; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 4
; RUN: llc -enable-subreg-liveness -mtriple=aarch64-linux-gnu -mattr=+sve < %s | FileCheck %s --check-prefixes=CHECK,NONSTREAMING
; RUN: llc -enable-subreg-liveness -mtriple=aarch64-linux-gnu -mattr=+sme -force-streaming < %s | FileCheck %s --check-prefixes=CHECK,STREAMING

; WITH VSCALE RANGE

define i32 @ctz_nxv32i1(<vscale x 32 x i1> %a) #0 {
; NONSTREAMING-LABEL: ctz_nxv32i1:
; NONSTREAMING:       // %bb.0:
; NONSTREAMING-NEXT:    ptrue p2.b
; NONSTREAMING-NEXT:    rdvl x8, #1
; NONSTREAMING-NEXT:    mov w10, w8
; NONSTREAMING-NEXT:    brkb p0.b, p2/z, p0.b
; NONSTREAMING-NEXT:    brkb p1.b, p2/z, p1.b
; NONSTREAMING-NEXT:    cntp x9, p0, p0.b
; NONSTREAMING-NEXT:    incp x8, p1.b
; NONSTREAMING-NEXT:    cmp w9, w10
; NONSTREAMING-NEXT:    csel w0, w9, w8, ne
; NONSTREAMING-NEXT:    ret
;
; STREAMING-LABEL: ctz_nxv32i1:
; STREAMING:       // %bb.0:
; STREAMING-NEXT:    ptrue p2.b
; STREAMING-NEXT:    rdvl x10, #1
; STREAMING-NEXT:    brkb p1.b, p2/z, p1.b
; STREAMING-NEXT:    brkb p0.b, p2/z, p0.b
; STREAMING-NEXT:    cntp x8, p1, p1.b
; STREAMING-NEXT:    cntp x9, p0, p0.b
; STREAMING-NEXT:    incb x8
; STREAMING-NEXT:    cmp w9, w10
; STREAMING-NEXT:    csel w0, w9, w8, ne
; STREAMING-NEXT:    ret
  %res = call i32 @llvm.experimental.cttz.elts.i32.nxv32i1(<vscale x 32 x i1> %a, i1 0)
  ret i32 %res
}

define i32 @ctz_nxv4i32(<vscale x 4 x i32> %a) #0 {
; CHECK-LABEL: ctz_nxv4i32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ptrue p0.s
; CHECK-NEXT:    cmpne p1.s, p0/z, z0.s, #0
; CHECK-NEXT:    brkb p1.b, p0/z, p1.b
; CHECK-NEXT:    cntp x0, p1, p1.s
; CHECK-NEXT:    ret
  %res = call i32 @llvm.experimental.cttz.elts.i32.nxv4i32(<vscale x 4 x i32> %a, i1 0)
  ret i32 %res
}

; VSCALE RANGE, ZERO IS POISON

define i64 @vscale_4096(<vscale x 16 x i8> %a) #1 {
; CHECK-LABEL: vscale_4096:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ptrue p0.b
; CHECK-NEXT:    cmpne p1.b, p0/z, z0.b, #0
; CHECK-NEXT:    brkb p1.b, p0/z, p1.b
; CHECK-NEXT:    cntp x0, p1, p1.b
; CHECK-NEXT:    ret
  %res = call i64 @llvm.experimental.cttz.elts.i64.nxv16i8(<vscale x 16 x i8> %a, i1 0)
  ret i64 %res
}

define i64 @vscale_4096_poison(<vscale x 16 x i8> %a) #1 {
; CHECK-LABEL: vscale_4096_poison:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ptrue p0.b
; CHECK-NEXT:    cmpne p1.b, p0/z, z0.b, #0
; CHECK-NEXT:    brkb p1.b, p0/z, p1.b
; CHECK-NEXT:    cntp x0, p1, p1.b
; CHECK-NEXT:    ret
  %res = call i64 @llvm.experimental.cttz.elts.i64.nxv16i8(<vscale x 16 x i8> %a, i1 1)
  ret i64 %res
}

; EFFICIENT LOWERING USING BRKB

define i32 @ctz_nxv2i1(<vscale x 2 x i1> %a) {
; CHECK-LABEL: ctz_nxv2i1:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ptrue p1.d
; CHECK-NEXT:    brkb p0.b, p1/z, p0.b
; CHECK-NEXT:    cntp x0, p0, p0.d
; CHECK-NEXT:    ret
  %res = call i32 @llvm.experimental.cttz.elts.i32.nxv2i1(<vscale x 2 x i1> %a, i1 0)
  ret i32 %res
}

define i32 @ctz_nxv2i1_poison(<vscale x 2 x i1> %a) {
; CHECK-LABEL: ctz_nxv2i1_poison:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ptrue p1.d
; CHECK-NEXT:    brkb p0.b, p1/z, p0.b
; CHECK-NEXT:    cntp x0, p0, p0.d
; CHECK-NEXT:    ret
  %res = call i32 @llvm.experimental.cttz.elts.i32.nxv2i1(<vscale x 2 x i1> %a, i1 1)
  ret i32 %res
}

define i64 @add_i64_ctz_nxv2i1_poison(<vscale x 2 x i1> %a, i64 %b) {
; CHECK-LABEL: add_i64_ctz_nxv2i1_poison:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ptrue p1.d
; CHECK-NEXT:    brkb p0.b, p1/z, p0.b
; CHECK-NEXT:    incp x0, p0.d
; CHECK-NEXT:    ret
  %res = call i64 @llvm.experimental.cttz.elts.i64.nxv2i1(<vscale x 2 x i1> %a, i1 1)
  %add = add i64 %res, %b
  ret i64 %add
}

define i32 @add_i32_ctz_nxv2i1_poison(<vscale x 2 x i1> %a, i32 %b) {
; CHECK-LABEL: add_i32_ctz_nxv2i1_poison:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ptrue p1.d
; CHECK-NEXT:    brkb p0.b, p1/z, p0.b
; CHECK-NEXT:    incp x0, p0.d
; CHECK-NEXT:    ret
  %res = call i64 @llvm.experimental.cttz.elts.i64.nxv2i1(<vscale x 2 x i1> %a, i1 1)
  %trunc = trunc i64 %res to i32
  %add = add i32 %trunc, %b
  ret i32 %add
}

define i32 @ctz_nxv4i1(<vscale x 4 x i1> %a) {
; CHECK-LABEL: ctz_nxv4i1:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ptrue p1.s
; CHECK-NEXT:    brkb p0.b, p1/z, p0.b
; CHECK-NEXT:    cntp x0, p0, p0.s
; CHECK-NEXT:    ret
  %res = call i32 @llvm.experimental.cttz.elts.i32.nxv4i1(<vscale x 4 x i1> %a, i1 0)
  ret i32 %res
}

define i32 @ctz_nxv4i1_poison(<vscale x 4 x i1> %a) {
; CHECK-LABEL: ctz_nxv4i1_poison:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ptrue p1.s
; CHECK-NEXT:    brkb p0.b, p1/z, p0.b
; CHECK-NEXT:    cntp x0, p0, p0.s
; CHECK-NEXT:    ret
  %res = call i32 @llvm.experimental.cttz.elts.i32.nxv4i1(<vscale x 4 x i1> %a, i1 1)
  ret i32 %res
}

define i64 @add_i64_ctz_nxv4i1_poison(<vscale x 4 x i1> %a, i64 %b) {
; CHECK-LABEL: add_i64_ctz_nxv4i1_poison:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ptrue p1.s
; CHECK-NEXT:    brkb p0.b, p1/z, p0.b
; CHECK-NEXT:    incp x0, p0.s
; CHECK-NEXT:    ret
  %res = call i64 @llvm.experimental.cttz.elts.i64.nxv4i1(<vscale x 4 x i1> %a, i1 1)
  %add = add i64 %res, %b
  ret i64 %add
}

define i32 @add_i32_ctz_nxv4i1_poison(<vscale x 4 x i1> %a, i32 %b) {
; CHECK-LABEL: add_i32_ctz_nxv4i1_poison:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ptrue p1.s
; CHECK-NEXT:    brkb p0.b, p1/z, p0.b
; CHECK-NEXT:    incp x0, p0.s
; CHECK-NEXT:    ret
  %res = call i64 @llvm.experimental.cttz.elts.i64.nxv4i1(<vscale x 4 x i1> %a, i1 1)
  %trunc = trunc i64 %res to i32
  %add = add i32 %trunc, %b
  ret i32 %add
}

define i32 @ctz_nxv8i1(<vscale x 8 x i1> %a) {
; CHECK-LABEL: ctz_nxv8i1:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ptrue p1.h
; CHECK-NEXT:    brkb p0.b, p1/z, p0.b
; CHECK-NEXT:    cntp x0, p0, p0.h
; CHECK-NEXT:    ret
  %res = call i32 @llvm.experimental.cttz.elts.i32.nxv8i1(<vscale x 8 x i1> %a, i1 0)
  ret i32 %res
}

define i32 @ctz_nxv8i1_poison(<vscale x 8 x i1> %a) {
; CHECK-LABEL: ctz_nxv8i1_poison:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ptrue p1.h
; CHECK-NEXT:    brkb p0.b, p1/z, p0.b
; CHECK-NEXT:    cntp x0, p0, p0.h
; CHECK-NEXT:    ret
  %res = call i32 @llvm.experimental.cttz.elts.i32.nxv8i1(<vscale x 8 x i1> %a, i1 1)
  ret i32 %res
}

define i64 @add_i64_ctz_nxv8i1_poison(<vscale x 8 x i1> %a, i64 %b) {
; CHECK-LABEL: add_i64_ctz_nxv8i1_poison:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ptrue p1.h
; CHECK-NEXT:    brkb p0.b, p1/z, p0.b
; CHECK-NEXT:    incp x0, p0.h
; CHECK-NEXT:    ret
  %res = call i64 @llvm.experimental.cttz.elts.i64.nxv8i1(<vscale x 8 x i1> %a, i1 1)
  %add = add i64 %res, %b
  ret i64 %add
}

define i32 @add_i32_ctz_nxv8i1_poison(<vscale x 8 x i1> %a, i32 %b) {
; CHECK-LABEL: add_i32_ctz_nxv8i1_poison:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ptrue p1.h
; CHECK-NEXT:    brkb p0.b, p1/z, p0.b
; CHECK-NEXT:    incp x0, p0.h
; CHECK-NEXT:    ret
  %res = call i64 @llvm.experimental.cttz.elts.i64.nxv8i1(<vscale x 8 x i1> %a, i1 1)
  %trunc = trunc i64 %res to i32
  %add = add i32 %trunc, %b
  ret i32 %add
}

define i32 @ctz_nxv16i1(<vscale x 16 x i1> %a) {
; CHECK-LABEL: ctz_nxv16i1:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ptrue p1.b
; CHECK-NEXT:    brkb p0.b, p1/z, p0.b
; CHECK-NEXT:    cntp x0, p0, p0.b
; CHECK-NEXT:    ret
  %res = call i32 @llvm.experimental.cttz.elts.i32.nxv16i1(<vscale x 16 x i1> %a, i1 0)
  ret i32 %res
}

define i32 @ctz_nxv16i1_poison(<vscale x 16 x i1> %a) {
; CHECK-LABEL: ctz_nxv16i1_poison:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ptrue p1.b
; CHECK-NEXT:    brkb p0.b, p1/z, p0.b
; CHECK-NEXT:    cntp x0, p0, p0.b
; CHECK-NEXT:    ret
  %res = call i32 @llvm.experimental.cttz.elts.i32.nxv16i1(<vscale x 16 x i1> %a, i1 1)
  ret i32 %res
}

define i32 @ctz_and_nxv16i1(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b) {
; CHECK-LABEL: ctz_and_nxv16i1:
; CHECK:       // %bb.0:
; CHECK-NEXT:    cmpne p1.b, p0/z, z0.b, z1.b
; CHECK-NEXT:    ptrue p0.b
; CHECK-NEXT:    brkb p1.b, p0/z, p1.b
; CHECK-NEXT:    cntp x0, p1, p1.b
; CHECK-NEXT:    ret
  %cmp = icmp ne <vscale x 16 x i8> %a, %b
  %select = select <vscale x 16 x i1> %pg, <vscale x 16 x i1> %cmp, <vscale x 16 x i1> zeroinitializer
  %and = and <vscale x 16 x i1> %pg, %select
  %res = call i32 @llvm.experimental.cttz.elts.i32.nxv16i1(<vscale x 16 x i1> %and, i1 0)
  ret i32 %res
}

define i64 @add_i64_ctz_nxv16i1_poison(<vscale x 16 x i1> %a, i64 %b) {
; CHECK-LABEL: add_i64_ctz_nxv16i1_poison:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ptrue p1.b
; CHECK-NEXT:    brkb p0.b, p1/z, p0.b
; CHECK-NEXT:    incp x0, p0.b
; CHECK-NEXT:    ret
  %res = call i64 @llvm.experimental.cttz.elts.i64.nxv16i1(<vscale x 16 x i1> %a, i1 1)
  %add = add i64 %res, %b
  ret i64 %add
}

define i32 @add_i32_ctz_nxv16i1_poison(<vscale x 16 x i1> %a, i32 %b) {
; CHECK-LABEL: add_i32_ctz_nxv16i1_poison:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ptrue p1.b
; CHECK-NEXT:    brkb p0.b, p1/z, p0.b
; CHECK-NEXT:    incp x0, p0.b
; CHECK-NEXT:    ret
  %res = call i64 @llvm.experimental.cttz.elts.i64.nxv16i1(<vscale x 16 x i1> %a, i1 1)
  %trunc = trunc i64 %res to i32
  %add = add i32 %trunc, %b
  ret i32 %add
}

; FIXED-WIDTH VECTOR TYPES

define i32 @ctz_v16i1(<16 x i1> %a) {
; NONSTREAMING-LABEL: ctz_v16i1:
; NONSTREAMING:       // %bb.0:
; NONSTREAMING-NEXT:    shl v0.16b, v0.16b, #7
; NONSTREAMING-NEXT:    ptrue p0.b, vl16
; NONSTREAMING-NEXT:    cmpne p1.b, p0/z, z0.b, #0
; NONSTREAMING-NEXT:    brkb p1.b, p0/z, p1.b
; NONSTREAMING-NEXT:    cntp x0, p1, p1.b
; NONSTREAMING-NEXT:    ret
;
; STREAMING-LABEL: ctz_v16i1:
; STREAMING:       // %bb.0:
; STREAMING-NEXT:    lsl z0.b, z0.b, #7
; STREAMING-NEXT:    ptrue p0.b, vl16
; STREAMING-NEXT:    asr z0.b, z0.b, #7
; STREAMING-NEXT:    cmpne p1.b, p0/z, z0.b, #0
; STREAMING-NEXT:    brkb p1.b, p0/z, p1.b
; STREAMING-NEXT:    cntp x0, p1, p1.b
; STREAMING-NEXT:    ret
  %res = call i32 @llvm.experimental.cttz.elts.i32.v16i1(<16 x i1> %a, i1 0)
  ret i32 %res
}

define i32 @ctz_v16i1_poison(<16 x i1> %a) {
; NONSTREAMING-LABEL: ctz_v16i1_poison:
; NONSTREAMING:       // %bb.0:
; NONSTREAMING-NEXT:    shl v0.16b, v0.16b, #7
; NONSTREAMING-NEXT:    ptrue p0.b, vl16
; NONSTREAMING-NEXT:    cmpne p1.b, p0/z, z0.b, #0
; NONSTREAMING-NEXT:    brkb p1.b, p0/z, p1.b
; NONSTREAMING-NEXT:    cntp x0, p1, p1.b
; NONSTREAMING-NEXT:    ret
;
; STREAMING-LABEL: ctz_v16i1_poison:
; STREAMING:       // %bb.0:
; STREAMING-NEXT:    lsl z0.b, z0.b, #7
; STREAMING-NEXT:    ptrue p0.b, vl16
; STREAMING-NEXT:    asr z0.b, z0.b, #7
; STREAMING-NEXT:    cmpne p1.b, p0/z, z0.b, #0
; STREAMING-NEXT:    brkb p1.b, p0/z, p1.b
; STREAMING-NEXT:    cntp x0, p1, p1.b
; STREAMING-NEXT:    ret
  %res = call i32 @llvm.experimental.cttz.elts.i32.v16i1(<16 x i1> %a, i1 1)
  ret i32 %res
}

define i64 @add_i64_ctz_v16i1_poison(<16 x i1> %a, i64 %b) {
; NONSTREAMING-LABEL: add_i64_ctz_v16i1_poison:
; NONSTREAMING:       // %bb.0:
; NONSTREAMING-NEXT:    shl v0.16b, v0.16b, #7
; NONSTREAMING-NEXT:    ptrue p0.b, vl16
; NONSTREAMING-NEXT:    cmpne p1.b, p0/z, z0.b, #0
; NONSTREAMING-NEXT:    brkb p1.b, p0/z, p1.b
; NONSTREAMING-NEXT:    incp x0, p1.b
; NONSTREAMING-NEXT:    ret
;
; STREAMING-LABEL: add_i64_ctz_v16i1_poison:
; STREAMING:       // %bb.0:
; STREAMING-NEXT:    lsl z0.b, z0.b, #7
; STREAMING-NEXT:    ptrue p0.b, vl16
; STREAMING-NEXT:    asr z0.b, z0.b, #7
; STREAMING-NEXT:    cmpne p1.b, p0/z, z0.b, #0
; STREAMING-NEXT:    brkb p1.b, p0/z, p1.b
; STREAMING-NEXT:    incp x0, p1.b
; STREAMING-NEXT:    ret
  %res = call i64 @llvm.experimental.cttz.elts.i64.v16i1(<16 x i1> %a, i1 1)
  %add = add i64 %res, %b
  ret i64 %add
}

define i32 @ctz_v8i1(<8 x i1> %a) {
; NONSTREAMING-LABEL: ctz_v8i1:
; NONSTREAMING:       // %bb.0:
; NONSTREAMING-NEXT:    shl v0.8b, v0.8b, #7
; NONSTREAMING-NEXT:    ptrue p0.b, vl8
; NONSTREAMING-NEXT:    cmpne p1.b, p0/z, z0.b, #0
; NONSTREAMING-NEXT:    brkb p1.b, p0/z, p1.b
; NONSTREAMING-NEXT:    cntp x0, p1, p1.b
; NONSTREAMING-NEXT:    ret
;
; STREAMING-LABEL: ctz_v8i1:
; STREAMING:       // %bb.0:
; STREAMING-NEXT:    lsl z0.b, z0.b, #7
; STREAMING-NEXT:    ptrue p0.b, vl8
; STREAMING-NEXT:    asr z0.b, z0.b, #7
; STREAMING-NEXT:    cmpne p1.b, p0/z, z0.b, #0
; STREAMING-NEXT:    brkb p1.b, p0/z, p1.b
; STREAMING-NEXT:    cntp x0, p1, p1.b
; STREAMING-NEXT:    ret
  %res = call i32 @llvm.experimental.cttz.elts.i32.v8i1(<8 x i1> %a, i1 0)
  ret i32 %res
}

define i32 @ctz_v8i1_poison(<8 x i1> %a) {
; NONSTREAMING-LABEL: ctz_v8i1_poison:
; NONSTREAMING:       // %bb.0:
; NONSTREAMING-NEXT:    shl v0.8b, v0.8b, #7
; NONSTREAMING-NEXT:    ptrue p0.b, vl8
; NONSTREAMING-NEXT:    cmpne p1.b, p0/z, z0.b, #0
; NONSTREAMING-NEXT:    brkb p1.b, p0/z, p1.b
; NONSTREAMING-NEXT:    cntp x0, p1, p1.b
; NONSTREAMING-NEXT:    ret
;
; STREAMING-LABEL: ctz_v8i1_poison:
; STREAMING:       // %bb.0:
; STREAMING-NEXT:    lsl z0.b, z0.b, #7
; STREAMING-NEXT:    ptrue p0.b, vl8
; STREAMING-NEXT:    asr z0.b, z0.b, #7
; STREAMING-NEXT:    cmpne p1.b, p0/z, z0.b, #0
; STREAMING-NEXT:    brkb p1.b, p0/z, p1.b
; STREAMING-NEXT:    cntp x0, p1, p1.b
; STREAMING-NEXT:    ret
  %res = call i32 @llvm.experimental.cttz.elts.i32.v8i1(<8 x i1> %a, i1 1)
  ret i32 %res
}

define i32 @ctz_v4i1(<4 x i1> %a) {
; NONSTREAMING-LABEL: ctz_v4i1:
; NONSTREAMING:       // %bb.0:
; NONSTREAMING-NEXT:    shl v0.4h, v0.4h, #15
; NONSTREAMING-NEXT:    ptrue p0.h, vl4
; NONSTREAMING-NEXT:    cmpne p1.h, p0/z, z0.h, #0
; NONSTREAMING-NEXT:    brkb p1.b, p0/z, p1.b
; NONSTREAMING-NEXT:    cntp x0, p1, p1.h
; NONSTREAMING-NEXT:    ret
;
; STREAMING-LABEL: ctz_v4i1:
; STREAMING:       // %bb.0:
; STREAMING-NEXT:    lsl z0.h, z0.h, #15
; STREAMING-NEXT:    ptrue p0.h, vl4
; STREAMING-NEXT:    asr z0.h, z0.h, #15
; STREAMING-NEXT:    cmpne p1.h, p0/z, z0.h, #0
; STREAMING-NEXT:    brkb p1.b, p0/z, p1.b
; STREAMING-NEXT:    cntp x0, p1, p1.h
; STREAMING-NEXT:    ret
  %res = call i32 @llvm.experimental.cttz.elts.i32.v4i1(<4 x i1> %a, i1 0)
  ret i32 %res
}

define i32 @ctz_v4i1_poison(<4 x i1> %a) {
; NONSTREAMING-LABEL: ctz_v4i1_poison:
; NONSTREAMING:       // %bb.0:
; NONSTREAMING-NEXT:    shl v0.4h, v0.4h, #15
; NONSTREAMING-NEXT:    ptrue p0.h, vl4
; NONSTREAMING-NEXT:    cmpne p1.h, p0/z, z0.h, #0
; NONSTREAMING-NEXT:    brkb p1.b, p0/z, p1.b
; NONSTREAMING-NEXT:    cntp x0, p1, p1.h
; NONSTREAMING-NEXT:    ret
;
; STREAMING-LABEL: ctz_v4i1_poison:
; STREAMING:       // %bb.0:
; STREAMING-NEXT:    lsl z0.h, z0.h, #15
; STREAMING-NEXT:    ptrue p0.h, vl4
; STREAMING-NEXT:    asr z0.h, z0.h, #15
; STREAMING-NEXT:    cmpne p1.h, p0/z, z0.h, #0
; STREAMING-NEXT:    brkb p1.b, p0/z, p1.b
; STREAMING-NEXT:    cntp x0, p1, p1.h
; STREAMING-NEXT:    ret
  %res = call i32 @llvm.experimental.cttz.elts.i32.v4i1(<4 x i1> %a, i1 1)
  ret i32 %res
}

define i32 @ctz_v2i1(<2 x i1> %a) {
; NONSTREAMING-LABEL: ctz_v2i1:
; NONSTREAMING:       // %bb.0:
; NONSTREAMING-NEXT:    shl v0.2s, v0.2s, #31
; NONSTREAMING-NEXT:    ptrue p0.s, vl2
; NONSTREAMING-NEXT:    cmpne p1.s, p0/z, z0.s, #0
; NONSTREAMING-NEXT:    brkb p1.b, p0/z, p1.b
; NONSTREAMING-NEXT:    cntp x0, p1, p1.s
; NONSTREAMING-NEXT:    ret
;
; STREAMING-LABEL: ctz_v2i1:
; STREAMING:       // %bb.0:
; STREAMING-NEXT:    lsl z0.s, z0.s, #31
; STREAMING-NEXT:    ptrue p0.s, vl2
; STREAMING-NEXT:    asr z0.s, z0.s, #31
; STREAMING-NEXT:    cmpne p1.s, p0/z, z0.s, #0
; STREAMING-NEXT:    brkb p1.b, p0/z, p1.b
; STREAMING-NEXT:    cntp x0, p1, p1.s
; STREAMING-NEXT:    ret
  %res = call i32 @llvm.experimental.cttz.elts.i32.v2i1(<2 x i1> %a, i1 0)
  ret i32 %res
}

define i32 @ctz_v2i1_poison(<2 x i1> %a) {
; NONSTREAMING-LABEL: ctz_v2i1_poison:
; NONSTREAMING:       // %bb.0:
; NONSTREAMING-NEXT:    shl v0.2s, v0.2s, #31
; NONSTREAMING-NEXT:    ptrue p0.s, vl2
; NONSTREAMING-NEXT:    cmpne p1.s, p0/z, z0.s, #0
; NONSTREAMING-NEXT:    brkb p1.b, p0/z, p1.b
; NONSTREAMING-NEXT:    cntp x0, p1, p1.s
; NONSTREAMING-NEXT:    ret
;
; STREAMING-LABEL: ctz_v2i1_poison:
; STREAMING:       // %bb.0:
; STREAMING-NEXT:    lsl z0.s, z0.s, #31
; STREAMING-NEXT:    ptrue p0.s, vl2
; STREAMING-NEXT:    asr z0.s, z0.s, #31
; STREAMING-NEXT:    cmpne p1.s, p0/z, z0.s, #0
; STREAMING-NEXT:    brkb p1.b, p0/z, p1.b
; STREAMING-NEXT:    cntp x0, p1, p1.s
; STREAMING-NEXT:    ret
  %res = call i32 @llvm.experimental.cttz.elts.i32.v2i1(<2 x i1> %a, i1 1)
  ret i32 %res
}

declare i32 @llvm.experimental.cttz.elts.i32.nxv8i1(<vscale x 8 x i1>, i1)
declare i64 @llvm.experimental.cttz.elts.i64.nxv8i1(<vscale x 8 x i1>, i1)
declare i64 @llvm.experimental.cttz.elts.i64.nxv16i1(<vscale x 16 x i1>, i1)
declare i32 @llvm.experimental.cttz.elts.i32.nxv16i1(<vscale x 16 x i1>, i1)
declare i32 @llvm.experimental.cttz.elts.i32.nxv32i1(<vscale x 32 x i1>, i1)
declare i32 @llvm.experimental.cttz.elts.i32.nxv4i32(<vscale x 4 x i32>, i1)

declare i64 @llvm.experimental.cttz.elts.i64.nxv16i8(<vscale x 16 x i8>, i1)

attributes #0 = { vscale_range(1,16) }
attributes #1 = { vscale_range(1,4096) }