1;RUN: llc < %s -march=amdgcn -mcpu=verde -verify-machineinstrs | FileCheck -check-prefixes=GCN,VERDE %s
2;RUN: llc < %s -march=amdgcn -mcpu=tonga -verify-machineinstrs | FileCheck -check-prefix=GCN %s
3
4; GCN-LABEL: {{^}}tbuffer_store:
5; GCN: v_mov_b32_e32 [[ZEROREG:v[0-9]+]], 0
6; GCN: tbuffer_store_format_xyzw v[0:3], [[ZEROREG]], s[0:3], dfmt:12, nfmt:2, 0 idxen
7; GCN: tbuffer_store_format_xyzw v[4:7], [[ZEROREG]], s[0:3], dfmt:13, nfmt:3, 0 idxen glc
8; GCN: tbuffer_store_format_xyzw v[8:11], [[ZEROREG]], s[0:3], dfmt:14, nfmt:4, 0 idxen slc
9; GCN: tbuffer_store_format_xyzw v[8:11], [[ZEROREG]], s[0:3], dfmt:14, nfmt:4, 0 idxen
10define amdgpu_ps void @tbuffer_store(<4 x i32> inreg, <4 x float>, <4 x float>, <4 x float>) {
11main_body:
12  %in1 = bitcast <4 x float> %1 to <4 x i32>
13  %in2 = bitcast <4 x float> %2 to <4 x i32>
14  %in3 = bitcast <4 x float> %3 to <4 x i32>
15  call void @llvm.amdgcn.struct.tbuffer.store.v4i32(<4 x i32> %in1, <4 x i32> %0, i32 0, i32 0, i32 0, i32 44, i32 0)
16  call void @llvm.amdgcn.struct.tbuffer.store.v4i32(<4 x i32> %in2, <4 x i32> %0, i32 0, i32 0, i32 0, i32 61, i32 1)
17  call void @llvm.amdgcn.struct.tbuffer.store.v4i32(<4 x i32> %in3, <4 x i32> %0, i32 0, i32 0, i32 0, i32 78, i32 2)
18  call void @llvm.amdgcn.struct.tbuffer.store.v4f32(<4 x float> %3, <4 x i32> %0, i32 0, i32 0, i32 0, i32 78, i32 0)
19  ret void
20}
21
22; GCN-LABEL: {{^}}tbuffer_store_immoffs:
23; GCN: v_mov_b32_e32 [[ZEROREG:v[0-9]+]], 0
24; GCN: tbuffer_store_format_xyzw v[0:3], [[ZEROREG]], s[0:3], dfmt:5, nfmt:7, 0 idxen offset:42
25define amdgpu_ps void @tbuffer_store_immoffs(<4 x i32> inreg, <4 x float>) {
26main_body:
27  %in1 = bitcast <4 x float> %1 to <4 x i32>
28  call void @llvm.amdgcn.struct.tbuffer.store.v4i32(<4 x i32> %in1, <4 x i32> %0, i32 0, i32 42, i32 0, i32 117, i32 0)
29  ret void
30}
31
32; GCN-LABEL: {{^}}tbuffer_store_scalar_and_imm_offs:
33; GCN: v_mov_b32_e32 [[ZEROREG:v[0-9]+]], 0
34; GCN: tbuffer_store_format_xyzw v[0:3], [[ZEROREG]], s[0:3], dfmt:5, nfmt:7, {{s[0-9]+}} idxen offset:42
35define amdgpu_ps void @tbuffer_store_scalar_and_imm_offs(<4 x i32> inreg, <4 x float> %vdata, i32 inreg %soffset) {
36main_body:
37  %in1 = bitcast <4 x float> %vdata to <4 x i32>
38  call void @llvm.amdgcn.struct.tbuffer.store.v4i32(<4 x i32> %in1, <4 x i32> %0, i32 0, i32 42, i32 %soffset, i32 117, i32 0)
39  ret void
40}
41
42; GCN-LABEL: {{^}}buffer_store_idx:
43; GCN: tbuffer_store_format_xyzw v[0:3], v4, s[0:3], dfmt:15, nfmt:2, 0 idxen
44define amdgpu_ps void @buffer_store_idx(<4 x i32> inreg, <4 x float> %vdata, i32 %vindex) {
45main_body:
46  %in1 = bitcast <4 x float> %vdata to <4 x i32>
47  call void @llvm.amdgcn.struct.tbuffer.store.v4i32(<4 x i32> %in1, <4 x i32> %0, i32 %vindex, i32 0, i32 0, i32 47, i32 0)
48  ret void
49}
50
51; GCN-LABEL: {{^}}buffer_store_ofs:
52; GCN: tbuffer_store_format_xyzw v[0:3], {{v\[[0-9]+:[0-9]+\]}}, s[0:3], dfmt:3, nfmt:7, 0 idxen offen
53define amdgpu_ps void @buffer_store_ofs(<4 x i32> inreg, <4 x float> %vdata, i32 %voffset) {
54main_body:
55  %in1 = bitcast <4 x float> %vdata to <4 x i32>
56  call void @llvm.amdgcn.struct.tbuffer.store.v4i32(<4 x i32> %in1, <4 x i32> %0, i32 0, i32 %voffset, i32 0, i32 115, i32 0)
57  ret void
58}
59
60; GCN-LABEL: {{^}}buffer_store_both:
61; GCN: tbuffer_store_format_xyzw v[0:3], v[4:5], s[0:3], dfmt:6, nfmt:4, 0 idxen offen
62define amdgpu_ps void @buffer_store_both(<4 x i32> inreg, <4 x float> %vdata, i32 %vindex, i32 %voffset) {
63main_body:
64  %in1 = bitcast <4 x float> %vdata to <4 x i32>
65  call void @llvm.amdgcn.struct.tbuffer.store.v4i32(<4 x i32> %in1, <4 x i32> %0, i32 %vindex, i32 %voffset, i32 0, i32 70, i32 0)
66  ret void
67}
68
69; Ideally, the register allocator would avoid the wait here
70;
71; GCN-LABEL: {{^}}buffer_store_wait:
72; GCN: tbuffer_store_format_xyzw v[0:3], v4, s[0:3], dfmt:15, nfmt:3, 0 idxen
73; VERDE: s_waitcnt expcnt(0)
74; GCN: buffer_load_format_xyzw v[0:3], v5, s[0:3], 0 idxen
75; GCN: s_waitcnt vmcnt(0)
76; GCN: tbuffer_store_format_xyzw v[0:3], v6, s[0:3], dfmt:14, nfmt:2, 0 idxen
77define amdgpu_ps void @buffer_store_wait(<4 x i32> inreg, <4 x float> %vdata, i32 %vindex.1, i32 %vindex.2, i32 %vindex.3) {
78main_body:
79  %in1 = bitcast <4 x float> %vdata to <4 x i32>
80  call void @llvm.amdgcn.struct.tbuffer.store.v4i32(<4 x i32> %in1, <4 x i32> %0, i32 %vindex.1, i32 0, i32 0, i32 63, i32 0)
81  %data = call <4 x float> @llvm.amdgcn.buffer.load.format.v4f32(<4 x i32> %0, i32 %vindex.2, i32 0, i1 0, i1 0)
82  %data.i = bitcast <4 x float> %data to <4 x i32>
83  call void @llvm.amdgcn.struct.tbuffer.store.v4i32(<4 x i32> %data.i, <4 x i32> %0, i32 %vindex.3, i32 0, i32 0, i32 46, i32 0)
84  ret void
85}
86
87; GCN-LABEL: {{^}}buffer_store_x1:
88; GCN: tbuffer_store_format_x v0, v1, s[0:3], dfmt:13, nfmt:7, 0 idxen
89define amdgpu_ps void @buffer_store_x1(<4 x i32> inreg %rsrc, float %data, i32 %vindex) {
90main_body:
91  %data.i = bitcast float %data to i32
92  call void @llvm.amdgcn.struct.tbuffer.store.i32(i32 %data.i, <4 x i32> %rsrc, i32 %vindex, i32 0, i32 0, i32 125, i32 0)
93  ret void
94}
95
96; GCN-LABEL: {{^}}buffer_store_x2:
97; GCN: tbuffer_store_format_xy v[0:1], v2, s[0:3], dfmt:1, nfmt:2, 0 idxen
98define amdgpu_ps void @buffer_store_x2(<4 x i32> inreg %rsrc, <2 x float> %data, i32 %vindex) {
99main_body:
100  %data.i = bitcast <2 x float> %data to <2 x i32>
101  call void @llvm.amdgcn.struct.tbuffer.store.v2i32(<2 x i32> %data.i, <4 x i32> %rsrc, i32 %vindex, i32 0, i32 0, i32 33, i32 0)
102  ret void
103}
104
105declare void @llvm.amdgcn.struct.tbuffer.store.i32(i32, <4 x i32>, i32, i32, i32, i32, i32) #0
106declare void @llvm.amdgcn.struct.tbuffer.store.v2i32(<2 x i32>, <4 x i32>, i32, i32, i32, i32, i32) #0
107declare void @llvm.amdgcn.struct.tbuffer.store.v4i32(<4 x i32>, <4 x i32>, i32, i32, i32, i32, i32) #0
108declare void @llvm.amdgcn.struct.tbuffer.store.v4f32(<4 x float>, <4 x i32>, i32, i32, i32, i32, i32) #0
109declare <4 x float> @llvm.amdgcn.buffer.load.format.v4f32(<4 x i32>, i32, i32, i1, i1) #1
110
111attributes #0 = { nounwind }
112attributes #1 = { nounwind readonly }
113
114
115