECK-NEXT:    vmov.f32 s2, s0
-; CHECK-NEXT:    vmov.f32 s3, s1
+; CHECK-NEXT:    vmov.f64 d1, d0
 ; CHECK-NEXT:    bx lr
   %b = bitcast <2 x double> %a to <2 x i64>
   %r = shufflevector <2 x i64> %b, <2 x i64> poison, <2 x i32> zeroinitializer
@@ -517,8 +513,7 @@ define arm_aapcs_vfpcc <2 x i64> @bitcast_v2f64_v2i64(<2 x double> %a) {
 define arm_aapcs_vfpcc <2 x i64> @bitcast_v8i16_v2i64(<8 x i16> %a) {
 ; CHECK-LABEL: bitcast_v8i16_v2i64:
 ; CHECK:       @ %bb.0:
-; CHECK-NEXT:    vmov.f32 s2, s0
-; CHECK-NEXT:    vmov.f32 s3, s1
+; CHECK-NEXT:    vmov.f64 d1, d0
 ; CHECK-NEXT:    bx lr
   %b = bitcast <8 x i16> %a to <2 x i64>
   %r = shufflevector <2 x i64> %b, <2 x i64> poison, <2 x i32> zeroinitializer
diff --git a/llvm/test/CodeGen/Thumb2/mve-vmovn.ll b/llvm/test/CodeGen/Thumb2/mve-vmovn.ll
index b005cb9..93848aa 100644
--- a/llvm/test/CodeGen/Thumb2/mve-vmovn.ll
+++ b/llvm/test/CodeGen/Thumb2/mve-vmovn.ll
@@ -192,14 +192,12 @@ entry:
 define arm_aapcs_vfpcc <2 x i64> @vmovn64_t1(<2 x i64> %src1, <2 x i64> %src2) {
 ; CHECK-LABEL: vmovn64_t1:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s2, s4
-; CHECK-NEXT:    vmov.f32 s3, s5
+; CHECK-NEXT:    vmov.f64 d1, d2
 ; CHECK-NEXT:    bx lr
 ;
 ; CHECKBE-LABEL: vmovn64_t1:
 ; CHECKBE:       @ %bb.0: @ %entry
-; CHECKBE-NEXT:    vmov.f32 s2, s4
-; CHECKBE-NEXT:    vmov.f32 s3, s5
+; CHECKBE-NEXT:    vmov.f64 d1, d2
 ; CHECKBE-NEXT:    bx lr
 entry:
   %out = shufflevector <2 x i64> %src1, <2 x i64> %src2, <2 x i32> <i32 0, i32 2>
@@ -209,15 +207,13 @@ entry:
 define arm_aapcs_vfpcc <2 x i64> @vmovn64_t2(<2 x i64> %src1, <2 x i64> %src2) {
 ; CHECK-LABEL: vmovn64_t2:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s6, s0
-; CHECK-NEXT:    vmov.f32 s7, s1
+; CHECK-NEXT:    vmov.f64 d3, d0
 ; CHECK-NEXT:    vmov q0, q1
 ; CHECK-NEXT:    bx lr
 ;
 ; CHECKBE-LABEL: vmovn64_t2:
 ; CHECKBE:       @ %bb.0: @ %entry
-; CHECKBE-NEXT:    vmov.f32 s6, s0
-; CHECKBE-NEXT:    vmov.f32 s7, s1
+; CHECKBE-NEXT:    vmov.f64 d3, d0
 ; CHECKBE-NEXT:    vmov q0, q1
 ; CHECKBE-NEXT:    bx lr
 entry:
@@ -228,14 +224,12 @@ entry:
 define arm_aapcs_vfpcc <2 x i64> @vmovn64_b1(<2 x i64> %src1, <2 x i64> %src2) {
 ; CHECK-LABEL: vmovn64_b1:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s2, s6
-; CHECK-NEXT:    vmov.f32 s3, s7
+; CHECK-NEXT:    vmov.f64 d1, d3
 ; CHECK-NEXT:    bx lr
 ;
 ; CHECKBE-LABEL: vmovn64_b1:
 ; CHECKBE:       @ %bb.0: @ %entry
-; CHECKBE-NEXT:    vmov.f32 s2, s6
-; CHECKBE-NEXT:    vmov.f32 s3, s7
+; CHECKBE-NEXT:    vmov.f64 d1, d3
 ; CHECKBE-NEXT:    bx lr
 entry:
   %out = shufflevector <2 x i64> %src1, <2 x i64> %src2, <2 x i32> <i32 0, i32 3>
@@ -245,19 +239,15 @@ entry:
 define arm_aapcs_vfpcc <2 x i64> @vmovn64_b2(<2 x i64> %src1, <2 x i64> %src2) {
 ; CHECK-LABEL: vmovn64_b2:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s4, s6
-; CHECK-NEXT:    vmov.f32 s6, s0
-; CHECK-NEXT:    vmov.f32 s5, s7
-; CHECK-NEXT:    vmov.f32 s7, s1
+; CHECK-NEXT:    vmov.f64 d2, d3
+; CHECK-NEXT:    vmov.f64 d3, d0
 ; CHECK-NEXT:    vmov q0, q1
 ; CHECK-NEXT:    bx lr
 ;
 ; CHECKBE-LABEL: vmovn64_b2:
 ; CHECKBE:       @ %bb.0: @ %entry
-; CHECKBE-NEXT:    vmov.f32 s4, s6
-; CHECKBE-NEXT:    vmov.f32 s6, s0
-; CHECKBE-NEXT:    vmov.f32 s5, s7
-; CHECKBE-NEXT:    vmov.f32 s7, s1
+; CHECKBE-NEXT:    vmov.f64 d2, d3
+; CHECKBE-NEXT:    vmov.f64 d3, d0
 ; CHECKBE-NEXT:    vmov q0, q1
 ; CHECKBE-NEXT:    bx lr
 entry:
@@ -268,18 +258,14 @@ entry:
 define arm_aapcs_vfpcc <2 x i64> @vmovn64_b3(<2 x i64> %src1, <2 x i64> %src2) {
 ; CHECK-LABEL: vmovn64_b3:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s0, s2
-; CHECK-NEXT:    vmov.f32 s2, s4
-; CHECK-NEXT:    vmov.f32 s1, s3
-; CHECK-NEXT:    vmov.f32 s3, s5
+; CHECK-NEXT:    vmov.f64 d0, d1
+; CHECK-NEXT:    vmov.f64 d1, d2
 ; CHECK-NEXT:    bx lr
 ;
 ; CHECKBE-LABEL: vmovn64_b3:
 ; CHECKBE:       @ %bb.0: @ %entry
-; CHECKBE-NEXT:    vmov.f32 s0, s2
-; CHECKBE-NEXT:    vmov.f32 s2, s4
-; CHECKBE-NEXT:    vmov.f32 s1, s3
-; CHECKBE-NEXT:    vmov.f32 s3, s5
+; CHECKBE-NEXT:    vmov.f64 d0, d1
+; CHECKBE-NEXT:    vmov.f64 d1, d2
 ; CHECKBE-NEXT:    bx lr
 entry:
   %out = shufflevector <2 x i64> %src1, <2 x i64> %src2, <2 x i32> <i32 1, i32 2>
@@ -289,15 +275,13 @@ entry:
 define arm_aapcs_vfpcc <2 x i64> @vmovn64_b4(<2 x i64> %src1, <2 x i64> %src2) {
 ; CHECK-LABEL: vmovn64_b4:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s6, s2
-; CHECK-NEXT:    vmov.f32 s7, s3
+; CHECK-NEXT:    vmov.f64 d3, d1
 ; CHECK-NEXT:    vmov q0, q1
 ; CHECK-NEXT:    bx lr
 ;
 ; CHECKBE-LABEL: vmovn64_b4:
 ; CHECKBE:       @ %bb.0: @ %entry
-; CHECKBE-NEXT:    vmov.f32 s6, s2
-; CHECKBE-NEXT:    vmov.f32 s7, s3
+; CHECKBE-NEXT:    vmov.f64 d3, d1
 ; CHECKBE-NEXT:    vmov q0, q1
 ; CHECKBE-NEXT:    bx lr
 entry:
diff --git a/llvm/test/CodeGen/Thumb2/mve-vmovnstore.ll b/llvm/test/CodeGen/Thumb2/mve-vmovnstore.ll
index 5850ad1..2428cd8 100644
--- a/llvm/test/CodeGen/Thumb2/mve-vmovnstore.ll
+++ b/llvm/test/CodeGen/Thumb2/mve-vmovnstore.ll
@@ -110,8 +110,7 @@ entry:
 define arm_aapcs_vfpcc void @vmovn64_t1(<2 x i64> %src1, <2 x i64> %src2, ptr %dest) {
 ; CHECK-LABEL: vmovn64_t1:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s2, s4
-; CHECK-NEXT:    vmov.f32 s3, s5
+; CHECK-NEXT:    vmov.f64 d1, d2
 ; CHECK-NEXT:    vstrw.32 q0, [r0]
 ; CHECK-NEXT:    bx lr
 entry:
@@ -123,8 +122,7 @@ entry:
 define arm_aapcs_vfpcc void @vmovn64_t2(<2 x i64> %src1, <2 x i64> %src2, ptr %dest) {
 ; CHECK-LABEL: vmovn64_t2:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s6, s0
-; CHECK-NEXT:    vmov.f32 s7, s1
+; CHECK-NEXT:    vmov.f64 d3, d0
 ; CHECK-NEXT:    vstrw.32 q1, [r0]
 ; CHECK-NEXT:    bx lr
 entry:
@@ -136,8 +134,7 @@ entry:
 define arm_aapcs_vfpcc void @vmovn64_b1(<2 x i64> %src1, <2 x i64> %src2, ptr %dest) {
 ; CHECK-LABEL: vmovn64_b1:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s2, s6
-; CHECK-NEXT:    vmov.f32 s3, s7
+; CHECK-NEXT:    vmov.f64 d1, d3
 ; CHECK-NEXT:    vstrw.32 q0, [r0]
 ; CHECK-NEXT:    bx lr
 entry:
@@ -149,10 +146,8 @@ entry:
 define arm_aapcs_vfpcc void @vmovn64_b2(<2 x i64> %src1, <2 x i64> %src2, ptr %dest) {
 ; CHECK-LABEL: vmovn64_b2:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s4, s6
-; CHECK-NEXT:    vmov.f32 s6, s0
-; CHECK-NEXT:    vmov.f32 s5, s7
-; CHECK-NEXT:    vmov.f32 s7, s1
+; CHECK-NEXT:    vmov.f64 d2, d3
+; CHECK-NEXT:    vmov.f64 d3, d0
 ; CHECK-NEXT:    vstrw.32 q1, [r0]
 ; CHECK-NEXT:    bx lr
 entry:
@@ -164,10 +159,8 @@ entry:
 define arm_aapcs_vfpcc void @vmovn64_b3(<2 x i64> %src1, <2 x i64> %src2, ptr %dest) {
 ; CHECK-LABEL: vmovn64_b3:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s0, s2
-; CHECK-NEXT:    vmov.f32 s2, s4
-; CHECK-NEXT:    vmov.f32 s1, s3
-; CHECK-NEXT:    vmov.f32 s3, s5
+; CHECK-NEXT:    vmov.f64 d0, d1
+; CHECK-NEXT:    vmov.f64 d1, d2
 ; CHECK-NEXT:    vstrw.32 q0, [r0]
 ; CHECK-NEXT:    bx lr
 entry:
@@ -179,8 +172,7 @@ entry:
 define arm_aapcs_vfpcc void @vmovn64_b4(<2 x i64> %src1, <2 x i64> %src2, ptr %dest) {
 ; CHECK-LABEL: vmovn64_b4:
 ; CHECK:       @ %bb.0: @ %entry
-; CHECK-NEXT:    vmov.f32 s6, s2
-; CHECK-NEXT:    vmov.f32 s7, s3
+; CHECK-NEXT:    vmov.f64 d3, d1
 ; CHECK-NEXT:    vstrw.32 q1, [r0]
 ; CHECK-NEXT:    bx lr
 entry:
diff --git a/llvm/test/CodeGen/Thumb2/vmovdrroffset.ll b/llvm/test/CodeGen/Thumb2/vmovdrroffset.ll
index 9d0c9c0..56f1ead 100644
--- a/llvm/test/CodeGen/Thumb2/vmovdrroffset.ll
+++ b/llvm/test/CodeGen/Thumb2/vmovdrroffset.ll
@@ -13,10 +13,8 @@ define arm_aapcs_vfpcc double @zero(double %a, double %b, double %c) {
 ; CHECK-NEXT:    vpush {d8, d9}
 ; CHECK-NEXT:    .pad #8
 ; CHECK-NEXT:    sub sp, #8
-; CHECK-NEXT:    vmov.f32 s16, s2
-; CHECK-NEXT:    vmov.f32 s18, s0
-; CHECK-NEXT:    vmov.f32 s17, s3
-; CHECK-NEXT:    vmov.f32 s19, s1
+; CHECK-NEXT:    vmov.f64 d8, d1
+; CHECK-NEXT:    vmov.f64 d9, d0
 ; CHECK-NEXT:    bl sqrt
 ; CHECK-NEXT:    vmov r2, r3, d8
 ; CHECK-NEXT:    vmov r0, r1, d0