1; Test that DAG->DAG ISel is able to pick up the S_LOAD_DWORDX4_SGPR instruction that fetches the offset
2; from a register.
3
4; RUN: llc -march=amdgcn -verify-machineinstrs -stop-after=amdgpu-isel -o - %s | FileCheck -check-prefix=GCN %s
5; RUN: llc -march=amdgcn -global-isel -verify-machineinstrs -stop-after=amdgpu-isel -o - %s | FileCheck -check-prefix=GISEL %s
6
7; GCN: %[[OFFSET:[0-9]+]]:sreg_32 = S_MOV_B32 target-flags(amdgpu-abs32-lo) @DescriptorBuffer
8; GCN: %{{[0-9]+}}:sgpr_128 = S_LOAD_DWORDX4_SGPR killed %{{[0-9]+}}, killed %[[OFFSET]], 0 :: (invariant load (s128) from %ir.13, addrspace 4)
9
10; GISEL: $[[OFFSET:.*]] = S_MOV_B32 target-flags(amdgpu-abs32-lo) @DescriptorBuffer
11; GISEL: S_LOAD_DWORDX4_SGPR killed renamable {{.*}}, killed renamable $[[OFFSET]], 0 :: (invariant load (<4 x s32>) from {{.*}}, addrspace 4)
12
13define amdgpu_cs void @test_load_zext(i32 inreg %0, i32 inreg %1, i32 inreg %resNode0, i32 inreg %resNode1, <3 x i32> inreg %2, i32 inreg %3, <3 x i32> %4) local_unnamed_addr #2 {
14.entry:
15  %5 = call i64 @llvm.amdgcn.s.getpc() #3
16  %6 = bitcast i64 %5 to <2 x i32>
17  %7 = insertelement <2 x i32> %6, i32 %resNode0, i32 0
18  %8 = bitcast <2 x i32> %7 to i64
19  %9 = inttoptr i64 %8 to [4294967295 x i8] addrspace(4)*
20  %10 = call i32 @llvm.amdgcn.reloc.constant(metadata !4)
21  %11 = zext i32 %10 to i64
22  %12 = getelementptr [4294967295 x i8], [4294967295 x i8] addrspace(4)* %9, i64 0, i64 %11
23  %13 = bitcast i8 addrspace(4)* %12 to <4 x i32> addrspace(4)*, !amdgpu.uniform !5
24  %14 = load <4 x i32>, <4 x i32> addrspace(4)* %13, align 16, !invariant.load !5
25  %15 = call <4 x i32> @llvm.amdgcn.s.buffer.load.v4i32(<4 x i32> %14, i32 0, i32 0)
26  call void @llvm.amdgcn.raw.buffer.store.v4i32(<4 x i32> %15, <4 x i32> %14, i32 0, i32 0, i32 0)
27  ret void
28}
29
30declare void @llvm.amdgcn.raw.buffer.store.v4i32(<4 x i32>, <4 x i32>, i32, i32, i32 immarg) #1
31
32; Function Attrs: nounwind readnone speculatable
33declare i32 @llvm.amdgcn.reloc.constant(metadata) #3
34
35; Function Attrs: nounwind readnone speculatable
36declare i64 @llvm.amdgcn.s.getpc() #3
37
38; Function Attrs: nounwind readnone
39declare <4 x i32> @llvm.amdgcn.s.buffer.load.v4i32(<4 x i32>, i32, i32 immarg) #1
40
41attributes #0 = { argmemonly nounwind willreturn }
42attributes #1 = { nounwind readnone }
43attributes #2 = { nounwind "amdgpu-unroll-threshold"="700" }
44attributes #3 = { nounwind readnone speculatable }
45attributes #4 = { nounwind writeonly }
46
47!llpc.compute.mode = !{!0}
48!llpc.options = !{!1}
49!llpc.options.CS = !{!2}
50!llpc.user.data.nodes = !{!3, !4, !5, !6}
51!amdgpu.pal.metadata.msgpack = !{!7}
52
53!0 = !{i32 2, i32 3, i32 1}
54!1 = !{i32 245227952, i32 996822128, i32 2024708198, i32 497230408}
55!2 = !{i32 1381820427, i32 1742110173, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 64}
56!3 = !{!"DescriptorTableVaPtr", i32 0, i32 1, i32 1}
57!4 = !{!"DescriptorBuffer", i32 4, i32 8, i32 0, i32 0}
58!5 = !{!"DescriptorTableVaPtr", i32 1, i32 1, i32 1}
59!6 = !{!"DescriptorBuffer", i32 4, i32 8, i32 1, i32 0}
60!7 = !{!"\82\B0amdpal.pipelines\91\88\A4.api\A6Vulkan\B0.hardware_stages\81\A3.cs\82\AB.sgpr_limith\AB.vgpr_limit\CD\01\00\B7.internal_pipeline_hash\92\CF;jLp\0E\9D\E1\B0\CF\1D\A3\22Hx\AE\98f\AA.registers\88\CD.\07\02\CD.\08\03\CD.\09\01\CD.\12\CE\00,\00\00\CD.\13\CD\0F\88\CD.@\CE\10\00\00\00\CD.B\00\CD.C\01\A8.shaders\81\A8.compute\82\B0.api_shader_hash\92\CFg\D6}\DDR\\\E8\0B\00\B1.hardware_mapping\91\A3.cs\B0.spill_threshold\CE\FF\FF\FF\FF\A5.type\A2Cs\B0.user_data_limit\02\AEamdpal.version\92\02\03"}
61!8 = !{i32 5}
62!9 = !{!"doff_0_0_b"}
63!10 = !{}
64!11 = !{!"doff_1_0_b"}
65