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
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
|
/*
* SPDX-License-Identifier: MIT
*
* Copyright © 2017-2018 Intel Corporation
*/
#include <linux/prime_numbers.h>
#include "gem/i915_gem_pm.h"
#include "intel_gt.h"
#include "../selftests/i915_random.h"
#include "../i915_selftest.h"
#include "../selftests/igt_flush_test.h"
#include "../selftests/mock_gem_device.h"
#include "selftests/mock_timeline.h"
static struct page *hwsp_page(struct intel_timeline *tl)
{
struct drm_i915_gem_object *obj = tl->hwsp_ggtt->obj;
GEM_BUG_ON(!i915_gem_object_has_pinned_pages(obj));
return sg_page(obj->mm.pages->sgl);
}
static unsigned long hwsp_cacheline(struct intel_timeline *tl)
{
unsigned long address = (unsigned long)page_address(hwsp_page(tl));
return (address + tl->hwsp_offset) / CACHELINE_BYTES;
}
#define CACHELINES_PER_PAGE (PAGE_SIZE / CACHELINE_BYTES)
struct mock_hwsp_freelist {
struct drm_i915_private *i915;
struct radix_tree_root cachelines;
struct intel_timeline **history;
unsigned long count, max;
struct rnd_state prng;
};
enum {
SHUFFLE = BIT(0),
};
static void __mock_hwsp_record(struct mock_hwsp_freelist *state,
unsigned int idx,
struct intel_timeline *tl)
{
tl = xchg(&state->history[idx], tl);
if (tl) {
radix_tree_delete(&state->cachelines, hwsp_cacheline(tl));
intel_timeline_put(tl);
}
}
static int __mock_hwsp_timeline(struct mock_hwsp_freelist *state,
unsigned int count,
unsigned int flags)
{
struct intel_timeline *tl;
unsigned int idx;
while (count--) {
unsigned long cacheline;
int err;
tl = intel_timeline_create(&state->i915->gt, NULL);
if (IS_ERR(tl))
return PTR_ERR(tl);
cacheline = hwsp_cacheline(tl);
err = radix_tree_insert(&state->cachelines, cacheline, tl);
if (err) {
if (err == -EEXIST) {
pr_err("HWSP cacheline %lu already used; duplicate allocation!\n",
cacheline);
}
intel_timeline_put(tl);
return err;
}
idx = state->count++ % state->max;
__mock_hwsp_record(state, idx, tl);
}
if (flags & SHUFFLE)
i915_prandom_shuffle(state->history,
sizeof(*state->history),
min(state->count, state->max),
&state->prng);
count = i915_prandom_u32_max_state(min(state->count, state->max),
&state->prng);
while (count--) {
idx = --state->count % state->max;
__mock_hwsp_record(state, idx, NULL);
}
return 0;
}
static int mock_hwsp_freelist(void *arg)
{
struct mock_hwsp_freelist state;
const struct {
const char *name;
unsigned int flags;
} phases[] = {
{ "linear", 0 },
{ "shuffled", SHUFFLE },
{ },
}, *p;
unsigned int na;
int err = 0;
INIT_RADIX_TREE(&state.cachelines, GFP_KERNEL);
state.prng = I915_RND_STATE_INITIALIZER(i915_selftest.random_seed);
state.i915 = mock_gem_device();
if (!state.i915)
return -ENOMEM;
/*
* Create a bunch of timelines and check that their HWSP do not overlap.
* Free some, and try again.
*/
state.max = PAGE_SIZE / sizeof(*state.history);
state.count = 0;
state.history = kcalloc(state.max, sizeof(*state.history), GFP_KERNEL);
if (!state.history) {
err = -ENOMEM;
goto err_put;
}
mutex_lock(&state.i915->drm.struct_mutex);
for (p = phases; p->name; p++) {
pr_debug("%s(%s)\n", __func__, p->name);
for_each_prime_number_from(na, 1, 2 * CACHELINES_PER_PAGE) {
err = __mock_hwsp_timeline(&state, na, p->flags);
if (err)
goto out;
}
}
out:
for (na = 0; na < state.max; na++)
__mock_hwsp_record(&state, na, NULL);
mutex_unlock(&state.i915->drm.struct_mutex);
kfree(state.history);
err_put:
drm_dev_put(&state.i915->drm);
return err;
}
struct __igt_sync {
const char *name;
u32 seqno;
bool expected;
bool set;
};
static int __igt_sync(struct intel_timeline *tl,
u64 ctx,
const struct __igt_sync *p,
const char *name)
{
int ret;
if (__intel_timeline_sync_is_later(tl, ctx, p->seqno) != p->expected) {
pr_err("%s: %s(ctx=%llu, seqno=%u) expected passed %s but failed\n",
name, p->name, ctx, p->seqno, yesno(p->expected));
return -EINVAL;
}
if (p->set) {
ret = __intel_timeline_sync_set(tl, ctx, p->seqno);
if (ret)
return ret;
}
return 0;
}
static int igt_sync(void *arg)
{
const struct __igt_sync pass[] = {
{ "unset", 0, false, false },
{ "new", 0, false, true },
{ "0a", 0, true, true },
{ "1a", 1, false, true },
{ "1b", 1, true, true },
{ "0b", 0, true, false },
{ "2a", 2, false, true },
{ "4", 4, false, true },
{ "INT_MAX", INT_MAX, false, true },
{ "INT_MAX-1", INT_MAX-1, true, false },
{ "INT_MAX+1", (u32)INT_MAX+1, false, true },
{ "INT_MAX", INT_MAX, true, false },
{ "UINT_MAX", UINT_MAX, false, true },
{ "wrap", 0, false, true },
{ "unwrap", UINT_MAX, true, false },
{},
}, *p;
struct intel_timeline tl;
int order, offset;
int ret = -ENODEV;
mock_timeline_init(&tl, 0);
for (p = pass; p->name; p++) {
for (order = 1; order < 64; order++) {
for (offset = -1; offset <= (order > 1); offset++) {
u64 ctx = BIT_ULL(order) + offset;
ret = __igt_sync(&tl, ctx, p, "1");
if (ret)
goto out;
}
}
}
mock_timeline_fini(&tl);
mock_timeline_init(&tl, 0);
for (order = 1; order < 64; order++) {
for (offset = -1; offset <= (order > 1); offset++) {
u64 ctx = BIT_ULL(order) + offset;
for (p = pass; p->name; p++) {
ret = __igt_sync(&tl, ctx, p, "2");
if (ret)
goto out;
}
}
}
out:
mock_timeline_fini(&tl);
return ret;
}
static unsigned int random_engine(struct rnd_state *rnd)
{
return i915_prandom_u32_max_state(I915_NUM_ENGINES, rnd);
}
static int bench_sync(void *arg)
{
struct rnd_state prng;
struct intel_timeline tl;
unsigned long end_time, count;
u64 prng32_1M;
ktime_t kt;
int order, last_order;
mock_timeline_init(&tl, 0);
/* Lookups from cache are very fast and so the random number generation
* and the loop itself becomes a significant factor in the per-iteration
* timings. We try to compensate the results by measuring the overhead
* of the prng and subtract it from the reported results.
*/
prandom_seed_state(&prng, i915_selftest.random_seed);
count = 0;
kt = ktime_get();
end_time = jiffies + HZ/10;
do {
u32 x;
/* Make sure the compiler doesn't optimise away the prng call */
WRITE_ONCE(x, prandom_u32_state(&prng));
count++;
} while (!time_after(jiffies, end_time));
kt = ktime_sub(ktime_get(), kt);
pr_debug("%s: %lu random evaluations, %lluns/prng\n",
__func__, count, (long long)div64_ul(ktime_to_ns(kt), count));
prng32_1M = div64_ul(ktime_to_ns(kt) << 20, count);
/* Benchmark (only) setting random context ids */
prandom_seed_state(&prng, i915_selftest.random_seed);
count = 0;
kt = ktime_get();
end_time = jiffies + HZ/10;
do {
u64 id = i915_prandom_u64_state(&prng);
__intel_timeline_sync_set(&tl, id, 0);
count++;
} while (!time_after(jiffies, end_time));
kt = ktime_sub(ktime_get(), kt);
kt = ktime_sub_ns(kt, (count * prng32_1M * 2) >> 20);
pr_info("%s: %lu random insertions, %lluns/insert\n",
__func__, count, (long long)div64_ul(ktime_to_ns(kt), count));
/* Benchmark looking up the exact same context ids as we just set */
prandom_seed_state(&prng, i915_selftest.random_seed);
end_time = count;
kt = ktime_get();
while (end_time--) {
u64 id = i915_prandom_u64_state(&prng);
if (!__intel_timeline_sync_is_later(&tl, id, 0)) {
mock_timeline_fini(&tl);
pr_err("Lookup of %llu failed\n", id);
return -EINVAL;
}
}
kt = ktime_sub(ktime_get(), kt);
kt = ktime_sub_ns(kt, (count * prng32_1M * 2) >> 20);
pr_info("%s: %lu random lookups, %lluns/lookup\n",
__func__, count, (long long)div64_ul(ktime_to_ns(kt), count));
mock_timeline_fini(&tl);
cond_resched();
mock_timeline_init(&tl, 0);
/* Benchmark setting the first N (in order) contexts */
count = 0;
kt = ktime_get();
end_time = jiffies + HZ/10;
do {
__intel_timeline_sync_set(&tl, count++, 0);
} while (!time_after(jiffies, end_time));
kt = ktime_sub(ktime_get(), kt);
pr_info("%s: %lu in-order insertions, %lluns/insert\n",
__func__, count, (long long)div64_ul(ktime_to_ns(kt), count));
/* Benchmark looking up the exact same context ids as we just set */
end_time = count;
kt = ktime_get();
while (end_time--) {
if (!__intel_timeline_sync_is_later(&tl, end_time, 0)) {
pr_err("Lookup of %lu failed\n", end_time);
mock_timeline_fini(&tl);
return -EINVAL;
}
}
kt = ktime_sub(ktime_get(), kt);
pr_info("%s: %lu in-order lookups, %lluns/lookup\n",
__func__, count, (long long)div64_ul(ktime_to_ns(kt), count));
mock_timeline_fini(&tl);
cond_resched();
mock_timeline_init(&tl, 0);
/* Benchmark searching for a random context id and maybe changing it */
prandom_seed_state(&prng, i915_selftest.random_seed);
count = 0;
kt = ktime_get();
end_time = jiffies + HZ/10;
do {
u32 id = random_engine(&prng);
u32 seqno = prandom_u32_state(&prng);
if (!__intel_timeline_sync_is_later(&tl, id, seqno))
__intel_timeline_sync_set(&tl, id, seqno);
count++;
} while (!time_after(jiffies, end_time));
kt = ktime_sub(ktime_get(), kt);
kt = ktime_sub_ns(kt, (count * prng32_1M * 2) >> 20);
pr_info("%s: %lu repeated insert/lookups, %lluns/op\n",
__func__, count, (long long)div64_ul(ktime_to_ns(kt), count));
mock_timeline_fini(&tl);
cond_resched();
/* Benchmark searching for a known context id and changing the seqno */
for (last_order = 1, order = 1; order < 32;
({ int tmp = last_order; last_order = order; order += tmp; })) {
unsigned int mask = BIT(order) - 1;
mock_timeline_init(&tl, 0);
count = 0;
kt = ktime_get();
end_time = jiffies + HZ/10;
do {
/* Without assuming too many details of the underlying
* implementation, try to identify its phase-changes
* (if any)!
*/
u64 id = (u64)(count & mask) << order;
__intel_timeline_sync_is_later(&tl, id, 0);
__intel_timeline_sync_set(&tl, id, 0);
count++;
} while (!time_after(jiffies, end_time));
kt = ktime_sub(ktime_get(), kt);
pr_info("%s: %lu cyclic/%d insert/lookups, %lluns/op\n",
__func__, count, order,
(long long)div64_ul(ktime_to_ns(kt), count));
mock_timeline_fini(&tl);
cond_resched();
}
return 0;
}
int intel_timeline_mock_selftests(void)
{
static const struct i915_subtest tests[] = {
SUBTEST(mock_hwsp_freelist),
SUBTEST(igt_sync),
SUBTEST(bench_sync),
};
return i915_subtests(tests, NULL);
}
static int emit_ggtt_store_dw(struct i915_request *rq, u32 addr, u32 value)
{
u32 *cs;
cs = intel_ring_begin(rq, 4);
if (IS_ERR(cs))
return PTR_ERR(cs);
if (INTEL_GEN(rq->i915) >= 8) {
*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
*cs++ = addr;
*cs++ = 0;
*cs++ = value;
} else if (INTEL_GEN(rq->i915) >= 4) {
*cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT;
*cs++ = 0;
*cs++ = addr;
*cs++ = value;
} else {
*cs++ = MI_STORE_DWORD_IMM | MI_MEM_VIRTUAL;
*cs++ = addr;
*cs++ = value;
*cs++ = MI_NOOP;
}
intel_ring_advance(rq, cs);
return 0;
}
static struct i915_request *
tl_write(struct intel_timeline *tl, struct intel_engine_cs *engine, u32 value)
{
struct i915_request *rq;
int err;
lockdep_assert_held(&tl->gt->i915->drm.struct_mutex); /* lazy rq refs */
err = intel_timeline_pin(tl);
if (err) {
rq = ERR_PTR(err);
goto out;
}
rq = i915_request_create(engine->kernel_context);
if (IS_ERR(rq))
goto out_unpin;
err = emit_ggtt_store_dw(rq, tl->hwsp_offset, value);
i915_request_add(rq);
if (err)
rq = ERR_PTR(err);
out_unpin:
intel_timeline_unpin(tl);
out:
if (IS_ERR(rq))
pr_err("Failed to write to timeline!\n");
return rq;
}
static struct intel_timeline *
checked_intel_timeline_create(struct drm_i915_private *i915)
{
struct intel_timeline *tl;
tl = intel_timeline_create(&i915->gt, NULL);
if (IS_ERR(tl))
return tl;
if (*tl->hwsp_seqno != tl->seqno) {
pr_err("Timeline created with incorrect breadcrumb, found %x, expected %x\n",
*tl->hwsp_seqno, tl->seqno);
intel_timeline_put(tl);
return ERR_PTR(-EINVAL);
}
return tl;
}
static int live_hwsp_engine(void *arg)
{
#define NUM_TIMELINES 4096
struct drm_i915_private *i915 = arg;
struct intel_timeline **timelines;
struct intel_engine_cs *engine;
enum intel_engine_id id;
intel_wakeref_t wakeref;
unsigned long count, n;
int err = 0;
/*
* Create a bunch of timelines and check we can write
* independently to each of their breadcrumb slots.
*/
timelines = kvmalloc_array(NUM_TIMELINES * I915_NUM_ENGINES,
sizeof(*timelines),
GFP_KERNEL);
if (!timelines)
return -ENOMEM;
mutex_lock(&i915->drm.struct_mutex);
wakeref = intel_runtime_pm_get(&i915->runtime_pm);
count = 0;
for_each_engine(engine, i915, id) {
if (!intel_engine_can_store_dword(engine))
continue;
for (n = 0; n < NUM_TIMELINES; n++) {
struct intel_timeline *tl;
struct i915_request *rq;
tl = checked_intel_timeline_create(i915);
if (IS_ERR(tl)) {
err = PTR_ERR(tl);
goto out;
}
rq = tl_write(tl, engine, count);
if (IS_ERR(rq)) {
intel_timeline_put(tl);
err = PTR_ERR(rq);
goto out;
}
timelines[count++] = tl;
}
}
out:
if (igt_flush_test(i915, I915_WAIT_LOCKED))
err = -EIO;
for (n = 0; n < count; n++) {
struct intel_timeline *tl = timelines[n];
if (!err && *tl->hwsp_seqno != n) {
pr_err("Invalid seqno stored in timeline %lu, found 0x%x\n",
n, *tl->hwsp_seqno);
err = -EINVAL;
}
intel_timeline_put(tl);
}
intel_runtime_pm_put(&i915->runtime_pm, wakeref);
mutex_unlock(&i915->drm.struct_mutex);
kvfree(timelines);
return err;
#undef NUM_TIMELINES
}
static int live_hwsp_alternate(void *arg)
{
#define NUM_TIMELINES 4096
struct drm_i915_private *i915 = arg;
struct intel_timeline **timelines;
struct intel_engine_cs *engine;
enum intel_engine_id id;
intel_wakeref_t wakeref;
unsigned long count, n;
int err = 0;
/*
* Create a bunch of timelines and check we can write
* independently to each of their breadcrumb slots with adjacent
* engines.
*/
timelines = kvmalloc_array(NUM_TIMELINES * I915_NUM_ENGINES,
sizeof(*timelines),
GFP_KERNEL);
if (!timelines)
return -ENOMEM;
mutex_lock(&i915->drm.struct_mutex);
wakeref = intel_runtime_pm_get(&i915->runtime_pm);
count = 0;
for (n = 0; n < NUM_TIMELINES; n++) {
for_each_engine(engine, i915, id) {
struct intel_timeline *tl;
struct i915_request *rq;
if (!intel_engine_can_store_dword(engine))
continue;
tl = checked_intel_timeline_create(i915);
if (IS_ERR(tl)) {
err = PTR_ERR(tl);
goto out;
}
rq = tl_write(tl, engine, count);
if (IS_ERR(rq)) {
intel_timeline_put(tl);
err = PTR_ERR(rq);
goto out;
}
timelines[count++] = tl;
}
}
out:
if (igt_flush_test(i915, I915_WAIT_LOCKED))
err = -EIO;
for (n = 0; n < count; n++) {
struct intel_timeline *tl = timelines[n];
if (!err && *tl->hwsp_seqno != n) {
pr_err("Invalid seqno stored in timeline %lu, found 0x%x\n",
n, *tl->hwsp_seqno);
err = -EINVAL;
}
intel_timeline_put(tl);
}
intel_runtime_pm_put(&i915->runtime_pm, wakeref);
mutex_unlock(&i915->drm.struct_mutex);
kvfree(timelines);
return err;
#undef NUM_TIMELINES
}
static int live_hwsp_wrap(void *arg)
{
struct drm_i915_private *i915 = arg;
struct intel_engine_cs *engine;
struct intel_timeline *tl;
enum intel_engine_id id;
intel_wakeref_t wakeref;
int err = 0;
/*
* Across a seqno wrap, we need to keep the old cacheline alive for
* foreign GPU references.
*/
mutex_lock(&i915->drm.struct_mutex);
wakeref = intel_runtime_pm_get(&i915->runtime_pm);
tl = intel_timeline_create(&i915->gt, NULL);
if (IS_ERR(tl)) {
err = PTR_ERR(tl);
goto out_rpm;
}
if (!tl->has_initial_breadcrumb || !tl->hwsp_cacheline)
goto out_free;
err = intel_timeline_pin(tl);
if (err)
goto out_free;
for_each_engine(engine, i915, id) {
const u32 *hwsp_seqno[2];
struct i915_request *rq;
u32 seqno[2];
if (!intel_engine_can_store_dword(engine))
continue;
rq = i915_request_create(engine->kernel_context);
if (IS_ERR(rq)) {
err = PTR_ERR(rq);
goto out;
}
tl->seqno = -4u;
mutex_lock_nested(&tl->mutex, SINGLE_DEPTH_NESTING);
err = intel_timeline_get_seqno(tl, rq, &seqno[0]);
mutex_unlock(&tl->mutex);
if (err) {
i915_request_add(rq);
goto out;
}
pr_debug("seqno[0]:%08x, hwsp_offset:%08x\n",
seqno[0], tl->hwsp_offset);
err = emit_ggtt_store_dw(rq, tl->hwsp_offset, seqno[0]);
if (err) {
i915_request_add(rq);
goto out;
}
hwsp_seqno[0] = tl->hwsp_seqno;
mutex_lock_nested(&tl->mutex, SINGLE_DEPTH_NESTING);
err = intel_timeline_get_seqno(tl, rq, &seqno[1]);
mutex_unlock(&tl->mutex);
if (err) {
i915_request_add(rq);
goto out;
}
pr_debug("seqno[1]:%08x, hwsp_offset:%08x\n",
seqno[1], tl->hwsp_offset);
err = emit_ggtt_store_dw(rq, tl->hwsp_offset, seqno[1]);
if (err) {
i915_request_add(rq);
goto out;
}
hwsp_seqno[1] = tl->hwsp_seqno;
/* With wrap should come a new hwsp */
GEM_BUG_ON(seqno[1] >= seqno[0]);
GEM_BUG_ON(hwsp_seqno[0] == hwsp_seqno[1]);
i915_request_add(rq);
if (i915_request_wait(rq, 0, HZ / 5) < 0) {
pr_err("Wait for timeline writes timed out!\n");
err = -EIO;
goto out;
}
if (*hwsp_seqno[0] != seqno[0] || *hwsp_seqno[1] != seqno[1]) {
pr_err("Bad timeline values: found (%x, %x), expected (%x, %x)\n",
*hwsp_seqno[0], *hwsp_seqno[1],
seqno[0], seqno[1]);
err = -EINVAL;
goto out;
}
i915_retire_requests(i915); /* recycle HWSP */
}
out:
if (igt_flush_test(i915, I915_WAIT_LOCKED))
err = -EIO;
intel_timeline_unpin(tl);
out_free:
intel_timeline_put(tl);
out_rpm:
intel_runtime_pm_put(&i915->runtime_pm, wakeref);
mutex_unlock(&i915->drm.struct_mutex);
return err;
}
static int live_hwsp_recycle(void *arg)
{
struct drm_i915_private *i915 = arg;
struct intel_engine_cs *engine;
enum intel_engine_id id;
intel_wakeref_t wakeref;
unsigned long count;
int err = 0;
/*
* Check seqno writes into one timeline at a time. We expect to
* recycle the breadcrumb slot between iterations and neither
* want to confuse ourselves or the GPU.
*/
mutex_lock(&i915->drm.struct_mutex);
wakeref = intel_runtime_pm_get(&i915->runtime_pm);
count = 0;
for_each_engine(engine, i915, id) {
IGT_TIMEOUT(end_time);
if (!intel_engine_can_store_dword(engine))
continue;
do {
struct intel_timeline *tl;
struct i915_request *rq;
tl = checked_intel_timeline_create(i915);
if (IS_ERR(tl)) {
err = PTR_ERR(tl);
goto out;
}
rq = tl_write(tl, engine, count);
if (IS_ERR(rq)) {
intel_timeline_put(tl);
err = PTR_ERR(rq);
goto out;
}
if (i915_request_wait(rq, 0, HZ / 5) < 0) {
pr_err("Wait for timeline writes timed out!\n");
intel_timeline_put(tl);
err = -EIO;
goto out;
}
if (*tl->hwsp_seqno != count) {
pr_err("Invalid seqno stored in timeline %lu, found 0x%x\n",
count, *tl->hwsp_seqno);
err = -EINVAL;
}
intel_timeline_put(tl);
count++;
if (err)
goto out;
} while (!__igt_timeout(end_time, NULL));
}
out:
intel_runtime_pm_put(&i915->runtime_pm, wakeref);
mutex_unlock(&i915->drm.struct_mutex);
return err;
}
int intel_timeline_live_selftests(struct drm_i915_private *i915)
{
static const struct i915_subtest tests[] = {
SUBTEST(live_hwsp_recycle),
SUBTEST(live_hwsp_engine),
SUBTEST(live_hwsp_alternate),
SUBTEST(live_hwsp_wrap),
};
if (intel_gt_is_wedged(&i915->gt))
return 0;
return i915_live_subtests(tests, i915);
}
|