; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 6
; RUN: llc -O3 < %s -mtriple=aarch64 | FileCheck %s

; The seemingly redundant wreg mov where src_reg == dst_reg shouldn't be
; removed, because it has the effect of zeroing the upper bits in the matching
; xreg.

define i32 @widget(i32 %arg, i32 %arg1, i1 %arg2, ptr %arg3, i1 %arg4) #0 nounwind {
; CHECK-LABEL: widget:
; CHECK:       // %bb.0: // %bb
; CHECK-NEXT:    tbz w2, #0, .LBB0_2
; CHECK-NEXT:  // %bb.1:
; CHECK-NEXT:    mov w0, #1 // =0x1
; CHECK-NEXT:    ret
; CHECK-NEXT:  .LBB0_2: // %bb5
; CHECK-NEXT:    tbz w4, #0, .LBB0_4
; CHECK-NEXT:  // %bb.3:
; CHECK-NEXT:    mov w0, #0 // =0x0
; CHECK-NEXT:    ret
; CHECK-NEXT:  .LBB0_4: // %bb6
; CHECK-NEXT:    str x30, [sp, #-48]! // 8-byte Folded Spill
; CHECK-NEXT:    stp x22, x21, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT:    stp x20, x19, [sp, #32] // 16-byte Folded Spill
; CHECK-NEXT:    mov x19, x3
; CHECK-NEXT:    mov x20, x0
; CHECK-NEXT:    mov x21, x1
; CHECK-NEXT:    bl baz
; CHECK-NEXT:    mov w0, #0 // =0x0
; CHECK-NEXT:    cbnz wzr, .LBB0_11
; CHECK-NEXT:  // %bb.5: // %bb6
; CHECK-NEXT:    mov w10, #1 // =0x1
; CHECK-NEXT:    cbnz w10, .LBB0_11
; CHECK-NEXT:  // %bb.6: // %bb7
; CHECK-NEXT:    cbnz w10, .LBB0_10
; CHECK-NEXT:  // %bb.7: // %bb8
; CHECK-NEXT:    mov x8, x21
; CHECK-NEXT:    mov x9, x20
; CHECK-NEXT:    mov w20, #0 // =0x0
; CHECK-NEXT:    mov w9, w9
; CHECK-NEXT:    mov x21, x9
; CHECK-NEXT:    mov w8, w8
; CHECK-NEXT:    mov x22, x8
; CHECK-NEXT:  .LBB0_8: // %bb10
; CHECK-NEXT:    // =>This Inner Loop Header: Depth=1
; CHECK-NEXT:    strb w20, [x19]
; CHECK-NEXT:    cbnz x21, .LBB0_8
; CHECK-NEXT:  // %bb.9: // %bb12
; CHECK-NEXT:    // in Loop: Header=BB0_8 Depth=1
; CHECK-NEXT:    bl snork
; CHECK-NEXT:    cbnz x22, .LBB0_8
; CHECK-NEXT:  .LBB0_10:
; CHECK-NEXT:    mov w0, #0 // =0x0
; CHECK-NEXT:  .LBB0_11:
; CHECK-NEXT:    ldp x20, x19, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT:    ldp x22, x21, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT:    ldr x30, [sp], #48 // 8-byte Folded Reload
; CHECK-NEXT:    ret
bb:
  br i1 %arg2, label %bb14, label %bb5

bb5:
  %load = load ptr, ptr null, align 8
  br i1 %arg4, label %bb14, label %bb6

bb6:
  %call = call i32 @baz() #1
  %or = or i1 false, true
  br i1 %or, label %bb14, label %bb7

bb7:
  %icmp = icmp eq i32 0, 0
  %zext = zext i32 %arg to i64
  br i1 %icmp, label %bb14, label %bb8

bb8:
  %zext9 = zext i32 %arg1 to i64
  br label %bb10

bb10:
  store i8 0, ptr %arg3, align 1
  %icmp11 = icmp eq i64 %zext, 0
  br i1 %icmp11, label %bb12, label %bb10

bb12:
  call void @snork()
  %icmp13 = icmp eq i64 0, %zext9
  br i1 %icmp13, label %bb14, label %bb10

bb14:
  %phi = phi i32 [ 0, %bb6 ], [ 0, %bb7 ], [ 0, %bb12 ], [ 1, %bb ], [ 0, %bb5 ]
  ret i32 %phi
}

declare i32 @baz()

declare void @snork()

attributes #0 = { "target-cpu"="apple-m1" }
attributes #1 = { returns_twice }