cc-update.ll
16.1 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx803 < %s | FileCheck --check-prefix=GFX803 %s
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 < %s | FileCheck --check-prefix=GFX900 %s
; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1010 < %s | FileCheck --check-prefix=GFX1010 %s
define amdgpu_kernel void @test_kern_empty() local_unnamed_addr #0 {
; GFX803-LABEL: test_kern_empty:
; GFX803: ; %bb.0: ; %entry
; GFX803-NEXT: s_endpgm
;
; GFX900-LABEL: test_kern_empty:
; GFX900: ; %bb.0: ; %entry
; GFX900-NEXT: s_endpgm
;
; GFX1010-LABEL: test_kern_empty:
; GFX1010: ; %bb.0: ; %entry
; GFX1010-NEXT: s_endpgm
entry:
ret void
}
define amdgpu_kernel void @test_kern_stack() local_unnamed_addr #0 {
; GFX803-LABEL: test_kern_stack:
; GFX803: ; %bb.0: ; %entry
; GFX803-NEXT: s_add_u32 s4, s4, s7
; GFX803-NEXT: s_lshr_b32 flat_scratch_hi, s4, 8
; GFX803-NEXT: s_add_u32 s0, s0, s7
; GFX803-NEXT: s_addc_u32 s1, s1, 0
; GFX803-NEXT: v_mov_b32_e32 v0, 0
; GFX803-NEXT: s_mov_b32 flat_scratch_lo, s5
; GFX803-NEXT: buffer_store_dword v0, off, s[0:3], 0 offset:4
; GFX803-NEXT: s_endpgm
;
; GFX900-LABEL: test_kern_stack:
; GFX900: ; %bb.0: ; %entry
; GFX900-NEXT: s_add_u32 flat_scratch_lo, s4, s7
; GFX900-NEXT: s_addc_u32 flat_scratch_hi, s5, 0
; GFX900-NEXT: s_add_u32 s0, s0, s7
; GFX900-NEXT: s_addc_u32 s1, s1, 0
; GFX900-NEXT: v_mov_b32_e32 v0, 0
; GFX900-NEXT: buffer_store_dword v0, off, s[0:3], 0 offset:4
; GFX900-NEXT: s_endpgm
;
; GFX1010-LABEL: test_kern_stack:
; GFX1010: ; %bb.0: ; %entry
; GFX1010-NEXT: s_add_u32 s4, s4, s7
; GFX1010-NEXT: s_addc_u32 s5, s5, 0
; GFX1010-NEXT: s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s4
; GFX1010-NEXT: s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s5
; GFX1010-NEXT: v_mov_b32_e32 v0, 0
; GFX1010-NEXT: s_add_u32 s0, s0, s7
; GFX1010-NEXT: s_addc_u32 s1, s1, 0
; GFX1010-NEXT: ; implicit-def: $vcc_hi
; GFX1010-NEXT: buffer_store_dword v0, off, s[0:3], 0 offset:4
; GFX1010-NEXT: s_endpgm
entry:
%x = alloca i32, align 4, addrspace(5)
store volatile i32 0, i32 addrspace(5)* %x, align 4
ret void
}
define amdgpu_kernel void @test_kern_call() local_unnamed_addr #0 {
; GFX803-LABEL: test_kern_call:
; GFX803: ; %bb.0: ; %entry
; GFX803-NEXT: s_add_u32 s4, s4, s7
; GFX803-NEXT: s_lshr_b32 flat_scratch_hi, s4, 8
; GFX803-NEXT: s_add_u32 s0, s0, s7
; GFX803-NEXT: s_addc_u32 s1, s1, 0
; GFX803-NEXT: s_mov_b32 flat_scratch_lo, s5
; GFX803-NEXT: s_getpc_b64 s[4:5]
; GFX803-NEXT: s_add_u32 s4, s4, ex@rel32@lo+4
; GFX803-NEXT: s_addc_u32 s5, s5, ex@rel32@hi+12
; GFX803-NEXT: s_mov_b32 s32, 0
; GFX803-NEXT: s_swappc_b64 s[30:31], s[4:5]
; GFX803-NEXT: s_endpgm
;
; GFX900-LABEL: test_kern_call:
; GFX900: ; %bb.0: ; %entry
; GFX900-NEXT: s_add_u32 flat_scratch_lo, s4, s7
; GFX900-NEXT: s_addc_u32 flat_scratch_hi, s5, 0
; GFX900-NEXT: s_add_u32 s0, s0, s7
; GFX900-NEXT: s_addc_u32 s1, s1, 0
; GFX900-NEXT: s_getpc_b64 s[4:5]
; GFX900-NEXT: s_add_u32 s4, s4, ex@rel32@lo+4
; GFX900-NEXT: s_addc_u32 s5, s5, ex@rel32@hi+12
; GFX900-NEXT: s_mov_b32 s32, 0
; GFX900-NEXT: s_swappc_b64 s[30:31], s[4:5]
; GFX900-NEXT: s_endpgm
;
; GFX1010-LABEL: test_kern_call:
; GFX1010: ; %bb.0: ; %entry
; GFX1010-NEXT: s_add_u32 s4, s4, s7
; GFX1010-NEXT: s_mov_b32 s32, 0
; GFX1010-NEXT: s_addc_u32 s5, s5, 0
; GFX1010-NEXT: s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s4
; GFX1010-NEXT: s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s5
; GFX1010-NEXT: s_add_u32 s0, s0, s7
; GFX1010-NEXT: s_addc_u32 s1, s1, 0
; GFX1010-NEXT: s_getpc_b64 s[4:5]
; GFX1010-NEXT: s_add_u32 s4, s4, ex@rel32@lo+4
; GFX1010-NEXT: s_addc_u32 s5, s5, ex@rel32@hi+12
; GFX1010-NEXT: ; implicit-def: $vcc_hi
; GFX1010-NEXT: s_swappc_b64 s[30:31], s[4:5]
; GFX1010-NEXT: s_endpgm
entry:
tail call void @ex() #0
ret void
}
define amdgpu_kernel void @test_kern_stack_and_call() local_unnamed_addr #0 {
; GFX803-LABEL: test_kern_stack_and_call:
; GFX803: ; %bb.0: ; %entry
; GFX803-NEXT: s_add_u32 s4, s4, s7
; GFX803-NEXT: s_lshr_b32 flat_scratch_hi, s4, 8
; GFX803-NEXT: s_add_u32 s0, s0, s7
; GFX803-NEXT: s_addc_u32 s1, s1, 0
; GFX803-NEXT: v_mov_b32_e32 v0, 0
; GFX803-NEXT: s_mov_b32 flat_scratch_lo, s5
; GFX803-NEXT: s_getpc_b64 s[4:5]
; GFX803-NEXT: s_add_u32 s4, s4, ex@rel32@lo+4
; GFX803-NEXT: s_addc_u32 s5, s5, ex@rel32@hi+12
; GFX803-NEXT: s_movk_i32 s32, 0x400
; GFX803-NEXT: buffer_store_dword v0, off, s[0:3], 0 offset:4
; GFX803-NEXT: s_swappc_b64 s[30:31], s[4:5]
; GFX803-NEXT: s_endpgm
;
; GFX900-LABEL: test_kern_stack_and_call:
; GFX900: ; %bb.0: ; %entry
; GFX900-NEXT: s_add_u32 flat_scratch_lo, s4, s7
; GFX900-NEXT: s_addc_u32 flat_scratch_hi, s5, 0
; GFX900-NEXT: s_add_u32 s0, s0, s7
; GFX900-NEXT: s_addc_u32 s1, s1, 0
; GFX900-NEXT: v_mov_b32_e32 v0, 0
; GFX900-NEXT: s_getpc_b64 s[4:5]
; GFX900-NEXT: s_add_u32 s4, s4, ex@rel32@lo+4
; GFX900-NEXT: s_addc_u32 s5, s5, ex@rel32@hi+12
; GFX900-NEXT: s_movk_i32 s32, 0x400
; GFX900-NEXT: buffer_store_dword v0, off, s[0:3], 0 offset:4
; GFX900-NEXT: s_swappc_b64 s[30:31], s[4:5]
; GFX900-NEXT: s_endpgm
;
; GFX1010-LABEL: test_kern_stack_and_call:
; GFX1010: ; %bb.0: ; %entry
; GFX1010-NEXT: s_add_u32 s4, s4, s7
; GFX1010-NEXT: s_movk_i32 s32, 0x200
; GFX1010-NEXT: s_addc_u32 s5, s5, 0
; GFX1010-NEXT: s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s4
; GFX1010-NEXT: s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s5
; GFX1010-NEXT: v_mov_b32_e32 v0, 0
; GFX1010-NEXT: s_add_u32 s0, s0, s7
; GFX1010-NEXT: s_addc_u32 s1, s1, 0
; GFX1010-NEXT: s_getpc_b64 s[4:5]
; GFX1010-NEXT: s_add_u32 s4, s4, ex@rel32@lo+4
; GFX1010-NEXT: s_addc_u32 s5, s5, ex@rel32@hi+12
; GFX1010-NEXT: ; implicit-def: $vcc_hi
; GFX1010-NEXT: buffer_store_dword v0, off, s[0:3], 0 offset:4
; GFX1010-NEXT: s_swappc_b64 s[30:31], s[4:5]
; GFX1010-NEXT: s_endpgm
entry:
%x = alloca i32, align 4, addrspace(5)
store volatile i32 0, i32 addrspace(5)* %x, align 4
tail call void @ex() #0
ret void
}
define amdgpu_kernel void @test_force_fp_kern_empty() local_unnamed_addr #2 {
; GFX803-LABEL: test_force_fp_kern_empty:
; GFX803: ; %bb.0: ; %entry
; GFX803-NEXT: s_mov_b32 s33, 0
; GFX803-NEXT: s_endpgm
;
; GFX900-LABEL: test_force_fp_kern_empty:
; GFX900: ; %bb.0: ; %entry
; GFX900-NEXT: s_mov_b32 s33, 0
; GFX900-NEXT: s_endpgm
;
; GFX1010-LABEL: test_force_fp_kern_empty:
; GFX1010: ; %bb.0: ; %entry
; GFX1010-NEXT: s_mov_b32 s33, 0
; GFX1010-NEXT: s_endpgm
entry:
ret void
}
define amdgpu_kernel void @test_force_fp_kern_stack() local_unnamed_addr #2 {
; GFX803-LABEL: test_force_fp_kern_stack:
; GFX803: ; %bb.0: ; %entry
; GFX803-NEXT: s_add_u32 s4, s4, s7
; GFX803-NEXT: s_lshr_b32 flat_scratch_hi, s4, 8
; GFX803-NEXT: s_add_u32 s0, s0, s7
; GFX803-NEXT: s_mov_b32 s33, 0
; GFX803-NEXT: s_addc_u32 s1, s1, 0
; GFX803-NEXT: v_mov_b32_e32 v0, 0
; GFX803-NEXT: s_mov_b32 flat_scratch_lo, s5
; GFX803-NEXT: buffer_store_dword v0, off, s[0:3], s33 offset:4
; GFX803-NEXT: s_endpgm
;
; GFX900-LABEL: test_force_fp_kern_stack:
; GFX900: ; %bb.0: ; %entry
; GFX900-NEXT: s_add_u32 flat_scratch_lo, s4, s7
; GFX900-NEXT: s_addc_u32 flat_scratch_hi, s5, 0
; GFX900-NEXT: s_add_u32 s0, s0, s7
; GFX900-NEXT: s_mov_b32 s33, 0
; GFX900-NEXT: s_addc_u32 s1, s1, 0
; GFX900-NEXT: v_mov_b32_e32 v0, 0
; GFX900-NEXT: buffer_store_dword v0, off, s[0:3], s33 offset:4
; GFX900-NEXT: s_endpgm
;
; GFX1010-LABEL: test_force_fp_kern_stack:
; GFX1010: ; %bb.0: ; %entry
; GFX1010-NEXT: s_add_u32 s4, s4, s7
; GFX1010-NEXT: s_mov_b32 s33, 0
; GFX1010-NEXT: s_addc_u32 s5, s5, 0
; GFX1010-NEXT: s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s4
; GFX1010-NEXT: s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s5
; GFX1010-NEXT: v_mov_b32_e32 v0, 0
; GFX1010-NEXT: s_add_u32 s0, s0, s7
; GFX1010-NEXT: s_addc_u32 s1, s1, 0
; GFX1010-NEXT: ; implicit-def: $vcc_hi
; GFX1010-NEXT: buffer_store_dword v0, off, s[0:3], s33 offset:4
; GFX1010-NEXT: s_endpgm
entry:
%x = alloca i32, align 4, addrspace(5)
store volatile i32 0, i32 addrspace(5)* %x, align 4
ret void
}
define amdgpu_kernel void @test_force_fp_kern_call() local_unnamed_addr #2 {
; GFX803-LABEL: test_force_fp_kern_call:
; GFX803: ; %bb.0: ; %entry
; GFX803-NEXT: s_add_u32 s4, s4, s7
; GFX803-NEXT: s_lshr_b32 flat_scratch_hi, s4, 8
; GFX803-NEXT: s_add_u32 s0, s0, s7
; GFX803-NEXT: s_addc_u32 s1, s1, 0
; GFX803-NEXT: s_mov_b32 flat_scratch_lo, s5
; GFX803-NEXT: s_getpc_b64 s[4:5]
; GFX803-NEXT: s_add_u32 s4, s4, ex@rel32@lo+4
; GFX803-NEXT: s_addc_u32 s5, s5, ex@rel32@hi+12
; GFX803-NEXT: s_mov_b32 s32, 0
; GFX803-NEXT: s_mov_b32 s33, 0
; GFX803-NEXT: s_swappc_b64 s[30:31], s[4:5]
; GFX803-NEXT: s_endpgm
;
; GFX900-LABEL: test_force_fp_kern_call:
; GFX900: ; %bb.0: ; %entry
; GFX900-NEXT: s_add_u32 flat_scratch_lo, s4, s7
; GFX900-NEXT: s_addc_u32 flat_scratch_hi, s5, 0
; GFX900-NEXT: s_add_u32 s0, s0, s7
; GFX900-NEXT: s_addc_u32 s1, s1, 0
; GFX900-NEXT: s_getpc_b64 s[4:5]
; GFX900-NEXT: s_add_u32 s4, s4, ex@rel32@lo+4
; GFX900-NEXT: s_addc_u32 s5, s5, ex@rel32@hi+12
; GFX900-NEXT: s_mov_b32 s32, 0
; GFX900-NEXT: s_mov_b32 s33, 0
; GFX900-NEXT: s_swappc_b64 s[30:31], s[4:5]
; GFX900-NEXT: s_endpgm
;
; GFX1010-LABEL: test_force_fp_kern_call:
; GFX1010: ; %bb.0: ; %entry
; GFX1010-NEXT: s_add_u32 s4, s4, s7
; GFX1010-NEXT: s_mov_b32 s32, 0
; GFX1010-NEXT: s_mov_b32 s33, 0
; GFX1010-NEXT: s_addc_u32 s5, s5, 0
; GFX1010-NEXT: s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s4
; GFX1010-NEXT: s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s5
; GFX1010-NEXT: s_add_u32 s0, s0, s7
; GFX1010-NEXT: s_addc_u32 s1, s1, 0
; GFX1010-NEXT: s_getpc_b64 s[4:5]
; GFX1010-NEXT: s_add_u32 s4, s4, ex@rel32@lo+4
; GFX1010-NEXT: s_addc_u32 s5, s5, ex@rel32@hi+12
; GFX1010-NEXT: ; implicit-def: $vcc_hi
; GFX1010-NEXT: s_swappc_b64 s[30:31], s[4:5]
; GFX1010-NEXT: s_endpgm
entry:
tail call void @ex() #2
ret void
}
define amdgpu_kernel void @test_force_fp_kern_stack_and_call() local_unnamed_addr #2 {
; GFX803-LABEL: test_force_fp_kern_stack_and_call:
; GFX803: ; %bb.0: ; %entry
; GFX803-NEXT: s_add_u32 s4, s4, s7
; GFX803-NEXT: s_lshr_b32 flat_scratch_hi, s4, 8
; GFX803-NEXT: s_add_u32 s0, s0, s7
; GFX803-NEXT: s_mov_b32 s33, 0
; GFX803-NEXT: s_addc_u32 s1, s1, 0
; GFX803-NEXT: v_mov_b32_e32 v0, 0
; GFX803-NEXT: s_mov_b32 flat_scratch_lo, s5
; GFX803-NEXT: s_getpc_b64 s[4:5]
; GFX803-NEXT: s_add_u32 s4, s4, ex@rel32@lo+4
; GFX803-NEXT: s_addc_u32 s5, s5, ex@rel32@hi+12
; GFX803-NEXT: s_movk_i32 s32, 0x400
; GFX803-NEXT: buffer_store_dword v0, off, s[0:3], s33 offset:4
; GFX803-NEXT: s_swappc_b64 s[30:31], s[4:5]
; GFX803-NEXT: s_endpgm
;
; GFX900-LABEL: test_force_fp_kern_stack_and_call:
; GFX900: ; %bb.0: ; %entry
; GFX900-NEXT: s_add_u32 flat_scratch_lo, s4, s7
; GFX900-NEXT: s_addc_u32 flat_scratch_hi, s5, 0
; GFX900-NEXT: s_add_u32 s0, s0, s7
; GFX900-NEXT: s_addc_u32 s1, s1, 0
; GFX900-NEXT: s_mov_b32 s33, 0
; GFX900-NEXT: v_mov_b32_e32 v0, 0
; GFX900-NEXT: s_getpc_b64 s[4:5]
; GFX900-NEXT: s_add_u32 s4, s4, ex@rel32@lo+4
; GFX900-NEXT: s_addc_u32 s5, s5, ex@rel32@hi+12
; GFX900-NEXT: s_movk_i32 s32, 0x400
; GFX900-NEXT: buffer_store_dword v0, off, s[0:3], s33 offset:4
; GFX900-NEXT: s_swappc_b64 s[30:31], s[4:5]
; GFX900-NEXT: s_endpgm
;
; GFX1010-LABEL: test_force_fp_kern_stack_and_call:
; GFX1010: ; %bb.0: ; %entry
; GFX1010-NEXT: s_add_u32 s4, s4, s7
; GFX1010-NEXT: s_movk_i32 s32, 0x200
; GFX1010-NEXT: s_mov_b32 s33, 0
; GFX1010-NEXT: s_addc_u32 s5, s5, 0
; GFX1010-NEXT: s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s4
; GFX1010-NEXT: s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s5
; GFX1010-NEXT: v_mov_b32_e32 v0, 0
; GFX1010-NEXT: s_add_u32 s0, s0, s7
; GFX1010-NEXT: s_addc_u32 s1, s1, 0
; GFX1010-NEXT: s_getpc_b64 s[4:5]
; GFX1010-NEXT: s_add_u32 s4, s4, ex@rel32@lo+4
; GFX1010-NEXT: s_addc_u32 s5, s5, ex@rel32@hi+12
; GFX1010-NEXT: ; implicit-def: $vcc_hi
; GFX1010-NEXT: buffer_store_dword v0, off, s[0:3], s33 offset:4
; GFX1010-NEXT: s_swappc_b64 s[30:31], s[4:5]
; GFX1010-NEXT: s_endpgm
entry:
%x = alloca i32, align 4, addrspace(5)
store volatile i32 0, i32 addrspace(5)* %x, align 4
tail call void @ex() #2
ret void
}
define amdgpu_kernel void @test_sgpr_offset_kernel() #1 {
; GFX803-LABEL: test_sgpr_offset_kernel:
; GFX803: ; %bb.0: ; %entry
; GFX803-NEXT: s_add_u32 s4, s4, s7
; GFX803-NEXT: s_lshr_b32 flat_scratch_hi, s4, 8
; GFX803-NEXT: s_add_u32 s0, s0, s7
; GFX803-NEXT: s_addc_u32 s1, s1, 0
; GFX803-NEXT: buffer_load_dword v0, off, s[0:3], 0 offset:8
; GFX803-NEXT: s_mov_b32 s4, 0x40000
; GFX803-NEXT: s_mov_b32 flat_scratch_lo, s5
; GFX803-NEXT: s_waitcnt vmcnt(0)
; GFX803-NEXT: buffer_store_dword v0, off, s[0:3], s4 ; 4-byte Folded Spill
; GFX803-NEXT: ;;#ASMSTART
; GFX803-NEXT: ;;#ASMEND
; GFX803-NEXT: s_mov_b32 s4, 0x40000
; GFX803-NEXT: buffer_load_dword v0, off, s[0:3], s4 ; 4-byte Folded Reload
; GFX803-NEXT: s_waitcnt vmcnt(0)
; GFX803-NEXT: buffer_store_dword v0, off, s[0:3], 0 offset:8
; GFX803-NEXT: s_endpgm
;
; GFX900-LABEL: test_sgpr_offset_kernel:
; GFX900: ; %bb.0: ; %entry
; GFX900-NEXT: s_add_u32 flat_scratch_lo, s4, s7
; GFX900-NEXT: s_addc_u32 flat_scratch_hi, s5, 0
; GFX900-NEXT: s_add_u32 s0, s0, s7
; GFX900-NEXT: s_addc_u32 s1, s1, 0
; GFX900-NEXT: buffer_load_dword v0, off, s[0:3], 0 offset:8
; GFX900-NEXT: s_mov_b32 s6, 0x40000
; GFX900-NEXT: s_waitcnt vmcnt(0)
; GFX900-NEXT: buffer_store_dword v0, off, s[0:3], s6 ; 4-byte Folded Spill
; GFX900-NEXT: ;;#ASMSTART
; GFX900-NEXT: ;;#ASMEND
; GFX900-NEXT: s_mov_b32 s6, 0x40000
; GFX900-NEXT: buffer_load_dword v0, off, s[0:3], s6 ; 4-byte Folded Reload
; GFX900-NEXT: s_waitcnt vmcnt(0)
; GFX900-NEXT: buffer_store_dword v0, off, s[0:3], 0 offset:8
; GFX900-NEXT: s_endpgm
;
; GFX1010-LABEL: test_sgpr_offset_kernel:
; GFX1010: ; %bb.0: ; %entry
; GFX1010-NEXT: s_add_u32 s4, s4, s7
; GFX1010-NEXT: s_addc_u32 s5, s5, 0
; GFX1010-NEXT: s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s4
; GFX1010-NEXT: s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s5
; GFX1010-NEXT: s_add_u32 s0, s0, s7
; GFX1010-NEXT: s_addc_u32 s1, s1, 0
; GFX1010-NEXT: s_mov_b32 s6, 0x20000
; GFX1010-NEXT: buffer_load_dword v0, off, s[0:3], 0 offset:8
; GFX1010-NEXT: ; implicit-def: $vcc_hi
; GFX1010-NEXT: s_waitcnt vmcnt(0)
; GFX1010-NEXT: buffer_store_dword v0, off, s[0:3], s6 ; 4-byte Folded Spill
; GFX1010-NEXT: s_waitcnt_depctr 0xffe3
; GFX1010-NEXT: s_mov_b32 s6, 0x20000
; GFX1010-NEXT: ;;#ASMSTART
; GFX1010-NEXT: ;;#ASMEND
; GFX1010-NEXT: buffer_load_dword v0, off, s[0:3], s6 ; 4-byte Folded Reload
; GFX1010-NEXT: s_waitcnt vmcnt(0)
; GFX1010-NEXT: buffer_store_dword v0, off, s[0:3], 0 offset:8
; GFX1010-NEXT: s_endpgm
entry:
; Occupy 4096 bytes of scratch, so the offset of the spill of %a does not
; fit in the instruction, and has to live in the SGPR offset.
%alloca = alloca i8, i32 4092, align 4, addrspace(5)
%buf = bitcast i8 addrspace(5)* %alloca to i32 addrspace(5)*
%aptr = getelementptr i32, i32 addrspace(5)* %buf, i32 1
; 0x40000 / 64 = 4096 (for wave64)
; CHECK: s_add_u32 s6, s7, 0x40000
; CHECK: buffer_store_dword v{{[0-9]+}}, off, s[{{[0-9]+:[0-9]+}}], s6 ; 4-byte Folded Spill
%a = load volatile i32, i32 addrspace(5)* %aptr
; Force %a to spill
call void asm sideeffect "", "~{v0},~{v1},~{v2},~{v3},~{v4},~{v5},~{v6},~{v7}" ()
%outptr = getelementptr i32, i32 addrspace(5)* %buf, i32 1
store volatile i32 %a, i32 addrspace(5)* %outptr
ret void
}
declare hidden void @ex() local_unnamed_addr #0
attributes #0 = { nounwind }
attributes #1 = { nounwind "amdgpu-num-vgpr"="8" }
attributes #2 = { nounwind "frame-pointer"="all" }