summaryrefslogtreecommitdiffstats
path: root/drivers/gpu/nvgpu/gk20a/platform_gk20a_tegra.c
blob: 2c9cfb63c301f636b8973f047016c732477f723c (plain) (blame)
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
/*
 * drivers/video/tegra/host/gk20a/platform_gk20a_tegra.c
 *
 * GK20A Tegra Platform Interface
 *
 * Copyright (c) 2014-2015, NVIDIA CORPORATION.  All rights reserved.
 *
 * This program is free software; you can redistribute it and/or modify it
 * under the terms and conditions of the GNU General Public License,
 * version 2, as published by the Free Software Foundation.
 *
 * This program is distributed in the hope it will be useful, but WITHOUT
 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
 * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
 * more details.
 */

#include <linux/of_platform.h>
#include <linux/nvhost.h>
#include <linux/debugfs.h>
#include <linux/tegra-powergate.h>
#include <linux/platform_data/tegra_edp.h>
#include <linux/delay.h>
#include <uapi/linux/nvgpu.h>
#include <linux/dma-buf.h>
#include <linux/nvmap.h>
#include <linux/tegra_pm_domains.h>
#include <linux/tegra_soctherm.h>
#include <linux/platform/tegra/clock.h>
#include <linux/platform/tegra/dvfs.h>
#include <linux/platform/tegra/common.h>
#include <linux/platform/tegra/mc.h>
#include <linux/clk/tegra.h>

#include <linux/platform/tegra/tegra_emc.h>

#include "gk20a.h"
#include "hal_gk20a.h"
#include "platform_gk20a.h"
#include "gk20a_scale.h"

#define TEGRA_GK20A_BW_PER_FREQ 32
#define TEGRA_GM20B_BW_PER_FREQ 64
#define TEGRA_DDR3_BW_PER_FREQ 16
#define TEGRA_DDR4_BW_PER_FREQ 16
#define MC_CLIENT_GPU 34
#define PMC_GPU_RG_CNTRL_0		0x2d4

extern struct device tegra_vpr_dev;

struct gk20a_emc_params {
	long bw_ratio;
	long freq_last_set;
};

static void __iomem *pmc = IO_ADDRESS(TEGRA_PMC_BASE);
static inline u32 pmc_read(unsigned long reg)
{
	return readl(pmc + reg);
}

static inline void pmc_write(u32 val, unsigned long reg)
{
	writel_relaxed(val, pmc + reg);
}
#define MHZ_TO_HZ(x) ((x) * 1000000)
#define HZ_TO_MHZ(x) ((x) / 1000000)

static void gk20a_tegra_secure_page_destroy(struct platform_device *pdev,
				       struct secure_page_buffer *secure_buffer)
{
	dma_free_attrs(&tegra_vpr_dev, secure_buffer->size,
			(void *)(uintptr_t)secure_buffer->iova,
			secure_buffer->iova, &secure_buffer->attrs);
}

int gk20a_tegra_secure_page_alloc(struct platform_device *pdev)
{
	struct gk20a_platform *platform = platform_get_drvdata(pdev);
	struct secure_page_buffer *secure_buffer = &platform->secure_buffer;
	DEFINE_DMA_ATTRS(attrs);
	dma_addr_t iova;
	size_t size = PAGE_SIZE;

	(void)dma_alloc_attrs(&tegra_vpr_dev, size, &iova,
				      DMA_MEMORY_NOMAP, &attrs);
	if (dma_mapping_error(&tegra_vpr_dev, iova))
		return -ENOMEM;

	secure_buffer->size = size;
	secure_buffer->iova = iova;
	secure_buffer->attrs = attrs;
	secure_buffer->destroy = gk20a_tegra_secure_page_destroy;

	return 0;
}

static void gk20a_tegra_secure_destroy(struct gk20a *g,
				       struct gr_ctx_buffer_desc *desc)
{
	DEFINE_DMA_ATTRS(attrs);

	if (desc->mem.sgt) {
		phys_addr_t pa = sg_phys(desc->mem.sgt->sgl);
		dma_free_attrs(&tegra_vpr_dev, desc->mem.size,
			(void *)(uintptr_t)pa,
			pa, &attrs);
		gk20a_free_sgtable(&desc->mem.sgt);
		desc->mem.sgt = NULL;
	}
}

int gk20a_tegra_secure_alloc(struct platform_device *pdev,
			     struct gr_ctx_buffer_desc *desc,
			     size_t size)
{
	struct gk20a_platform *platform = platform_get_drvdata(pdev);
	struct device *dev = &pdev->dev;
	DEFINE_DMA_ATTRS(attrs);
	dma_addr_t iova;
	struct sg_table *sgt;
	struct page *page;
	int err = 0;

	if (!platform->secure_alloc_ready)
		return -EINVAL;

	(void)dma_alloc_attrs(&tegra_vpr_dev, size, &iova,
				      DMA_MEMORY_NOMAP, &attrs);
	if (dma_mapping_error(&tegra_vpr_dev, iova))
		return -ENOMEM;

	desc->mem.size = size;
	desc->destroy = gk20a_tegra_secure_destroy;

	sgt = kzalloc(sizeof(*sgt), GFP_KERNEL);
	if (!sgt) {
		gk20a_err(dev, "failed to allocate memory\n");
		goto fail;
	}
	err = sg_alloc_table(sgt, 1, GFP_KERNEL);
	if (err) {
		gk20a_err(dev, "failed to allocate sg_table\n");
		goto fail_sgt;
	}
	page = phys_to_page(iova);
	sg_set_page(sgt->sgl, page, size, 0);
	/* This bypasses SMMU for VPR during gmmu_map. */
	sg_dma_address(sgt->sgl) = 0;

	desc->mem.sgt = sgt;

	return err;

fail_sgt:
	kfree(sgt);
fail:
	dma_free_attrs(&tegra_vpr_dev, desc->mem.size,
			(void *)(uintptr_t)iova, iova, &attrs);
	return err;
}

/*
 * gk20a_tegra_get_emc_rate()
 *
 * This function returns the minimum emc clock based on gpu frequency
 */

static unsigned long gk20a_tegra_get_emc_rate(struct gk20a *g,
				struct gk20a_emc_params *emc_params)
{
	unsigned long gpu_freq, gpu_fmax_at_vmin;
	unsigned long emc_rate, emc_scale;

	gpu_freq = clk_get_rate(g->clk.tegra_clk);
	gpu_fmax_at_vmin = tegra_dvfs_get_fmax_at_vmin_safe_t(
		clk_get_parent(g->clk.tegra_clk));

	/* When scaling emc, account for the gpu load when the
	 * gpu frequency is less than or equal to fmax@vmin. */
	if (gpu_freq <= gpu_fmax_at_vmin)
		emc_scale = min(g->pmu.load_avg, g->emc3d_ratio);
	else
		emc_scale = g->emc3d_ratio;

	emc_rate =
		(HZ_TO_MHZ(gpu_freq) * emc_params->bw_ratio * emc_scale) / 1000;

	return MHZ_TO_HZ(emc_rate);
}

/*
 * gk20a_tegra_postscale(profile, freq)
 *
 * This function sets emc frequency based on current gpu frequency
 */

static void gk20a_tegra_postscale(struct platform_device *pdev,
				  unsigned long freq)
{
	struct gk20a_platform *platform = platform_get_drvdata(pdev);
	struct gk20a_scale_profile *profile = platform->g->scale_profile;
	struct gk20a_emc_params *emc_params = profile->private_data;
	struct gk20a *g = get_gk20a(pdev);
	struct clk *emc_clk = platform->clk[2];
	enum tegra_chipid chip_id = tegra_get_chip_id();
	unsigned long emc_target;
	long emc_freq_lower, emc_freq_upper, emc_freq_rounded;

	emc_target = gk20a_tegra_get_emc_rate(g, emc_params);

	switch (chip_id) {
	case TEGRA_CHIPID_TEGRA12:
	case TEGRA_CHIPID_TEGRA13:
		/* T124 and T132 don't apply any rounding. The resulting
		 * emc frequency gets implicitly rounded up after issuing
		 * the clock_set_request.
		 * So explicitly round up the emc target here to achieve
		 * the same outcome. */
		emc_freq_rounded =
			tegra_emc_round_rate_updown(emc_target, true);
		break;

	case TEGRA_CHIPID_TEGRA21:
		emc_freq_lower = tegra_emc_round_rate_updown(emc_target, false);
		emc_freq_upper = tegra_emc_round_rate_updown(emc_target, true);

		/* round to the nearest frequency step */
		if (emc_target < (emc_freq_lower + emc_freq_upper) / 2)
			emc_freq_rounded = emc_freq_lower;
		else
			emc_freq_rounded = emc_freq_upper;
		break;

	case TEGRA_CHIPID_UNKNOWN:
	default:
		/* a proper rounding function needs to be implemented
		 * for emc in t18x */
		emc_freq_rounded = clk_round_rate(emc_clk, emc_target);
		break;
	}

	/* only change the emc clock if new rounded frequency is different
	 * from previously set emc rate */
	if (emc_freq_rounded != emc_params->freq_last_set) {
		clk_set_rate(emc_clk, emc_freq_rounded);
		emc_params->freq_last_set = emc_freq_rounded;
	}
}

/*
 * gk20a_tegra_prescale(profile, freq)
 *
 * This function informs EDP about changed constraints.
 */

static void gk20a_tegra_prescale(struct platform_device *pdev)
{
	struct gk20a *g = get_gk20a(pdev);
	u32 avg = 0;

	gk20a_pmu_load_norm(g, &avg);
	tegra_edp_notify_gpu_load(avg, clk_get_rate(g->clk.tegra_clk));
}

/*
 * gk20a_tegra_calibrate_emc()
 *
 */

static void gk20a_tegra_calibrate_emc(struct platform_device *pdev,
			       struct gk20a_emc_params *emc_params)
{
	enum tegra_chipid cid = tegra_get_chipid();
	long gpu_bw, emc_bw;

	/* store gpu bw based on soc */
	switch (cid) {
	case TEGRA_CHIPID_TEGRA21:
		gpu_bw = TEGRA_GM20B_BW_PER_FREQ;
		break;
	case TEGRA_CHIPID_TEGRA12:
	case TEGRA_CHIPID_TEGRA13:
		gpu_bw = TEGRA_GK20A_BW_PER_FREQ;
		break;
	case TEGRA_CHIPID_UNKNOWN:
	default:
		gpu_bw = 0;
		break;
	}

	/* TODO detect DDR type.
	 * Okay for now since DDR3 and DDR4 have the same BW ratio */
	emc_bw = TEGRA_DDR3_BW_PER_FREQ;

	/* Calculate the bandwidth ratio of gpu_freq <-> emc_freq
	 *   NOTE the ratio must come out as an integer */
	emc_params->bw_ratio = (gpu_bw / emc_bw);
}

#ifdef CONFIG_TEGRA_CLK_FRAMEWORK
/*
 * gk20a_tegra_is_railgated()
 *
 * Check status of gk20a power rail
 */

static bool gk20a_tegra_is_railgated(struct platform_device *pdev)
{
	struct gk20a_platform *platform = platform_get_drvdata(pdev);
	bool ret = false;

	if (!tegra_platform_is_linsim())
		ret = !tegra_dvfs_is_rail_up(platform->gpu_rail);

	return ret;
}

/*
 * gk20a_tegra_railgate()
 *
 * Gate (disable) gk20a power rail
 */

static int gk20a_tegra_railgate(struct platform_device *pdev)
{
	struct gk20a_platform *platform = platform_get_drvdata(pdev);
	int ret = 0;

	if (tegra_platform_is_linsim() ||
	    !tegra_dvfs_is_rail_up(platform->gpu_rail))
		return 0;

	tegra_mc_flush(MC_CLIENT_GPU);

	udelay(10);

	/* enable clamp */
	pmc_write(0x1, PMC_GPU_RG_CNTRL_0);
	pmc_read(PMC_GPU_RG_CNTRL_0);

	udelay(10);

	platform->reset_assert(pdev);

	udelay(10);

	/*
	 * GPCPLL is already disabled before entering this function; reference
	 * clocks are enabled until now - disable them just before rail gating
	 */
	clk_disable(platform->clk[0]);
	clk_disable(platform->clk[1]);

	udelay(10);

	if (tegra_dvfs_is_rail_up(platform->gpu_rail)) {
		ret = tegra_dvfs_rail_power_down(platform->gpu_rail);
		if (ret)
			goto err_power_off;
	} else
		pr_info("No GPU regulator?\n");

	return 0;

err_power_off:
	gk20a_err(&pdev->dev, "Could not railgate GPU");
	return ret;
}

/*
 * gm20b_tegra_railgate()
 *
 * Gate (disable) gm20b power rail
 */

static int gm20b_tegra_railgate(struct platform_device *pdev)
{
	struct gk20a_platform *platform = platform_get_drvdata(pdev);
	int ret = 0;

	if (tegra_platform_is_linsim() ||
	    !tegra_dvfs_is_rail_up(platform->gpu_rail))
		return 0;

	tegra_mc_flush(MC_CLIENT_GPU);

	udelay(10);

	/* enable clamp */
	pmc_write(0x1, PMC_GPU_RG_CNTRL_0);
	pmc_read(PMC_GPU_RG_CNTRL_0);

	udelay(10);

	platform->reset_assert(pdev);

	udelay(10);

	/*
	 * GPCPLL is already disabled before entering this function; reference
	 * clocks are enabled until now - disable them just before rail gating
	 */
	clk_disable(platform->clk_reset);
	clk_disable(platform->clk[0]);
	clk_disable(platform->clk[1]);

	udelay(10);

	tegra_soctherm_gpu_tsens_invalidate(1);

	if (tegra_dvfs_is_rail_up(platform->gpu_rail)) {
		ret = tegra_dvfs_rail_power_down(platform->gpu_rail);
		if (ret)
			goto err_power_off;
	} else
		pr_info("No GPU regulator?\n");

	return 0;

err_power_off:
	gk20a_err(&pdev->dev, "Could not railgate GPU");
	return ret;
}

/*
 * gk20a_tegra_unrailgate()
 *
 * Ungate (enable) gk20a power rail
 */

static int gk20a_tegra_unrailgate(struct platform_device *pdev)
{
	struct gk20a_platform *platform = platform_get_drvdata(pdev);
	int ret = 0;
	bool first = false;

	if (tegra_platform_is_linsim())
		return 0;

	if (!platform->gpu_rail) {
		platform->gpu_rail = tegra_dvfs_get_rail_by_name("vdd_gpu");
		if (IS_ERR_OR_NULL(platform->gpu_rail)) {
			WARN(1, "No GPU regulator?\n");
			return -EINVAL;
		}
		first = true;
	}

	ret = tegra_dvfs_rail_power_up(platform->gpu_rail);
	if (ret)
		return ret;

	if (!first) {
		ret = clk_enable(platform->clk[0]);
		if (ret) {
			gk20a_err(&pdev->dev, "could not turn on gpu pll");
			goto err_clk_on;
		}
		ret = clk_enable(platform->clk[1]);
		if (ret) {
			gk20a_err(&pdev->dev, "could not turn on pwr clock");
			goto err_clk_on;
		}
	}

	udelay(10);

	platform->reset_assert(pdev);

	udelay(10);

	pmc_write(0, PMC_GPU_RG_CNTRL_0);
	pmc_read(PMC_GPU_RG_CNTRL_0);

	udelay(10);

	platform->reset_deassert(pdev);

	/* Flush MC after boot/railgate/SC7 */
	tegra_mc_flush(MC_CLIENT_GPU);

	udelay(10);

	tegra_mc_flush_done(MC_CLIENT_GPU);

	udelay(10);

	return 0;

err_clk_on:
	tegra_dvfs_rail_power_down(platform->gpu_rail);

	return ret;
}

/*
 * gm20b_tegra_unrailgate()
 *
 * Ungate (enable) gm20b power rail
 */

static int gm20b_tegra_unrailgate(struct platform_device *pdev)
{
	struct gk20a_platform *platform = platform_get_drvdata(pdev);
	int ret = 0;
	bool first = false;

	if (tegra_platform_is_linsim())
		return 0;

	if (!platform->gpu_rail) {
		platform->gpu_rail = tegra_dvfs_get_rail_by_name("vdd_gpu");
		if (IS_ERR_OR_NULL(platform->gpu_rail)) {
			WARN(1, "No GPU regulator?\n");
			return -EINVAL;
		}
		first = true;
	}

	ret = tegra_dvfs_rail_power_up(platform->gpu_rail);
	if (ret)
		return ret;

	tegra_soctherm_gpu_tsens_invalidate(0);

	if (!first) {
		ret = clk_enable(platform->clk_reset);
		if (ret) {
			gk20a_err(&pdev->dev, "could not turn on gpu_gate");
			goto err_clk_on;
		}

		ret = clk_enable(platform->clk[0]);
		if (ret) {
			gk20a_err(&pdev->dev, "could not turn on gpu pll");
			goto err_clk_on;
		}
		ret = clk_enable(platform->clk[1]);
		if (ret) {
			gk20a_err(&pdev->dev, "could not turn on pwr clock");
			goto err_clk_on;
		}
	}

	udelay(10);

	platform->reset_assert(pdev);

	udelay(10);

	pmc_write(0, PMC_GPU_RG_CNTRL_0);
	pmc_read(PMC_GPU_RG_CNTRL_0);

	udelay(10);

	clk_disable(platform->clk_reset);
	platform->reset_deassert(pdev);
	clk_enable(platform->clk_reset);

	/* Flush MC after boot/railgate/SC7 */
	tegra_mc_flush(MC_CLIENT_GPU);

	udelay(10);

	tegra_mc_flush_done(MC_CLIENT_GPU);

	udelay(10);

	return 0;

err_clk_on:
	tegra_dvfs_rail_power_down(platform->gpu_rail);

	return ret;
}
#endif

static struct {
	char *name;
	unsigned long default_rate;
} tegra_gk20a_clocks[] = {
	{"PLLG_ref", UINT_MAX},
	{"pwr", 204000000},
	{"emc", UINT_MAX} };

/*
 * gk20a_tegra_get_clocks()
 *
 * This function finds clocks in tegra platform and populates
 * the clock information to gk20a platform data.
 */

static int gk20a_tegra_get_clocks(struct platform_device *pdev)
{
	struct gk20a_platform *platform = platform_get_drvdata(pdev);
	char devname[16];
	int i;
	int ret = 0;

	snprintf(devname, sizeof(devname),
		 (pdev->id <= 0) ? "tegra_%s" : "tegra_%s.%d\n",
		 pdev->name, pdev->id);

	platform->num_clks = 0;
	for (i = 0; i < ARRAY_SIZE(tegra_gk20a_clocks); i++) {
		long rate = tegra_gk20a_clocks[i].default_rate;
		struct clk *c;

		c = clk_get_sys(devname, tegra_gk20a_clocks[i].name);
		if (IS_ERR(c)) {
			ret = PTR_ERR(c);
			goto err_get_clock;
		}
		rate = clk_round_rate(c, rate);
		clk_set_rate(c, rate);
		platform->clk[i] = c;
	}
	platform->num_clks = i;

	return 0;

err_get_clock:

	while (i--)
		clk_put(platform->clk[i]);
	return ret;
}

static int gk20a_tegra_reset_assert(struct platform_device *dev)
{
	struct gk20a_platform *platform = gk20a_get_platform(dev);

	if (!platform->clk_reset)
		platform->clk_reset = platform->clk[0];

	tegra_periph_reset_assert(platform->clk_reset);

	return 0;
}

static int gk20a_tegra_reset_deassert(struct platform_device *dev)
{
	struct gk20a_platform *platform = gk20a_get_platform(dev);

	if (!platform->clk_reset)
		return -EINVAL;

	tegra_periph_reset_deassert(platform->clk_reset);

	return 0;
}

static int gm20b_tegra_reset_assert(struct platform_device *dev)
{
	struct gk20a_platform *platform = gk20a_get_platform(dev);

	if (!platform->clk_reset) {
		platform->clk_reset = clk_get(&dev->dev, "gpu_gate");
		if (IS_ERR(platform->clk_reset)) {
			gk20a_err(&dev->dev, "fail to get gpu reset clk\n");
			return PTR_ERR(platform->clk_reset);
		}
	}

	tegra_periph_reset_assert(platform->clk_reset);

	return 0;
}

static void gk20a_tegra_scale_init(struct platform_device *pdev)
{
	struct gk20a_platform *platform = gk20a_get_platform(pdev);
	struct gk20a_scale_profile *profile = platform->g->scale_profile;
	struct gk20a_emc_params *emc_params;

	if (!profile)
		return;

	emc_params = kzalloc(sizeof(*emc_params), GFP_KERNEL);
	if (!emc_params)
		return;

	emc_params->freq_last_set = -1;
	gk20a_tegra_calibrate_emc(pdev, emc_params);

	profile->private_data = emc_params;
}

static void gk20a_tegra_scale_exit(struct platform_device *pdev)
{
	struct gk20a_platform *platform = gk20a_get_platform(pdev);
	struct gk20a_scale_profile *profile = platform->g->scale_profile;

	if (profile)
		kfree(profile->private_data);
}

void gk20a_tegra_debug_dump(struct platform_device *pdev)
{
	struct gk20a_platform *platform = gk20a_get_platform(pdev);
	struct gk20a *g = platform->g;
	nvhost_debug_dump_device(g->host1x_dev);
}

int gk20a_tegra_busy(struct platform_device *dev)
{
	struct gk20a_platform *platform = gk20a_get_platform(dev);
	struct gk20a *g = platform->g;

	if (g->host1x_dev)
		return nvhost_module_busy_ext(g->host1x_dev);
	return 0;
}

void gk20a_tegra_idle(struct platform_device *dev)
{
	struct gk20a_platform *platform = gk20a_get_platform(dev);
	struct gk20a *g = platform->g;

	if (g->host1x_dev)
		nvhost_module_idle_ext(g->host1x_dev);
}

static int gk20a_tegra_probe(struct platform_device *dev)
{
	struct gk20a_platform *platform = gk20a_get_platform(dev);
	struct device_node *np = dev->dev.of_node;
	const __be32 *host1x_ptr;
	struct platform_device *host1x_pdev = NULL;

	host1x_ptr = of_get_property(np, "nvidia,host1x", NULL);
	if (host1x_ptr) {
		struct device_node *host1x_node =
			of_find_node_by_phandle(be32_to_cpup(host1x_ptr));

		host1x_pdev = of_find_device_by_node(host1x_node);
		if (!host1x_pdev) {
			dev_warn(&dev->dev, "host1x device not available");
			return -EPROBE_DEFER;
		}

	} else {
		host1x_pdev = to_platform_device(dev->dev.parent);
		dev_warn(&dev->dev, "host1x reference not found. assuming host1x to be parent");
	}

	platform->g->host1x_dev = host1x_pdev;

	if (platform->g->host1x_dev)
		nvhost_register_dump_device(platform->g->host1x_dev,
					gk20a_debug_dump_device);

	/* WAR for bug 1547668: Disable railgating and scaling irrespective of
	 * platform data if the rework has not been made. */

	if (tegra_get_chipid() == TEGRA_CHIPID_TEGRA21) {
		np = of_find_node_by_path("/gpu-dvfs-rework");
		if (!(np && of_device_is_available(np))) {
			platform->devfreq_governor = "";
			dev_warn(&dev->dev, "board does not support scaling");
		}
	}

	gk20a_tegra_get_clocks(dev);

	return 0;
}

static int gk20a_tegra_late_probe(struct platform_device *dev)
{
	/* Make gk20a power domain a subdomain of host1x */
	nvhost_register_client_domain(dev_to_genpd(&dev->dev));

	/* Initialise tegra specific scaling quirks */
	gk20a_tegra_scale_init(dev);

	return 0;
}

static int gk20a_tegra_remove(struct platform_device *dev)
{
	struct gk20a_platform *platform = gk20a_get_platform(dev);

	if (platform->g->host1x_dev)
		nvhost_unregister_dump_device(platform->g->host1x_dev);

	/* remove gk20a power subdomain from host1x */
	nvhost_unregister_client_domain(dev_to_genpd(&dev->dev));

	/* deinitialise tegra specific scaling quirks */
	gk20a_tegra_scale_exit(dev);

	return 0;
}

static int gk20a_tegra_suspend(struct device *dev)
{
	tegra_edp_notify_gpu_load(0, 0);
	return 0;
}

struct gk20a_platform gk20a_tegra_platform = {
	.has_syncpoints = true,

	/* power management configuration */
	.railgate_delay		= 500,
	.clockgate_delay	= 50,
	.can_railgate		= true,
	.enable_slcg            = true,
	.enable_blcg            = true,
	.enable_elcg            = true,
	.enable_elpg            = true,
	.enable_aelpg           = true,
	.ptimerscaling10x	= 26,

	.force_reset_in_do_idle = false,

	.default_big_page_size	= SZ_128K,

	.ch_wdt_timeout_ms = 7000,

	.probe = gk20a_tegra_probe,
	.late_probe = gk20a_tegra_late_probe,
	.remove = gk20a_tegra_remove,

	/* power management callbacks */
	.suspend = gk20a_tegra_suspend,
#ifdef CONFIG_TEGRA_CLK_FRAMEWORK
	.railgate = gk20a_tegra_railgate,
	.unrailgate = gk20a_tegra_unrailgate,
	.is_railgated = gk20a_tegra_is_railgated,
#endif

	.busy = gk20a_tegra_busy,
	.idle = gk20a_tegra_idle,

	.reset_assert = gk20a_tegra_reset_assert,
	.reset_deassert = gk20a_tegra_reset_deassert,

	/* frequency scaling configuration */
	.prescale = gk20a_tegra_prescale,
	.postscale = gk20a_tegra_postscale,
	.devfreq_governor = "nvhost_podgov",
	.qos_id = PM_QOS_GPU_FREQ_MIN,

	.secure_alloc = gk20a_tegra_secure_alloc,
	.secure_page_alloc = gk20a_tegra_secure_page_alloc,
	.dump_platform_dependencies = gk20a_tegra_debug_dump,
};

struct gk20a_platform gm20b_tegra_platform = {
	.has_syncpoints = true,

	/* power management configuration */
	.railgate_delay		= 500,
	.clockgate_delay	= 50,
	.can_railgate           = true,
	.enable_slcg            = true,
	.enable_blcg            = true,
	.enable_elcg            = true,
	.enable_elpg            = true,
	.enable_aelpg           = true,
	.ptimerscaling10x	= 26,

	.force_reset_in_do_idle = false,

	.default_big_page_size	= SZ_128K,

	.ch_wdt_timeout_ms = 5000,

	.probe = gk20a_tegra_probe,
	.late_probe = gk20a_tegra_late_probe,
	.remove = gk20a_tegra_remove,

	/* power management callbacks */
	.suspend = gk20a_tegra_suspend,
#ifdef CONFIG_TEGRA_CLK_FRAMEWORK
	.railgate = gm20b_tegra_railgate,
	.unrailgate = gm20b_tegra_unrailgate,
	.is_railgated = gk20a_tegra_is_railgated,
#endif

	.busy = gk20a_tegra_busy,
	.idle = gk20a_tegra_idle,

	.reset_assert = gm20b_tegra_reset_assert,
	.reset_deassert = gk20a_tegra_reset_deassert,

	/* frequency scaling configuration */
	.prescale = gk20a_tegra_prescale,
	.postscale = gk20a_tegra_postscale,
	.devfreq_governor = "nvhost_podgov",
	.qos_id = PM_QOS_GPU_FREQ_MIN,

	.secure_alloc = gk20a_tegra_secure_alloc,
	.secure_page_alloc = gk20a_tegra_secure_page_alloc,
	.dump_platform_dependencies = gk20a_tegra_debug_dump,

	.has_cde = true,
};