; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
; RUN: llc -mtriple=amdgcn -mcpu=bonaire < %s | FileCheck -enable-var-scope -check-prefix=CI %s
; RUN: llc -mtriple=amdgcn -mcpu=tonga -mattr=-flat-for-global < %s | FileCheck -enable-var-scope -check-prefix=VI %s
; RUN: llc -mtriple=amdgcn -mcpu=gfx900 < %s | FileCheck -enable-var-scope -check-prefix=GFX9 %s

declare i32 @llvm.amdgcn.atomic.inc.i32.p1(ptr addrspace(1) nocapture, i32, i32, i32, i1) #2
declare i32 @llvm.amdgcn.atomic.inc.i32.p3(ptr addrspace(3) nocapture, i32, i32, i32, i1) #2
declare i32 @llvm.amdgcn.atomic.inc.i32.p0(ptr nocapture, i32, i32, i32, i1) #2

declare i64 @llvm.amdgcn.atomic.inc.i64.p1(ptr addrspace(1) nocapture, i64, i32, i32, i1) #2
declare i64 @llvm.amdgcn.atomic.inc.i64.p3(ptr addrspace(3) nocapture, i64, i32, i32, i1) #2
declare i64 @llvm.amdgcn.atomic.inc.i64.p0(ptr nocapture, i64, i32, i32, i1) #2

declare i32 @llvm.amdgcn.workitem.id.x() #1

define amdgpu_kernel void @lds_atomic_inc_ret_i32(ptr addrspace(1) %out, ptr addrspace(3) %ptr) #0 {
; CI-LABEL: lds_atomic_inc_ret_i32:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dword s2, s[4:5], 0xb
; CI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x9
; CI-NEXT:    v_mov_b32_e32 v0, 42
; CI-NEXT:    s_mov_b32 m0, -1
; CI-NEXT:    s_mov_b32 s3, 0xf000
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    v_mov_b32_e32 v1, s2
; CI-NEXT:    ds_inc_rtn_u32 v0, v1, v0
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    s_mov_b32 s2, -1
; CI-NEXT:    buffer_store_dword v0, off, s[0:3], 0
; CI-NEXT:    s_endpgm
;
; VI-LABEL: lds_atomic_inc_ret_i32:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dword s2, s[4:5], 0x2c
; VI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; VI-NEXT:    v_mov_b32_e32 v0, 42
; VI-NEXT:    s_mov_b32 m0, -1
; VI-NEXT:    s_mov_b32 s3, 0xf000
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    v_mov_b32_e32 v1, s2
; VI-NEXT:    ds_inc_rtn_u32 v0, v1, v0
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    s_mov_b32 s2, -1
; VI-NEXT:    buffer_store_dword v0, off, s[0:3], 0
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: lds_atomic_inc_ret_i32:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dword s2, s[4:5], 0x2c
; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v0, 42
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v1, s2
; GFX9-NEXT:    ds_inc_rtn_u32 v0, v1, v0
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v1, 0
; GFX9-NEXT:    global_store_dword v1, v0, s[0:1]
; GFX9-NEXT:    s_endpgm
  %result = call i32 @llvm.amdgcn.atomic.inc.i32.p3(ptr addrspace(3) %ptr, i32 42, i32 0, i32 0, i1 false)
  store i32 %result, ptr addrspace(1) %out
  ret void
}

define amdgpu_kernel void @lds_atomic_inc_ret_i32_offset(ptr addrspace(1) %out, ptr addrspace(3) %ptr) #0 {
; CI-LABEL: lds_atomic_inc_ret_i32_offset:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dword s2, s[4:5], 0xb
; CI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x9
; CI-NEXT:    v_mov_b32_e32 v0, 42
; CI-NEXT:    s_mov_b32 m0, -1
; CI-NEXT:    s_mov_b32 s3, 0xf000
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    v_mov_b32_e32 v1, s2
; CI-NEXT:    ds_inc_rtn_u32 v0, v1, v0 offset:16
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    s_mov_b32 s2, -1
; CI-NEXT:    buffer_store_dword v0, off, s[0:3], 0
; CI-NEXT:    s_endpgm
;
; VI-LABEL: lds_atomic_inc_ret_i32_offset:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dword s2, s[4:5], 0x2c
; VI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; VI-NEXT:    v_mov_b32_e32 v0, 42
; VI-NEXT:    s_mov_b32 m0, -1
; VI-NEXT:    s_mov_b32 s3, 0xf000
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    v_mov_b32_e32 v1, s2
; VI-NEXT:    ds_inc_rtn_u32 v0, v1, v0 offset:16
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    s_mov_b32 s2, -1
; VI-NEXT:    buffer_store_dword v0, off, s[0:3], 0
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: lds_atomic_inc_ret_i32_offset:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dword s2, s[4:5], 0x2c
; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v0, 42
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v1, s2
; GFX9-NEXT:    ds_inc_rtn_u32 v0, v1, v0 offset:16
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v1, 0
; GFX9-NEXT:    global_store_dword v1, v0, s[0:1]
; GFX9-NEXT:    s_endpgm
  %gep = getelementptr i32, ptr addrspace(3) %ptr, i32 4
  %result = call i32 @llvm.amdgcn.atomic.inc.i32.p3(ptr addrspace(3) %gep, i32 42, i32 0, i32 0, i1 false)
  store i32 %result, ptr addrspace(1) %out
  ret void
}

define amdgpu_kernel void @lds_atomic_inc_noret_i32(ptr addrspace(3) %ptr) nounwind {
; CI-LABEL: lds_atomic_inc_noret_i32:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dword s0, s[4:5], 0x9
; CI-NEXT:    v_mov_b32_e32 v0, 42
; CI-NEXT:    s_mov_b32 m0, -1
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    v_mov_b32_e32 v1, s0
; CI-NEXT:    ds_inc_u32 v1, v0
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    s_endpgm
;
; VI-LABEL: lds_atomic_inc_noret_i32:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dword s0, s[4:5], 0x24
; VI-NEXT:    v_mov_b32_e32 v0, 42
; VI-NEXT:    s_mov_b32 m0, -1
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    v_mov_b32_e32 v1, s0
; VI-NEXT:    ds_inc_u32 v1, v0
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: lds_atomic_inc_noret_i32:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dword s0, s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v0, 42
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v1, s0
; GFX9-NEXT:    ds_inc_u32 v1, v0
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    s_endpgm
  %result = call i32 @llvm.amdgcn.atomic.inc.i32.p3(ptr addrspace(3) %ptr, i32 42, i32 0, i32 0, i1 false)
  ret void
}

define amdgpu_kernel void @lds_atomic_inc_noret_i32_offset(ptr addrspace(3) %ptr) nounwind {
; CI-LABEL: lds_atomic_inc_noret_i32_offset:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dword s0, s[4:5], 0x9
; CI-NEXT:    v_mov_b32_e32 v0, 42
; CI-NEXT:    s_mov_b32 m0, -1
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    v_mov_b32_e32 v1, s0
; CI-NEXT:    ds_inc_u32 v1, v0 offset:16
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    s_endpgm
;
; VI-LABEL: lds_atomic_inc_noret_i32_offset:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dword s0, s[4:5], 0x24
; VI-NEXT:    v_mov_b32_e32 v0, 42
; VI-NEXT:    s_mov_b32 m0, -1
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    v_mov_b32_e32 v1, s0
; VI-NEXT:    ds_inc_u32 v1, v0 offset:16
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: lds_atomic_inc_noret_i32_offset:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dword s0, s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v0, 42
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v1, s0
; GFX9-NEXT:    ds_inc_u32 v1, v0 offset:16
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    s_endpgm
  %gep = getelementptr i32, ptr addrspace(3) %ptr, i32 4
  %result = call i32 @llvm.amdgcn.atomic.inc.i32.p3(ptr addrspace(3) %gep, i32 42, i32 0, i32 0, i1 false)
  ret void
}

define amdgpu_kernel void @global_atomic_inc_ret_i32(ptr addrspace(1) %out, ptr addrspace(1) %ptr) #0 {
; CI-LABEL: global_atomic_inc_ret_i32:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x9
; CI-NEXT:    s_mov_b32 s7, 0xf000
; CI-NEXT:    s_mov_b32 s6, -1
; CI-NEXT:    s_mov_b32 s10, s6
; CI-NEXT:    s_mov_b32 s11, s7
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    s_mov_b32 s8, s2
; CI-NEXT:    s_mov_b32 s9, s3
; CI-NEXT:    v_mov_b32_e32 v0, 42
; CI-NEXT:    buffer_atomic_inc v0, off, s[8:11], 0 glc
; CI-NEXT:    s_waitcnt vmcnt(0)
; CI-NEXT:    buffer_wbinvl1_vol
; CI-NEXT:    s_mov_b32 s4, s0
; CI-NEXT:    s_mov_b32 s5, s1
; CI-NEXT:    buffer_store_dword v0, off, s[4:7], 0
; CI-NEXT:    s_endpgm
;
; VI-LABEL: global_atomic_inc_ret_i32:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; VI-NEXT:    s_mov_b32 s7, 0xf000
; VI-NEXT:    s_mov_b32 s6, -1
; VI-NEXT:    s_mov_b32 s10, s6
; VI-NEXT:    s_mov_b32 s11, s7
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    s_mov_b32 s8, s2
; VI-NEXT:    s_mov_b32 s9, s3
; VI-NEXT:    v_mov_b32_e32 v0, 42
; VI-NEXT:    buffer_atomic_inc v0, off, s[8:11], 0 glc
; VI-NEXT:    s_waitcnt vmcnt(0)
; VI-NEXT:    buffer_wbinvl1_vol
; VI-NEXT:    s_mov_b32 s4, s0
; VI-NEXT:    s_mov_b32 s5, s1
; VI-NEXT:    buffer_store_dword v0, off, s[4:7], 0
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: global_atomic_inc_ret_i32:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v0, 0
; GFX9-NEXT:    v_mov_b32_e32 v1, 42
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    global_atomic_inc v1, v0, v1, s[2:3] glc
; GFX9-NEXT:    s_waitcnt vmcnt(0)
; GFX9-NEXT:    buffer_wbinvl1_vol
; GFX9-NEXT:    global_store_dword v0, v1, s[0:1]
; GFX9-NEXT:    s_endpgm
  %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1(ptr addrspace(1) %ptr, i32 42, i32 0, i32 0, i1 false)
  store i32 %result, ptr addrspace(1) %out
  ret void
}

define amdgpu_kernel void @global_atomic_inc_ret_i32_offset(ptr addrspace(1) %out, ptr addrspace(1) %ptr) #0 {
; CI-LABEL: global_atomic_inc_ret_i32_offset:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x9
; CI-NEXT:    s_mov_b32 s7, 0xf000
; CI-NEXT:    s_mov_b32 s6, -1
; CI-NEXT:    s_mov_b32 s10, s6
; CI-NEXT:    s_mov_b32 s11, s7
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    s_mov_b32 s8, s2
; CI-NEXT:    s_mov_b32 s9, s3
; CI-NEXT:    v_mov_b32_e32 v0, 42
; CI-NEXT:    buffer_atomic_inc v0, off, s[8:11], 0 offset:16 glc
; CI-NEXT:    s_waitcnt vmcnt(0)
; CI-NEXT:    buffer_wbinvl1_vol
; CI-NEXT:    s_mov_b32 s4, s0
; CI-NEXT:    s_mov_b32 s5, s1
; CI-NEXT:    buffer_store_dword v0, off, s[4:7], 0
; CI-NEXT:    s_endpgm
;
; VI-LABEL: global_atomic_inc_ret_i32_offset:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; VI-NEXT:    s_mov_b32 s7, 0xf000
; VI-NEXT:    s_mov_b32 s6, -1
; VI-NEXT:    s_mov_b32 s10, s6
; VI-NEXT:    s_mov_b32 s11, s7
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    s_mov_b32 s8, s2
; VI-NEXT:    s_mov_b32 s9, s3
; VI-NEXT:    v_mov_b32_e32 v0, 42
; VI-NEXT:    buffer_atomic_inc v0, off, s[8:11], 0 offset:16 glc
; VI-NEXT:    s_waitcnt vmcnt(0)
; VI-NEXT:    buffer_wbinvl1_vol
; VI-NEXT:    s_mov_b32 s4, s0
; VI-NEXT:    s_mov_b32 s5, s1
; VI-NEXT:    buffer_store_dword v0, off, s[4:7], 0
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: global_atomic_inc_ret_i32_offset:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v0, 0
; GFX9-NEXT:    v_mov_b32_e32 v1, 42
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    global_atomic_inc v1, v0, v1, s[2:3] offset:16 glc
; GFX9-NEXT:    s_waitcnt vmcnt(0)
; GFX9-NEXT:    buffer_wbinvl1_vol
; GFX9-NEXT:    global_store_dword v0, v1, s[0:1]
; GFX9-NEXT:    s_endpgm
  %gep = getelementptr i32, ptr addrspace(1) %ptr, i32 4
  %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1(ptr addrspace(1) %gep, i32 42, i32 0, i32 0, i1 false)
  store i32 %result, ptr addrspace(1) %out
  ret void
}

define amdgpu_kernel void @global_atomic_inc_noret_i32(ptr addrspace(1) %ptr) nounwind {
; CI-LABEL: global_atomic_inc_noret_i32:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x9
; CI-NEXT:    s_mov_b32 s3, 0xf000
; CI-NEXT:    s_mov_b32 s2, -1
; CI-NEXT:    v_mov_b32_e32 v0, 42
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    buffer_atomic_inc v0, off, s[0:3], 0
; CI-NEXT:    s_waitcnt vmcnt(0)
; CI-NEXT:    buffer_wbinvl1_vol
; CI-NEXT:    s_endpgm
;
; VI-LABEL: global_atomic_inc_noret_i32:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; VI-NEXT:    s_mov_b32 s3, 0xf000
; VI-NEXT:    s_mov_b32 s2, -1
; VI-NEXT:    v_mov_b32_e32 v0, 42
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    buffer_atomic_inc v0, off, s[0:3], 0
; VI-NEXT:    s_waitcnt vmcnt(0)
; VI-NEXT:    buffer_wbinvl1_vol
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: global_atomic_inc_noret_i32:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v0, 0
; GFX9-NEXT:    v_mov_b32_e32 v1, 42
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    global_atomic_inc v0, v1, s[0:1]
; GFX9-NEXT:    s_waitcnt vmcnt(0)
; GFX9-NEXT:    buffer_wbinvl1_vol
; GFX9-NEXT:    s_endpgm
  %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1(ptr addrspace(1) %ptr, i32 42, i32 0, i32 0, i1 false)
  ret void
}

define amdgpu_kernel void @global_atomic_inc_noret_i32_offset(ptr addrspace(1) %ptr) nounwind {
; CI-LABEL: global_atomic_inc_noret_i32_offset:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x9
; CI-NEXT:    s_mov_b32 s3, 0xf000
; CI-NEXT:    s_mov_b32 s2, -1
; CI-NEXT:    v_mov_b32_e32 v0, 42
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    buffer_atomic_inc v0, off, s[0:3], 0 offset:16
; CI-NEXT:    s_waitcnt vmcnt(0)
; CI-NEXT:    buffer_wbinvl1_vol
; CI-NEXT:    s_endpgm
;
; VI-LABEL: global_atomic_inc_noret_i32_offset:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; VI-NEXT:    s_mov_b32 s3, 0xf000
; VI-NEXT:    s_mov_b32 s2, -1
; VI-NEXT:    v_mov_b32_e32 v0, 42
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    buffer_atomic_inc v0, off, s[0:3], 0 offset:16
; VI-NEXT:    s_waitcnt vmcnt(0)
; VI-NEXT:    buffer_wbinvl1_vol
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: global_atomic_inc_noret_i32_offset:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v0, 0
; GFX9-NEXT:    v_mov_b32_e32 v1, 42
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    global_atomic_inc v0, v1, s[0:1] offset:16
; GFX9-NEXT:    s_waitcnt vmcnt(0)
; GFX9-NEXT:    buffer_wbinvl1_vol
; GFX9-NEXT:    s_endpgm
  %gep = getelementptr i32, ptr addrspace(1) %ptr, i32 4
  %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1(ptr addrspace(1) %gep, i32 42, i32 0, i32 0, i1 false)
  ret void
}

define amdgpu_kernel void @global_atomic_inc_ret_i32_offset_addr64(ptr addrspace(1) %out, ptr addrspace(1) %ptr) #0 {
; CI-LABEL: global_atomic_inc_ret_i32_offset_addr64:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x9
; CI-NEXT:    s_mov_b32 s7, 0xf000
; CI-NEXT:    s_mov_b32 s6, 0
; CI-NEXT:    v_lshlrev_b32_e32 v0, 2, v0
; CI-NEXT:    v_mov_b32_e32 v1, 0
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    s_mov_b64 s[8:9], s[2:3]
; CI-NEXT:    s_mov_b64 s[10:11], s[6:7]
; CI-NEXT:    v_mov_b32_e32 v2, 42
; CI-NEXT:    buffer_atomic_inc v2, v[0:1], s[8:11], 0 addr64 offset:20 glc
; CI-NEXT:    s_waitcnt vmcnt(0)
; CI-NEXT:    buffer_wbinvl1_vol
; CI-NEXT:    s_mov_b64 s[4:5], s[0:1]
; CI-NEXT:    buffer_store_dword v2, v[0:1], s[4:7], 0 addr64
; CI-NEXT:    s_endpgm
;
; VI-LABEL: global_atomic_inc_ret_i32_offset_addr64:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; VI-NEXT:    v_lshlrev_b32_e32 v2, 2, v0
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    v_mov_b32_e32 v0, s3
; VI-NEXT:    v_add_u32_e32 v1, vcc, s2, v2
; VI-NEXT:    v_addc_u32_e32 v3, vcc, 0, v0, vcc
; VI-NEXT:    v_add_u32_e32 v0, vcc, 20, v1
; VI-NEXT:    v_addc_u32_e32 v1, vcc, 0, v3, vcc
; VI-NEXT:    v_mov_b32_e32 v3, 42
; VI-NEXT:    flat_atomic_inc v3, v[0:1], v3 glc
; VI-NEXT:    s_waitcnt vmcnt(0)
; VI-NEXT:    buffer_wbinvl1_vol
; VI-NEXT:    v_mov_b32_e32 v1, s1
; VI-NEXT:    v_add_u32_e32 v0, vcc, s0, v2
; VI-NEXT:    v_addc_u32_e32 v1, vcc, 0, v1, vcc
; VI-NEXT:    flat_store_dword v[0:1], v3
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: global_atomic_inc_ret_i32_offset_addr64:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; GFX9-NEXT:    v_lshlrev_b32_e32 v0, 2, v0
; GFX9-NEXT:    v_mov_b32_e32 v1, 42
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    global_atomic_inc v1, v0, v1, s[2:3] offset:20 glc
; GFX9-NEXT:    s_waitcnt vmcnt(0)
; GFX9-NEXT:    buffer_wbinvl1_vol
; GFX9-NEXT:    global_store_dword v0, v1, s[0:1]
; GFX9-NEXT:    s_endpgm
  %id = call i32 @llvm.amdgcn.workitem.id.x()
  %gep.tid = getelementptr i32, ptr addrspace(1) %ptr, i32 %id
  %out.gep = getelementptr i32, ptr addrspace(1) %out, i32 %id
  %gep = getelementptr i32, ptr addrspace(1) %gep.tid, i32 5
  %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1(ptr addrspace(1) %gep, i32 42, i32 0, i32 0, i1 false)
  store i32 %result, ptr addrspace(1) %out.gep
  ret void
}

define amdgpu_kernel void @global_atomic_inc_noret_i32_offset_addr64(ptr addrspace(1) %ptr) #0 {
; CI-LABEL: global_atomic_inc_noret_i32_offset_addr64:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x9
; CI-NEXT:    v_lshlrev_b32_e32 v0, 2, v0
; CI-NEXT:    v_mov_b32_e32 v1, 0
; CI-NEXT:    s_mov_b32 s3, 0xf000
; CI-NEXT:    s_mov_b32 s2, 0
; CI-NEXT:    v_mov_b32_e32 v2, 42
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    buffer_atomic_inc v2, v[0:1], s[0:3], 0 addr64 offset:20
; CI-NEXT:    s_waitcnt vmcnt(0)
; CI-NEXT:    buffer_wbinvl1_vol
; CI-NEXT:    s_endpgm
;
; VI-LABEL: global_atomic_inc_noret_i32_offset_addr64:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; VI-NEXT:    v_lshlrev_b32_e32 v0, 2, v0
; VI-NEXT:    v_mov_b32_e32 v2, 42
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    v_mov_b32_e32 v1, s1
; VI-NEXT:    v_add_u32_e32 v0, vcc, s0, v0
; VI-NEXT:    v_addc_u32_e32 v1, vcc, 0, v1, vcc
; VI-NEXT:    v_add_u32_e32 v0, vcc, 20, v0
; VI-NEXT:    v_addc_u32_e32 v1, vcc, 0, v1, vcc
; VI-NEXT:    flat_atomic_inc v[0:1], v2
; VI-NEXT:    s_waitcnt vmcnt(0)
; VI-NEXT:    buffer_wbinvl1_vol
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: global_atomic_inc_noret_i32_offset_addr64:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; GFX9-NEXT:    v_lshlrev_b32_e32 v0, 2, v0
; GFX9-NEXT:    v_mov_b32_e32 v1, 42
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    global_atomic_inc v0, v1, s[0:1] offset:20
; GFX9-NEXT:    s_waitcnt vmcnt(0)
; GFX9-NEXT:    buffer_wbinvl1_vol
; GFX9-NEXT:    s_endpgm
  %id = call i32 @llvm.amdgcn.workitem.id.x()
  %gep.tid = getelementptr i32, ptr addrspace(1) %ptr, i32 %id
  %gep = getelementptr i32, ptr addrspace(1) %gep.tid, i32 5
  %result = call i32 @llvm.amdgcn.atomic.inc.i32.p1(ptr addrspace(1) %gep, i32 42, i32 0, i32 0, i1 false)
  ret void
}

@lds0 = addrspace(3) global [512 x i32] poison, align 4

define amdgpu_kernel void @atomic_inc_shl_base_lds_0_i32(ptr addrspace(1) %out, ptr addrspace(1) %add_use) #0 {
; CI-LABEL: atomic_inc_shl_base_lds_0_i32:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x9
; CI-NEXT:    v_lshlrev_b32_e32 v1, 2, v0
; CI-NEXT:    v_mov_b32_e32 v2, 9
; CI-NEXT:    s_mov_b32 m0, -1
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    ds_inc_rtn_u32 v1, v1, v2 offset:8
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    s_mov_b32 s7, 0xf000
; CI-NEXT:    s_mov_b32 s6, -1
; CI-NEXT:    s_mov_b32 s4, s0
; CI-NEXT:    s_mov_b32 s5, s1
; CI-NEXT:    s_mov_b32 s0, s2
; CI-NEXT:    s_mov_b32 s1, s3
; CI-NEXT:    s_mov_b32 s2, s6
; CI-NEXT:    s_mov_b32 s3, s7
; CI-NEXT:    v_add_i32_e32 v0, vcc, 2, v0
; CI-NEXT:    buffer_store_dword v0, off, s[0:3], 0
; CI-NEXT:    buffer_store_dword v1, off, s[4:7], 0
; CI-NEXT:    s_endpgm
;
; VI-LABEL: atomic_inc_shl_base_lds_0_i32:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; VI-NEXT:    v_lshlrev_b32_e32 v1, 2, v0
; VI-NEXT:    v_mov_b32_e32 v2, 9
; VI-NEXT:    s_mov_b32 m0, -1
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    ds_inc_rtn_u32 v1, v1, v2 offset:8
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    s_mov_b32 s7, 0xf000
; VI-NEXT:    s_mov_b32 s6, -1
; VI-NEXT:    s_mov_b32 s4, s0
; VI-NEXT:    s_mov_b32 s5, s1
; VI-NEXT:    s_mov_b32 s0, s2
; VI-NEXT:    s_mov_b32 s1, s3
; VI-NEXT:    s_mov_b32 s2, s6
; VI-NEXT:    s_mov_b32 s3, s7
; VI-NEXT:    v_add_u32_e32 v0, vcc, 2, v0
; VI-NEXT:    buffer_store_dword v0, off, s[0:3], 0
; VI-NEXT:    buffer_store_dword v1, off, s[4:7], 0
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: atomic_inc_shl_base_lds_0_i32:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; GFX9-NEXT:    v_lshlrev_b32_e32 v1, 2, v0
; GFX9-NEXT:    v_mov_b32_e32 v2, 9
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    ds_inc_rtn_u32 v1, v1, v2 offset:8
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v2, 0
; GFX9-NEXT:    v_add_u32_e32 v0, 2, v0
; GFX9-NEXT:    global_store_dword v2, v0, s[2:3]
; GFX9-NEXT:    global_store_dword v2, v1, s[0:1]
; GFX9-NEXT:    s_endpgm
  %tid.x = tail call i32 @llvm.amdgcn.workitem.id.x() #1
  %idx.0 = add nsw i32 %tid.x, 2
  %arrayidx0 = getelementptr inbounds [512 x i32], ptr addrspace(3) @lds0, i32 0, i32 %idx.0
  %val0 = call i32 @llvm.amdgcn.atomic.inc.i32.p3(ptr addrspace(3) %arrayidx0, i32 9, i32 0, i32 0, i1 false)
  store i32 %idx.0, ptr addrspace(1) %add_use
  store i32 %val0, ptr addrspace(1) %out
  ret void
}

define amdgpu_kernel void @lds_atomic_inc_ret_i64(ptr addrspace(1) %out, ptr addrspace(3) %ptr) #0 {
; CI-LABEL: lds_atomic_inc_ret_i64:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dword s2, s[4:5], 0xb
; CI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x9
; CI-NEXT:    v_mov_b32_e32 v0, 42
; CI-NEXT:    v_mov_b32_e32 v1, 0
; CI-NEXT:    s_mov_b32 m0, -1
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    v_mov_b32_e32 v2, s2
; CI-NEXT:    ds_inc_rtn_u64 v[0:1], v2, v[0:1]
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    s_mov_b32 s3, 0xf000
; CI-NEXT:    s_mov_b32 s2, -1
; CI-NEXT:    buffer_store_dwordx2 v[0:1], off, s[0:3], 0
; CI-NEXT:    s_endpgm
;
; VI-LABEL: lds_atomic_inc_ret_i64:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dword s2, s[4:5], 0x2c
; VI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; VI-NEXT:    v_mov_b32_e32 v0, 42
; VI-NEXT:    v_mov_b32_e32 v1, 0
; VI-NEXT:    s_mov_b32 m0, -1
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    v_mov_b32_e32 v2, s2
; VI-NEXT:    ds_inc_rtn_u64 v[0:1], v2, v[0:1]
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    s_mov_b32 s3, 0xf000
; VI-NEXT:    s_mov_b32 s2, -1
; VI-NEXT:    buffer_store_dwordx2 v[0:1], off, s[0:3], 0
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: lds_atomic_inc_ret_i64:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dword s2, s[4:5], 0x2c
; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v0, 42
; GFX9-NEXT:    v_mov_b32_e32 v1, 0
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v2, s2
; GFX9-NEXT:    ds_inc_rtn_u64 v[0:1], v2, v[0:1]
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v2, 0
; GFX9-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
; GFX9-NEXT:    s_endpgm
  %result = call i64 @llvm.amdgcn.atomic.inc.i64.p3(ptr addrspace(3) %ptr, i64 42, i32 0, i32 0, i1 false)
  store i64 %result, ptr addrspace(1) %out
  ret void
}

define amdgpu_kernel void @lds_atomic_inc_ret_i64_offset(ptr addrspace(1) %out, ptr addrspace(3) %ptr) #0 {
; CI-LABEL: lds_atomic_inc_ret_i64_offset:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dword s2, s[4:5], 0xb
; CI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x9
; CI-NEXT:    v_mov_b32_e32 v0, 42
; CI-NEXT:    v_mov_b32_e32 v1, 0
; CI-NEXT:    s_mov_b32 m0, -1
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    v_mov_b32_e32 v2, s2
; CI-NEXT:    ds_inc_rtn_u64 v[0:1], v2, v[0:1] offset:32
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    s_mov_b32 s3, 0xf000
; CI-NEXT:    s_mov_b32 s2, -1
; CI-NEXT:    buffer_store_dwordx2 v[0:1], off, s[0:3], 0
; CI-NEXT:    s_endpgm
;
; VI-LABEL: lds_atomic_inc_ret_i64_offset:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dword s2, s[4:5], 0x2c
; VI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; VI-NEXT:    v_mov_b32_e32 v0, 42
; VI-NEXT:    v_mov_b32_e32 v1, 0
; VI-NEXT:    s_mov_b32 m0, -1
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    v_mov_b32_e32 v2, s2
; VI-NEXT:    ds_inc_rtn_u64 v[0:1], v2, v[0:1] offset:32
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    s_mov_b32 s3, 0xf000
; VI-NEXT:    s_mov_b32 s2, -1
; VI-NEXT:    buffer_store_dwordx2 v[0:1], off, s[0:3], 0
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: lds_atomic_inc_ret_i64_offset:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dword s2, s[4:5], 0x2c
; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v0, 42
; GFX9-NEXT:    v_mov_b32_e32 v1, 0
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v2, s2
; GFX9-NEXT:    ds_inc_rtn_u64 v[0:1], v2, v[0:1] offset:32
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v2, 0
; GFX9-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
; GFX9-NEXT:    s_endpgm
  %gep = getelementptr i64, ptr addrspace(3) %ptr, i32 4
  %result = call i64 @llvm.amdgcn.atomic.inc.i64.p3(ptr addrspace(3) %gep, i64 42, i32 0, i32 0, i1 false)
  store i64 %result, ptr addrspace(1) %out
  ret void
}

define amdgpu_kernel void @lds_atomic_inc_noret_i64(ptr addrspace(3) %ptr) nounwind {
; CI-LABEL: lds_atomic_inc_noret_i64:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dword s0, s[4:5], 0x9
; CI-NEXT:    v_mov_b32_e32 v0, 42
; CI-NEXT:    v_mov_b32_e32 v1, 0
; CI-NEXT:    s_mov_b32 m0, -1
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    v_mov_b32_e32 v2, s0
; CI-NEXT:    ds_inc_u64 v2, v[0:1]
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    s_endpgm
;
; VI-LABEL: lds_atomic_inc_noret_i64:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dword s0, s[4:5], 0x24
; VI-NEXT:    v_mov_b32_e32 v0, 42
; VI-NEXT:    v_mov_b32_e32 v1, 0
; VI-NEXT:    s_mov_b32 m0, -1
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    v_mov_b32_e32 v2, s0
; VI-NEXT:    ds_inc_u64 v2, v[0:1]
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: lds_atomic_inc_noret_i64:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dword s0, s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v0, 42
; GFX9-NEXT:    v_mov_b32_e32 v1, 0
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v2, s0
; GFX9-NEXT:    ds_inc_u64 v2, v[0:1]
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    s_endpgm
  %result = call i64 @llvm.amdgcn.atomic.inc.i64.p3(ptr addrspace(3) %ptr, i64 42, i32 0, i32 0, i1 false)
  ret void
}

define amdgpu_kernel void @lds_atomic_inc_noret_i64_offset(ptr addrspace(3) %ptr) nounwind {
; CI-LABEL: lds_atomic_inc_noret_i64_offset:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dword s0, s[4:5], 0x9
; CI-NEXT:    v_mov_b32_e32 v0, 42
; CI-NEXT:    v_mov_b32_e32 v1, 0
; CI-NEXT:    s_mov_b32 m0, -1
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    v_mov_b32_e32 v2, s0
; CI-NEXT:    ds_inc_u64 v2, v[0:1] offset:32
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    s_endpgm
;
; VI-LABEL: lds_atomic_inc_noret_i64_offset:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dword s0, s[4:5], 0x24
; VI-NEXT:    v_mov_b32_e32 v0, 42
; VI-NEXT:    v_mov_b32_e32 v1, 0
; VI-NEXT:    s_mov_b32 m0, -1
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    v_mov_b32_e32 v2, s0
; VI-NEXT:    ds_inc_u64 v2, v[0:1] offset:32
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: lds_atomic_inc_noret_i64_offset:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dword s0, s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v0, 42
; GFX9-NEXT:    v_mov_b32_e32 v1, 0
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v2, s0
; GFX9-NEXT:    ds_inc_u64 v2, v[0:1] offset:32
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    s_endpgm
  %gep = getelementptr i64, ptr addrspace(3) %ptr, i32 4
  %result = call i64 @llvm.amdgcn.atomic.inc.i64.p3(ptr addrspace(3) %gep, i64 42, i32 0, i32 0, i1 false)
  ret void
}

define amdgpu_kernel void @global_atomic_inc_ret_i64(ptr addrspace(1) %out, ptr addrspace(1) %ptr) #0 {
; CI-LABEL: global_atomic_inc_ret_i64:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x9
; CI-NEXT:    s_mov_b32 s7, 0xf000
; CI-NEXT:    s_mov_b32 s6, -1
; CI-NEXT:    v_mov_b32_e32 v0, 42
; CI-NEXT:    s_mov_b32 s10, s6
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    s_mov_b32 s8, s2
; CI-NEXT:    s_mov_b32 s9, s3
; CI-NEXT:    v_mov_b32_e32 v1, 0
; CI-NEXT:    s_mov_b32 s11, s7
; CI-NEXT:    buffer_atomic_inc_x2 v[0:1], off, s[8:11], 0 glc
; CI-NEXT:    s_waitcnt vmcnt(0)
; CI-NEXT:    buffer_wbinvl1_vol
; CI-NEXT:    s_mov_b32 s4, s0
; CI-NEXT:    s_mov_b32 s5, s1
; CI-NEXT:    buffer_store_dwordx2 v[0:1], off, s[4:7], 0
; CI-NEXT:    s_endpgm
;
; VI-LABEL: global_atomic_inc_ret_i64:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; VI-NEXT:    s_mov_b32 s7, 0xf000
; VI-NEXT:    s_mov_b32 s6, -1
; VI-NEXT:    v_mov_b32_e32 v0, 42
; VI-NEXT:    s_mov_b32 s10, s6
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    s_mov_b32 s8, s2
; VI-NEXT:    s_mov_b32 s9, s3
; VI-NEXT:    v_mov_b32_e32 v1, 0
; VI-NEXT:    s_mov_b32 s11, s7
; VI-NEXT:    buffer_atomic_inc_x2 v[0:1], off, s[8:11], 0 glc
; VI-NEXT:    s_waitcnt vmcnt(0)
; VI-NEXT:    buffer_wbinvl1_vol
; VI-NEXT:    s_mov_b32 s4, s0
; VI-NEXT:    s_mov_b32 s5, s1
; VI-NEXT:    buffer_store_dwordx2 v[0:1], off, s[4:7], 0
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: global_atomic_inc_ret_i64:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v0, 42
; GFX9-NEXT:    v_mov_b32_e32 v1, 0
; GFX9-NEXT:    v_mov_b32_e32 v2, 0
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    global_atomic_inc_x2 v[0:1], v2, v[0:1], s[2:3] glc
; GFX9-NEXT:    s_waitcnt vmcnt(0)
; GFX9-NEXT:    buffer_wbinvl1_vol
; GFX9-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
; GFX9-NEXT:    s_endpgm
  %result = call i64 @llvm.amdgcn.atomic.inc.i64.p1(ptr addrspace(1) %ptr, i64 42, i32 0, i32 0, i1 false)
  store i64 %result, ptr addrspace(1) %out
  ret void
}

define amdgpu_kernel void @global_atomic_inc_ret_i64_offset(ptr addrspace(1) %out, ptr addrspace(1) %ptr) #0 {
; CI-LABEL: global_atomic_inc_ret_i64_offset:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x9
; CI-NEXT:    s_mov_b32 s7, 0xf000
; CI-NEXT:    s_mov_b32 s6, -1
; CI-NEXT:    v_mov_b32_e32 v0, 42
; CI-NEXT:    s_mov_b32 s10, s6
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    s_mov_b32 s8, s2
; CI-NEXT:    s_mov_b32 s9, s3
; CI-NEXT:    v_mov_b32_e32 v1, 0
; CI-NEXT:    s_mov_b32 s11, s7
; CI-NEXT:    buffer_atomic_inc_x2 v[0:1], off, s[8:11], 0 offset:32 glc
; CI-NEXT:    s_waitcnt vmcnt(0)
; CI-NEXT:    buffer_wbinvl1_vol
; CI-NEXT:    s_mov_b32 s4, s0
; CI-NEXT:    s_mov_b32 s5, s1
; CI-NEXT:    buffer_store_dwordx2 v[0:1], off, s[4:7], 0
; CI-NEXT:    s_endpgm
;
; VI-LABEL: global_atomic_inc_ret_i64_offset:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; VI-NEXT:    s_mov_b32 s7, 0xf000
; VI-NEXT:    s_mov_b32 s6, -1
; VI-NEXT:    v_mov_b32_e32 v0, 42
; VI-NEXT:    s_mov_b32 s10, s6
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    s_mov_b32 s8, s2
; VI-NEXT:    s_mov_b32 s9, s3
; VI-NEXT:    v_mov_b32_e32 v1, 0
; VI-NEXT:    s_mov_b32 s11, s7
; VI-NEXT:    buffer_atomic_inc_x2 v[0:1], off, s[8:11], 0 offset:32 glc
; VI-NEXT:    s_waitcnt vmcnt(0)
; VI-NEXT:    buffer_wbinvl1_vol
; VI-NEXT:    s_mov_b32 s4, s0
; VI-NEXT:    s_mov_b32 s5, s1
; VI-NEXT:    buffer_store_dwordx2 v[0:1], off, s[4:7], 0
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: global_atomic_inc_ret_i64_offset:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v0, 42
; GFX9-NEXT:    v_mov_b32_e32 v1, 0
; GFX9-NEXT:    v_mov_b32_e32 v2, 0
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    global_atomic_inc_x2 v[0:1], v2, v[0:1], s[2:3] offset:32 glc
; GFX9-NEXT:    s_waitcnt vmcnt(0)
; GFX9-NEXT:    buffer_wbinvl1_vol
; GFX9-NEXT:    global_store_dwordx2 v2, v[0:1], s[0:1]
; GFX9-NEXT:    s_endpgm
  %gep = getelementptr i64, ptr addrspace(1) %ptr, i32 4
  %result = call i64 @llvm.amdgcn.atomic.inc.i64.p1(ptr addrspace(1) %gep, i64 42, i32 0, i32 0, i1 false)
  store i64 %result, ptr addrspace(1) %out
  ret void
}

define amdgpu_kernel void @global_atomic_inc_noret_i64(ptr addrspace(1) %ptr) nounwind {
; CI-LABEL: global_atomic_inc_noret_i64:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x9
; CI-NEXT:    v_mov_b32_e32 v0, 42
; CI-NEXT:    s_mov_b32 s3, 0xf000
; CI-NEXT:    v_mov_b32_e32 v1, 0
; CI-NEXT:    s_mov_b32 s2, -1
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    buffer_atomic_inc_x2 v[0:1], off, s[0:3], 0
; CI-NEXT:    s_waitcnt vmcnt(0)
; CI-NEXT:    buffer_wbinvl1_vol
; CI-NEXT:    s_endpgm
;
; VI-LABEL: global_atomic_inc_noret_i64:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; VI-NEXT:    v_mov_b32_e32 v0, 42
; VI-NEXT:    s_mov_b32 s3, 0xf000
; VI-NEXT:    v_mov_b32_e32 v1, 0
; VI-NEXT:    s_mov_b32 s2, -1
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    buffer_atomic_inc_x2 v[0:1], off, s[0:3], 0
; VI-NEXT:    s_waitcnt vmcnt(0)
; VI-NEXT:    buffer_wbinvl1_vol
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: global_atomic_inc_noret_i64:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v0, 42
; GFX9-NEXT:    v_mov_b32_e32 v1, 0
; GFX9-NEXT:    v_mov_b32_e32 v2, 0
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    global_atomic_inc_x2 v2, v[0:1], s[0:1]
; GFX9-NEXT:    s_waitcnt vmcnt(0)
; GFX9-NEXT:    buffer_wbinvl1_vol
; GFX9-NEXT:    s_endpgm
  %result = call i64 @llvm.amdgcn.atomic.inc.i64.p1(ptr addrspace(1) %ptr, i64 42, i32 0, i32 0, i1 false)
  ret void
}

define amdgpu_kernel void @global_atomic_inc_noret_i64_offset(ptr addrspace(1) %ptr) nounwind {
; CI-LABEL: global_atomic_inc_noret_i64_offset:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x9
; CI-NEXT:    v_mov_b32_e32 v0, 42
; CI-NEXT:    s_mov_b32 s3, 0xf000
; CI-NEXT:    v_mov_b32_e32 v1, 0
; CI-NEXT:    s_mov_b32 s2, -1
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    buffer_atomic_inc_x2 v[0:1], off, s[0:3], 0 offset:32
; CI-NEXT:    s_waitcnt vmcnt(0)
; CI-NEXT:    buffer_wbinvl1_vol
; CI-NEXT:    s_endpgm
;
; VI-LABEL: global_atomic_inc_noret_i64_offset:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; VI-NEXT:    v_mov_b32_e32 v0, 42
; VI-NEXT:    s_mov_b32 s3, 0xf000
; VI-NEXT:    v_mov_b32_e32 v1, 0
; VI-NEXT:    s_mov_b32 s2, -1
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    buffer_atomic_inc_x2 v[0:1], off, s[0:3], 0 offset:32
; VI-NEXT:    s_waitcnt vmcnt(0)
; VI-NEXT:    buffer_wbinvl1_vol
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: global_atomic_inc_noret_i64_offset:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v0, 42
; GFX9-NEXT:    v_mov_b32_e32 v1, 0
; GFX9-NEXT:    v_mov_b32_e32 v2, 0
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    global_atomic_inc_x2 v2, v[0:1], s[0:1] offset:32
; GFX9-NEXT:    s_waitcnt vmcnt(0)
; GFX9-NEXT:    buffer_wbinvl1_vol
; GFX9-NEXT:    s_endpgm
  %gep = getelementptr i64, ptr addrspace(1) %ptr, i32 4
  %result = call i64 @llvm.amdgcn.atomic.inc.i64.p1(ptr addrspace(1) %gep, i64 42, i32 0, i32 0, i1 false)
  ret void
}

define amdgpu_kernel void @global_atomic_inc_ret_i64_offset_addr64(ptr addrspace(1) %out, ptr addrspace(1) %ptr) #0 {
; CI-LABEL: global_atomic_inc_ret_i64_offset_addr64:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x9
; CI-NEXT:    s_mov_b32 s7, 0xf000
; CI-NEXT:    s_mov_b32 s6, 0
; CI-NEXT:    v_mov_b32_e32 v2, 42
; CI-NEXT:    v_lshlrev_b32_e32 v0, 3, v0
; CI-NEXT:    v_mov_b32_e32 v1, 0
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    s_mov_b64 s[8:9], s[2:3]
; CI-NEXT:    v_mov_b32_e32 v3, 0
; CI-NEXT:    s_mov_b64 s[10:11], s[6:7]
; CI-NEXT:    buffer_atomic_inc_x2 v[2:3], v[0:1], s[8:11], 0 addr64 offset:40 glc
; CI-NEXT:    s_waitcnt vmcnt(0)
; CI-NEXT:    buffer_wbinvl1_vol
; CI-NEXT:    s_mov_b64 s[4:5], s[0:1]
; CI-NEXT:    buffer_store_dwordx2 v[2:3], v[0:1], s[4:7], 0 addr64
; CI-NEXT:    s_endpgm
;
; VI-LABEL: global_atomic_inc_ret_i64_offset_addr64:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; VI-NEXT:    v_lshlrev_b32_e32 v4, 3, v0
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    v_mov_b32_e32 v0, s3
; VI-NEXT:    v_add_u32_e32 v2, vcc, s2, v4
; VI-NEXT:    v_addc_u32_e32 v3, vcc, 0, v0, vcc
; VI-NEXT:    v_mov_b32_e32 v0, 42
; VI-NEXT:    v_add_u32_e32 v2, vcc, 40, v2
; VI-NEXT:    v_mov_b32_e32 v1, 0
; VI-NEXT:    v_addc_u32_e32 v3, vcc, 0, v3, vcc
; VI-NEXT:    flat_atomic_inc_x2 v[0:1], v[2:3], v[0:1] glc
; VI-NEXT:    s_waitcnt vmcnt(0)
; VI-NEXT:    buffer_wbinvl1_vol
; VI-NEXT:    v_mov_b32_e32 v3, s1
; VI-NEXT:    v_add_u32_e32 v2, vcc, s0, v4
; VI-NEXT:    v_addc_u32_e32 v3, vcc, 0, v3, vcc
; VI-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: global_atomic_inc_ret_i64_offset_addr64:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v1, 42
; GFX9-NEXT:    v_mov_b32_e32 v2, 0
; GFX9-NEXT:    v_lshlrev_b32_e32 v3, 3, v0
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    global_atomic_inc_x2 v[0:1], v3, v[1:2], s[2:3] offset:40 glc
; GFX9-NEXT:    s_waitcnt vmcnt(0)
; GFX9-NEXT:    buffer_wbinvl1_vol
; GFX9-NEXT:    global_store_dwordx2 v3, v[0:1], s[0:1]
; GFX9-NEXT:    s_endpgm
  %id = call i32 @llvm.amdgcn.workitem.id.x()
  %gep.tid = getelementptr i64, ptr addrspace(1) %ptr, i32 %id
  %out.gep = getelementptr i64, ptr addrspace(1) %out, i32 %id
  %gep = getelementptr i64, ptr addrspace(1) %gep.tid, i32 5
  %result = call i64 @llvm.amdgcn.atomic.inc.i64.p1(ptr addrspace(1) %gep, i64 42, i32 0, i32 0, i1 false)
  store i64 %result, ptr addrspace(1) %out.gep
  ret void
}

define amdgpu_kernel void @global_atomic_inc_noret_i64_offset_addr64(ptr addrspace(1) %ptr) #0 {
; CI-LABEL: global_atomic_inc_noret_i64_offset_addr64:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x9
; CI-NEXT:    v_mov_b32_e32 v2, 42
; CI-NEXT:    v_lshlrev_b32_e32 v0, 3, v0
; CI-NEXT:    v_mov_b32_e32 v1, 0
; CI-NEXT:    s_mov_b32 s3, 0xf000
; CI-NEXT:    v_mov_b32_e32 v3, 0
; CI-NEXT:    s_mov_b32 s2, 0
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    buffer_atomic_inc_x2 v[2:3], v[0:1], s[0:3], 0 addr64 offset:40
; CI-NEXT:    s_waitcnt vmcnt(0)
; CI-NEXT:    buffer_wbinvl1_vol
; CI-NEXT:    s_endpgm
;
; VI-LABEL: global_atomic_inc_noret_i64_offset_addr64:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; VI-NEXT:    v_lshlrev_b32_e32 v0, 3, v0
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    v_mov_b32_e32 v1, s1
; VI-NEXT:    v_add_u32_e32 v2, vcc, s0, v0
; VI-NEXT:    v_addc_u32_e32 v3, vcc, 0, v1, vcc
; VI-NEXT:    v_mov_b32_e32 v0, 42
; VI-NEXT:    v_add_u32_e32 v2, vcc, 40, v2
; VI-NEXT:    v_mov_b32_e32 v1, 0
; VI-NEXT:    v_addc_u32_e32 v3, vcc, 0, v3, vcc
; VI-NEXT:    flat_atomic_inc_x2 v[2:3], v[0:1]
; VI-NEXT:    s_waitcnt vmcnt(0)
; VI-NEXT:    buffer_wbinvl1_vol
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: global_atomic_inc_noret_i64_offset_addr64:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v1, 42
; GFX9-NEXT:    v_mov_b32_e32 v2, 0
; GFX9-NEXT:    v_lshlrev_b32_e32 v0, 3, v0
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    global_atomic_inc_x2 v0, v[1:2], s[0:1] offset:40
; GFX9-NEXT:    s_waitcnt vmcnt(0)
; GFX9-NEXT:    buffer_wbinvl1_vol
; GFX9-NEXT:    s_endpgm
  %id = call i32 @llvm.amdgcn.workitem.id.x()
  %gep.tid = getelementptr i64, ptr addrspace(1) %ptr, i32 %id
  %gep = getelementptr i64, ptr addrspace(1) %gep.tid, i32 5
  %result = call i64 @llvm.amdgcn.atomic.inc.i64.p1(ptr addrspace(1) %gep, i64 42, i32 0, i32 0, i1 false)
  ret void
}

define amdgpu_kernel void @flat_atomic_inc_ret_i32(ptr %out, ptr %ptr) #0 {
; CI-LABEL: flat_atomic_inc_ret_i32:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x9
; CI-NEXT:    v_mov_b32_e32 v2, 42
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    v_mov_b32_e32 v0, s2
; CI-NEXT:    v_mov_b32_e32 v1, s3
; CI-NEXT:    flat_atomic_inc v2, v[0:1], v2 glc
; CI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CI-NEXT:    buffer_wbinvl1_vol
; CI-NEXT:    v_mov_b32_e32 v0, s0
; CI-NEXT:    v_mov_b32_e32 v1, s1
; CI-NEXT:    flat_store_dword v[0:1], v2
; CI-NEXT:    s_endpgm
;
; VI-LABEL: flat_atomic_inc_ret_i32:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; VI-NEXT:    v_mov_b32_e32 v2, 42
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    v_mov_b32_e32 v0, s2
; VI-NEXT:    v_mov_b32_e32 v1, s3
; VI-NEXT:    flat_atomic_inc v2, v[0:1], v2 glc
; VI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; VI-NEXT:    buffer_wbinvl1_vol
; VI-NEXT:    v_mov_b32_e32 v0, s0
; VI-NEXT:    v_mov_b32_e32 v1, s1
; VI-NEXT:    flat_store_dword v[0:1], v2
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: flat_atomic_inc_ret_i32:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v2, 42
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v0, s2
; GFX9-NEXT:    v_mov_b32_e32 v1, s3
; GFX9-NEXT:    flat_atomic_inc v2, v[0:1], v2 glc
; GFX9-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX9-NEXT:    buffer_wbinvl1_vol
; GFX9-NEXT:    v_mov_b32_e32 v0, s0
; GFX9-NEXT:    v_mov_b32_e32 v1, s1
; GFX9-NEXT:    flat_store_dword v[0:1], v2
; GFX9-NEXT:    s_endpgm
  %result = call i32 @llvm.amdgcn.atomic.inc.i32.p0(ptr %ptr, i32 42, i32 0, i32 0, i1 false)
  store i32 %result, ptr %out
  ret void
}

define amdgpu_kernel void @flat_atomic_inc_ret_i32_offset(ptr %out, ptr %ptr) #0 {
; CI-LABEL: flat_atomic_inc_ret_i32_offset:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x9
; CI-NEXT:    v_mov_b32_e32 v2, 42
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    s_add_u32 s2, s2, 16
; CI-NEXT:    s_addc_u32 s3, s3, 0
; CI-NEXT:    v_mov_b32_e32 v0, s2
; CI-NEXT:    v_mov_b32_e32 v1, s3
; CI-NEXT:    flat_atomic_inc v2, v[0:1], v2 glc
; CI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CI-NEXT:    buffer_wbinvl1_vol
; CI-NEXT:    v_mov_b32_e32 v0, s0
; CI-NEXT:    v_mov_b32_e32 v1, s1
; CI-NEXT:    flat_store_dword v[0:1], v2
; CI-NEXT:    s_endpgm
;
; VI-LABEL: flat_atomic_inc_ret_i32_offset:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; VI-NEXT:    v_mov_b32_e32 v2, 42
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    s_add_u32 s2, s2, 16
; VI-NEXT:    s_addc_u32 s3, s3, 0
; VI-NEXT:    v_mov_b32_e32 v0, s2
; VI-NEXT:    v_mov_b32_e32 v1, s3
; VI-NEXT:    flat_atomic_inc v2, v[0:1], v2 glc
; VI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; VI-NEXT:    buffer_wbinvl1_vol
; VI-NEXT:    v_mov_b32_e32 v0, s0
; VI-NEXT:    v_mov_b32_e32 v1, s1
; VI-NEXT:    flat_store_dword v[0:1], v2
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: flat_atomic_inc_ret_i32_offset:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v2, 42
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v0, s2
; GFX9-NEXT:    v_mov_b32_e32 v1, s3
; GFX9-NEXT:    flat_atomic_inc v2, v[0:1], v2 offset:16 glc
; GFX9-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX9-NEXT:    buffer_wbinvl1_vol
; GFX9-NEXT:    v_mov_b32_e32 v0, s0
; GFX9-NEXT:    v_mov_b32_e32 v1, s1
; GFX9-NEXT:    flat_store_dword v[0:1], v2
; GFX9-NEXT:    s_endpgm
  %gep = getelementptr i32, ptr %ptr, i32 4
  %result = call i32 @llvm.amdgcn.atomic.inc.i32.p0(ptr %gep, i32 42, i32 0, i32 0, i1 false)
  store i32 %result, ptr %out
  ret void
}

define amdgpu_kernel void @flat_atomic_inc_noret_i32(ptr %ptr) nounwind {
; CI-LABEL: flat_atomic_inc_noret_i32:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x9
; CI-NEXT:    v_mov_b32_e32 v2, 42
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    v_mov_b32_e32 v0, s0
; CI-NEXT:    v_mov_b32_e32 v1, s1
; CI-NEXT:    flat_atomic_inc v[0:1], v2
; CI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CI-NEXT:    buffer_wbinvl1_vol
; CI-NEXT:    s_endpgm
;
; VI-LABEL: flat_atomic_inc_noret_i32:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; VI-NEXT:    v_mov_b32_e32 v2, 42
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    v_mov_b32_e32 v0, s0
; VI-NEXT:    v_mov_b32_e32 v1, s1
; VI-NEXT:    flat_atomic_inc v[0:1], v2
; VI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; VI-NEXT:    buffer_wbinvl1_vol
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: flat_atomic_inc_noret_i32:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v2, 42
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v0, s0
; GFX9-NEXT:    v_mov_b32_e32 v1, s1
; GFX9-NEXT:    flat_atomic_inc v[0:1], v2
; GFX9-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX9-NEXT:    buffer_wbinvl1_vol
; GFX9-NEXT:    s_endpgm
  %result = call i32 @llvm.amdgcn.atomic.inc.i32.p0(ptr %ptr, i32 42, i32 0, i32 0, i1 false)
  ret void
}

define amdgpu_kernel void @flat_atomic_inc_noret_i32_offset(ptr %ptr) nounwind {
; CI-LABEL: flat_atomic_inc_noret_i32_offset:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x9
; CI-NEXT:    v_mov_b32_e32 v2, 42
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    s_add_u32 s0, s0, 16
; CI-NEXT:    s_addc_u32 s1, s1, 0
; CI-NEXT:    v_mov_b32_e32 v0, s0
; CI-NEXT:    v_mov_b32_e32 v1, s1
; CI-NEXT:    flat_atomic_inc v[0:1], v2
; CI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CI-NEXT:    buffer_wbinvl1_vol
; CI-NEXT:    s_endpgm
;
; VI-LABEL: flat_atomic_inc_noret_i32_offset:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; VI-NEXT:    v_mov_b32_e32 v2, 42
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    s_add_u32 s0, s0, 16
; VI-NEXT:    s_addc_u32 s1, s1, 0
; VI-NEXT:    v_mov_b32_e32 v0, s0
; VI-NEXT:    v_mov_b32_e32 v1, s1
; VI-NEXT:    flat_atomic_inc v[0:1], v2
; VI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; VI-NEXT:    buffer_wbinvl1_vol
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: flat_atomic_inc_noret_i32_offset:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v2, 42
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v0, s0
; GFX9-NEXT:    v_mov_b32_e32 v1, s1
; GFX9-NEXT:    flat_atomic_inc v[0:1], v2 offset:16
; GFX9-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX9-NEXT:    buffer_wbinvl1_vol
; GFX9-NEXT:    s_endpgm
  %gep = getelementptr i32, ptr %ptr, i32 4
  %result = call i32 @llvm.amdgcn.atomic.inc.i32.p0(ptr %gep, i32 42, i32 0, i32 0, i1 false)
  ret void
}

define amdgpu_kernel void @flat_atomic_inc_ret_i32_offset_addr64(ptr %out, ptr %ptr) #0 {
; CI-LABEL: flat_atomic_inc_ret_i32_offset_addr64:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x9
; CI-NEXT:    v_lshlrev_b32_e32 v2, 2, v0
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    v_mov_b32_e32 v0, s3
; CI-NEXT:    v_add_i32_e32 v1, vcc, s2, v2
; CI-NEXT:    v_addc_u32_e32 v3, vcc, 0, v0, vcc
; CI-NEXT:    v_add_i32_e32 v0, vcc, 20, v1
; CI-NEXT:    v_addc_u32_e32 v1, vcc, 0, v3, vcc
; CI-NEXT:    v_mov_b32_e32 v3, 42
; CI-NEXT:    flat_atomic_inc v3, v[0:1], v3 glc
; CI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CI-NEXT:    buffer_wbinvl1_vol
; CI-NEXT:    v_mov_b32_e32 v1, s1
; CI-NEXT:    v_add_i32_e32 v0, vcc, s0, v2
; CI-NEXT:    v_addc_u32_e32 v1, vcc, 0, v1, vcc
; CI-NEXT:    flat_store_dword v[0:1], v3
; CI-NEXT:    s_endpgm
;
; VI-LABEL: flat_atomic_inc_ret_i32_offset_addr64:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; VI-NEXT:    v_lshlrev_b32_e32 v2, 2, v0
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    v_mov_b32_e32 v0, s3
; VI-NEXT:    v_add_u32_e32 v1, vcc, s2, v2
; VI-NEXT:    v_addc_u32_e32 v3, vcc, 0, v0, vcc
; VI-NEXT:    v_add_u32_e32 v0, vcc, 20, v1
; VI-NEXT:    v_addc_u32_e32 v1, vcc, 0, v3, vcc
; VI-NEXT:    v_mov_b32_e32 v3, 42
; VI-NEXT:    flat_atomic_inc v3, v[0:1], v3 glc
; VI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; VI-NEXT:    buffer_wbinvl1_vol
; VI-NEXT:    v_mov_b32_e32 v1, s1
; VI-NEXT:    v_add_u32_e32 v0, vcc, s0, v2
; VI-NEXT:    v_addc_u32_e32 v1, vcc, 0, v1, vcc
; VI-NEXT:    flat_store_dword v[0:1], v3
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: flat_atomic_inc_ret_i32_offset_addr64:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; GFX9-NEXT:    v_lshlrev_b32_e32 v2, 2, v0
; GFX9-NEXT:    v_mov_b32_e32 v3, 42
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v1, s3
; GFX9-NEXT:    v_add_co_u32_e32 v0, vcc, s2, v2
; GFX9-NEXT:    v_addc_co_u32_e32 v1, vcc, 0, v1, vcc
; GFX9-NEXT:    flat_atomic_inc v3, v[0:1], v3 offset:20 glc
; GFX9-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX9-NEXT:    buffer_wbinvl1_vol
; GFX9-NEXT:    v_mov_b32_e32 v1, s1
; GFX9-NEXT:    v_add_co_u32_e32 v0, vcc, s0, v2
; GFX9-NEXT:    v_addc_co_u32_e32 v1, vcc, 0, v1, vcc
; GFX9-NEXT:    flat_store_dword v[0:1], v3
; GFX9-NEXT:    s_endpgm
  %id = call i32 @llvm.amdgcn.workitem.id.x()
  %gep.tid = getelementptr i32, ptr %ptr, i32 %id
  %out.gep = getelementptr i32, ptr %out, i32 %id
  %gep = getelementptr i32, ptr %gep.tid, i32 5
  %result = call i32 @llvm.amdgcn.atomic.inc.i32.p0(ptr %gep, i32 42, i32 0, i32 0, i1 false)
  store i32 %result, ptr %out.gep
  ret void
}

define amdgpu_kernel void @flat_atomic_inc_noret_i32_offset_addr64(ptr %ptr) #0 {
; CI-LABEL: flat_atomic_inc_noret_i32_offset_addr64:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x9
; CI-NEXT:    v_lshlrev_b32_e32 v0, 2, v0
; CI-NEXT:    v_mov_b32_e32 v2, 42
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    v_mov_b32_e32 v1, s1
; CI-NEXT:    v_add_i32_e32 v0, vcc, s0, v0
; CI-NEXT:    v_addc_u32_e32 v1, vcc, 0, v1, vcc
; CI-NEXT:    v_add_i32_e32 v0, vcc, 20, v0
; CI-NEXT:    v_addc_u32_e32 v1, vcc, 0, v1, vcc
; CI-NEXT:    flat_atomic_inc v[0:1], v2
; CI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CI-NEXT:    buffer_wbinvl1_vol
; CI-NEXT:    s_endpgm
;
; VI-LABEL: flat_atomic_inc_noret_i32_offset_addr64:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; VI-NEXT:    v_lshlrev_b32_e32 v0, 2, v0
; VI-NEXT:    v_mov_b32_e32 v2, 42
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    v_mov_b32_e32 v1, s1
; VI-NEXT:    v_add_u32_e32 v0, vcc, s0, v0
; VI-NEXT:    v_addc_u32_e32 v1, vcc, 0, v1, vcc
; VI-NEXT:    v_add_u32_e32 v0, vcc, 20, v0
; VI-NEXT:    v_addc_u32_e32 v1, vcc, 0, v1, vcc
; VI-NEXT:    flat_atomic_inc v[0:1], v2
; VI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; VI-NEXT:    buffer_wbinvl1_vol
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: flat_atomic_inc_noret_i32_offset_addr64:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; GFX9-NEXT:    v_lshlrev_b32_e32 v0, 2, v0
; GFX9-NEXT:    v_mov_b32_e32 v2, 42
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v1, s1
; GFX9-NEXT:    v_add_co_u32_e32 v0, vcc, s0, v0
; GFX9-NEXT:    v_addc_co_u32_e32 v1, vcc, 0, v1, vcc
; GFX9-NEXT:    flat_atomic_inc v[0:1], v2 offset:20
; GFX9-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX9-NEXT:    buffer_wbinvl1_vol
; GFX9-NEXT:    s_endpgm
  %id = call i32 @llvm.amdgcn.workitem.id.x()
  %gep.tid = getelementptr i32, ptr %ptr, i32 %id
  %gep = getelementptr i32, ptr %gep.tid, i32 5
  %result = call i32 @llvm.amdgcn.atomic.inc.i32.p0(ptr %gep, i32 42, i32 0, i32 0, i1 false)
  ret void
}

@lds1 = addrspace(3) global [512 x i64] poison, align 8

define amdgpu_kernel void @atomic_inc_shl_base_lds_0_i64(ptr addrspace(1) %out, ptr addrspace(1) %add_use) #0 {
; CI-LABEL: atomic_inc_shl_base_lds_0_i64:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x9
; CI-NEXT:    v_mov_b32_e32 v1, 9
; CI-NEXT:    v_lshlrev_b32_e32 v3, 3, v0
; CI-NEXT:    v_mov_b32_e32 v2, 0
; CI-NEXT:    s_mov_b32 m0, -1
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    ds_inc_rtn_u64 v[1:2], v3, v[1:2] offset:16
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    s_mov_b32 s7, 0xf000
; CI-NEXT:    s_mov_b32 s6, -1
; CI-NEXT:    s_mov_b32 s4, s0
; CI-NEXT:    s_mov_b32 s5, s1
; CI-NEXT:    s_mov_b32 s0, s2
; CI-NEXT:    s_mov_b32 s1, s3
; CI-NEXT:    s_mov_b32 s2, s6
; CI-NEXT:    s_mov_b32 s3, s7
; CI-NEXT:    v_add_i32_e32 v0, vcc, 2, v0
; CI-NEXT:    buffer_store_dword v0, off, s[0:3], 0
; CI-NEXT:    buffer_store_dwordx2 v[1:2], off, s[4:7], 0
; CI-NEXT:    s_endpgm
;
; VI-LABEL: atomic_inc_shl_base_lds_0_i64:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; VI-NEXT:    v_mov_b32_e32 v1, 9
; VI-NEXT:    v_lshlrev_b32_e32 v3, 3, v0
; VI-NEXT:    v_mov_b32_e32 v2, 0
; VI-NEXT:    s_mov_b32 m0, -1
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    ds_inc_rtn_u64 v[1:2], v3, v[1:2] offset:16
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    s_mov_b32 s7, 0xf000
; VI-NEXT:    s_mov_b32 s6, -1
; VI-NEXT:    s_mov_b32 s4, s0
; VI-NEXT:    s_mov_b32 s5, s1
; VI-NEXT:    s_mov_b32 s0, s2
; VI-NEXT:    s_mov_b32 s1, s3
; VI-NEXT:    s_mov_b32 s2, s6
; VI-NEXT:    s_mov_b32 s3, s7
; VI-NEXT:    v_add_u32_e32 v0, vcc, 2, v0
; VI-NEXT:    buffer_store_dword v0, off, s[0:3], 0
; VI-NEXT:    buffer_store_dwordx2 v[1:2], off, s[4:7], 0
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: atomic_inc_shl_base_lds_0_i64:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    v_mov_b32_e32 v1, 9
; GFX9-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v2, 0
; GFX9-NEXT:    v_lshlrev_b32_e32 v3, 3, v0
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    ds_inc_rtn_u64 v[1:2], v3, v[1:2] offset:16
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v3, 0
; GFX9-NEXT:    v_add_u32_e32 v0, 2, v0
; GFX9-NEXT:    global_store_dword v3, v0, s[2:3]
; GFX9-NEXT:    global_store_dwordx2 v3, v[1:2], s[0:1]
; GFX9-NEXT:    s_endpgm
  %tid.x = tail call i32 @llvm.amdgcn.workitem.id.x() #1
  %idx.0 = add nsw i32 %tid.x, 2
  %arrayidx0 = getelementptr inbounds [512 x i64], ptr addrspace(3) @lds1, i32 0, i32 %idx.0
  %val0 = call i64 @llvm.amdgcn.atomic.inc.i64.p3(ptr addrspace(3) %arrayidx0, i64 9, i32 0, i32 0, i1 false)
  store i32 %idx.0, ptr addrspace(1) %add_use
  store i64 %val0, ptr addrspace(1) %out
  ret void
}

define amdgpu_kernel void @flat_atomic_inc_ret_i64(ptr %out, ptr %ptr) #0 {
; CI-LABEL: flat_atomic_inc_ret_i64:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x9
; CI-NEXT:    v_mov_b32_e32 v0, 42
; CI-NEXT:    v_mov_b32_e32 v1, 0
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    v_mov_b32_e32 v2, s2
; CI-NEXT:    v_mov_b32_e32 v3, s3
; CI-NEXT:    flat_atomic_inc_x2 v[0:1], v[2:3], v[0:1] glc
; CI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CI-NEXT:    buffer_wbinvl1_vol
; CI-NEXT:    v_mov_b32_e32 v2, s0
; CI-NEXT:    v_mov_b32_e32 v3, s1
; CI-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
; CI-NEXT:    s_endpgm
;
; VI-LABEL: flat_atomic_inc_ret_i64:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; VI-NEXT:    v_mov_b32_e32 v0, 42
; VI-NEXT:    v_mov_b32_e32 v1, 0
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    v_mov_b32_e32 v2, s2
; VI-NEXT:    v_mov_b32_e32 v3, s3
; VI-NEXT:    flat_atomic_inc_x2 v[0:1], v[2:3], v[0:1] glc
; VI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; VI-NEXT:    buffer_wbinvl1_vol
; VI-NEXT:    v_mov_b32_e32 v2, s0
; VI-NEXT:    v_mov_b32_e32 v3, s1
; VI-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: flat_atomic_inc_ret_i64:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v0, 42
; GFX9-NEXT:    v_mov_b32_e32 v1, 0
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v2, s2
; GFX9-NEXT:    v_mov_b32_e32 v3, s3
; GFX9-NEXT:    flat_atomic_inc_x2 v[0:1], v[2:3], v[0:1] glc
; GFX9-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX9-NEXT:    buffer_wbinvl1_vol
; GFX9-NEXT:    v_mov_b32_e32 v2, s0
; GFX9-NEXT:    v_mov_b32_e32 v3, s1
; GFX9-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
; GFX9-NEXT:    s_endpgm
  %result = call i64 @llvm.amdgcn.atomic.inc.i64.p0(ptr %ptr, i64 42, i32 0, i32 0, i1 false)
  store i64 %result, ptr %out
  ret void
}

define amdgpu_kernel void @flat_atomic_inc_ret_i64_offset(ptr %out, ptr %ptr) #0 {
; CI-LABEL: flat_atomic_inc_ret_i64_offset:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x9
; CI-NEXT:    v_mov_b32_e32 v0, 42
; CI-NEXT:    v_mov_b32_e32 v1, 0
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    s_add_u32 s2, s2, 32
; CI-NEXT:    s_addc_u32 s3, s3, 0
; CI-NEXT:    v_mov_b32_e32 v2, s2
; CI-NEXT:    v_mov_b32_e32 v3, s3
; CI-NEXT:    flat_atomic_inc_x2 v[0:1], v[2:3], v[0:1] glc
; CI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CI-NEXT:    buffer_wbinvl1_vol
; CI-NEXT:    v_mov_b32_e32 v2, s0
; CI-NEXT:    v_mov_b32_e32 v3, s1
; CI-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
; CI-NEXT:    s_endpgm
;
; VI-LABEL: flat_atomic_inc_ret_i64_offset:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; VI-NEXT:    v_mov_b32_e32 v0, 42
; VI-NEXT:    v_mov_b32_e32 v1, 0
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    s_add_u32 s2, s2, 32
; VI-NEXT:    s_addc_u32 s3, s3, 0
; VI-NEXT:    v_mov_b32_e32 v2, s2
; VI-NEXT:    v_mov_b32_e32 v3, s3
; VI-NEXT:    flat_atomic_inc_x2 v[0:1], v[2:3], v[0:1] glc
; VI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; VI-NEXT:    buffer_wbinvl1_vol
; VI-NEXT:    v_mov_b32_e32 v2, s0
; VI-NEXT:    v_mov_b32_e32 v3, s1
; VI-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: flat_atomic_inc_ret_i64_offset:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v0, 42
; GFX9-NEXT:    v_mov_b32_e32 v1, 0
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v2, s2
; GFX9-NEXT:    v_mov_b32_e32 v3, s3
; GFX9-NEXT:    flat_atomic_inc_x2 v[0:1], v[2:3], v[0:1] offset:32 glc
; GFX9-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX9-NEXT:    buffer_wbinvl1_vol
; GFX9-NEXT:    v_mov_b32_e32 v2, s0
; GFX9-NEXT:    v_mov_b32_e32 v3, s1
; GFX9-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
; GFX9-NEXT:    s_endpgm
  %gep = getelementptr i64, ptr %ptr, i32 4
  %result = call i64 @llvm.amdgcn.atomic.inc.i64.p0(ptr %gep, i64 42, i32 0, i32 0, i1 false)
  store i64 %result, ptr %out
  ret void
}

define amdgpu_kernel void @flat_atomic_inc_noret_i64(ptr %ptr) nounwind {
; CI-LABEL: flat_atomic_inc_noret_i64:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x9
; CI-NEXT:    v_mov_b32_e32 v0, 42
; CI-NEXT:    v_mov_b32_e32 v1, 0
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    v_mov_b32_e32 v3, s1
; CI-NEXT:    v_mov_b32_e32 v2, s0
; CI-NEXT:    flat_atomic_inc_x2 v[2:3], v[0:1]
; CI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CI-NEXT:    buffer_wbinvl1_vol
; CI-NEXT:    s_endpgm
;
; VI-LABEL: flat_atomic_inc_noret_i64:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; VI-NEXT:    v_mov_b32_e32 v0, 42
; VI-NEXT:    v_mov_b32_e32 v1, 0
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    v_mov_b32_e32 v3, s1
; VI-NEXT:    v_mov_b32_e32 v2, s0
; VI-NEXT:    flat_atomic_inc_x2 v[2:3], v[0:1]
; VI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; VI-NEXT:    buffer_wbinvl1_vol
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: flat_atomic_inc_noret_i64:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v0, 42
; GFX9-NEXT:    v_mov_b32_e32 v1, 0
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v3, s1
; GFX9-NEXT:    v_mov_b32_e32 v2, s0
; GFX9-NEXT:    flat_atomic_inc_x2 v[2:3], v[0:1]
; GFX9-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX9-NEXT:    buffer_wbinvl1_vol
; GFX9-NEXT:    s_endpgm
  %result = call i64 @llvm.amdgcn.atomic.inc.i64.p0(ptr %ptr, i64 42, i32 0, i32 0, i1 false)
  ret void
}

define amdgpu_kernel void @flat_atomic_inc_noret_i64_offset(ptr %ptr) nounwind {
; CI-LABEL: flat_atomic_inc_noret_i64_offset:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x9
; CI-NEXT:    v_mov_b32_e32 v0, 42
; CI-NEXT:    v_mov_b32_e32 v1, 0
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    s_add_u32 s0, s0, 32
; CI-NEXT:    s_addc_u32 s1, s1, 0
; CI-NEXT:    v_mov_b32_e32 v3, s1
; CI-NEXT:    v_mov_b32_e32 v2, s0
; CI-NEXT:    flat_atomic_inc_x2 v[2:3], v[0:1]
; CI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CI-NEXT:    buffer_wbinvl1_vol
; CI-NEXT:    s_endpgm
;
; VI-LABEL: flat_atomic_inc_noret_i64_offset:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; VI-NEXT:    v_mov_b32_e32 v0, 42
; VI-NEXT:    v_mov_b32_e32 v1, 0
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    s_add_u32 s0, s0, 32
; VI-NEXT:    s_addc_u32 s1, s1, 0
; VI-NEXT:    v_mov_b32_e32 v3, s1
; VI-NEXT:    v_mov_b32_e32 v2, s0
; VI-NEXT:    flat_atomic_inc_x2 v[2:3], v[0:1]
; VI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; VI-NEXT:    buffer_wbinvl1_vol
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: flat_atomic_inc_noret_i64_offset:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v0, 42
; GFX9-NEXT:    v_mov_b32_e32 v1, 0
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v3, s1
; GFX9-NEXT:    v_mov_b32_e32 v2, s0
; GFX9-NEXT:    flat_atomic_inc_x2 v[2:3], v[0:1] offset:32
; GFX9-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX9-NEXT:    buffer_wbinvl1_vol
; GFX9-NEXT:    s_endpgm
  %gep = getelementptr i64, ptr %ptr, i32 4
  %result = call i64 @llvm.amdgcn.atomic.inc.i64.p0(ptr %gep, i64 42, i32 0, i32 0, i1 false)
  ret void
}

define amdgpu_kernel void @flat_atomic_inc_ret_i64_offset_addr64(ptr %out, ptr %ptr) #0 {
; CI-LABEL: flat_atomic_inc_ret_i64_offset_addr64:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x9
; CI-NEXT:    v_lshlrev_b32_e32 v4, 3, v0
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    v_mov_b32_e32 v0, s3
; CI-NEXT:    v_add_i32_e32 v2, vcc, s2, v4
; CI-NEXT:    v_addc_u32_e32 v3, vcc, 0, v0, vcc
; CI-NEXT:    v_mov_b32_e32 v0, 42
; CI-NEXT:    v_add_i32_e32 v2, vcc, 40, v2
; CI-NEXT:    v_mov_b32_e32 v1, 0
; CI-NEXT:    v_addc_u32_e32 v3, vcc, 0, v3, vcc
; CI-NEXT:    flat_atomic_inc_x2 v[0:1], v[2:3], v[0:1] glc
; CI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CI-NEXT:    buffer_wbinvl1_vol
; CI-NEXT:    v_mov_b32_e32 v3, s1
; CI-NEXT:    v_add_i32_e32 v2, vcc, s0, v4
; CI-NEXT:    v_addc_u32_e32 v3, vcc, 0, v3, vcc
; CI-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
; CI-NEXT:    s_endpgm
;
; VI-LABEL: flat_atomic_inc_ret_i64_offset_addr64:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; VI-NEXT:    v_lshlrev_b32_e32 v4, 3, v0
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    v_mov_b32_e32 v0, s3
; VI-NEXT:    v_add_u32_e32 v2, vcc, s2, v4
; VI-NEXT:    v_addc_u32_e32 v3, vcc, 0, v0, vcc
; VI-NEXT:    v_mov_b32_e32 v0, 42
; VI-NEXT:    v_add_u32_e32 v2, vcc, 40, v2
; VI-NEXT:    v_mov_b32_e32 v1, 0
; VI-NEXT:    v_addc_u32_e32 v3, vcc, 0, v3, vcc
; VI-NEXT:    flat_atomic_inc_x2 v[0:1], v[2:3], v[0:1] glc
; VI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; VI-NEXT:    buffer_wbinvl1_vol
; VI-NEXT:    v_mov_b32_e32 v3, s1
; VI-NEXT:    v_add_u32_e32 v2, vcc, s0, v4
; VI-NEXT:    v_addc_u32_e32 v3, vcc, 0, v3, vcc
; VI-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: flat_atomic_inc_ret_i64_offset_addr64:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; GFX9-NEXT:    v_lshlrev_b32_e32 v5, 3, v0
; GFX9-NEXT:    v_mov_b32_e32 v1, 42
; GFX9-NEXT:    v_mov_b32_e32 v2, 0
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v0, s3
; GFX9-NEXT:    v_add_co_u32_e32 v3, vcc, s2, v5
; GFX9-NEXT:    v_addc_co_u32_e32 v4, vcc, 0, v0, vcc
; GFX9-NEXT:    flat_atomic_inc_x2 v[0:1], v[3:4], v[1:2] offset:40 glc
; GFX9-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX9-NEXT:    buffer_wbinvl1_vol
; GFX9-NEXT:    v_mov_b32_e32 v3, s1
; GFX9-NEXT:    v_add_co_u32_e32 v2, vcc, s0, v5
; GFX9-NEXT:    v_addc_co_u32_e32 v3, vcc, 0, v3, vcc
; GFX9-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
; GFX9-NEXT:    s_endpgm
  %id = call i32 @llvm.amdgcn.workitem.id.x()
  %gep.tid = getelementptr i64, ptr %ptr, i32 %id
  %out.gep = getelementptr i64, ptr %out, i32 %id
  %gep = getelementptr i64, ptr %gep.tid, i32 5
  %result = call i64 @llvm.amdgcn.atomic.inc.i64.p0(ptr %gep, i64 42, i32 0, i32 0, i1 false)
  store i64 %result, ptr %out.gep
  ret void
}

define amdgpu_kernel void @flat_atomic_inc_noret_i64_offset_addr64(ptr %ptr) #0 {
; CI-LABEL: flat_atomic_inc_noret_i64_offset_addr64:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x9
; CI-NEXT:    v_lshlrev_b32_e32 v0, 3, v0
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    v_mov_b32_e32 v1, s1
; CI-NEXT:    v_add_i32_e32 v2, vcc, s0, v0
; CI-NEXT:    v_addc_u32_e32 v3, vcc, 0, v1, vcc
; CI-NEXT:    v_mov_b32_e32 v0, 42
; CI-NEXT:    v_add_i32_e32 v2, vcc, 40, v2
; CI-NEXT:    v_mov_b32_e32 v1, 0
; CI-NEXT:    v_addc_u32_e32 v3, vcc, 0, v3, vcc
; CI-NEXT:    flat_atomic_inc_x2 v[2:3], v[0:1]
; CI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; CI-NEXT:    buffer_wbinvl1_vol
; CI-NEXT:    s_endpgm
;
; VI-LABEL: flat_atomic_inc_noret_i64_offset_addr64:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; VI-NEXT:    v_lshlrev_b32_e32 v0, 3, v0
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    v_mov_b32_e32 v1, s1
; VI-NEXT:    v_add_u32_e32 v2, vcc, s0, v0
; VI-NEXT:    v_addc_u32_e32 v3, vcc, 0, v1, vcc
; VI-NEXT:    v_mov_b32_e32 v0, 42
; VI-NEXT:    v_add_u32_e32 v2, vcc, 40, v2
; VI-NEXT:    v_mov_b32_e32 v1, 0
; VI-NEXT:    v_addc_u32_e32 v3, vcc, 0, v3, vcc
; VI-NEXT:    flat_atomic_inc_x2 v[2:3], v[0:1]
; VI-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; VI-NEXT:    buffer_wbinvl1_vol
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: flat_atomic_inc_noret_i64_offset_addr64:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
; GFX9-NEXT:    v_lshlrev_b32_e32 v0, 3, v0
; GFX9-NEXT:    v_mov_b32_e32 v1, 42
; GFX9-NEXT:    v_mov_b32_e32 v2, 0
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v4, s1
; GFX9-NEXT:    v_add_co_u32_e32 v3, vcc, s0, v0
; GFX9-NEXT:    v_addc_co_u32_e32 v4, vcc, 0, v4, vcc
; GFX9-NEXT:    flat_atomic_inc_x2 v[3:4], v[1:2] offset:40
; GFX9-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX9-NEXT:    buffer_wbinvl1_vol
; GFX9-NEXT:    s_endpgm
  %id = call i32 @llvm.amdgcn.workitem.id.x()
  %gep.tid = getelementptr i64, ptr %ptr, i32 %id
  %gep = getelementptr i64, ptr %gep.tid, i32 5
  %result = call i64 @llvm.amdgcn.atomic.inc.i64.p0(ptr %gep, i64 42, i32 0, i32 0, i1 false)
  ret void
}

define amdgpu_kernel void @nocse_lds_atomic_inc_ret_i32(ptr addrspace(1) %out0, ptr addrspace(1) %out1, ptr addrspace(3) %ptr) #0 {
; CI-LABEL: nocse_lds_atomic_inc_ret_i32:
; CI:       ; %bb.0:
; CI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x9
; CI-NEXT:    s_load_dword s8, s[4:5], 0xd
; CI-NEXT:    v_mov_b32_e32 v0, 42
; CI-NEXT:    s_mov_b32 m0, -1
; CI-NEXT:    s_mov_b32 s7, 0xf000
; CI-NEXT:    s_mov_b32 s6, -1
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    v_mov_b32_e32 v1, s8
; CI-NEXT:    ds_inc_rtn_u32 v2, v1, v0
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    ds_inc_rtn_u32 v0, v1, v0
; CI-NEXT:    s_waitcnt lgkmcnt(0)
; CI-NEXT:    s_mov_b32 s4, s0
; CI-NEXT:    s_mov_b32 s5, s1
; CI-NEXT:    s_mov_b32 s0, s2
; CI-NEXT:    s_mov_b32 s1, s3
; CI-NEXT:    s_mov_b32 s2, s6
; CI-NEXT:    s_mov_b32 s3, s7
; CI-NEXT:    buffer_store_dword v2, off, s[4:7], 0
; CI-NEXT:    buffer_store_dword v0, off, s[0:3], 0
; CI-NEXT:    s_endpgm
;
; VI-LABEL: nocse_lds_atomic_inc_ret_i32:
; VI:       ; %bb.0:
; VI-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; VI-NEXT:    s_load_dword s8, s[4:5], 0x34
; VI-NEXT:    v_mov_b32_e32 v0, 42
; VI-NEXT:    s_mov_b32 m0, -1
; VI-NEXT:    s_mov_b32 s7, 0xf000
; VI-NEXT:    s_mov_b32 s6, -1
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    v_mov_b32_e32 v1, s8
; VI-NEXT:    ds_inc_rtn_u32 v2, v1, v0
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    ds_inc_rtn_u32 v0, v1, v0
; VI-NEXT:    s_waitcnt lgkmcnt(0)
; VI-NEXT:    s_mov_b32 s4, s0
; VI-NEXT:    s_mov_b32 s5, s1
; VI-NEXT:    s_mov_b32 s0, s2
; VI-NEXT:    s_mov_b32 s1, s3
; VI-NEXT:    s_mov_b32 s2, s6
; VI-NEXT:    s_mov_b32 s3, s7
; VI-NEXT:    buffer_store_dword v2, off, s[4:7], 0
; VI-NEXT:    buffer_store_dword v0, off, s[0:3], 0
; VI-NEXT:    s_endpgm
;
; GFX9-LABEL: nocse_lds_atomic_inc_ret_i32:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_load_dword s6, s[4:5], 0x34
; GFX9-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; GFX9-NEXT:    v_mov_b32_e32 v0, 42
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v1, s6
; GFX9-NEXT:    ds_inc_rtn_u32 v2, v1, v0
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    ds_inc_rtn_u32 v0, v1, v0
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    v_mov_b32_e32 v1, 0
; GFX9-NEXT:    global_store_dword v1, v2, s[0:1]
; GFX9-NEXT:    global_store_dword v1, v0, s[2:3]
; GFX9-NEXT:    s_endpgm
  %result0 = call i32 @llvm.amdgcn.atomic.inc.i32.p3(ptr addrspace(3) %ptr, i32 42, i32 0, i32 0, i1 false)
  %result1 = call i32 @llvm.amdgcn.atomic.inc.i32.p3(ptr addrspace(3) %ptr, i32 42, i32 0, i32 0, i1 false)

  store i32 %result0, ptr addrspace(1) %out0
  store i32 %result1, ptr addrspace(1) %out1
  ret void
}

attributes #0 = { nounwind }
attributes #1 = { nounwind readnone }
attributes #2 = { nounwind argmemonly }