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
|
/*
* Copyright © 2016 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 "i915_vma.h"
#include "i915_drv.h"
#include "intel_ringbuffer.h"
#include "intel_frontbuffer.h"
#include <drm/drm_gem.h>
static void
i915_vma_retire(struct i915_gem_active *active, struct i915_request *rq)
{
const unsigned int idx = rq->engine->id;
struct i915_vma *vma =
container_of(active, struct i915_vma, last_read[idx]);
struct drm_i915_gem_object *obj = vma->obj;
GEM_BUG_ON(!i915_vma_has_active_engine(vma, idx));
i915_vma_clear_active(vma, idx);
if (i915_vma_is_active(vma))
return;
GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
list_move_tail(&vma->vm_link, &vma->vm->inactive_list);
GEM_BUG_ON(!i915_gem_object_is_active(obj));
if (--obj->active_count)
return;
/* Prune the shared fence arrays iff completely idle (inc. external) */
if (reservation_object_trylock(obj->resv)) {
if (reservation_object_test_signaled_rcu(obj->resv, true))
reservation_object_add_excl_fence(obj->resv, NULL);
reservation_object_unlock(obj->resv);
}
/* Bump our place on the bound list to keep it roughly in LRU order
* so that we don't steal from recently used but inactive objects
* (unless we are forced to ofc!)
*/
spin_lock(&rq->i915->mm.obj_lock);
if (obj->bind_count)
list_move_tail(&obj->mm.link, &rq->i915->mm.bound_list);
spin_unlock(&rq->i915->mm.obj_lock);
obj->mm.dirty = true; /* be paranoid */
if (i915_gem_object_has_active_reference(obj)) {
i915_gem_object_clear_active_reference(obj);
i915_gem_object_put(obj);
}
}
static struct i915_vma *
vma_create(struct drm_i915_gem_object *obj,
struct i915_address_space *vm,
const struct i915_ggtt_view *view)
{
struct i915_vma *vma;
struct rb_node *rb, **p;
int i;
/* The aliasing_ppgtt should never be used directly! */
GEM_BUG_ON(vm == &vm->i915->mm.aliasing_ppgtt->base);
vma = kmem_cache_zalloc(vm->i915->vmas, GFP_KERNEL);
if (vma == NULL)
return ERR_PTR(-ENOMEM);
for (i = 0; i < ARRAY_SIZE(vma->last_read); i++)
init_request_active(&vma->last_read[i], i915_vma_retire);
init_request_active(&vma->last_fence, NULL);
vma->vm = vm;
vma->obj = obj;
vma->resv = obj->resv;
vma->size = obj->base.size;
vma->display_alignment = I915_GTT_MIN_ALIGNMENT;
if (view && view->type != I915_GGTT_VIEW_NORMAL) {
vma->ggtt_view = *view;
if (view->type == I915_GGTT_VIEW_PARTIAL) {
GEM_BUG_ON(range_overflows_t(u64,
view->partial.offset,
view->partial.size,
obj->base.size >> PAGE_SHIFT));
vma->size = view->partial.size;
vma->size <<= PAGE_SHIFT;
GEM_BUG_ON(vma->size > obj->base.size);
} else if (view->type == I915_GGTT_VIEW_ROTATED) {
vma->size = intel_rotation_info_size(&view->rotated);
vma->size <<= PAGE_SHIFT;
}
}
if (unlikely(vma->size > vm->total))
goto err_vma;
GEM_BUG_ON(!IS_ALIGNED(vma->size, I915_GTT_PAGE_SIZE));
if (i915_is_ggtt(vm)) {
if (unlikely(overflows_type(vma->size, u32)))
goto err_vma;
vma->fence_size = i915_gem_fence_size(vm->i915, vma->size,
i915_gem_object_get_tiling(obj),
i915_gem_object_get_stride(obj));
if (unlikely(vma->fence_size < vma->size || /* overflow */
vma->fence_size > vm->total))
goto err_vma;
GEM_BUG_ON(!IS_ALIGNED(vma->fence_size, I915_GTT_MIN_ALIGNMENT));
vma->fence_alignment = i915_gem_fence_alignment(vm->i915, vma->size,
i915_gem_object_get_tiling(obj),
i915_gem_object_get_stride(obj));
GEM_BUG_ON(!is_power_of_2(vma->fence_alignment));
/*
* We put the GGTT vma at the start of the vma-list, followed
* by the ppGGTT vma. This allows us to break early when
* iterating over only the GGTT vma for an object, see
* for_each_ggtt_vma()
*/
vma->flags |= I915_VMA_GGTT;
list_add(&vma->obj_link, &obj->vma_list);
} else {
i915_ppgtt_get(i915_vm_to_ppgtt(vm));
list_add_tail(&vma->obj_link, &obj->vma_list);
}
rb = NULL;
p = &obj->vma_tree.rb_node;
while (*p) {
struct i915_vma *pos;
rb = *p;
pos = rb_entry(rb, struct i915_vma, obj_node);
if (i915_vma_compare(pos, vm, view) < 0)
p = &rb->rb_right;
else
p = &rb->rb_left;
}
rb_link_node(&vma->obj_node, rb, p);
rb_insert_color(&vma->obj_node, &obj->vma_tree);
list_add(&vma->vm_link, &vm->unbound_list);
return vma;
err_vma:
kmem_cache_free(vm->i915->vmas, vma);
return ERR_PTR(-E2BIG);
}
static struct i915_vma *
vma_lookup(struct drm_i915_gem_object *obj,
struct i915_address_space *vm,
const struct i915_ggtt_view *view)
{
struct rb_node *rb;
rb = obj->vma_tree.rb_node;
while (rb) {
struct i915_vma *vma = rb_entry(rb, struct i915_vma, obj_node);
long cmp;
cmp = i915_vma_compare(vma, vm, view);
if (cmp == 0)
return vma;
if (cmp < 0)
rb = rb->rb_right;
else
rb = rb->rb_left;
}
return NULL;
}
/**
* i915_vma_instance - return the singleton instance of the VMA
* @obj: parent &struct drm_i915_gem_object to be mapped
* @vm: address space in which the mapping is located
* @view: additional mapping requirements
*
* i915_vma_instance() looks up an existing VMA of the @obj in the @vm with
* the same @view characteristics. If a match is not found, one is created.
* Once created, the VMA is kept until either the object is freed, or the
* address space is closed.
*
* Must be called with struct_mutex held.
*
* Returns the vma, or an error pointer.
*/
struct i915_vma *
i915_vma_instance(struct drm_i915_gem_object *obj,
struct i915_address_space *vm,
const struct i915_ggtt_view *view)
{
struct i915_vma *vma;
lockdep_assert_held(&obj->base.dev->struct_mutex);
GEM_BUG_ON(view && !i915_is_ggtt(vm));
GEM_BUG_ON(vm->closed);
vma = vma_lookup(obj, vm, view);
if (!vma)
vma = vma_create(obj, vm, view);
GEM_BUG_ON(!IS_ERR(vma) && i915_vma_compare(vma, vm, view));
GEM_BUG_ON(!IS_ERR(vma) && vma_lookup(obj, vm, view) != vma);
return vma;
}
/**
* i915_vma_bind - Sets up PTEs for an VMA in it's corresponding address space.
* @vma: VMA to map
* @cache_level: mapping cache level
* @flags: flags like global or local mapping
*
* DMA addresses are taken from the scatter-gather table of this object (or of
* this VMA in case of non-default GGTT views) and PTE entries set up.
* Note that DMA addresses are also the only part of the SG table we care about.
*/
int i915_vma_bind(struct i915_vma *vma, enum i915_cache_level cache_level,
u32 flags)
{
u32 bind_flags;
u32 vma_flags;
int ret;
GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
GEM_BUG_ON(vma->size > vma->node.size);
if (GEM_WARN_ON(range_overflows(vma->node.start,
vma->node.size,
vma->vm->total)))
return -ENODEV;
if (GEM_WARN_ON(!flags))
return -EINVAL;
bind_flags = 0;
if (flags & PIN_GLOBAL)
bind_flags |= I915_VMA_GLOBAL_BIND;
if (flags & PIN_USER)
bind_flags |= I915_VMA_LOCAL_BIND;
vma_flags = vma->flags & (I915_VMA_GLOBAL_BIND | I915_VMA_LOCAL_BIND);
if (flags & PIN_UPDATE)
bind_flags |= vma_flags;
else
bind_flags &= ~vma_flags;
if (bind_flags == 0)
return 0;
GEM_BUG_ON(!vma->pages);
trace_i915_vma_bind(vma, bind_flags);
ret = vma->vm->bind_vma(vma, cache_level, bind_flags);
if (ret)
return ret;
vma->flags |= bind_flags;
return 0;
}
void __iomem *i915_vma_pin_iomap(struct i915_vma *vma)
{
void __iomem *ptr;
int err;
/* Access through the GTT requires the device to be awake. */
assert_rpm_wakelock_held(vma->vm->i915);
lockdep_assert_held(&vma->vm->i915->drm.struct_mutex);
if (WARN_ON(!i915_vma_is_map_and_fenceable(vma))) {
err = -ENODEV;
goto err;
}
GEM_BUG_ON(!i915_vma_is_ggtt(vma));
GEM_BUG_ON((vma->flags & I915_VMA_GLOBAL_BIND) == 0);
ptr = vma->iomap;
if (ptr == NULL) {
ptr = io_mapping_map_wc(&i915_vm_to_ggtt(vma->vm)->iomap,
vma->node.start,
vma->node.size);
if (ptr == NULL) {
err = -ENOMEM;
goto err;
}
vma->iomap = ptr;
}
__i915_vma_pin(vma);
err = i915_vma_pin_fence(vma);
if (err)
goto err_unpin;
i915_vma_set_ggtt_write(vma);
return ptr;
err_unpin:
__i915_vma_unpin(vma);
err:
return IO_ERR_PTR(err);
}
void i915_vma_flush_writes(struct i915_vma *vma)
{
if (!i915_vma_has_ggtt_write(vma))
return;
i915_gem_flush_ggtt_writes(vma->vm->i915);
i915_vma_unset_ggtt_write(vma);
}
void i915_vma_unpin_iomap(struct i915_vma *vma)
{
lockdep_assert_held(&vma->obj->base.dev->struct_mutex);
GEM_BUG_ON(vma->iomap == NULL);
i915_vma_flush_writes(vma);
i915_vma_unpin_fence(vma);
i915_vma_unpin(vma);
}
void i915_vma_unpin_and_release(struct i915_vma **p_vma)
{
struct i915_vma *vma;
struct drm_i915_gem_object *obj;
vma = fetch_and_zero(p_vma);
if (!vma)
return;
obj = vma->obj;
i915_vma_unpin(vma);
i915_vma_close(vma);
__i915_gem_object_release_unless_active(obj);
}
bool i915_vma_misplaced(const struct i915_vma *vma,
u64 size, u64 alignment, u64 flags)
{
if (!drm_mm_node_allocated(&vma->node))
return false;
if (vma->node.size < size)
return true;
GEM_BUG_ON(alignment && !is_power_of_2(alignment));
if (alignment && !IS_ALIGNED(vma->node.start, alignment))
return true;
if (flags & PIN_MAPPABLE && !i915_vma_is_map_and_fenceable(vma))
return true;
if (flags & PIN_OFFSET_BIAS &&
vma->node.start < (flags & PIN_OFFSET_MASK))
return true;
if (flags & PIN_OFFSET_FIXED &&
vma->node.start != (flags & PIN_OFFSET_MASK))
return true;
return false;
}
void __i915_vma_set_map_and_fenceable(struct i915_vma *vma)
{
bool mappable, fenceable;
GEM_BUG_ON(!i915_vma_is_ggtt(vma));
GEM_BUG_ON(!vma->fence_size);
/*
* Explicitly disable for rotated VMA since the display does not
* need the fence and the VMA is not accessible to other users.
*/
if (vma->ggtt_view.type == I915_GGTT_VIEW_ROTATED)
return;
fenceable = (vma->node.size >= vma->fence_size &&
IS_ALIGNED(vma->node.start, vma->fence_alignment));
mappable = vma->node.start + vma->fence_size <= i915_vm_to_ggtt(vma->vm)->mappable_end;
if (mappable && fenceable)
vma->flags |= I915_VMA_CAN_FENCE;
else
vma->flags &= ~I915_VMA_CAN_FENCE;
}
static bool color_differs(struct drm_mm_node *node, unsigned long color)
{
return node->allocated && node->color != color;
}
bool i915_gem_valid_gtt_space(struct i915_vma *vma, unsigned long cache_level)
{
struct drm_mm_node *node = &vma->node;
struct drm_mm_node *other;
/*
* On some machines we have to be careful when putting differing types
* of snoopable memory together to avoid the prefetcher crossing memory
* domains and dying. During vm initialisation, we decide whether or not
* these constraints apply and set the drm_mm.color_adjust
* appropriately.
*/
if (vma->vm->mm.color_adjust == NULL)
return true;
/* Only valid to be called on an already inserted vma */
GEM_BUG_ON(!drm_mm_node_allocated(node));
GEM_BUG_ON(list_empty(&node->node_list));
other = list_prev_entry(node, node_list);
if (color_differs(other, cache_level) && !drm_mm_hole_follows(other))
return false;
other = list_next_entry(node, node_list);
if (color_differs(other, cache_level) && !drm_mm_hole_follows(node))
return false;
return true;
}
/**
* i915_vma_insert - finds a slot for the vma in its address space
* @vma: the vma
* @size: requested size in bytes (can be larger than the VMA)
* @alignment: required alignment
* @flags: mask of PIN_* flags to use
*
* First we try to allocate some free space that meets the requirements for
* the VMA. Failiing that, if the flags permit, it will evict an old VMA,
* preferrably the oldest idle entry to make room for the new VMA.
*
* Returns:
* 0 on success, negative error code otherwise.
*/
static int
i915_vma_insert(struct i915_vma *vma, u64 size, u64 alignment, u64 flags)
{
struct drm_i915_private *dev_priv = vma->vm->i915;
struct drm_i915_gem_object *obj = vma->obj;
u64 start, end;
int ret;
GEM_BUG_ON(i915_vma_is_closed(vma));
GEM_BUG_ON(vma->flags & (I915_VMA_GLOBAL_BIND | I915_VMA_LOCAL_BIND));
GEM_BUG_ON(drm_mm_node_allocated(&vma->node));
size = max(size, vma->size);
alignment = max(alignment, vma->display_alignment);
if (flags & PIN_MAPPABLE) {
size = max_t(typeof(size), size, vma->fence_size);
alignment = max_t(typeof(alignment),
alignment, vma->fence_alignment);
}
GEM_BUG_ON(!IS_ALIGNED(size, I915_GTT_PAGE_SIZE));
GEM_BUG_ON(!IS_ALIGNED(alignment, I915_GTT_MIN_ALIGNMENT));
GEM_BUG_ON(!is_power_of_2(alignment));
start = flags & PIN_OFFSET_BIAS ? flags & PIN_OFFSET_MASK : 0;
GEM_BUG_ON(!IS_ALIGNED(start, I915_GTT_PAGE_SIZE));
end = vma->vm->total;
if (flags & PIN_MAPPABLE)
end = min_t(u64, end, dev_priv->ggtt.mappable_end);
if (flags & PIN_ZONE_4G)
end = min_t(u64, end, (1ULL << 32) - I915_GTT_PAGE_SIZE);
GEM_BUG_ON(!IS_ALIGNED(end, I915_GTT_PAGE_SIZE));
/* If binding the object/GGTT view requires more space than the entire
* aperture has, reject it early before evicting everything in a vain
* attempt to find space.
*/
if (size > end) {
DRM_DEBUG("Attempting to bind an object larger than the aperture: request=%llu [object=%zd] > %s aperture=%llu\n",
size, obj->base.size,
flags & PIN_MAPPABLE ? "mappable" : "total",
end);
return -ENOSPC;
}
ret = i915_gem_object_pin_pages(obj);
if (ret)
return ret;
GEM_BUG_ON(vma->pages);
ret = vma->vm->set_pages(vma);
if (ret)
goto err_unpin;
if (flags & PIN_OFFSET_FIXED) {
u64 offset = flags & PIN_OFFSET_MASK;
if (!IS_ALIGNED(offset, alignment) ||
range_overflows(offset, size, end)) {
ret = -EINVAL;
goto err_clear;
}
ret = i915_gem_gtt_reserve(vma->vm, &vma->node,
size, offset, obj->cache_level,
flags);
if (ret)
goto err_clear;
} else {
/*
* We only support huge gtt pages through the 48b PPGTT,
* however we also don't want to force any alignment for
* objects which need to be tightly packed into the low 32bits.
*
* Note that we assume that GGTT are limited to 4GiB for the
* forseeable future. See also i915_ggtt_offset().
*/
if (upper_32_bits(end - 1) &&
vma->page_sizes.sg > I915_GTT_PAGE_SIZE) {
/*
* We can't mix 64K and 4K PTEs in the same page-table
* (2M block), and so to avoid the ugliness and
* complexity of coloring we opt for just aligning 64K
* objects to 2M.
*/
u64 page_alignment =
rounddown_pow_of_two(vma->page_sizes.sg |
I915_GTT_PAGE_SIZE_2M);
/*
* Check we don't expand for the limited Global GTT
* (mappable aperture is even more precious!). This
* also checks that we exclude the aliasing-ppgtt.
*/
GEM_BUG_ON(i915_vma_is_ggtt(vma));
alignment = max(alignment, page_alignment);
if (vma->page_sizes.sg & I915_GTT_PAGE_SIZE_64K)
size = round_up(size, I915_GTT_PAGE_SIZE_2M);
}
ret = i915_gem_gtt_insert(vma->vm, &vma->node,
size, alignment, obj->cache_level,
start, end, flags);
if (ret)
goto err_clear;
GEM_BUG_ON(vma->node.start < start);
GEM_BUG_ON(vma->node.start + vma->node.size > end);
}
GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
GEM_BUG_ON(!i915_gem_valid_gtt_space(vma, obj->cache_level));
list_move_tail(&vma->vm_link, &vma->vm->inactive_list);
spin_lock(&dev_priv->mm.obj_lock);
list_move_tail(&obj->mm.link, &dev_priv->mm.bound_list);
obj->bind_count++;
spin_unlock(&dev_priv->mm.obj_lock);
GEM_BUG_ON(atomic_read(&obj->mm.pages_pin_count) < obj->bind_count);
return 0;
err_clear:
vma->vm->clear_pages(vma);
err_unpin:
i915_gem_object_unpin_pages(obj);
return ret;
}
static void
i915_vma_remove(struct i915_vma *vma)
{
struct drm_i915_private *i915 = vma->vm->i915;
struct drm_i915_gem_object *obj = vma->obj;
GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
GEM_BUG_ON(vma->flags & (I915_VMA_GLOBAL_BIND | I915_VMA_LOCAL_BIND));
vma->vm->clear_pages(vma);
drm_mm_remove_node(&vma->node);
list_move_tail(&vma->vm_link, &vma->vm->unbound_list);
/* Since the unbound list is global, only move to that list if
* no more VMAs exist.
*/
spin_lock(&i915->mm.obj_lock);
if (--obj->bind_count == 0)
list_move_tail(&obj->mm.link, &i915->mm.unbound_list);
spin_unlock(&i915->mm.obj_lock);
/* And finally now the object is completely decoupled from this vma,
* we can drop its hold on the backing storage and allow it to be
* reaped by the shrinker.
*/
i915_gem_object_unpin_pages(obj);
GEM_BUG_ON(atomic_read(&obj->mm.pages_pin_count) < obj->bind_count);
}
int __i915_vma_do_pin(struct i915_vma *vma,
u64 size, u64 alignment, u64 flags)
{
const unsigned int bound = vma->flags;
int ret;
lockdep_assert_held(&vma->vm->i915->drm.struct_mutex);
GEM_BUG_ON((flags & (PIN_GLOBAL | PIN_USER)) == 0);
GEM_BUG_ON((flags & PIN_GLOBAL) && !i915_vma_is_ggtt(vma));
if (WARN_ON(bound & I915_VMA_PIN_OVERFLOW)) {
ret = -EBUSY;
goto err_unpin;
}
if ((bound & I915_VMA_BIND_MASK) == 0) {
ret = i915_vma_insert(vma, size, alignment, flags);
if (ret)
goto err_unpin;
}
GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
ret = i915_vma_bind(vma, vma->obj->cache_level, flags);
if (ret)
goto err_remove;
GEM_BUG_ON((vma->flags & I915_VMA_BIND_MASK) == 0);
if ((bound ^ vma->flags) & I915_VMA_GLOBAL_BIND)
__i915_vma_set_map_and_fenceable(vma);
GEM_BUG_ON(i915_vma_misplaced(vma, size, alignment, flags));
return 0;
err_remove:
if ((bound & I915_VMA_BIND_MASK) == 0) {
i915_vma_remove(vma);
GEM_BUG_ON(vma->pages);
GEM_BUG_ON(vma->flags & I915_VMA_BIND_MASK);
}
err_unpin:
__i915_vma_unpin(vma);
return ret;
}
void i915_vma_close(struct i915_vma *vma)
{
lockdep_assert_held(&vma->vm->i915->drm.struct_mutex);
GEM_BUG_ON(i915_vma_is_closed(vma));
vma->flags |= I915_VMA_CLOSED;
/*
* We defer actually closing, unbinding and destroying the VMA until
* the next idle point, or if the object is freed in the meantime. By
* postponing the unbind, we allow for it to be resurrected by the
* client, avoiding the work required to rebind the VMA. This is
* advantageous for DRI, where the client/server pass objects
* between themselves, temporarily opening a local VMA to the
* object, and then closing it again. The same object is then reused
* on the next frame (or two, depending on the depth of the swap queue)
* causing us to rebind the VMA once more. This ends up being a lot
* of wasted work for the steady state.
*/
list_add_tail(&vma->closed_link, &vma->vm->i915->gt.closed_vma);
}
void i915_vma_reopen(struct i915_vma *vma)
{
lockdep_assert_held(&vma->vm->i915->drm.struct_mutex);
if (vma->flags & I915_VMA_CLOSED) {
vma->flags &= ~I915_VMA_CLOSED;
list_del(&vma->closed_link);
}
}
static void __i915_vma_destroy(struct i915_vma *vma)
{
int i;
GEM_BUG_ON(vma->node.allocated);
GEM_BUG_ON(vma->fence);
for (i = 0; i < ARRAY_SIZE(vma->last_read); i++)
GEM_BUG_ON(i915_gem_active_isset(&vma->last_read[i]));
GEM_BUG_ON(i915_gem_active_isset(&vma->last_fence));
list_del(&vma->obj_link);
list_del(&vma->vm_link);
rb_erase(&vma->obj_node, &vma->obj->vma_tree);
if (!i915_vma_is_ggtt(vma))
i915_ppgtt_put(i915_vm_to_ppgtt(vma->vm));
kmem_cache_free(to_i915(vma->obj->base.dev)->vmas, vma);
}
void i915_vma_destroy(struct i915_vma *vma)
{
lockdep_assert_held(&vma->vm->i915->drm.struct_mutex);
GEM_BUG_ON(i915_vma_is_active(vma));
GEM_BUG_ON(i915_vma_is_pinned(vma));
if (i915_vma_is_closed(vma))
list_del(&vma->closed_link);
WARN_ON(i915_vma_unbind(vma));
__i915_vma_destroy(vma);
}
void i915_vma_parked(struct drm_i915_private *i915)
{
struct i915_vma *vma, *next;
list_for_each_entry_safe(vma, next, &i915->gt.closed_vma, closed_link) {
GEM_BUG_ON(!i915_vma_is_closed(vma));
i915_vma_destroy(vma);
}
GEM_BUG_ON(!list_empty(&i915->gt.closed_vma));
}
static void __i915_vma_iounmap(struct i915_vma *vma)
{
GEM_BUG_ON(i915_vma_is_pinned(vma));
if (vma->iomap == NULL)
return;
io_mapping_unmap(vma->iomap);
vma->iomap = NULL;
}
void i915_vma_revoke_mmap(struct i915_vma *vma)
{
struct drm_vma_offset_node *node = &vma->obj->base.vma_node;
u64 vma_offset;
lockdep_assert_held(&vma->vm->i915->drm.struct_mutex);
if (!i915_vma_has_userfault(vma))
return;
GEM_BUG_ON(!i915_vma_is_map_and_fenceable(vma));
GEM_BUG_ON(!vma->obj->userfault_count);
vma_offset = vma->ggtt_view.partial.offset << PAGE_SHIFT;
unmap_mapping_range(vma->vm->i915->drm.anon_inode->i_mapping,
drm_vma_node_offset_addr(node) + vma_offset,
vma->size,
1);
i915_vma_unset_userfault(vma);
if (!--vma->obj->userfault_count)
list_del(&vma->obj->userfault_link);
}
int i915_vma_unbind(struct i915_vma *vma)
{
struct drm_i915_gem_object *obj = vma->obj;
unsigned long active;
int ret;
lockdep_assert_held(&obj->base.dev->struct_mutex);
/* First wait upon any activity as retiring the request may
* have side-effects such as unpinning or even unbinding this vma.
*/
might_sleep();
active = i915_vma_get_active(vma);
if (active) {
int idx;
/* When a closed VMA is retired, it is unbound - eek.
* In order to prevent it from being recursively closed,
* take a pin on the vma so that the second unbind is
* aborted.
*
* Even more scary is that the retire callback may free
* the object (last active vma). To prevent the explosion
* we defer the actual object free to a worker that can
* only proceed once it acquires the struct_mutex (which
* we currently hold, therefore it cannot free this object
* before we are finished).
*/
__i915_vma_pin(vma);
for_each_active(active, idx) {
ret = i915_gem_active_retire(&vma->last_read[idx],
&vma->vm->i915->drm.struct_mutex);
if (ret)
break;
}
if (!ret) {
ret = i915_gem_active_retire(&vma->last_fence,
&vma->vm->i915->drm.struct_mutex);
}
__i915_vma_unpin(vma);
if (ret)
return ret;
}
GEM_BUG_ON(i915_vma_is_active(vma));
if (i915_vma_is_pinned(vma))
return -EBUSY;
if (!drm_mm_node_allocated(&vma->node))
return 0;
GEM_BUG_ON(obj->bind_count == 0);
GEM_BUG_ON(!i915_gem_object_has_pinned_pages(obj));
if (i915_vma_is_map_and_fenceable(vma)) {
/*
* Check that we have flushed all writes through the GGTT
* before the unbind, other due to non-strict nature of those
* indirect writes they may end up referencing the GGTT PTE
* after the unbind.
*/
i915_vma_flush_writes(vma);
GEM_BUG_ON(i915_vma_has_ggtt_write(vma));
/* release the fence reg _after_ flushing */
ret = i915_vma_put_fence(vma);
if (ret)
return ret;
/* Force a pagefault for domain tracking on next user access */
i915_vma_revoke_mmap(vma);
__i915_vma_iounmap(vma);
vma->flags &= ~I915_VMA_CAN_FENCE;
}
GEM_BUG_ON(vma->fence);
GEM_BUG_ON(i915_vma_has_userfault(vma));
if (likely(!vma->vm->closed)) {
trace_i915_vma_unbind(vma);
vma->vm->unbind_vma(vma);
}
vma->flags &= ~(I915_VMA_GLOBAL_BIND | I915_VMA_LOCAL_BIND);
i915_vma_remove(vma);
return 0;
}
#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
#include "selftests/i915_vma.c"
#endif
|