extent_map.c 66.9 KB
Newer Older
1 2 3 4 5 6 7 8 9 10
#include <linux/bitops.h>
#include <linux/slab.h>
#include <linux/bio.h>
#include <linux/mm.h>
#include <linux/gfp.h>
#include <linux/pagemap.h>
#include <linux/page-flags.h>
#include <linux/module.h>
#include <linux/spinlock.h>
#include <linux/blkdev.h>
11
#include <linux/swap.h>
12 13
#include "extent_map.h"

14 15 16 17 18 19
/* temporary define until extent_map moves out of btrfs */
struct kmem_cache *btrfs_cache_create(const char *name, size_t size,
				       unsigned long extra_flags,
				       void (*ctor)(void *, struct kmem_cache *,
						    unsigned long));

20 21
static struct kmem_cache *extent_map_cache;
static struct kmem_cache *extent_state_cache;
22
static struct kmem_cache *extent_buffer_cache;
23 24 25 26 27

static LIST_HEAD(buffers);
static LIST_HEAD(states);

static spinlock_t state_lock = SPIN_LOCK_UNLOCKED;
28
#define BUFFER_LRU_MAX 64
29 30 31 32 33 34 35 36 37 38

struct tree_entry {
	u64 start;
	u64 end;
	int in_tree;
	struct rb_node rb_node;
};

void __init extent_map_init(void)
{
39
	extent_map_cache = btrfs_cache_create("extent_map",
40
					    sizeof(struct extent_map), 0,
41
					    NULL);
42
	extent_state_cache = btrfs_cache_create("extent_state",
43
					    sizeof(struct extent_state), 0,
44
					    NULL);
45 46 47
	extent_buffer_cache = btrfs_cache_create("extent_buffers",
					    sizeof(struct extent_buffer), 0,
					    NULL);
48 49 50 51
}

void __exit extent_map_exit(void)
{
52
	struct extent_state *state;
53

54 55 56 57 58 59 60 61
	while (!list_empty(&states)) {
		state = list_entry(states.next, struct extent_state, list);
		printk("state leak: start %Lu end %Lu state %lu in tree %d refs %d\n", state->start, state->end, state->state, state->in_tree, atomic_read(&state->refs));
		list_del(&state->list);
		kmem_cache_free(extent_state_cache, state);

	}

62 63 64 65
	if (extent_map_cache)
		kmem_cache_destroy(extent_map_cache);
	if (extent_state_cache)
		kmem_cache_destroy(extent_state_cache);
66 67
	if (extent_buffer_cache)
		kmem_cache_destroy(extent_buffer_cache);
68 69 70 71 72 73 74
}

void extent_map_tree_init(struct extent_map_tree *tree,
			  struct address_space *mapping, gfp_t mask)
{
	tree->map.rb_node = NULL;
	tree->state.rb_node = NULL;
75
	tree->ops = NULL;
76
	rwlock_init(&tree->lock);
77
	spin_lock_init(&tree->lru_lock);
78
	tree->mapping = mapping;
79 80
	INIT_LIST_HEAD(&tree->buffer_lru);
	tree->lru_size = 0;
81 82 83
}
EXPORT_SYMBOL(extent_map_tree_init);

84
void extent_map_tree_empty_lru(struct extent_map_tree *tree)
85 86 87 88 89 90 91 92 93
{
	struct extent_buffer *eb;
	while(!list_empty(&tree->buffer_lru)) {
		eb = list_entry(tree->buffer_lru.next, struct extent_buffer,
				lru);
		list_del(&eb->lru);
		free_extent_buffer(eb);
	}
}
94
EXPORT_SYMBOL(extent_map_tree_empty_lru);
95

96 97 98 99 100 101 102 103 104 105 106 107 108 109
struct extent_map *alloc_extent_map(gfp_t mask)
{
	struct extent_map *em;
	em = kmem_cache_alloc(extent_map_cache, mask);
	if (!em || IS_ERR(em))
		return em;
	em->in_tree = 0;
	atomic_set(&em->refs, 1);
	return em;
}
EXPORT_SYMBOL(alloc_extent_map);

void free_extent_map(struct extent_map *em)
{
C
Chris Mason 已提交
110 111
	if (!em)
		return;
112 113 114 115 116 117 118 119 120 121 122
	if (atomic_dec_and_test(&em->refs)) {
		WARN_ON(em->in_tree);
		kmem_cache_free(extent_map_cache, em);
	}
}
EXPORT_SYMBOL(free_extent_map);


struct extent_state *alloc_extent_state(gfp_t mask)
{
	struct extent_state *state;
123 124
	unsigned long flags;

125 126 127 128 129
	state = kmem_cache_alloc(extent_state_cache, mask);
	if (!state || IS_ERR(state))
		return state;
	state->state = 0;
	state->in_tree = 0;
130
	state->private = 0;
131 132 133 134 135

	spin_lock_irqsave(&state_lock, flags);
	list_add(&state->list, &states);
	spin_unlock_irqrestore(&state_lock, flags);

136 137 138 139 140 141 142 143
	atomic_set(&state->refs, 1);
	init_waitqueue_head(&state->wq);
	return state;
}
EXPORT_SYMBOL(alloc_extent_state);

void free_extent_state(struct extent_state *state)
{
144
	unsigned long flags;
C
Chris Mason 已提交
145 146
	if (!state)
		return;
147 148
	if (atomic_dec_and_test(&state->refs)) {
		WARN_ON(state->in_tree);
149 150 151
		spin_lock_irqsave(&state_lock, flags);
		list_del(&state->list);
		spin_unlock_irqrestore(&state_lock, flags);
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
		kmem_cache_free(extent_state_cache, state);
	}
}
EXPORT_SYMBOL(free_extent_state);

static struct rb_node *tree_insert(struct rb_root *root, u64 offset,
				   struct rb_node *node)
{
	struct rb_node ** p = &root->rb_node;
	struct rb_node * parent = NULL;
	struct tree_entry *entry;

	while(*p) {
		parent = *p;
		entry = rb_entry(parent, struct tree_entry, rb_node);

		if (offset < entry->start)
			p = &(*p)->rb_left;
		else if (offset > entry->end)
			p = &(*p)->rb_right;
		else
			return parent;
	}

	entry = rb_entry(node, struct tree_entry, rb_node);
	entry->in_tree = 1;
	rb_link_node(node, parent, p);
	rb_insert_color(node, root);
	return NULL;
}

static struct rb_node *__tree_search(struct rb_root *root, u64 offset,
				   struct rb_node **prev_ret)
{
	struct rb_node * n = root->rb_node;
	struct rb_node *prev = NULL;
	struct tree_entry *entry;
	struct tree_entry *prev_entry = NULL;

	while(n) {
		entry = rb_entry(n, struct tree_entry, rb_node);
		prev = n;
		prev_entry = entry;

		if (offset < entry->start)
			n = n->rb_left;
		else if (offset > entry->end)
			n = n->rb_right;
		else
			return n;
	}
	if (!prev_ret)
		return NULL;
	while(prev && offset > prev_entry->end) {
		prev = rb_next(prev);
		prev_entry = rb_entry(prev, struct tree_entry, rb_node);
	}
	*prev_ret = prev;
	return NULL;
}

static inline struct rb_node *tree_search(struct rb_root *root, u64 offset)
{
	struct rb_node *prev;
	struct rb_node *ret;
	ret = __tree_search(root, offset, &prev);
	if (!ret)
		return prev;
	return ret;
}

static int tree_delete(struct rb_root *root, u64 offset)
{
	struct rb_node *node;
	struct tree_entry *entry;

	node = __tree_search(root, offset, NULL);
	if (!node)
		return -ENOENT;
	entry = rb_entry(node, struct tree_entry, rb_node);
	entry->in_tree = 0;
	rb_erase(node, root);
	return 0;
}

/*
 * add_extent_mapping tries a simple backward merge with existing
 * mappings.  The extent_map struct passed in will be inserted into
 * the tree directly (no copies made, just a reference taken).
 */
int add_extent_mapping(struct extent_map_tree *tree,
		       struct extent_map *em)
{
	int ret = 0;
	struct extent_map *prev = NULL;
	struct rb_node *rb;

	write_lock_irq(&tree->lock);
	rb = tree_insert(&tree->map, em->end, &em->rb_node);
	if (rb) {
		prev = rb_entry(rb, struct extent_map, rb_node);
		printk("found extent map %Lu %Lu on insert of %Lu %Lu\n", prev->start, prev->end, em->start, em->end);
		ret = -EEXIST;
		goto out;
	}
	atomic_inc(&em->refs);
	if (em->start != 0) {
		rb = rb_prev(&em->rb_node);
		if (rb)
			prev = rb_entry(rb, struct extent_map, rb_node);
		if (prev && prev->end + 1 == em->start &&
263 264
		    ((em->block_start == EXTENT_MAP_HOLE &&
		      prev->block_start == EXTENT_MAP_HOLE) ||
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
			     (em->block_start == prev->block_end + 1))) {
			em->start = prev->start;
			em->block_start = prev->block_start;
			rb_erase(&prev->rb_node, &tree->map);
			prev->in_tree = 0;
			free_extent_map(prev);
		}
	 }
out:
	write_unlock_irq(&tree->lock);
	return ret;
}
EXPORT_SYMBOL(add_extent_mapping);

/*
 * lookup_extent_mapping returns the first extent_map struct in the
 * tree that intersects the [start, end] (inclusive) range.  There may
 * be additional objects in the tree that intersect, so check the object
 * returned carefully to make sure you don't need additional lookups.
 */
struct extent_map *lookup_extent_mapping(struct extent_map_tree *tree,
					 u64 start, u64 end)
{
	struct extent_map *em;
	struct rb_node *rb_node;

	read_lock_irq(&tree->lock);
	rb_node = tree_search(&tree->map, start);
	if (!rb_node) {
		em = NULL;
		goto out;
	}
	if (IS_ERR(rb_node)) {
		em = ERR_PTR(PTR_ERR(rb_node));
		goto out;
	}
	em = rb_entry(rb_node, struct extent_map, rb_node);
	if (em->end < start || em->start > end) {
		em = NULL;
		goto out;
	}
	atomic_inc(&em->refs);
out:
	read_unlock_irq(&tree->lock);
	return em;
}
EXPORT_SYMBOL(lookup_extent_mapping);

/*
 * removes an extent_map struct from the tree.  No reference counts are
 * dropped, and no checks are done to  see if the range is in use
 */
int remove_extent_mapping(struct extent_map_tree *tree, struct extent_map *em)
{
	int ret;

	write_lock_irq(&tree->lock);
	ret = tree_delete(&tree->map, em->end);
	write_unlock_irq(&tree->lock);
	return ret;
}
EXPORT_SYMBOL(remove_extent_mapping);

/*
 * utility function to look for merge candidates inside a given range.
 * Any extents with matching state are merged together into a single
 * extent in the tree.  Extents with EXTENT_IO in their state field
 * are not merged because the end_io handlers need to be able to do
 * operations on them without sleeping (or doing allocations/splits).
 *
 * This should be called with the tree lock held.
 */
static int merge_state(struct extent_map_tree *tree,
		       struct extent_state *state)
{
	struct extent_state *other;
	struct rb_node *other_node;

	if (state->state & EXTENT_IOBITS)
		return 0;

	other_node = rb_prev(&state->rb_node);
	if (other_node) {
		other = rb_entry(other_node, struct extent_state, rb_node);
		if (other->end == state->start - 1 &&
		    other->state == state->state) {
			state->start = other->start;
			other->in_tree = 0;
			rb_erase(&other->rb_node, &tree->state);
			free_extent_state(other);
		}
	}
	other_node = rb_next(&state->rb_node);
	if (other_node) {
		other = rb_entry(other_node, struct extent_state, rb_node);
		if (other->start == state->end + 1 &&
		    other->state == state->state) {
			other->start = state->start;
			state->in_tree = 0;
			rb_erase(&state->rb_node, &tree->state);
			free_extent_state(state);
		}
	}
	return 0;
}

/*
 * insert an extent_state struct into the tree.  'bits' are set on the
 * struct before it is inserted.
 *
 * This may return -EEXIST if the extent is already there, in which case the
 * state struct is freed.
 *
 * The tree lock is not taken internally.  This is a utility function and
 * probably isn't what you want to call (see set/clear_extent_bit).
 */
static int insert_state(struct extent_map_tree *tree,
			struct extent_state *state, u64 start, u64 end,
			int bits)
{
	struct rb_node *node;

	if (end < start) {
		printk("end < start %Lu %Lu\n", end, start);
		WARN_ON(1);
	}
	state->state |= bits;
	state->start = start;
	state->end = end;
	node = tree_insert(&tree->state, end, &state->rb_node);
	if (node) {
		struct extent_state *found;
		found = rb_entry(node, struct extent_state, rb_node);
398
		printk("found node %Lu %Lu on insert of %Lu %Lu\n", found->start, found->end, start, end);
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
		free_extent_state(state);
		return -EEXIST;
	}
	merge_state(tree, state);
	return 0;
}

/*
 * split a given extent state struct in two, inserting the preallocated
 * struct 'prealloc' as the newly created second half.  'split' indicates an
 * offset inside 'orig' where it should be split.
 *
 * Before calling,
 * the tree has 'orig' at [orig->start, orig->end].  After calling, there
 * are two extent state structs in the tree:
 * prealloc: [orig->start, split - 1]
 * orig: [ split, orig->end ]
 *
 * The tree locks are not taken by this function. They need to be held
 * by the caller.
 */
static int split_state(struct extent_map_tree *tree, struct extent_state *orig,
		       struct extent_state *prealloc, u64 split)
{
	struct rb_node *node;
	prealloc->start = orig->start;
	prealloc->end = split - 1;
	prealloc->state = orig->state;
	orig->start = split;
428

429 430 431 432
	node = tree_insert(&tree->state, prealloc->end, &prealloc->rb_node);
	if (node) {
		struct extent_state *found;
		found = rb_entry(node, struct extent_state, rb_node);
433
		printk("found node %Lu %Lu on insert of %Lu %Lu\n", found->start, found->end, prealloc->start, prealloc->end);
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
		free_extent_state(prealloc);
		return -EEXIST;
	}
	return 0;
}

/*
 * utility function to clear some bits in an extent state struct.
 * it will optionally wake up any one waiting on this state (wake == 1), or
 * forcibly remove the state from the tree (delete == 1).
 *
 * If no bits are set on the state struct after clearing things, the
 * struct is freed and removed from the tree
 */
static int clear_state_bit(struct extent_map_tree *tree,
			    struct extent_state *state, int bits, int wake,
			    int delete)
{
	int ret = state->state & bits;
	state->state &= ~bits;
	if (wake)
		wake_up(&state->wq);
	if (delete || state->state == 0) {
		if (state->in_tree) {
			rb_erase(&state->rb_node, &tree->state);
			state->in_tree = 0;
			free_extent_state(state);
		} else {
			WARN_ON(1);
		}
	} else {
		merge_state(tree, state);
	}
	return ret;
}

/*
 * clear some bits on a range in the tree.  This may require splitting
 * or inserting elements in the tree, so the gfp mask is used to
 * indicate which allocations or sleeping are allowed.
 *
 * pass 'wake' == 1 to kick any sleepers, and 'delete' == 1 to remove
 * the given range from the tree regardless of state (ie for truncate).
 *
 * the range [start, end] is inclusive.
 *
 * This takes the tree lock, and returns < 0 on error, > 0 if any of the
 * bits were already set, or zero if none of the bits were already set.
 */
int clear_extent_bit(struct extent_map_tree *tree, u64 start, u64 end,
		     int bits, int wake, int delete, gfp_t mask)
{
	struct extent_state *state;
	struct extent_state *prealloc = NULL;
	struct rb_node *node;
489
	unsigned long flags;
490 491 492 493 494 495 496 497 498 499
	int err;
	int set = 0;

again:
	if (!prealloc && (mask & __GFP_WAIT)) {
		prealloc = alloc_extent_state(mask);
		if (!prealloc)
			return -ENOMEM;
	}

500
	write_lock_irqsave(&tree->lock, flags);
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
	/*
	 * this search will find the extents that end after
	 * our range starts
	 */
	node = tree_search(&tree->state, start);
	if (!node)
		goto out;
	state = rb_entry(node, struct extent_state, rb_node);
	if (state->start > end)
		goto out;
	WARN_ON(state->end < start);

	/*
	 *     | ---- desired range ---- |
	 *  | state | or
	 *  | ------------- state -------------- |
	 *
	 * We need to split the extent we found, and may flip
	 * bits on second half.
	 *
	 * If the extent we found extends past our range, we
	 * just split and search again.  It'll get split again
	 * the next time though.
	 *
	 * If the extent we found is inside our range, we clear
	 * the desired bit on it.
	 */

	if (state->start < start) {
		err = split_state(tree, state, prealloc, start);
		BUG_ON(err == -EEXIST);
		prealloc = NULL;
		if (err)
			goto out;
		if (state->end <= end) {
			start = state->end + 1;
			set |= clear_state_bit(tree, state, bits,
					wake, delete);
		} else {
			start = state->start;
		}
		goto search_again;
	}
	/*
	 * | ---- desired range ---- |
	 *                        | state |
	 * We need to split the extent, and clear the bit
	 * on the first half
	 */
	if (state->start <= end && state->end > end) {
		err = split_state(tree, state, prealloc, end + 1);
		BUG_ON(err == -EEXIST);

		if (wake)
			wake_up(&state->wq);
		set |= clear_state_bit(tree, prealloc, bits,
				       wake, delete);
		prealloc = NULL;
		goto out;
	}

	start = state->end + 1;
	set |= clear_state_bit(tree, state, bits, wake, delete);
	goto search_again;

out:
567
	write_unlock_irqrestore(&tree->lock, flags);
568 569 570 571 572 573
	if (prealloc)
		free_extent_state(prealloc);

	return set;

search_again:
574
	if (start > end)
575
		goto out;
576
	write_unlock_irqrestore(&tree->lock, flags);
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
	if (mask & __GFP_WAIT)
		cond_resched();
	goto again;
}
EXPORT_SYMBOL(clear_extent_bit);

static int wait_on_state(struct extent_map_tree *tree,
			 struct extent_state *state)
{
	DEFINE_WAIT(wait);
	prepare_to_wait(&state->wq, &wait, TASK_UNINTERRUPTIBLE);
	read_unlock_irq(&tree->lock);
	schedule();
	read_lock_irq(&tree->lock);
	finish_wait(&state->wq, &wait);
	return 0;
}

/*
 * waits for one or more bits to clear on a range in the state tree.
 * The range [start, end] is inclusive.
 * The tree lock is taken by this function
 */
int wait_extent_bit(struct extent_map_tree *tree, u64 start, u64 end, int bits)
{
	struct extent_state *state;
	struct rb_node *node;

	read_lock_irq(&tree->lock);
again:
	while (1) {
		/*
		 * this search will find all the extents that end after
		 * our range starts
		 */
		node = tree_search(&tree->state, start);
		if (!node)
			break;

		state = rb_entry(node, struct extent_state, rb_node);

		if (state->start > end)
			goto out;

		if (state->state & bits) {
			start = state->start;
			atomic_inc(&state->refs);
			wait_on_state(tree, state);
			free_extent_state(state);
			goto again;
		}
		start = state->end + 1;

		if (start > end)
			break;

		if (need_resched()) {
			read_unlock_irq(&tree->lock);
			cond_resched();
			read_lock_irq(&tree->lock);
		}
	}
out:
	read_unlock_irq(&tree->lock);
	return 0;
}
EXPORT_SYMBOL(wait_extent_bit);

/*
 * set some bits on a range in the tree.  This may require allocations
 * or sleeping, so the gfp mask is used to indicate what is allowed.
 *
 * If 'exclusive' == 1, this will fail with -EEXIST if some part of the
 * range already has the desired bits set.  The start of the existing
 * range is returned in failed_start in this case.
 *
 * [start, end] is inclusive
 * This takes the tree lock.
 */
int set_extent_bit(struct extent_map_tree *tree, u64 start, u64 end, int bits,
		   int exclusive, u64 *failed_start, gfp_t mask)
{
	struct extent_state *state;
	struct extent_state *prealloc = NULL;
	struct rb_node *node;
662
	unsigned long flags;
663 664 665 666 667 668 669 670 671 672 673
	int err = 0;
	int set;
	u64 last_start;
	u64 last_end;
again:
	if (!prealloc && (mask & __GFP_WAIT)) {
		prealloc = alloc_extent_state(mask);
		if (!prealloc)
			return -ENOMEM;
	}

674
	write_lock_irqsave(&tree->lock, flags);
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
	/*
	 * this search will find all the extents that end after
	 * our range starts.
	 */
	node = tree_search(&tree->state, start);
	if (!node) {
		err = insert_state(tree, prealloc, start, end, bits);
		prealloc = NULL;
		BUG_ON(err == -EEXIST);
		goto out;
	}

	state = rb_entry(node, struct extent_state, rb_node);
	last_start = state->start;
	last_end = state->end;

	/*
	 * | ---- desired range ---- |
	 * | state |
	 *
	 * Just lock what we found and keep going
	 */
	if (state->start == start && state->end <= end) {
		set = state->state & bits;
		if (set && exclusive) {
			*failed_start = state->start;
			err = -EEXIST;
			goto out;
		}
		state->state |= bits;
		start = state->end + 1;
		merge_state(tree, state);
		goto search_again;
	}

	/*
	 *     | ---- desired range ---- |
	 * | state |
	 *   or
	 * | ------------- state -------------- |
	 *
	 * We need to split the extent we found, and may flip bits on
	 * second half.
	 *
	 * If the extent we found extends past our
	 * range, we just split and search again.  It'll get split
	 * again the next time though.
	 *
	 * If the extent we found is inside our range, we set the
	 * desired bit on it.
	 */
	if (state->start < start) {
		set = state->state & bits;
		if (exclusive && set) {
			*failed_start = start;
			err = -EEXIST;
			goto out;
		}
		err = split_state(tree, state, prealloc, start);
		BUG_ON(err == -EEXIST);
		prealloc = NULL;
		if (err)
			goto out;
		if (state->end <= end) {
			state->state |= bits;
			start = state->end + 1;
			merge_state(tree, state);
		} else {
			start = state->start;
		}
		goto search_again;
	}
	/*
	 * | ---- desired range ---- |
	 *     | state | or               | state |
	 *
	 * There's a hole, we need to insert something in it and
	 * ignore the extent we found.
	 */
	if (state->start > start) {
		u64 this_end;
		if (end < last_start)
			this_end = end;
		else
			this_end = last_start -1;
		err = insert_state(tree, prealloc, start, this_end,
				   bits);
		prealloc = NULL;
		BUG_ON(err == -EEXIST);
		if (err)
			goto out;
		start = this_end + 1;
		goto search_again;
	}
769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790
	/*
	 * | ---- desired range ---- |
	 *                        | state |
	 * We need to split the extent, and set the bit
	 * on the first half
	 */
	if (state->start <= end && state->end > end) {
		set = state->state & bits;
		if (exclusive && set) {
			*failed_start = start;
			err = -EEXIST;
			goto out;
		}
		err = split_state(tree, state, prealloc, end + 1);
		BUG_ON(err == -EEXIST);

		prealloc->state |= bits;
		merge_state(tree, prealloc);
		prealloc = NULL;
		goto out;
	}

791 792 793
	goto search_again;

out:
794
	write_unlock_irqrestore(&tree->lock, flags);
795 796 797 798 799 800 801 802
	if (prealloc)
		free_extent_state(prealloc);

	return err;

search_again:
	if (start > end)
		goto out;
803
	write_unlock_irqrestore(&tree->lock, flags);
804 805 806 807 808 809 810 811 812 813 814 815 816 817 818
	if (mask & __GFP_WAIT)
		cond_resched();
	goto again;
}
EXPORT_SYMBOL(set_extent_bit);

/* wrappers around set/clear extent bit */
int set_extent_dirty(struct extent_map_tree *tree, u64 start, u64 end,
		     gfp_t mask)
{
	return set_extent_bit(tree, start, end, EXTENT_DIRTY, 0, NULL,
			      mask);
}
EXPORT_SYMBOL(set_extent_dirty);

819 820 821 822 823 824 825 826 827 828 829 830 831 832 833
int set_extent_bits(struct extent_map_tree *tree, u64 start, u64 end,
		    int bits, gfp_t mask)
{
	return set_extent_bit(tree, start, end, bits, 0, NULL,
			      mask);
}
EXPORT_SYMBOL(set_extent_bits);

int clear_extent_bits(struct extent_map_tree *tree, u64 start, u64 end,
		      int bits, gfp_t mask)
{
	return clear_extent_bit(tree, start, end, bits, 0, 0, mask);
}
EXPORT_SYMBOL(clear_extent_bits);

834 835 836 837 838 839 840 841 842
int set_extent_delalloc(struct extent_map_tree *tree, u64 start, u64 end,
		     gfp_t mask)
{
	return set_extent_bit(tree, start, end,
			      EXTENT_DELALLOC | EXTENT_DIRTY, 0, NULL,
			      mask);
}
EXPORT_SYMBOL(set_extent_delalloc);

843 844 845
int clear_extent_dirty(struct extent_map_tree *tree, u64 start, u64 end,
		       gfp_t mask)
{
846 847
	return clear_extent_bit(tree, start, end,
				EXTENT_DIRTY | EXTENT_DELALLOC, 0, 0, mask);
848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973
}
EXPORT_SYMBOL(clear_extent_dirty);

int set_extent_new(struct extent_map_tree *tree, u64 start, u64 end,
		     gfp_t mask)
{
	return set_extent_bit(tree, start, end, EXTENT_NEW, 0, NULL,
			      mask);
}
EXPORT_SYMBOL(set_extent_new);

int clear_extent_new(struct extent_map_tree *tree, u64 start, u64 end,
		       gfp_t mask)
{
	return clear_extent_bit(tree, start, end, EXTENT_NEW, 0, 0, mask);
}
EXPORT_SYMBOL(clear_extent_new);

int set_extent_uptodate(struct extent_map_tree *tree, u64 start, u64 end,
			gfp_t mask)
{
	return set_extent_bit(tree, start, end, EXTENT_UPTODATE, 0, NULL,
			      mask);
}
EXPORT_SYMBOL(set_extent_uptodate);

int clear_extent_uptodate(struct extent_map_tree *tree, u64 start, u64 end,
			  gfp_t mask)
{
	return clear_extent_bit(tree, start, end, EXTENT_UPTODATE, 0, 0, mask);
}
EXPORT_SYMBOL(clear_extent_uptodate);

int set_extent_writeback(struct extent_map_tree *tree, u64 start, u64 end,
			 gfp_t mask)
{
	return set_extent_bit(tree, start, end, EXTENT_WRITEBACK,
			      0, NULL, mask);
}
EXPORT_SYMBOL(set_extent_writeback);

int clear_extent_writeback(struct extent_map_tree *tree, u64 start, u64 end,
			   gfp_t mask)
{
	return clear_extent_bit(tree, start, end, EXTENT_WRITEBACK, 1, 0, mask);
}
EXPORT_SYMBOL(clear_extent_writeback);

int wait_on_extent_writeback(struct extent_map_tree *tree, u64 start, u64 end)
{
	return wait_extent_bit(tree, start, end, EXTENT_WRITEBACK);
}
EXPORT_SYMBOL(wait_on_extent_writeback);

/*
 * locks a range in ascending order, waiting for any locked regions
 * it hits on the way.  [start,end] are inclusive, and this will sleep.
 */
int lock_extent(struct extent_map_tree *tree, u64 start, u64 end, gfp_t mask)
{
	int err;
	u64 failed_start;
	while (1) {
		err = set_extent_bit(tree, start, end, EXTENT_LOCKED, 1,
				     &failed_start, mask);
		if (err == -EEXIST && (mask & __GFP_WAIT)) {
			wait_extent_bit(tree, failed_start, end, EXTENT_LOCKED);
			start = failed_start;
		} else {
			break;
		}
		WARN_ON(start > end);
	}
	return err;
}
EXPORT_SYMBOL(lock_extent);

int unlock_extent(struct extent_map_tree *tree, u64 start, u64 end,
		  gfp_t mask)
{
	return clear_extent_bit(tree, start, end, EXTENT_LOCKED, 1, 0, mask);
}
EXPORT_SYMBOL(unlock_extent);

/*
 * helper function to set pages and extents in the tree dirty
 */
int set_range_dirty(struct extent_map_tree *tree, u64 start, u64 end)
{
	unsigned long index = start >> PAGE_CACHE_SHIFT;
	unsigned long end_index = end >> PAGE_CACHE_SHIFT;
	struct page *page;

	while (index <= end_index) {
		page = find_get_page(tree->mapping, index);
		BUG_ON(!page);
		__set_page_dirty_nobuffers(page);
		page_cache_release(page);
		index++;
	}
	set_extent_dirty(tree, start, end, GFP_NOFS);
	return 0;
}
EXPORT_SYMBOL(set_range_dirty);

/*
 * helper function to set both pages and extents in the tree writeback
 */
int set_range_writeback(struct extent_map_tree *tree, u64 start, u64 end)
{
	unsigned long index = start >> PAGE_CACHE_SHIFT;
	unsigned long end_index = end >> PAGE_CACHE_SHIFT;
	struct page *page;

	while (index <= end_index) {
		page = find_get_page(tree->mapping, index);
		BUG_ON(!page);
		set_page_writeback(page);
		page_cache_release(page);
		index++;
	}
	set_extent_writeback(tree, start, end, GFP_NOFS);
	return 0;
}
EXPORT_SYMBOL(set_range_writeback);

974 975 976 977 978 979 980
int find_first_extent_bit(struct extent_map_tree *tree, u64 start,
			  u64 *start_ret, u64 *end_ret, int bits)
{
	struct rb_node *node;
	struct extent_state *state;
	int ret = 1;

981
	read_lock_irq(&tree->lock);
982 983 984 985 986 987 988 989 990 991 992
	/*
	 * this search will find all the extents that end after
	 * our range starts.
	 */
	node = tree_search(&tree->state, start);
	if (!node || IS_ERR(node)) {
		goto out;
	}

	while(1) {
		state = rb_entry(node, struct extent_state, rb_node);
993
		if (state->end >= start && (state->state & bits)) {
994 995 996
			*start_ret = state->start;
			*end_ret = state->end;
			ret = 0;
997
			break;
998 999 1000 1001 1002 1003
		}
		node = rb_next(node);
		if (!node)
			break;
	}
out:
1004
	read_unlock_irq(&tree->lock);
1005 1006 1007 1008
	return ret;
}
EXPORT_SYMBOL(find_first_extent_bit);

1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064
u64 find_lock_delalloc_range(struct extent_map_tree *tree,
			     u64 start, u64 lock_start, u64 *end, u64 max_bytes)
{
	struct rb_node *node;
	struct extent_state *state;
	u64 cur_start = start;
	u64 found = 0;
	u64 total_bytes = 0;

	write_lock_irq(&tree->lock);
	/*
	 * this search will find all the extents that end after
	 * our range starts.
	 */
search_again:
	node = tree_search(&tree->state, cur_start);
	if (!node || IS_ERR(node)) {
		goto out;
	}

	while(1) {
		state = rb_entry(node, struct extent_state, rb_node);
		if (state->start != cur_start) {
			goto out;
		}
		if (!(state->state & EXTENT_DELALLOC)) {
			goto out;
		}
		if (state->start >= lock_start) {
			if (state->state & EXTENT_LOCKED) {
				DEFINE_WAIT(wait);
				atomic_inc(&state->refs);
				write_unlock_irq(&tree->lock);
				schedule();
				write_lock_irq(&tree->lock);
				finish_wait(&state->wq, &wait);
				free_extent_state(state);
				goto search_again;
			}
			state->state |= EXTENT_LOCKED;
		}
		found++;
		*end = state->end;
		cur_start = state->end + 1;
		node = rb_next(node);
		if (!node)
			break;
		total_bytes = state->end - state->start + 1;
		if (total_bytes >= max_bytes)
			break;
	}
out:
	write_unlock_irq(&tree->lock);
	return found;
}

1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127
/*
 * helper function to lock both pages and extents in the tree.
 * pages must be locked first.
 */
int lock_range(struct extent_map_tree *tree, u64 start, u64 end)
{
	unsigned long index = start >> PAGE_CACHE_SHIFT;
	unsigned long end_index = end >> PAGE_CACHE_SHIFT;
	struct page *page;
	int err;

	while (index <= end_index) {
		page = grab_cache_page(tree->mapping, index);
		if (!page) {
			err = -ENOMEM;
			goto failed;
		}
		if (IS_ERR(page)) {
			err = PTR_ERR(page);
			goto failed;
		}
		index++;
	}
	lock_extent(tree, start, end, GFP_NOFS);
	return 0;

failed:
	/*
	 * we failed above in getting the page at 'index', so we undo here
	 * up to but not including the page at 'index'
	 */
	end_index = index;
	index = start >> PAGE_CACHE_SHIFT;
	while (index < end_index) {
		page = find_get_page(tree->mapping, index);
		unlock_page(page);
		page_cache_release(page);
		index++;
	}
	return err;
}
EXPORT_SYMBOL(lock_range);

/*
 * helper function to unlock both pages and extents in the tree.
 */
int unlock_range(struct extent_map_tree *tree, u64 start, u64 end)
{
	unsigned long index = start >> PAGE_CACHE_SHIFT;
	unsigned long end_index = end >> PAGE_CACHE_SHIFT;
	struct page *page;

	while (index <= end_index) {
		page = find_get_page(tree->mapping, index);
		unlock_page(page);
		page_cache_release(page);
		index++;
	}
	unlock_extent(tree, start, end, GFP_NOFS);
	return 0;
}
EXPORT_SYMBOL(unlock_range);

1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181
int set_state_private(struct extent_map_tree *tree, u64 start, u64 private)
{
	struct rb_node *node;
	struct extent_state *state;
	int ret = 0;

	write_lock_irq(&tree->lock);
	/*
	 * this search will find all the extents that end after
	 * our range starts.
	 */
	node = tree_search(&tree->state, start);
	if (!node || IS_ERR(node)) {
		ret = -ENOENT;
		goto out;
	}
	state = rb_entry(node, struct extent_state, rb_node);
	if (state->start != start) {
		ret = -ENOENT;
		goto out;
	}
	state->private = private;
out:
	write_unlock_irq(&tree->lock);
	return ret;
}

int get_state_private(struct extent_map_tree *tree, u64 start, u64 *private)
{
	struct rb_node *node;
	struct extent_state *state;
	int ret = 0;

	read_lock_irq(&tree->lock);
	/*
	 * this search will find all the extents that end after
	 * our range starts.
	 */
	node = tree_search(&tree->state, start);
	if (!node || IS_ERR(node)) {
		ret = -ENOENT;
		goto out;
	}
	state = rb_entry(node, struct extent_state, rb_node);
	if (state->start != start) {
		ret = -ENOENT;
		goto out;
	}
	*private = state->private;
out:
	read_unlock_irq(&tree->lock);
	return ret;
}

1182 1183 1184 1185 1186 1187
/*
 * searches a range in the state tree for a given mask.
 * If 'filled' == 1, this returns 1 only if ever extent in the tree
 * has the bits set.  Otherwise, 1 is returned if any bit in the
 * range is found set.
 */
1188 1189
int test_range_bit(struct extent_map_tree *tree, u64 start, u64 end,
		   int bits, int filled)
1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221
{
	struct extent_state *state = NULL;
	struct rb_node *node;
	int bitset = 0;

	read_lock_irq(&tree->lock);
	node = tree_search(&tree->state, start);
	while (node && start <= end) {
		state = rb_entry(node, struct extent_state, rb_node);
		if (state->start > end)
			break;

		if (filled && state->start > start) {
			bitset = 0;
			break;
		}
		if (state->state & bits) {
			bitset = 1;
			if (!filled)
				break;
		} else if (filled) {
			bitset = 0;
			break;
		}
		start = state->end + 1;
		if (start > end)
			break;
		node = rb_next(node);
	}
	read_unlock_irq(&tree->lock);
	return bitset;
}
1222
EXPORT_SYMBOL(test_range_bit);
1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313

/*
 * helper function to set a given page up to date if all the
 * extents in the tree for that page are up to date
 */
static int check_page_uptodate(struct extent_map_tree *tree,
			       struct page *page)
{
	u64 start = page->index << PAGE_CACHE_SHIFT;
	u64 end = start + PAGE_CACHE_SIZE - 1;
	if (test_range_bit(tree, start, end, EXTENT_UPTODATE, 1))
		SetPageUptodate(page);
	return 0;
}

/*
 * helper function to unlock a page if all the extents in the tree
 * for that page are unlocked
 */
static int check_page_locked(struct extent_map_tree *tree,
			     struct page *page)
{
	u64 start = page->index << PAGE_CACHE_SHIFT;
	u64 end = start + PAGE_CACHE_SIZE - 1;
	if (!test_range_bit(tree, start, end, EXTENT_LOCKED, 0))
		unlock_page(page);
	return 0;
}

/*
 * helper function to end page writeback if all the extents
 * in the tree for that page are done with writeback
 */
static int check_page_writeback(struct extent_map_tree *tree,
			     struct page *page)
{
	u64 start = page->index << PAGE_CACHE_SHIFT;
	u64 end = start + PAGE_CACHE_SIZE - 1;
	if (!test_range_bit(tree, start, end, EXTENT_WRITEBACK, 0))
		end_page_writeback(page);
	return 0;
}

/* lots and lots of room for performance fixes in the end_bio funcs */

/*
 * after a writepage IO is done, we need to:
 * clear the uptodate bits on error
 * clear the writeback bits in the extent tree for this IO
 * end_page_writeback if the page has no more pending IO
 *
 * Scheduling is not allowed, so the extent state tree is expected
 * to have one and only one object corresponding to this IO.
 */
static int end_bio_extent_writepage(struct bio *bio,
				   unsigned int bytes_done, int err)
{
	const int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
	struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
	struct extent_map_tree *tree = bio->bi_private;
	u64 start;
	u64 end;
	int whole_page;

	if (bio->bi_size)
		return 1;

	do {
		struct page *page = bvec->bv_page;
		start = (page->index << PAGE_CACHE_SHIFT) + bvec->bv_offset;
		end = start + bvec->bv_len - 1;

		if (bvec->bv_offset == 0 && bvec->bv_len == PAGE_CACHE_SIZE)
			whole_page = 1;
		else
			whole_page = 0;

		if (--bvec >= bio->bi_io_vec)
			prefetchw(&bvec->bv_page->flags);

		if (!uptodate) {
			clear_extent_uptodate(tree, start, end, GFP_ATOMIC);
			ClearPageUptodate(page);
			SetPageError(page);
		}
		clear_extent_writeback(tree, start, end, GFP_ATOMIC);

		if (whole_page)
			end_page_writeback(page);
		else
			check_page_writeback(tree, page);
1314 1315
		if (tree->ops && tree->ops->writepage_end_io_hook)
			tree->ops->writepage_end_io_hook(page, start, end);
1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335
	} while (bvec >= bio->bi_io_vec);

	bio_put(bio);
	return 0;
}

/*
 * after a readpage IO is done, we need to:
 * clear the uptodate bits on error
 * set the uptodate bits if things worked
 * set the page up to date if all extents in the tree are uptodate
 * clear the lock bit in the extent tree
 * unlock the page if there are no other extents locked for it
 *
 * Scheduling is not allowed, so the extent state tree is expected
 * to have one and only one object corresponding to this IO.
 */
static int end_bio_extent_readpage(struct bio *bio,
				   unsigned int bytes_done, int err)
{
1336
	int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
1337 1338 1339 1340 1341
	struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
	struct extent_map_tree *tree = bio->bi_private;
	u64 start;
	u64 end;
	int whole_page;
1342
	int ret;
1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359

	if (bio->bi_size)
		return 1;

	do {
		struct page *page = bvec->bv_page;
		start = (page->index << PAGE_CACHE_SHIFT) + bvec->bv_offset;
		end = start + bvec->bv_len - 1;

		if (bvec->bv_offset == 0 && bvec->bv_len == PAGE_CACHE_SIZE)
			whole_page = 1;
		else
			whole_page = 0;

		if (--bvec >= bio->bi_io_vec)
			prefetchw(&bvec->bv_page->flags);

1360 1361 1362 1363 1364
		if (uptodate && tree->ops && tree->ops->readpage_end_io_hook) {
			ret = tree->ops->readpage_end_io_hook(page, start, end);
			if (ret)
				uptodate = 0;
		}
1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461
		if (uptodate) {
			set_extent_uptodate(tree, start, end, GFP_ATOMIC);
			if (whole_page)
				SetPageUptodate(page);
			else
				check_page_uptodate(tree, page);
		} else {
			ClearPageUptodate(page);
			SetPageError(page);
		}

		unlock_extent(tree, start, end, GFP_ATOMIC);

		if (whole_page)
			unlock_page(page);
		else
			check_page_locked(tree, page);
	} while (bvec >= bio->bi_io_vec);

	bio_put(bio);
	return 0;
}

/*
 * IO done from prepare_write is pretty simple, we just unlock
 * the structs in the extent tree when done, and set the uptodate bits
 * as appropriate.
 */
static int end_bio_extent_preparewrite(struct bio *bio,
				       unsigned int bytes_done, int err)
{
	const int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
	struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
	struct extent_map_tree *tree = bio->bi_private;
	u64 start;
	u64 end;

	if (bio->bi_size)
		return 1;

	do {
		struct page *page = bvec->bv_page;
		start = (page->index << PAGE_CACHE_SHIFT) + bvec->bv_offset;
		end = start + bvec->bv_len - 1;

		if (--bvec >= bio->bi_io_vec)
			prefetchw(&bvec->bv_page->flags);

		if (uptodate) {
			set_extent_uptodate(tree, start, end, GFP_ATOMIC);
		} else {
			ClearPageUptodate(page);
			SetPageError(page);
		}

		unlock_extent(tree, start, end, GFP_ATOMIC);

	} while (bvec >= bio->bi_io_vec);

	bio_put(bio);
	return 0;
}

static int submit_extent_page(int rw, struct extent_map_tree *tree,
			      struct page *page, sector_t sector,
			      size_t size, unsigned long offset,
			      struct block_device *bdev,
			      bio_end_io_t end_io_func)
{
	struct bio *bio;
	int ret = 0;

	bio = bio_alloc(GFP_NOIO, 1);

	bio->bi_sector = sector;
	bio->bi_bdev = bdev;
	bio->bi_io_vec[0].bv_page = page;
	bio->bi_io_vec[0].bv_len = size;
	bio->bi_io_vec[0].bv_offset = offset;

	bio->bi_vcnt = 1;
	bio->bi_idx = 0;
	bio->bi_size = size;

	bio->bi_end_io = end_io_func;
	bio->bi_private = tree;

	bio_get(bio);
	submit_bio(rw, bio);

	if (bio_flagged(bio, BIO_EOPNOTSUPP))
		ret = -EOPNOTSUPP;

	bio_put(bio);
	return ret;
}

1462 1463 1464 1465 1466
void set_page_extent_mapped(struct page *page)
{
	if (!PagePrivate(page)) {
		SetPagePrivate(page);
		WARN_ON(!page->mapping->a_ops->invalidatepage);
1467
		set_page_private(page, EXTENT_PAGE_PRIVATE);
1468 1469 1470 1471
		page_cache_get(page);
	}
}

1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497
/*
 * basic readpage implementation.  Locked extent state structs are inserted
 * into the tree that are removed when the IO is done (by the end_io
 * handlers)
 */
int extent_read_full_page(struct extent_map_tree *tree, struct page *page,
			  get_extent_t *get_extent)
{
	struct inode *inode = page->mapping->host;
	u64 start = page->index << PAGE_CACHE_SHIFT;
	u64 page_end = start + PAGE_CACHE_SIZE - 1;
	u64 end;
	u64 cur = start;
	u64 extent_offset;
	u64 last_byte = i_size_read(inode);
	u64 block_start;
	u64 cur_end;
	sector_t sector;
	struct extent_map *em;
	struct block_device *bdev;
	int ret;
	int nr = 0;
	size_t page_offset = 0;
	size_t iosize;
	size_t blocksize = inode->i_sb->s_blocksize;

1498
	set_page_extent_mapped(page);
1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532

	end = page_end;
	lock_extent(tree, start, end, GFP_NOFS);

	while (cur <= end) {
		if (cur >= last_byte) {
			iosize = PAGE_CACHE_SIZE - page_offset;
			zero_user_page(page, page_offset, iosize, KM_USER0);
			set_extent_uptodate(tree, cur, cur + iosize - 1,
					    GFP_NOFS);
			unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS);
			break;
		}
		em = get_extent(inode, page, page_offset, cur, end, 0);
		if (IS_ERR(em) || !em) {
			SetPageError(page);
			unlock_extent(tree, cur, end, GFP_NOFS);
			break;
		}

		extent_offset = cur - em->start;
		BUG_ON(em->end < cur);
		BUG_ON(end < cur);

		iosize = min(em->end - cur, end - cur) + 1;
		cur_end = min(em->end, end);
		iosize = (iosize + blocksize - 1) & ~((u64)blocksize - 1);
		sector = (em->block_start + extent_offset) >> 9;
		bdev = em->bdev;
		block_start = em->block_start;
		free_extent_map(em);
		em = NULL;

		/* we've found a hole, just zero and go on */
1533
		if (block_start == EXTENT_MAP_HOLE) {
1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549
			zero_user_page(page, page_offset, iosize, KM_USER0);
			set_extent_uptodate(tree, cur, cur + iosize - 1,
					    GFP_NOFS);
			unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS);
			cur = cur + iosize;
			page_offset += iosize;
			continue;
		}
		/* the get_extent function already copied into the page */
		if (test_range_bit(tree, cur, cur_end, EXTENT_UPTODATE, 1)) {
			unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS);
			cur = cur + iosize;
			page_offset += iosize;
			continue;
		}

1550 1551 1552 1553 1554 1555 1556 1557 1558 1559
		ret = 0;
		if (tree->ops && tree->ops->readpage_io_hook) {
			ret = tree->ops->readpage_io_hook(page, cur,
							  cur + iosize - 1);
		}
		if (!ret) {
			ret = submit_extent_page(READ, tree, page,
						 sector, iosize, page_offset,
						 bdev, end_bio_extent_readpage);
		}
1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602
		if (ret)
			SetPageError(page);
		cur = cur + iosize;
		page_offset += iosize;
		nr++;
	}
	if (!nr) {
		if (!PageError(page))
			SetPageUptodate(page);
		unlock_page(page);
	}
	return 0;
}
EXPORT_SYMBOL(extent_read_full_page);

/*
 * the writepage semantics are similar to regular writepage.  extent
 * records are inserted to lock ranges in the tree, and as dirty areas
 * are found, they are marked writeback.  Then the lock bits are removed
 * and the end_io handler clears the writeback ranges
 */
int extent_write_full_page(struct extent_map_tree *tree, struct page *page,
			  get_extent_t *get_extent,
			  struct writeback_control *wbc)
{
	struct inode *inode = page->mapping->host;
	u64 start = page->index << PAGE_CACHE_SHIFT;
	u64 page_end = start + PAGE_CACHE_SIZE - 1;
	u64 end;
	u64 cur = start;
	u64 extent_offset;
	u64 last_byte = i_size_read(inode);
	u64 block_start;
	sector_t sector;
	struct extent_map *em;
	struct block_device *bdev;
	int ret;
	int nr = 0;
	size_t page_offset = 0;
	size_t iosize;
	size_t blocksize;
	loff_t i_size = i_size_read(inode);
	unsigned long end_index = i_size >> PAGE_CACHE_SHIFT;
1603 1604
	u64 nr_delalloc;
	u64 delalloc_end;
1605

1606
	WARN_ON(!PageLocked(page));
1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618
	if (page->index > end_index) {
		clear_extent_dirty(tree, start, page_end, GFP_NOFS);
		unlock_page(page);
		return 0;
	}

	if (page->index == end_index) {
		size_t offset = i_size & (PAGE_CACHE_SIZE - 1);
		zero_user_page(page, offset,
			       PAGE_CACHE_SIZE - offset, KM_USER0);
	}

1619
	set_page_extent_mapped(page);
1620 1621

	lock_extent(tree, start, page_end, GFP_NOFS);
1622 1623 1624 1625
	nr_delalloc = find_lock_delalloc_range(tree, start, page_end + 1,
					       &delalloc_end,
					       128 * 1024 * 1024);
	if (nr_delalloc) {
1626
		tree->ops->fill_delalloc(inode, start, delalloc_end);
1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644
		if (delalloc_end >= page_end + 1) {
			clear_extent_bit(tree, page_end + 1, delalloc_end,
					 EXTENT_LOCKED | EXTENT_DELALLOC,
					 1, 0, GFP_NOFS);
		}
		clear_extent_bit(tree, start, page_end, EXTENT_DELALLOC,
				 0, 0, GFP_NOFS);
		if (test_range_bit(tree, start, page_end, EXTENT_DELALLOC, 0)) {
			printk("found delalloc bits after clear extent_bit\n");
		}
	} else if (test_range_bit(tree, start, page_end, EXTENT_DELALLOC, 0)) {
		printk("found delalloc bits after find_delalloc_range returns 0\n");
	}

	end = page_end;
	if (test_range_bit(tree, start, page_end, EXTENT_DELALLOC, 0)) {
		printk("found delalloc bits after lock_extent\n");
	}
1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658

	if (last_byte <= start) {
		clear_extent_dirty(tree, start, page_end, GFP_NOFS);
		goto done;
	}

	set_extent_uptodate(tree, start, page_end, GFP_NOFS);
	blocksize = inode->i_sb->s_blocksize;

	while (cur <= end) {
		if (cur >= last_byte) {
			clear_extent_dirty(tree, cur, page_end, GFP_NOFS);
			break;
		}
1659
		em = get_extent(inode, page, page_offset, cur, end, 0);
1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675
		if (IS_ERR(em) || !em) {
			SetPageError(page);
			break;
		}

		extent_offset = cur - em->start;
		BUG_ON(em->end < cur);
		BUG_ON(end < cur);
		iosize = min(em->end - cur, end - cur) + 1;
		iosize = (iosize + blocksize - 1) & ~((u64)blocksize - 1);
		sector = (em->block_start + extent_offset) >> 9;
		bdev = em->bdev;
		block_start = em->block_start;
		free_extent_map(em);
		em = NULL;

1676 1677
		if (block_start == EXTENT_MAP_HOLE ||
		    block_start == EXTENT_MAP_INLINE) {
1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692
			clear_extent_dirty(tree, cur,
					   cur + iosize - 1, GFP_NOFS);
			cur = cur + iosize;
			page_offset += iosize;
			continue;
		}

		/* leave this out until we have a page_mkwrite call */
		if (0 && !test_range_bit(tree, cur, cur + iosize - 1,
				   EXTENT_DIRTY, 0)) {
			cur = cur + iosize;
			page_offset += iosize;
			continue;
		}
		clear_extent_dirty(tree, cur, cur + iosize - 1, GFP_NOFS);
1693 1694 1695 1696 1697 1698
		if (tree->ops && tree->ops->writepage_io_hook) {
			ret = tree->ops->writepage_io_hook(page, cur,
						cur + iosize - 1);
		} else {
			ret = 0;
		}
1699 1700
		if (ret)
			SetPageError(page);
1701 1702 1703 1704 1705 1706 1707 1708
		else {
			set_range_writeback(tree, cur, cur + iosize - 1);
			ret = submit_extent_page(WRITE, tree, page, sector,
						 iosize, page_offset, bdev,
						 end_bio_extent_writepage);
			if (ret)
				SetPageError(page);
		}
1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737
		cur = cur + iosize;
		page_offset += iosize;
		nr++;
	}
done:
	unlock_extent(tree, start, page_end, GFP_NOFS);
	unlock_page(page);
	return 0;
}
EXPORT_SYMBOL(extent_write_full_page);

/*
 * basic invalidatepage code, this waits on any locked or writeback
 * ranges corresponding to the page, and then deletes any extent state
 * records from the tree
 */
int extent_invalidatepage(struct extent_map_tree *tree,
			  struct page *page, unsigned long offset)
{
	u64 start = (page->index << PAGE_CACHE_SHIFT);
	u64 end = start + PAGE_CACHE_SIZE - 1;
	size_t blocksize = page->mapping->host->i_sb->s_blocksize;

	start += (offset + blocksize -1) & ~(blocksize - 1);
	if (start > end)
		return 0;

	lock_extent(tree, start, end, GFP_NOFS);
	wait_on_extent_writeback(tree, start, end);
C
Chris Mason 已提交
1738 1739
	clear_extent_bit(tree, start, end,
			 EXTENT_LOCKED | EXTENT_DIRTY | EXTENT_DELALLOC,
1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754
			 1, 1, GFP_NOFS);
	return 0;
}
EXPORT_SYMBOL(extent_invalidatepage);

/*
 * simple commit_write call, set_range_dirty is used to mark both
 * the pages and the extent records as dirty
 */
int extent_commit_write(struct extent_map_tree *tree,
			struct inode *inode, struct page *page,
			unsigned from, unsigned to)
{
	loff_t pos = ((loff_t)page->index << PAGE_CACHE_SHIFT) + to;

1755
	set_page_extent_mapped(page);
1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785
	set_page_dirty(page);

	if (pos > inode->i_size) {
		i_size_write(inode, pos);
		mark_inode_dirty(inode);
	}
	return 0;
}
EXPORT_SYMBOL(extent_commit_write);

int extent_prepare_write(struct extent_map_tree *tree,
			 struct inode *inode, struct page *page,
			 unsigned from, unsigned to, get_extent_t *get_extent)
{
	u64 page_start = page->index << PAGE_CACHE_SHIFT;
	u64 page_end = page_start + PAGE_CACHE_SIZE - 1;
	u64 block_start;
	u64 orig_block_start;
	u64 block_end;
	u64 cur_end;
	struct extent_map *em;
	unsigned blocksize = 1 << inode->i_blkbits;
	size_t page_offset = 0;
	size_t block_off_start;
	size_t block_off_end;
	int err = 0;
	int iocount = 0;
	int ret = 0;
	int isnew;

1786 1787
	set_page_extent_mapped(page);

1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870
	block_start = (page_start + from) & ~((u64)blocksize - 1);
	block_end = (page_start + to - 1) | (blocksize - 1);
	orig_block_start = block_start;

	lock_extent(tree, page_start, page_end, GFP_NOFS);
	while(block_start <= block_end) {
		em = get_extent(inode, page, page_offset, block_start,
				block_end, 1);
		if (IS_ERR(em) || !em) {
			goto err;
		}
		cur_end = min(block_end, em->end);
		block_off_start = block_start & (PAGE_CACHE_SIZE - 1);
		block_off_end = block_off_start + blocksize;
		isnew = clear_extent_new(tree, block_start, cur_end, GFP_NOFS);

		if (!PageUptodate(page) && isnew &&
		    (block_off_end > to || block_off_start < from)) {
			void *kaddr;

			kaddr = kmap_atomic(page, KM_USER0);
			if (block_off_end > to)
				memset(kaddr + to, 0, block_off_end - to);
			if (block_off_start < from)
				memset(kaddr + block_off_start, 0,
				       from - block_off_start);
			flush_dcache_page(page);
			kunmap_atomic(kaddr, KM_USER0);
		}
		if (!isnew && !PageUptodate(page) &&
		    (block_off_end > to || block_off_start < from) &&
		    !test_range_bit(tree, block_start, cur_end,
				    EXTENT_UPTODATE, 1)) {
			u64 sector;
			u64 extent_offset = block_start - em->start;
			size_t iosize;
			sector = (em->block_start + extent_offset) >> 9;
			iosize = (cur_end - block_start + blocksize - 1) &
				~((u64)blocksize - 1);
			/*
			 * we've already got the extent locked, but we
			 * need to split the state such that our end_bio
			 * handler can clear the lock.
			 */
			set_extent_bit(tree, block_start,
				       block_start + iosize - 1,
				       EXTENT_LOCKED, 0, NULL, GFP_NOFS);
			ret = submit_extent_page(READ, tree, page,
					 sector, iosize, page_offset, em->bdev,
					 end_bio_extent_preparewrite);
			iocount++;
			block_start = block_start + iosize;
		} else {
			set_extent_uptodate(tree, block_start, cur_end,
					    GFP_NOFS);
			unlock_extent(tree, block_start, cur_end, GFP_NOFS);
			block_start = cur_end + 1;
		}
		page_offset = block_start & (PAGE_CACHE_SIZE - 1);
		free_extent_map(em);
	}
	if (iocount) {
		wait_extent_bit(tree, orig_block_start,
				block_end, EXTENT_LOCKED);
	}
	check_page_uptodate(tree, page);
err:
	/* FIXME, zero out newly allocated blocks on error */
	return err;
}
EXPORT_SYMBOL(extent_prepare_write);

/*
 * a helper for releasepage.  As long as there are no locked extents
 * in the range corresponding to the page, both state records and extent
 * map records are removed
 */
int try_release_extent_mapping(struct extent_map_tree *tree, struct page *page)
{
	struct extent_map *em;
	u64 start = page->index << PAGE_CACHE_SHIFT;
	u64 end = start + PAGE_CACHE_SIZE - 1;
	u64 orig_start = start;
1871
	int ret = 1;
1872 1873 1874 1875 1876

	while (start <= end) {
		em = lookup_extent_mapping(tree, start, end);
		if (!em || IS_ERR(em))
			break;
1877 1878 1879 1880
		if (!test_range_bit(tree, em->start, em->end,
				    EXTENT_LOCKED, 0)) {
			remove_extent_mapping(tree, em);
			/* once for the rb tree */
1881 1882 1883 1884 1885 1886
			free_extent_map(em);
		}
		start = em->end + 1;
		/* once for us */
		free_extent_map(em);
	}
1887 1888 1889 1890 1891 1892
	if (test_range_bit(tree, orig_start, end, EXTENT_LOCKED, 0))
		ret = 0;
	else
		clear_extent_bit(tree, orig_start, end, EXTENT_UPTODATE,
				 1, 1, GFP_NOFS);
	return ret;
1893 1894 1895
}
EXPORT_SYMBOL(try_release_extent_mapping);

1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908
sector_t extent_bmap(struct address_space *mapping, sector_t iblock,
		get_extent_t *get_extent)
{
	struct inode *inode = mapping->host;
	u64 start = iblock << inode->i_blkbits;
	u64 end = start + (1 << inode->i_blkbits) - 1;
	struct extent_map *em;

	em = get_extent(inode, NULL, 0, start, end, 0);
	if (!em || IS_ERR(em))
		return 0;

	if (em->block_start == EXTENT_MAP_INLINE ||
1909 1910
	    em->block_start == EXTENT_MAP_HOLE)
		return 0;
1911 1912 1913

	return (em->block_start + start - em->start) >> inode->i_blkbits;
}
1914

1915
static int add_lru(struct extent_map_tree *tree, struct extent_buffer *eb)
1916
{
1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938
	if (list_empty(&eb->lru)) {
		extent_buffer_get(eb);
		list_add(&eb->lru, &tree->buffer_lru);
		tree->lru_size++;
		if (tree->lru_size >= BUFFER_LRU_MAX) {
			struct extent_buffer *rm;
			rm = list_entry(tree->buffer_lru.prev,
					struct extent_buffer, lru);
			tree->lru_size--;
			list_del(&rm->lru);
			free_extent_buffer(rm);
		}
	} else
		list_move(&eb->lru, &tree->buffer_lru);
	return 0;
}
static struct extent_buffer *find_lru(struct extent_map_tree *tree,
				      u64 start, unsigned long len)
{
	struct list_head *lru = &tree->buffer_lru;
	struct list_head *cur = lru->next;
	struct extent_buffer *eb;
1939

1940 1941
	if (list_empty(lru))
		return NULL;
1942

1943 1944 1945 1946 1947 1948 1949 1950 1951
	do {
		eb = list_entry(cur, struct extent_buffer, lru);
		if (eb->start == start && eb->len == len) {
			extent_buffer_get(eb);
			return eb;
		}
		cur = cur->next;
	} while (cur != lru);
	return NULL;
1952 1953
}

1954
static inline unsigned long num_extent_pages(u64 start, u64 len)
1955
{
1956 1957
	return ((start + len + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT) -
		(start >> PAGE_CACHE_SHIFT);
1958 1959
}

1960 1961
static inline struct page *extent_buffer_page(struct extent_buffer *eb,
					      unsigned long i)
1962 1963
{
	struct page *p;
1964

1965
	if (i == 0)
1966
		return eb->first_page;
1967
	i += eb->start >> PAGE_CACHE_SHIFT;
1968
	p = find_get_page(eb->first_page->mapping, i);
1969 1970 1971 1972
	page_cache_release(p);
	return p;
}

1973 1974 1975 1976
static struct extent_buffer *__alloc_extent_buffer(struct extent_map_tree *tree,
						   u64 start,
						   unsigned long len,
						   gfp_t mask)
1977
{
1978 1979 1980 1981
	struct extent_buffer *eb = NULL;

	spin_lock(&tree->lru_lock);
	eb = find_lru(tree, start, len);
1982
	if (eb) {
1983
		goto lru_add;
1984
	}
1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
	spin_unlock(&tree->lru_lock);

	if (eb) {
		memset(eb, 0, sizeof(*eb));
	} else {
		eb = kmem_cache_zalloc(extent_buffer_cache, mask);
	}
	INIT_LIST_HEAD(&eb->lru);
	eb->start = start;
	eb->len = len;
	atomic_set(&eb->refs, 1);

	spin_lock(&tree->lru_lock);
lru_add:
	add_lru(tree, eb);
	spin_unlock(&tree->lru_lock);
	return eb;
}

static void __free_extent_buffer(struct extent_buffer *eb)
{
	kmem_cache_free(extent_buffer_cache, eb);
2007
}
2008

2009 2010
struct extent_buffer *alloc_extent_buffer(struct extent_map_tree *tree,
					  u64 start, unsigned long len,
2011
					  struct page *page0,
2012 2013
					  gfp_t mask)
{
2014
	unsigned long num_pages = num_extent_pages(start, len);
2015 2016 2017 2018 2019 2020 2021
	unsigned long i;
	unsigned long index = start >> PAGE_CACHE_SHIFT;
	struct extent_buffer *eb;
	struct page *p;
	struct address_space *mapping = tree->mapping;
	int uptodate = 0;

2022
	eb = __alloc_extent_buffer(tree, start, len, mask);
2023 2024 2025
	if (!eb || IS_ERR(eb))
		return NULL;

2026 2027
	if (eb->flags & EXTENT_BUFFER_FILLED)
		return eb;
2028

2029 2030 2031 2032 2033
	if (page0) {
		eb->first_page = page0;
		i = 1;
		index++;
		page_cache_get(page0);
2034
		mark_page_accessed(page0);
2035 2036 2037 2038 2039 2040 2041
		set_page_extent_mapped(page0);
		set_page_private(page0, EXTENT_PAGE_PRIVATE_FIRST_PAGE |
				 len << 2);
	} else {
		i = 0;
	}
	for (; i < num_pages; i++, index++) {
2042
		p = find_or_create_page(mapping, index, mask | __GFP_HIGHMEM);
2043
		if (!p) {
2044
			WARN_ON(1);
2045 2046 2047 2048 2049
			/* make sure the free only frees the pages we've
			 * grabbed a reference on
			 */
			eb->len = i << PAGE_CACHE_SHIFT;
			eb->start &= ~((u64)PAGE_CACHE_SIZE - 1);
2050
			goto fail;
2051
		}
2052
		set_page_extent_mapped(p);
2053
		mark_page_accessed(p);
2054
		if (i == 0) {
2055
			eb->first_page = p;
2056 2057 2058 2059 2060
			set_page_private(p, EXTENT_PAGE_PRIVATE_FIRST_PAGE |
					 len << 2);
		} else {
			set_page_private(p, EXTENT_PAGE_PRIVATE);
		}
2061 2062 2063 2064 2065 2066
		if (!PageUptodate(p))
			uptodate = 0;
		unlock_page(p);
	}
	if (uptodate)
		eb->flags |= EXTENT_UPTODATE;
2067
	eb->flags |= EXTENT_BUFFER_FILLED;
2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078
	return eb;
fail:
	free_extent_buffer(eb);
	return NULL;
}
EXPORT_SYMBOL(alloc_extent_buffer);

struct extent_buffer *find_extent_buffer(struct extent_map_tree *tree,
					 u64 start, unsigned long len,
					  gfp_t mask)
{
2079
	unsigned long num_pages = num_extent_pages(start, len);
2080
	unsigned long i; unsigned long index = start >> PAGE_CACHE_SHIFT;
2081 2082 2083
	struct extent_buffer *eb;
	struct page *p;
	struct address_space *mapping = tree->mapping;
2084
	int uptodate = 1;
2085

2086
	eb = __alloc_extent_buffer(tree, start, len, mask);
2087 2088 2089
	if (!eb || IS_ERR(eb))
		return NULL;

2090 2091
	if (eb->flags & EXTENT_BUFFER_FILLED)
		return eb;
2092 2093

	for (i = 0; i < num_pages; i++, index++) {
2094
		p = find_lock_page(mapping, index);
2095 2096 2097 2098 2099 2100
		if (!p) {
			/* make sure the free only frees the pages we've
			 * grabbed a reference on
			 */
			eb->len = i << PAGE_CACHE_SHIFT;
			eb->start &= ~((u64)PAGE_CACHE_SIZE - 1);
2101
			goto fail;
2102
		}
2103
		set_page_extent_mapped(p);
2104
		mark_page_accessed(p);
2105 2106

		if (i == 0) {
2107
			eb->first_page = p;
2108 2109 2110 2111 2112 2113
			set_page_private(p, EXTENT_PAGE_PRIVATE_FIRST_PAGE |
					 len << 2);
		} else {
			set_page_private(p, EXTENT_PAGE_PRIVATE);
		}

2114 2115 2116
		if (!PageUptodate(p))
			uptodate = 0;
		unlock_page(p);
2117
	}
2118 2119
	if (uptodate)
		eb->flags |= EXTENT_UPTODATE;
2120
	eb->flags |= EXTENT_BUFFER_FILLED;
2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138
	return eb;
fail:
	free_extent_buffer(eb);
	return NULL;
}
EXPORT_SYMBOL(find_extent_buffer);

void free_extent_buffer(struct extent_buffer *eb)
{
	unsigned long i;
	unsigned long num_pages;

	if (!eb)
		return;

	if (!atomic_dec_and_test(&eb->refs))
		return;

2139
	num_pages = num_extent_pages(eb->start, eb->len);
2140

2141
	for (i = 0; i < num_pages; i++) {
2142
		page_cache_release(extent_buffer_page(eb, i));
2143
	}
2144
	__free_extent_buffer(eb);
2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159
}
EXPORT_SYMBOL(free_extent_buffer);

int clear_extent_buffer_dirty(struct extent_map_tree *tree,
			      struct extent_buffer *eb)
{
	int set;
	unsigned long i;
	unsigned long num_pages;
	struct page *page;

	u64 start = eb->start;
	u64 end = start + eb->len - 1;

	set = clear_extent_dirty(tree, start, end, GFP_NOFS);
2160
	num_pages = num_extent_pages(eb->start, eb->len);
2161 2162

	for (i = 0; i < num_pages; i++) {
2163
		page = extent_buffer_page(eb, i);
2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198
		lock_page(page);
		/*
		 * if we're on the last page or the first page and the
		 * block isn't aligned on a page boundary, do extra checks
		 * to make sure we don't clean page that is partially dirty
		 */
		if ((i == 0 && (eb->start & (PAGE_CACHE_SIZE - 1))) ||
		    ((i == num_pages - 1) &&
		     ((eb->start + eb->len - 1) & (PAGE_CACHE_SIZE - 1)))) {
			start = page->index << PAGE_CACHE_SHIFT;
			end  = start + PAGE_CACHE_SIZE - 1;
			if (test_range_bit(tree, start, end,
					   EXTENT_DIRTY, 0)) {
				unlock_page(page);
				continue;
			}
		}
		clear_page_dirty_for_io(page);
		unlock_page(page);
	}
	return 0;
}
EXPORT_SYMBOL(clear_extent_buffer_dirty);

int wait_on_extent_buffer_writeback(struct extent_map_tree *tree,
				    struct extent_buffer *eb)
{
	return wait_on_extent_writeback(tree, eb->start,
					eb->start + eb->len - 1);
}
EXPORT_SYMBOL(wait_on_extent_buffer_writeback);

int set_extent_buffer_dirty(struct extent_map_tree *tree,
			     struct extent_buffer *eb)
{
2199 2200 2201 2202 2203
	unsigned long i;
	unsigned long num_pages;

	num_pages = num_extent_pages(eb->start, eb->len);
	for (i = 0; i < num_pages; i++) {
2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215
		struct page *page = extent_buffer_page(eb, i);
		/* writepage may need to do something special for the
		 * first page, we have to make sure page->private is
		 * properly set.  releasepage may drop page->private
		 * on us if the page isn't already dirty.
		 */
		if (i == 0) {
			lock_page(page);
			set_page_private(page,
					 EXTENT_PAGE_PRIVATE_FIRST_PAGE |
					 eb->len << 2);
		}
2216
		__set_page_dirty_nobuffers(extent_buffer_page(eb, i));
2217 2218
		if (i == 0)
			unlock_page(page);
2219 2220 2221
	}
	return set_extent_dirty(tree, eb->start,
				eb->start + eb->len - 1, GFP_NOFS);
2222 2223 2224 2225 2226 2227 2228 2229 2230 2231
}
EXPORT_SYMBOL(set_extent_buffer_dirty);

int set_extent_buffer_uptodate(struct extent_map_tree *tree,
				struct extent_buffer *eb)
{
	unsigned long i;
	struct page *page;
	unsigned long num_pages;

2232
	num_pages = num_extent_pages(eb->start, eb->len);
2233 2234 2235 2236

	set_extent_uptodate(tree, eb->start, eb->start + eb->len - 1,
			    GFP_NOFS);
	for (i = 0; i < num_pages; i++) {
2237
		page = extent_buffer_page(eb, i);
2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260
		if ((i == 0 && (eb->start & (PAGE_CACHE_SIZE - 1))) ||
		    ((i == num_pages - 1) &&
		     ((eb->start + eb->len - 1) & (PAGE_CACHE_SIZE - 1)))) {
			check_page_uptodate(tree, page);
			continue;
		}
		SetPageUptodate(page);
	}
	return 0;
}
EXPORT_SYMBOL(set_extent_buffer_uptodate);

int extent_buffer_uptodate(struct extent_map_tree *tree,
			     struct extent_buffer *eb)
{
	if (eb->flags & EXTENT_UPTODATE)
		return 1;
	return test_range_bit(tree, eb->start, eb->start + eb->len - 1,
			   EXTENT_UPTODATE, 1);
}
EXPORT_SYMBOL(extent_buffer_uptodate);

int read_extent_buffer_pages(struct extent_map_tree *tree,
2261 2262 2263
			     struct extent_buffer *eb,
			     u64 start,
			     int wait)
2264 2265
{
	unsigned long i;
2266
	unsigned long start_i;
2267 2268 2269 2270 2271 2272 2273 2274
	struct page *page;
	int err;
	int ret = 0;
	unsigned long num_pages;

	if (eb->flags & EXTENT_UPTODATE)
		return 0;

2275
	if (0 && test_range_bit(tree, eb->start, eb->start + eb->len - 1,
2276 2277 2278
			   EXTENT_UPTODATE, 1)) {
		return 0;
	}
2279 2280 2281 2282 2283 2284 2285
	if (start) {
		WARN_ON(start < eb->start);
		start_i = (start >> PAGE_CACHE_SHIFT) -
			(eb->start >> PAGE_CACHE_SHIFT);
	} else {
		start_i = 0;
	}
2286

2287
	num_pages = num_extent_pages(eb->start, eb->len);
2288
	for (i = start_i; i < num_pages; i++) {
2289
		page = extent_buffer_page(eb, i);
2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313
		if (PageUptodate(page)) {
			continue;
		}
		if (!wait) {
			if (TestSetPageLocked(page)) {
				continue;
			}
		} else {
			lock_page(page);
		}
		if (!PageUptodate(page)) {
			err = page->mapping->a_ops->readpage(NULL, page);
			if (err) {
				ret = err;
			}
		} else {
			unlock_page(page);
		}
	}

	if (ret || !wait) {
		return ret;
	}

2314
	for (i = start_i; i < num_pages; i++) {
2315
		page = extent_buffer_page(eb, i);
2316 2317 2318 2319 2320
		wait_on_page_locked(page);
		if (!PageUptodate(page)) {
			ret = -EIO;
		}
	}
2321 2322
	if (!ret)
		eb->flags |= EXTENT_UPTODATE;
2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337
	return ret;
}
EXPORT_SYMBOL(read_extent_buffer_pages);

void read_extent_buffer(struct extent_buffer *eb, void *dstv,
			unsigned long start,
			unsigned long len)
{
	size_t cur;
	size_t offset;
	struct page *page;
	char *kaddr;
	char *dst = (char *)dstv;
	size_t start_offset = eb->start & ((u64)PAGE_CACHE_SIZE - 1);
	unsigned long i = (start_offset + start) >> PAGE_CACHE_SHIFT;
2338
	unsigned long num_pages = num_extent_pages(eb->start, eb->len);
2339 2340 2341 2342 2343 2344 2345 2346 2347

	WARN_ON(start > eb->len);
	WARN_ON(start + len > eb->start + eb->len);

	offset = start & ((unsigned long)PAGE_CACHE_SIZE - 1);
	if (i == 0)
		offset += start_offset;

	while(len > 0) {
2348
		page = extent_buffer_page(eb, i);
2349 2350 2351 2352
		if (!PageUptodate(page)) {
			printk("page %lu not up to date i %lu, total %lu, len %lu\n", page->index, i, num_pages, eb->len);
			WARN_ON(1);
		}
2353 2354 2355
		WARN_ON(!PageUptodate(page));

		cur = min(len, (PAGE_CACHE_SIZE - offset));
2356
		kaddr = kmap_atomic(page, KM_USER0);
2357
		memcpy(dst, kaddr + offset, cur);
2358
		kunmap_atomic(kaddr, KM_USER0);
2359 2360 2361 2362 2363 2364 2365 2366 2367

		dst += cur;
		len -= cur;
		offset = 0;
		i++;
	}
}
EXPORT_SYMBOL(read_extent_buffer);

2368
int map_private_extent_buffer(struct extent_buffer *eb, unsigned long start,
2369 2370 2371
			       unsigned long min_len, char **token, char **map,
			       unsigned long *map_start,
			       unsigned long *map_len, int km)
2372
{
2373
	size_t offset = start & (PAGE_CACHE_SIZE - 1);
2374
	char *kaddr;
2375
	struct page *p;
2376 2377
	size_t start_offset = eb->start & ((u64)PAGE_CACHE_SIZE - 1);
	unsigned long i = (start_offset + start) >> PAGE_CACHE_SHIFT;
2378
	unsigned long end_i = (start_offset + start + min_len) >>
2379
		PAGE_CACHE_SHIFT;
2380 2381 2382

	if (i != end_i)
		return -EINVAL;
2383 2384 2385 2386 2387

	if (i == 0) {
		offset = start_offset;
		*map_start = 0;
	} else {
2388
		offset = 0;
2389
		*map_start = (i << PAGE_CACHE_SHIFT) - start_offset;
2390
	}
2391 2392 2393 2394
	if (start + min_len >= eb->len) {
printk("bad mapping eb start %Lu len %lu, wanted %lu %lu\n", eb->start, eb->len, start, min_len);
		WARN_ON(1);
	}
2395

2396 2397 2398
	p = extent_buffer_page(eb, i);
	WARN_ON(!PageUptodate(p));
	kaddr = kmap_atomic(p, km);
2399 2400 2401 2402 2403
	*token = kaddr;
	*map = kaddr + offset;
	*map_len = PAGE_CACHE_SIZE - offset;
	return 0;
}
2404
EXPORT_SYMBOL(map_private_extent_buffer);
2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418

int map_extent_buffer(struct extent_buffer *eb, unsigned long start,
		      unsigned long min_len,
		      char **token, char **map,
		      unsigned long *map_start,
		      unsigned long *map_len, int km)
{
	int err;
	int save = 0;
	if (eb->map_token) {
		unmap_extent_buffer(eb, eb->map_token, km);
		eb->map_token = NULL;
		save = 1;
	}
2419 2420
	err = map_private_extent_buffer(eb, start, min_len, token, map,
				       map_start, map_len, km);
2421 2422 2423 2424 2425 2426 2427 2428
	if (!err && save) {
		eb->map_token = *token;
		eb->kaddr = *map;
		eb->map_start = *map_start;
		eb->map_len = *map_len;
	}
	return err;
}
2429 2430 2431 2432
EXPORT_SYMBOL(map_extent_buffer);

void unmap_extent_buffer(struct extent_buffer *eb, char *token, int km)
{
2433
	kunmap_atomic(token, km);
2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457
}
EXPORT_SYMBOL(unmap_extent_buffer);

int memcmp_extent_buffer(struct extent_buffer *eb, const void *ptrv,
			  unsigned long start,
			  unsigned long len)
{
	size_t cur;
	size_t offset;
	struct page *page;
	char *kaddr;
	char *ptr = (char *)ptrv;
	size_t start_offset = eb->start & ((u64)PAGE_CACHE_SIZE - 1);
	unsigned long i = (start_offset + start) >> PAGE_CACHE_SHIFT;
	int ret = 0;

	WARN_ON(start > eb->len);
	WARN_ON(start + len > eb->start + eb->len);

	offset = start & ((unsigned long)PAGE_CACHE_SIZE - 1);
	if (i == 0)
		offset += start_offset;

	while(len > 0) {
2458
		page = extent_buffer_page(eb, i);
2459 2460 2461 2462
		WARN_ON(!PageUptodate(page));

		cur = min(len, (PAGE_CACHE_SIZE - offset));

2463
		kaddr = kmap_atomic(page, KM_USER0);
2464
		ret = memcmp(ptr, kaddr + offset, cur);
2465
		kunmap_atomic(kaddr, KM_USER0);
2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496
		if (ret)
			break;

		ptr += cur;
		len -= cur;
		offset = 0;
		i++;
	}
	return ret;
}
EXPORT_SYMBOL(memcmp_extent_buffer);

void write_extent_buffer(struct extent_buffer *eb, const void *srcv,
			 unsigned long start, unsigned long len)
{
	size_t cur;
	size_t offset;
	struct page *page;
	char *kaddr;
	char *src = (char *)srcv;
	size_t start_offset = eb->start & ((u64)PAGE_CACHE_SIZE - 1);
	unsigned long i = (start_offset + start) >> PAGE_CACHE_SHIFT;

	WARN_ON(start > eb->len);
	WARN_ON(start + len > eb->start + eb->len);

	offset = start & ((unsigned long)PAGE_CACHE_SIZE - 1);
	if (i == 0)
		offset += start_offset;

	while(len > 0) {
2497
		page = extent_buffer_page(eb, i);
2498 2499 2500
		WARN_ON(!PageUptodate(page));

		cur = min(len, PAGE_CACHE_SIZE - offset);
2501
		kaddr = kmap_atomic(page, KM_USER0);
2502
		memcpy(kaddr + offset, src, cur);
2503
		kunmap_atomic(kaddr, KM_USER0);
2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530

		src += cur;
		len -= cur;
		offset = 0;
		i++;
	}
}
EXPORT_SYMBOL(write_extent_buffer);

void memset_extent_buffer(struct extent_buffer *eb, char c,
			  unsigned long start, unsigned long len)
{
	size_t cur;
	size_t offset;
	struct page *page;
	char *kaddr;
	size_t start_offset = eb->start & ((u64)PAGE_CACHE_SIZE - 1);
	unsigned long i = (start_offset + start) >> PAGE_CACHE_SHIFT;

	WARN_ON(start > eb->len);
	WARN_ON(start + len > eb->start + eb->len);

	offset = start & ((unsigned long)PAGE_CACHE_SIZE - 1);
	if (i == 0)
		offset += start_offset;

	while(len > 0) {
2531
		page = extent_buffer_page(eb, i);
2532 2533 2534
		WARN_ON(!PageUptodate(page));

		cur = min(len, PAGE_CACHE_SIZE - offset);
2535
		kaddr = kmap_atomic(page, KM_USER0);
2536
		memset(kaddr + offset, c, cur);
2537
		kunmap_atomic(kaddr, KM_USER0);
2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564

		len -= cur;
		offset = 0;
		i++;
	}
}
EXPORT_SYMBOL(memset_extent_buffer);

void copy_extent_buffer(struct extent_buffer *dst, struct extent_buffer *src,
			unsigned long dst_offset, unsigned long src_offset,
			unsigned long len)
{
	u64 dst_len = dst->len;
	size_t cur;
	size_t offset;
	struct page *page;
	char *kaddr;
	size_t start_offset = dst->start & ((u64)PAGE_CACHE_SIZE - 1);
	unsigned long i = (start_offset + dst_offset) >> PAGE_CACHE_SHIFT;

	WARN_ON(src->len != dst_len);

	offset = dst_offset & ((unsigned long)PAGE_CACHE_SIZE - 1);
	if (i == 0)
		offset += start_offset;

	while(len > 0) {
2565
		page = extent_buffer_page(dst, i);
2566 2567 2568 2569
		WARN_ON(!PageUptodate(page));

		cur = min(len, (unsigned long)(PAGE_CACHE_SIZE - offset));

2570
		kaddr = kmap_atomic(page, KM_USER1);
2571
		read_extent_buffer(src, kaddr + offset, src_offset, cur);
2572
		kunmap_atomic(kaddr, KM_USER1);
2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585

		src_offset += cur;
		len -= cur;
		offset = 0;
		i++;
	}
}
EXPORT_SYMBOL(copy_extent_buffer);

static void move_pages(struct page *dst_page, struct page *src_page,
		       unsigned long dst_off, unsigned long src_off,
		       unsigned long len)
{
2586
	char *dst_kaddr = kmap_atomic(dst_page, KM_USER0);
2587 2588 2589
	if (dst_page == src_page) {
		memmove(dst_kaddr + dst_off, dst_kaddr + src_off, len);
	} else {
2590
		char *src_kaddr = kmap_atomic(src_page, KM_USER1);
2591 2592 2593 2594 2595 2596
		char *p = dst_kaddr + dst_off + len;
		char *s = src_kaddr + src_off + len;

		while (len--)
			*--p = *--s;

2597
		kunmap_atomic(src_kaddr, KM_USER1);
2598
	}
2599
	kunmap_atomic(dst_kaddr, KM_USER0);
2600 2601 2602 2603 2604 2605
}

static void copy_pages(struct page *dst_page, struct page *src_page,
		       unsigned long dst_off, unsigned long src_off,
		       unsigned long len)
{
2606
	char *dst_kaddr = kmap_atomic(dst_page, KM_USER0);
2607 2608 2609
	char *src_kaddr;

	if (dst_page != src_page)
2610
		src_kaddr = kmap_atomic(src_page, KM_USER1);
2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656
	else
		src_kaddr = dst_kaddr;

	memcpy(dst_kaddr + dst_off, src_kaddr + src_off, len);
	kunmap_atomic(dst_kaddr, KM_USER0);
	if (dst_page != src_page)
		kunmap_atomic(src_kaddr, KM_USER1);
}

void memcpy_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset,
			   unsigned long src_offset, unsigned long len)
{
	size_t cur;
	size_t dst_off_in_page;
	size_t src_off_in_page;
	size_t start_offset = dst->start & ((u64)PAGE_CACHE_SIZE - 1);
	unsigned long dst_i;
	unsigned long src_i;

	if (src_offset + len > dst->len) {
		printk("memmove bogus src_offset %lu move len %lu len %lu\n",
		       src_offset, len, dst->len);
		BUG_ON(1);
	}
	if (dst_offset + len > dst->len) {
		printk("memmove bogus dst_offset %lu move len %lu len %lu\n",
		       dst_offset, len, dst->len);
		BUG_ON(1);
	}

	while(len > 0) {
		dst_off_in_page = dst_offset &
			((unsigned long)PAGE_CACHE_SIZE - 1);
		src_off_in_page = src_offset &
			((unsigned long)PAGE_CACHE_SIZE - 1);

		dst_i = (start_offset + dst_offset) >> PAGE_CACHE_SHIFT;
		src_i = (start_offset + src_offset) >> PAGE_CACHE_SHIFT;

		if (src_i == 0)
			src_off_in_page += start_offset;
		if (dst_i == 0)
			dst_off_in_page += start_offset;

		cur = min(len, (unsigned long)(PAGE_CACHE_SIZE -
					       src_off_in_page));
J
Jens Axboe 已提交
2657 2658
		cur = min_t(unsigned long, cur,
			(unsigned long)(PAGE_CACHE_SIZE - dst_off_in_page));
2659

2660 2661
		copy_pages(extent_buffer_page(dst, dst_i),
			   extent_buffer_page(dst, src_i),
2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709
			   dst_off_in_page, src_off_in_page, cur);

		src_offset += cur;
		dst_offset += cur;
		len -= cur;
	}
}
EXPORT_SYMBOL(memcpy_extent_buffer);

void memmove_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset,
			   unsigned long src_offset, unsigned long len)
{
	size_t cur;
	size_t dst_off_in_page;
	size_t src_off_in_page;
	unsigned long dst_end = dst_offset + len - 1;
	unsigned long src_end = src_offset + len - 1;
	size_t start_offset = dst->start & ((u64)PAGE_CACHE_SIZE - 1);
	unsigned long dst_i;
	unsigned long src_i;

	if (src_offset + len > dst->len) {
		printk("memmove bogus src_offset %lu move len %lu len %lu\n",
		       src_offset, len, dst->len);
		BUG_ON(1);
	}
	if (dst_offset + len > dst->len) {
		printk("memmove bogus dst_offset %lu move len %lu len %lu\n",
		       dst_offset, len, dst->len);
		BUG_ON(1);
	}
	if (dst_offset < src_offset) {
		memcpy_extent_buffer(dst, dst_offset, src_offset, len);
		return;
	}
	while(len > 0) {
		dst_i = (start_offset + dst_end) >> PAGE_CACHE_SHIFT;
		src_i = (start_offset + src_end) >> PAGE_CACHE_SHIFT;

		dst_off_in_page = dst_end &
			((unsigned long)PAGE_CACHE_SIZE - 1);
		src_off_in_page = src_end &
			((unsigned long)PAGE_CACHE_SIZE - 1);
		if (src_i == 0)
			src_off_in_page += start_offset;
		if (dst_i == 0)
			dst_off_in_page += start_offset;

J
Jens Axboe 已提交
2710
		cur = min_t(unsigned long, len, src_off_in_page + 1);
2711
		cur = min(cur, dst_off_in_page + 1);
2712 2713
		move_pages(extent_buffer_page(dst, dst_i),
			   extent_buffer_page(dst, src_i),
2714 2715 2716
			   dst_off_in_page - cur + 1,
			   src_off_in_page - cur + 1, cur);

2717 2718
		dst_end -= cur;
		src_end -= cur;
2719 2720 2721 2722
		len -= cur;
	}
}
EXPORT_SYMBOL(memmove_extent_buffer);