; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 2
; RUN: llc -mtriple aarch64-none-linux-gnu < %s | FileCheck %s

; Verify that hasNearbyPairedStore sees through non-inbounds GEPs
; (as produced by LoopStrengthReduce) and avoids unprofitable ST2
; lowering in favor of zip+stp.

define void @st2_noninbounds_gep(ptr %buf, <8 x i16> %a) {
; CHECK-LABEL: st2_noninbounds_gep:
; CHECK:       // %bb.0: // %entry
; CHECK-NEXT:    movi v1.2d, #0000000000000000
; CHECK-NEXT:    zip1 v2.8h, v0.8h, v1.8h
; CHECK-NEXT:    zip2 v0.8h, v0.8h, v1.8h
; CHECK-NEXT:    stp q2, q0, [x0]
; CHECK-NEXT:    ret
entry:
  %vzip.i = shufflevector <8 x i16> %a, <8 x i16> <i16 0, i16 0, i16 0, i16 0, i16 poison, i16 poison, i16 poison, i16 poison>, <8 x i32> <i32 0, i32 8, i32 1, i32 9, i32 2, i32 10, i32 3, i32 11>
  %vzip1.i = shufflevector <8 x i16> %a, <8 x i16> <i16 poison, i16 poison, i16 poison, i16 poison, i16 0, i16 0, i16 0, i16 0>, <8 x i32> <i32 4, i32 12, i32 5, i32 13, i32 6, i32 14, i32 7, i32 15>
  store <8 x i16> %vzip.i, ptr %buf, align 4
  %add.ptr = getelementptr i32, ptr %buf, i64 4
  store <8 x i16> %vzip1.i, ptr %add.ptr, align 4
  ret void
}

; Same pattern but with a negative non-inbounds offset, mimicking LSR output
; where the base pointer is advanced past the stores and offsets go negative.
define void @st2_noninbounds_negative_offset(ptr %buf, <8 x i16> %a) {
; CHECK-LABEL: st2_noninbounds_negative_offset:
; CHECK:       // %bb.0: // %entry
; CHECK-NEXT:    movi v1.2d, #0000000000000000
; CHECK-NEXT:    zip1 v2.8h, v0.8h, v1.8h
; CHECK-NEXT:    zip2 v0.8h, v0.8h, v1.8h
; CHECK-NEXT:    stp q2, q0, [x0]
; CHECK-NEXT:    ret
entry:
  %advanced = getelementptr i8, ptr %buf, i64 32
  %vzip.i = shufflevector <8 x i16> %a, <8 x i16> <i16 0, i16 0, i16 0, i16 0, i16 poison, i16 poison, i16 poison, i16 poison>, <8 x i32> <i32 0, i32 8, i32 1, i32 9, i32 2, i32 10, i32 3, i32 11>
  %vzip1.i = shufflevector <8 x i16> %a, <8 x i16> <i16 poison, i16 poison, i16 poison, i16 poison, i16 0, i16 0, i16 0, i16 0>, <8 x i32> <i32 4, i32 12, i32 5, i32 13, i32 6, i32 14, i32 7, i32 15>
  %ptr0 = getelementptr i8, ptr %advanced, i64 -32
  store <8 x i16> %vzip.i, ptr %ptr0, align 4
  %ptr1 = getelementptr i8, ptr %advanced, i64 -16
  store <8 x i16> %vzip1.i, ptr %ptr1, align 4
  ret void
}