1; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py 2; RUN: llc -global-isel -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 -stop-after=irtranslator -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s 3 4declare hidden void @external_void_func_sret_struct_i8_i32_byval_struct_i8_i32(ptr addrspace(5) sret({ i8, i32 }), ptr addrspace(5) byval({ i8, i32 })) #0 5 6define amdgpu_kernel void @test_call_external_void_func_sret_struct_i8_i32_byval_struct_i8_i32(i32) #0 { 7 ; GCN-LABEL: name: test_call_external_void_func_sret_struct_i8_i32_byval_struct_i8_i32 8 ; GCN: bb.1 (%ir-block.1): 9 ; GCN-NEXT: liveins: $sgpr14, $sgpr15, $sgpr16, $vgpr0, $vgpr1, $vgpr2, $sgpr4_sgpr5, $sgpr6_sgpr7, $sgpr8_sgpr9, $sgpr10_sgpr11 10 ; GCN-NEXT: {{ $}} 11 ; GCN-NEXT: [[COPY:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr2 12 ; GCN-NEXT: [[COPY1:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr1 13 ; GCN-NEXT: [[COPY2:%[0-9]+]]:vgpr_32(s32) = COPY $vgpr0 14 ; GCN-NEXT: [[COPY3:%[0-9]+]]:sgpr_32 = COPY $sgpr16 15 ; GCN-NEXT: [[COPY4:%[0-9]+]]:sgpr_32 = COPY $sgpr15 16 ; GCN-NEXT: [[COPY5:%[0-9]+]]:sgpr_32 = COPY $sgpr14 17 ; GCN-NEXT: [[COPY6:%[0-9]+]]:sgpr_64 = COPY $sgpr10_sgpr11 18 ; GCN-NEXT: [[COPY7:%[0-9]+]]:sgpr_64 = COPY $sgpr6_sgpr7 19 ; GCN-NEXT: [[COPY8:%[0-9]+]]:sgpr_64 = COPY $sgpr4_sgpr5 20 ; GCN-NEXT: [[COPY9:%[0-9]+]]:_(p4) = COPY $sgpr8_sgpr9 21 ; GCN-NEXT: [[C:%[0-9]+]]:_(s8) = G_CONSTANT i8 3 22 ; GCN-NEXT: [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 8 23 ; GCN-NEXT: [[DEF:%[0-9]+]]:_(p1) = G_IMPLICIT_DEF 24 ; GCN-NEXT: [[FRAME_INDEX:%[0-9]+]]:_(p5) = G_FRAME_INDEX %stack.0.in.val 25 ; GCN-NEXT: [[FRAME_INDEX1:%[0-9]+]]:_(p5) = G_FRAME_INDEX %stack.1.out.val 26 ; GCN-NEXT: [[INT:%[0-9]+]]:_(p4) = G_INTRINSIC intrinsic(@llvm.amdgcn.kernarg.segment.ptr) 27 ; GCN-NEXT: [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 4 28 ; GCN-NEXT: %18:_(p5) = nuw nusw G_PTR_ADD [[FRAME_INDEX]], [[C2]](s32) 29 ; GCN-NEXT: G_STORE [[C]](s8), [[FRAME_INDEX]](p5) :: (store (s8) into %ir.in.val, addrspace 5) 30 ; GCN-NEXT: G_STORE [[C1]](s32), %18(p5) :: (store (s32) into %ir.in.gep1, addrspace 5) 31 ; GCN-NEXT: ADJCALLSTACKUP 0, 0, implicit-def $scc 32 ; GCN-NEXT: [[GV:%[0-9]+]]:_(p0) = G_GLOBAL_VALUE @external_void_func_sret_struct_i8_i32_byval_struct_i8_i32 33 ; GCN-NEXT: [[COPY10:%[0-9]+]]:_(p4) = COPY [[COPY8]] 34 ; GCN-NEXT: [[COPY11:%[0-9]+]]:_(p4) = COPY [[COPY7]] 35 ; GCN-NEXT: [[COPY12:%[0-9]+]]:_(p4) = COPY [[COPY9]](p4) 36 ; GCN-NEXT: [[C3:%[0-9]+]]:_(s64) = G_CONSTANT i64 8 37 ; GCN-NEXT: [[PTR_ADD:%[0-9]+]]:_(p4) = G_PTR_ADD [[COPY12]], [[C3]](s64) 38 ; GCN-NEXT: [[COPY13:%[0-9]+]]:_(s64) = COPY [[COPY6]] 39 ; GCN-NEXT: [[COPY14:%[0-9]+]]:_(s32) = COPY [[COPY5]] 40 ; GCN-NEXT: [[COPY15:%[0-9]+]]:_(s32) = COPY [[COPY4]] 41 ; GCN-NEXT: [[COPY16:%[0-9]+]]:_(s32) = COPY [[COPY3]] 42 ; GCN-NEXT: [[DEF1:%[0-9]+]]:_(s32) = G_IMPLICIT_DEF 43 ; GCN-NEXT: [[COPY17:%[0-9]+]]:_(s32) = COPY [[COPY2]](s32) 44 ; GCN-NEXT: [[COPY18:%[0-9]+]]:_(s32) = COPY [[COPY1]](s32) 45 ; GCN-NEXT: [[C4:%[0-9]+]]:_(s32) = G_CONSTANT i32 10 46 ; GCN-NEXT: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[COPY18]], [[C4]](s32) 47 ; GCN-NEXT: [[OR:%[0-9]+]]:_(s32) = G_OR [[COPY17]], [[SHL]] 48 ; GCN-NEXT: [[COPY19:%[0-9]+]]:_(s32) = COPY [[COPY]](s32) 49 ; GCN-NEXT: [[C5:%[0-9]+]]:_(s32) = G_CONSTANT i32 20 50 ; GCN-NEXT: [[SHL1:%[0-9]+]]:_(s32) = G_SHL [[COPY19]], [[C5]](s32) 51 ; GCN-NEXT: [[OR1:%[0-9]+]]:_(s32) = G_OR [[OR]], [[SHL1]] 52 ; GCN-NEXT: [[AMDGPU_WAVE_ADDRESS:%[0-9]+]]:_(p5) = G_AMDGPU_WAVE_ADDRESS $sp_reg 53 ; GCN-NEXT: [[C6:%[0-9]+]]:_(s32) = G_CONSTANT i32 0 54 ; GCN-NEXT: [[PTR_ADD1:%[0-9]+]]:_(p5) = G_PTR_ADD [[AMDGPU_WAVE_ADDRESS]], [[C6]](s32) 55 ; GCN-NEXT: [[C7:%[0-9]+]]:_(s32) = G_CONSTANT i32 8 56 ; GCN-NEXT: G_MEMCPY [[PTR_ADD1]](p5), [[FRAME_INDEX]](p5), [[C7]](s32), 0 :: (dereferenceable store (s64) into stack, align 4, addrspace 5), (dereferenceable load (s64) from %ir.in.val, align 4, addrspace 5) 57 ; GCN-NEXT: $vgpr0 = COPY [[FRAME_INDEX1]](p5) 58 ; GCN-NEXT: [[COPY20:%[0-9]+]]:_(<4 x s32>) = COPY $private_rsrc_reg 59 ; GCN-NEXT: $sgpr0_sgpr1_sgpr2_sgpr3 = COPY [[COPY20]](<4 x s32>) 60 ; GCN-NEXT: $sgpr4_sgpr5 = COPY [[COPY10]](p4) 61 ; GCN-NEXT: $sgpr6_sgpr7 = COPY [[COPY11]](p4) 62 ; GCN-NEXT: $sgpr8_sgpr9 = COPY [[PTR_ADD]](p4) 63 ; GCN-NEXT: $sgpr10_sgpr11 = COPY [[COPY13]](s64) 64 ; GCN-NEXT: $sgpr12 = COPY [[COPY14]](s32) 65 ; GCN-NEXT: $sgpr13 = COPY [[COPY15]](s32) 66 ; GCN-NEXT: $sgpr14 = COPY [[COPY16]](s32) 67 ; GCN-NEXT: $sgpr15 = COPY [[DEF1]](s32) 68 ; GCN-NEXT: $vgpr31 = COPY [[OR1]](s32) 69 ; GCN-NEXT: $sgpr30_sgpr31 = noconvergent G_SI_CALL [[GV]](p0), @external_void_func_sret_struct_i8_i32_byval_struct_i8_i32, csr_amdgpu, implicit $vgpr0, implicit $sgpr0_sgpr1_sgpr2_sgpr3, implicit $sgpr4_sgpr5, implicit $sgpr6_sgpr7, implicit $sgpr8_sgpr9, implicit $sgpr10_sgpr11, implicit $sgpr12, implicit $sgpr13, implicit $sgpr14, implicit $sgpr15, implicit $vgpr31 70 ; GCN-NEXT: ADJCALLSTACKDOWN 0, 8, implicit-def $scc 71 ; GCN-NEXT: %46:_(p5) = nuw nusw G_PTR_ADD [[FRAME_INDEX1]], [[C2]](s32) 72 ; GCN-NEXT: [[LOAD:%[0-9]+]]:_(s8) = G_LOAD [[FRAME_INDEX1]](p5) :: (dereferenceable load (s8) from %ir.out.val, addrspace 5) 73 ; GCN-NEXT: [[LOAD1:%[0-9]+]]:_(s32) = G_LOAD %46(p5) :: (dereferenceable load (s32) from %ir.out.gep1, addrspace 5) 74 ; GCN-NEXT: G_STORE [[LOAD]](s8), [[DEF]](p1) :: (volatile store (s8) into `ptr addrspace(1) undef`, addrspace 1) 75 ; GCN-NEXT: G_STORE [[LOAD1]](s32), [[DEF]](p1) :: (volatile store (s32) into `ptr addrspace(1) undef`, addrspace 1) 76 ; GCN-NEXT: S_ENDPGM 0 77 %in.val = alloca { i8, i32 }, align 4, addrspace(5) 78 %out.val = alloca { i8, i32 }, align 4, addrspace(5) 79 %in.gep0 = getelementptr inbounds { i8, i32 }, ptr addrspace(5) %in.val, i32 0, i32 0 80 %in.gep1 = getelementptr inbounds { i8, i32 }, ptr addrspace(5) %in.val, i32 0, i32 1 81 store i8 3, ptr addrspace(5) %in.gep0 82 store i32 8, ptr addrspace(5) %in.gep1 83 call void @external_void_func_sret_struct_i8_i32_byval_struct_i8_i32(ptr addrspace(5) %out.val, ptr addrspace(5) %in.val) 84 %out.gep0 = getelementptr inbounds { i8, i32 }, ptr addrspace(5) %out.val, i32 0, i32 0 85 %out.gep1 = getelementptr inbounds { i8, i32 }, ptr addrspace(5) %out.val, i32 0, i32 1 86 %out.val0 = load i8, ptr addrspace(5) %out.gep0 87 %out.val1 = load i32, ptr addrspace(5) %out.gep1 88 store volatile i8 %out.val0, ptr addrspace(1) undef 89 store volatile i32 %out.val1, ptr addrspace(1) undef 90 ret void 91} 92 93!llvm.module.flags = !{!0} 94!0 = !{i32 1, !"amdhsa_code_object_version", i32 500} 95