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
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
; RUN: llc -mtriple=aarch64-linux-gnu -aarch64-streaming-hazard-size=0 -mattr=+sve,+sme2 < %s | FileCheck %s
declare void @callee()
declare void @callee_farg(float)
declare float @callee_farg_fret(float)
; normal caller -> streaming callees
define void @test0() nounwind {
; CHECK-LABEL: test0:
; CHECK: // %bb.0:
; CHECK-NEXT: stp d15, d14, [sp, #-80]! // 16-byte Folded Spill
; CHECK-NEXT: cntd x9
; CHECK-NEXT: stp d13, d12, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: stp d11, d10, [sp, #32] // 16-byte Folded Spill
; CHECK-NEXT: stp d9, d8, [sp, #48] // 16-byte Folded Spill
; CHECK-NEXT: stp x30, x9, [sp, #64] // 16-byte Folded Spill
; CHECK-NEXT: smstart sm
; CHECK-NEXT: bl callee
; CHECK-NEXT: bl callee
; CHECK-NEXT: smstop sm
; CHECK-NEXT: ldp d9, d8, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: ldr x30, [sp, #64] // 8-byte Folded Reload
; CHECK-NEXT: ldp d11, d10, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: ldp d13, d12, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: ldp d15, d14, [sp], #80 // 16-byte Folded Reload
; CHECK-NEXT: ret
call void @callee() "aarch64_pstate_sm_enabled"
call void @callee() "aarch64_pstate_sm_enabled"
ret void
}
; streaming caller -> normal callees
define void @test1() nounwind "aarch64_pstate_sm_enabled" {
; CHECK-LABEL: test1:
; CHECK: // %bb.0:
; CHECK-NEXT: stp d15, d14, [sp, #-80]! // 16-byte Folded Spill
; CHECK-NEXT: cntd x9
; CHECK-NEXT: stp d13, d12, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: stp d11, d10, [sp, #32] // 16-byte Folded Spill
; CHECK-NEXT: stp d9, d8, [sp, #48] // 16-byte Folded Spill
; CHECK-NEXT: stp x30, x9, [sp, #64] // 16-byte Folded Spill
; CHECK-NEXT: smstop sm
; CHECK-NEXT: bl callee
; CHECK-NEXT: bl callee
; CHECK-NEXT: smstart sm
; CHECK-NEXT: ldp d9, d8, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: ldr x30, [sp, #64] // 8-byte Folded Reload
; CHECK-NEXT: ldp d11, d10, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: ldp d13, d12, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: ldp d15, d14, [sp], #80 // 16-byte Folded Reload
; CHECK-NEXT: ret
call void @callee()
call void @callee()
ret void
}
; streaming-compatible caller -> normal callees
; these conditional smstart/smstop are not yet optimized away.
define void @test2() nounwind "aarch64_pstate_sm_compatible" {
; CHECK-LABEL: test2:
; CHECK: // %bb.0:
; CHECK-NEXT: stp d15, d14, [sp, #-96]! // 16-byte Folded Spill
; CHECK-NEXT: cntd x9
; CHECK-NEXT: stp d13, d12, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: stp d11, d10, [sp, #32] // 16-byte Folded Spill
; CHECK-NEXT: stp d9, d8, [sp, #48] // 16-byte Folded Spill
; CHECK-NEXT: stp x30, x9, [sp, #64] // 16-byte Folded Spill
; CHECK-NEXT: str x19, [sp, #80] // 8-byte Folded Spill
; CHECK-NEXT: bl __arm_sme_state
; CHECK-NEXT: and x19, x0, #0x1
; CHECK-NEXT: tbz w19, #0, .LBB2_2
; CHECK-NEXT: // %bb.1:
; CHECK-NEXT: smstop sm
; CHECK-NEXT: .LBB2_2:
; CHECK-NEXT: bl callee
; CHECK-NEXT: tbz w19, #0, .LBB2_4
; CHECK-NEXT: // %bb.3:
; CHECK-NEXT: smstart sm
; CHECK-NEXT: .LBB2_4:
; CHECK-NEXT: bl __arm_sme_state
; CHECK-NEXT: and x19, x0, #0x1
; CHECK-NEXT: tbz w19, #0, .LBB2_6
; CHECK-NEXT: // %bb.5:
; CHECK-NEXT: smstop sm
; CHECK-NEXT: .LBB2_6:
; CHECK-NEXT: bl callee
; CHECK-NEXT: tbz w19, #0, .LBB2_8
; CHECK-NEXT: // %bb.7:
; CHECK-NEXT: smstart sm
; CHECK-NEXT: .LBB2_8:
; CHECK-NEXT: ldp d9, d8, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: ldr x19, [sp, #80] // 8-byte Folded Reload
; CHECK-NEXT: ldp d11, d10, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: ldr x30, [sp, #64] // 8-byte Folded Reload
; CHECK-NEXT: ldp d13, d12, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: ldp d15, d14, [sp], #96 // 16-byte Folded Reload
; CHECK-NEXT: ret
call void @callee()
call void @callee()
ret void
}
; streaming-compatible caller -> mixed callees
define void @test3() nounwind "aarch64_pstate_sm_compatible" {
; CHECK-LABEL: test3:
; CHECK: // %bb.0:
; CHECK-NEXT: stp d15, d14, [sp, #-96]! // 16-byte Folded Spill
; CHECK-NEXT: cntd x9
; CHECK-NEXT: stp d13, d12, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: stp d11, d10, [sp, #32] // 16-byte Folded Spill
; CHECK-NEXT: stp d9, d8, [sp, #48] // 16-byte Folded Spill
; CHECK-NEXT: stp x30, x9, [sp, #64] // 16-byte Folded Spill
; CHECK-NEXT: str x19, [sp, #80] // 8-byte Folded Spill
; CHECK-NEXT: bl __arm_sme_state
; CHECK-NEXT: and x19, x0, #0x1
; CHECK-NEXT: tbnz w19, #0, .LBB3_2
; CHECK-NEXT: // %bb.1:
; CHECK-NEXT: smstart sm
; CHECK-NEXT: .LBB3_2:
; CHECK-NEXT: bl callee
; CHECK-NEXT: tbnz w19, #0, .LBB3_4
; CHECK-NEXT: // %bb.3:
; CHECK-NEXT: smstop sm
; CHECK-NEXT: .LBB3_4:
; CHECK-NEXT: bl __arm_sme_state
; CHECK-NEXT: and x19, x0, #0x1
; CHECK-NEXT: tbz w19, #0, .LBB3_6
; CHECK-NEXT: // %bb.5:
; CHECK-NEXT: smstop sm
; CHECK-NEXT: .LBB3_6:
; CHECK-NEXT: bl callee
; CHECK-NEXT: tbz w19, #0, .LBB3_8
; CHECK-NEXT: // %bb.7:
; CHECK-NEXT: smstart sm
; CHECK-NEXT: .LBB3_8:
; CHECK-NEXT: bl __arm_sme_state
; CHECK-NEXT: and x19, x0, #0x1
; CHECK-NEXT: tbnz w19, #0, .LBB3_10
; CHECK-NEXT: // %bb.9:
; CHECK-NEXT: smstart sm
; CHECK-NEXT: .LBB3_10:
; CHECK-NEXT: bl callee
; CHECK-NEXT: tbnz w19, #0, .LBB3_12
; CHECK-NEXT: // %bb.11:
; CHECK-NEXT: smstop sm
; CHECK-NEXT: .LBB3_12:
; CHECK-NEXT: ldp d9, d8, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: ldr x19, [sp, #80] // 8-byte Folded Reload
; CHECK-NEXT: ldp d11, d10, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: ldr x30, [sp, #64] // 8-byte Folded Reload
; CHECK-NEXT: ldp d13, d12, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: ldp d15, d14, [sp], #96 // 16-byte Folded Reload
; CHECK-NEXT: ret
call void @callee() "aarch64_pstate_sm_enabled"
call void @callee()
call void @callee() "aarch64_pstate_sm_enabled"
ret void
}
; streaming caller -> normal callees (pass 0.0f)
define void @test4() nounwind "aarch64_pstate_sm_enabled" {
; CHECK-LABEL: test4:
; CHECK: // %bb.0:
; CHECK-NEXT: stp d15, d14, [sp, #-80]! // 16-byte Folded Spill
; CHECK-NEXT: cntd x9
; CHECK-NEXT: stp d13, d12, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: stp d11, d10, [sp, #32] // 16-byte Folded Spill
; CHECK-NEXT: stp d9, d8, [sp, #48] // 16-byte Folded Spill
; CHECK-NEXT: stp x30, x9, [sp, #64] // 16-byte Folded Spill
; CHECK-NEXT: smstop sm
; CHECK-NEXT: fmov s0, wzr
; CHECK-NEXT: bl callee_farg
; CHECK-NEXT: fmov s0, wzr
; CHECK-NEXT: bl callee_farg
; CHECK-NEXT: smstart sm
; CHECK-NEXT: ldp d9, d8, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: ldr x30, [sp, #64] // 8-byte Folded Reload
; CHECK-NEXT: ldp d11, d10, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: ldp d13, d12, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: ldp d15, d14, [sp], #80 // 16-byte Folded Reload
; CHECK-NEXT: ret
call void @callee_farg(float zeroinitializer)
call void @callee_farg(float zeroinitializer)
ret void
}
; streaming caller -> normal callees (pass fp arg)
define void @test5(float %f) nounwind "aarch64_pstate_sm_enabled" {
; CHECK-LABEL: test5:
; CHECK: // %bb.0:
; CHECK-NEXT: sub sp, sp, #96
; CHECK-NEXT: cntd x9
; CHECK-NEXT: stp d15, d14, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: stp d13, d12, [sp, #32] // 16-byte Folded Spill
; CHECK-NEXT: stp d11, d10, [sp, #48] // 16-byte Folded Spill
; CHECK-NEXT: stp d9, d8, [sp, #64] // 16-byte Folded Spill
; CHECK-NEXT: stp x30, x9, [sp, #80] // 16-byte Folded Spill
; CHECK-NEXT: str s0, [sp, #12] // 4-byte Folded Spill
; CHECK-NEXT: smstop sm
; CHECK-NEXT: ldr s0, [sp, #12] // 4-byte Folded Reload
; CHECK-NEXT: bl callee_farg
; CHECK-NEXT: ldr s0, [sp, #12] // 4-byte Folded Reload
; CHECK-NEXT: bl callee_farg
; CHECK-NEXT: smstart sm
; CHECK-NEXT: ldp d9, d8, [sp, #64] // 16-byte Folded Reload
; CHECK-NEXT: ldr x30, [sp, #80] // 8-byte Folded Reload
; CHECK-NEXT: ldp d11, d10, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: ldp d13, d12, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: ldp d15, d14, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: add sp, sp, #96
; CHECK-NEXT: ret
call void @callee_farg(float %f)
call void @callee_farg(float %f)
ret void
}
define float @test6(float %f) nounwind "aarch64_pstate_sm_enabled" {
; CHECK-LABEL: test6:
; CHECK: // %bb.0:
; CHECK-NEXT: sub sp, sp, #96
; CHECK-NEXT: cntd x9
; CHECK-NEXT: stp d15, d14, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: stp d13, d12, [sp, #32] // 16-byte Folded Spill
; CHECK-NEXT: stp d11, d10, [sp, #48] // 16-byte Folded Spill
; CHECK-NEXT: stp d9, d8, [sp, #64] // 16-byte Folded Spill
; CHECK-NEXT: stp x30, x9, [sp, #80] // 16-byte Folded Spill
; CHECK-NEXT: str s0, [sp, #12] // 4-byte Folded Spill
; CHECK-NEXT: smstop sm
; CHECK-NEXT: ldr s0, [sp, #12] // 4-byte Folded Reload
; CHECK-NEXT: bl callee_farg_fret
; CHECK-NEXT: bl callee_farg_fret
; CHECK-NEXT: str s0, [sp, #12] // 4-byte Folded Spill
; CHECK-NEXT: smstart sm
; CHECK-NEXT: ldr s0, [sp, #12] // 4-byte Folded Reload
; CHECK-NEXT: ldp d9, d8, [sp, #64] // 16-byte Folded Reload
; CHECK-NEXT: ldr x30, [sp, #80] // 8-byte Folded Reload
; CHECK-NEXT: ldp d11, d10, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: ldp d13, d12, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: ldp d15, d14, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: add sp, sp, #96
; CHECK-NEXT: ret
%res0 = call float @callee_farg_fret(float %f)
%res1 = call float @callee_farg_fret(float %res0)
ret float %res1
}
; save/restore zt0 to stack is not yet optimised away by the pass,
; because of the ldr/str of zt0, which will need some further analysis
; to make sure if the redundant str can be removed.
define void @test7() nounwind "aarch64_inout_zt0" {
; CHECK-LABEL: test7:
; CHECK: // %bb.0:
; CHECK-NEXT: sub sp, sp, #144
; CHECK-NEXT: stp x30, x19, [sp, #128] // 16-byte Folded Spill
; CHECK-NEXT: add x19, sp, #64
; CHECK-NEXT: str zt0, [x19]
; CHECK-NEXT: smstop za
; CHECK-NEXT: bl callee
; CHECK-NEXT: smstart za
; CHECK-NEXT: ldr zt0, [x19]
; CHECK-NEXT: mov x19, sp
; CHECK-NEXT: str zt0, [x19]
; CHECK-NEXT: smstop za
; CHECK-NEXT: bl callee
; CHECK-NEXT: smstart za
; CHECK-NEXT: ldr zt0, [x19]
; CHECK-NEXT: ldp x30, x19, [sp, #128] // 16-byte Folded Reload
; CHECK-NEXT: add sp, sp, #144
; CHECK-NEXT: ret
call void @callee()
call void @callee()
ret void
}
; test that 'smstop za' is not cancelled out with 'smstart sm'.
define void @test8() nounwind "aarch64_pstate_sm_enabled" {
; CHECK-LABEL: test8:
; CHECK: // %bb.0:
; CHECK-NEXT: stp d15, d14, [sp, #-80]! // 16-byte Folded Spill
; CHECK-NEXT: cntd x9
; CHECK-NEXT: stp d13, d12, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: stp d11, d10, [sp, #32] // 16-byte Folded Spill
; CHECK-NEXT: stp d9, d8, [sp, #48] // 16-byte Folded Spill
; CHECK-NEXT: stp x30, x9, [sp, #64] // 16-byte Folded Spill
; CHECK-NEXT: smstop sm
; CHECK-NEXT: bl callee
; CHECK-NEXT: smstart sm
; CHECK-NEXT: smstop za
; CHECK-NEXT: ldp d9, d8, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: ldr x30, [sp, #64] // 8-byte Folded Reload
; CHECK-NEXT: ldp d11, d10, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: ldp d13, d12, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: ldp d15, d14, [sp], #80 // 16-byte Folded Reload
; CHECK-NEXT: ret
call void @callee()
call void @llvm.aarch64.sme.za.disable()
ret void
}
; test that the 'smstart' and 'smstop' are entirely removed,
; along with any code to read 'vg' for the CFI.
define void @test9() "aarch64_pstate_sm_body" {
; CHECK-LABEL: test9:
; CHECK: // %bb.0:
; CHECK-NEXT: str x30, [sp, #-16]! // 8-byte Folded Spill
; CHECK-NEXT: .cfi_def_cfa_offset 16
; CHECK-NEXT: .cfi_offset w30, -16
; CHECK-NEXT: bl callee
; CHECK-NEXT: ldr x30, [sp], #16 // 8-byte Folded Reload
; CHECK-NEXT: ret
call void @callee()
ret void
}
; similar to above, but in this case only the first
; 'smstart, smstop' pair can be removed and the code required
; for the CFI is still needed.
define void @test10() "aarch64_pstate_sm_body" {
; CHECK-LABEL: test10:
; CHECK: // %bb.0:
; CHECK-NEXT: stp d15, d14, [sp, #-96]! // 16-byte Folded Spill
; CHECK-NEXT: .cfi_def_cfa_offset 96
; CHECK-NEXT: rdsvl x9, #1
; CHECK-NEXT: stp d13, d12, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: lsr x9, x9, #3
; CHECK-NEXT: stp d11, d10, [sp, #32] // 16-byte Folded Spill
; CHECK-NEXT: stp d9, d8, [sp, #48] // 16-byte Folded Spill
; CHECK-NEXT: stp x30, x9, [sp, #64] // 16-byte Folded Spill
; CHECK-NEXT: cntd x9
; CHECK-NEXT: str x9, [sp, #80] // 8-byte Folded Spill
; CHECK-NEXT: .cfi_offset vg, -16
; CHECK-NEXT: .cfi_offset w30, -32
; CHECK-NEXT: .cfi_offset b8, -40
; CHECK-NEXT: .cfi_offset b9, -48
; CHECK-NEXT: .cfi_offset b10, -56
; CHECK-NEXT: .cfi_offset b11, -64
; CHECK-NEXT: .cfi_offset b12, -72
; CHECK-NEXT: .cfi_offset b13, -80
; CHECK-NEXT: .cfi_offset b14, -88
; CHECK-NEXT: .cfi_offset b15, -96
; CHECK-NEXT: bl callee
; CHECK-NEXT: smstart sm
; CHECK-NEXT: .cfi_restore vg
; CHECK-NEXT: bl callee
; CHECK-NEXT: .cfi_offset vg, -24
; CHECK-NEXT: smstop sm
; CHECK-NEXT: bl callee
; CHECK-NEXT: ldp d9, d8, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: ldr x30, [sp, #64] // 8-byte Folded Reload
; CHECK-NEXT: ldp d11, d10, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: ldp d13, d12, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: ldp d15, d14, [sp], #96 // 16-byte Folded Reload
; CHECK-NEXT: .cfi_def_cfa_offset 0
; CHECK-NEXT: .cfi_restore w30
; CHECK-NEXT: .cfi_restore b8
; CHECK-NEXT: .cfi_restore b9
; CHECK-NEXT: .cfi_restore b10
; CHECK-NEXT: .cfi_restore b11
; CHECK-NEXT: .cfi_restore b12
; CHECK-NEXT: .cfi_restore b13
; CHECK-NEXT: .cfi_restore b14
; CHECK-NEXT: .cfi_restore b15
; CHECK-NEXT: ret
call void @callee()
call void @callee() "aarch64_pstate_sm_enabled"
call void @callee()
ret void
}
; test that an operation like a store is executed in the right
; streaming mode and blocks the optimization.
define void @test11(ptr %p) nounwind "aarch64_pstate_sm_enabled" {
; CHECK-LABEL: test11:
; CHECK: // %bb.0:
; CHECK-NEXT: stp d15, d14, [sp, #-96]! // 16-byte Folded Spill
; CHECK-NEXT: cntd x9
; CHECK-NEXT: stp d13, d12, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: stp d11, d10, [sp, #32] // 16-byte Folded Spill
; CHECK-NEXT: stp d9, d8, [sp, #48] // 16-byte Folded Spill
; CHECK-NEXT: stp x30, x9, [sp, #64] // 16-byte Folded Spill
; CHECK-NEXT: str x19, [sp, #80] // 8-byte Folded Spill
; CHECK-NEXT: mov x19, x0
; CHECK-NEXT: smstop sm
; CHECK-NEXT: bl callee
; CHECK-NEXT: smstart sm
; CHECK-NEXT: mov z0.b, #0 // =0x0
; CHECK-NEXT: ptrue p0.b
; CHECK-NEXT: st1b { z0.b }, p0, [x19]
; CHECK-NEXT: smstop sm
; CHECK-NEXT: bl callee
; CHECK-NEXT: smstart sm
; CHECK-NEXT: ldp d9, d8, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: ldr x19, [sp, #80] // 8-byte Folded Reload
; CHECK-NEXT: ldp d11, d10, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: ldr x30, [sp, #64] // 8-byte Folded Reload
; CHECK-NEXT: ldp d13, d12, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: ldp d15, d14, [sp], #96 // 16-byte Folded Reload
; CHECK-NEXT: ret
call void @callee()
store <vscale x 16 x i8> zeroinitializer, ptr %p
call void @callee()
ret void
}
; test that 'smstart sm' and 'smstop za' don't get folded away together.
; we can further optimize this test by considering streaming mode
; separately from ZA.
define void @test12() "aarch64_pstate_sm_body" {
; CHECK-LABEL: test12:
; CHECK: // %bb.0:
; CHECK-NEXT: stp d15, d14, [sp, #-96]! // 16-byte Folded Spill
; CHECK-NEXT: .cfi_def_cfa_offset 96
; CHECK-NEXT: rdsvl x9, #1
; CHECK-NEXT: stp d13, d12, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: lsr x9, x9, #3
; CHECK-NEXT: stp d11, d10, [sp, #32] // 16-byte Folded Spill
; CHECK-NEXT: stp d9, d8, [sp, #48] // 16-byte Folded Spill
; CHECK-NEXT: stp x30, x9, [sp, #64] // 16-byte Folded Spill
; CHECK-NEXT: cntd x9
; CHECK-NEXT: str x9, [sp, #80] // 8-byte Folded Spill
; CHECK-NEXT: .cfi_offset vg, -16
; CHECK-NEXT: .cfi_offset w30, -32
; CHECK-NEXT: .cfi_offset b8, -40
; CHECK-NEXT: .cfi_offset b9, -48
; CHECK-NEXT: .cfi_offset b10, -56
; CHECK-NEXT: .cfi_offset b11, -64
; CHECK-NEXT: .cfi_offset b12, -72
; CHECK-NEXT: .cfi_offset b13, -80
; CHECK-NEXT: .cfi_offset b14, -88
; CHECK-NEXT: .cfi_offset b15, -96
; CHECK-NEXT: smstart sm
; CHECK-NEXT: smstop za
; CHECK-NEXT: .cfi_offset vg, -24
; CHECK-NEXT: smstop sm
; CHECK-NEXT: bl callee
; CHECK-NEXT: smstart sm
; CHECK-NEXT: .cfi_restore vg
; CHECK-NEXT: smstart za
; CHECK-NEXT: smstop sm
; CHECK-NEXT: ldp d9, d8, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: ldr x30, [sp, #64] // 8-byte Folded Reload
; CHECK-NEXT: ldp d11, d10, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: ldp d13, d12, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: ldp d15, d14, [sp], #96 // 16-byte Folded Reload
; CHECK-NEXT: .cfi_def_cfa_offset 0
; CHECK-NEXT: .cfi_restore w30
; CHECK-NEXT: .cfi_restore b8
; CHECK-NEXT: .cfi_restore b9
; CHECK-NEXT: .cfi_restore b10
; CHECK-NEXT: .cfi_restore b11
; CHECK-NEXT: .cfi_restore b12
; CHECK-NEXT: .cfi_restore b13
; CHECK-NEXT: .cfi_restore b14
; CHECK-NEXT: .cfi_restore b15
; CHECK-NEXT: ret
call void @llvm.aarch64.sme.za.disable()
call void @callee()
call void @llvm.aarch64.sme.za.enable()
ret void
}
; We conservatively don't remove the smstart/smstop pair yet when there are COPY
; instructions that copy SVE registers, because we can't yet conclusively prove
; that this is safe (although for this example, it would be).
define void @test13(ptr %ptr) nounwind "aarch64_pstate_sm_enabled" {
; CHECK-LABEL: test13:
; CHECK: // %bb.0:
; CHECK-NEXT: stp d15, d14, [sp, #-96]! // 16-byte Folded Spill
; CHECK-NEXT: cntd x9
; CHECK-NEXT: stp d13, d12, [sp, #16] // 16-byte Folded Spill
; CHECK-NEXT: stp d11, d10, [sp, #32] // 16-byte Folded Spill
; CHECK-NEXT: stp d9, d8, [sp, #48] // 16-byte Folded Spill
; CHECK-NEXT: stp x29, x30, [sp, #64] // 16-byte Folded Spill
; CHECK-NEXT: stp x9, x19, [sp, #80] // 16-byte Folded Spill
; CHECK-NEXT: addvl sp, sp, #-1
; CHECK-NEXT: mov z0.s, #0 // =0x0
; CHECK-NEXT: mov x19, x0
; CHECK-NEXT: str z0, [sp] // 16-byte Folded Spill
; CHECK-NEXT: smstop sm
; CHECK-NEXT: ldr z0, [sp] // 16-byte Folded Reload
; CHECK-NEXT: bl callee_farg_fret
; CHECK-NEXT: str z0, [sp] // 16-byte Folded Spill
; CHECK-NEXT: smstart sm
; CHECK-NEXT: smstop sm
; CHECK-NEXT: ldr z0, [sp] // 16-byte Folded Reload
; CHECK-NEXT: bl callee_farg_fret
; CHECK-NEXT: str z0, [sp] // 16-byte Folded Spill
; CHECK-NEXT: smstart sm
; CHECK-NEXT: ptrue p0.s
; CHECK-NEXT: ldr z0, [sp] // 16-byte Folded Reload
; CHECK-NEXT: st1w { z0.s }, p0, [x19]
; CHECK-NEXT: addvl sp, sp, #1
; CHECK-NEXT: ldp x29, x30, [sp, #64] // 16-byte Folded Reload
; CHECK-NEXT: ldr x19, [sp, #88] // 8-byte Folded Reload
; CHECK-NEXT: ldp d9, d8, [sp, #48] // 16-byte Folded Reload
; CHECK-NEXT: ldp d11, d10, [sp, #32] // 16-byte Folded Reload
; CHECK-NEXT: ldp d13, d12, [sp, #16] // 16-byte Folded Reload
; CHECK-NEXT: ldp d15, d14, [sp], #96 // 16-byte Folded Reload
; CHECK-NEXT: ret
%res0 = call <vscale x 4 x float> @callee_farg_fret(<vscale x 4 x float> zeroinitializer)
%res1 = call <vscale x 4 x float> @callee_farg_fret(<vscale x 4 x float> %res0)
store <vscale x 4 x float> %res1, ptr %ptr
ret void
}
|
