# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1010 -run-pass=none -o - %s | FileCheck %s
--- |
  ; ModuleID = 'test/CodeGen/AMDGPU/memcpy-scoped-aa.ll'
  source_filename = "test/CodeGen/AMDGPU/memcpy-scoped-aa.ll"
  target datalayout = "e-p:64:64-p1:64:64-p2:32:32-p3:32:32-p4:64:64-p5:32:32-p6:32:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-v2048:2048-n32:64-S32-A5-G1-ni:7"
  target triple = "amdgcn-amd-amdhsa"

  define i32 @test_memcpy(i32 addrspace(1)* nocapture %p, i32 addrspace(1)* nocapture readonly %q) #0 {
    %p0 = bitcast i32 addrspace(1)* %p to i8 addrspace(1)*
    %add.ptr = getelementptr inbounds i32, i32 addrspace(1)* %p, i64 4
    %p1 = bitcast i32 addrspace(1)* %add.ptr to i8 addrspace(1)*
    tail call void @llvm.memcpy.p1i8.p1i8.i64(i8 addrspace(1)* noundef nonnull align 4 dereferenceable(16) %p0, i8 addrspace(1)* noundef nonnull align 4 dereferenceable(16) %p1, i64 16, i1 false), !alias.scope !0, !noalias !3
    %1 = bitcast i32 addrspace(1)* %q to <2 x i32> addrspace(1)*
    %2 = load <2 x i32>, <2 x i32> addrspace(1)* %1, align 4, !alias.scope !3, !noalias !0
    %v01 = extractelement <2 x i32> %2, i32 0
    %v12 = extractelement <2 x i32> %2, i32 1
    %add = add i32 %v01, %v12
    ret i32 %add
  }

  define i32 @test_memcpy_inline(i32 addrspace(1)* nocapture %p, i32 addrspace(1)* nocapture readonly %q) #0 {
    %p0 = bitcast i32 addrspace(1)* %p to i8 addrspace(1)*
    %add.ptr = getelementptr inbounds i32, i32 addrspace(1)* %p, i64 4
    %p1 = bitcast i32 addrspace(1)* %add.ptr to i8 addrspace(1)*
    tail call void @llvm.memcpy.inline.p1i8.p1i8.i64(i8 addrspace(1)* noundef nonnull align 4 dereferenceable(16) %p0, i8 addrspace(1)* noundef nonnull align 4 dereferenceable(16) %p1, i64 16, i1 false), !alias.scope !0, !noalias !3
    %1 = bitcast i32 addrspace(1)* %q to <2 x i32> addrspace(1)*
    %2 = load <2 x i32>, <2 x i32> addrspace(1)* %1, align 4, !alias.scope !3, !noalias !0
    %v01 = extractelement <2 x i32> %2, i32 0
    %v12 = extractelement <2 x i32> %2, i32 1
    %add = add i32 %v01, %v12
    ret i32 %add
  }

  ; Function Attrs: argmemonly nofree nounwind willreturn
  declare void @llvm.memcpy.p1i8.p1i8.i64(i8 addrspace(1)* noalias nocapture writeonly, i8 addrspace(1)* noalias nocapture readonly, i64, i1 immarg) #1

  ; Function Attrs: argmemonly nofree nounwind willreturn
  declare void @llvm.memcpy.inline.p1i8.p1i8.i64(i8 addrspace(1)* noalias nocapture writeonly, i8 addrspace(1)* noalias nocapture readonly, i64 immarg, i1 immarg) #1

  ; Function Attrs: convergent nounwind willreturn
  declare { i1, i32 } @llvm.amdgcn.if.i32(i1) #2

  ; Function Attrs: convergent nounwind willreturn
  declare { i1, i32 } @llvm.amdgcn.else.i32.i32(i32) #2

  ; Function Attrs: convergent nounwind readnone willreturn
  declare i32 @llvm.amdgcn.if.break.i32(i1, i32) #3

  ; Function Attrs: convergent nounwind willreturn
  declare i1 @llvm.amdgcn.loop.i32(i32) #2

  ; Function Attrs: convergent nounwind willreturn
  declare void @llvm.amdgcn.end.cf.i32(i32) #2

  attributes #0 = { "target-cpu"="gfx1010" }
  attributes #1 = { argmemonly nofree nounwind willreturn "target-cpu"="gfx1010" }
  attributes #2 = { convergent nounwind willreturn }
  attributes #3 = { convergent nounwind readnone willreturn }

  !0 = !{!1}
  !1 = distinct !{!1, !2, !"bax: %p"}
  !2 = distinct !{!2, !"bax"}
  !3 = !{!4}
  !4 = distinct !{!4, !2, !"bax: %q"}

...
---
name:            test_memcpy
machineMetadataNodes:
  - '!9 = distinct !{!9, !7, !"Dst"}'
  - '!6 = distinct !{!6, !7, !"Src"}'
  - '!11 = !{!4, !6}'
  - '!5 = !{!1, !6}'
  - '!8 = !{!4, !9}'
  - '!10 = !{!1, !9}'
  - '!7 = distinct !{!7, !"MemcpyLoweringDomain"}'
body:             |
  bb.0 (%ir-block.0):
    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $sgpr30_sgpr31

    ; CHECK-LABEL: name: test_memcpy
    ; CHECK: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr30_sgpr31
    ; CHECK: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr3
    ; CHECK: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
    ; CHECK: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr1
    ; CHECK: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
    ; CHECK: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY2]], %subreg.sub0, [[COPY1]], %subreg.sub1
    ; CHECK: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY3]], %subreg.sub1
    ; CHECK: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]]
    ; CHECK: [[GLOBAL_LOAD_DWORDX4_:%[0-9]+]]:vreg_128 = GLOBAL_LOAD_DWORDX4 [[COPY5]], 16, 0, implicit $exec :: (load (s128) from %ir.p1, align 4, !alias.scope !5, !noalias !8, addrspace 1)
    ; CHECK: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]]
    ; CHECK: GLOBAL_STORE_DWORDX4 [[COPY6]], killed [[GLOBAL_LOAD_DWORDX4_]], 0, 0, implicit $exec :: (store (s128) into %ir.p0, align 4, !alias.scope !10, !noalias !11, addrspace 1)
    ; CHECK: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]]
    ; CHECK: [[GLOBAL_LOAD_DWORDX2_:%[0-9]+]]:vreg_64 = GLOBAL_LOAD_DWORDX2 killed [[COPY7]], 0, 0, implicit $exec :: (load (s64) from %ir.1, align 4, !alias.scope !3, !noalias !0, addrspace 1)
    ; CHECK: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[GLOBAL_LOAD_DWORDX2_]].sub0
    ; CHECK: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[GLOBAL_LOAD_DWORDX2_]].sub1
    ; CHECK: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[COPY8]], killed [[COPY9]], 0, implicit $exec
    ; CHECK: [[COPY10:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
    ; CHECK: $vgpr0 = COPY [[V_ADD_U32_e64_]]
    ; CHECK: [[COPY11:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY10]]
    ; CHECK: S_SETPC_B64_return [[COPY11]], implicit $vgpr0
    %4:sreg_64 = COPY $sgpr30_sgpr31
    %3:vgpr_32 = COPY $vgpr3
    %2:vgpr_32 = COPY $vgpr2
    %1:vgpr_32 = COPY $vgpr1
    %0:vgpr_32 = COPY $vgpr0
    %17:vreg_64 = REG_SEQUENCE %2, %subreg.sub0, %3, %subreg.sub1
    %18:vreg_64 = REG_SEQUENCE %0, %subreg.sub0, %1, %subreg.sub1
    %9:vreg_64 = COPY %18
    %8:vreg_128 = GLOBAL_LOAD_DWORDX4 %9, 16, 0, implicit $exec :: (load (s128) from %ir.p1, align 4, !alias.scope !5, !noalias !8, addrspace 1)
    %10:vreg_64 = COPY %18
    GLOBAL_STORE_DWORDX4 %10, killed %8, 0, 0, implicit $exec :: (store (s128) into %ir.p0, align 4, !alias.scope !10, !noalias !11, addrspace 1)
    %12:vreg_64 = COPY %17
    %11:vreg_64 = GLOBAL_LOAD_DWORDX2 killed %12, 0, 0, implicit $exec :: (load (s64) from %ir.1, align 4, !alias.scope !3, !noalias !0, addrspace 1)
    %13:vgpr_32 = COPY %11.sub0
    %14:vgpr_32 = COPY %11.sub1
    %15:vgpr_32 = V_ADD_U32_e64 killed %13, killed %14, 0, implicit $exec
    %5:ccr_sgpr_64 = COPY %4
    $vgpr0 = COPY %15
    %16:ccr_sgpr_64 = COPY %5
    S_SETPC_B64_return %16, implicit $vgpr0

...
---
name:            test_memcpy_inline
machineMetadataNodes:
  - '!6 = distinct !{!6, !7, !"Src"}'
  - '!7 = distinct !{!7, !"MemcpyLoweringDomain"}'
  - '!9 = distinct !{!9, !7, !"Dst"}'
  - '!11 = !{!4, !6}'
  - '!5 = !{!1, !6}'
  - '!8 = !{!4, !9}'
  - '!10 = !{!1, !9}'
body:             |
  bb.0 (%ir-block.0):
    liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $sgpr30_sgpr31

    ; CHECK-LABEL: name: test_memcpy_inline
    ; CHECK: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr30_sgpr31
    ; CHECK: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr3
    ; CHECK: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr2
    ; CHECK: [[COPY3:%[0-9]+]]:vgpr_32 = COPY $vgpr1
    ; CHECK: [[COPY4:%[0-9]+]]:vgpr_32 = COPY $vgpr0
    ; CHECK: [[REG_SEQUENCE:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY2]], %subreg.sub0, [[COPY1]], %subreg.sub1
    ; CHECK: [[REG_SEQUENCE1:%[0-9]+]]:vreg_64 = REG_SEQUENCE [[COPY4]], %subreg.sub0, [[COPY3]], %subreg.sub1
    ; CHECK: [[COPY5:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]]
    ; CHECK: [[GLOBAL_LOAD_DWORDX4_:%[0-9]+]]:vreg_128 = GLOBAL_LOAD_DWORDX4 [[COPY5]], 16, 0, implicit $exec :: (load (s128) from %ir.p1, align 4, !alias.scope !5, !noalias !8, addrspace 1)
    ; CHECK: [[COPY6:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE1]]
    ; CHECK: GLOBAL_STORE_DWORDX4 [[COPY6]], killed [[GLOBAL_LOAD_DWORDX4_]], 0, 0, implicit $exec :: (store (s128) into %ir.p0, align 4, !alias.scope !10, !noalias !11, addrspace 1)
    ; CHECK: [[COPY7:%[0-9]+]]:vreg_64 = COPY [[REG_SEQUENCE]]
    ; CHECK: [[GLOBAL_LOAD_DWORDX2_:%[0-9]+]]:vreg_64 = GLOBAL_LOAD_DWORDX2 killed [[COPY7]], 0, 0, implicit $exec :: (load (s64) from %ir.1, align 4, !alias.scope !3, !noalias !0, addrspace 1)
    ; CHECK: [[COPY8:%[0-9]+]]:vgpr_32 = COPY [[GLOBAL_LOAD_DWORDX2_]].sub0
    ; CHECK: [[COPY9:%[0-9]+]]:vgpr_32 = COPY [[GLOBAL_LOAD_DWORDX2_]].sub1
    ; CHECK: [[V_ADD_U32_e64_:%[0-9]+]]:vgpr_32 = V_ADD_U32_e64 killed [[COPY8]], killed [[COPY9]], 0, implicit $exec
    ; CHECK: [[COPY10:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY]]
    ; CHECK: $vgpr0 = COPY [[V_ADD_U32_e64_]]
    ; CHECK: [[COPY11:%[0-9]+]]:ccr_sgpr_64 = COPY [[COPY10]]
    ; CHECK: S_SETPC_B64_return [[COPY11]], implicit $vgpr0
    %4:sreg_64 = COPY $sgpr30_sgpr31
    %3:vgpr_32 = COPY $vgpr3
    %2:vgpr_32 = COPY $vgpr2
    %1:vgpr_32 = COPY $vgpr1
    %0:vgpr_32 = COPY $vgpr0
    %17:vreg_64 = REG_SEQUENCE %2, %subreg.sub0, %3, %subreg.sub1
    %18:vreg_64 = REG_SEQUENCE %0, %subreg.sub0, %1, %subreg.sub1
    %9:vreg_64 = COPY %18
    %8:vreg_128 = GLOBAL_LOAD_DWORDX4 %9, 16, 0, implicit $exec :: (load (s128) from %ir.p1, align 4, !alias.scope !5, !noalias !8, addrspace 1)
    %10:vreg_64 = COPY %18
    GLOBAL_STORE_DWORDX4 %10, killed %8, 0, 0, implicit $exec :: (store (s128) into %ir.p0, align 4, !alias.scope !10, !noalias !11, addrspace 1)
    %12:vreg_64 = COPY %17
    %11:vreg_64 = GLOBAL_LOAD_DWORDX2 killed %12, 0, 0, implicit $exec :: (load (s64) from %ir.1, align 4, !alias.scope !3, !noalias !0, addrspace 1)
    %13:vgpr_32 = COPY %11.sub0
    %14:vgpr_32 = COPY %11.sub1
    %15:vgpr_32 = V_ADD_U32_e64 killed %13, killed %14, 0, implicit $exec
    %5:ccr_sgpr_64 = COPY %4
    $vgpr0 = COPY %15
    %16:ccr_sgpr_64 = COPY %5
    S_SETPC_B64_return %16, implicit $vgpr0

...