llvm/test/CodeGen/AMDGPU/pal-simple-indirect-call.ll

; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --function-signature --check-globals
; Check that no attributes are added to graphics functions
; RUN: opt -S -mtriple=amdgcn-amd-amdpal -passes=amdgpu-attributor %s | FileCheck -check-prefixes=ATTRIBUTOR_GCN %s

; Check that it doesn't crash
; RUN: llc -mtriple=amdgcn-amd-amdpal -mcpu=gfx900 < %s | FileCheck -check-prefixes=GFX9 %s
; RUN: llc -mtriple=amdgcn-amd-amdpal -mcpu=gfx1010 < %s | FileCheck -check-prefixes=GFX10 %s
; RUN: llc -mtriple=amdgcn-amd-amdpal -mcpu=gfx1010 -global-isel < %s | FileCheck -check-prefixes=GFX10 %s

define amdgpu_cs void @test_simple_indirect_call() {
; ATTRIBUTOR_GCN-LABEL: define {{[^@]+}}@test_simple_indirect_call
; ATTRIBUTOR_GCN-SAME: () #[[ATTR0:[0-9]+]] {
; ATTRIBUTOR_GCN-NEXT:    [[PC:%.*]] = call i64 @llvm.amdgcn.s.getpc()
; ATTRIBUTOR_GCN-NEXT:    [[FUN:%.*]] = inttoptr i64 [[PC]] to ptr
; ATTRIBUTOR_GCN-NEXT:    call amdgpu_gfx void [[FUN]]()
; ATTRIBUTOR_GCN-NEXT:    ret void
;
; Attributor adds work-group-size attribute. This should be ok.
; GFX9-LABEL: test_simple_indirect_call:
; GFX9:       ; %bb.0:
; GFX9-NEXT:    s_getpc_b64 s[8:9]
; GFX9-NEXT:    s_mov_b32 s8, s0
; GFX9-NEXT:    s_load_dwordx4 s[8:11], s[8:9], 0x10
; GFX9-NEXT:    s_getpc_b64 s[4:5]
; GFX9-NEXT:    s_mov_b32 s32, 0
; GFX9-NEXT:    s_waitcnt lgkmcnt(0)
; GFX9-NEXT:    s_add_u32 s8, s8, s0
; GFX9-NEXT:    s_addc_u32 s9, s9, 0
; GFX9-NEXT:    s_mov_b64 s[0:1], s[8:9]
; GFX9-NEXT:    s_mov_b64 s[2:3], s[10:11]
; GFX9-NEXT:    s_swappc_b64 s[30:31], s[4:5]
; GFX9-NEXT:    s_endpgm
; GFX10-LABEL: test_simple_indirect_call:
; GFX10:       ; %bb.0:
; GFX10-NEXT:    s_getpc_b64 s[8:9]
; GFX10-NEXT:    s_mov_b32 s8, s0
; GFX10-NEXT:    s_getpc_b64 s[4:5]
; GFX10-NEXT:    s_load_dwordx4 s[8:11], s[8:9], 0x10
; GFX10-NEXT:    s_mov_b32 s32, 0
; GFX10-NEXT:    s_waitcnt lgkmcnt(0)
; GFX10-NEXT:    s_bitset0_b32 s11, 21
; GFX10-NEXT:    s_add_u32 s8, s8, s0
; GFX10-NEXT:    s_addc_u32 s9, s9, 0
; GFX10-NEXT:    s_mov_b64 s[0:1], s[8:9]
; GFX10-NEXT:    s_mov_b64 s[2:3], s[10:11]
; GFX10-NEXT:    s_swappc_b64 s[30:31], s[4:5]
; GFX10-NEXT:    s_endpgm


  %pc = call i64 @llvm.amdgcn.s.getpc()
  %fun = inttoptr i64 %pc to ptr
  call amdgpu_gfx void %fun()
  ret void
}

; Function Attrs: nounwind readnone speculatable willreturn
declare i64 @llvm.amdgcn.s.getpc() #0

attributes #0 = { nounwind readnone speculatable willreturn }
;.
; ATTRIBUTOR_GCN: attributes #[[ATTR0]] = { "amdgpu-waves-per-eu"="4,10" "uniform-work-group-size"="false" }
; ATTRIBUTOR_GCN: attributes #[[ATTR1:[0-9]+]] = { nocallback nofree nosync nounwind speculatable willreturn memory(none) }
;.