; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 4
; RUN: opt -S -passes=instcombine -mtriple=aarch64 < %s | FileCheck %s

; Transform the SVE idiom used to concatenate two vectors into target agnostic IR.

define <vscale x 8 x i1> @reinterpt_uzp1_1(<vscale x 4 x i1> %cmp0, <vscale x 4 x i1> %cmp1) {
; CHECK-LABEL: define <vscale x 8 x i1> @reinterpt_uzp1_1(
; CHECK-SAME: <vscale x 4 x i1> [[CMP0:%.*]], <vscale x 4 x i1> [[CMP1:%.*]]) {
; CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.vector.insert.nxv8i1.nxv4i1(<vscale x 8 x i1> poison, <vscale x 4 x i1> [[CMP0]], i64 0)
; CHECK-NEXT:    [[UZ1:%.*]] = call <vscale x 8 x i1> @llvm.vector.insert.nxv8i1.nxv4i1(<vscale x 8 x i1> [[TMP1]], <vscale x 4 x i1> [[CMP1]], i64 4)
; CHECK-NEXT:    ret <vscale x 8 x i1> [[UZ1]]
;
  %1 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %cmp0)
  %2 = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %1)
  %3 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %cmp1)
  %4 = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %3)
  %uz1 = tail call <vscale x 8 x i1> @llvm.aarch64.sve.uzp1.nxv8i1(<vscale x 8 x i1> %2, <vscale x 8 x i1> %4)
  ret <vscale x 8 x i1> %uz1
}

define <vscale x 8 x i1> @reinterpt_uzp1_2(<vscale x 2 x i1> %cmp0, <vscale x 2 x i1> %cmp1) {
; CHECK-LABEL: define <vscale x 8 x i1> @reinterpt_uzp1_2(
; CHECK-SAME: <vscale x 2 x i1> [[CMP0:%.*]], <vscale x 2 x i1> [[CMP1:%.*]]) {
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[CMP0]])
; CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[TMP1]])
; CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[CMP1]])
; CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> [[TMP3]])
; CHECK-NEXT:    [[TMP5:%.*]] = call <vscale x 8 x i1> @llvm.vector.insert.nxv8i1.nxv4i1(<vscale x 8 x i1> poison, <vscale x 4 x i1> [[TMP2]], i64 0)
; CHECK-NEXT:    [[UZ1:%.*]] = call <vscale x 8 x i1> @llvm.vector.insert.nxv8i1.nxv4i1(<vscale x 8 x i1> [[TMP5]], <vscale x 4 x i1> [[TMP4]], i64 4)
; CHECK-NEXT:    ret <vscale x 8 x i1> [[UZ1]]
;
  %1 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %cmp0)
  %2 = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %1)
  %3 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %cmp1)
  %4 = tail call <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1> %3)
  %5 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %2)
  %6 = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %5)
  %7 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %4)
  %8 = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %7)
  %uz1 = tail call <vscale x 8 x i1> @llvm.aarch64.sve.uzp1.nxv8i1(<vscale x 8 x i1> %6, <vscale x 8 x i1> %8)
  ret <vscale x 8 x i1> %uz1
}

define <vscale x 16 x i1> @reinterpt_uzp1_3(<vscale x 4 x i1> %cmp0, <vscale x 4 x i1> %cmp1, <vscale x 4 x i1> %cmp2, <vscale x 4 x i1> %cmp3) {
; CHECK-LABEL: define <vscale x 16 x i1> @reinterpt_uzp1_3(
; CHECK-SAME: <vscale x 4 x i1> [[CMP0:%.*]], <vscale x 4 x i1> [[CMP1:%.*]], <vscale x 4 x i1> [[CMP2:%.*]], <vscale x 4 x i1> [[CMP3:%.*]]) {
; CHECK-NEXT:    [[TMP1:%.*]] = call <vscale x 8 x i1> @llvm.vector.insert.nxv8i1.nxv4i1(<vscale x 8 x i1> poison, <vscale x 4 x i1> [[CMP0]], i64 0)
; CHECK-NEXT:    [[UZ1_1:%.*]] = call <vscale x 8 x i1> @llvm.vector.insert.nxv8i1.nxv4i1(<vscale x 8 x i1> [[TMP1]], <vscale x 4 x i1> [[CMP1]], i64 4)
; CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 8 x i1> @llvm.vector.insert.nxv8i1.nxv4i1(<vscale x 8 x i1> poison, <vscale x 4 x i1> [[CMP2]], i64 0)
; CHECK-NEXT:    [[UZ1_2:%.*]] = call <vscale x 8 x i1> @llvm.vector.insert.nxv8i1.nxv4i1(<vscale x 8 x i1> [[TMP2]], <vscale x 4 x i1> [[CMP3]], i64 4)
; CHECK-NEXT:    [[TMP3:%.*]] = call <vscale x 16 x i1> @llvm.vector.insert.nxv16i1.nxv8i1(<vscale x 16 x i1> poison, <vscale x 8 x i1> [[UZ1_1]], i64 0)
; CHECK-NEXT:    [[UZ3:%.*]] = call <vscale x 16 x i1> @llvm.vector.insert.nxv16i1.nxv8i1(<vscale x 16 x i1> [[TMP3]], <vscale x 8 x i1> [[UZ1_2]], i64 8)
; CHECK-NEXT:    ret <vscale x 16 x i1> [[UZ3]]
;
  %1 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %cmp0)
  %2 = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %1)
  %3 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %cmp1)
  %4 = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %3)
  %uz1_1 = tail call <vscale x 8 x i1> @llvm.aarch64.sve.uzp1.nxv8i1(<vscale x 8 x i1> %2, <vscale x 8 x i1> %4)
  %5 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %cmp2)
  %6 = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %5)
  %7 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %cmp3)
  %8 = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %7)
  %uz1_2 = tail call <vscale x 8 x i1> @llvm.aarch64.sve.uzp1.nxv8i1(<vscale x 8 x i1> %6, <vscale x 8 x i1> %8)
  %9 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %uz1_1)
  %10 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1> %uz1_2)
  %uz3 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.uzp1.nxv16i1(<vscale x 16 x i1> %9, <vscale x 16 x i1> %10)
  ret <vscale x 16 x i1> %uz3
}

define <vscale x 4 x i1> @neg1(<vscale x 4 x i1> %cmp0, <vscale x 4 x i1> %cmp1) {
; CHECK-LABEL: define <vscale x 4 x i1> @neg1(
; CHECK-SAME: <vscale x 4 x i1> [[CMP0:%.*]], <vscale x 4 x i1> [[CMP1:%.*]]) {
; CHECK-NEXT:    [[UZ1:%.*]] = tail call <vscale x 4 x i1> @llvm.aarch64.sve.uzp1.nxv4i1(<vscale x 4 x i1> [[CMP0]], <vscale x 4 x i1> [[CMP1]])
; CHECK-NEXT:    ret <vscale x 4 x i1> [[UZ1]]
;
  %uz1 = tail call <vscale x 4 x i1> @llvm.aarch64.sve.uzp1.nxv4i1(<vscale x 4 x i1> %cmp0, <vscale x 4 x i1> %cmp1)
  ret <vscale x 4 x i1> %uz1
}

define <vscale x 8 x i1> @neg2(<vscale x 2 x i1> %cmp0, <vscale x 4 x i1> %cmp1) {
; CHECK-LABEL: define <vscale x 8 x i1> @neg2(
; CHECK-SAME: <vscale x 2 x i1> [[CMP0:%.*]], <vscale x 4 x i1> [[CMP1:%.*]]) {
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[CMP0]])
; CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[TMP1]])
; CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[CMP1]])
; CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[TMP3]])
; CHECK-NEXT:    [[UZ1:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.uzp1.nxv8i1(<vscale x 8 x i1> [[TMP2]], <vscale x 8 x i1> [[TMP4]])
; CHECK-NEXT:    ret <vscale x 8 x i1> [[UZ1]]
;
  %1 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %cmp0)
  %2 = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %1)
  %3 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %cmp1)
  %4 = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %3)
  %uz1 = tail call <vscale x 8 x i1> @llvm.aarch64.sve.uzp1.nxv8i1(<vscale x 8 x i1> %2, <vscale x 8 x i1> %4)
  ret <vscale x 8 x i1> %uz1
}

define <vscale x 8 x i1> @neg3(<vscale x 8 x i1> %cmp0, <vscale x 4 x i1> %cmp1) {
; CHECK-LABEL: define <vscale x 8 x i1> @neg3(
; CHECK-SAME: <vscale x 8 x i1> [[CMP0:%.*]], <vscale x 4 x i1> [[CMP1:%.*]]) {
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> [[CMP1]])
; CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[TMP1]])
; CHECK-NEXT:    [[UZ1:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.uzp1.nxv8i1(<vscale x 8 x i1> [[CMP0]], <vscale x 8 x i1> [[TMP2]])
; CHECK-NEXT:    ret <vscale x 8 x i1> [[UZ1]]
;
  %1 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1> %cmp1)
  %2 = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %1)
  %uz1 = tail call <vscale x 8 x i1> @llvm.aarch64.sve.uzp1.nxv8i1(<vscale x 8 x i1> %cmp0, <vscale x 8 x i1> %2)
  ret <vscale x 8 x i1> %uz1
}

define <vscale x 8 x i1> @neg4(<vscale x 2 x i1> %cmp0, <vscale x 2 x i1> %cmp1) {
; CHECK-LABEL: define <vscale x 8 x i1> @neg4(
; CHECK-SAME: <vscale x 2 x i1> [[CMP0:%.*]], <vscale x 2 x i1> [[CMP1:%.*]]) {
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[CMP0]])
; CHECK-NEXT:    [[TMP2:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[TMP1]])
; CHECK-NEXT:    [[TMP3:%.*]] = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> [[CMP1]])
; CHECK-NEXT:    [[TMP4:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> [[TMP3]])
; CHECK-NEXT:    [[UZ1:%.*]] = tail call <vscale x 8 x i1> @llvm.aarch64.sve.uzp1.nxv8i1(<vscale x 8 x i1> [[TMP2]], <vscale x 8 x i1> [[TMP4]])
; CHECK-NEXT:    ret <vscale x 8 x i1> [[UZ1]]
;
  %1 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %cmp0)
  %2 = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %1)
  %3 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1> %cmp1)
  %4 = tail call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %3)
  %uz1 = tail call <vscale x 8 x i1> @llvm.aarch64.sve.uzp1.nxv8i1(<vscale x 8 x i1> %2, <vscale x 8 x i1> %4)
  ret <vscale x 8 x i1> %uz1
}

declare <vscale x 4 x i1> @llvm.aarch64.sve.uzp1.nxv4i1(<vscale x 4 x i1>, <vscale x 4 x i1>)
declare <vscale x 8 x i1> @llvm.aarch64.sve.uzp1.nxv8i1(<vscale x 8 x i1>, <vscale x 8 x i1>)
declare <vscale x 16 x i1> @llvm.aarch64.sve.uzp1.nxv16i1(<vscale x 16 x i1>, <vscale x 16 x i1>)

declare <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv8i1(<vscale x 8 x i1>)
declare <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv4i1(<vscale x 4 x i1>)
declare <vscale x 16 x i1> @llvm.aarch64.sve.convert.to.svbool.nxv2i1(<vscale x 2 x i1>)
declare <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1>)
declare <vscale x 4 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv4i1(<vscale x 16 x i1>)