summaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
blob: bcc1cbeb8799713512e57da7671a9f1b795c10f7 (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
/*
 * Copyright 2014-2018 Advanced Micro Devices, Inc.
 *
 * Permission is hereby granted, free of charge, to any person obtaining a
 * copy of this software and associated documentation files (the "Software"),
 * to deal in the Software without restriction, including without limitation
 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
 * and/or sell copies of the Software, and to permit persons to whom the
 * Software is furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in
 * all copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
 * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
 * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
 * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
 * OTHER DEALINGS IN THE SOFTWARE.
 */
#include "amdgpu.h"
#include "amdgpu_amdkfd.h"
#include "gc/gc_9_0_offset.h"
#include "gc/gc_9_0_sh_mask.h"
#include "vega10_enum.h"
#include "sdma0/sdma0_4_0_offset.h"
#include "sdma0/sdma0_4_0_sh_mask.h"
#include "sdma1/sdma1_4_0_offset.h"
#include "sdma1/sdma1_4_0_sh_mask.h"
#include "athub/athub_1_0_offset.h"
#include "athub/athub_1_0_sh_mask.h"
#include "oss/osssys_4_0_offset.h"
#include "oss/osssys_4_0_sh_mask.h"
#include "soc15_common.h"
#include "v9_structs.h"
#include "soc15.h"
#include "soc15d.h"
#include "gfx_v9_0.h"
#include "amdgpu_amdkfd_gfx_v9.h"

enum hqd_dequeue_request_type {
	NO_ACTION = 0,
	DRAIN_PIPE,
	RESET_WAVES,
	SAVE_WAVES
};

static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd)
{
	return (struct amdgpu_device *)kgd;
}

static void lock_srbm(struct kgd_dev *kgd, uint32_t mec, uint32_t pipe,
			uint32_t queue, uint32_t vmid)
{
	struct amdgpu_device *adev = get_amdgpu_device(kgd);

	mutex_lock(&adev->srbm_mutex);
	soc15_grbm_select(adev, mec, pipe, queue, vmid);
}

static void unlock_srbm(struct kgd_dev *kgd)
{
	struct amdgpu_device *adev = get_amdgpu_device(kgd);

	soc15_grbm_select(adev, 0, 0, 0, 0);
	mutex_unlock(&adev->srbm_mutex);
}

static void acquire_queue(struct kgd_dev *kgd, uint32_t pipe_id,
				uint32_t queue_id)
{
	struct amdgpu_device *adev = get_amdgpu_device(kgd);

	uint32_t mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
	uint32_t pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);

	lock_srbm(kgd, mec, pipe, queue_id, 0);
}

static uint64_t get_queue_mask(struct amdgpu_device *adev,
			       uint32_t pipe_id, uint32_t queue_id)
{
	unsigned int bit = pipe_id * adev->gfx.mec.num_queue_per_pipe +
			queue_id;

	return 1ull << bit;
}

static void release_queue(struct kgd_dev *kgd)
{
	unlock_srbm(kgd);
}

void kgd_gfx_v9_program_sh_mem_settings(struct kgd_dev *kgd, uint32_t vmid,
					uint32_t sh_mem_config,
					uint32_t sh_mem_ape1_base,
					uint32_t sh_mem_ape1_limit,
					uint32_t sh_mem_bases)
{
	struct amdgpu_device *adev = get_amdgpu_device(kgd);

	lock_srbm(kgd, 0, 0, 0, vmid);

	WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmSH_MEM_CONFIG), sh_mem_config);
	WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmSH_MEM_BASES), sh_mem_bases);
	/* APE1 no longer exists on GFX9 */

	unlock_srbm(kgd);
}

int kgd_gfx_v9_set_pasid_vmid_mapping(struct kgd_dev *kgd, u32 pasid,
					unsigned int vmid)
{
	struct amdgpu_device *adev = get_amdgpu_device(kgd);

	/*
	 * We have to assume that there is no outstanding mapping.
	 * The ATC_VMID_PASID_MAPPING_UPDATE_STATUS bit could be 0 because
	 * a mapping is in progress or because a mapping finished
	 * and the SW cleared it.
	 * So the protocol is to always wait & clear.
	 */
	uint32_t pasid_mapping = (pasid == 0) ? 0 : (uint32_t)pasid |
			ATC_VMID0_PASID_MAPPING__VALID_MASK;

	/*
	 * need to do this twice, once for gfx and once for mmhub
	 * for ATC add 16 to VMID for mmhub, for IH different registers.
	 * ATC_VMID0..15 registers are separate from ATC_VMID16..31.
	 */

	WREG32(SOC15_REG_OFFSET(ATHUB, 0, mmATC_VMID0_PASID_MAPPING) + vmid,
	       pasid_mapping);

	while (!(RREG32(SOC15_REG_OFFSET(
				ATHUB, 0,
				mmATC_VMID_PASID_MAPPING_UPDATE_STATUS)) &
		 (1U << vmid)))
		cpu_relax();

	WREG32(SOC15_REG_OFFSET(ATHUB, 0,
				mmATC_VMID_PASID_MAPPING_UPDATE_STATUS),
	       1U << vmid);

	/* Mapping vmid to pasid also for IH block */
	WREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_VMID_0_LUT) + vmid,
	       pasid_mapping);

	WREG32(SOC15_REG_OFFSET(ATHUB, 0, mmATC_VMID16_PASID_MAPPING) + vmid,
	       pasid_mapping);

	while (!(RREG32(SOC15_REG_OFFSET(
				ATHUB, 0,
				mmATC_VMID_PASID_MAPPING_UPDATE_STATUS)) &
		 (1U << (vmid + 16))))
		cpu_relax();

	WREG32(SOC15_REG_OFFSET(ATHUB, 0,
				mmATC_VMID_PASID_MAPPING_UPDATE_STATUS),
	       1U << (vmid + 16));

	/* Mapping vmid to pasid also for IH block */
	WREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_VMID_0_LUT_MM) + vmid,
	       pasid_mapping);
	return 0;
}

/* TODO - RING0 form of field is obsolete, seems to date back to SI
 * but still works
 */

int kgd_gfx_v9_init_interrupts(struct kgd_dev *kgd, uint32_t pipe_id)
{
	struct amdgpu_device *adev = get_amdgpu_device(kgd);
	uint32_t mec;
	uint32_t pipe;

	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);

	lock_srbm(kgd, mec, pipe, 0, 0);

	WREG32(SOC15_REG_OFFSET(GC, 0, mmCPC_INT_CNTL),
		CP_INT_CNTL_RING0__TIME_STAMP_INT_ENABLE_MASK |
		CP_INT_CNTL_RING0__OPCODE_ERROR_INT_ENABLE_MASK);

	unlock_srbm(kgd);

	return 0;
}

static uint32_t get_sdma_rlc_reg_offset(struct amdgpu_device *adev,
				unsigned int engine_id,
				unsigned int queue_id)
{
	uint32_t sdma_engine_reg_base = 0;
	uint32_t sdma_rlc_reg_offset;

	switch (engine_id) {
	default:
		dev_warn(adev->dev,
			 "Invalid sdma engine id (%d), using engine id 0\n",
			 engine_id);
		fallthrough;
	case 0:
		sdma_engine_reg_base = SOC15_REG_OFFSET(SDMA0, 0,
				mmSDMA0_RLC0_RB_CNTL) - mmSDMA0_RLC0_RB_CNTL;
		break;
	case 1:
		sdma_engine_reg_base = SOC15_REG_OFFSET(SDMA1, 0,
				mmSDMA1_RLC0_RB_CNTL) - mmSDMA0_RLC0_RB_CNTL;
		break;
	}

	sdma_rlc_reg_offset = sdma_engine_reg_base
		+ queue_id * (mmSDMA0_RLC1_RB_CNTL - mmSDMA0_RLC0_RB_CNTL);

	pr_debug("RLC register offset for SDMA%d RLC%d: 0x%x\n", engine_id,
		 queue_id, sdma_rlc_reg_offset);

	return sdma_rlc_reg_offset;
}

static inline struct v9_mqd *get_mqd(void *mqd)
{
	return (struct v9_mqd *)mqd;
}

static inline struct v9_sdma_mqd *get_sdma_mqd(void *mqd)
{
	return (struct v9_sdma_mqd *)mqd;
}

int kgd_gfx_v9_hqd_load(struct kgd_dev *kgd, void *mqd, uint32_t pipe_id,
			uint32_t queue_id, uint32_t __user *wptr,
			uint32_t wptr_shift, uint32_t wptr_mask,
			struct mm_struct *mm)
{
	struct amdgpu_device *adev = get_amdgpu_device(kgd);
	struct v9_mqd *m;
	uint32_t *mqd_hqd;
	uint32_t reg, hqd_base, data;

	m = get_mqd(mqd);

	acquire_queue(kgd, pipe_id, queue_id);

	/* HQD registers extend from CP_MQD_BASE_ADDR to CP_HQD_EOP_WPTR_MEM. */
	mqd_hqd = &m->cp_mqd_base_addr_lo;
	hqd_base = SOC15_REG_OFFSET(GC, 0, mmCP_MQD_BASE_ADDR);

	for (reg = hqd_base;
	     reg <= SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI); reg++)
		WREG32_RLC(reg, mqd_hqd[reg - hqd_base]);


	/* Activate doorbell logic before triggering WPTR poll. */
	data = REG_SET_FIELD(m->cp_hqd_pq_doorbell_control,
			     CP_HQD_PQ_DOORBELL_CONTROL, DOORBELL_EN, 1);
	WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL), data);

	if (wptr) {
		/* Don't read wptr with get_user because the user
		 * context may not be accessible (if this function
		 * runs in a work queue). Instead trigger a one-shot
		 * polling read from memory in the CP. This assumes
		 * that wptr is GPU-accessible in the queue's VMID via
		 * ATC or SVM. WPTR==RPTR before starting the poll so
		 * the CP starts fetching new commands from the right
		 * place.
		 *
		 * Guessing a 64-bit WPTR from a 32-bit RPTR is a bit
		 * tricky. Assume that the queue didn't overflow. The
		 * number of valid bits in the 32-bit RPTR depends on
		 * the queue size. The remaining bits are taken from
		 * the saved 64-bit WPTR. If the WPTR wrapped, add the
		 * queue size.
		 */
		uint32_t queue_size =
			2 << REG_GET_FIELD(m->cp_hqd_pq_control,
					   CP_HQD_PQ_CONTROL, QUEUE_SIZE);
		uint64_t guessed_wptr = m->cp_hqd_pq_rptr & (queue_size - 1);

		if ((m->cp_hqd_pq_wptr_lo & (queue_size - 1)) < guessed_wptr)
			guessed_wptr += queue_size;
		guessed_wptr += m->cp_hqd_pq_wptr_lo & ~(queue_size - 1);
		guessed_wptr += (uint64_t)m->cp_hqd_pq_wptr_hi << 32;

		WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_LO),
		       lower_32_bits(guessed_wptr));
		WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI),
		       upper_32_bits(guessed_wptr));
		WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR),
		       lower_32_bits((uintptr_t)wptr));
		WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR_HI),
		       upper_32_bits((uintptr_t)wptr));
		WREG32(SOC15_REG_OFFSET(GC, 0, mmCP_PQ_WPTR_POLL_CNTL1),
		       (uint32_t)get_queue_mask(adev, pipe_id, queue_id));
	}

	/* Start the EOP fetcher */
	WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_EOP_RPTR),
	       REG_SET_FIELD(m->cp_hqd_eop_rptr,
			     CP_HQD_EOP_RPTR, INIT_FETCHER, 1));

	data = REG_SET_FIELD(m->cp_hqd_active, CP_HQD_ACTIVE, ACTIVE, 1);
	WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_ACTIVE), data);

	release_queue(kgd);

	return 0;
}

int kgd_gfx_v9_hiq_mqd_load(struct kgd_dev *kgd, void *mqd,
			    uint32_t pipe_id, uint32_t queue_id,
			    uint32_t doorbell_off)
{
	struct amdgpu_device *adev = get_amdgpu_device(kgd);
	struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring;
	struct v9_mqd *m;
	uint32_t mec, pipe;
	int r;

	m = get_mqd(mqd);

	acquire_queue(kgd, pipe_id, queue_id);

	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);

	pr_debug("kfd: set HIQ, mec:%d, pipe:%d, queue:%d.\n",
		 mec, pipe, queue_id);

	spin_lock(&adev->gfx.kiq.ring_lock);
	r = amdgpu_ring_alloc(kiq_ring, 7);
	if (r) {
		pr_err("Failed to alloc KIQ (%d).\n", r);
		goto out_unlock;
	}

	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_MAP_QUEUES, 5));
	amdgpu_ring_write(kiq_ring,
			  PACKET3_MAP_QUEUES_QUEUE_SEL(0) | /* Queue_Sel */
			  PACKET3_MAP_QUEUES_VMID(m->cp_hqd_vmid) | /* VMID */
			  PACKET3_MAP_QUEUES_QUEUE(queue_id) |
			  PACKET3_MAP_QUEUES_PIPE(pipe) |
			  PACKET3_MAP_QUEUES_ME((mec - 1)) |
			  PACKET3_MAP_QUEUES_QUEUE_TYPE(0) | /*queue_type: normal compute queue */
			  PACKET3_MAP_QUEUES_ALLOC_FORMAT(0) | /* alloc format: all_on_one_pipe */
			  PACKET3_MAP_QUEUES_ENGINE_SEL(1) | /* engine_sel: hiq */
			  PACKET3_MAP_QUEUES_NUM_QUEUES(1)); /* num_queues: must be 1 */
	amdgpu_ring_write(kiq_ring,
			  PACKET3_MAP_QUEUES_DOORBELL_OFFSET(doorbell_off));
	amdgpu_ring_write(kiq_ring, m->cp_mqd_base_addr_lo);
	amdgpu_ring_write(kiq_ring, m->cp_mqd_base_addr_hi);
	amdgpu_ring_write(kiq_ring, m->cp_hqd_pq_wptr_poll_addr_lo);
	amdgpu_ring_write(kiq_ring, m->cp_hqd_pq_wptr_poll_addr_hi);
	amdgpu_ring_commit(kiq_ring);

out_unlock:
	spin_unlock(&adev->gfx.kiq.ring_lock);
	release_queue(kgd);

	return r;
}

int kgd_gfx_v9_hqd_dump(struct kgd_dev *kgd,
			uint32_t pipe_id, uint32_t queue_id,
			uint32_t (**dump)[2], uint32_t *n_regs)
{
	struct amdgpu_device *adev = get_amdgpu_device(kgd);
	uint32_t i = 0, reg;
#define HQD_N_REGS 56
#define DUMP_REG(addr) do {				\
		if (WARN_ON_ONCE(i >= HQD_N_REGS))	\
			break;				\
		(*dump)[i][0] = (addr) << 2;		\
		(*dump)[i++][1] = RREG32(addr);		\
	} while (0)

	*dump = kmalloc_array(HQD_N_REGS * 2, sizeof(uint32_t), GFP_KERNEL);
	if (*dump == NULL)
		return -ENOMEM;

	acquire_queue(kgd, pipe_id, queue_id);

	for (reg = SOC15_REG_OFFSET(GC, 0, mmCP_MQD_BASE_ADDR);
	     reg <= SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI); reg++)
		DUMP_REG(reg);

	release_queue(kgd);

	WARN_ON_ONCE(i != HQD_N_REGS);
	*n_regs = i;

	return 0;
}

static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
			     uint32_t __user *wptr, struct mm_struct *mm)
{
	struct amdgpu_device *adev = get_amdgpu_device(kgd);
	struct v9_sdma_mqd *m;
	uint32_t sdma_rlc_reg_offset;
	unsigned long end_jiffies;
	uint32_t data;
	uint64_t data64;
	uint64_t __user *wptr64 = (uint64_t __user *)wptr;

	m = get_sdma_mqd(mqd);
	sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev, m->sdma_engine_id,
					    m->sdma_queue_id);

	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_CNTL,
		m->sdmax_rlcx_rb_cntl & (~SDMA0_RLC0_RB_CNTL__RB_ENABLE_MASK));

	end_jiffies = msecs_to_jiffies(2000) + jiffies;
	while (true) {
		data = RREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_CONTEXT_STATUS);
		if (data & SDMA0_RLC0_CONTEXT_STATUS__IDLE_MASK)
			break;
		if (time_after(jiffies, end_jiffies)) {
			pr_err("SDMA RLC not idle in %s\n", __func__);
			return -ETIME;
		}
		usleep_range(500, 1000);
	}

	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_DOORBELL_OFFSET,
	       m->sdmax_rlcx_doorbell_offset);

	data = REG_SET_FIELD(m->sdmax_rlcx_doorbell, SDMA0_RLC0_DOORBELL,
			     ENABLE, 1);
	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_DOORBELL, data);
	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_RPTR,
				m->sdmax_rlcx_rb_rptr);
	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_RPTR_HI,
				m->sdmax_rlcx_rb_rptr_hi);

	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_MINOR_PTR_UPDATE, 1);
	if (read_user_wptr(mm, wptr64, data64)) {
		WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_WPTR,
		       lower_32_bits(data64));
		WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_WPTR_HI,
		       upper_32_bits(data64));
	} else {
		WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_WPTR,
		       m->sdmax_rlcx_rb_rptr);
		WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_WPTR_HI,
		       m->sdmax_rlcx_rb_rptr_hi);
	}
	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_MINOR_PTR_UPDATE, 0);

	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_BASE, m->sdmax_rlcx_rb_base);
	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_BASE_HI,
			m->sdmax_rlcx_rb_base_hi);
	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_RPTR_ADDR_LO,
			m->sdmax_rlcx_rb_rptr_addr_lo);
	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_RPTR_ADDR_HI,
			m->sdmax_rlcx_rb_rptr_addr_hi);

	data = REG_SET_FIELD(m->sdmax_rlcx_rb_cntl, SDMA0_RLC0_RB_CNTL,
			     RB_ENABLE, 1);
	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_CNTL, data);

	return 0;
}

static int kgd_hqd_sdma_dump(struct kgd_dev *kgd,
			     uint32_t engine_id, uint32_t queue_id,
			     uint32_t (**dump)[2], uint32_t *n_regs)
{
	struct amdgpu_device *adev = get_amdgpu_device(kgd);
	uint32_t sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev,
			engine_id, queue_id);
	uint32_t i = 0, reg;
#undef HQD_N_REGS
#define HQD_N_REGS (19+6+7+10)

	*dump = kmalloc_array(HQD_N_REGS * 2, sizeof(uint32_t), GFP_KERNEL);
	if (*dump == NULL)
		return -ENOMEM;

	for (reg = mmSDMA0_RLC0_RB_CNTL; reg <= mmSDMA0_RLC0_DOORBELL; reg++)
		DUMP_REG(sdma_rlc_reg_offset + reg);
	for (reg = mmSDMA0_RLC0_STATUS; reg <= mmSDMA0_RLC0_CSA_ADDR_HI; reg++)
		DUMP_REG(sdma_rlc_reg_offset + reg);
	for (reg = mmSDMA0_RLC0_IB_SUB_REMAIN;
	     reg <= mmSDMA0_RLC0_MINOR_PTR_UPDATE; reg++)
		DUMP_REG(sdma_rlc_reg_offset + reg);
	for (reg = mmSDMA0_RLC0_MIDCMD_DATA0;
	     reg <= mmSDMA0_RLC0_MIDCMD_CNTL; reg++)
		DUMP_REG(sdma_rlc_reg_offset + reg);

	WARN_ON_ONCE(i != HQD_N_REGS);
	*n_regs = i;

	return 0;
}

bool kgd_gfx_v9_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address,
				uint32_t pipe_id, uint32_t queue_id)
{
	struct amdgpu_device *adev = get_amdgpu_device(kgd);
	uint32_t act;
	bool retval = false;
	uint32_t low, high;

	acquire_queue(kgd, pipe_id, queue_id);
	act = RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_ACTIVE));
	if (act) {
		low = lower_32_bits(queue_address >> 8);
		high = upper_32_bits(queue_address >> 8);

		if (low == RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_BASE)) &&
		   high == RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_BASE_HI)))
			retval = true;
	}
	release_queue(kgd);
	return retval;
}

static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
{
	struct amdgpu_device *adev = get_amdgpu_device(kgd);
	struct v9_sdma_mqd *m;
	uint32_t sdma_rlc_reg_offset;
	uint32_t sdma_rlc_rb_cntl;

	m = get_sdma_mqd(mqd);
	sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev, m->sdma_engine_id,
					    m->sdma_queue_id);

	sdma_rlc_rb_cntl = RREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_CNTL);

	if (sdma_rlc_rb_cntl & SDMA0_RLC0_RB_CNTL__RB_ENABLE_MASK)
		return true;

	return false;
}

int kgd_gfx_v9_hqd_destroy(struct kgd_dev *kgd, void *mqd,
				enum kfd_preempt_type reset_type,
				unsigned int utimeout, uint32_t pipe_id,
				uint32_t queue_id)
{
	struct amdgpu_device *adev = get_amdgpu_device(kgd);
	enum hqd_dequeue_request_type type;
	unsigned long end_jiffies;
	uint32_t temp;
	struct v9_mqd *m = get_mqd(mqd);

	if (amdgpu_in_reset(adev))
		return -EIO;

	acquire_queue(kgd, pipe_id, queue_id);

	if (m->cp_hqd_vmid == 0)
		WREG32_FIELD15_RLC(GC, 0, RLC_CP_SCHEDULERS, scheduler1, 0);

	switch (reset_type) {
	case KFD_PREEMPT_TYPE_WAVEFRONT_DRAIN:
		type = DRAIN_PIPE;
		break;
	case KFD_PREEMPT_TYPE_WAVEFRONT_RESET:
		type = RESET_WAVES;
		break;
	case KFD_PREEMPT_TYPE_WAVEFRONT_SAVE:
		type = SAVE_WAVES;
		break;
	default:
		type = DRAIN_PIPE;
		break;
	}

	WREG32_RLC(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_DEQUEUE_REQUEST), type);

	end_jiffies = (utimeout * HZ / 1000) + jiffies;
	while (true) {
		temp = RREG32(SOC15_REG_OFFSET(GC, 0, mmCP_HQD_ACTIVE));
		if (!(temp & CP_HQD_ACTIVE__ACTIVE_MASK))
			break;
		if (time_after(jiffies, end_jiffies)) {
			pr_err("cp queue preemption time out.\n");
			release_queue(kgd);
			return -ETIME;
		}
		usleep_range(500, 1000);
	}

	release_queue(kgd);
	return 0;
}

static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
				unsigned int utimeout)
{
	struct amdgpu_device *adev = get_amdgpu_device(kgd);
	struct v9_sdma_mqd *m;
	uint32_t sdma_rlc_reg_offset;
	uint32_t temp;
	unsigned long end_jiffies = (utimeout * HZ / 1000) + jiffies;

	m = get_sdma_mqd(mqd);
	sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev, m->sdma_engine_id,
					    m->sdma_queue_id);

	temp = RREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_CNTL);
	temp = temp & ~SDMA0_RLC0_RB_CNTL__RB_ENABLE_MASK;
	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_CNTL, temp);

	while (true) {
		temp = RREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_CONTEXT_STATUS);
		if (temp & SDMA0_RLC0_CONTEXT_STATUS__IDLE_MASK)
			break;
		if (time_after(jiffies, end_jiffies)) {
			pr_err("SDMA RLC not idle in %s\n", __func__);
			return -ETIME;
		}
		usleep_range(500, 1000);
	}

	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_DOORBELL, 0);
	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_CNTL,
		RREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_CNTL) |
		SDMA0_RLC0_RB_CNTL__RB_ENABLE_MASK);

	m->sdmax_rlcx_rb_rptr = RREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_RPTR);
	m->sdmax_rlcx_rb_rptr_hi =
		RREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_RPTR_HI);

	return 0;
}

bool kgd_gfx_v9_get_atc_vmid_pasid_mapping_info(struct kgd_dev *kgd,
					uint8_t vmid, uint16_t *p_pasid)
{
	uint32_t value;
	struct amdgpu_device *adev = (struct amdgpu_device *) kgd;

	value = RREG32(SOC15_REG_OFFSET(ATHUB, 0, mmATC_VMID0_PASID_MAPPING)
		     + vmid);
	*p_pasid = value & ATC_VMID0_PASID_MAPPING__PASID_MASK;

	return !!(value & ATC_VMID0_PASID_MAPPING__VALID_MASK);
}

int kgd_gfx_v9_address_watch_disable(struct kgd_dev *kgd)
{
	return 0;
}

int kgd_gfx_v9_address_watch_execute(struct kgd_dev *kgd,
					unsigned int watch_point_id,
					uint32_t cntl_val,
					uint32_t addr_hi,
					uint32_t addr_lo)
{
	return 0;
}

int kgd_gfx_v9_wave_control_execute(struct kgd_dev *kgd,
					uint32_t gfx_index_val,
					uint32_t sq_cmd)
{
	struct amdgpu_device *adev = get_amdgpu_device(kgd);
	uint32_t data = 0;

	mutex_lock(&adev->grbm_idx_mutex);

	WREG32_SOC15_RLC_SHADOW(GC, 0, mmGRBM_GFX_INDEX, gfx_index_val);
	WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_CMD), sq_cmd);

	data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
		INSTANCE_BROADCAST_WRITES, 1);
	data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
		SH_BROADCAST_WRITES, 1);
	data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
		SE_BROADCAST_WRITES, 1);

	WREG32_SOC15_RLC_SHADOW(GC, 0, mmGRBM_GFX_INDEX, data);
	mutex_unlock(&adev->grbm_idx_mutex);

	return 0;
}

uint32_t kgd_gfx_v9_address_watch_get_offset(struct kgd_dev *kgd,
					unsigned int watch_point_id,
					unsigned int reg_offset)
{
	return 0;
}

void kgd_gfx_v9_set_vm_context_page_table_base(struct kgd_dev *kgd,
			uint32_t vmid, uint64_t page_table_base)
{
	struct amdgpu_device *adev = get_amdgpu_device(kgd);

	if (!amdgpu_amdkfd_is_kfd_vmid(adev, vmid)) {
		pr_err("trying to set page table base for wrong VMID %u\n",
		       vmid);
		return;
	}

	adev->mmhub.funcs->setup_vm_pt_regs(adev, vmid, page_table_base);

	adev->gfxhub.funcs->setup_vm_pt_regs(adev, vmid, page_table_base);
}

static void lock_spi_csq_mutexes(struct amdgpu_device *adev)
{
	mutex_lock(&adev->srbm_mutex);
	mutex_lock(&adev->grbm_idx_mutex);

}

static void unlock_spi_csq_mutexes(struct amdgpu_device *adev)
{
	mutex_unlock(&adev->grbm_idx_mutex);
	mutex_unlock(&adev->srbm_mutex);
}

/**
 * get_wave_count: Read device registers to get number of waves in flight for
 * a particular queue. The method also returns the VMID associated with the
 * queue.
 *
 * @adev: Handle of device whose registers are to be read
 * @queue_idx: Index of queue in the queue-map bit-field
 * @wave_cnt: Output parameter updated with number of waves in flight
 * @vmid: Output parameter updated with VMID of queue whose wave count
 * is being collected
 */
static void get_wave_count(struct amdgpu_device *adev, int queue_idx,
		int *wave_cnt, int *vmid)
{
	int pipe_idx;
	int queue_slot;
	unsigned int reg_val;

	/*
	 * Program GRBM with appropriate MEID, PIPEID, QUEUEID and VMID
	 * parameters to read out waves in flight. Get VMID if there are
	 * non-zero waves in flight.
	 */
	*vmid = 0xFF;
	*wave_cnt = 0;
	pipe_idx = queue_idx / adev->gfx.mec.num_queue_per_pipe;
	queue_slot = queue_idx % adev->gfx.mec.num_queue_per_pipe;
	soc15_grbm_select(adev, 1, pipe_idx, queue_slot, 0);
	reg_val = RREG32(SOC15_REG_OFFSET(GC, 0, mmSPI_CSQ_WF_ACTIVE_COUNT_0) +
			 queue_slot);
	*wave_cnt = reg_val & SPI_CSQ_WF_ACTIVE_COUNT_0__COUNT_MASK;
	if (*wave_cnt != 0)
		*vmid = (RREG32_SOC15(GC, 0, mmCP_HQD_VMID) &
			 CP_HQD_VMID__VMID_MASK) >> CP_HQD_VMID__VMID__SHIFT;
}

/**
 * kgd_gfx_v9_get_cu_occupancy: Reads relevant registers associated with each
 * shader engine and aggregates the number of waves that are in flight for the
 * process whose pasid is provided as a parameter. The process could have ZERO
 * or more queues running and submitting waves to compute units.
 *
 * @kgd: Handle of device from which to get number of waves in flight
 * @pasid: Identifies the process for which this query call is invoked
 * @pasid_wave_cnt: Output parameter updated with number of waves in flight that
 * belong to process with given pasid
 * @max_waves_per_cu: Output parameter updated with maximum number of waves
 * possible per Compute Unit
 *
 * Note: It's possible that the device has too many queues (oversubscription)
 * in which case a VMID could be remapped to a different PASID. This could lead
 * to an iaccurate wave count. Following is a high-level sequence:
 *    Time T1: vmid = getVmid(); vmid is associated with Pasid P1
 *    Time T2: passId = getPasId(vmid); vmid is associated with Pasid P2
 * In the sequence above wave count obtained from time T1 will be incorrectly
 * lost or added to total wave count.
 *
 * The registers that provide the waves in flight are:
 *
 *  SPI_CSQ_WF_ACTIVE_STATUS - bit-map of queues per pipe. The bit is ON if a
 *  queue is slotted, OFF if there is no queue. A process could have ZERO or
 *  more queues slotted and submitting waves to be run on compute units. Even
 *  when there is a queue it is possible there could be zero wave fronts, this
 *  can happen when queue is waiting on top-of-pipe events - e.g. waitRegMem
 *  command
 *
 *  For each bit that is ON from above:
 *
 *    Read (SPI_CSQ_WF_ACTIVE_COUNT_0 + queue_idx) register. It provides the
 *    number of waves that are in flight for the queue at specified index. The
 *    index ranges from 0 to 7.
 *
 *    If non-zero waves are in flight, read CP_HQD_VMID register to obtain VMID
 *    of the wave(s).
 *
 *    Determine if VMID from above step maps to pasid provided as parameter. If
 *    it matches agrregate the wave count. That the VMID will not match pasid is
 *    a normal condition i.e. a device is expected to support multiple queues
 *    from multiple proceses.
 *
 *  Reading registers referenced above involves programming GRBM appropriately
 */
void kgd_gfx_v9_get_cu_occupancy(struct kgd_dev *kgd, int pasid,
		int *pasid_wave_cnt, int *max_waves_per_cu)
{
	int qidx;
	int vmid;
	int se_idx;
	int sh_idx;
	int se_cnt;
	int sh_cnt;
	int wave_cnt;
	int queue_map;
	int pasid_tmp;
	int max_queue_cnt;
	int vmid_wave_cnt = 0;
	struct amdgpu_device *adev;
	DECLARE_BITMAP(cp_queue_bitmap, KGD_MAX_QUEUES);

	adev = get_amdgpu_device(kgd);
	lock_spi_csq_mutexes(adev);
	soc15_grbm_select(adev, 1, 0, 0, 0);

	/*
	 * Iterate through the shader engines and arrays of the device
	 * to get number of waves in flight
	 */
	bitmap_complement(cp_queue_bitmap, adev->gfx.mec.queue_bitmap,
			  KGD_MAX_QUEUES);
	max_queue_cnt = adev->gfx.mec.num_pipe_per_mec *
			adev->gfx.mec.num_queue_per_pipe;
	sh_cnt = adev->gfx.config.max_sh_per_se;
	se_cnt = adev->gfx.config.max_shader_engines;
	for (se_idx = 0; se_idx < se_cnt; se_idx++) {
		for (sh_idx = 0; sh_idx < sh_cnt; sh_idx++) {

			gfx_v9_0_select_se_sh(adev, se_idx, sh_idx, 0xffffffff);
			queue_map = RREG32(SOC15_REG_OFFSET(GC, 0,
					   mmSPI_CSQ_WF_ACTIVE_STATUS));

			/*
			 * Assumption: queue map encodes following schema: four
			 * pipes per each micro-engine, with each pipe mapping
			 * eight queues. This schema is true for GFX9 devices
			 * and must be verified for newer device families
			 */
			for (qidx = 0; qidx < max_queue_cnt; qidx++) {

				/* Skip qeueus that are not associated with
				 * compute functions
				 */
				if (!test_bit(qidx, cp_queue_bitmap))
					continue;

				if (!(queue_map & (1 << qidx)))
					continue;

				/* Get number of waves in flight and aggregate them */
				get_wave_count(adev, qidx, &wave_cnt, &vmid);
				if (wave_cnt != 0) {
					pasid_tmp =
					  RREG32(SOC15_REG_OFFSET(OSSSYS, 0,
						 mmIH_VMID_0_LUT) + vmid);
					if (pasid_tmp == pasid)
						vmid_wave_cnt += wave_cnt;
				}
			}
		}
	}

	gfx_v9_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
	soc15_grbm_select(adev, 0, 0, 0, 0);
	unlock_spi_csq_mutexes(adev);

	/* Update the output parameters and return */
	*pasid_wave_cnt = vmid_wave_cnt;
	*max_waves_per_cu = adev->gfx.cu_info.simd_per_cu *
				adev->gfx.cu_info.max_waves_per_simd;
}

void kgd_gfx_v9_program_trap_handler_settings(struct kgd_dev *kgd,
                        uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr)
{
	struct amdgpu_device *adev = get_amdgpu_device(kgd);

	lock_srbm(kgd, 0, 0, 0, vmid);

	/*
	 * Program TBA registers
	 */
	WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_SHADER_TBA_LO),
                        lower_32_bits(tba_addr >> 8));
	WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_SHADER_TBA_HI),
                        upper_32_bits(tba_addr >> 8));

	/*
	 * Program TMA registers
	 */
	WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_SHADER_TMA_LO),
			lower_32_bits(tma_addr >> 8));
	WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_SHADER_TMA_HI),
			upper_32_bits(tma_addr >> 8));

	unlock_srbm(kgd);
}

const struct kfd2kgd_calls gfx_v9_kfd2kgd = {
	.program_sh_mem_settings = kgd_gfx_v9_program_sh_mem_settings,
	.set_pasid_vmid_mapping = kgd_gfx_v9_set_pasid_vmid_mapping,
	.init_interrupts = kgd_gfx_v9_init_interrupts,
	.hqd_load = kgd_gfx_v9_hqd_load,
	.hiq_mqd_load = kgd_gfx_v9_hiq_mqd_load,
	.hqd_sdma_load = kgd_hqd_sdma_load,
	.hqd_dump = kgd_gfx_v9_hqd_dump,
	.hqd_sdma_dump = kgd_hqd_sdma_dump,
	.hqd_is_occupied = kgd_gfx_v9_hqd_is_occupied,
	.hqd_sdma_is_occupied = kgd_hqd_sdma_is_occupied,
	.hqd_destroy = kgd_gfx_v9_hqd_destroy,
	.hqd_sdma_destroy = kgd_hqd_sdma_destroy,
	.address_watch_disable = kgd_gfx_v9_address_watch_disable,
	.address_watch_execute = kgd_gfx_v9_address_watch_execute,
	.wave_control_execute = kgd_gfx_v9_wave_control_execute,
	.address_watch_get_offset = kgd_gfx_v9_address_watch_get_offset,
	.get_atc_vmid_pasid_mapping_info =
			kgd_gfx_v9_get_atc_vmid_pasid_mapping_info,
	.set_vm_context_page_table_base = kgd_gfx_v9_set_vm_context_page_table_base,
	.get_cu_occupancy = kgd_gfx_v9_get_cu_occupancy,
	.program_trap_handler_settings = kgd_gfx_v9_program_trap_handler_settings,
};