summaryrefslogtreecommitdiffstats
path: root/drivers/gpu/drm/tegra/submit.c
blob: 776f825df52fa704f3a5de73d9752309b0ea89e0 (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
// SPDX-License-Identifier: GPL-2.0-only
/* Copyright (c) 2020 NVIDIA Corporation */

#include <linux/dma-fence-array.h>
#include <linux/dma-mapping.h>
#include <linux/file.h>
#include <linux/host1x.h>
#include <linux/iommu.h>
#include <linux/kref.h>
#include <linux/list.h>
#include <linux/nospec.h>
#include <linux/pm_runtime.h>
#include <linux/scatterlist.h>
#include <linux/slab.h>
#include <linux/sync_file.h>

#include <drm/drm_drv.h>
#include <drm/drm_file.h>
#include <drm/drm_syncobj.h>

#include "drm.h"
#include "gem.h"
#include "submit.h"
#include "uapi.h"

#define SUBMIT_ERR(context, fmt, ...) \
	dev_err_ratelimited(context->client->base.dev, \
		"%s: job submission failed: " fmt "\n", \
		current->comm, ##__VA_ARGS__)

struct gather_bo {
	struct host1x_bo base;

	struct kref ref;

	struct device *dev;
	u32 *gather_data;
	dma_addr_t gather_data_dma;
	size_t gather_data_words;
};

static struct host1x_bo *gather_bo_get(struct host1x_bo *host_bo)
{
	struct gather_bo *bo = container_of(host_bo, struct gather_bo, base);

	kref_get(&bo->ref);

	return host_bo;
}

static void gather_bo_release(struct kref *ref)
{
	struct gather_bo *bo = container_of(ref, struct gather_bo, ref);

	dma_free_attrs(bo->dev, bo->gather_data_words * 4, bo->gather_data, bo->gather_data_dma,
		       0);
	kfree(bo);
}

static void gather_bo_put(struct host1x_bo *host_bo)
{
	struct gather_bo *bo = container_of(host_bo, struct gather_bo, base);

	kref_put(&bo->ref, gather_bo_release);
}

static struct sg_table *
gather_bo_pin(struct device *dev, struct host1x_bo *host_bo, dma_addr_t *phys)
{
	struct gather_bo *bo = container_of(host_bo, struct gather_bo, base);
	struct sg_table *sgt;
	int err;

	sgt = kzalloc(sizeof(*sgt), GFP_KERNEL);
	if (!sgt)
		return ERR_PTR(-ENOMEM);

	err = dma_get_sgtable(bo->dev, sgt, bo->gather_data, bo->gather_data_dma,
			      bo->gather_data_words * 4);
	if (err) {
		kfree(sgt);
		return ERR_PTR(err);
	}

	return sgt;
}

static void gather_bo_unpin(struct device *dev, struct sg_table *sgt)
{
	if (sgt) {
		sg_free_table(sgt);
		kfree(sgt);
	}
}

static void *gather_bo_mmap(struct host1x_bo *host_bo)
{
	struct gather_bo *bo = container_of(host_bo, struct gather_bo, base);

	return bo->gather_data;
}

static void gather_bo_munmap(struct host1x_bo *host_bo, void *addr)
{
}

const struct host1x_bo_ops gather_bo_ops = {
	.get = gather_bo_get,
	.put = gather_bo_put,
	.pin = gather_bo_pin,
	.unpin = gather_bo_unpin,
	.mmap = gather_bo_mmap,
	.munmap = gather_bo_munmap,
};

static struct tegra_drm_mapping *
tegra_drm_mapping_get(struct tegra_drm_context *context, u32 id)
{
	struct tegra_drm_mapping *mapping;

	xa_lock(&context->mappings);

	mapping = xa_load(&context->mappings, id);
	if (mapping)
		kref_get(&mapping->ref);

	xa_unlock(&context->mappings);

	return mapping;
}

static void *alloc_copy_user_array(void __user *from, size_t count, size_t size)
{
	size_t copy_len;
	void *data;

	if (check_mul_overflow(count, size, &copy_len))
		return ERR_PTR(-EINVAL);

	if (copy_len > 0x4000)
		return ERR_PTR(-E2BIG);

	data = kvmalloc(copy_len, GFP_KERNEL);
	if (!data)
		return ERR_PTR(-ENOMEM);

	if (copy_from_user(data, from, copy_len)) {
		kvfree(data);
		return ERR_PTR(-EFAULT);
	}

	return data;
}

static int submit_copy_gather_data(struct gather_bo **pbo, struct device *dev,
				   struct tegra_drm_context *context,
				   struct drm_tegra_channel_submit *args)
{
	struct gather_bo *bo;
	size_t copy_len;

	if (args->gather_data_words == 0) {
		SUBMIT_ERR(context, "gather_data_words cannot be zero");
		return -EINVAL;
	}

	if (check_mul_overflow((size_t)args->gather_data_words, (size_t)4, &copy_len)) {
		SUBMIT_ERR(context, "gather_data_words is too large");
		return -EINVAL;
	}

	bo = kzalloc(sizeof(*bo), GFP_KERNEL);
	if (!bo) {
		SUBMIT_ERR(context, "failed to allocate memory for bo info");
		return -ENOMEM;
	}

	host1x_bo_init(&bo->base, &gather_bo_ops);
	kref_init(&bo->ref);
	bo->dev = dev;

	bo->gather_data = dma_alloc_attrs(dev, copy_len, &bo->gather_data_dma,
					  GFP_KERNEL | __GFP_NOWARN, 0);
	if (!bo->gather_data) {
		SUBMIT_ERR(context, "failed to allocate memory for gather data");
		kfree(bo);
		return -ENOMEM;
	}

	if (copy_from_user(bo->gather_data, u64_to_user_ptr(args->gather_data_ptr), copy_len)) {
		SUBMIT_ERR(context, "failed to copy gather data from userspace");
		dma_free_attrs(dev, copy_len, bo->gather_data, bo->gather_data_dma, 0);
		kfree(bo);
		return -EFAULT;
	}

	bo->gather_data_words = args->gather_data_words;

	*pbo = bo;

	return 0;
}

static int submit_write_reloc(struct tegra_drm_context *context, struct gather_bo *bo,
			      struct drm_tegra_submit_buf *buf, struct tegra_drm_mapping *mapping)
{
	/* TODO check that target_offset is within bounds */
	dma_addr_t iova = mapping->iova + buf->reloc.target_offset;
	u32 written_ptr;

#ifdef CONFIG_ARCH_DMA_ADDR_T_64BIT
	if (buf->flags & DRM_TEGRA_SUBMIT_RELOC_SECTOR_LAYOUT)
		iova |= BIT_ULL(39);
#endif

	written_ptr = iova >> buf->reloc.shift;

	if (buf->reloc.gather_offset_words >= bo->gather_data_words) {
		SUBMIT_ERR(context,
			   "relocation has too large gather offset (%u vs gather length %zu)",
			   buf->reloc.gather_offset_words, bo->gather_data_words);
		return -EINVAL;
	}

	buf->reloc.gather_offset_words = array_index_nospec(buf->reloc.gather_offset_words,
							    bo->gather_data_words);

	bo->gather_data[buf->reloc.gather_offset_words] = written_ptr;

	return 0;
}

static int submit_process_bufs(struct tegra_drm_context *context, struct gather_bo *bo,
			       struct drm_tegra_channel_submit *args,
			       struct tegra_drm_submit_data *job_data)
{
	struct tegra_drm_used_mapping *mappings;
	struct drm_tegra_submit_buf *bufs;
	int err;
	u32 i;

	bufs = alloc_copy_user_array(u64_to_user_ptr(args->bufs_ptr), args->num_bufs,
				     sizeof(*bufs));
	if (IS_ERR(bufs)) {
		SUBMIT_ERR(context, "failed to copy bufs array from userspace");
		return PTR_ERR(bufs);
	}

	mappings = kcalloc(args->num_bufs, sizeof(*mappings), GFP_KERNEL);
	if (!mappings) {
		SUBMIT_ERR(context, "failed to allocate memory for mapping info");
		err = -ENOMEM;
		goto done;
	}

	for (i = 0; i < args->num_bufs; i++) {
		struct drm_tegra_submit_buf *buf = &bufs[i];
		struct tegra_drm_mapping *mapping;

		if (buf->flags & ~DRM_TEGRA_SUBMIT_RELOC_SECTOR_LAYOUT) {
			SUBMIT_ERR(context, "invalid flag specified for buffer");
			err = -EINVAL;
			goto drop_refs;
		}

		mapping = tegra_drm_mapping_get(context, buf->mapping);
		if (!mapping) {
			SUBMIT_ERR(context, "invalid mapping ID '%u' for buffer", buf->mapping);
			err = -EINVAL;
			goto drop_refs;
		}

		err = submit_write_reloc(context, bo, buf, mapping);
		if (err) {
			tegra_drm_mapping_put(mapping);
			goto drop_refs;
		}

		mappings[i].mapping = mapping;
		mappings[i].flags = buf->flags;
	}

	job_data->used_mappings = mappings;
	job_data->num_used_mappings = i;

	err = 0;

	goto done;

drop_refs:
	while (i--)
		tegra_drm_mapping_put(mappings[i].mapping);

	kfree(mappings);
	job_data->used_mappings = NULL;

done:
	kvfree(bufs);

	return err;
}

static int submit_get_syncpt(struct tegra_drm_context *context, struct host1x_job *job,
			     struct xarray *syncpoints, struct drm_tegra_channel_submit *args)
{
	struct host1x_syncpt *sp;

	if (args->syncpt.flags) {
		SUBMIT_ERR(context, "invalid flag specified for syncpt");
		return -EINVAL;
	}

	/* Syncpt ref will be dropped on job release */
	sp = xa_load(syncpoints, args->syncpt.id);
	if (!sp) {
		SUBMIT_ERR(context, "syncpoint specified in syncpt was not allocated");
		return -EINVAL;
	}

	job->syncpt = host1x_syncpt_get(sp);
	job->syncpt_incrs = args->syncpt.increments;

	return 0;
}

static int submit_job_add_gather(struct host1x_job *job, struct tegra_drm_context *context,
				 struct drm_tegra_submit_cmd_gather_uptr *cmd,
				 struct gather_bo *bo, u32 *offset,
				 struct tegra_drm_submit_data *job_data,
				 u32 *class)
{
	u32 next_offset;

	if (cmd->reserved[0] || cmd->reserved[1] || cmd->reserved[2]) {
		SUBMIT_ERR(context, "non-zero reserved field in GATHER_UPTR command");
		return -EINVAL;
	}

	/* Check for maximum gather size */
	if (cmd->words > 16383) {
		SUBMIT_ERR(context, "too many words in GATHER_UPTR command");
		return -EINVAL;
	}

	if (check_add_overflow(*offset, cmd->words, &next_offset)) {
		SUBMIT_ERR(context, "too many total words in job");
		return -EINVAL;
	}

	if (next_offset > bo->gather_data_words) {
		SUBMIT_ERR(context, "GATHER_UPTR command overflows gather data");
		return -EINVAL;
	}

	if (tegra_drm_fw_validate(context->client, bo->gather_data, *offset,
				  cmd->words, job_data, class)) {
		SUBMIT_ERR(context, "job was rejected by firewall");
		return -EINVAL;
	}

	host1x_job_add_gather(job, &bo->base, cmd->words, *offset * 4);

	*offset = next_offset;

	return 0;
}

static struct host1x_job *
submit_create_job(struct tegra_drm_context *context, struct gather_bo *bo,
		  struct drm_tegra_channel_submit *args, struct tegra_drm_submit_data *job_data,
		  struct xarray *syncpoints)
{
	struct drm_tegra_submit_cmd *cmds;
	u32 i, gather_offset = 0, class;
	struct host1x_job *job;
	int err;

	/* Set initial class for firewall. */
	class = context->client->base.class;

	cmds = alloc_copy_user_array(u64_to_user_ptr(args->cmds_ptr), args->num_cmds,
				     sizeof(*cmds));
	if (IS_ERR(cmds)) {
		SUBMIT_ERR(context, "failed to copy cmds array from userspace");
		return ERR_CAST(cmds);
	}

	job = host1x_job_alloc(context->channel, args->num_cmds, 0, true);
	if (!job) {
		SUBMIT_ERR(context, "failed to allocate memory for job");
		job = ERR_PTR(-ENOMEM);
		goto done;
	}

	err = submit_get_syncpt(context, job, syncpoints, args);
	if (err < 0)
		goto free_job;

	job->client = &context->client->base;
	job->class = context->client->base.class;
	job->serialize = true;

	for (i = 0; i < args->num_cmds; i++) {
		struct drm_tegra_submit_cmd *cmd = &cmds[i];

		if (cmd->flags) {
			SUBMIT_ERR(context, "unknown flags given for cmd");
			err = -EINVAL;
			goto free_job;
		}

		if (cmd->type == DRM_TEGRA_SUBMIT_CMD_GATHER_UPTR) {
			err = submit_job_add_gather(job, context, &cmd->gather_uptr, bo,
						    &gather_offset, job_data, &class);
			if (err)
				goto free_job;
		} else if (cmd->type == DRM_TEGRA_SUBMIT_CMD_WAIT_SYNCPT) {
			if (cmd->wait_syncpt.reserved[0] || cmd->wait_syncpt.reserved[1]) {
				SUBMIT_ERR(context, "non-zero reserved value");
				err = -EINVAL;
				goto free_job;
			}

			host1x_job_add_wait(job, cmd->wait_syncpt.id, cmd->wait_syncpt.value,
					    false, class);
		} else if (cmd->type == DRM_TEGRA_SUBMIT_CMD_WAIT_SYNCPT_RELATIVE) {
			if (cmd->wait_syncpt.reserved[0] || cmd->wait_syncpt.reserved[1]) {
				SUBMIT_ERR(context, "non-zero reserved value");
				err = -EINVAL;
				goto free_job;
			}

			if (cmd->wait_syncpt.id != args->syncpt.id) {
				SUBMIT_ERR(context, "syncpoint ID in CMD_WAIT_SYNCPT_RELATIVE is not used by the job");
				err = -EINVAL;
				goto free_job;
			}

			host1x_job_add_wait(job, cmd->wait_syncpt.id, cmd->wait_syncpt.value,
					    true, class);
		} else {
			SUBMIT_ERR(context, "unknown cmd type");
			err = -EINVAL;
			goto free_job;
		}
	}

	if (gather_offset == 0) {
		SUBMIT_ERR(context, "job must have at least one gather");
		err = -EINVAL;
		goto free_job;
	}

	goto done;

free_job:
	host1x_job_put(job);
	job = ERR_PTR(err);

done:
	kvfree(cmds);

	return job;
}

static void release_job(struct host1x_job *job)
{
	struct tegra_drm_client *client = container_of(job->client, struct tegra_drm_client, base);
	struct tegra_drm_submit_data *job_data = job->user_data;
	u32 i;

	for (i = 0; i < job_data->num_used_mappings; i++)
		tegra_drm_mapping_put(job_data->used_mappings[i].mapping);

	kfree(job_data->used_mappings);
	kfree(job_data);

	if (pm_runtime_enabled(client->base.dev))
		pm_runtime_put_autosuspend(client->base.dev);
}

int tegra_drm_ioctl_channel_submit(struct drm_device *drm, void *data,
				   struct drm_file *file)
{
	struct tegra_drm_file *fpriv = file->driver_priv;
	struct drm_tegra_channel_submit *args = data;
	struct tegra_drm_submit_data *job_data;
	struct drm_syncobj *syncobj = NULL;
	struct tegra_drm_context *context;
	struct host1x_job *job;
	struct gather_bo *bo;
	u32 i;
	int err;

	mutex_lock(&fpriv->lock);

	context = xa_load(&fpriv->contexts, args->context);
	if (!context) {
		mutex_unlock(&fpriv->lock);
		pr_err_ratelimited("%s: %s: invalid channel context '%#x'", __func__,
				   current->comm, args->context);
		return -EINVAL;
	}

	if (args->syncobj_in) {
		struct dma_fence *fence;

		err = drm_syncobj_find_fence(file, args->syncobj_in, 0, 0, &fence);
		if (err) {
			SUBMIT_ERR(context, "invalid syncobj_in '%#x'", args->syncobj_in);
			goto unlock;
		}

		err = dma_fence_wait_timeout(fence, true, msecs_to_jiffies(10000));
		dma_fence_put(fence);
		if (err) {
			SUBMIT_ERR(context, "wait for syncobj_in timed out");
			goto unlock;
		}
	}

	if (args->syncobj_out) {
		syncobj = drm_syncobj_find(file, args->syncobj_out);
		if (!syncobj) {
			SUBMIT_ERR(context, "invalid syncobj_out '%#x'", args->syncobj_out);
			err = -ENOENT;
			goto unlock;
		}
	}

	/* Allocate gather BO and copy gather words in. */
	err = submit_copy_gather_data(&bo, drm->dev, context, args);
	if (err)
		goto unlock;

	job_data = kzalloc(sizeof(*job_data), GFP_KERNEL);
	if (!job_data) {
		SUBMIT_ERR(context, "failed to allocate memory for job data");
		err = -ENOMEM;
		goto put_bo;
	}

	/* Get data buffer mappings and do relocation patching. */
	err = submit_process_bufs(context, bo, args, job_data);
	if (err)
		goto free_job_data;

	/* Allocate host1x_job and add gathers and waits to it. */
	job = submit_create_job(context, bo, args, job_data, &fpriv->syncpoints);
	if (IS_ERR(job)) {
		err = PTR_ERR(job);
		goto free_job_data;
	}

	/* Map gather data for Host1x. */
	err = host1x_job_pin(job, context->client->base.dev);
	if (err) {
		SUBMIT_ERR(context, "failed to pin job: %d", err);
		goto put_job;
	}

	/* Boot engine. */
	if (pm_runtime_enabled(context->client->base.dev)) {
		err = pm_runtime_resume_and_get(context->client->base.dev);
		if (err < 0) {
			SUBMIT_ERR(context, "could not power up engine: %d", err);
			goto unpin_job;
		}
	}

	job->user_data = job_data;
	job->release = release_job;
	job->timeout = 10000;

	/*
	 * job_data is now part of job reference counting, so don't release
	 * it from here.
	 */
	job_data = NULL;

	/* Submit job to hardware. */
	err = host1x_job_submit(job);
	if (err) {
		SUBMIT_ERR(context, "host1x job submission failed: %d", err);
		goto unpin_job;
	}

	/* Return postfences to userspace and add fences to DMA reservations. */
	args->syncpt.value = job->syncpt_end;

	if (syncobj) {
		struct dma_fence *fence = host1x_fence_create(job->syncpt, job->syncpt_end);
		if (IS_ERR(fence)) {
			err = PTR_ERR(fence);
			SUBMIT_ERR(context, "failed to create postfence: %d", err);
		}

		drm_syncobj_replace_fence(syncobj, fence);
	}

	goto put_job;

unpin_job:
	host1x_job_unpin(job);
put_job:
	host1x_job_put(job);
free_job_data:
	if (job_data && job_data->used_mappings) {
		for (i = 0; i < job_data->num_used_mappings; i++)
			tegra_drm_mapping_put(job_data->used_mappings[i].mapping);

		kfree(job_data->used_mappings);
	}

	if (job_data)
		kfree(job_data);
put_bo:
	gather_bo_put(&bo->base);
unlock:
	if (syncobj)
		drm_syncobj_put(syncobj);

	mutex_unlock(&fpriv->lock);
	return err;
}