Free Electrons

Embedded Linux Experts

  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
/*
 * Copyright © 2012-2014 Intel Corporation
 *
 * 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 (including the next
 * paragraph) 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 AUTHORS OR COPYRIGHT HOLDERS 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 <drm/drmP.h>
#include <drm/i915_drm.h>
#include "i915_drv.h"
#include "i915_trace.h"
#include "intel_drv.h"
#include <linux/mmu_context.h>
#include <linux/mmu_notifier.h>
#include <linux/mempolicy.h>
#include <linux/swap.h>
#include <linux/sched/mm.h>

struct i915_mm_struct {
	struct mm_struct *mm;
	struct drm_i915_private *i915;
	struct i915_mmu_notifier *mn;
	struct hlist_node node;
	struct kref kref;
	struct work_struct work;
};

#if defined(CONFIG_MMU_NOTIFIER)
#include <linux/interval_tree.h>

struct i915_mmu_notifier {
	spinlock_t lock;
	struct hlist_node node;
	struct mmu_notifier mn;
	struct rb_root_cached objects;
	struct workqueue_struct *wq;
};

struct i915_mmu_object {
	struct i915_mmu_notifier *mn;
	struct drm_i915_gem_object *obj;
	struct interval_tree_node it;
	struct list_head link;
	struct work_struct work;
	bool attached;
};

static void cancel_userptr(struct work_struct *work)
{
	struct i915_mmu_object *mo = container_of(work, typeof(*mo), work);
	struct drm_i915_gem_object *obj = mo->obj;
	struct work_struct *active;

	/* Cancel any active worker and force us to re-evaluate gup */
	mutex_lock(&obj->mm.lock);
	active = fetch_and_zero(&obj->userptr.work);
	mutex_unlock(&obj->mm.lock);
	if (active)
		goto out;

	i915_gem_object_wait(obj, I915_WAIT_ALL, MAX_SCHEDULE_TIMEOUT, NULL);

	mutex_lock(&obj->base.dev->struct_mutex);

	/* We are inside a kthread context and can't be interrupted */
	if (i915_gem_object_unbind(obj) == 0)
		__i915_gem_object_put_pages(obj, I915_MM_NORMAL);
	WARN_ONCE(obj->mm.pages,
		  "Failed to release pages: bind_count=%d, pages_pin_count=%d, pin_display=%d\n",
		  obj->bind_count,
		  atomic_read(&obj->mm.pages_pin_count),
		  obj->pin_display);

	mutex_unlock(&obj->base.dev->struct_mutex);

out:
	i915_gem_object_put(obj);
}

static void add_object(struct i915_mmu_object *mo)
{
	if (mo->attached)
		return;

	interval_tree_insert(&mo->it, &mo->mn->objects);
	mo->attached = true;
}

static void del_object(struct i915_mmu_object *mo)
{
	if (!mo->attached)
		return;

	interval_tree_remove(&mo->it, &mo->mn->objects);
	mo->attached = false;
}

static void i915_gem_userptr_mn_invalidate_range_start(struct mmu_notifier *_mn,
						       struct mm_struct *mm,
						       unsigned long start,
						       unsigned long end)
{
	struct i915_mmu_notifier *mn =
		container_of(_mn, struct i915_mmu_notifier, mn);
	struct i915_mmu_object *mo;
	struct interval_tree_node *it;
	LIST_HEAD(cancelled);

	if (RB_EMPTY_ROOT(&mn->objects.rb_root))
		return;

	/* interval ranges are inclusive, but invalidate range is exclusive */
	end--;

	spin_lock(&mn->lock);
	it = interval_tree_iter_first(&mn->objects, start, end);
	while (it) {
		/* The mmu_object is released late when destroying the
		 * GEM object so it is entirely possible to gain a
		 * reference on an object in the process of being freed
		 * since our serialisation is via the spinlock and not
		 * the struct_mutex - and consequently use it after it
		 * is freed and then double free it. To prevent that
		 * use-after-free we only acquire a reference on the
		 * object if it is not in the process of being destroyed.
		 */
		mo = container_of(it, struct i915_mmu_object, it);
		if (kref_get_unless_zero(&mo->obj->base.refcount))
			queue_work(mn->wq, &mo->work);

		list_add(&mo->link, &cancelled);
		it = interval_tree_iter_next(it, start, end);
	}
	list_for_each_entry(mo, &cancelled, link)
		del_object(mo);
	spin_unlock(&mn->lock);

	if (!list_empty(&cancelled))
		flush_workqueue(mn->wq);
}

static const struct mmu_notifier_ops i915_gem_userptr_notifier = {
	.invalidate_range_start = i915_gem_userptr_mn_invalidate_range_start,
};

static struct i915_mmu_notifier *
i915_mmu_notifier_create(struct mm_struct *mm)
{
	struct i915_mmu_notifier *mn;
	int ret;

	mn = kmalloc(sizeof(*mn), GFP_KERNEL);
	if (mn == NULL)
		return ERR_PTR(-ENOMEM);

	spin_lock_init(&mn->lock);
	mn->mn.ops = &i915_gem_userptr_notifier;
	mn->objects = RB_ROOT_CACHED;
	mn->wq = alloc_workqueue("i915-userptr-release", WQ_UNBOUND, 0);
	if (mn->wq == NULL) {
		kfree(mn);
		return ERR_PTR(-ENOMEM);
	}

	 /* Protected by mmap_sem (write-lock) */
	ret = __mmu_notifier_register(&mn->mn, mm);
	if (ret) {
		destroy_workqueue(mn->wq);
		kfree(mn);
		return ERR_PTR(ret);
	}

	return mn;
}

static void
i915_gem_userptr_release__mmu_notifier(struct drm_i915_gem_object *obj)
{
	struct i915_mmu_object *mo;

	mo = obj->userptr.mmu_object;
	if (mo == NULL)
		return;

	spin_lock(&mo->mn->lock);
	del_object(mo);
	spin_unlock(&mo->mn->lock);
	kfree(mo);

	obj->userptr.mmu_object = NULL;
}

static struct i915_mmu_notifier *
i915_mmu_notifier_find(struct i915_mm_struct *mm)
{
	struct i915_mmu_notifier *mn = mm->mn;

	mn = mm->mn;
	if (mn)
		return mn;

	down_write(&mm->mm->mmap_sem);
	mutex_lock(&mm->i915->mm_lock);
	if ((mn = mm->mn) == NULL) {
		mn = i915_mmu_notifier_create(mm->mm);
		if (!IS_ERR(mn))
			mm->mn = mn;
	}
	mutex_unlock(&mm->i915->mm_lock);
	up_write(&mm->mm->mmap_sem);

	return mn;
}

static int
i915_gem_userptr_init__mmu_notifier(struct drm_i915_gem_object *obj,
				    unsigned flags)
{
	struct i915_mmu_notifier *mn;
	struct i915_mmu_object *mo;

	if (flags & I915_USERPTR_UNSYNCHRONIZED)
		return capable(CAP_SYS_ADMIN) ? 0 : -EPERM;

	if (WARN_ON(obj->userptr.mm == NULL))
		return -EINVAL;

	mn = i915_mmu_notifier_find(obj->userptr.mm);
	if (IS_ERR(mn))
		return PTR_ERR(mn);

	mo = kzalloc(sizeof(*mo), GFP_KERNEL);
	if (mo == NULL)
		return -ENOMEM;

	mo->mn = mn;
	mo->obj = obj;
	mo->it.start = obj->userptr.ptr;
	mo->it.last = obj->userptr.ptr + obj->base.size - 1;
	INIT_WORK(&mo->work, cancel_userptr);

	obj->userptr.mmu_object = mo;
	return 0;
}

static void
i915_mmu_notifier_free(struct i915_mmu_notifier *mn,
		       struct mm_struct *mm)
{
	if (mn == NULL)
		return;

	mmu_notifier_unregister(&mn->mn, mm);
	destroy_workqueue(mn->wq);
	kfree(mn);
}

#else

static void
i915_gem_userptr_release__mmu_notifier(struct drm_i915_gem_object *obj)
{
}

static int
i915_gem_userptr_init__mmu_notifier(struct drm_i915_gem_object *obj,
				    unsigned flags)
{
	if ((flags & I915_USERPTR_UNSYNCHRONIZED) == 0)
		return -ENODEV;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	return 0;
}

static void
i915_mmu_notifier_free(struct i915_mmu_notifier *mn,
		       struct mm_struct *mm)
{
}

#endif

static struct i915_mm_struct *
__i915_mm_struct_find(struct drm_i915_private *dev_priv, struct mm_struct *real)
{
	struct i915_mm_struct *mm;

	/* Protected by dev_priv->mm_lock */
	hash_for_each_possible(dev_priv->mm_structs, mm, node, (unsigned long)real)
		if (mm->mm == real)
			return mm;

	return NULL;
}

static int
i915_gem_userptr_init__mm_struct(struct drm_i915_gem_object *obj)
{
	struct drm_i915_private *dev_priv = to_i915(obj->base.dev);
	struct i915_mm_struct *mm;
	int ret = 0;

	/* During release of the GEM object we hold the struct_mutex. This
	 * precludes us from calling mmput() at that time as that may be
	 * the last reference and so call exit_mmap(). exit_mmap() will
	 * attempt to reap the vma, and if we were holding a GTT mmap
	 * would then call drm_gem_vm_close() and attempt to reacquire
	 * the struct mutex. So in order to avoid that recursion, we have
	 * to defer releasing the mm reference until after we drop the
	 * struct_mutex, i.e. we need to schedule a worker to do the clean
	 * up.
	 */
	mutex_lock(&dev_priv->mm_lock);
	mm = __i915_mm_struct_find(dev_priv, current->mm);
	if (mm == NULL) {
		mm = kmalloc(sizeof(*mm), GFP_KERNEL);
		if (mm == NULL) {
			ret = -ENOMEM;
			goto out;
		}

		kref_init(&mm->kref);
		mm->i915 = to_i915(obj->base.dev);

		mm->mm = current->mm;
		mmgrab(current->mm);

		mm->mn = NULL;

		/* Protected by dev_priv->mm_lock */
		hash_add(dev_priv->mm_structs,
			 &mm->node, (unsigned long)mm->mm);
	} else
		kref_get(&mm->kref);

	obj->userptr.mm = mm;
out:
	mutex_unlock(&dev_priv->mm_lock);
	return ret;
}

static void
__i915_mm_struct_free__worker(struct work_struct *work)
{
	struct i915_mm_struct *mm = container_of(work, typeof(*mm), work);
	i915_mmu_notifier_free(mm->mn, mm->mm);
	mmdrop(mm->mm);
	kfree(mm);
}

static void
__i915_mm_struct_free(struct kref *kref)
{
	struct i915_mm_struct *mm = container_of(kref, typeof(*mm), kref);

	/* Protected by dev_priv->mm_lock */
	hash_del(&mm->node);
	mutex_unlock(&mm->i915->mm_lock);

	INIT_WORK(&mm->work, __i915_mm_struct_free__worker);
	queue_work(mm->i915->mm.userptr_wq, &mm->work);
}

static void
i915_gem_userptr_release__mm_struct(struct drm_i915_gem_object *obj)
{
	if (obj->userptr.mm == NULL)
		return;

	kref_put_mutex(&obj->userptr.mm->kref,
		       __i915_mm_struct_free,
		       &to_i915(obj->base.dev)->mm_lock);
	obj->userptr.mm = NULL;
}

struct get_pages_work {
	struct work_struct work;
	struct drm_i915_gem_object *obj;
	struct task_struct *task;
};

#if IS_ENABLED(CONFIG_SWIOTLB)
#define swiotlb_active() swiotlb_nr_tbl()
#else
#define swiotlb_active() 0
#endif

static int
st_set_pages(struct sg_table **st, struct page **pvec, int num_pages)
{
	struct scatterlist *sg;
	int ret, n;

	*st = kmalloc(sizeof(**st), GFP_KERNEL);
	if (*st == NULL)
		return -ENOMEM;

	if (swiotlb_active()) {
		ret = sg_alloc_table(*st, num_pages, GFP_KERNEL);
		if (ret)
			goto err;

		for_each_sg((*st)->sgl, sg, num_pages, n)
			sg_set_page(sg, pvec[n], PAGE_SIZE, 0);
	} else {
		ret = sg_alloc_table_from_pages(*st, pvec, num_pages,
						0, num_pages << PAGE_SHIFT,
						GFP_KERNEL);
		if (ret)
			goto err;
	}

	return 0;

err:
	kfree(*st);
	*st = NULL;
	return ret;
}

static struct sg_table *
__i915_gem_userptr_set_pages(struct drm_i915_gem_object *obj,
			     struct page **pvec, int num_pages)
{
	struct sg_table *pages;
	int ret;

	ret = st_set_pages(&pages, pvec, num_pages);
	if (ret)
		return ERR_PTR(ret);

	ret = i915_gem_gtt_prepare_pages(obj, pages);
	if (ret) {
		sg_free_table(pages);
		kfree(pages);
		return ERR_PTR(ret);
	}

	return pages;
}

static int
__i915_gem_userptr_set_active(struct drm_i915_gem_object *obj,
			      bool value)
{
	int ret = 0;

	/* During mm_invalidate_range we need to cancel any userptr that
	 * overlaps the range being invalidated. Doing so requires the
	 * struct_mutex, and that risks recursion. In order to cause
	 * recursion, the user must alias the userptr address space with
	 * a GTT mmapping (possible with a MAP_FIXED) - then when we have
	 * to invalidate that mmaping, mm_invalidate_range is called with
	 * the userptr address *and* the struct_mutex held.  To prevent that
	 * we set a flag under the i915_mmu_notifier spinlock to indicate
	 * whether this object is valid.
	 */
#if defined(CONFIG_MMU_NOTIFIER)
	if (obj->userptr.mmu_object == NULL)
		return 0;

	spin_lock(&obj->userptr.mmu_object->mn->lock);
	/* In order to serialise get_pages with an outstanding
	 * cancel_userptr, we must drop the struct_mutex and try again.
	 */
	if (!value)
		del_object(obj->userptr.mmu_object);
	else if (!work_pending(&obj->userptr.mmu_object->work))
		add_object(obj->userptr.mmu_object);
	else
		ret = -EAGAIN;
	spin_unlock(&obj->userptr.mmu_object->mn->lock);
#endif

	return ret;
}

static void
__i915_gem_userptr_get_pages_worker(struct work_struct *_work)
{
	struct get_pages_work *work = container_of(_work, typeof(*work), work);
	struct drm_i915_gem_object *obj = work->obj;
	const int npages = obj->base.size >> PAGE_SHIFT;
	struct page **pvec;
	int pinned, ret;

	ret = -ENOMEM;
	pinned = 0;

	pvec = kvmalloc_array(npages, sizeof(struct page *), GFP_KERNEL);
	if (pvec != NULL) {
		struct mm_struct *mm = obj->userptr.mm->mm;
		unsigned int flags = 0;

		if (!obj->userptr.read_only)
			flags |= FOLL_WRITE;

		ret = -EFAULT;
		if (mmget_not_zero(mm)) {
			down_read(&mm->mmap_sem);
			while (pinned < npages) {
				ret = get_user_pages_remote
					(work->task, mm,
					 obj->userptr.ptr + pinned * PAGE_SIZE,
					 npages - pinned,
					 flags,
					 pvec + pinned, NULL, NULL);
				if (ret < 0)
					break;

				pinned += ret;
			}
			up_read(&mm->mmap_sem);
			mmput(mm);
		}
	}

	mutex_lock(&obj->mm.lock);
	if (obj->userptr.work == &work->work) {
		struct sg_table *pages = ERR_PTR(ret);

		if (pinned == npages) {
			pages = __i915_gem_userptr_set_pages(obj, pvec, npages);
			if (!IS_ERR(pages)) {
				__i915_gem_object_set_pages(obj, pages);
				pinned = 0;
				pages = NULL;
			}
		}

		obj->userptr.work = ERR_CAST(pages);
		if (IS_ERR(pages))
			__i915_gem_userptr_set_active(obj, false);
	}
	mutex_unlock(&obj->mm.lock);

	release_pages(pvec, pinned, 0);
	kvfree(pvec);

	i915_gem_object_put(obj);
	put_task_struct(work->task);
	kfree(work);
}

static struct sg_table *
__i915_gem_userptr_get_pages_schedule(struct drm_i915_gem_object *obj)
{
	struct get_pages_work *work;

	/* Spawn a worker so that we can acquire the
	 * user pages without holding our mutex. Access
	 * to the user pages requires mmap_sem, and we have
	 * a strict lock ordering of mmap_sem, struct_mutex -
	 * we already hold struct_mutex here and so cannot
	 * call gup without encountering a lock inversion.
	 *
	 * Userspace will keep on repeating the operation
	 * (thanks to EAGAIN) until either we hit the fast
	 * path or the worker completes. If the worker is
	 * cancelled or superseded, the task is still run
	 * but the results ignored. (This leads to
	 * complications that we may have a stray object
	 * refcount that we need to be wary of when
	 * checking for existing objects during creation.)
	 * If the worker encounters an error, it reports
	 * that error back to this function through
	 * obj->userptr.work = ERR_PTR.
	 */
	work = kmalloc(sizeof(*work), GFP_KERNEL);
	if (work == NULL)
		return ERR_PTR(-ENOMEM);

	obj->userptr.work = &work->work;

	work->obj = i915_gem_object_get(obj);

	work->task = current;
	get_task_struct(work->task);

	INIT_WORK(&work->work, __i915_gem_userptr_get_pages_worker);
	queue_work(to_i915(obj->base.dev)->mm.userptr_wq, &work->work);

	return ERR_PTR(-EAGAIN);
}

static struct sg_table *
i915_gem_userptr_get_pages(struct drm_i915_gem_object *obj)
{
	const int num_pages = obj->base.size >> PAGE_SHIFT;
	struct mm_struct *mm = obj->userptr.mm->mm;
	struct page **pvec;
	struct sg_table *pages;
	bool active;
	int pinned;

	/* If userspace should engineer that these pages are replaced in
	 * the vma between us binding this page into the GTT and completion
	 * of rendering... Their loss. If they change the mapping of their
	 * pages they need to create a new bo to point to the new vma.
	 *
	 * However, that still leaves open the possibility of the vma
	 * being copied upon fork. Which falls under the same userspace
	 * synchronisation issue as a regular bo, except that this time
	 * the process may not be expecting that a particular piece of
	 * memory is tied to the GPU.
	 *
	 * Fortunately, we can hook into the mmu_notifier in order to
	 * discard the page references prior to anything nasty happening
	 * to the vma (discard or cloning) which should prevent the more
	 * egregious cases from causing harm.
	 */

	if (obj->userptr.work) {
		/* active flag should still be held for the pending work */
		if (IS_ERR(obj->userptr.work))
			return ERR_CAST(obj->userptr.work);
		else
			return ERR_PTR(-EAGAIN);
	}

	pvec = NULL;
	pinned = 0;

	if (mm == current->mm) {
		pvec = kvmalloc_array(num_pages, sizeof(struct page *),
				      GFP_KERNEL |
				      __GFP_NORETRY |
				      __GFP_NOWARN);
		if (pvec) /* defer to worker if malloc fails */
			pinned = __get_user_pages_fast(obj->userptr.ptr,
						       num_pages,
						       !obj->userptr.read_only,
						       pvec);
	}

	active = false;
	if (pinned < 0) {
		pages = ERR_PTR(pinned);
		pinned = 0;
	} else if (pinned < num_pages) {
		pages = __i915_gem_userptr_get_pages_schedule(obj);
		active = pages == ERR_PTR(-EAGAIN);
	} else {
		pages = __i915_gem_userptr_set_pages(obj, pvec, num_pages);
		active = !IS_ERR(pages);
	}
	if (active)
		__i915_gem_userptr_set_active(obj, true);

	if (IS_ERR(pages))
		release_pages(pvec, pinned, 0);
	kvfree(pvec);

	return pages;
}

static void
i915_gem_userptr_put_pages(struct drm_i915_gem_object *obj,
			   struct sg_table *pages)
{
	struct sgt_iter sgt_iter;
	struct page *page;

	BUG_ON(obj->userptr.work != NULL);
	__i915_gem_userptr_set_active(obj, false);

	if (obj->mm.madv != I915_MADV_WILLNEED)
		obj->mm.dirty = false;

	i915_gem_gtt_finish_pages(obj, pages);

	for_each_sgt_page(page, sgt_iter, pages) {
		if (obj->mm.dirty)
			set_page_dirty(page);

		mark_page_accessed(page);
		put_page(page);
	}
	obj->mm.dirty = false;

	sg_free_table(pages);
	kfree(pages);
}

static void
i915_gem_userptr_release(struct drm_i915_gem_object *obj)
{
	i915_gem_userptr_release__mmu_notifier(obj);
	i915_gem_userptr_release__mm_struct(obj);
}

static int
i915_gem_userptr_dmabuf_export(struct drm_i915_gem_object *obj)
{
	if (obj->userptr.mmu_object)
		return 0;

	return i915_gem_userptr_init__mmu_notifier(obj, 0);
}

static const struct drm_i915_gem_object_ops i915_gem_userptr_ops = {
	.flags = I915_GEM_OBJECT_HAS_STRUCT_PAGE |
		 I915_GEM_OBJECT_IS_SHRINKABLE,
	.get_pages = i915_gem_userptr_get_pages,
	.put_pages = i915_gem_userptr_put_pages,
	.dmabuf_export = i915_gem_userptr_dmabuf_export,
	.release = i915_gem_userptr_release,
};

/**
 * Creates a new mm object that wraps some normal memory from the process
 * context - user memory.
 *
 * We impose several restrictions upon the memory being mapped
 * into the GPU.
 * 1. It must be page aligned (both start/end addresses, i.e ptr and size).
 * 2. It must be normal system memory, not a pointer into another map of IO
 *    space (e.g. it must not be a GTT mmapping of another object).
 * 3. We only allow a bo as large as we could in theory map into the GTT,
 *    that is we limit the size to the total size of the GTT.
 * 4. The bo is marked as being snoopable. The backing pages are left
 *    accessible directly by the CPU, but reads and writes by the GPU may
 *    incur the cost of a snoop (unless you have an LLC architecture).
 *
 * Synchronisation between multiple users and the GPU is left to userspace
 * through the normal set-domain-ioctl. The kernel will enforce that the
 * GPU relinquishes the VMA before it is returned back to the system
 * i.e. upon free(), munmap() or process termination. However, the userspace
 * malloc() library may not immediately relinquish the VMA after free() and
 * instead reuse it whilst the GPU is still reading and writing to the VMA.
 * Caveat emptor.
 *
 * Also note, that the object created here is not currently a "first class"
 * object, in that several ioctls are banned. These are the CPU access
 * ioctls: mmap(), pwrite and pread. In practice, you are expected to use
 * direct access via your pointer rather than use those ioctls. Another
 * restriction is that we do not allow userptr surfaces to be pinned to the
 * hardware and so we reject any attempt to create a framebuffer out of a
 * userptr.
 *
 * If you think this is a good interface to use to pass GPU memory between
 * drivers, please use dma-buf instead. In fact, wherever possible use
 * dma-buf instead.
 */
int
i915_gem_userptr_ioctl(struct drm_device *dev, void *data, struct drm_file *file)
{
	struct drm_i915_private *dev_priv = to_i915(dev);
	struct drm_i915_gem_userptr *args = data;
	struct drm_i915_gem_object *obj;
	int ret;
	u32 handle;

	if (!HAS_LLC(dev_priv) && !HAS_SNOOP(dev_priv)) {
		/* We cannot support coherent userptr objects on hw without
		 * LLC and broken snooping.
		 */
		return -ENODEV;
	}

	if (args->flags & ~(I915_USERPTR_READ_ONLY |
			    I915_USERPTR_UNSYNCHRONIZED))
		return -EINVAL;

	if (offset_in_page(args->user_ptr | args->user_size))
		return -EINVAL;

	if (!access_ok(args->flags & I915_USERPTR_READ_ONLY ? VERIFY_READ : VERIFY_WRITE,
		       (char __user *)(unsigned long)args->user_ptr, args->user_size))
		return -EFAULT;

	if (args->flags & I915_USERPTR_READ_ONLY) {
		/* On almost all of the current hw, we cannot tell the GPU that a
		 * page is readonly, so this is just a placeholder in the uAPI.
		 */
		return -ENODEV;
	}

	obj = i915_gem_object_alloc(dev_priv);
	if (obj == NULL)
		return -ENOMEM;

	drm_gem_private_object_init(dev, &obj->base, args->user_size);
	i915_gem_object_init(obj, &i915_gem_userptr_ops);
	obj->base.read_domains = I915_GEM_DOMAIN_CPU;
	obj->base.write_domain = I915_GEM_DOMAIN_CPU;
	i915_gem_object_set_cache_coherency(obj, I915_CACHE_LLC);

	obj->userptr.ptr = args->user_ptr;
	obj->userptr.read_only = !!(args->flags & I915_USERPTR_READ_ONLY);

	/* And keep a pointer to the current->mm for resolving the user pages
	 * at binding. This means that we need to hook into the mmu_notifier
	 * in order to detect if the mmu is destroyed.
	 */
	ret = i915_gem_userptr_init__mm_struct(obj);
	if (ret == 0)
		ret = i915_gem_userptr_init__mmu_notifier(obj, args->flags);
	if (ret == 0)
		ret = drm_gem_handle_create(file, &obj->base, &handle);

	/* drop reference from allocate - handle holds it now */
	i915_gem_object_put(obj);
	if (ret)
		return ret;

	args->handle = handle;
	return 0;
}

int i915_gem_init_userptr(struct drm_i915_private *dev_priv)
{
	mutex_init(&dev_priv->mm_lock);
	hash_init(dev_priv->mm_structs);

	dev_priv->mm.userptr_wq =
		alloc_workqueue("i915-userptr-acquire", WQ_HIGHPRI, 0);
	if (!dev_priv->mm.userptr_wq)
		return -ENOMEM;

	return 0;
}

void i915_gem_cleanup_userptr(struct drm_i915_private *dev_priv)
{
	destroy_workqueue(dev_priv->mm.userptr_wq);
}