ttm_bo.c 44.0 KB
Newer Older
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
/**************************************************************************
 *
 * Copyright (c) 2006-2009 VMware, Inc., Palo Alto, CA., USA
 * All Rights Reserved.
 *
 * 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, sub license, 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 NON-INFRINGEMENT. IN NO EVENT SHALL
 * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS 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.
 *
 **************************************************************************/
/*
 * Authors: Thomas Hellstrom <thellstrom-at-vmware-dot-com>
 */

J
Joe Perches 已提交
31 32
#define pr_fmt(fmt) "[TTM] " fmt

33 34 35
#include <drm/ttm/ttm_module.h>
#include <drm/ttm/ttm_bo_driver.h>
#include <drm/ttm/ttm_placement.h>
36 37 38 39 40 41
#include <linux/jiffies.h>
#include <linux/slab.h>
#include <linux/sched.h>
#include <linux/mm.h>
#include <linux/file.h>
#include <linux/module.h>
A
Arun Sharma 已提交
42
#include <linux/atomic.h>
43
#include <linux/reservation.h>
44 45 46 47 48 49

#define TTM_ASSERT_LOCKED(param)
#define TTM_DEBUG(fmt, arg...)
#define TTM_BO_HASH_ORDER 13

static int ttm_bo_swapout(struct ttm_mem_shrink *shrink);
50 51 52 53 54 55 56
static void ttm_bo_global_kobj_release(struct kobject *kobj);

static struct attribute ttm_bo_count = {
	.name = "bo_count",
	.mode = S_IRUGO
};

57 58
static inline int ttm_mem_type_from_place(const struct ttm_place *place,
					  uint32_t *mem_type)
59
{
60
	int pos;
61

62 63 64 65 66 67
	pos = ffs(place->flags & TTM_PL_MASK_MEM);
	if (unlikely(!pos))
		return -EINVAL;

	*mem_type = pos - 1;
	return 0;
68 69
}

70
static void ttm_mem_type_debug(struct ttm_bo_device *bdev, int mem_type)
71
{
72
	struct ttm_mem_type_manager *man = &bdev->man[mem_type];
73
	struct drm_printer p = drm_debug_printer(TTM_PFX);
74

J
Joe Perches 已提交
75 76 77
	pr_err("    has_type: %d\n", man->has_type);
	pr_err("    use_type: %d\n", man->use_type);
	pr_err("    flags: 0x%08X\n", man->flags);
78
	pr_err("    gpu_offset: 0x%08llX\n", man->gpu_offset);
J
Joe Perches 已提交
79 80 81
	pr_err("    size: %llu\n", man->size);
	pr_err("    available_caching: 0x%08X\n", man->available_caching);
	pr_err("    default_caching: 0x%08X\n", man->default_caching);
82
	if (mem_type != TTM_PL_SYSTEM)
83
		(*man->func->debug)(man, &p);
84 85 86 87 88 89 90
}

static void ttm_bo_mem_space_debug(struct ttm_buffer_object *bo,
					struct ttm_placement *placement)
{
	int i, ret, mem_type;

J
Joe Perches 已提交
91 92 93
	pr_err("No space for %p (%lu pages, %luK, %luM)\n",
	       bo, bo->mem.num_pages, bo->mem.size >> 10,
	       bo->mem.size >> 20);
94
	for (i = 0; i < placement->num_placement; i++) {
95
		ret = ttm_mem_type_from_place(&placement->placement[i],
96 97 98
						&mem_type);
		if (ret)
			return;
J
Joe Perches 已提交
99
		pr_err("  placement[%d]=0x%08X (%d)\n",
100
		       i, placement->placement[i].flags, mem_type);
101
		ttm_mem_type_debug(bo->bdev, mem_type);
102 103 104
	}
}

105 106 107 108 109 110 111
static ssize_t ttm_bo_global_show(struct kobject *kobj,
				  struct attribute *attr,
				  char *buffer)
{
	struct ttm_bo_global *glob =
		container_of(kobj, struct ttm_bo_global, kobj);

112 113
	return snprintf(buffer, PAGE_SIZE, "%d\n",
				atomic_read(&glob->bo_count));
114 115 116 117 118 119 120
}

static struct attribute *ttm_bo_global_attrs[] = {
	&ttm_bo_count,
	NULL
};

121
static const struct sysfs_ops ttm_bo_global_ops = {
122 123 124 125 126 127 128 129 130
	.show = &ttm_bo_global_show
};

static struct kobj_type ttm_bo_glob_kobj_type  = {
	.release = &ttm_bo_global_kobj_release,
	.sysfs_ops = &ttm_bo_global_ops,
	.default_attrs = ttm_bo_global_attrs
};

131 132 133 134 135 136 137 138 139 140 141

static inline uint32_t ttm_bo_type_flags(unsigned type)
{
	return 1 << (type);
}

static void ttm_bo_release_list(struct kref *list_kref)
{
	struct ttm_buffer_object *bo =
	    container_of(list_kref, struct ttm_buffer_object, list_kref);
	struct ttm_bo_device *bdev = bo->bdev;
142
	size_t acc_size = bo->acc_size;
143

144 145
	BUG_ON(kref_read(&bo->list_kref));
	BUG_ON(kref_read(&bo->kref));
146 147 148 149
	BUG_ON(atomic_read(&bo->cpu_writers));
	BUG_ON(bo->mem.mm_node != NULL);
	BUG_ON(!list_empty(&bo->lru));
	BUG_ON(!list_empty(&bo->ddestroy));
150
	ttm_tt_destroy(bo->ttm);
151
	atomic_dec(&bo->glob->bo_count);
152
	dma_fence_put(bo->moving);
153
	reservation_object_fini(&bo->ttm_resv);
154
	mutex_destroy(&bo->wu_mutex);
155 156 157 158 159
	if (bo->destroy)
		bo->destroy(bo);
	else {
		kfree(bo);
	}
160
	ttm_mem_global_free(bdev->glob->mem_glob, acc_size);
161 162
}

163
void ttm_bo_add_to_lru(struct ttm_buffer_object *bo)
164 165
{
	struct ttm_bo_device *bdev = bo->bdev;
166
	struct ttm_mem_type_manager *man;
167

168
	lockdep_assert_held(&bo->resv->lock.base);
169 170 171 172 173

	if (!(bo->mem.placement & TTM_PL_FLAG_NO_EVICT)) {

		BUG_ON(!list_empty(&bo->lru));

174 175
		man = &bdev->man[bo->mem.mem_type];
		list_add_tail(&bo->lru, &man->lru[bo->priority]);
176 177
		kref_get(&bo->list_kref);

178
		if (bo->ttm && !(bo->ttm->page_flags & TTM_PAGE_FLAG_SG)) {
179 180
			list_add_tail(&bo->swap,
				      &bo->glob->swap_lru[bo->priority]);
181 182 183 184
			kref_get(&bo->list_kref);
		}
	}
}
185
EXPORT_SYMBOL(ttm_bo_add_to_lru);
186

187 188 189 190 191 192
static void ttm_bo_ref_bug(struct kref *list_kref)
{
	BUG();
}

void ttm_bo_del_from_lru(struct ttm_buffer_object *bo)
193 194 195
{
	if (!list_empty(&bo->swap)) {
		list_del_init(&bo->swap);
196
		kref_put(&bo->list_kref, ttm_bo_ref_bug);
197 198 199
	}
	if (!list_empty(&bo->lru)) {
		list_del_init(&bo->lru);
200
		kref_put(&bo->list_kref, ttm_bo_ref_bug);
201 202
	}

203 204 205 206
	/*
	 * TODO: Add a driver hook to delete from
	 * driver-specific LRU's here.
	 */
207 208
}

209
void ttm_bo_del_sub_from_lru(struct ttm_buffer_object *bo)
210
{
211
	spin_lock(&bo->glob->lru_lock);
212
	ttm_bo_del_from_lru(bo);
213
	spin_unlock(&bo->glob->lru_lock);
214
}
215
EXPORT_SYMBOL(ttm_bo_del_sub_from_lru);
216

217 218 219 220
void ttm_bo_move_to_lru_tail(struct ttm_buffer_object *bo)
{
	lockdep_assert_held(&bo->resv->lock.base);

221
	ttm_bo_del_from_lru(bo);
222
	ttm_bo_add_to_lru(bo);
223 224 225
}
EXPORT_SYMBOL(ttm_bo_move_to_lru_tail);

226 227 228 229 230 231
/*
 * Call bo->mutex locked.
 */
static int ttm_bo_add_ttm(struct ttm_buffer_object *bo, bool zero_alloc)
{
	struct ttm_bo_device *bdev = bo->bdev;
232
	struct ttm_bo_global *glob = bo->glob;
233 234 235 236 237 238
	int ret = 0;
	uint32_t page_flags = 0;

	TTM_ASSERT_LOCKED(&bo->mutex);
	bo->ttm = NULL;

D
Dave Airlie 已提交
239 240 241
	if (bdev->need_dma32)
		page_flags |= TTM_PAGE_FLAG_DMA32;

242 243 244 245 246
	switch (bo->type) {
	case ttm_bo_type_device:
		if (zero_alloc)
			page_flags |= TTM_PAGE_FLAG_ZERO_ALLOC;
	case ttm_bo_type_kernel:
247 248
		bo->ttm = bdev->driver->ttm_tt_create(bdev, bo->num_pages << PAGE_SHIFT,
						      page_flags, glob->dummy_read_page);
249 250 251
		if (unlikely(bo->ttm == NULL))
			ret = -ENOMEM;
		break;
252 253 254 255 256 257 258 259 260 261
	case ttm_bo_type_sg:
		bo->ttm = bdev->driver->ttm_tt_create(bdev, bo->num_pages << PAGE_SHIFT,
						      page_flags | TTM_PAGE_FLAG_SG,
						      glob->dummy_read_page);
		if (unlikely(bo->ttm == NULL)) {
			ret = -ENOMEM;
			break;
		}
		bo->ttm->sg = bo->sg;
		break;
262
	default:
J
Joe Perches 已提交
263
		pr_err("Illegal buffer object type\n");
264 265 266 267 268 269 270 271 272
		ret = -EINVAL;
		break;
	}

	return ret;
}

static int ttm_bo_handle_move_mem(struct ttm_buffer_object *bo,
				  struct ttm_mem_reg *mem,
273
				  bool evict, bool interruptible,
274
				  bool no_wait_gpu)
275 276 277 278 279 280 281 282 283
{
	struct ttm_bo_device *bdev = bo->bdev;
	bool old_is_pci = ttm_mem_reg_is_pci(bdev, &bo->mem);
	bool new_is_pci = ttm_mem_reg_is_pci(bdev, mem);
	struct ttm_mem_type_manager *old_man = &bdev->man[bo->mem.mem_type];
	struct ttm_mem_type_manager *new_man = &bdev->man[mem->mem_type];
	int ret = 0;

	if (old_is_pci || new_is_pci ||
284 285 286 287 288 289 290
	    ((mem->placement & bo->mem.placement & TTM_PL_MASK_CACHING) == 0)) {
		ret = ttm_mem_io_lock(old_man, true);
		if (unlikely(ret != 0))
			goto out_err;
		ttm_bo_unmap_virtual_locked(bo);
		ttm_mem_io_unlock(old_man);
	}
291 292 293 294 295

	/*
	 * Create and bind a ttm if required.
	 */

296 297
	if (!(new_man->flags & TTM_MEMTYPE_FLAG_FIXED)) {
		if (bo->ttm == NULL) {
298 299
			bool zero = !(old_man->flags & TTM_MEMTYPE_FLAG_FIXED);
			ret = ttm_bo_add_ttm(bo, zero);
300 301 302
			if (ret)
				goto out_err;
		}
303 304 305

		ret = ttm_tt_set_placement_caching(bo->ttm, mem->placement);
		if (ret)
306
			goto out_err;
307 308 309 310 311 312 313 314

		if (mem->mem_type != TTM_PL_SYSTEM) {
			ret = ttm_tt_bind(bo->ttm, mem);
			if (ret)
				goto out_err;
		}

		if (bo->mem.mem_type == TTM_PL_SYSTEM) {
315
			if (bdev->driver->move_notify)
316
				bdev->driver->move_notify(bo, evict, mem);
317
			bo->mem = *mem;
318 319 320 321 322
			mem->mm_node = NULL;
			goto moved;
		}
	}

323
	if (bdev->driver->move_notify)
324
		bdev->driver->move_notify(bo, evict, mem);
325

326 327
	if (!(old_man->flags & TTM_MEMTYPE_FLAG_FIXED) &&
	    !(new_man->flags & TTM_MEMTYPE_FLAG_FIXED))
328
		ret = ttm_bo_move_ttm(bo, interruptible, no_wait_gpu, mem);
329 330
	else if (bdev->driver->move)
		ret = bdev->driver->move(bo, evict, interruptible,
331
					 no_wait_gpu, mem);
332
	else
333
		ret = ttm_bo_move_memcpy(bo, interruptible, no_wait_gpu, mem);
334

335 336 337 338 339
	if (ret) {
		if (bdev->driver->move_notify) {
			struct ttm_mem_reg tmp_mem = *mem;
			*mem = bo->mem;
			bo->mem = tmp_mem;
340
			bdev->driver->move_notify(bo, false, mem);
341
			bo->mem = *mem;
342
			*mem = tmp_mem;
343
		}
344

345 346
		goto out_err;
	}
347

348 349
moved:
	if (bo->evicted) {
350 351 352 353 354
		if (bdev->driver->invalidate_caches) {
			ret = bdev->driver->invalidate_caches(bdev, bo->mem.placement);
			if (ret)
				pr_err("Can not flush read caches\n");
		}
355 356 357
		bo->evicted = false;
	}

C
Christian König 已提交
358
	if (bo->mem.mm_node)
359
		bo->offset = (bo->mem.start << PAGE_SHIFT) +
360
		    bdev->man[bo->mem.mem_type].gpu_offset;
C
Christian König 已提交
361
	else
362
		bo->offset = 0;
363 364 365 366 367

	return 0;

out_err:
	new_man = &bdev->man[bo->mem.mem_type];
368
	if (new_man->flags & TTM_MEMTYPE_FLAG_FIXED) {
369 370 371 372 373 374 375
		ttm_tt_destroy(bo->ttm);
		bo->ttm = NULL;
	}

	return ret;
}

376
/**
377
 * Call bo::reserved.
378
 * Will release GPU memory type usage on destruction.
379 380 381
 * This is the place to put in driver specific hooks to release
 * driver private resources.
 * Will release the bo::reserved lock.
382 383 384 385
 */

static void ttm_bo_cleanup_memtype_use(struct ttm_buffer_object *bo)
{
386
	if (bo->bdev->driver->move_notify)
387
		bo->bdev->driver->move_notify(bo, false, NULL);
388

389 390
	ttm_tt_destroy(bo->ttm);
	bo->ttm = NULL;
391
	ttm_bo_mem_put(bo, &bo->mem);
392 393
}

394 395 396 397 398 399 400
static int ttm_bo_individualize_resv(struct ttm_buffer_object *bo)
{
	int r;

	if (bo->resv == &bo->ttm_resv)
		return 0;

401
	BUG_ON(!reservation_object_trylock(&bo->ttm_resv));
402 403

	r = reservation_object_copy_fences(&bo->ttm_resv, bo->resv);
404
	if (r)
405 406 407 408 409
		reservation_object_unlock(&bo->ttm_resv);

	return r;
}

410 411 412
static void ttm_bo_flush_all_fences(struct ttm_buffer_object *bo)
{
	struct reservation_object_list *fobj;
413
	struct dma_fence *fence;
414 415
	int i;

416 417
	fobj = reservation_object_get_list(&bo->ttm_resv);
	fence = reservation_object_get_excl(&bo->ttm_resv);
418
	if (fence && !fence->ops->signaled)
419
		dma_fence_enable_sw_signaling(fence);
420 421 422 423 424 425

	for (i = 0; fobj && i < fobj->shared_count; ++i) {
		fence = rcu_dereference_protected(fobj->shared[i],
					reservation_object_held(bo->resv));

		if (!fence->ops->signaled)
426
			dma_fence_enable_sw_signaling(fence);
427 428 429
	}
}

430
static void ttm_bo_cleanup_refs_or_queue(struct ttm_buffer_object *bo)
431 432
{
	struct ttm_bo_device *bdev = bo->bdev;
433
	struct ttm_bo_global *glob = bo->glob;
434 435
	int ret;

436 437 438 439 440 441 442 443 444 445 446
	ret = ttm_bo_individualize_resv(bo);
	if (ret) {
		/* Last resort, if we fail to allocate memory for the
		 * fences block for the BO to become idle
		 */
		reservation_object_wait_timeout_rcu(bo->resv, true, false,
						    30 * HZ);
		spin_lock(&glob->lru_lock);
		goto error;
	}

447
	spin_lock(&glob->lru_lock);
448
	ret = reservation_object_trylock(bo->resv) ? 0 : -EBUSY;
M
Maarten Lankhorst 已提交
449
	if (!ret) {
450
		if (reservation_object_test_signaled_rcu(&bo->ttm_resv, true)) {
451
			ttm_bo_del_from_lru(bo);
M
Maarten Lankhorst 已提交
452
			spin_unlock(&glob->lru_lock);
453
			if (bo->resv != &bo->ttm_resv)
454
				reservation_object_unlock(&bo->ttm_resv);
455

M
Maarten Lankhorst 已提交
456
			ttm_bo_cleanup_memtype_use(bo);
457
			reservation_object_unlock(bo->resv);
M
Maarten Lankhorst 已提交
458
			return;
459 460 461
		}

		ttm_bo_flush_all_fences(bo);
462 463 464 465 466 467 468 469 470 471 472

		/*
		 * Make NO_EVICT bos immediately available to
		 * shrinkers, now that they are queued for
		 * destruction.
		 */
		if (bo->mem.placement & TTM_PL_FLAG_NO_EVICT) {
			bo->mem.placement &= ~TTM_PL_FLAG_NO_EVICT;
			ttm_bo_add_to_lru(bo);
		}

473
		reservation_object_unlock(bo->resv);
474
	}
475 476
	if (bo->resv != &bo->ttm_resv)
		reservation_object_unlock(&bo->ttm_resv);
477

478
error:
479 480 481 482 483 484 485 486 487
	kref_get(&bo->list_kref);
	list_add_tail(&bo->ddestroy, &bdev->ddestroy);
	spin_unlock(&glob->lru_lock);

	schedule_delayed_work(&bdev->wq,
			      ((HZ / 100) < 1) ? 1 : HZ / 100);
}

/**
488
 * function ttm_bo_cleanup_refs
489 490 491
 * If bo idle, remove from delayed- and lru lists, and unref.
 * If not idle, do nothing.
 *
492
 * Must be called with lru_lock and reservation held, this function
493
 * will drop the lru lock and optionally the reservation lock before returning.
494
 *
495 496
 * @interruptible         Any sleeps should occur interruptibly.
 * @no_wait_gpu           Never wait for gpu. Return -EBUSY instead.
497
 * @unlock_resv           Unlock the reservation lock as well.
498 499
 */

500 501 502
static int ttm_bo_cleanup_refs(struct ttm_buffer_object *bo,
			       bool interruptible, bool no_wait_gpu,
			       bool unlock_resv)
503 504
{
	struct ttm_bo_global *glob = bo->glob;
505
	struct reservation_object *resv;
506
	int ret;
507

508 509 510 511 512 513 514 515 516
	if (unlikely(list_empty(&bo->ddestroy)))
		resv = bo->resv;
	else
		resv = &bo->ttm_resv;

	if (reservation_object_test_signaled_rcu(resv, true))
		ret = 0;
	else
		ret = -EBUSY;
517

518
	if (ret && !no_wait_gpu) {
M
Maarten Lankhorst 已提交
519
		long lret;
520

521 522
		if (unlock_resv)
			reservation_object_unlock(bo->resv);
M
Maarten Lankhorst 已提交
523 524
		spin_unlock(&glob->lru_lock);

525
		lret = reservation_object_wait_timeout_rcu(resv, true,
M
Maarten Lankhorst 已提交
526 527 528 529 530 531 532
							   interruptible,
							   30 * HZ);

		if (lret < 0)
			return lret;
		else if (lret == 0)
			return -EBUSY;
533

534
		spin_lock(&glob->lru_lock);
535 536 537 538 539 540 541 542 543
		if (unlock_resv && !reservation_object_trylock(bo->resv)) {
			/*
			 * We raced, and lost, someone else holds the reservation now,
			 * and is probably busy in ttm_bo_cleanup_memtype_use.
			 *
			 * Even if it's not the case, because we finished waiting any
			 * delayed destruction would succeed, so just return success
			 * here.
			 */
544 545 546
			spin_unlock(&glob->lru_lock);
			return 0;
		}
547
		ret = 0;
548
	}
549

550
	if (ret || unlikely(list_empty(&bo->ddestroy))) {
551 552
		if (unlock_resv)
			reservation_object_unlock(bo->resv);
553
		spin_unlock(&glob->lru_lock);
554
		return ret;
555 556
	}

557
	ttm_bo_del_from_lru(bo);
558
	list_del_init(&bo->ddestroy);
559
	kref_put(&bo->list_kref, ttm_bo_ref_bug);
560 561 562

	spin_unlock(&glob->lru_lock);
	ttm_bo_cleanup_memtype_use(bo);
563 564 565

	if (unlock_resv)
		reservation_object_unlock(bo->resv);
566 567

	return 0;
568 569 570 571 572 573
}

/**
 * Traverse the delayed list, and call ttm_bo_cleanup_refs on all
 * encountered buffers.
 */
574
static bool ttm_bo_delayed_delete(struct ttm_bo_device *bdev, bool remove_all)
575
{
576
	struct ttm_bo_global *glob = bdev->glob;
577 578
	struct list_head removed;
	bool empty;
579

580
	INIT_LIST_HEAD(&removed);
581

582 583 584
	spin_lock(&glob->lru_lock);
	while (!list_empty(&bdev->ddestroy)) {
		struct ttm_buffer_object *bo;
585

586 587 588 589 590
		bo = list_first_entry(&bdev->ddestroy, struct ttm_buffer_object,
				      ddestroy);
		kref_get(&bo->list_kref);
		list_move_tail(&bo->ddestroy, &removed);
		spin_unlock(&glob->lru_lock);
591

592
		reservation_object_lock(bo->resv, NULL);
593

594 595
		spin_lock(&glob->lru_lock);
		ttm_bo_cleanup_refs(bo, false, !remove_all, true);
596

597
		kref_put(&bo->list_kref, ttm_bo_release_list);
598
		spin_lock(&glob->lru_lock);
599
	}
600 601
	list_splice_tail(&removed, &bdev->ddestroy);
	empty = list_empty(&bdev->ddestroy);
602
	spin_unlock(&glob->lru_lock);
603 604

	return empty;
605 606 607 608 609 610 611
}

static void ttm_bo_delayed_workqueue(struct work_struct *work)
{
	struct ttm_bo_device *bdev =
	    container_of(work, struct ttm_bo_device, wq.work);

612
	if (!ttm_bo_delayed_delete(bdev, false)) {
613 614 615 616 617 618 619 620 621 622
		schedule_delayed_work(&bdev->wq,
				      ((HZ / 100) < 1) ? 1 : HZ / 100);
	}
}

static void ttm_bo_release(struct kref *kref)
{
	struct ttm_buffer_object *bo =
	    container_of(kref, struct ttm_buffer_object, kref);
	struct ttm_bo_device *bdev = bo->bdev;
623
	struct ttm_mem_type_manager *man = &bdev->man[bo->mem.mem_type];
624

625
	drm_vma_offset_remove(&bdev->vma_manager, &bo->vma_node);
626 627 628
	ttm_mem_io_lock(man, false);
	ttm_mem_io_free_vm(bo);
	ttm_mem_io_unlock(man);
629
	ttm_bo_cleanup_refs_or_queue(bo);
630 631 632 633 634 635 636 637 638 639 640 641
	kref_put(&bo->list_kref, ttm_bo_release_list);
}

void ttm_bo_unref(struct ttm_buffer_object **p_bo)
{
	struct ttm_buffer_object *bo = *p_bo;

	*p_bo = NULL;
	kref_put(&bo->kref, ttm_bo_release);
}
EXPORT_SYMBOL(ttm_bo_unref);

642 643 644 645 646 647 648 649 650 651 652 653 654 655
int ttm_bo_lock_delayed_workqueue(struct ttm_bo_device *bdev)
{
	return cancel_delayed_work_sync(&bdev->wq);
}
EXPORT_SYMBOL(ttm_bo_lock_delayed_workqueue);

void ttm_bo_unlock_delayed_workqueue(struct ttm_bo_device *bdev, int resched)
{
	if (resched)
		schedule_delayed_work(&bdev->wq,
				      ((HZ / 100) < 1) ? 1 : HZ / 100);
}
EXPORT_SYMBOL(ttm_bo_unlock_delayed_workqueue);

656
static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible,
657
			bool no_wait_gpu)
658 659 660
{
	struct ttm_bo_device *bdev = bo->bdev;
	struct ttm_mem_reg evict_mem;
661 662
	struct ttm_placement placement;
	int ret = 0;
663

664
	lockdep_assert_held(&bo->resv->lock.base);
665 666 667

	evict_mem = bo->mem;
	evict_mem.mm_node = NULL;
668 669
	evict_mem.bus.io_reserved_vm = false;
	evict_mem.bus.io_reserved_count = 0;
670

671 672
	placement.num_placement = 0;
	placement.num_busy_placement = 0;
673 674
	bdev->driver->evict_flags(bo, &placement);
	ret = ttm_bo_mem_space(bo, &placement, &evict_mem, interruptible,
675
				no_wait_gpu);
676
	if (ret) {
677
		if (ret != -ERESTARTSYS) {
J
Joe Perches 已提交
678 679
			pr_err("Failed to find memory space for buffer 0x%p eviction\n",
			       bo);
680 681
			ttm_bo_mem_space_debug(bo, &placement);
		}
682 683 684 685
		goto out;
	}

	ret = ttm_bo_handle_move_mem(bo, &evict_mem, true, interruptible,
686
				     no_wait_gpu);
687
	if (unlikely(ret)) {
688
		if (ret != -ERESTARTSYS)
J
Joe Perches 已提交
689
			pr_err("Buffer eviction failed\n");
690
		ttm_bo_mem_put(bo, &evict_mem);
691 692
		goto out;
	}
693 694 695 696 697
	bo->evicted = true;
out:
	return ret;
}

698 699 700 701 702 703 704 705 706 707 708 709 710 711
bool ttm_bo_eviction_valuable(struct ttm_buffer_object *bo,
			      const struct ttm_place *place)
{
	/* Don't evict this BO if it's outside of the
	 * requested placement range
	 */
	if (place->fpfn >= (bo->mem.start + bo->mem.size) ||
	    (place->lpfn && place->lpfn <= bo->mem.start))
		return false;

	return true;
}
EXPORT_SYMBOL(ttm_bo_eviction_valuable);

712
static int ttm_mem_evict_first(struct ttm_bo_device *bdev,
713 714 715 716 717
			       struct reservation_object *resv,
			       uint32_t mem_type,
			       const struct ttm_place *place,
			       bool interruptible,
			       bool no_wait_gpu)
718 719 720
{
	struct ttm_bo_global *glob = bdev->glob;
	struct ttm_mem_type_manager *man = &bdev->man[mem_type];
721 722
	struct ttm_buffer_object *bo = NULL;
	bool locked = false;
723
	unsigned i;
724
	int ret;
725

726
	spin_lock(&glob->lru_lock);
727 728
	for (i = 0; i < TTM_MAX_BO_PRIORITY; ++i) {
		list_for_each_entry(bo, &man->lru[i], lru) {
729 730 731 732 733 734 735 736
			if (bo->resv == resv) {
				if (list_empty(&bo->ddestroy))
					continue;
			} else {
				locked = reservation_object_trylock(bo->resv);
				if (!locked)
					continue;
			}
737

738 739
			if (place && !bdev->driver->eviction_valuable(bo,
								      place)) {
740 741
				if (locked)
					reservation_object_unlock(bo->resv);
742 743 744
				continue;
			}
			break;
745
		}
746

747 748
		/* If the inner loop terminated early, we have our candidate */
		if (&bo->lru != &man->lru[i])
749
			break;
750 751

		bo = NULL;
752 753
	}

754
	if (!bo) {
755
		spin_unlock(&glob->lru_lock);
756
		return -EBUSY;
757 758
	}

759
	kref_get(&bo->list_kref);
760

761
	if (!list_empty(&bo->ddestroy)) {
762 763
		ret = ttm_bo_cleanup_refs(bo, interruptible, no_wait_gpu,
					  locked);
764
		kref_put(&bo->list_kref, ttm_bo_release_list);
765
		return ret;
766 767
	}

768
	ttm_bo_del_from_lru(bo);
769
	spin_unlock(&glob->lru_lock);
770

771
	ret = ttm_bo_evict(bo, interruptible, no_wait_gpu);
772
	if (locked) {
773
		ttm_bo_unreserve(bo);
774 775
	} else {
		spin_lock(&glob->lru_lock);
776
		ttm_bo_add_to_lru(bo);
777 778
		spin_unlock(&glob->lru_lock);
	}
779

780
	kref_put(&bo->list_kref, ttm_bo_release_list);
781 782 783
	return ret;
}

784 785
void ttm_bo_mem_put(struct ttm_buffer_object *bo, struct ttm_mem_reg *mem)
{
786
	struct ttm_mem_type_manager *man = &bo->bdev->man[mem->mem_type];
787

788 789
	if (mem->mm_node)
		(*man->func->put_node)(man, mem);
790 791 792
}
EXPORT_SYMBOL(ttm_bo_mem_put);

793 794 795 796 797 798 799
/**
 * Add the last move fence to the BO and reserve a new shared slot.
 */
static int ttm_bo_add_move_fence(struct ttm_buffer_object *bo,
				 struct ttm_mem_type_manager *man,
				 struct ttm_mem_reg *mem)
{
800
	struct dma_fence *fence;
801 802 803
	int ret;

	spin_lock(&man->move_lock);
804
	fence = dma_fence_get(man->move);
805 806 807 808 809 810 811 812 813
	spin_unlock(&man->move_lock);

	if (fence) {
		reservation_object_add_shared_fence(bo->resv, fence);

		ret = reservation_object_reserve_shared(bo->resv);
		if (unlikely(ret))
			return ret;

814
		dma_fence_put(bo->moving);
815 816 817 818 819 820
		bo->moving = fence;
	}

	return 0;
}

821 822 823 824
/**
 * Repeatedly evict memory from the LRU for @mem_type until we create enough
 * space, or we've evicted everything and there isn't enough space.
 */
825 826
static int ttm_bo_mem_force_space(struct ttm_buffer_object *bo,
					uint32_t mem_type,
827
					const struct ttm_place *place,
828
					struct ttm_mem_reg *mem,
829 830
					bool interruptible,
					bool no_wait_gpu)
831
{
832
	struct ttm_bo_device *bdev = bo->bdev;
833 834 835 836
	struct ttm_mem_type_manager *man = &bdev->man[mem_type];
	int ret;

	do {
837
		ret = (*man->func->get_node)(man, bo, place, mem);
838 839
		if (unlikely(ret != 0))
			return ret;
840
		if (mem->mm_node)
841
			break;
842
		ret = ttm_mem_evict_first(bdev, bo->resv, mem_type, place,
843
					  interruptible, no_wait_gpu);
844 845 846 847
		if (unlikely(ret != 0))
			return ret;
	} while (1);
	mem->mem_type = mem_type;
848
	return ttm_bo_add_move_fence(bo, man, mem);
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
static uint32_t ttm_bo_select_caching(struct ttm_mem_type_manager *man,
				      uint32_t cur_placement,
				      uint32_t proposed_placement)
{
	uint32_t caching = proposed_placement & TTM_PL_MASK_CACHING;
	uint32_t result = proposed_placement & ~TTM_PL_MASK_CACHING;

	/**
	 * Keep current caching if possible.
	 */

	if ((cur_placement & caching) != 0)
		result |= (cur_placement & caching);
	else if ((man->default_caching & caching) != 0)
		result |= man->default_caching;
	else if ((TTM_PL_FLAG_CACHED & caching) != 0)
		result |= TTM_PL_FLAG_CACHED;
	else if ((TTM_PL_FLAG_WC & caching) != 0)
		result |= TTM_PL_FLAG_WC;
	else if ((TTM_PL_FLAG_UNCACHED & caching) != 0)
		result |= TTM_PL_FLAG_UNCACHED;

	return result;
}

876 877
static bool ttm_bo_mt_compatible(struct ttm_mem_type_manager *man,
				 uint32_t mem_type,
878
				 const struct ttm_place *place,
879
				 uint32_t *masked_placement)
880 881 882
{
	uint32_t cur_flags = ttm_bo_type_flags(mem_type);

883
	if ((cur_flags & place->flags & TTM_PL_MASK_MEM) == 0)
884 885
		return false;

886
	if ((place->flags & man->available_caching) == 0)
887 888
		return false;

889
	cur_flags |= (place->flags & man->available_caching);
890 891

	*masked_placement = cur_flags;
892 893 894 895 896 897 898 899 900 901 902 903
	return true;
}

/**
 * Creates space for memory region @mem according to its type.
 *
 * This function first searches for free space in compatible memory types in
 * the priority order defined by the driver.  If free space isn't found, then
 * ttm_bo_mem_force_space is attempted in priority order to evict and find
 * space.
 */
int ttm_bo_mem_space(struct ttm_buffer_object *bo,
904 905
			struct ttm_placement *placement,
			struct ttm_mem_reg *mem,
906
			bool interruptible,
907
			bool no_wait_gpu)
908 909 910 911 912 913 914
{
	struct ttm_bo_device *bdev = bo->bdev;
	struct ttm_mem_type_manager *man;
	uint32_t mem_type = TTM_PL_SYSTEM;
	uint32_t cur_flags = 0;
	bool type_found = false;
	bool type_ok = false;
915
	bool has_erestartsys = false;
916
	int i, ret;
917

918 919 920 921
	ret = reservation_object_reserve_shared(bo->resv);
	if (unlikely(ret))
		return ret;

922
	mem->mm_node = NULL;
923
	for (i = 0; i < placement->num_placement; ++i) {
924 925 926
		const struct ttm_place *place = &placement->placement[i];

		ret = ttm_mem_type_from_place(place, &mem_type);
927 928
		if (ret)
			return ret;
929
		man = &bdev->man[mem_type];
930 931
		if (!man->has_type || !man->use_type)
			continue;
932

933
		type_ok = ttm_bo_mt_compatible(man, mem_type, place,
934
						&cur_flags);
935 936 937 938

		if (!type_ok)
			continue;

939
		type_found = true;
940 941
		cur_flags = ttm_bo_select_caching(man, bo->mem.placement,
						  cur_flags);
942 943 944 945
		/*
		 * Use the access and other non-mapping-related flag bits from
		 * the memory placement flags to the current flags
		 */
946
		ttm_flag_masked(&cur_flags, place->flags,
947
				~TTM_PL_MASK_MEMTYPE);
948

949 950 951
		if (mem_type == TTM_PL_SYSTEM)
			break;

952 953 954
		ret = (*man->func->get_node)(man, bo, place, mem);
		if (unlikely(ret))
			return ret;
955 956 957 958 959 960 961

		if (mem->mm_node) {
			ret = ttm_bo_add_move_fence(bo, man, mem);
			if (unlikely(ret)) {
				(*man->func->put_node)(man, mem);
				return ret;
			}
962
			break;
963
		}
964 965
	}

966
	if ((type_ok && (mem_type == TTM_PL_SYSTEM)) || mem->mm_node) {
967 968 969 970 971
		mem->mem_type = mem_type;
		mem->placement = cur_flags;
		return 0;
	}

972
	for (i = 0; i < placement->num_busy_placement; ++i) {
973 974 975
		const struct ttm_place *place = &placement->busy_placement[i];

		ret = ttm_mem_type_from_place(place, &mem_type);
976 977
		if (ret)
			return ret;
978
		man = &bdev->man[mem_type];
979
		if (!man->has_type || !man->use_type)
980
			continue;
981
		if (!ttm_bo_mt_compatible(man, mem_type, place, &cur_flags))
982 983
			continue;

984
		type_found = true;
985 986
		cur_flags = ttm_bo_select_caching(man, bo->mem.placement,
						  cur_flags);
987 988 989 990
		/*
		 * Use the access and other non-mapping-related flag bits from
		 * the memory placement flags to the current flags
		 */
991
		ttm_flag_masked(&cur_flags, place->flags,
992
				~TTM_PL_MASK_MEMTYPE);
993

994 995 996 997 998 999 1000
		if (mem_type == TTM_PL_SYSTEM) {
			mem->mem_type = mem_type;
			mem->placement = cur_flags;
			mem->mm_node = NULL;
			return 0;
		}

1001
		ret = ttm_bo_mem_force_space(bo, mem_type, place, mem,
1002
						interruptible, no_wait_gpu);
1003 1004 1005 1006
		if (ret == 0 && mem->mm_node) {
			mem->placement = cur_flags;
			return 0;
		}
1007 1008
		if (ret == -ERESTARTSYS)
			has_erestartsys = true;
1009
	}
1010 1011

	if (!type_found) {
1012
		pr_err(TTM_PFX "No compatible memory type found\n");
1013 1014 1015 1016
		return -EINVAL;
	}

	return (has_erestartsys) ? -ERESTARTSYS : -ENOMEM;
1017 1018 1019
}
EXPORT_SYMBOL(ttm_bo_mem_space);

1020
static int ttm_bo_move_buffer(struct ttm_buffer_object *bo,
1021
			struct ttm_placement *placement,
1022
			bool interruptible,
1023
			bool no_wait_gpu)
1024 1025 1026 1027
{
	int ret = 0;
	struct ttm_mem_reg mem;

1028
	lockdep_assert_held(&bo->resv->lock.base);
1029 1030 1031 1032

	mem.num_pages = bo->num_pages;
	mem.size = mem.num_pages << PAGE_SHIFT;
	mem.page_alignment = bo->mem.page_alignment;
1033 1034
	mem.bus.io_reserved_vm = false;
	mem.bus.io_reserved_count = 0;
1035 1036 1037
	/*
	 * Determine where to move the buffer.
	 */
1038 1039
	ret = ttm_bo_mem_space(bo, placement, &mem,
			       interruptible, no_wait_gpu);
1040 1041
	if (ret)
		goto out_unlock;
1042 1043
	ret = ttm_bo_handle_move_mem(bo, &mem, false,
				     interruptible, no_wait_gpu);
1044
out_unlock:
1045 1046
	if (ret && mem.mm_node)
		ttm_bo_mem_put(bo, &mem);
1047 1048 1049
	return ret;
}

1050 1051 1052 1053
static bool ttm_bo_places_compat(const struct ttm_place *places,
				 unsigned num_placement,
				 struct ttm_mem_reg *mem,
				 uint32_t *new_flags)
1054
{
1055
	unsigned i;
1056

1057 1058
	for (i = 0; i < num_placement; i++) {
		const struct ttm_place *heap = &places[i];
1059

1060
		if (mem->mm_node && (mem->start < heap->fpfn ||
1061
		     (heap->lpfn != 0 && (mem->start + mem->num_pages) > heap->lpfn)))
1062 1063 1064
			continue;

		*new_flags = heap->flags;
1065
		if ((*new_flags & mem->placement & TTM_PL_MASK_CACHING) &&
1066 1067 1068
		    (*new_flags & mem->placement & TTM_PL_MASK_MEM) &&
		    (!(*new_flags & TTM_PL_FLAG_CONTIGUOUS) ||
		     (mem->placement & TTM_PL_FLAG_CONTIGUOUS)))
1069
			return true;
1070
	}
1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087
	return false;
}

bool ttm_bo_mem_compat(struct ttm_placement *placement,
		       struct ttm_mem_reg *mem,
		       uint32_t *new_flags)
{
	if (ttm_bo_places_compat(placement->placement, placement->num_placement,
				 mem, new_flags))
		return true;

	if ((placement->busy_placement != placement->placement ||
	     placement->num_busy_placement > placement->num_placement) &&
	    ttm_bo_places_compat(placement->busy_placement,
				 placement->num_busy_placement,
				 mem, new_flags))
		return true;
1088 1089

	return false;
1090
}
1091
EXPORT_SYMBOL(ttm_bo_mem_compat);
1092

1093
int ttm_bo_validate(struct ttm_buffer_object *bo,
1094 1095
		    struct ttm_placement *placement,
		    struct ttm_operation_ctx *ctx)
1096 1097
{
	int ret;
1098
	uint32_t new_flags;
1099

1100
	lockdep_assert_held(&bo->resv->lock.base);
1101 1102 1103
	/*
	 * Check whether we need to move buffer.
	 */
1104
	if (!ttm_bo_mem_compat(placement, &bo->mem, &new_flags)) {
1105 1106
		ret = ttm_bo_move_buffer(bo, placement, ctx->interruptible,
					 ctx->no_wait_gpu);
1107
		if (ret)
1108
			return ret;
1109 1110 1111 1112 1113
	} else {
		/*
		 * Use the access and other non-mapping-related flag bits from
		 * the compatible memory placement flags to the active flags
		 */
1114
		ttm_flag_masked(&bo->mem.placement, new_flags,
1115
				~TTM_PL_MASK_MEMTYPE);
1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126
	}
	/*
	 * We might need to add a TTM.
	 */
	if (bo->mem.mem_type == TTM_PL_SYSTEM && bo->ttm == NULL) {
		ret = ttm_bo_add_ttm(bo, true);
		if (ret)
			return ret;
	}
	return 0;
}
1127
EXPORT_SYMBOL(ttm_bo_validate);
1128

1129 1130 1131 1132 1133 1134
int ttm_bo_init_reserved(struct ttm_bo_device *bdev,
			 struct ttm_buffer_object *bo,
			 unsigned long size,
			 enum ttm_bo_type type,
			 struct ttm_placement *placement,
			 uint32_t page_alignment,
1135
			 struct ttm_operation_ctx *ctx,
1136 1137 1138 1139 1140
			 struct file *persistent_swap_storage,
			 size_t acc_size,
			 struct sg_table *sg,
			 struct reservation_object *resv,
			 void (*destroy) (struct ttm_buffer_object *))
1141
{
1142
	int ret = 0;
1143
	unsigned long num_pages;
1144
	struct ttm_mem_global *mem_glob = bdev->glob->mem_glob;
1145
	bool locked;
1146 1147 1148

	ret = ttm_mem_global_alloc(mem_glob, acc_size, false, false);
	if (ret) {
J
Joe Perches 已提交
1149
		pr_err("Out of kernel memory\n");
1150 1151 1152 1153 1154 1155
		if (destroy)
			(*destroy)(bo);
		else
			kfree(bo);
		return -ENOMEM;
	}
1156 1157 1158

	num_pages = (size + PAGE_SIZE - 1) >> PAGE_SHIFT;
	if (num_pages == 0) {
J
Joe Perches 已提交
1159
		pr_err("Illegal buffer object size\n");
1160 1161 1162 1163
		if (destroy)
			(*destroy)(bo);
		else
			kfree(bo);
1164
		ttm_mem_global_free(mem_glob, acc_size);
1165 1166 1167 1168 1169 1170 1171 1172 1173 1174
		return -EINVAL;
	}
	bo->destroy = destroy;

	kref_init(&bo->kref);
	kref_init(&bo->list_kref);
	atomic_set(&bo->cpu_writers, 0);
	INIT_LIST_HEAD(&bo->lru);
	INIT_LIST_HEAD(&bo->ddestroy);
	INIT_LIST_HEAD(&bo->swap);
1175
	INIT_LIST_HEAD(&bo->io_reserve_lru);
1176
	mutex_init(&bo->wu_mutex);
1177
	bo->bdev = bdev;
1178
	bo->glob = bdev->glob;
1179 1180
	bo->type = type;
	bo->num_pages = num_pages;
1181
	bo->mem.size = num_pages << PAGE_SHIFT;
1182 1183 1184 1185
	bo->mem.mem_type = TTM_PL_SYSTEM;
	bo->mem.num_pages = bo->num_pages;
	bo->mem.mm_node = NULL;
	bo->mem.page_alignment = page_alignment;
1186 1187
	bo->mem.bus.io_reserved_vm = false;
	bo->mem.bus.io_reserved_count = 0;
1188
	bo->moving = NULL;
1189
	bo->mem.placement = (TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED);
J
Jan Engelhardt 已提交
1190
	bo->persistent_swap_storage = persistent_swap_storage;
1191
	bo->acc_size = acc_size;
1192
	bo->sg = sg;
1193 1194 1195 1196 1197 1198
	if (resv) {
		bo->resv = resv;
		lockdep_assert_held(&bo->resv->lock.base);
	} else {
		bo->resv = &bo->ttm_resv;
	}
1199
	reservation_object_init(&bo->ttm_resv);
1200
	atomic_inc(&bo->glob->bo_count);
1201
	drm_vma_node_reset(&bo->vma_node);
1202
	bo->priority = 0;
1203 1204 1205 1206 1207

	/*
	 * For ttm_bo_type_device buffers, allocate
	 * address space from the device.
	 */
1208 1209
	if (bo->type == ttm_bo_type_device ||
	    bo->type == ttm_bo_type_sg)
1210 1211
		ret = drm_vma_offset_add(&bdev->vma_manager, &bo->vma_node,
					 bo->mem.num_pages);
1212

1213 1214 1215 1216 1217 1218 1219
	/* passed reservation objects should already be locked,
	 * since otherwise lockdep will be angered in radeon.
	 */
	if (!resv) {
		locked = ww_mutex_trylock(&bo->resv->lock);
		WARN_ON(!locked);
	}
1220

1221 1222
	if (likely(!ret))
		ret = ttm_bo_validate(bo, placement, ctx);
1223

1224
	if (unlikely(ret)) {
1225 1226 1227
		if (!resv)
			ttm_bo_unreserve(bo);

1228 1229 1230 1231 1232
		ttm_bo_unref(&bo);
		return ret;
	}

	if (resv && !(bo->mem.placement & TTM_PL_FLAG_NO_EVICT)) {
1233 1234 1235 1236 1237
		spin_lock(&bo->glob->lru_lock);
		ttm_bo_add_to_lru(bo);
		spin_unlock(&bo->glob->lru_lock);
	}

1238 1239
	return ret;
}
1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254
EXPORT_SYMBOL(ttm_bo_init_reserved);

int ttm_bo_init(struct ttm_bo_device *bdev,
		struct ttm_buffer_object *bo,
		unsigned long size,
		enum ttm_bo_type type,
		struct ttm_placement *placement,
		uint32_t page_alignment,
		bool interruptible,
		struct file *persistent_swap_storage,
		size_t acc_size,
		struct sg_table *sg,
		struct reservation_object *resv,
		void (*destroy) (struct ttm_buffer_object *))
{
1255
	struct ttm_operation_ctx ctx = { interruptible, false };
1256 1257 1258
	int ret;

	ret = ttm_bo_init_reserved(bdev, bo, size, type, placement,
1259
				   page_alignment, &ctx,
1260 1261 1262 1263 1264 1265 1266 1267 1268 1269
				   persistent_swap_storage, acc_size,
				   sg, resv, destroy);
	if (ret)
		return ret;

	if (!resv)
		ttm_bo_unreserve(bo);

	return 0;
}
1270
EXPORT_SYMBOL(ttm_bo_init);
1271

1272 1273 1274
size_t ttm_bo_acc_size(struct ttm_bo_device *bdev,
		       unsigned long bo_size,
		       unsigned struct_size)
1275
{
1276 1277
	unsigned npages = (PAGE_ALIGN(bo_size)) >> PAGE_SHIFT;
	size_t size = 0;
1278

1279
	size += ttm_round_pot(struct_size);
F
Felix Kuehling 已提交
1280
	size += ttm_round_pot(npages * sizeof(void *));
1281 1282
	size += ttm_round_pot(sizeof(struct ttm_tt));
	return size;
1283
}
1284 1285 1286 1287 1288 1289 1290 1291 1292 1293
EXPORT_SYMBOL(ttm_bo_acc_size);

size_t ttm_bo_dma_acc_size(struct ttm_bo_device *bdev,
			   unsigned long bo_size,
			   unsigned struct_size)
{
	unsigned npages = (PAGE_ALIGN(bo_size)) >> PAGE_SHIFT;
	size_t size = 0;

	size += ttm_round_pot(struct_size);
F
Felix Kuehling 已提交
1294
	size += ttm_round_pot(npages * (2*sizeof(void *) + sizeof(dma_addr_t)));
1295 1296 1297 1298
	size += ttm_round_pot(sizeof(struct ttm_dma_tt));
	return size;
}
EXPORT_SYMBOL(ttm_bo_dma_acc_size);
1299

1300 1301 1302 1303 1304 1305
int ttm_bo_create(struct ttm_bo_device *bdev,
			unsigned long size,
			enum ttm_bo_type type,
			struct ttm_placement *placement,
			uint32_t page_alignment,
			bool interruptible,
J
Jan Engelhardt 已提交
1306
			struct file *persistent_swap_storage,
1307
			struct ttm_buffer_object **p_bo)
1308 1309
{
	struct ttm_buffer_object *bo;
1310
	size_t acc_size;
1311
	int ret;
1312 1313

	bo = kzalloc(sizeof(*bo), GFP_KERNEL);
1314
	if (unlikely(bo == NULL))
1315 1316
		return -ENOMEM;

1317
	acc_size = ttm_bo_acc_size(bdev, size, sizeof(struct ttm_buffer_object));
1318
	ret = ttm_bo_init(bdev, bo, size, type, placement, page_alignment,
1319
			  interruptible, persistent_swap_storage, acc_size,
1320
			  NULL, NULL, NULL);
1321 1322 1323 1324 1325
	if (likely(ret == 0))
		*p_bo = bo;

	return ret;
}
T
Thomas Hellstrom 已提交
1326
EXPORT_SYMBOL(ttm_bo_create);
1327 1328

static int ttm_bo_force_list_clean(struct ttm_bo_device *bdev,
1329
				   unsigned mem_type)
1330
{
1331
	struct ttm_mem_type_manager *man = &bdev->man[mem_type];
1332
	struct ttm_bo_global *glob = bdev->glob;
1333
	struct dma_fence *fence;
1334
	int ret;
1335
	unsigned i;
1336 1337 1338 1339 1340

	/*
	 * Can't use standard list traversal since we're unlocking.
	 */

1341
	spin_lock(&glob->lru_lock);
1342 1343 1344
	for (i = 0; i < TTM_MAX_BO_PRIORITY; ++i) {
		while (!list_empty(&man->lru[i])) {
			spin_unlock(&glob->lru_lock);
1345 1346
			ret = ttm_mem_evict_first(bdev, NULL, mem_type, NULL,
						  false, false);
1347
			if (ret)
1348
				return ret;
1349
			spin_lock(&glob->lru_lock);
1350
		}
1351
	}
1352
	spin_unlock(&glob->lru_lock);
1353 1354

	spin_lock(&man->move_lock);
1355
	fence = dma_fence_get(man->move);
1356 1357 1358
	spin_unlock(&man->move_lock);

	if (fence) {
1359 1360
		ret = dma_fence_wait(fence, false);
		dma_fence_put(fence);
1361 1362
		if (ret)
			return ret;
1363 1364
	}

1365 1366 1367 1368 1369
	return 0;
}

int ttm_bo_clean_mm(struct ttm_bo_device *bdev, unsigned mem_type)
{
R
Roel Kluin 已提交
1370
	struct ttm_mem_type_manager *man;
1371 1372 1373
	int ret = -EINVAL;

	if (mem_type >= TTM_NUM_MEM_TYPES) {
J
Joe Perches 已提交
1374
		pr_err("Illegal memory type %d\n", mem_type);
1375 1376
		return ret;
	}
R
Roel Kluin 已提交
1377
	man = &bdev->man[mem_type];
1378 1379

	if (!man->has_type) {
J
Joe Perches 已提交
1380 1381
		pr_err("Trying to take down uninitialized memory manager type %u\n",
		       mem_type);
1382 1383 1384 1385 1386 1387 1388 1389
		return ret;
	}

	man->use_type = false;
	man->has_type = false;

	ret = 0;
	if (mem_type > 0) {
1390 1391 1392 1393 1394
		ret = ttm_bo_force_list_clean(bdev, mem_type);
		if (ret) {
			pr_err("Cleanup eviction failed\n");
			return ret;
		}
1395

1396
		ret = (*man->func->takedown)(man);
1397 1398
	}

1399 1400 1401
	dma_fence_put(man->move);
	man->move = NULL;

1402 1403 1404 1405 1406 1407 1408 1409 1410
	return ret;
}
EXPORT_SYMBOL(ttm_bo_clean_mm);

int ttm_bo_evict_mm(struct ttm_bo_device *bdev, unsigned mem_type)
{
	struct ttm_mem_type_manager *man = &bdev->man[mem_type];

	if (mem_type == 0 || mem_type >= TTM_NUM_MEM_TYPES) {
J
Joe Perches 已提交
1411
		pr_err("Illegal memory manager memory type %u\n", mem_type);
1412 1413 1414 1415
		return -EINVAL;
	}

	if (!man->has_type) {
J
Joe Perches 已提交
1416
		pr_err("Memory type %u has not been initialized\n", mem_type);
1417 1418 1419
		return 0;
	}

1420
	return ttm_bo_force_list_clean(bdev, mem_type);
1421 1422 1423 1424
}
EXPORT_SYMBOL(ttm_bo_evict_mm);

int ttm_bo_init_mm(struct ttm_bo_device *bdev, unsigned type,
1425
			unsigned long p_size)
1426
{
H
Huang Rui 已提交
1427
	int ret;
1428
	struct ttm_mem_type_manager *man;
1429
	unsigned i;
1430

1431
	BUG_ON(type >= TTM_NUM_MEM_TYPES);
1432
	man = &bdev->man[type];
1433
	BUG_ON(man->has_type);
1434 1435 1436
	man->io_reserve_fastpath = true;
	man->use_io_reserve_lru = false;
	mutex_init(&man->io_reserve_mutex);
1437
	spin_lock_init(&man->move_lock);
1438
	INIT_LIST_HEAD(&man->io_reserve_lru);
1439 1440 1441 1442

	ret = bdev->driver->init_mem_type(bdev, type, man);
	if (ret)
		return ret;
1443
	man->bdev = bdev;
1444 1445

	if (type != TTM_PL_SYSTEM) {
1446
		ret = (*man->func->init)(man, p_size);
1447 1448 1449 1450 1451 1452 1453
		if (ret)
			return ret;
	}
	man->has_type = true;
	man->use_type = true;
	man->size = p_size;

1454 1455
	for (i = 0; i < TTM_MAX_BO_PRIORITY; ++i)
		INIT_LIST_HEAD(&man->lru[i]);
1456
	man->move = NULL;
1457 1458 1459 1460 1461

	return 0;
}
EXPORT_SYMBOL(ttm_bo_init_mm);

1462 1463 1464 1465 1466 1467 1468 1469 1470 1471
static void ttm_bo_global_kobj_release(struct kobject *kobj)
{
	struct ttm_bo_global *glob =
		container_of(kobj, struct ttm_bo_global, kobj);

	ttm_mem_unregister_shrink(glob->mem_glob, &glob->shrink);
	__free_page(glob->dummy_read_page);
	kfree(glob);
}

1472
void ttm_bo_global_release(struct drm_global_reference *ref)
1473 1474 1475 1476 1477 1478 1479 1480
{
	struct ttm_bo_global *glob = ref->object;

	kobject_del(&glob->kobj);
	kobject_put(&glob->kobj);
}
EXPORT_SYMBOL(ttm_bo_global_release);

1481
int ttm_bo_global_init(struct drm_global_reference *ref)
1482 1483 1484 1485 1486
{
	struct ttm_bo_global_ref *bo_ref =
		container_of(ref, struct ttm_bo_global_ref, ref);
	struct ttm_bo_global *glob = ref->object;
	int ret;
1487
	unsigned i;
1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498

	mutex_init(&glob->device_list_mutex);
	spin_lock_init(&glob->lru_lock);
	glob->mem_glob = bo_ref->mem_glob;
	glob->dummy_read_page = alloc_page(__GFP_ZERO | GFP_DMA32);

	if (unlikely(glob->dummy_read_page == NULL)) {
		ret = -ENOMEM;
		goto out_no_drp;
	}

1499 1500
	for (i = 0; i < TTM_MAX_BO_PRIORITY; ++i)
		INIT_LIST_HEAD(&glob->swap_lru[i]);
1501 1502 1503 1504 1505
	INIT_LIST_HEAD(&glob->device_list);

	ttm_mem_init_shrink(&glob->shrink, ttm_bo_swapout);
	ret = ttm_mem_register_shrink(glob->mem_glob, &glob->shrink);
	if (unlikely(ret != 0)) {
J
Joe Perches 已提交
1506
		pr_err("Could not register buffer object swapout\n");
1507 1508 1509 1510 1511
		goto out_no_shrink;
	}

	atomic_set(&glob->bo_count, 0);

1512 1513
	ret = kobject_init_and_add(
		&glob->kobj, &ttm_bo_glob_kobj_type, ttm_get_kobj(), "buffer_objects");
1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525
	if (unlikely(ret != 0))
		kobject_put(&glob->kobj);
	return ret;
out_no_shrink:
	__free_page(glob->dummy_read_page);
out_no_drp:
	kfree(glob);
	return ret;
}
EXPORT_SYMBOL(ttm_bo_global_init);


1526 1527 1528 1529 1530
int ttm_bo_device_release(struct ttm_bo_device *bdev)
{
	int ret = 0;
	unsigned i = TTM_NUM_MEM_TYPES;
	struct ttm_mem_type_manager *man;
1531
	struct ttm_bo_global *glob = bdev->glob;
1532 1533 1534 1535 1536 1537 1538

	while (i--) {
		man = &bdev->man[i];
		if (man->has_type) {
			man->use_type = false;
			if ((i != TTM_PL_SYSTEM) && ttm_bo_clean_mm(bdev, i)) {
				ret = -EBUSY;
J
Joe Perches 已提交
1539 1540
				pr_err("DRM memory manager type %d is not clean\n",
				       i);
1541 1542 1543 1544 1545
			}
			man->has_type = false;
		}
	}

1546 1547 1548 1549
	mutex_lock(&glob->device_list_mutex);
	list_del(&bdev->device_list);
	mutex_unlock(&glob->device_list_mutex);

1550
	cancel_delayed_work_sync(&bdev->wq);
1551

1552
	if (ttm_bo_delayed_delete(bdev, true))
1553 1554
		TTM_DEBUG("Delayed destroy list was clean\n");

1555
	spin_lock(&glob->lru_lock);
1556 1557 1558
	for (i = 0; i < TTM_MAX_BO_PRIORITY; ++i)
		if (list_empty(&bdev->man[0].lru[0]))
			TTM_DEBUG("Swap list %d was clean\n", i);
1559
	spin_unlock(&glob->lru_lock);
1560

1561
	drm_vma_offset_manager_destroy(&bdev->vma_manager);
1562 1563 1564 1565 1566 1567

	return ret;
}
EXPORT_SYMBOL(ttm_bo_device_release);

int ttm_bo_device_init(struct ttm_bo_device *bdev,
1568 1569
		       struct ttm_bo_global *glob,
		       struct ttm_bo_driver *driver,
1570
		       struct address_space *mapping,
D
Dave Airlie 已提交
1571
		       uint64_t file_page_offset,
D
Dave Airlie 已提交
1572
		       bool need_dma32)
1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583
{
	int ret = -EINVAL;

	bdev->driver = driver;

	memset(bdev->man, 0, sizeof(bdev->man));

	/*
	 * Initialize the system memory buffer type.
	 * Other types need to be driver / IOCTL initialized.
	 */
1584
	ret = ttm_bo_init_mm(bdev, TTM_PL_SYSTEM, 0);
1585
	if (unlikely(ret != 0))
1586
		goto out_no_sys;
1587

1588 1589
	drm_vma_offset_manager_init(&bdev->vma_manager, file_page_offset,
				    0x10000000);
1590 1591
	INIT_DELAYED_WORK(&bdev->wq, ttm_bo_delayed_workqueue);
	INIT_LIST_HEAD(&bdev->ddestroy);
1592
	bdev->dev_mapping = mapping;
1593
	bdev->glob = glob;
D
Dave Airlie 已提交
1594
	bdev->need_dma32 = need_dma32;
1595 1596 1597
	mutex_lock(&glob->device_list_mutex);
	list_add_tail(&bdev->device_list, &glob->device_list);
	mutex_unlock(&glob->device_list_mutex);
1598 1599

	return 0;
1600
out_no_sys:
1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625
	return ret;
}
EXPORT_SYMBOL(ttm_bo_device_init);

/*
 * buffer object vm functions.
 */

bool ttm_mem_reg_is_pci(struct ttm_bo_device *bdev, struct ttm_mem_reg *mem)
{
	struct ttm_mem_type_manager *man = &bdev->man[mem->mem_type];

	if (!(man->flags & TTM_MEMTYPE_FLAG_FIXED)) {
		if (mem->mem_type == TTM_PL_SYSTEM)
			return false;

		if (man->flags & TTM_MEMTYPE_FLAG_CMA)
			return false;

		if (mem->placement & TTM_PL_FLAG_CACHED)
			return false;
	}
	return true;
}

1626
void ttm_bo_unmap_virtual_locked(struct ttm_buffer_object *bo)
1627 1628 1629
{
	struct ttm_bo_device *bdev = bo->bdev;

1630
	drm_vma_node_unmap(&bo->vma_node, bdev->dev_mapping);
1631
	ttm_mem_io_free_vm(bo);
1632
}
1633 1634 1635 1636 1637 1638 1639 1640 1641

void ttm_bo_unmap_virtual(struct ttm_buffer_object *bo)
{
	struct ttm_bo_device *bdev = bo->bdev;
	struct ttm_mem_type_manager *man = &bdev->man[bo->mem.mem_type];

	ttm_mem_io_lock(man, false);
	ttm_bo_unmap_virtual_locked(bo);
	ttm_mem_io_unlock(man);
1642
}
1643 1644


1645
EXPORT_SYMBOL(ttm_bo_unmap_virtual);
1646 1647

int ttm_bo_wait(struct ttm_buffer_object *bo,
1648
		bool interruptible, bool no_wait)
1649
{
C
Christian König 已提交
1650 1651 1652 1653 1654 1655 1656 1657
	long timeout = 15 * HZ;

	if (no_wait) {
		if (reservation_object_test_signaled_rcu(bo->resv, true))
			return 0;
		else
			return -EBUSY;
	}
1658

C
Christian König 已提交
1659 1660
	timeout = reservation_object_wait_timeout_rcu(bo->resv, true,
						      interruptible, timeout);
1661 1662 1663 1664 1665 1666
	if (timeout < 0)
		return timeout;

	if (timeout == 0)
		return -EBUSY;

C
Christian König 已提交
1667
	reservation_object_add_excl_fence(bo->resv, NULL);
1668
	return 0;
1669 1670 1671 1672 1673 1674 1675 1676
}
EXPORT_SYMBOL(ttm_bo_wait);

int ttm_bo_synccpu_write_grab(struct ttm_buffer_object *bo, bool no_wait)
{
	int ret = 0;

	/*
1677
	 * Using ttm_bo_reserve makes sure the lru lists are updated.
1678 1679
	 */

1680
	ret = ttm_bo_reserve(bo, true, no_wait, NULL);
1681 1682
	if (unlikely(ret != 0))
		return ret;
1683
	ret = ttm_bo_wait(bo, true, no_wait);
1684 1685 1686 1687 1688
	if (likely(ret == 0))
		atomic_inc(&bo->cpu_writers);
	ttm_bo_unreserve(bo);
	return ret;
}
1689
EXPORT_SYMBOL(ttm_bo_synccpu_write_grab);
1690 1691 1692

void ttm_bo_synccpu_write_release(struct ttm_buffer_object *bo)
{
1693
	atomic_dec(&bo->cpu_writers);
1694
}
1695
EXPORT_SYMBOL(ttm_bo_synccpu_write_release);
1696 1697 1698 1699 1700 1701 1702 1703

/**
 * A buffer object shrink method that tries to swap out the first
 * buffer object on the bo_global::swap_lru list.
 */

static int ttm_bo_swapout(struct ttm_mem_shrink *shrink)
{
1704 1705
	struct ttm_bo_global *glob =
	    container_of(shrink, struct ttm_bo_global, shrink);
1706 1707
	struct ttm_buffer_object *bo;
	int ret = -EBUSY;
1708
	unsigned i;
1709

1710
	spin_lock(&glob->lru_lock);
1711 1712
	for (i = 0; i < TTM_MAX_BO_PRIORITY; ++i) {
		list_for_each_entry(bo, &glob->swap_lru[i], swap) {
1713
			ret = reservation_object_trylock(bo->resv) ? 0 : -EBUSY;
1714 1715 1716
			if (!ret)
				break;
		}
1717 1718 1719
		if (!ret)
			break;
	}
1720

1721 1722 1723 1724
	if (ret) {
		spin_unlock(&glob->lru_lock);
		return ret;
	}
1725

1726
	kref_get(&bo->list_kref);
1727

1728
	if (!list_empty(&bo->ddestroy)) {
1729
		ret = ttm_bo_cleanup_refs(bo, false, false, true);
1730 1731
		kref_put(&bo->list_kref, ttm_bo_release_list);
		return ret;
1732 1733
	}

1734
	ttm_bo_del_from_lru(bo);
1735
	spin_unlock(&glob->lru_lock);
1736 1737

	/**
1738
	 * Move to system cached
1739 1740
	 */

1741 1742
	if (bo->mem.mem_type != TTM_PL_SYSTEM ||
	    bo->ttm->caching_state != tt_cached) {
1743 1744 1745 1746 1747 1748 1749 1750
		struct ttm_mem_reg evict_mem;

		evict_mem = bo->mem;
		evict_mem.mm_node = NULL;
		evict_mem.placement = TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED;
		evict_mem.mem_type = TTM_PL_SYSTEM;

		ret = ttm_bo_handle_move_mem(bo, &evict_mem, true,
1751
					     false, false);
1752 1753 1754 1755
		if (unlikely(ret != 0))
			goto out;
	}

1756 1757 1758 1759 1760 1761 1762 1763
	/**
	 * Make sure BO is idle.
	 */

	ret = ttm_bo_wait(bo, false, false);
	if (unlikely(ret != 0))
		goto out;

1764 1765 1766 1767 1768 1769 1770
	ttm_bo_unmap_virtual(bo);

	/**
	 * Swap out. Buffer will be swapped in again as soon as
	 * anyone tries to access a ttm page.
	 */

1771 1772 1773
	if (bo->bdev->driver->swap_notify)
		bo->bdev->driver->swap_notify(bo);

J
Jan Engelhardt 已提交
1774
	ret = ttm_tt_swapout(bo->ttm, bo->persistent_swap_storage);
1775 1776 1777 1778 1779 1780 1781 1782
out:

	/**
	 *
	 * Unreserve without putting on LRU to avoid swapping out an
	 * already swapped buffer.
	 */

1783
	reservation_object_unlock(bo->resv);
1784 1785 1786 1787 1788 1789
	kref_put(&bo->list_kref, ttm_bo_release_list);
	return ret;
}

void ttm_bo_swapout_all(struct ttm_bo_device *bdev)
{
1790
	while (ttm_bo_swapout(&bdev->glob->shrink) == 0)
1791 1792
		;
}
1793
EXPORT_SYMBOL(ttm_bo_swapout_all);
1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816

/**
 * ttm_bo_wait_unreserved - interruptible wait for a buffer object to become
 * unreserved
 *
 * @bo: Pointer to buffer
 */
int ttm_bo_wait_unreserved(struct ttm_buffer_object *bo)
{
	int ret;

	/*
	 * In the absense of a wait_unlocked API,
	 * Use the bo::wu_mutex to avoid triggering livelocks due to
	 * concurrent use of this function. Note that this use of
	 * bo::wu_mutex can go away if we change locking order to
	 * mmap_sem -> bo::reserve.
	 */
	ret = mutex_lock_interruptible(&bo->wu_mutex);
	if (unlikely(ret != 0))
		return -ERESTARTSYS;
	if (!ww_mutex_is_locked(&bo->resv->lock))
		goto out_unlock;
1817 1818 1819
	ret = reservation_object_lock_interruptible(bo->resv, NULL);
	if (ret == -EINTR)
		ret = -ERESTARTSYS;
1820 1821
	if (unlikely(ret != 0))
		goto out_unlock;
1822
	reservation_object_unlock(bo->resv);
1823 1824 1825 1826 1827

out_unlock:
	mutex_unlock(&bo->wu_mutex);
	return ret;
}