mve-vecreduce-add.ll | Explore in Territory

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=thumbv8.1m.main-none-none-eabi -mattr=+mve -verify-machineinstrs %s -o - | FileCheck %s --check-prefix=CHECK

define arm_aapcs_vfpcc i32 @add_v4i32_v4i32(<4 x i32> %x) {
; CHECK-LABEL: add_v4i32_v4i32:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vaddv.u32 r0, q0
; CHECK-NEXT:    bx lr
entry:
  %z = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %x)
  ret i32 %z
}

define arm_aapcs_vfpcc i64 @add_v4i32_v4i64_zext(<4 x i32> %x) {
; CHECK-LABEL: add_v4i32_v4i64_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vaddlv.u32 r0, r1, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <4 x i32> %x to <4 x i64>
  %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %xx)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v4i32_v4i64_sext(<4 x i32> %x) {
; CHECK-LABEL: add_v4i32_v4i64_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vaddlv.s32 r0, r1, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <4 x i32> %x to <4 x i64>
  %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %xx)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v2i32_v2i64_zext(<2 x i32> %x) {
; CHECK-LABEL: add_v2i32_v2i64_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov.i64 q1, #0xffffffff
; CHECK-NEXT:    vand q0, q0, q1
; CHECK-NEXT:    vmov r0, r1, d1
; CHECK-NEXT:    vmov r2, r3, d0
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <2 x i32> %x to <2 x i64>
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %xx)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v2i32_v2i64_sext(<2 x i32> %x) {
; CHECK-LABEL: add_v2i32_v2i64_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov r0, s0
; CHECK-NEXT:    vmov r2, s2
; CHECK-NEXT:    asrs r1, r0, #31
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <2 x i32> %x to <2 x i64>
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %xx)
  ret i64 %z
}

define arm_aapcs_vfpcc i32 @add_v8i16_v8i32_zext(<8 x i16> %x) {
; CHECK-LABEL: add_v8i16_v8i32_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vaddv.u16 r0, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <8 x i16> %x to <8 x i32>
  %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %xx)
  ret i32 %z
}

define arm_aapcs_vfpcc i32 @add_v8i16_v8i32_sext(<8 x i16> %x) {
; CHECK-LABEL: add_v8i16_v8i32_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vaddv.s16 r0, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <8 x i16> %x to <8 x i32>
  %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %xx)
  ret i32 %z
}

define arm_aapcs_vfpcc i32 @add_v4i16_v4i32_zext(<4 x i16> %x) {
; CHECK-LABEL: add_v4i16_v4i32_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.u16 q0, q0
; CHECK-NEXT:    vaddv.u32 r0, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <4 x i16> %x to <4 x i32>
  %z = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %xx)
  ret i32 %z
}

define arm_aapcs_vfpcc i32 @add_v4i16_v4i32_sext(<4 x i16> %x) {
; CHECK-LABEL: add_v4i16_v4i32_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s16 q0, q0
; CHECK-NEXT:    vaddv.u32 r0, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <4 x i16> %x to <4 x i32>
  %z = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %xx)
  ret i32 %z
}

define arm_aapcs_vfpcc zeroext i16 @add_v8i16_v8i16(<8 x i16> %x) {
; CHECK-LABEL: add_v8i16_v8i16:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vaddv.u16 r0, q0
; CHECK-NEXT:    uxth r0, r0
; CHECK-NEXT:    bx lr
entry:
  %z = call i16 @llvm.vector.reduce.add.v8i16(<8 x i16> %x)
  ret i16 %z
}

define arm_aapcs_vfpcc i64 @add_v8i16_v8i64_zext(<8 x i16> %x) {
; CHECK-LABEL: add_v8i16_v8i64_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov.u16 r0, q0[1]
; CHECK-NEXT:    vmov.u16 r1, q0[0]
; CHECK-NEXT:    vmov q2[2], q2[0], r1, r0
; CHECK-NEXT:    vmov.i64 q1, #0xffff
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov.u16 r3, q0[2]
; CHECK-NEXT:    vmov r0, s10
; CHECK-NEXT:    vmov r1, r2, d4
; CHECK-NEXT:    add r0, r1
; CHECK-NEXT:    vmov.u16 r1, q0[3]
; CHECK-NEXT:    vmov q2[2], q2[0], r3, r1
; CHECK-NEXT:    vmov.u16 r3, q0[4]
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov r1, s8
; CHECK-NEXT:    add r0, r1
; CHECK-NEXT:    vmov r1, s10
; CHECK-NEXT:    add r0, r1
; CHECK-NEXT:    vmov.u16 r1, q0[5]
; CHECK-NEXT:    vmov q2[2], q2[0], r3, r1
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov r1, s8
; CHECK-NEXT:    add r0, r1
; CHECK-NEXT:    vmov r1, r3, d5
; CHECK-NEXT:    adds r0, r0, r1
; CHECK-NEXT:    adc.w r1, r2, r3
; CHECK-NEXT:    vmov.u16 r2, q0[7]
; CHECK-NEXT:    vmov.u16 r3, q0[6]
; CHECK-NEXT:    vmov q0[2], q0[0], r3, r2
; CHECK-NEXT:    vand q0, q0, q1
; CHECK-NEXT:    vmov r2, r3, d0
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    vmov r2, r3, d1
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <8 x i16> %x to <8 x i64>
  %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %xx)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v8i16_v8i64_sext(<8 x i16> %x) {
; CHECK-LABEL: add_v8i16_v8i64_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov.s16 r0, q0[0]
; CHECK-NEXT:    vmov.s16 r2, q0[1]
; CHECK-NEXT:    asrs r1, r0, #31
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.s16 r2, q0[2]
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.s16 r2, q0[3]
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.s16 r2, q0[4]
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.s16 r2, q0[5]
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.s16 r2, q0[6]
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.s16 r2, q0[7]
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <8 x i16> %x to <8 x i64>
  %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %xx)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v4i16_v4i64_zext(<4 x i16> %x) {
; CHECK-LABEL: add_v4i16_v4i64_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.u16 q0, q0
; CHECK-NEXT:    vaddlv.u32 r0, r1, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <4 x i16> %x to <4 x i64>
  %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %xx)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v4i16_v4i64_sext(<4 x i16> %x) {
; CHECK-LABEL: add_v4i16_v4i64_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s16 q0, q0
; CHECK-NEXT:    vaddlv.s32 r0, r1, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <4 x i16> %x to <4 x i64>
  %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %xx)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v2i16_v2i64_zext(<2 x i16> %x) {
; CHECK-LABEL: add_v2i16_v2i64_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov.i64 q1, #0xffff
; CHECK-NEXT:    vand q0, q0, q1
; CHECK-NEXT:    vmov r0, s2
; CHECK-NEXT:    vmov r2, r1, d0
; CHECK-NEXT:    add r0, r2
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <2 x i16> %x to <2 x i64>
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %xx)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v2i16_v2i64_sext(<2 x i16> %x) {
; CHECK-LABEL: add_v2i16_v2i64_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov r0, s0
; CHECK-NEXT:    vmov r2, s2
; CHECK-NEXT:    sxth r0, r0
; CHECK-NEXT:    asrs r1, r0, #31
; CHECK-NEXT:    sxth r2, r2
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <2 x i16> %x to <2 x i64>
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %xx)
  ret i64 %z
}

define arm_aapcs_vfpcc i32 @add_v16i8_v16i32_zext(<16 x i8> %x) {
; CHECK-LABEL: add_v16i8_v16i32_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vaddv.u8 r0, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <16 x i8> %x to <16 x i32>
  %z = call i32 @llvm.vector.reduce.add.v16i32(<16 x i32> %xx)
  ret i32 %z
}

define arm_aapcs_vfpcc i32 @add_v16i8_v16i32_sext(<16 x i8> %x) {
; CHECK-LABEL: add_v16i8_v16i32_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vaddv.s8 r0, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <16 x i8> %x to <16 x i32>
  %z = call i32 @llvm.vector.reduce.add.v16i32(<16 x i32> %xx)
  ret i32 %z
}

define arm_aapcs_vfpcc i32 @add_v8i8_v8i32_zext(<8 x i8> %x) {
; CHECK-LABEL: add_v8i8_v8i32_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.u8 q0, q0
; CHECK-NEXT:    vaddv.u16 r0, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <8 x i8> %x to <8 x i32>
  %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %xx)
  ret i32 %z
}

define arm_aapcs_vfpcc i32 @add_v8i8_v8i32_sext(<8 x i8> %x) {
; CHECK-LABEL: add_v8i8_v8i32_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s8 q0, q0
; CHECK-NEXT:    vaddv.s16 r0, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <8 x i8> %x to <8 x i32>
  %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %xx)
  ret i32 %z
}

define arm_aapcs_vfpcc i32 @add_v4i8_v4i32_zext(<4 x i8> %x) {
; CHECK-LABEL: add_v4i8_v4i32_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov.i32 q1, #0xff
; CHECK-NEXT:    vand q0, q0, q1
; CHECK-NEXT:    vaddv.u32 r0, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <4 x i8> %x to <4 x i32>
  %z = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %xx)
  ret i32 %z
}

define arm_aapcs_vfpcc i32 @add_v4i8_v4i32_sext(<4 x i8> %x) {
; CHECK-LABEL: add_v4i8_v4i32_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s8 q0, q0
; CHECK-NEXT:    vmovlb.s16 q0, q0
; CHECK-NEXT:    vaddv.u32 r0, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <4 x i8> %x to <4 x i32>
  %z = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %xx)
  ret i32 %z
}

define arm_aapcs_vfpcc zeroext i16 @add_v16i8_v16i16_zext(<16 x i8> %x) {
; CHECK-LABEL: add_v16i8_v16i16_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vaddv.u8 r0, q0
; CHECK-NEXT:    uxth r0, r0
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <16 x i8> %x to <16 x i16>
  %z = call i16 @llvm.vector.reduce.add.v16i16(<16 x i16> %xx)
  ret i16 %z
}

define arm_aapcs_vfpcc signext i16 @add_v16i8_v16i16_sext(<16 x i8> %x) {
; CHECK-LABEL: add_v16i8_v16i16_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vaddv.s8 r0, q0
; CHECK-NEXT:    sxth r0, r0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <16 x i8> %x to <16 x i16>
  %z = call i16 @llvm.vector.reduce.add.v16i16(<16 x i16> %xx)
  ret i16 %z
}

define arm_aapcs_vfpcc zeroext i16 @add_v8i8_v8i16_zext(<8 x i8> %x) {
; CHECK-LABEL: add_v8i8_v8i16_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.u8 q0, q0
; CHECK-NEXT:    vaddv.u16 r0, q0
; CHECK-NEXT:    uxth r0, r0
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <8 x i8> %x to <8 x i16>
  %z = call i16 @llvm.vector.reduce.add.v8i16(<8 x i16> %xx)
  ret i16 %z
}

define arm_aapcs_vfpcc signext i16 @add_v8i8_v8i16_sext(<8 x i8> %x) {
; CHECK-LABEL: add_v8i8_v8i16_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s8 q0, q0
; CHECK-NEXT:    vaddv.u16 r0, q0
; CHECK-NEXT:    sxth r0, r0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <8 x i8> %x to <8 x i16>
  %z = call i16 @llvm.vector.reduce.add.v8i16(<8 x i16> %xx)
  ret i16 %z
}

define arm_aapcs_vfpcc zeroext i8 @add_v16i8_v16i8(<16 x i8> %x) {
; CHECK-LABEL: add_v16i8_v16i8:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vaddv.u8 r0, q0
; CHECK-NEXT:    uxtb r0, r0
; CHECK-NEXT:    bx lr
entry:
  %z = call i8 @llvm.vector.reduce.add.v16i8(<16 x i8> %x)
  ret i8 %z
}

define arm_aapcs_vfpcc i64 @add_v16i8_v16i64_zext(<16 x i8> %x) {
; CHECK-LABEL: add_v16i8_v16i64_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov.u8 r0, q0[1]
; CHECK-NEXT:    vmov.u8 r1, q0[0]
; CHECK-NEXT:    vmov q2[2], q2[0], r1, r0
; CHECK-NEXT:    vmov.i64 q1, #0xff
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov.u8 r3, q0[2]
; CHECK-NEXT:    vmov r0, s10
; CHECK-NEXT:    vmov r1, r2, d4
; CHECK-NEXT:    add r0, r1
; CHECK-NEXT:    vmov.u8 r1, q0[3]
; CHECK-NEXT:    vmov q2[2], q2[0], r3, r1
; CHECK-NEXT:    vmov.u8 r3, q0[4]
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov r1, s8
; CHECK-NEXT:    add r0, r1
; CHECK-NEXT:    vmov r1, s10
; CHECK-NEXT:    add r0, r1
; CHECK-NEXT:    vmov.u8 r1, q0[5]
; CHECK-NEXT:    vmov q2[2], q2[0], r3, r1
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov r1, s8
; CHECK-NEXT:    add r0, r1
; CHECK-NEXT:    vmov r1, r3, d5
; CHECK-NEXT:    adds r0, r0, r1
; CHECK-NEXT:    adc.w r1, r2, r3
; CHECK-NEXT:    vmov.u8 r2, q0[7]
; CHECK-NEXT:    vmov.u8 r3, q0[6]
; CHECK-NEXT:    vmov q2[2], q2[0], r3, r2
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov r2, r3, d4
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    vmov r2, r3, d5
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    vmov.u8 r2, q0[9]
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    vmov.u8 r3, q0[8]
; CHECK-NEXT:    vmov q2[2], q2[0], r3, r2
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov r2, r3, d4
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    vmov r2, r3, d5
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    vmov.u8 r2, q0[11]
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    vmov.u8 r3, q0[10]
; CHECK-NEXT:    vmov q2[2], q2[0], r3, r2
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov r2, r3, d4
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    vmov r2, r3, d5
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    vmov.u8 r2, q0[13]
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    vmov.u8 r3, q0[12]
; CHECK-NEXT:    vmov q2[2], q2[0], r3, r2
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov r2, r3, d4
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    vmov r2, r3, d5
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    vmov.u8 r2, q0[15]
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    vmov.u8 r3, q0[14]
; CHECK-NEXT:    vmov q0[2], q0[0], r3, r2
; CHECK-NEXT:    vand q0, q0, q1
; CHECK-NEXT:    vmov r2, r3, d0
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    vmov r2, r3, d1
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <16 x i8> %x to <16 x i64>
  %z = call i64 @llvm.vector.reduce.add.v16i64(<16 x i64> %xx)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v16i8_v16i64_sext(<16 x i8> %x) {
; CHECK-LABEL: add_v16i8_v16i64_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov.s8 r0, q0[0]
; CHECK-NEXT:    vmov.s8 r2, q0[1]
; CHECK-NEXT:    asrs r1, r0, #31
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.s8 r2, q0[2]
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.s8 r2, q0[3]
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.s8 r2, q0[4]
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.s8 r2, q0[5]
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.s8 r2, q0[6]
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.s8 r2, q0[7]
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.s8 r2, q0[8]
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.s8 r2, q0[9]
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.s8 r2, q0[10]
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.s8 r2, q0[11]
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.s8 r2, q0[12]
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.s8 r2, q0[13]
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.s8 r2, q0[14]
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.s8 r2, q0[15]
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <16 x i8> %x to <16 x i64>
  %z = call i64 @llvm.vector.reduce.add.v16i64(<16 x i64> %xx)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v8i8_v8i64_zext(<8 x i8> %x) {
; CHECK-LABEL: add_v8i8_v8i64_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.u8 q0, q0
; CHECK-NEXT:    vmov.i64 q1, #0xffff
; CHECK-NEXT:    vmov.u16 r0, q0[1]
; CHECK-NEXT:    vmov.u16 r1, q0[0]
; CHECK-NEXT:    vmov q2[2], q2[0], r1, r0
; CHECK-NEXT:    vmov.u16 r3, q0[2]
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov r0, s10
; CHECK-NEXT:    vmov r1, r2, d4
; CHECK-NEXT:    add r0, r1
; CHECK-NEXT:    vmov.u16 r1, q0[3]
; CHECK-NEXT:    vmov q2[2], q2[0], r3, r1
; CHECK-NEXT:    vmov.u16 r3, q0[4]
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov r1, s8
; CHECK-NEXT:    add r0, r1
; CHECK-NEXT:    vmov r1, s10
; CHECK-NEXT:    add r0, r1
; CHECK-NEXT:    vmov.u16 r1, q0[5]
; CHECK-NEXT:    vmov q2[2], q2[0], r3, r1
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov r1, s8
; CHECK-NEXT:    add r0, r1
; CHECK-NEXT:    vmov r1, r3, d5
; CHECK-NEXT:    adds r0, r0, r1
; CHECK-NEXT:    adc.w r1, r2, r3
; CHECK-NEXT:    vmov.u16 r2, q0[7]
; CHECK-NEXT:    vmov.u16 r3, q0[6]
; CHECK-NEXT:    vmov q0[2], q0[0], r3, r2
; CHECK-NEXT:    vand q0, q0, q1
; CHECK-NEXT:    vmov r2, r3, d0
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    vmov r2, r3, d1
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <8 x i8> %x to <8 x i64>
  %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %xx)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v8i8_v8i64_sext(<8 x i8> %x) {
; CHECK-LABEL: add_v8i8_v8i64_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov.u16 r0, q0[0]
; CHECK-NEXT:    vmov.u16 r2, q0[1]
; CHECK-NEXT:    sxtb r0, r0
; CHECK-NEXT:    sxtb r2, r2
; CHECK-NEXT:    asrs r1, r0, #31
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.u16 r2, q0[2]
; CHECK-NEXT:    sxtb r2, r2
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.u16 r2, q0[3]
; CHECK-NEXT:    sxtb r2, r2
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.u16 r2, q0[4]
; CHECK-NEXT:    sxtb r2, r2
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.u16 r2, q0[5]
; CHECK-NEXT:    sxtb r2, r2
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.u16 r2, q0[6]
; CHECK-NEXT:    sxtb r2, r2
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    vmov.u16 r2, q0[7]
; CHECK-NEXT:    sxtb r2, r2
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <8 x i8> %x to <8 x i64>
  %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %xx)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v4i8_v4i64_zext(<4 x i8> %x) {
; CHECK-LABEL: add_v4i8_v4i64_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov.i32 q1, #0xff
; CHECK-NEXT:    vand q0, q0, q1
; CHECK-NEXT:    vaddlv.u32 r0, r1, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <4 x i8> %x to <4 x i64>
  %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %xx)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v4i8_v4i64_sext(<4 x i8> %x) {
; CHECK-LABEL: add_v4i8_v4i64_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s8 q0, q0
; CHECK-NEXT:    vmovlb.s16 q0, q0
; CHECK-NEXT:    vaddlv.s32 r0, r1, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <4 x i8> %x to <4 x i64>
  %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %xx)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v2i8_v2i64_zext(<2 x i8> %x) {
; CHECK-LABEL: add_v2i8_v2i64_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov.i64 q1, #0xff
; CHECK-NEXT:    vand q0, q0, q1
; CHECK-NEXT:    vmov r0, s2
; CHECK-NEXT:    vmov r2, r1, d0
; CHECK-NEXT:    add r0, r2
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <2 x i8> %x to <2 x i64>
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %xx)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v2i8_v2i64_sext(<2 x i8> %x) {
; CHECK-LABEL: add_v2i8_v2i64_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov r0, s0
; CHECK-NEXT:    vmov r2, s2
; CHECK-NEXT:    sxtb r0, r0
; CHECK-NEXT:    asrs r1, r0, #31
; CHECK-NEXT:    sxtb r2, r2
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r2, asr #31
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <2 x i8> %x to <2 x i64>
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %xx)
  ret i64 %z
}

define arm_aapcs_vfpcc i64 @add_v2i64_v2i64(<2 x i64> %x) {
; CHECK-LABEL: add_v2i64_v2i64:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov r0, r1, d1
; CHECK-NEXT:    vmov r2, r3, d0
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    bx lr
entry:
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %x)
  ret i64 %z
}

define arm_aapcs_vfpcc i32 @add_v4i32_v4i32_acc(<4 x i32> %x, i32 %a) {
; CHECK-LABEL: add_v4i32_v4i32_acc:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vaddva.u32 r0, q0
; CHECK-NEXT:    bx lr
entry:
  %z = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %x)
  %r = add i32 %z, %a
  ret i32 %r
}

define arm_aapcs_vfpcc i64 @add_v4i32_v4i64_acc_zext(<4 x i32> %x, i64 %a) {
; CHECK-LABEL: add_v4i32_v4i64_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vaddlva.u32 r0, r1, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <4 x i32> %x to <4 x i64>
  %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %xx)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v4i32_v4i64_acc_sext(<4 x i32> %x, i64 %a) {
; CHECK-LABEL: add_v4i32_v4i64_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vaddlva.s32 r0, r1, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <4 x i32> %x to <4 x i64>
  %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %xx)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v2i32_v2i64_acc_zext(<2 x i32> %x, i64 %a) {
; CHECK-LABEL: add_v2i32_v2i64_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r7, lr}
; CHECK-NEXT:    push {r7, lr}
; CHECK-NEXT:    vmov.i64 q1, #0xffffffff
; CHECK-NEXT:    vand q0, q0, q1
; CHECK-NEXT:    vmov lr, r12, d1
; CHECK-NEXT:    vmov r3, r2, d0
; CHECK-NEXT:    adds.w r3, r3, lr
; CHECK-NEXT:    adc.w r2, r2, r12
; CHECK-NEXT:    adds r0, r0, r3
; CHECK-NEXT:    adcs r1, r2
; CHECK-NEXT:    pop {r7, pc}
entry:
  %xx = zext <2 x i32> %x to <2 x i64>
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %xx)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v2i32_v2i64_acc_sext(<2 x i32> %x, i64 %a) {
; CHECK-LABEL: add_v2i32_v2i64_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov r2, s0
; CHECK-NEXT:    vmov r3, s2
; CHECK-NEXT:    asr.w r12, r2, #31
; CHECK-NEXT:    adds r2, r2, r3
; CHECK-NEXT:    adc.w r3, r12, r3, asr #31
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <2 x i32> %x to <2 x i64>
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %xx)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i32 @add_v8i16_v8i32_acc_zext(<8 x i16> %x, i32 %a) {
; CHECK-LABEL: add_v8i16_v8i32_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vaddva.u16 r0, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <8 x i16> %x to <8 x i32>
  %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %xx)
  %r = add i32 %z, %a
  ret i32 %r
}

define arm_aapcs_vfpcc i32 @add_v8i16_v8i32_acc_sext(<8 x i16> %x, i32 %a) {
; CHECK-LABEL: add_v8i16_v8i32_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vaddva.s16 r0, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <8 x i16> %x to <8 x i32>
  %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %xx)
  %r = add i32 %z, %a
  ret i32 %r
}

define arm_aapcs_vfpcc i32 @add_v4i16_v4i32_acc_zext(<4 x i16> %x, i32 %a) {
; CHECK-LABEL: add_v4i16_v4i32_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.u16 q0, q0
; CHECK-NEXT:    vaddva.u32 r0, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <4 x i16> %x to <4 x i32>
  %z = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %xx)
  %r = add i32 %z, %a
  ret i32 %r
}

define arm_aapcs_vfpcc i32 @add_v4i16_v4i32_acc_sext(<4 x i16> %x, i32 %a) {
; CHECK-LABEL: add_v4i16_v4i32_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s16 q0, q0
; CHECK-NEXT:    vaddva.u32 r0, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <4 x i16> %x to <4 x i32>
  %z = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %xx)
  %r = add i32 %z, %a
  ret i32 %r
}

define arm_aapcs_vfpcc zeroext i16 @add_v8i16_v8i16_acc(<8 x i16> %x, i16 %a) {
; CHECK-LABEL: add_v8i16_v8i16_acc:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vaddva.u16 r0, q0
; CHECK-NEXT:    uxth r0, r0
; CHECK-NEXT:    bx lr
entry:
  %z = call i16 @llvm.vector.reduce.add.v8i16(<8 x i16> %x)
  %r = add i16 %z, %a
  ret i16 %r
}

define arm_aapcs_vfpcc i64 @add_v8i16_v8i64_acc_zext(<8 x i16> %x, i64 %a) {
; CHECK-LABEL: add_v8i16_v8i64_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r7, lr}
; CHECK-NEXT:    push {r7, lr}
; CHECK-NEXT:    vmov.u16 r2, q0[1]
; CHECK-NEXT:    vmov.u16 r3, q0[0]
; CHECK-NEXT:    vmov q2[2], q2[0], r3, r2
; CHECK-NEXT:    vmov.i64 q1, #0xffff
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov r2, s10
; CHECK-NEXT:    vmov r3, r12, d4
; CHECK-NEXT:    add.w lr, r3, r2
; CHECK-NEXT:    vmov.u16 r3, q0[3]
; CHECK-NEXT:    vmov.u16 r2, q0[2]
; CHECK-NEXT:    vmov q2[2], q2[0], r2, r3
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov r2, s8
; CHECK-NEXT:    vmov r3, s10
; CHECK-NEXT:    add r2, lr
; CHECK-NEXT:    add.w lr, r2, r3
; CHECK-NEXT:    vmov.u16 r3, q0[5]
; CHECK-NEXT:    vmov.u16 r2, q0[4]
; CHECK-NEXT:    vmov q2[2], q2[0], r2, r3
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov r2, s8
; CHECK-NEXT:    add lr, r2
; CHECK-NEXT:    vmov r3, r2, d5
; CHECK-NEXT:    adds.w lr, lr, r3
; CHECK-NEXT:    vmov.u16 r3, q0[6]
; CHECK-NEXT:    adc.w r12, r12, r2
; CHECK-NEXT:    vmov.u16 r2, q0[7]
; CHECK-NEXT:    vmov q0[2], q0[0], r3, r2
; CHECK-NEXT:    vand q0, q0, q1
; CHECK-NEXT:    vmov r2, r3, d0
; CHECK-NEXT:    adds.w lr, lr, r2
; CHECK-NEXT:    adc.w r12, r12, r3
; CHECK-NEXT:    vmov r2, r3, d1
; CHECK-NEXT:    adds.w r2, r2, lr
; CHECK-NEXT:    adc.w r3, r3, r12
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    pop {r7, pc}
entry:
  %xx = zext <8 x i16> %x to <8 x i64>
  %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %xx)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v8i16_v8i64_acc_sext(<8 x i16> %x, i64 %a) {
; CHECK-LABEL: add_v8i16_v8i64_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r7, lr}
; CHECK-NEXT:    push {r7, lr}
; CHECK-NEXT:    vmov.s16 r2, q0[0]
; CHECK-NEXT:    vmov.s16 r3, q0[1]
; CHECK-NEXT:    asr.w r12, r2, #31
; CHECK-NEXT:    adds.w lr, r2, r3
; CHECK-NEXT:    vmov.s16 r2, q0[2]
; CHECK-NEXT:    adc.w r3, r12, r3, asr #31
; CHECK-NEXT:    adds.w r12, lr, r2
; CHECK-NEXT:    adc.w r2, r3, r2, asr #31
; CHECK-NEXT:    vmov.s16 r3, q0[3]
; CHECK-NEXT:    adds.w r12, r12, r3
; CHECK-NEXT:    adc.w r2, r2, r3, asr #31
; CHECK-NEXT:    vmov.s16 r3, q0[4]
; CHECK-NEXT:    adds.w r12, r12, r3
; CHECK-NEXT:    adc.w r2, r2, r3, asr #31
; CHECK-NEXT:    vmov.s16 r3, q0[5]
; CHECK-NEXT:    adds.w r12, r12, r3
; CHECK-NEXT:    adc.w r2, r2, r3, asr #31
; CHECK-NEXT:    vmov.s16 r3, q0[6]
; CHECK-NEXT:    adds.w r12, r12, r3
; CHECK-NEXT:    adc.w lr, r2, r3, asr #31
; CHECK-NEXT:    vmov.s16 r3, q0[7]
; CHECK-NEXT:    adds.w r2, r12, r3
; CHECK-NEXT:    adc.w r3, lr, r3, asr #31
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    pop {r7, pc}
entry:
  %xx = sext <8 x i16> %x to <8 x i64>
  %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %xx)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v4i16_v4i64_acc_zext(<4 x i16> %x, i64 %a) {
; CHECK-LABEL: add_v4i16_v4i64_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.u16 q0, q0
; CHECK-NEXT:    vaddlva.u32 r0, r1, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <4 x i16> %x to <4 x i64>
  %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %xx)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v4i16_v4i64_acc_sext(<4 x i16> %x, i64 %a) {
; CHECK-LABEL: add_v4i16_v4i64_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s16 q0, q0
; CHECK-NEXT:    vaddlva.s32 r0, r1, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <4 x i16> %x to <4 x i64>
  %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %xx)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v2i16_v2i64_acc_zext(<2 x i16> %x, i64 %a) {
; CHECK-LABEL: add_v2i16_v2i64_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov.i64 q1, #0xffff
; CHECK-NEXT:    vand q0, q0, q1
; CHECK-NEXT:    vmov r2, s2
; CHECK-NEXT:    vmov r3, r12, d0
; CHECK-NEXT:    add r2, r3
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r12
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <2 x i16> %x to <2 x i64>
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %xx)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v2i16_v2i64_acc_sext(<2 x i16> %x, i64 %a) {
; CHECK-LABEL: add_v2i16_v2i64_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov r2, s0
; CHECK-NEXT:    vmov r3, s2
; CHECK-NEXT:    sxth r2, r2
; CHECK-NEXT:    asr.w r12, r2, #31
; CHECK-NEXT:    sxth r3, r3
; CHECK-NEXT:    adds r2, r2, r3
; CHECK-NEXT:    adc.w r3, r12, r3, asr #31
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <2 x i16> %x to <2 x i64>
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %xx)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i32 @add_v16i8_v16i32_acc_zext(<16 x i8> %x, i32 %a) {
; CHECK-LABEL: add_v16i8_v16i32_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vaddva.u8 r0, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <16 x i8> %x to <16 x i32>
  %z = call i32 @llvm.vector.reduce.add.v16i32(<16 x i32> %xx)
  %r = add i32 %z, %a
  ret i32 %r
}

define arm_aapcs_vfpcc i32 @add_v16i8_v16i32_acc_sext(<16 x i8> %x, i32 %a) {
; CHECK-LABEL: add_v16i8_v16i32_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vaddva.s8 r0, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <16 x i8> %x to <16 x i32>
  %z = call i32 @llvm.vector.reduce.add.v16i32(<16 x i32> %xx)
  %r = add i32 %z, %a
  ret i32 %r
}

define arm_aapcs_vfpcc i32 @add_v8i8_v8i32_acc_zext(<8 x i8> %x, i32 %a) {
; CHECK-LABEL: add_v8i8_v8i32_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.u8 q0, q0
; CHECK-NEXT:    vaddva.u16 r0, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <8 x i8> %x to <8 x i32>
  %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %xx)
  %r = add i32 %z, %a
  ret i32 %r
}

define arm_aapcs_vfpcc i32 @add_v8i8_v8i32_acc_sext(<8 x i8> %x, i32 %a) {
; CHECK-LABEL: add_v8i8_v8i32_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s8 q0, q0
; CHECK-NEXT:    vaddva.s16 r0, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <8 x i8> %x to <8 x i32>
  %z = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %xx)
  %r = add i32 %z, %a
  ret i32 %r
}

define arm_aapcs_vfpcc i32 @add_v4i8_v4i32_acc_zext(<4 x i8> %x, i32 %a) {
; CHECK-LABEL: add_v4i8_v4i32_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov.i32 q1, #0xff
; CHECK-NEXT:    vand q0, q0, q1
; CHECK-NEXT:    vaddva.u32 r0, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <4 x i8> %x to <4 x i32>
  %z = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %xx)
  %r = add i32 %z, %a
  ret i32 %r
}

define arm_aapcs_vfpcc i32 @add_v4i8_v4i32_acc_sext(<4 x i8> %x, i32 %a) {
; CHECK-LABEL: add_v4i8_v4i32_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s8 q0, q0
; CHECK-NEXT:    vmovlb.s16 q0, q0
; CHECK-NEXT:    vaddva.u32 r0, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <4 x i8> %x to <4 x i32>
  %z = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %xx)
  %r = add i32 %z, %a
  ret i32 %r
}

define arm_aapcs_vfpcc zeroext i16 @add_v16i8_v16i16_acc_zext(<16 x i8> %x, i16 %a) {
; CHECK-LABEL: add_v16i8_v16i16_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vaddva.u8 r0, q0
; CHECK-NEXT:    uxth r0, r0
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <16 x i8> %x to <16 x i16>
  %z = call i16 @llvm.vector.reduce.add.v16i16(<16 x i16> %xx)
  %r = add i16 %z, %a
  ret i16 %r
}

define arm_aapcs_vfpcc signext i16 @add_v16i8_v16i16_acc_sext(<16 x i8> %x, i16 %a) {
; CHECK-LABEL: add_v16i8_v16i16_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vaddva.s8 r0, q0
; CHECK-NEXT:    sxth r0, r0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <16 x i8> %x to <16 x i16>
  %z = call i16 @llvm.vector.reduce.add.v16i16(<16 x i16> %xx)
  %r = add i16 %z, %a
  ret i16 %r
}

define arm_aapcs_vfpcc zeroext i16 @add_v8i8_v8i16_acc_zext(<8 x i8> %x, i16 %a) {
; CHECK-LABEL: add_v8i8_v8i16_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.u8 q0, q0
; CHECK-NEXT:    vaddva.u16 r0, q0
; CHECK-NEXT:    uxth r0, r0
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <8 x i8> %x to <8 x i16>
  %z = call i16 @llvm.vector.reduce.add.v8i16(<8 x i16> %xx)
  %r = add i16 %z, %a
  ret i16 %r
}

define arm_aapcs_vfpcc signext i16 @add_v8i8_v8i16_acc_sext(<8 x i8> %x, i16 %a) {
; CHECK-LABEL: add_v8i8_v8i16_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s8 q0, q0
; CHECK-NEXT:    vaddva.u16 r0, q0
; CHECK-NEXT:    sxth r0, r0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <8 x i8> %x to <8 x i16>
  %z = call i16 @llvm.vector.reduce.add.v8i16(<8 x i16> %xx)
  %r = add i16 %z, %a
  ret i16 %r
}

define arm_aapcs_vfpcc zeroext i8 @add_v16i8_v16i8_acc(<16 x i8> %x, i8 %a) {
; CHECK-LABEL: add_v16i8_v16i8_acc:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vaddva.u8 r0, q0
; CHECK-NEXT:    uxtb r0, r0
; CHECK-NEXT:    bx lr
entry:
  %z = call i8 @llvm.vector.reduce.add.v16i8(<16 x i8> %x)
  %r = add i8 %z, %a
  ret i8 %r
}

define arm_aapcs_vfpcc i64 @add_v16i8_v16i64_acc_zext(<16 x i8> %x, i64 %a) {
; CHECK-LABEL: add_v16i8_v16i64_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r7, lr}
; CHECK-NEXT:    push {r7, lr}
; CHECK-NEXT:    vmov.u8 r2, q0[1]
; CHECK-NEXT:    vmov.u8 r3, q0[0]
; CHECK-NEXT:    vmov q2[2], q2[0], r3, r2
; CHECK-NEXT:    vmov.i64 q1, #0xff
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov r2, s10
; CHECK-NEXT:    vmov r3, r12, d4
; CHECK-NEXT:    add.w lr, r3, r2
; CHECK-NEXT:    vmov.u8 r3, q0[3]
; CHECK-NEXT:    vmov.u8 r2, q0[2]
; CHECK-NEXT:    vmov q2[2], q2[0], r2, r3
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov r2, s8
; CHECK-NEXT:    vmov r3, s10
; CHECK-NEXT:    add r2, lr
; CHECK-NEXT:    add.w lr, r2, r3
; CHECK-NEXT:    vmov.u8 r3, q0[5]
; CHECK-NEXT:    vmov.u8 r2, q0[4]
; CHECK-NEXT:    vmov q2[2], q2[0], r2, r3
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov r2, s8
; CHECK-NEXT:    add lr, r2
; CHECK-NEXT:    vmov r3, r2, d5
; CHECK-NEXT:    adds.w lr, lr, r3
; CHECK-NEXT:    vmov.u8 r3, q0[6]
; CHECK-NEXT:    adc.w r12, r12, r2
; CHECK-NEXT:    vmov.u8 r2, q0[7]
; CHECK-NEXT:    vmov q2[2], q2[0], r3, r2
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov r2, r3, d4
; CHECK-NEXT:    adds.w lr, lr, r2
; CHECK-NEXT:    adc.w r12, r12, r3
; CHECK-NEXT:    vmov r2, r3, d5
; CHECK-NEXT:    adds.w lr, lr, r2
; CHECK-NEXT:    vmov.u8 r2, q0[9]
; CHECK-NEXT:    adc.w r12, r12, r3
; CHECK-NEXT:    vmov.u8 r3, q0[8]
; CHECK-NEXT:    vmov q2[2], q2[0], r3, r2
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov r2, r3, d4
; CHECK-NEXT:    adds.w lr, lr, r2
; CHECK-NEXT:    adc.w r12, r12, r3
; CHECK-NEXT:    vmov r2, r3, d5
; CHECK-NEXT:    adds.w lr, lr, r2
; CHECK-NEXT:    vmov.u8 r2, q0[11]
; CHECK-NEXT:    adc.w r12, r12, r3
; CHECK-NEXT:    vmov.u8 r3, q0[10]
; CHECK-NEXT:    vmov q2[2], q2[0], r3, r2
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov r2, r3, d4
; CHECK-NEXT:    adds.w lr, lr, r2
; CHECK-NEXT:    adc.w r12, r12, r3
; CHECK-NEXT:    vmov r2, r3, d5
; CHECK-NEXT:    adds.w lr, lr, r2
; CHECK-NEXT:    vmov.u8 r2, q0[13]
; CHECK-NEXT:    adc.w r12, r12, r3
; CHECK-NEXT:    vmov.u8 r3, q0[12]
; CHECK-NEXT:    vmov q2[2], q2[0], r3, r2
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov r2, r3, d4
; CHECK-NEXT:    adds.w lr, lr, r2
; CHECK-NEXT:    adc.w r12, r12, r3
; CHECK-NEXT:    vmov r2, r3, d5
; CHECK-NEXT:    adds.w lr, lr, r2
; CHECK-NEXT:    vmov.u8 r2, q0[15]
; CHECK-NEXT:    adc.w r12, r12, r3
; CHECK-NEXT:    vmov.u8 r3, q0[14]
; CHECK-NEXT:    vmov q0[2], q0[0], r3, r2
; CHECK-NEXT:    vand q0, q0, q1
; CHECK-NEXT:    vmov r2, r3, d0
; CHECK-NEXT:    adds.w lr, lr, r2
; CHECK-NEXT:    adc.w r12, r12, r3
; CHECK-NEXT:    vmov r2, r3, d1
; CHECK-NEXT:    adds.w r2, r2, lr
; CHECK-NEXT:    adc.w r3, r3, r12
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    pop {r7, pc}
entry:
  %xx = zext <16 x i8> %x to <16 x i64>
  %z = call i64 @llvm.vector.reduce.add.v16i64(<16 x i64> %xx)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v16i8_v16i64_acc_sext(<16 x i8> %x, i64 %a) {
; CHECK-LABEL: add_v16i8_v16i64_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r7, lr}
; CHECK-NEXT:    push {r7, lr}
; CHECK-NEXT:    vmov.s8 r2, q0[0]
; CHECK-NEXT:    vmov.s8 r3, q0[1]
; CHECK-NEXT:    asr.w r12, r2, #31
; CHECK-NEXT:    adds.w lr, r2, r3
; CHECK-NEXT:    vmov.s8 r2, q0[2]
; CHECK-NEXT:    adc.w r3, r12, r3, asr #31
; CHECK-NEXT:    adds.w r12, lr, r2
; CHECK-NEXT:    adc.w r2, r3, r2, asr #31
; CHECK-NEXT:    vmov.s8 r3, q0[3]
; CHECK-NEXT:    adds.w r12, r12, r3
; CHECK-NEXT:    adc.w r2, r2, r3, asr #31
; CHECK-NEXT:    vmov.s8 r3, q0[4]
; CHECK-NEXT:    adds.w r12, r12, r3
; CHECK-NEXT:    adc.w r2, r2, r3, asr #31
; CHECK-NEXT:    vmov.s8 r3, q0[5]
; CHECK-NEXT:    adds.w r12, r12, r3
; CHECK-NEXT:    adc.w r2, r2, r3, asr #31
; CHECK-NEXT:    vmov.s8 r3, q0[6]
; CHECK-NEXT:    adds.w r12, r12, r3
; CHECK-NEXT:    adc.w r2, r2, r3, asr #31
; CHECK-NEXT:    vmov.s8 r3, q0[7]
; CHECK-NEXT:    adds.w r12, r12, r3
; CHECK-NEXT:    adc.w r2, r2, r3, asr #31
; CHECK-NEXT:    vmov.s8 r3, q0[8]
; CHECK-NEXT:    adds.w r12, r12, r3
; CHECK-NEXT:    adc.w r2, r2, r3, asr #31
; CHECK-NEXT:    vmov.s8 r3, q0[9]
; CHECK-NEXT:    adds.w r12, r12, r3
; CHECK-NEXT:    adc.w r2, r2, r3, asr #31
; CHECK-NEXT:    vmov.s8 r3, q0[10]
; CHECK-NEXT:    adds.w r12, r12, r3
; CHECK-NEXT:    adc.w r2, r2, r3, asr #31
; CHECK-NEXT:    vmov.s8 r3, q0[11]
; CHECK-NEXT:    adds.w r12, r12, r3
; CHECK-NEXT:    adc.w r2, r2, r3, asr #31
; CHECK-NEXT:    vmov.s8 r3, q0[12]
; CHECK-NEXT:    adds.w r12, r12, r3
; CHECK-NEXT:    adc.w r2, r2, r3, asr #31
; CHECK-NEXT:    vmov.s8 r3, q0[13]
; CHECK-NEXT:    adds.w r12, r12, r3
; CHECK-NEXT:    adc.w r2, r2, r3, asr #31
; CHECK-NEXT:    vmov.s8 r3, q0[14]
; CHECK-NEXT:    adds.w r12, r12, r3
; CHECK-NEXT:    adc.w lr, r2, r3, asr #31
; CHECK-NEXT:    vmov.s8 r3, q0[15]
; CHECK-NEXT:    adds.w r2, r12, r3
; CHECK-NEXT:    adc.w r3, lr, r3, asr #31
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    pop {r7, pc}
entry:
  %xx = sext <16 x i8> %x to <16 x i64>
  %z = call i64 @llvm.vector.reduce.add.v16i64(<16 x i64> %xx)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v8i8_v8i64_acc_zext(<8 x i8> %x, i64 %a) {
; CHECK-LABEL: add_v8i8_v8i64_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r7, lr}
; CHECK-NEXT:    push {r7, lr}
; CHECK-NEXT:    vmovlb.u8 q0, q0
; CHECK-NEXT:    vmov.i64 q1, #0xffff
; CHECK-NEXT:    vmov.u16 r2, q0[1]
; CHECK-NEXT:    vmov.u16 r3, q0[0]
; CHECK-NEXT:    vmov q2[2], q2[0], r3, r2
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov r2, s10
; CHECK-NEXT:    vmov r3, r12, d4
; CHECK-NEXT:    add.w lr, r3, r2
; CHECK-NEXT:    vmov.u16 r3, q0[3]
; CHECK-NEXT:    vmov.u16 r2, q0[2]
; CHECK-NEXT:    vmov q2[2], q2[0], r2, r3
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov r2, s8
; CHECK-NEXT:    vmov r3, s10
; CHECK-NEXT:    add r2, lr
; CHECK-NEXT:    add.w lr, r2, r3
; CHECK-NEXT:    vmov.u16 r3, q0[5]
; CHECK-NEXT:    vmov.u16 r2, q0[4]
; CHECK-NEXT:    vmov q2[2], q2[0], r2, r3
; CHECK-NEXT:    vand q2, q2, q1
; CHECK-NEXT:    vmov r2, s8
; CHECK-NEXT:    add lr, r2
; CHECK-NEXT:    vmov r3, r2, d5
; CHECK-NEXT:    adds.w lr, lr, r3
; CHECK-NEXT:    vmov.u16 r3, q0[6]
; CHECK-NEXT:    adc.w r12, r12, r2
; CHECK-NEXT:    vmov.u16 r2, q0[7]
; CHECK-NEXT:    vmov q0[2], q0[0], r3, r2
; CHECK-NEXT:    vand q0, q0, q1
; CHECK-NEXT:    vmov r2, r3, d0
; CHECK-NEXT:    adds.w lr, lr, r2
; CHECK-NEXT:    adc.w r12, r12, r3
; CHECK-NEXT:    vmov r2, r3, d1
; CHECK-NEXT:    adds.w r2, r2, lr
; CHECK-NEXT:    adc.w r3, r3, r12
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    pop {r7, pc}
entry:
  %xx = zext <8 x i8> %x to <8 x i64>
  %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %xx)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v8i8_v8i64_acc_sext(<8 x i8> %x, i64 %a) {
; CHECK-LABEL: add_v8i8_v8i64_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r7, lr}
; CHECK-NEXT:    push {r7, lr}
; CHECK-NEXT:    vmov.u16 r2, q0[0]
; CHECK-NEXT:    vmov.u16 r3, q0[1]
; CHECK-NEXT:    sxtb r2, r2
; CHECK-NEXT:    sxtb r3, r3
; CHECK-NEXT:    asr.w r12, r2, #31
; CHECK-NEXT:    adds.w lr, r2, r3
; CHECK-NEXT:    vmov.u16 r2, q0[2]
; CHECK-NEXT:    adc.w r3, r12, r3, asr #31
; CHECK-NEXT:    sxtb r2, r2
; CHECK-NEXT:    adds.w r12, lr, r2
; CHECK-NEXT:    adc.w r2, r3, r2, asr #31
; CHECK-NEXT:    vmov.u16 r3, q0[3]
; CHECK-NEXT:    sxtb r3, r3
; CHECK-NEXT:    adds.w r12, r12, r3
; CHECK-NEXT:    adc.w r2, r2, r3, asr #31
; CHECK-NEXT:    vmov.u16 r3, q0[4]
; CHECK-NEXT:    sxtb r3, r3
; CHECK-NEXT:    adds.w r12, r12, r3
; CHECK-NEXT:    adc.w r2, r2, r3, asr #31
; CHECK-NEXT:    vmov.u16 r3, q0[5]
; CHECK-NEXT:    sxtb r3, r3
; CHECK-NEXT:    adds.w r12, r12, r3
; CHECK-NEXT:    adc.w r2, r2, r3, asr #31
; CHECK-NEXT:    vmov.u16 r3, q0[6]
; CHECK-NEXT:    sxtb r3, r3
; CHECK-NEXT:    adds.w r12, r12, r3
; CHECK-NEXT:    adc.w lr, r2, r3, asr #31
; CHECK-NEXT:    vmov.u16 r3, q0[7]
; CHECK-NEXT:    sxtb r3, r3
; CHECK-NEXT:    adds.w r2, r12, r3
; CHECK-NEXT:    adc.w r3, lr, r3, asr #31
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    pop {r7, pc}
entry:
  %xx = sext <8 x i8> %x to <8 x i64>
  %z = call i64 @llvm.vector.reduce.add.v8i64(<8 x i64> %xx)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v4i8_v4i64_acc_zext(<4 x i8> %x, i64 %a) {
; CHECK-LABEL: add_v4i8_v4i64_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov.i32 q1, #0xff
; CHECK-NEXT:    vand q0, q0, q1
; CHECK-NEXT:    vaddlva.u32 r0, r1, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <4 x i8> %x to <4 x i64>
  %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %xx)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v4i8_v4i64_acc_sext(<4 x i8> %x, i64 %a) {
; CHECK-LABEL: add_v4i8_v4i64_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmovlb.s8 q0, q0
; CHECK-NEXT:    vmovlb.s16 q0, q0
; CHECK-NEXT:    vaddlva.s32 r0, r1, q0
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <4 x i8> %x to <4 x i64>
  %z = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %xx)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v2i8_v2i64_acc_zext(<2 x i8> %x, i64 %a) {
; CHECK-LABEL: add_v2i8_v2i64_acc_zext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov.i64 q1, #0xff
; CHECK-NEXT:    vand q0, q0, q1
; CHECK-NEXT:    vmov r2, s2
; CHECK-NEXT:    vmov r3, r12, d0
; CHECK-NEXT:    add r2, r3
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adc.w r1, r1, r12
; CHECK-NEXT:    bx lr
entry:
  %xx = zext <2 x i8> %x to <2 x i64>
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %xx)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v2i8_v2i64_acc_sext(<2 x i8> %x, i64 %a) {
; CHECK-LABEL: add_v2i8_v2i64_acc_sext:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    vmov r2, s0
; CHECK-NEXT:    vmov r3, s2
; CHECK-NEXT:    sxtb r2, r2
; CHECK-NEXT:    asr.w r12, r2, #31
; CHECK-NEXT:    sxtb r3, r3
; CHECK-NEXT:    adds r2, r2, r3
; CHECK-NEXT:    adc.w r3, r12, r3, asr #31
; CHECK-NEXT:    adds r0, r0, r2
; CHECK-NEXT:    adcs r1, r3
; CHECK-NEXT:    bx lr
entry:
  %xx = sext <2 x i8> %x to <2 x i64>
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %xx)
  %r = add i64 %z, %a
  ret i64 %r
}

define arm_aapcs_vfpcc i64 @add_v2i64_v2i64_acc(<2 x i64> %x, i64 %a) {
; CHECK-LABEL: add_v2i64_v2i64_acc:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r7, lr}
; CHECK-NEXT:    push {r7, lr}
; CHECK-NEXT:    vmov lr, r12, d1
; CHECK-NEXT:    vmov r3, r2, d0
; CHECK-NEXT:    adds.w r3, r3, lr
; CHECK-NEXT:    adc.w r2, r2, r12
; CHECK-NEXT:    adds r0, r0, r3
; CHECK-NEXT:    adcs r1, r2
; CHECK-NEXT:    pop {r7, pc}
entry:
  %z = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %x)
  %r = add i64 %z, %a
  ret i64 %r
}

declare i16 @llvm.vector.reduce.add.v16i16(<16 x i16>)
declare i16 @llvm.vector.reduce.add.v8i16(<8 x i16>)
declare i32 @llvm.vector.reduce.add.v16i32(<16 x i32>)
declare i32 @llvm.vector.reduce.add.v4i32(<4 x i32>)
declare i32 @llvm.vector.reduce.add.v8i32(<8 x i32>)
declare i64 @llvm.vector.reduce.add.v16i64(<16 x i64>)
declare i64 @llvm.vector.reduce.add.v2i64(<2 x i64>)
declare i64 @llvm.vector.reduce.add.v4i64(<4 x i64>)
declare i64 @llvm.vector.reduce.add.v8i64(<8 x i64>)
declare i8 @llvm.vector.reduce.add.v16i8(<16 x i8>)
llvm/llvm/test/CodeGen/Thumb2/mve-vecreduce-add.ll
