summaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/msm/msm_gpu.c
blob: 4edb874548b322be130d8c755a01f3730ea42b42 (plain)
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
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
// SPDX-License-Identifier: GPL-2.0-only
/*
 * Copyright (C) 2013 Red Hat
 * Author: Rob Clark <robdclark@gmail.com>
 */

#include "msm_gpu.h"
#include "msm_gem.h"
#include "msm_mmu.h"
#include "msm_fence.h"
#include "msm_gpu_trace.h"
#include "adreno/adreno_gpu.h"

#include <generated/utsrelease.h>
#include <linux/string_helpers.h>
#include <linux/pm_opp.h>
#include <linux/devfreq.h>
#include <linux/devcoredump.h>

/*
 * Power Management:
 */

static int msm_devfreq_target(struct device *dev, unsigned long *freq,
		u32 flags)
{
	struct msm_gpu *gpu = platform_get_drvdata(to_platform_device(dev));
	struct dev_pm_opp *opp;

	opp = devfreq_recommended_opp(dev, freq, flags);

	if (IS_ERR(opp))
		return PTR_ERR(opp);

	if (gpu->funcs->gpu_set_freq)
		gpu->funcs->gpu_set_freq(gpu, (u64)*freq);
	else
		clk_set_rate(gpu->core_clk, *freq);

	dev_pm_opp_put(opp);

	return 0;
}

static int msm_devfreq_get_dev_status(struct device *dev,
		struct devfreq_dev_status *status)
{
	struct msm_gpu *gpu = platform_get_drvdata(to_platform_device(dev));
	ktime_t time;

	if (gpu->funcs->gpu_get_freq)
		status->current_frequency = gpu->funcs->gpu_get_freq(gpu);
	else
		status->current_frequency = clk_get_rate(gpu->core_clk);

	status->busy_time = gpu->funcs->gpu_busy(gpu);

	time = ktime_get();
	status->total_time = ktime_us_delta(time, gpu->devfreq.time);
	gpu->devfreq.time = time;

	return 0;
}

static int msm_devfreq_get_cur_freq(struct device *dev, unsigned long *freq)
{
	struct msm_gpu *gpu = platform_get_drvdata(to_platform_device(dev));

	if (gpu->funcs->gpu_get_freq)
		*freq = gpu->funcs->gpu_get_freq(gpu);
	else
		*freq = clk_get_rate(gpu->core_clk);

	return 0;
}

static struct devfreq_dev_profile msm_devfreq_profile = {
	.polling_ms = 10,
	.target = msm_devfreq_target,
	.get_dev_status = msm_devfreq_get_dev_status,
	.get_cur_freq = msm_devfreq_get_cur_freq,
};

static void msm_devfreq_init(struct msm_gpu *gpu)
{
	/* We need target support to do devfreq */
	if (!gpu->funcs->gpu_busy)
		return;

	msm_devfreq_profile.initial_freq = gpu->fast_rate;

	/*
	 * Don't set the freq_table or max_state and let devfreq build the table
	 * from OPP
	 */

	gpu->devfreq.devfreq = devm_devfreq_add_device(&gpu->pdev->dev,
			&msm_devfreq_profile, "simple_ondemand", NULL);

	if (IS_ERR(gpu->devfreq.devfreq)) {
		DRM_DEV_ERROR(&gpu->pdev->dev, "Couldn't initialize GPU devfreq\n");
		gpu->devfreq.devfreq = NULL;
	}

	devfreq_suspend_device(gpu->devfreq.devfreq);
}

static int enable_pwrrail(struct msm_gpu *gpu)
{
	struct drm_device *dev = gpu->dev;
	int ret = 0;

	if (gpu->gpu_reg) {
		ret = regulator_enable(gpu->gpu_reg);
		if (ret) {
			DRM_DEV_ERROR(dev->dev, "failed to enable 'gpu_reg': %d\n", ret);
			return ret;
		}
	}

	if (gpu->gpu_cx) {
		ret = regulator_enable(gpu->gpu_cx);
		if (ret) {
			DRM_DEV_ERROR(dev->dev, "failed to enable 'gpu_cx': %d\n", ret);
			return ret;
		}
	}

	return 0;
}

static int disable_pwrrail(struct msm_gpu *gpu)
{
	if (gpu->gpu_cx)
		regulator_disable(gpu->gpu_cx);
	if (gpu->gpu_reg)
		regulator_disable(gpu->gpu_reg);
	return 0;
}

static int enable_clk(struct msm_gpu *gpu)
{
	if (gpu->core_clk && gpu->fast_rate)
		clk_set_rate(gpu->core_clk, gpu->fast_rate);

	/* Set the RBBM timer rate to 19.2Mhz */
	if (gpu->rbbmtimer_clk)
		clk_set_rate(gpu->rbbmtimer_clk, 19200000);

	return clk_bulk_prepare_enable(gpu->nr_clocks, gpu->grp_clks);
}

static int disable_clk(struct msm_gpu *gpu)
{
	clk_bulk_disable_unprepare(gpu->nr_clocks, gpu->grp_clks);

	/*
	 * Set the clock to a deliberately low rate. On older targets the clock
	 * speed had to be non zero to avoid problems. On newer targets this
	 * will be rounded down to zero anyway so it all works out.
	 */
	if (gpu->core_clk)
		clk_set_rate(gpu->core_clk, 27000000);

	if (gpu->rbbmtimer_clk)
		clk_set_rate(gpu->rbbmtimer_clk, 0);

	return 0;
}

static int enable_axi(struct msm_gpu *gpu)
{
	if (gpu->ebi1_clk)
		clk_prepare_enable(gpu->ebi1_clk);
	return 0;
}

static int disable_axi(struct msm_gpu *gpu)
{
	if (gpu->ebi1_clk)
		clk_disable_unprepare(gpu->ebi1_clk);
	return 0;
}

void msm_gpu_resume_devfreq(struct msm_gpu *gpu)
{
	gpu->devfreq.busy_cycles = 0;
	gpu->devfreq.time = ktime_get();

	devfreq_resume_device(gpu->devfreq.devfreq);
}

int msm_gpu_pm_resume(struct msm_gpu *gpu)
{
	int ret;

	DBG("%s", gpu->name);

	ret = enable_pwrrail(gpu);
	if (ret)
		return ret;

	ret = enable_clk(gpu);
	if (ret)
		return ret;

	ret = enable_axi(gpu);
	if (ret)
		return ret;

	msm_gpu_resume_devfreq(gpu);

	gpu->needs_hw_init = true;

	return 0;
}

int msm_gpu_pm_suspend(struct msm_gpu *gpu)
{
	int ret;

	DBG("%s", gpu->name);

	devfreq_suspend_device(gpu->devfreq.devfreq);

	ret = disable_axi(gpu);
	if (ret)
		return ret;

	ret = disable_clk(gpu);
	if (ret)
		return ret;

	ret = disable_pwrrail(gpu);
	if (ret)
		return ret;

	return 0;
}

int msm_gpu_hw_init(struct msm_gpu *gpu)
{
	int ret;

	WARN_ON(!mutex_is_locked(&gpu->dev->struct_mutex));

	if (!gpu->needs_hw_init)
		return 0;

	disable_irq(gpu->irq);
	ret = gpu->funcs->hw_init(gpu);
	if (!ret)
		gpu->needs_hw_init = false;
	enable_irq(gpu->irq);

	return ret;
}

#ifdef CONFIG_DEV_COREDUMP
static ssize_t msm_gpu_devcoredump_read(char *buffer, loff_t offset,
		size_t count, void *data, size_t datalen)
{
	struct msm_gpu *gpu = data;
	struct drm_print_iterator iter;
	struct drm_printer p;
	struct msm_gpu_state *state;

	state = msm_gpu_crashstate_get(gpu);
	if (!state)
		return 0;

	iter.data = buffer;
	iter.offset = 0;
	iter.start = offset;
	iter.remain = count;

	p = drm_coredump_printer(&iter);

	drm_printf(&p, "---\n");
	drm_printf(&p, "kernel: " UTS_RELEASE "\n");
	drm_printf(&p, "module: " KBUILD_MODNAME "\n");
	drm_printf(&p, "time: %lld.%09ld\n",
		state->time.tv_sec, state->time.tv_nsec);
	if (state->comm)
		drm_printf(&p, "comm: %s\n", state->comm);
	if (state->cmd)
		drm_printf(&p, "cmdline: %s\n", state->cmd);

	gpu->funcs->show(gpu, state, &p);

	msm_gpu_crashstate_put(gpu);

	return count - iter.remain;
}

static void msm_gpu_devcoredump_free(void *data)
{
	struct msm_gpu *gpu = data;

	msm_gpu_crashstate_put(gpu);
}

static void msm_gpu_crashstate_get_bo(struct msm_gpu_state *state,
		struct msm_gem_object *obj, u64 iova, u32 flags)
{
	struct msm_gpu_state_bo *state_bo = &state->bos[state->nr_bos];

	/* Don't record write only objects */
	state_bo->size = obj->base.size;
	state_bo->iova = iova;

	/* Only store data for non imported buffer objects marked for read */
	if ((flags & MSM_SUBMIT_BO_READ) && !obj->base.import_attach) {
		void *ptr;

		state_bo->data = kvmalloc(obj->base.size, GFP_KERNEL);
		if (!state_bo->data)
			goto out;

		ptr = msm_gem_get_vaddr_active(&obj->base);
		if (IS_ERR(ptr)) {
			kvfree(state_bo->data);
			state_bo->data = NULL;
			goto out;
		}

		memcpy(state_bo->data, ptr, obj->base.size);
		msm_gem_put_vaddr(&obj->base);
	}
out:
	state->nr_bos++;
}

static void msm_gpu_crashstate_capture(struct msm_gpu *gpu,
		struct msm_gem_submit *submit, char *comm, char *cmd)
{
	struct msm_gpu_state *state;

	/* Check if the target supports capturing crash state */
	if (!gpu->funcs->gpu_state_get)
		return;

	/* Only save one crash state at a time */
	if (gpu->crashstate)
		return;

	state = gpu->funcs->gpu_state_get(gpu);
	if (IS_ERR_OR_NULL(state))
		return;

	/* Fill in the additional crash state information */
	state->comm = kstrdup(comm, GFP_KERNEL);
	state->cmd = kstrdup(cmd, GFP_KERNEL);

	if (submit) {
		int i;

		state->bos = kcalloc(submit->nr_cmds,
			sizeof(struct msm_gpu_state_bo), GFP_KERNEL);

		for (i = 0; state->bos && i < submit->nr_cmds; i++) {
			int idx = submit->cmd[i].idx;

			msm_gpu_crashstate_get_bo(state, submit->bos[idx].obj,
				submit->bos[idx].iova, submit->bos[idx].flags);
		}
	}

	/* Set the active crash state to be dumped on failure */
	gpu->crashstate = state;

	/* FIXME: Release the crashstate if this errors out? */
	dev_coredumpm(gpu->dev->dev, THIS_MODULE, gpu, 0, GFP_KERNEL,
		msm_gpu_devcoredump_read, msm_gpu_devcoredump_free);
}
#else
static void msm_gpu_crashstate_capture(struct msm_gpu *gpu,
		struct msm_gem_submit *submit, char *comm, char *cmd)
{
}
#endif

/*
 * Hangcheck detection for locked gpu:
 */

static void update_fences(struct msm_gpu *gpu, struct msm_ringbuffer *ring,
		uint32_t fence)
{
	struct msm_gem_submit *submit;

	list_for_each_entry(submit, &ring->submits, node) {
		if (submit->seqno > fence)
			break;

		msm_update_fence(submit->ring->fctx,
			submit->fence->seqno);
	}
}

static struct msm_gem_submit *
find_submit(struct msm_ringbuffer *ring, uint32_t fence)
{
	struct msm_gem_submit *submit;

	WARN_ON(!mutex_is_locked(&ring->gpu->dev->struct_mutex));

	list_for_each_entry(submit, &ring->submits, node)
		if (submit->seqno == fence)
			return submit;

	return NULL;
}

static void retire_submits(struct msm_gpu *gpu);

static void recover_worker(struct work_struct *work)
{
	struct msm_gpu *gpu = container_of(work, struct msm_gpu, recover_work);
	struct drm_device *dev = gpu->dev;
	struct msm_drm_private *priv = dev->dev_private;
	struct msm_gem_submit *submit;
	struct msm_ringbuffer *cur_ring = gpu->funcs->active_ring(gpu);
	char *comm = NULL, *cmd = NULL;
	int i;

	mutex_lock(&dev->struct_mutex);

	DRM_DEV_ERROR(dev->dev, "%s: hangcheck recover!\n", gpu->name);

	submit = find_submit(cur_ring, cur_ring->memptrs->fence + 1);
	if (submit) {
		struct task_struct *task;

		/* Increment the fault counts */
		gpu->global_faults++;
		submit->queue->faults++;

		task = get_pid_task(submit->pid, PIDTYPE_PID);
		if (task) {
			comm = kstrdup(task->comm, GFP_KERNEL);
			cmd = kstrdup_quotable_cmdline(task, GFP_KERNEL);
			put_task_struct(task);
		}

		if (comm && cmd) {
			DRM_DEV_ERROR(dev->dev, "%s: offending task: %s (%s)\n",
				gpu->name, comm, cmd);

			msm_rd_dump_submit(priv->hangrd, submit,
				"offending task: %s (%s)", comm, cmd);
		} else
			msm_rd_dump_submit(priv->hangrd, submit, NULL);
	}

	/* Record the crash state */
	pm_runtime_get_sync(&gpu->pdev->dev);
	msm_gpu_crashstate_capture(gpu, submit, comm, cmd);
	pm_runtime_put_sync(&gpu->pdev->dev);

	kfree(cmd);
	kfree(comm);

	/*
	 * Update all the rings with the latest and greatest fence.. this
	 * needs to happen after msm_rd_dump_submit() to ensure that the
	 * bo's referenced by the offending submit are still around.
	 */
	for (i = 0; i < gpu->nr_rings; i++) {
		struct msm_ringbuffer *ring = gpu->rb[i];

		uint32_t fence = ring->memptrs->fence;

		/*
		 * For the current (faulting?) ring/submit advance the fence by
		 * one more to clear the faulting submit
		 */
		if (ring == cur_ring)
			fence++;

		update_fences(gpu, ring, fence);
	}

	if (msm_gpu_active(gpu)) {
		/* retire completed submits, plus the one that hung: */
		retire_submits(gpu);

		pm_runtime_get_sync(&gpu->pdev->dev);
		gpu->funcs->recover(gpu);
		pm_runtime_put_sync(&gpu->pdev->dev);

		/*
		 * Replay all remaining submits starting with highest priority
		 * ring
		 */
		for (i = 0; i < gpu->nr_rings; i++) {
			struct msm_ringbuffer *ring = gpu->rb[i];

			list_for_each_entry(submit, &ring->submits, node)
				gpu->funcs->submit(gpu, submit, NULL);
		}
	}

	mutex_unlock(&dev->struct_mutex);

	msm_gpu_retire(gpu);
}

static void hangcheck_timer_reset(struct msm_gpu *gpu)
{
	DBG("%s", gpu->name);
	mod_timer(&gpu->hangcheck_timer,
			round_jiffies_up(jiffies + DRM_MSM_HANGCHECK_JIFFIES));
}

static void hangcheck_handler(struct timer_list *t)
{
	struct msm_gpu *gpu = from_timer(gpu, t, hangcheck_timer);
	struct drm_device *dev = gpu->dev;
	struct msm_drm_private *priv = dev->dev_private;
	struct msm_ringbuffer *ring = gpu->funcs->active_ring(gpu);
	uint32_t fence = ring->memptrs->fence;

	if (fence != ring->hangcheck_fence) {
		/* some progress has been made.. ya! */
		ring->hangcheck_fence = fence;
	} else if (fence < ring->seqno) {
		/* no progress and not done.. hung! */
		ring->hangcheck_fence = fence;
		DRM_DEV_ERROR(dev->dev, "%s: hangcheck detected gpu lockup rb %d!\n",
				gpu->name, ring->id);
		DRM_DEV_ERROR(dev->dev, "%s:     completed fence: %u\n",
				gpu->name, fence);
		DRM_DEV_ERROR(dev->dev, "%s:     submitted fence: %u\n",
				gpu->name, ring->seqno);

		queue_work(priv->wq, &gpu->recover_work);
	}

	/* if still more pending work, reset the hangcheck timer: */
	if (ring->seqno > ring->hangcheck_fence)
		hangcheck_timer_reset(gpu);

	/* workaround for missing irq: */
	queue_work(priv->wq, &gpu->retire_work);
}

/*
 * Performance Counters:
 */

/* called under perf_lock */
static int update_hw_cntrs(struct msm_gpu *gpu, uint32_t ncntrs, uint32_t *cntrs)
{
	uint32_t current_cntrs[ARRAY_SIZE(gpu->last_cntrs)];
	int i, n = min(ncntrs, gpu->num_perfcntrs);

	/* read current values: */
	for (i = 0; i < gpu->num_perfcntrs; i++)
		current_cntrs[i] = gpu_read(gpu, gpu->perfcntrs[i].sample_reg);

	/* update cntrs: */
	for (i = 0; i < n; i++)
		cntrs[i] = current_cntrs[i] - gpu->last_cntrs[i];

	/* save current values: */
	for (i = 0; i < gpu->num_perfcntrs; i++)
		gpu->last_cntrs[i] = current_cntrs[i];

	return n;
}

static void update_sw_cntrs(struct msm_gpu *gpu)
{
	ktime_t time;
	uint32_t elapsed;
	unsigned long flags;

	spin_lock_irqsave(&gpu->perf_lock, flags);
	if (!gpu->perfcntr_active)
		goto out;

	time = ktime_get();
	elapsed = ktime_to_us(ktime_sub(time, gpu->last_sample.time));

	gpu->totaltime += elapsed;
	if (gpu->last_sample.active)
		gpu->activetime += elapsed;

	gpu->last_sample.active = msm_gpu_active(gpu);
	gpu->last_sample.time = time;

out:
	spin_unlock_irqrestore(&gpu->perf_lock, flags);
}

void msm_gpu_perfcntr_start(struct msm_gpu *gpu)
{
	unsigned long flags;

	pm_runtime_get_sync(&gpu->pdev->dev);

	spin_lock_irqsave(&gpu->perf_lock, flags);
	/* we could dynamically enable/disable perfcntr registers too.. */
	gpu->last_sample.active = msm_gpu_active(gpu);
	gpu->last_sample.time = ktime_get();
	gpu->activetime = gpu->totaltime = 0;
	gpu->perfcntr_active = true;
	update_hw_cntrs(gpu, 0, NULL);
	spin_unlock_irqrestore(&gpu->perf_lock, flags);
}

void msm_gpu_perfcntr_stop(struct msm_gpu *gpu)
{
	gpu->perfcntr_active = false;
	pm_runtime_put_sync(&gpu->pdev->dev);
}

/* returns -errno or # of cntrs sampled */
int msm_gpu_perfcntr_sample(struct msm_gpu *gpu, uint32_t *activetime,
		uint32_t *totaltime, uint32_t ncntrs, uint32_t *cntrs)
{
	unsigned long flags;
	int ret;

	spin_lock_irqsave(&gpu->perf_lock, flags);

	if (!gpu->perfcntr_active) {
		ret = -EINVAL;
		goto out;
	}

	*activetime = gpu->activetime;
	*totaltime = gpu->totaltime;

	gpu->activetime = gpu->totaltime = 0;

	ret = update_hw_cntrs(gpu, ncntrs, cntrs);

out:
	spin_unlock_irqrestore(&gpu->perf_lock, flags);

	return ret;
}

/*
 * Cmdstream submission/retirement:
 */

static void retire_submit(struct msm_gpu *gpu, struct msm_ringbuffer *ring,
		struct msm_gem_submit *submit)
{
	int index = submit->seqno % MSM_GPU_SUBMIT_STATS_COUNT;
	volatile struct msm_gpu_submit_stats *stats;
	u64 elapsed, clock = 0;
	int i;

	stats = &ring->memptrs->stats[index];
	/* Convert 19.2Mhz alwayson ticks to nanoseconds for elapsed time */
	elapsed = (stats->alwayson_end - stats->alwayson_start) * 10000;
	do_div(elapsed, 192);

	/* Calculate the clock frequency from the number of CP cycles */
	if (elapsed) {
		clock = (stats->cpcycles_end - stats->cpcycles_start) * 1000;
		do_div(clock, elapsed);
	}

	trace_msm_gpu_submit_retired(submit, elapsed, clock,
		stats->alwayson_start, stats->alwayson_end);

	for (i = 0; i < submit->nr_bos; i++) {
		struct msm_gem_object *msm_obj = submit->bos[i].obj;
		/* move to inactive: */
		msm_gem_move_to_inactive(&msm_obj->base);
		msm_gem_unpin_iova(&msm_obj->base, submit->aspace);
		drm_gem_object_put(&msm_obj->base);
	}

	pm_runtime_mark_last_busy(&gpu->pdev->dev);
	pm_runtime_put_autosuspend(&gpu->pdev->dev);
	msm_gem_submit_free(submit);
}

static void retire_submits(struct msm_gpu *gpu)
{
	struct drm_device *dev = gpu->dev;
	struct msm_gem_submit *submit, *tmp;
	int i;

	WARN_ON(!mutex_is_locked(&dev->struct_mutex));

	/* Retire the commits starting with highest priority */
	for (i = 0; i < gpu->nr_rings; i++) {
		struct msm_ringbuffer *ring = gpu->rb[i];

		list_for_each_entry_safe(submit, tmp, &ring->submits, node) {
			if (dma_fence_is_signaled(submit->fence))
				retire_submit(gpu, ring, submit);
		}
	}
}

static void retire_worker(struct work_struct *work)
{
	struct msm_gpu *gpu = container_of(work, struct msm_gpu, retire_work);
	struct drm_device *dev = gpu->dev;
	int i;

	for (i = 0; i < gpu->nr_rings; i++)
		update_fences(gpu, gpu->rb[i], gpu->rb[i]->memptrs->fence);

	mutex_lock(&dev->struct_mutex);
	retire_submits(gpu);
	mutex_unlock(&dev->struct_mutex);
}

/* call from irq handler to schedule work to retire bo's */
void msm_gpu_retire(struct msm_gpu *gpu)
{
	struct msm_drm_private *priv = gpu->dev->dev_private;
	queue_work(priv->wq, &gpu->retire_work);
	update_sw_cntrs(gpu);
}

/* add bo's to gpu's ring, and kick gpu: */
void msm_gpu_submit(struct msm_gpu *gpu, struct msm_gem_submit *submit,
		struct msm_file_private *ctx)
{
	struct drm_device *dev = gpu->dev;
	struct msm_drm_private *priv = dev->dev_private;
	struct msm_ringbuffer *ring = submit->ring;
	int i;

	WARN_ON(!mutex_is_locked(&dev->struct_mutex));

	pm_runtime_get_sync(&gpu->pdev->dev);

	msm_gpu_hw_init(gpu);

	submit->seqno = ++ring->seqno;

	list_add_tail(&submit->node, &ring->submits);

	msm_rd_dump_submit(priv->rd, submit, NULL);

	update_sw_cntrs(gpu);

	for (i = 0; i < submit->nr_bos; i++) {
		struct msm_gem_object *msm_obj = submit->bos[i].obj;
		uint64_t iova;

		/* can't happen yet.. but when we add 2d support we'll have
		 * to deal w/ cross-ring synchronization:
		 */
		WARN_ON(is_active(msm_obj) && (msm_obj->gpu != gpu));

		/* submit takes a reference to the bo and iova until retired: */
		drm_gem_object_get(&msm_obj->base);
		msm_gem_get_and_pin_iova(&msm_obj->base, submit->aspace, &iova);

		if (submit->bos[i].flags & MSM_SUBMIT_BO_WRITE)
			msm_gem_move_to_active(&msm_obj->base, gpu, true, submit->fence);
		else if (submit->bos[i].flags & MSM_SUBMIT_BO_READ)
			msm_gem_move_to_active(&msm_obj->base, gpu, false, submit->fence);
	}

	gpu->funcs->submit(gpu, submit, ctx);
	priv->lastctx = ctx;

	hangcheck_timer_reset(gpu);
}

/*
 * Init/Cleanup:
 */

static irqreturn_t irq_handler(int irq, void *data)
{
	struct msm_gpu *gpu = data;
	return gpu->funcs->irq(gpu);
}

static int get_clocks(struct platform_device *pdev, struct msm_gpu *gpu)
{
	int ret = msm_clk_bulk_get(&pdev->dev, &gpu->grp_clks);

	if (ret < 1) {
		gpu->nr_clocks = 0;
		return ret;
	}

	gpu->nr_clocks = ret;

	gpu->core_clk = msm_clk_bulk_get_clock(gpu->grp_clks,
		gpu->nr_clocks, "core");

	gpu->rbbmtimer_clk = msm_clk_bulk_get_clock(gpu->grp_clks,
		gpu->nr_clocks, "rbbmtimer");

	return 0;
}

static struct msm_gem_address_space *
msm_gpu_create_address_space(struct msm_gpu *gpu, struct platform_device *pdev,
		uint64_t va_start, uint64_t va_end)
{
	struct msm_gem_address_space *aspace;
	int ret;

	/*
	 * Setup IOMMU.. eventually we will (I think) do this once per context
	 * and have separate page tables per context.  For now, to keep things
	 * simple and to get something working, just use a single address space:
	 */
	if (!adreno_is_a2xx(to_adreno_gpu(gpu))) {
		struct iommu_domain *iommu = iommu_domain_alloc(&platform_bus_type);
		if (!iommu)
			return NULL;

		iommu->geometry.aperture_start = va_start;
		iommu->geometry.aperture_end = va_end;

		DRM_DEV_INFO(gpu->dev->dev, "%s: using IOMMU\n", gpu->name);

		aspace = msm_gem_address_space_create(&pdev->dev, iommu, "gpu");
		if (IS_ERR(aspace))
			iommu_domain_free(iommu);
	} else {
		aspace = msm_gem_address_space_create_a2xx(&pdev->dev, gpu, "gpu",
			va_start, va_end);
	}

	if (IS_ERR(aspace)) {
		DRM_DEV_ERROR(gpu->dev->dev, "failed to init mmu: %ld\n",
			PTR_ERR(aspace));
		return ERR_CAST(aspace);
	}

	ret = aspace->mmu->funcs->attach(aspace->mmu, NULL, 0);
	if (ret) {
		msm_gem_address_space_put(aspace);
		return ERR_PTR(ret);
	}

	return aspace;
}

int msm_gpu_init(struct drm_device *drm, struct platform_device *pdev,
		struct msm_gpu *gpu, const struct msm_gpu_funcs *funcs,
		const char *name, struct msm_gpu_config *config)
{
	int i, ret, nr_rings = config->nr_rings;
	void *memptrs;
	uint64_t memptrs_iova;

	if (WARN_ON(gpu->num_perfcntrs > ARRAY_SIZE(gpu->last_cntrs)))
		gpu->num_perfcntrs = ARRAY_SIZE(gpu->last_cntrs);

	gpu->dev = drm;
	gpu->funcs = funcs;
	gpu->name = name;

	INIT_LIST_HEAD(&gpu->active_list);
	INIT_WORK(&gpu->retire_work, retire_worker);
	INIT_WORK(&gpu->recover_work, recover_worker);


	timer_setup(&gpu->hangcheck_timer, hangcheck_handler, 0);

	spin_lock_init(&gpu->perf_lock);


	/* Map registers: */
	gpu->mmio = msm_ioremap(pdev, config->ioname, name);
	if (IS_ERR(gpu->mmio)) {
		ret = PTR_ERR(gpu->mmio);
		goto fail;
	}

	/* Get Interrupt: */
	gpu->irq = platform_get_irq(pdev, 0);
	if (gpu->irq < 0) {
		ret = gpu->irq;
		DRM_DEV_ERROR(drm->dev, "failed to get irq: %d\n", ret);
		goto fail;
	}

	ret = devm_request_irq(&pdev->dev, gpu->irq, irq_handler,
			IRQF_TRIGGER_HIGH, gpu->name, gpu);
	if (ret) {
		DRM_DEV_ERROR(drm->dev, "failed to request IRQ%u: %d\n", gpu->irq, ret);
		goto fail;
	}

	ret = get_clocks(pdev, gpu);
	if (ret)
		goto fail;

	gpu->ebi1_clk = msm_clk_get(pdev, "bus");
	DBG("ebi1_clk: %p", gpu->ebi1_clk);
	if (IS_ERR(gpu->ebi1_clk))
		gpu->ebi1_clk = NULL;

	/* Acquire regulators: */
	gpu->gpu_reg = devm_regulator_get(&pdev->dev, "vdd");
	DBG("gpu_reg: %p", gpu->gpu_reg);
	if (IS_ERR(gpu->gpu_reg))
		gpu->gpu_reg = NULL;

	gpu->gpu_cx = devm_regulator_get(&pdev->dev, "vddcx");
	DBG("gpu_cx: %p", gpu->gpu_cx);
	if (IS_ERR(gpu->gpu_cx))
		gpu->gpu_cx = NULL;

	gpu->pdev = pdev;
	platform_set_drvdata(pdev, gpu);

	msm_devfreq_init(gpu);

	gpu->aspace = msm_gpu_create_address_space(gpu, pdev,
		config->va_start, config->va_end);

	if (gpu->aspace == NULL)
		DRM_DEV_INFO(drm->dev, "%s: no IOMMU, fallback to VRAM carveout!\n", name);
	else if (IS_ERR(gpu->aspace)) {
		ret = PTR_ERR(gpu->aspace);
		goto fail;
	}

	memptrs = msm_gem_kernel_new(drm,
		sizeof(struct msm_rbmemptrs) * nr_rings,
		MSM_BO_UNCACHED, gpu->aspace, &gpu->memptrs_bo,
		&memptrs_iova);

	if (IS_ERR(memptrs)) {
		ret = PTR_ERR(memptrs);
		DRM_DEV_ERROR(drm->dev, "could not allocate memptrs: %d\n", ret);
		goto fail;
	}

	msm_gem_object_set_name(gpu->memptrs_bo, "memptrs");

	if (nr_rings > ARRAY_SIZE(gpu->rb)) {
		DRM_DEV_INFO_ONCE(drm->dev, "Only creating %zu ringbuffers\n",
			ARRAY_SIZE(gpu->rb));
		nr_rings = ARRAY_SIZE(gpu->rb);
	}

	/* Create ringbuffer(s): */
	for (i = 0; i < nr_rings; i++) {
		gpu->rb[i] = msm_ringbuffer_new(gpu, i, memptrs, memptrs_iova);

		if (IS_ERR(gpu->rb[i])) {
			ret = PTR_ERR(gpu->rb[i]);
			DRM_DEV_ERROR(drm->dev,
				"could not create ringbuffer %d: %d\n", i, ret);
			goto fail;
		}

		memptrs += sizeof(struct msm_rbmemptrs);
		memptrs_iova += sizeof(struct msm_rbmemptrs);
	}

	gpu->nr_rings = nr_rings;

	return 0;

fail:
	for (i = 0; i < ARRAY_SIZE(gpu->rb); i++)  {
		msm_ringbuffer_destroy(gpu->rb[i]);
		gpu->rb[i] = NULL;
	}

	msm_gem_kernel_put(gpu->memptrs_bo, gpu->aspace, false);

	platform_set_drvdata(pdev, NULL);
	return ret;
}

void msm_gpu_cleanup(struct msm_gpu *gpu)
{
	int i;

	DBG("%s", gpu->name);

	WARN_ON(!list_empty(&gpu->active_list));

	for (i = 0; i < ARRAY_SIZE(gpu->rb); i++) {
		msm_ringbuffer_destroy(gpu->rb[i]);
		gpu->rb[i] = NULL;
	}

	msm_gem_kernel_put(gpu->memptrs_bo, gpu->aspace, false);

	if (!IS_ERR_OR_NULL(gpu->aspace)) {
		gpu->aspace->mmu->funcs->detach(gpu->aspace->mmu,
			NULL, 0);
		msm_gem_address_space_put(gpu->aspace);
	}
}