extent_io.c 86.3 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
#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>
#include <linux/swap.h>
#include <linux/version.h>
#include <linux/writeback.h>
#include <linux/pagevec.h>
#include "extent_io.h"
#include "extent_map.h"
17
#include "compat.h"
18 19 20 21 22 23 24 25 26 27 28 29

/* 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));

static struct kmem_cache *extent_state_cache;
static struct kmem_cache *extent_buffer_cache;

static LIST_HEAD(buffers);
static LIST_HEAD(states);
30
static spinlock_t leak_lock = SPIN_LOCK_UNLOCKED;
31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68

#define BUFFER_LRU_MAX 64

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

struct extent_page_data {
	struct bio *bio;
	struct extent_io_tree *tree;
	get_extent_t *get_extent;
};

int __init extent_io_init(void)
{
	extent_state_cache = btrfs_cache_create("extent_state",
					    sizeof(struct extent_state), 0,
					    NULL);
	if (!extent_state_cache)
		return -ENOMEM;

	extent_buffer_cache = btrfs_cache_create("extent_buffers",
					    sizeof(struct extent_buffer), 0,
					    NULL);
	if (!extent_buffer_cache)
		goto free_state_cache;
	return 0;

free_state_cache:
	kmem_cache_destroy(extent_state_cache);
	return -ENOMEM;
}

void extent_io_exit(void)
{
	struct extent_state *state;
69
	struct extent_buffer *eb;
70 71

	while (!list_empty(&states)) {
72
		state = list_entry(states.next, struct extent_state, leak_list);
73
		printk("state leak: start %Lu end %Lu state %lu in tree %p refs %d\n", state->start, state->end, state->state, state->tree, atomic_read(&state->refs));
74
		list_del(&state->leak_list);
75 76 77 78
		kmem_cache_free(extent_state_cache, state);

	}

79 80 81 82 83 84
	while (!list_empty(&buffers)) {
		eb = list_entry(buffers.next, struct extent_buffer, leak_list);
		printk("buffer leak start %Lu len %lu refs %d\n", eb->start, eb->len, atomic_read(&eb->refs));
		list_del(&eb->leak_list);
		kmem_cache_free(extent_buffer_cache, eb);
	}
85 86 87 88 89 90 91 92 93 94
	if (extent_state_cache)
		kmem_cache_destroy(extent_state_cache);
	if (extent_buffer_cache)
		kmem_cache_destroy(extent_buffer_cache);
}

void extent_io_tree_init(struct extent_io_tree *tree,
			  struct address_space *mapping, gfp_t mask)
{
	tree->state.rb_node = NULL;
95
	tree->buffer.rb_node = NULL;
96 97
	tree->ops = NULL;
	tree->dirty_bytes = 0;
98
	spin_lock_init(&tree->lock);
99
	spin_lock_init(&tree->buffer_lock);
100 101 102 103 104 105 106
	tree->mapping = mapping;
}
EXPORT_SYMBOL(extent_io_tree_init);

struct extent_state *alloc_extent_state(gfp_t mask)
{
	struct extent_state *state;
107
	unsigned long flags;
108 109

	state = kmem_cache_alloc(extent_state_cache, mask);
110
	if (!state)
111 112 113
		return state;
	state->state = 0;
	state->private = 0;
114
	state->tree = NULL;
115 116 117
	spin_lock_irqsave(&leak_lock, flags);
	list_add(&state->leak_list, &states);
	spin_unlock_irqrestore(&leak_lock, flags);
118 119 120 121 122 123 124 125 126 127 128 129

	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)
{
	if (!state)
		return;
	if (atomic_dec_and_test(&state->refs)) {
130
		unsigned long flags;
131
		WARN_ON(state->tree);
132 133 134
		spin_lock_irqsave(&leak_lock, flags);
		list_del(&state->leak_list);
		spin_unlock_irqrestore(&leak_lock, flags);
135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164
		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);
	rb_link_node(node, parent, p);
	rb_insert_color(node, root);
	return NULL;
}

165
static struct rb_node *__etree_search(struct extent_io_tree *tree, u64 offset,
166 167 168
				     struct rb_node **prev_ret,
				     struct rb_node **next_ret)
{
169
	struct rb_root *root = &tree->state;
170 171 172 173 174 175 176 177 178 179 180 181 182 183 184
	struct rb_node * n = root->rb_node;
	struct rb_node *prev = NULL;
	struct rb_node *orig_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;
185
		else {
186
			return n;
187
		}
188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210
	}

	if (prev_ret) {
		orig_prev = prev;
		while(prev && offset > prev_entry->end) {
			prev = rb_next(prev);
			prev_entry = rb_entry(prev, struct tree_entry, rb_node);
		}
		*prev_ret = prev;
		prev = orig_prev;
	}

	if (next_ret) {
		prev_entry = rb_entry(prev, struct tree_entry, rb_node);
		while(prev && offset < prev_entry->start) {
			prev = rb_prev(prev);
			prev_entry = rb_entry(prev, struct tree_entry, rb_node);
		}
		*next_ret = prev;
	}
	return NULL;
}

211 212
static inline struct rb_node *tree_search(struct extent_io_tree *tree,
					  u64 offset)
213
{
214
	struct rb_node *prev = NULL;
215
	struct rb_node *ret;
216

217 218
	ret = __etree_search(tree, offset, &prev, NULL);
	if (!ret) {
219
		return prev;
220
	}
221 222 223
	return ret;
}

224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267
static struct extent_buffer *buffer_tree_insert(struct extent_io_tree *tree,
					  u64 offset, struct rb_node *node)
{
	struct rb_root *root = &tree->buffer;
	struct rb_node ** p = &root->rb_node;
	struct rb_node * parent = NULL;
	struct extent_buffer *eb;

	while(*p) {
		parent = *p;
		eb = rb_entry(parent, struct extent_buffer, rb_node);

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

	rb_link_node(node, parent, p);
	rb_insert_color(node, root);
	return NULL;
}

static struct extent_buffer *buffer_search(struct extent_io_tree *tree,
					   u64 offset)
{
	struct rb_root *root = &tree->buffer;
	struct rb_node * n = root->rb_node;
	struct extent_buffer *eb;

	while(n) {
		eb = rb_entry(n, struct extent_buffer, rb_node);
		if (offset < eb->start)
			n = n->rb_left;
		else if (offset > eb->start)
			n = n->rb_right;
		else
			return eb;
	}
	return NULL;
}

268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291
/*
 * 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_io_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;
292
			other->tree = NULL;
293 294 295 296 297 298 299 300 301 302
			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;
303
			state->tree = NULL;
304 305 306 307 308 309 310
			rb_erase(&state->rb_node, &tree->state);
			free_extent_state(state);
		}
	}
	return 0;
}

311 312 313 314 315 316
static void set_state_cb(struct extent_io_tree *tree,
			 struct extent_state *state,
			 unsigned long bits)
{
	if (tree->ops && tree->ops->set_bit_hook) {
		tree->ops->set_bit_hook(tree->mapping->host, state->start,
C
Chris Mason 已提交
317
					state->end, state->state, bits);
318 319 320 321 322 323 324 325 326
	}
}

static void clear_state_cb(struct extent_io_tree *tree,
			   struct extent_state *state,
			   unsigned long bits)
{
	if (tree->ops && tree->ops->set_bit_hook) {
		tree->ops->clear_bit_hook(tree->mapping->host, state->start,
C
Chris Mason 已提交
327
					  state->end, state->state, bits);
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
/*
 * 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_io_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);
	}
	if (bits & EXTENT_DIRTY)
		tree->dirty_bytes += end - start + 1;
C
Chris Mason 已提交
353
	set_state_cb(tree, state, bits);
354 355 356 357 358 359 360 361 362 363 364
	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);
		printk("found node %Lu %Lu on insert of %Lu %Lu\n", found->start, found->end, start, end);
		free_extent_state(state);
		return -EEXIST;
	}
365
	state->tree = tree;
366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400
	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_io_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;

	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);
		printk("found node %Lu %Lu on insert of %Lu %Lu\n", found->start, found->end, prealloc->start, prealloc->end);
		free_extent_state(prealloc);
		return -EEXIST;
	}
401
	prealloc->tree = tree;
402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423
	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_io_tree *tree,
			    struct extent_state *state, int bits, int wake,
			    int delete)
{
	int ret = state->state & bits;

	if ((bits & EXTENT_DIRTY) && (state->state & EXTENT_DIRTY)) {
		u64 range = state->end - state->start + 1;
		WARN_ON(range > tree->dirty_bytes);
		tree->dirty_bytes -= range;
	}
424
	clear_state_cb(tree, state, bits);
C
Chris Mason 已提交
425
	state->state &= ~bits;
426 427 428
	if (wake)
		wake_up(&state->wq);
	if (delete || state->state == 0) {
429
		if (state->tree) {
430
			clear_state_cb(tree, state, state->state);
431
			rb_erase(&state->rb_node, &tree->state);
432
			state->tree = NULL;
433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472
			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_io_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;
	unsigned long flags;
	int err;
	int set = 0;

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

473
	spin_lock_irqsave(&tree->lock, flags);
474 475 476 477
	/*
	 * this search will find the extents that end after
	 * our range starts
	 */
478
	node = tree_search(tree, start);
479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502
	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) {
503 504
		if (!prealloc)
			prealloc = alloc_extent_state(GFP_ATOMIC);
505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525
		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) {
526 527
		if (!prealloc)
			prealloc = alloc_extent_state(GFP_ATOMIC);
528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543
		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:
544
	spin_unlock_irqrestore(&tree->lock, flags);
545 546 547 548 549 550 551 552
	if (prealloc)
		free_extent_state(prealloc);

	return set;

search_again:
	if (start > end)
		goto out;
553
	spin_unlock_irqrestore(&tree->lock, flags);
554 555 556 557 558 559 560 561 562 563 564
	if (mask & __GFP_WAIT)
		cond_resched();
	goto again;
}
EXPORT_SYMBOL(clear_extent_bit);

static int wait_on_state(struct extent_io_tree *tree,
			 struct extent_state *state)
{
	DEFINE_WAIT(wait);
	prepare_to_wait(&state->wq, &wait, TASK_UNINTERRUPTIBLE);
565
	spin_unlock_irq(&tree->lock);
566
	schedule();
567
	spin_lock_irq(&tree->lock);
568 569 570 571 572 573 574 575 576 577 578 579 580 581
	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_io_tree *tree, u64 start, u64 end, int bits)
{
	struct extent_state *state;
	struct rb_node *node;

582
	spin_lock_irq(&tree->lock);
583 584 585 586 587 588
again:
	while (1) {
		/*
		 * this search will find all the extents that end after
		 * our range starts
		 */
589
		node = tree_search(tree, start);
590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610
		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()) {
611
			spin_unlock_irq(&tree->lock);
612
			cond_resched();
613
			spin_lock_irq(&tree->lock);
614 615 616
		}
	}
out:
617
	spin_unlock_irq(&tree->lock);
618 619 620 621 622 623 624 625 626 627 628 629
	return 0;
}
EXPORT_SYMBOL(wait_extent_bit);

static void set_state_bits(struct extent_io_tree *tree,
			   struct extent_state *state,
			   int bits)
{
	if ((bits & EXTENT_DIRTY) && !(state->state & EXTENT_DIRTY)) {
		u64 range = state->end - state->start + 1;
		tree->dirty_bytes += range;
	}
630
	set_state_cb(tree, state, bits);
C
Chris Mason 已提交
631
	state->state |= bits;
632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662
}

/*
 * 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_io_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;
	unsigned long flags;
	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;
	}

663
	spin_lock_irqsave(&tree->lock, flags);
664 665 666 667
	/*
	 * this search will find all the extents that end after
	 * our range starts.
	 */
668
	node = tree_search(tree, start);
669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782
	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;
		}
		set_state_bits(tree, 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) {
			set_state_bits(tree, 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;
	}
	/*
	 * | ---- 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);

		set_state_bits(tree, prealloc, bits);
		merge_state(tree, prealloc);
		prealloc = NULL;
		goto out;
	}

	goto search_again;

out:
783
	spin_unlock_irqrestore(&tree->lock, flags);
784 785 786 787 788 789 790 791
	if (prealloc)
		free_extent_state(prealloc);

	return err;

search_again:
	if (start > end)
		goto out;
792
	spin_unlock_irqrestore(&tree->lock, flags);
793 794 795 796 797 798 799 800 801 802 803 804 805 806 807
	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_io_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);

808 809 810 811 812 813 814
int set_extent_ordered(struct extent_io_tree *tree, u64 start, u64 end,
		       gfp_t mask)
{
	return set_extent_bit(tree, start, end, EXTENT_ORDERED, 0, NULL, mask);
}
EXPORT_SYMBOL(set_extent_ordered);

815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833
int set_extent_bits(struct extent_io_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_io_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);

int set_extent_delalloc(struct extent_io_tree *tree, u64 start, u64 end,
		     gfp_t mask)
{
	return set_extent_bit(tree, start, end,
834 835
			      EXTENT_DELALLOC | EXTENT_DIRTY,
			      0, NULL, mask);
836 837 838 839 840 841 842 843 844 845 846
}
EXPORT_SYMBOL(set_extent_delalloc);

int clear_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end,
		       gfp_t mask)
{
	return clear_extent_bit(tree, start, end,
				EXTENT_DIRTY | EXTENT_DELALLOC, 0, 0, mask);
}
EXPORT_SYMBOL(clear_extent_dirty);

847 848 849 850 851 852 853
int clear_extent_ordered(struct extent_io_tree *tree, u64 start, u64 end,
			 gfp_t mask)
{
	return clear_extent_bit(tree, start, end, EXTENT_ORDERED, 1, 0, mask);
}
EXPORT_SYMBOL(clear_extent_ordered);

854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979
int set_extent_new(struct extent_io_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_io_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_io_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_io_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_io_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_io_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_io_tree *tree, u64 start, u64 end)
{
	return wait_extent_bit(tree, start, end, EXTENT_WRITEBACK);
}
EXPORT_SYMBOL(wait_on_extent_writeback);

int lock_extent(struct extent_io_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_io_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_io_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_io_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);

int find_first_extent_bit(struct extent_io_tree *tree, u64 start,
			  u64 *start_ret, u64 *end_ret, int bits)
{
	struct rb_node *node;
	struct extent_state *state;
	int ret = 1;

980
	spin_lock_irq(&tree->lock);
981 982 983 984
	/*
	 * this search will find all the extents that end after
	 * our range starts.
	 */
985
	node = tree_search(tree, start);
986
	if (!node) {
987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002
		goto out;
	}

	while(1) {
		state = rb_entry(node, struct extent_state, rb_node);
		if (state->end >= start && (state->state & bits)) {
			*start_ret = state->start;
			*end_ret = state->end;
			ret = 0;
			break;
		}
		node = rb_next(node);
		if (!node)
			break;
	}
out:
1003
	spin_unlock_irq(&tree->lock);
1004 1005 1006 1007
	return ret;
}
EXPORT_SYMBOL(find_first_extent_bit);

C
Chris Mason 已提交
1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018
struct extent_state *find_first_extent_bit_state(struct extent_io_tree *tree,
						 u64 start, int bits)
{
	struct rb_node *node;
	struct extent_state *state;

	/*
	 * this search will find all the extents that end after
	 * our range starts.
	 */
	node = tree_search(tree, start);
1019
	if (!node) {
C
Chris Mason 已提交
1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036
		goto out;
	}

	while(1) {
		state = rb_entry(node, struct extent_state, rb_node);
		if (state->end >= start && (state->state & bits)) {
			return state;
		}
		node = rb_next(node);
		if (!node)
			break;
	}
out:
	return NULL;
}
EXPORT_SYMBOL(find_first_extent_bit_state);

1037 1038 1039 1040 1041 1042 1043 1044 1045
u64 find_lock_delalloc_range(struct extent_io_tree *tree,
			     u64 *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;

1046
	spin_lock_irq(&tree->lock);
1047 1048 1049 1050 1051
	/*
	 * this search will find all the extents that end after
	 * our range starts.
	 */
search_again:
1052
	node = tree_search(tree, cur_start);
1053
	if (!node) {
1054 1055
		if (!found)
			*end = (u64)-1;
1056 1057 1058 1059 1060 1061 1062 1063 1064 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
		goto out;
	}

	while(1) {
		state = rb_entry(node, struct extent_state, rb_node);
		if (found && state->start != cur_start) {
			goto out;
		}
		if (!(state->state & EXTENT_DELALLOC)) {
			if (!found)
				*end = state->end;
			goto out;
		}
		if (!found) {
			struct extent_state *prev_state;
			struct rb_node *prev_node = node;
			while(1) {
				prev_node = rb_prev(prev_node);
				if (!prev_node)
					break;
				prev_state = rb_entry(prev_node,
						      struct extent_state,
						      rb_node);
				if (!(prev_state->state & EXTENT_DELALLOC))
					break;
				state = prev_state;
				node = prev_node;
			}
		}
		if (state->state & EXTENT_LOCKED) {
			DEFINE_WAIT(wait);
			atomic_inc(&state->refs);
			prepare_to_wait(&state->wq, &wait,
					TASK_UNINTERRUPTIBLE);
1090
			spin_unlock_irq(&tree->lock);
1091
			schedule();
1092
			spin_lock_irq(&tree->lock);
1093 1094 1095 1096
			finish_wait(&state->wq, &wait);
			free_extent_state(state);
			goto search_again;
		}
1097
		set_state_cb(tree, state, EXTENT_LOCKED);
C
Chris Mason 已提交
1098
		state->state |= EXTENT_LOCKED;
1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111
		if (!found)
			*start = state->start;
		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:
1112
	spin_unlock_irq(&tree->lock);
1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131
	return found;
}

u64 count_range_bits(struct extent_io_tree *tree,
		     u64 *start, u64 search_end, u64 max_bytes,
		     unsigned long bits)
{
	struct rb_node *node;
	struct extent_state *state;
	u64 cur_start = *start;
	u64 total_bytes = 0;
	int found = 0;

	if (search_end <= cur_start) {
		printk("search_end %Lu start %Lu\n", search_end, cur_start);
		WARN_ON(1);
		return 0;
	}

1132
	spin_lock_irq(&tree->lock);
1133 1134 1135 1136 1137 1138 1139 1140
	if (cur_start == 0 && bits == EXTENT_DIRTY) {
		total_bytes = tree->dirty_bytes;
		goto out;
	}
	/*
	 * this search will find all the extents that end after
	 * our range starts.
	 */
1141
	node = tree_search(tree, cur_start);
1142
	if (!node) {
1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164
		goto out;
	}

	while(1) {
		state = rb_entry(node, struct extent_state, rb_node);
		if (state->start > search_end)
			break;
		if (state->end >= cur_start && (state->state & bits)) {
			total_bytes += min(search_end, state->end) + 1 -
				       max(cur_start, state->start);
			if (total_bytes >= max_bytes)
				break;
			if (!found) {
				*start = state->start;
				found = 1;
			}
		}
		node = rb_next(node);
		if (!node)
			break;
	}
out:
1165
	spin_unlock_irq(&tree->lock);
1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 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 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236
	return total_bytes;
}
/*
 * helper function to lock both pages and extents in the tree.
 * pages must be locked first.
 */
int lock_range(struct extent_io_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_io_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);

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

1237
	spin_lock_irq(&tree->lock);
1238 1239 1240 1241
	/*
	 * this search will find all the extents that end after
	 * our range starts.
	 */
1242
	node = tree_search(tree, start);
1243
	if (!node) {
1244 1245 1246 1247 1248 1249 1250 1251 1252 1253
		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:
1254
	spin_unlock_irq(&tree->lock);
1255 1256 1257 1258 1259 1260 1261 1262 1263
	return ret;
}

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

1264
	spin_lock_irq(&tree->lock);
1265 1266 1267 1268
	/*
	 * this search will find all the extents that end after
	 * our range starts.
	 */
1269
	node = tree_search(tree, start);
1270
	if (!node) {
1271 1272 1273 1274 1275 1276 1277 1278 1279 1280
		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:
1281
	spin_unlock_irq(&tree->lock);
1282 1283 1284 1285 1286
	return ret;
}

/*
 * searches a range in the state tree for a given mask.
1287
 * If 'filled' == 1, this returns 1 only if every extent in the tree
1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298
 * has the bits set.  Otherwise, 1 is returned if any bit in the
 * range is found set.
 */
int test_range_bit(struct extent_io_tree *tree, u64 start, u64 end,
		   int bits, int filled)
{
	struct extent_state *state = NULL;
	struct rb_node *node;
	int bitset = 0;
	unsigned long flags;

1299
	spin_lock_irqsave(&tree->lock, flags);
1300
	node = tree_search(tree, start);
1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329
	while (node && start <= end) {
		state = rb_entry(node, struct extent_state, rb_node);

		if (filled && state->start > start) {
			bitset = 0;
			break;
		}

		if (state->start > end)
			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);
		if (!node) {
			if (filled)
				bitset = 0;
			break;
		}
	}
1330
	spin_unlock_irqrestore(&tree->lock, flags);
1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 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
	return bitset;
}
EXPORT_SYMBOL(test_range_bit);

/*
 * 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_io_tree *tree,
			       struct page *page)
{
	u64 start = (u64)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_io_tree *tree,
			     struct page *page)
{
	u64 start = (u64)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_io_tree *tree,
			     struct page *page)
{
	u64 start = (u64)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.
 */
#if LINUX_VERSION_CODE > KERNEL_VERSION(2,6,23)
static void end_bio_extent_writepage(struct bio *bio, int err)
#else
static int end_bio_extent_writepage(struct bio *bio,
				   unsigned int bytes_done, int err)
#endif
{
1395
	int uptodate = err == 0;
1396
	struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
1397 1398 1399
	struct extent_state *state = bio->bi_private;
	struct extent_io_tree *tree = state->tree;
	struct rb_node *node;
1400 1401
	u64 start;
	u64 end;
1402
	u64 cur;
1403
	int whole_page;
1404
	int ret;
1405
	unsigned long flags;
1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423

#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,23)
	if (bio->bi_size)
		return 1;
#endif
	do {
		struct page *page = bvec->bv_page;
		start = ((u64)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);
1424 1425
		if (tree->ops && tree->ops->writepage_end_io_hook) {
			ret = tree->ops->writepage_end_io_hook(page, start,
1426
						       end, state, uptodate);
1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441
			if (ret)
				uptodate = 0;
		}

		if (!uptodate && tree->ops &&
		    tree->ops->writepage_io_failed_hook) {
			ret = tree->ops->writepage_io_failed_hook(bio, page,
							 start, end, state);
			if (ret == 0) {
				state = NULL;
				uptodate = (err == 0);
				continue;
			}
		}

1442 1443 1444 1445 1446
		if (!uptodate) {
			clear_extent_uptodate(tree, start, end, GFP_ATOMIC);
			ClearPageUptodate(page);
			SetPageError(page);
		}
1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460

		/*
		 * bios can get merged in funny ways, and so we need to
		 * be careful with the state variable.  We know the
		 * state won't be merged with others because it has
		 * WRITEBACK set, but we can't be sure each biovec is
		 * sequential in the file.  So, if our cached state
		 * doesn't match the expected end, search the tree
		 * for the correct one.
		 */

		spin_lock_irqsave(&tree->lock, flags);
		if (!state || state->end != end) {
			state = NULL;
1461
			node = __etree_search(tree, start, NULL, NULL);
1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 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 1498 1499 1500 1501 1502 1503 1504
			if (node) {
				state = rb_entry(node, struct extent_state,
						 rb_node);
				if (state->end != end ||
				    !(state->state & EXTENT_WRITEBACK))
					state = NULL;
			}
			if (!state) {
				spin_unlock_irqrestore(&tree->lock, flags);
				clear_extent_writeback(tree, start,
						       end, GFP_ATOMIC);
				goto next_io;
			}
		}
		cur = end;
		while(1) {
			struct extent_state *clear = state;
			cur = state->start;
			node = rb_prev(&state->rb_node);
			if (node) {
				state = rb_entry(node,
						 struct extent_state,
						 rb_node);
			} else {
				state = NULL;
			}

			clear_state_bit(tree, clear, EXTENT_WRITEBACK,
					1, 0);
			if (cur == start)
				break;
			if (cur < start) {
				WARN_ON(1);
				break;
			}
			if (!node)
				break;
		}
		/* before releasing the lock, make sure the next state
		 * variable has the expected bits set and corresponds
		 * to the correct offsets in the file
		 */
		if (state && (state->end + 1 != start ||
Y
Yan 已提交
1505
		    !(state->state & EXTENT_WRITEBACK))) {
1506 1507 1508 1509
			state = NULL;
		}
		spin_unlock_irqrestore(&tree->lock, flags);
next_io:
1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541

		if (whole_page)
			end_page_writeback(page);
		else
			check_page_writeback(tree, page);
	} while (bvec >= bio->bi_io_vec);
	bio_put(bio);
#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,23)
	return 0;
#endif
}

/*
 * 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.
 */
#if LINUX_VERSION_CODE > KERNEL_VERSION(2,6,23)
static void end_bio_extent_readpage(struct bio *bio, int err)
#else
static int end_bio_extent_readpage(struct bio *bio,
				   unsigned int bytes_done, int err)
#endif
{
	int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
	struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
1542 1543 1544
	struct extent_state *state = bio->bi_private;
	struct extent_io_tree *tree = state->tree;
	struct rb_node *node;
1545 1546
	u64 start;
	u64 end;
1547 1548
	u64 cur;
	unsigned long flags;
1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571
	int whole_page;
	int ret;

#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,23)
	if (bio->bi_size)
		return 1;
#endif

	do {
		struct page *page = bvec->bv_page;
		start = ((u64)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 && tree->ops && tree->ops->readpage_end_io_hook) {
1572 1573
			ret = tree->ops->readpage_end_io_hook(page, start, end,
							      state);
1574 1575 1576
			if (ret)
				uptodate = 0;
		}
1577 1578 1579 1580 1581 1582
		if (!uptodate && tree->ops &&
		    tree->ops->readpage_io_failed_hook) {
			ret = tree->ops->readpage_io_failed_hook(bio, page,
							 start, end, state);
			if (ret == 0) {
				state = NULL;
1583 1584
				uptodate =
					test_bit(BIO_UPTODATE, &bio->bi_flags);
1585 1586 1587
				continue;
			}
		}
1588

1589 1590 1591
		spin_lock_irqsave(&tree->lock, flags);
		if (!state || state->end != end) {
			state = NULL;
1592
			node = __etree_search(tree, start, NULL, NULL);
1593 1594 1595 1596 1597 1598 1599
			if (node) {
				state = rb_entry(node, struct extent_state,
						 rb_node);
				if (state->end != end ||
				    !(state->state & EXTENT_LOCKED))
					state = NULL;
			}
1600
			if (!state) {
1601
				spin_unlock_irqrestore(&tree->lock, flags);
1602 1603 1604
				if (uptodate)
					set_extent_uptodate(tree, start, end,
							    GFP_ATOMIC);
1605 1606 1607 1608
				unlock_extent(tree, start, end, GFP_ATOMIC);
				goto next_io;
			}
		}
1609

1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621
		cur = end;
		while(1) {
			struct extent_state *clear = state;
			cur = state->start;
			node = rb_prev(&state->rb_node);
			if (node) {
				state = rb_entry(node,
					 struct extent_state,
					 rb_node);
			} else {
				state = NULL;
			}
1622 1623 1624 1625
			if (uptodate) {
				set_state_cb(tree, clear, EXTENT_UPTODATE);
				clear->state |= EXTENT_UPTODATE;
			}
1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641
			clear_state_bit(tree, clear, EXTENT_LOCKED,
					1, 0);
			if (cur == start)
				break;
			if (cur < start) {
				WARN_ON(1);
				break;
			}
			if (!node)
				break;
		}
		/* before releasing the lock, make sure the next state
		 * variable has the expected bits set and corresponds
		 * to the correct offsets in the file
		 */
		if (state && (state->end + 1 != start ||
Y
Yan 已提交
1642
		    !(state->state & EXTENT_LOCKED))) {
1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653
			state = NULL;
		}
		spin_unlock_irqrestore(&tree->lock, flags);
next_io:
		if (whole_page) {
			if (uptodate) {
				SetPageUptodate(page);
			} else {
				ClearPageUptodate(page);
				SetPageError(page);
			}
1654
			unlock_page(page);
1655 1656 1657 1658 1659 1660 1661
		} else {
			if (uptodate) {
				check_page_uptodate(tree, page);
			} else {
				ClearPageUptodate(page);
				SetPageError(page);
			}
1662
			check_page_locked(tree, page);
1663
		}
1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685
	} while (bvec >= bio->bi_io_vec);

	bio_put(bio);
#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,23)
	return 0;
#endif
}

/*
 * 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.
 */
#if LINUX_VERSION_CODE > KERNEL_VERSION(2,6,23)
static void end_bio_extent_preparewrite(struct bio *bio, int err)
#else
static int end_bio_extent_preparewrite(struct bio *bio,
				       unsigned int bytes_done, int err)
#endif
{
	const int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
	struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
1686 1687
	struct extent_state *state = bio->bi_private;
	struct extent_io_tree *tree = state->tree;
1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 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
	u64 start;
	u64 end;

#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,23)
	if (bio->bi_size)
		return 1;
#endif

	do {
		struct page *page = bvec->bv_page;
		start = ((u64)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);
#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,23)
	return 0;
#endif
}

static struct bio *
extent_bio_alloc(struct block_device *bdev, u64 first_sector, int nr_vecs,
		 gfp_t gfp_flags)
{
	struct bio *bio;

	bio = bio_alloc(gfp_flags, nr_vecs);

	if (bio == NULL && (current->flags & PF_MEMALLOC)) {
		while (!bio && (nr_vecs /= 2))
			bio = bio_alloc(gfp_flags, nr_vecs);
	}

	if (bio) {
1736
		bio->bi_size = 0;
1737 1738 1739 1740 1741 1742
		bio->bi_bdev = bdev;
		bio->bi_sector = first_sector;
	}
	return bio;
}

1743
static int submit_one_bio(int rw, struct bio *bio, int mirror_num)
1744 1745
{
	int ret = 0;
1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757
	struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
	struct page *page = bvec->bv_page;
	struct extent_io_tree *tree = bio->bi_private;
	struct rb_node *node;
	struct extent_state *state;
	u64 start;
	u64 end;

	start = ((u64)page->index << PAGE_CACHE_SHIFT) + bvec->bv_offset;
	end = start + bvec->bv_len - 1;

	spin_lock_irq(&tree->lock);
1758
	node = __etree_search(tree, start, NULL, NULL);
1759 1760 1761 1762 1763 1764 1765 1766 1767 1768
	BUG_ON(!node);
	state = rb_entry(node, struct extent_state, rb_node);
	while(state->end < end) {
		node = rb_next(node);
		state = rb_entry(node, struct extent_state, rb_node);
	}
	BUG_ON(state->end != end);
	spin_unlock_irq(&tree->lock);

	bio->bi_private = state;
1769 1770 1771

	bio_get(bio);

1772
	if (tree->ops && tree->ops->submit_bio_hook)
1773 1774
		tree->ops->submit_bio_hook(page->mapping->host, rw, bio,
					   mirror_num);
1775 1776
	else
		submit_bio(rw, bio);
1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788
	if (bio_flagged(bio, BIO_EOPNOTSUPP))
		ret = -EOPNOTSUPP;
	bio_put(bio);
	return ret;
}

static int submit_extent_page(int rw, struct extent_io_tree *tree,
			      struct page *page, sector_t sector,
			      size_t size, unsigned long offset,
			      struct block_device *bdev,
			      struct bio **bio_ret,
			      unsigned long max_pages,
1789 1790
			      bio_end_io_t end_io_func,
			      int mirror_num)
1791 1792 1793 1794 1795 1796 1797 1798
{
	int ret = 0;
	struct bio *bio;
	int nr;

	if (bio_ret && *bio_ret) {
		bio = *bio_ret;
		if (bio->bi_sector + (bio->bi_size >> 9) != sector ||
1799 1800
		    (tree->ops && tree->ops->merge_bio_hook &&
		     tree->ops->merge_bio_hook(page, offset, size, bio)) ||
1801
		    bio_add_page(bio, page, size, offset) < size) {
1802
			ret = submit_one_bio(rw, bio, mirror_num);
1803 1804 1805 1806 1807
			bio = NULL;
		} else {
			return 0;
		}
	}
1808
	nr = bio_get_nr_vecs(bdev);
1809 1810 1811 1812
	bio = extent_bio_alloc(bdev, sector, nr, GFP_NOFS | __GFP_HIGH);
	if (!bio) {
		printk("failed to allocate bio nr %d\n", nr);
	}
1813 1814


1815 1816 1817
	bio_add_page(bio, page, size, offset);
	bio->bi_end_io = end_io_func;
	bio->bi_private = tree;
1818

1819 1820 1821
	if (bio_ret) {
		*bio_ret = bio;
	} else {
1822
		ret = submit_one_bio(rw, bio, mirror_num);
1823 1824 1825 1826 1827 1828 1829 1830 1831 1832
	}

	return ret;
}

void set_page_extent_mapped(struct page *page)
{
	if (!PagePrivate(page)) {
		SetPagePrivate(page);
		page_cache_get(page);
1833
		set_page_private(page, EXTENT_PAGE_PRIVATE);
1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849
	}
}

void set_page_extent_head(struct page *page, unsigned long len)
{
	set_page_private(page, EXTENT_PAGE_PRIVATE_FIRST_PAGE | len << 2);
}

/*
 * 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)
 */
static int __extent_read_full_page(struct extent_io_tree *tree,
				   struct page *page,
				   get_extent_t *get_extent,
1850
				   struct bio **bio, int mirror_num)
1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895
{
	struct inode *inode = page->mapping->host;
	u64 start = (u64)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;

	set_page_extent_mapped(page);

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

	while (cur <= end) {
		if (cur >= last_byte) {
			char *userpage;
			iosize = PAGE_CACHE_SIZE - page_offset;
			userpage = kmap_atomic(page, KM_USER0);
			memset(userpage + page_offset, 0, iosize);
			flush_dcache_page(page);
			kunmap_atomic(userpage, 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 - cur + 1, 0);
		if (IS_ERR(em) || !em) {
			SetPageError(page);
			unlock_extent(tree, cur, end, GFP_NOFS);
			break;
		}
		extent_offset = cur - em->start;
1896 1897 1898
		if (extent_map_end(em) <= cur) {
printk("bad mapping em [%Lu %Lu] cur %Lu\n", em->start, extent_map_end(em), cur);
		}
1899
		BUG_ON(extent_map_end(em) <= cur);
1900 1901 1902
		if (end < cur) {
printk("2bad mapping end %Lu cur %Lu\n", end, cur);
		}
1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935
		BUG_ON(end < cur);

		iosize = min(extent_map_end(em) - cur, end - cur + 1);
		cur_end = min(extent_map_end(em) - 1, 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 */
		if (block_start == EXTENT_MAP_HOLE) {
			char *userpage;
			userpage = kmap_atomic(page, KM_USER0);
			memset(userpage + page_offset, 0, iosize);
			flush_dcache_page(page);
			kunmap_atomic(userpage, 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;
		}
1936 1937 1938 1939 1940 1941 1942 1943 1944 1945
		/* we have an inline extent but it didn't get marked up
		 * to date.  Error out
		 */
		if (block_start == EXTENT_MAP_INLINE) {
			SetPageError(page);
			unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS);
			cur = cur + iosize;
			page_offset += iosize;
			continue;
		}
1946 1947 1948 1949 1950 1951 1952

		ret = 0;
		if (tree->ops && tree->ops->readpage_io_hook) {
			ret = tree->ops->readpage_io_hook(page, cur,
							  cur + iosize - 1);
		}
		if (!ret) {
1953 1954
			unsigned long pnr = (last_byte >> PAGE_CACHE_SHIFT) + 1;
			pnr -= page->index;
1955 1956
			ret = submit_extent_page(READ, tree, page,
					 sector, iosize, page_offset,
1957
					 bdev, bio, pnr,
1958
					 end_bio_extent_readpage, mirror_num);
1959
			nr++;
1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979
		}
		if (ret)
			SetPageError(page);
		cur = cur + iosize;
		page_offset += iosize;
	}
	if (!nr) {
		if (!PageError(page))
			SetPageUptodate(page);
		unlock_page(page);
	}
	return 0;
}

int extent_read_full_page(struct extent_io_tree *tree, struct page *page,
			    get_extent_t *get_extent)
{
	struct bio *bio = NULL;
	int ret;

1980
	ret = __extent_read_full_page(tree, page, get_extent, &bio, 0);
1981
	if (bio)
1982
		submit_one_bio(READ, bio, 0);
1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
	return ret;
}
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
 */
static int __extent_writepage(struct page *page, struct writeback_control *wbc,
			      void *data)
{
	struct inode *inode = page->mapping->host;
	struct extent_page_data *epd = data;
	struct extent_io_tree *tree = epd->tree;
	u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
	u64 delalloc_start;
	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 iosize;
2008
	u64 unlock_start;
2009 2010 2011 2012 2013
	sector_t sector;
	struct extent_map *em;
	struct block_device *bdev;
	int ret;
	int nr = 0;
2014
	size_t pg_offset = 0;
2015 2016 2017 2018 2019 2020 2021
	size_t blocksize;
	loff_t i_size = i_size_read(inode);
	unsigned long end_index = i_size >> PAGE_CACHE_SHIFT;
	u64 nr_delalloc;
	u64 delalloc_end;

	WARN_ON(!PageLocked(page));
2022
	pg_offset = i_size & (PAGE_CACHE_SIZE - 1);
2023
	if (page->index > end_index ||
2024
	   (page->index == end_index && !pg_offset)) {
2025
		page->mapping->a_ops->invalidatepage(page, 0);
2026 2027 2028 2029 2030 2031 2032 2033
		unlock_page(page);
		return 0;
	}

	if (page->index == end_index) {
		char *userpage;

		userpage = kmap_atomic(page, KM_USER0);
2034 2035
		memset(userpage + pg_offset, 0,
		       PAGE_CACHE_SIZE - pg_offset);
2036
		kunmap_atomic(userpage, KM_USER0);
2037
		flush_dcache_page(page);
2038
	}
2039
	pg_offset = 0;
2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061

	set_page_extent_mapped(page);

	delalloc_start = start;
	delalloc_end = 0;
	while(delalloc_end < page_end) {
		nr_delalloc = find_lock_delalloc_range(tree, &delalloc_start,
						       &delalloc_end,
						       128 * 1024 * 1024);
		if (nr_delalloc == 0) {
			delalloc_start = delalloc_end + 1;
			continue;
		}
		tree->ops->fill_delalloc(inode, delalloc_start,
					 delalloc_end);
		clear_extent_bit(tree, delalloc_start,
				 delalloc_end,
				 EXTENT_LOCKED | EXTENT_DELALLOC,
				 1, 0, GFP_NOFS);
		delalloc_start = delalloc_end + 1;
	}
	lock_extent(tree, start, page_end, GFP_NOFS);
2062
	unlock_start = start;
2063

2064 2065 2066 2067 2068 2069 2070 2071 2072 2073
	if (tree->ops && tree->ops->writepage_start_hook) {
		ret = tree->ops->writepage_start_hook(page, start, page_end);
		if (ret == -EAGAIN) {
			unlock_extent(tree, start, page_end, GFP_NOFS);
			redirty_page_for_writepage(wbc, page);
			unlock_page(page);
			return 0;
		}
	}

2074 2075 2076 2077 2078 2079 2080
	end = page_end;
	if (test_range_bit(tree, start, page_end, EXTENT_DELALLOC, 0)) {
		printk("found delalloc bits after lock_extent\n");
	}

	if (last_byte <= start) {
		clear_extent_dirty(tree, start, page_end, GFP_NOFS);
2081 2082 2083 2084 2085
		unlock_extent(tree, start, page_end, GFP_NOFS);
		if (tree->ops && tree->ops->writepage_end_io_hook)
			tree->ops->writepage_end_io_hook(page, start,
							 page_end, NULL, 1);
		unlock_start = page_end + 1;
2086 2087 2088 2089 2090 2091 2092 2093 2094
		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);
2095 2096 2097 2098 2099
			unlock_extent(tree, unlock_start, page_end, GFP_NOFS);
			if (tree->ops && tree->ops->writepage_end_io_hook)
				tree->ops->writepage_end_io_hook(page, cur,
							 page_end, NULL, 1);
			unlock_start = page_end + 1;
2100 2101
			break;
		}
2102
		em = epd->get_extent(inode, page, pg_offset, cur,
2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123
				     end - cur + 1, 1);
		if (IS_ERR(em) || !em) {
			SetPageError(page);
			break;
		}

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

		if (block_start == EXTENT_MAP_HOLE ||
		    block_start == EXTENT_MAP_INLINE) {
			clear_extent_dirty(tree, cur,
					   cur + iosize - 1, GFP_NOFS);
2124 2125 2126

			unlock_extent(tree, unlock_start, cur + iosize -1,
				      GFP_NOFS);
2127

2128 2129 2130 2131
			if (tree->ops && tree->ops->writepage_end_io_hook)
				tree->ops->writepage_end_io_hook(page, cur,
							 cur + iosize - 1,
							 NULL, 1);
2132
			cur = cur + iosize;
2133
			pg_offset += iosize;
2134
			unlock_start = cur;
2135 2136 2137 2138 2139 2140 2141
			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;
2142
			pg_offset += iosize;
2143 2144 2145 2146 2147 2148 2149 2150 2151
			continue;
		}
		clear_extent_dirty(tree, cur, cur + iosize - 1, GFP_NOFS);
		if (tree->ops && tree->ops->writepage_io_hook) {
			ret = tree->ops->writepage_io_hook(page, cur,
						cur + iosize - 1);
		} else {
			ret = 0;
		}
2152
		if (ret) {
2153
			SetPageError(page);
2154
		} else {
2155
			unsigned long max_nr = end_index + 1;
2156

2157 2158 2159 2160 2161 2162 2163 2164 2165
			set_range_writeback(tree, cur, cur + iosize - 1);
			if (!PageWriteback(page)) {
				printk("warning page %lu not writeback, "
				       "cur %llu end %llu\n", page->index,
				       (unsigned long long)cur,
				       (unsigned long long)end);
			}

			ret = submit_extent_page(WRITE, tree, page, sector,
2166
						 iosize, pg_offset, bdev,
2167
						 &epd->bio, max_nr,
2168
						 end_bio_extent_writepage, 0);
2169 2170 2171 2172
			if (ret)
				SetPageError(page);
		}
		cur = cur + iosize;
2173
		pg_offset += iosize;
2174 2175 2176 2177 2178 2179 2180 2181
		nr++;
	}
done:
	if (nr == 0) {
		/* make sure the mapping tag for page dirty gets cleared */
		set_page_writeback(page);
		end_page_writeback(page);
	}
2182 2183
	if (unlock_start <= page_end)
		unlock_extent(tree, unlock_start, page_end, GFP_NOFS);
2184 2185 2186 2187
	unlock_page(page);
	return 0;
}

2188
#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,22)
2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334
/* Taken directly from 2.6.23 for 2.6.18 back port */
typedef int (*writepage_t)(struct page *page, struct writeback_control *wbc,
                                void *data);

/**
 * write_cache_pages - walk the list of dirty pages of the given address space
 * and write all of them.
 * @mapping: address space structure to write
 * @wbc: subtract the number of written pages from *@wbc->nr_to_write
 * @writepage: function called for each page
 * @data: data passed to writepage function
 *
 * If a page is already under I/O, write_cache_pages() skips it, even
 * if it's dirty.  This is desirable behaviour for memory-cleaning writeback,
 * but it is INCORRECT for data-integrity system calls such as fsync().  fsync()
 * and msync() need to guarantee that all the data which was dirty at the time
 * the call was made get new I/O started against them.  If wbc->sync_mode is
 * WB_SYNC_ALL then we were called for data integrity and we must wait for
 * existing IO to complete.
 */
static int write_cache_pages(struct address_space *mapping,
		      struct writeback_control *wbc, writepage_t writepage,
		      void *data)
{
	struct backing_dev_info *bdi = mapping->backing_dev_info;
	int ret = 0;
	int done = 0;
	struct pagevec pvec;
	int nr_pages;
	pgoff_t index;
	pgoff_t end;		/* Inclusive */
	int scanned = 0;
	int range_whole = 0;

	if (wbc->nonblocking && bdi_write_congested(bdi)) {
		wbc->encountered_congestion = 1;
		return 0;
	}

	pagevec_init(&pvec, 0);
	if (wbc->range_cyclic) {
		index = mapping->writeback_index; /* Start from prev offset */
		end = -1;
	} else {
		index = wbc->range_start >> PAGE_CACHE_SHIFT;
		end = wbc->range_end >> PAGE_CACHE_SHIFT;
		if (wbc->range_start == 0 && wbc->range_end == LLONG_MAX)
			range_whole = 1;
		scanned = 1;
	}
retry:
	while (!done && (index <= end) &&
	       (nr_pages = pagevec_lookup_tag(&pvec, mapping, &index,
					      PAGECACHE_TAG_DIRTY,
					      min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1))) {
		unsigned i;

		scanned = 1;
		for (i = 0; i < nr_pages; i++) {
			struct page *page = pvec.pages[i];

			/*
			 * At this point we hold neither mapping->tree_lock nor
			 * lock on the page itself: the page may be truncated or
			 * invalidated (changing page->mapping to NULL), or even
			 * swizzled back from swapper_space to tmpfs file
			 * mapping
			 */
			lock_page(page);

			if (unlikely(page->mapping != mapping)) {
				unlock_page(page);
				continue;
			}

			if (!wbc->range_cyclic && page->index > end) {
				done = 1;
				unlock_page(page);
				continue;
			}

			if (wbc->sync_mode != WB_SYNC_NONE)
				wait_on_page_writeback(page);

			if (PageWriteback(page) ||
			    !clear_page_dirty_for_io(page)) {
				unlock_page(page);
				continue;
			}

			ret = (*writepage)(page, wbc, data);

			if (unlikely(ret == AOP_WRITEPAGE_ACTIVATE)) {
				unlock_page(page);
				ret = 0;
			}
			if (ret || (--(wbc->nr_to_write) <= 0))
				done = 1;
			if (wbc->nonblocking && bdi_write_congested(bdi)) {
				wbc->encountered_congestion = 1;
				done = 1;
			}
		}
		pagevec_release(&pvec);
		cond_resched();
	}
	if (!scanned && !done) {
		/*
		 * We hit the last page and there is more work to be done: wrap
		 * back to the start of the file
		 */
		scanned = 1;
		index = 0;
		goto retry;
	}
	if (wbc->range_cyclic || (range_whole && wbc->nr_to_write > 0))
		mapping->writeback_index = index;
	return ret;
}
#endif

int extent_write_full_page(struct extent_io_tree *tree, struct page *page,
			  get_extent_t *get_extent,
			  struct writeback_control *wbc)
{
	int ret;
	struct address_space *mapping = page->mapping;
	struct extent_page_data epd = {
		.bio = NULL,
		.tree = tree,
		.get_extent = get_extent,
	};
	struct writeback_control wbc_writepages = {
		.bdi		= wbc->bdi,
		.sync_mode	= WB_SYNC_NONE,
		.older_than_this = NULL,
		.nr_to_write	= 64,
		.range_start	= page_offset(page) + PAGE_CACHE_SIZE,
		.range_end	= (loff_t)-1,
	};


	ret = __extent_writepage(page, wbc, &epd);

	write_cache_pages(mapping, &wbc_writepages, __extent_writepage, &epd);
	if (epd.bio) {
2335
		submit_one_bio(WRITE, epd.bio, 0);
2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355
	}
	return ret;
}
EXPORT_SYMBOL(extent_write_full_page);


int extent_writepages(struct extent_io_tree *tree,
		      struct address_space *mapping,
		      get_extent_t *get_extent,
		      struct writeback_control *wbc)
{
	int ret = 0;
	struct extent_page_data epd = {
		.bio = NULL,
		.tree = tree,
		.get_extent = get_extent,
	};

	ret = write_cache_pages(mapping, wbc, __extent_writepage, &epd);
	if (epd.bio) {
2356
		submit_one_bio(WRITE, epd.bio, 0);
2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387
	}
	return ret;
}
EXPORT_SYMBOL(extent_writepages);

int extent_readpages(struct extent_io_tree *tree,
		     struct address_space *mapping,
		     struct list_head *pages, unsigned nr_pages,
		     get_extent_t get_extent)
{
	struct bio *bio = NULL;
	unsigned page_idx;
	struct pagevec pvec;

	pagevec_init(&pvec, 0);
	for (page_idx = 0; page_idx < nr_pages; page_idx++) {
		struct page *page = list_entry(pages->prev, struct page, lru);

		prefetchw(&page->flags);
		list_del(&page->lru);
		/*
		 * what we want to do here is call add_to_page_cache_lru,
		 * but that isn't exported, so we reproduce it here
		 */
		if (!add_to_page_cache(page, mapping,
					page->index, GFP_KERNEL)) {

			/* open coding of lru_cache_add, also not exported */
			page_cache_get(page);
			if (!pagevec_add(&pvec, page))
				__pagevec_lru_add(&pvec);
2388 2389
			__extent_read_full_page(tree, page, get_extent,
						&bio, 0);
2390 2391 2392 2393 2394 2395 2396
		}
		page_cache_release(page);
	}
	if (pagevec_count(&pvec))
		__pagevec_lru_add(&pvec);
	BUG_ON(!list_empty(pages));
	if (bio)
2397
		submit_one_bio(READ, bio, 0);
2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 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 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521
	return 0;
}
EXPORT_SYMBOL(extent_readpages);

/*
 * 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_io_tree *tree,
			  struct page *page, unsigned long offset)
{
	u64 start = ((u64)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);
	clear_extent_bit(tree, start, end,
			 EXTENT_LOCKED | EXTENT_DIRTY | EXTENT_DELALLOC,
			 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_io_tree *tree,
			struct inode *inode, struct page *page,
			unsigned from, unsigned to)
{
	loff_t pos = ((loff_t)page->index << PAGE_CACHE_SHIFT) + to;

	set_page_extent_mapped(page);
	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_io_tree *tree,
			 struct inode *inode, struct page *page,
			 unsigned from, unsigned to, get_extent_t *get_extent)
{
	u64 page_start = (u64)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;

	set_page_extent_mapped(page);

	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 - block_start + 1, 1);
		if (IS_ERR(em) || !em) {
			goto err;
		}
		cur_end = min(block_end, extent_map_end(em) - 1);
		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 ((em->block_start != EXTENT_MAP_HOLE &&
		     em->block_start != EXTENT_MAP_INLINE) &&
		    !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) &
				~((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,
					 NULL, 1,
2522
					 end_bio_extent_preparewrite, 0);
2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544
			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);

2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557
/*
 * a helper for releasepage, this tests for areas of the page that
 * are locked or under IO and drops the related state bits if it is safe
 * to drop the page.
 */
int try_release_extent_state(struct extent_map_tree *map,
			     struct extent_io_tree *tree, struct page *page,
			     gfp_t mask)
{
	u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
	u64 end = start + PAGE_CACHE_SIZE - 1;
	int ret = 1;

2558 2559
	if (test_range_bit(tree, start, end,
			   EXTENT_IOBITS | EXTENT_ORDERED, 0))
2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570
		ret = 0;
	else {
		if ((mask & GFP_NOFS) == GFP_NOFS)
			mask = GFP_NOFS;
		clear_extent_bit(tree, start, end, EXTENT_UPTODATE,
				 1, 1, mask);
	}
	return ret;
}
EXPORT_SYMBOL(try_release_extent_state);

2571 2572 2573 2574 2575 2576
/*
 * 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 *map,
2577 2578
			       struct extent_io_tree *tree, struct page *page,
			       gfp_t mask)
2579 2580 2581 2582
{
	struct extent_map *em;
	u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
	u64 end = start + PAGE_CACHE_SIZE - 1;
2583

2584 2585
	if ((mask & __GFP_WAIT) &&
	    page->mapping->host->i_size > 16 * 1024 * 1024) {
2586
		u64 len;
2587
		while (start <= end) {
2588
			len = end - start + 1;
2589
			spin_lock(&map->lock);
2590
			em = lookup_extent_mapping(map, start, len);
2591 2592 2593 2594
			if (!em || IS_ERR(em)) {
				spin_unlock(&map->lock);
				break;
			}
2595 2596
			if (test_bit(EXTENT_FLAG_PINNED, &em->flags) ||
			    em->start != start) {
2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608
				spin_unlock(&map->lock);
				free_extent_map(em);
				break;
			}
			if (!test_range_bit(tree, em->start,
					    extent_map_end(em) - 1,
					    EXTENT_LOCKED, 0)) {
				remove_extent_mapping(map, em);
				/* once for the rb tree */
				free_extent_map(em);
			}
			start = extent_map_end(em);
2609
			spin_unlock(&map->lock);
2610 2611

			/* once for us */
2612 2613 2614
			free_extent_map(em);
		}
	}
2615
	return try_release_extent_state(map, tree, page, mask);
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
}
EXPORT_SYMBOL(try_release_extent_mapping);

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;
	sector_t sector = 0;
	struct extent_map *em;

	em = get_extent(inode, NULL, 0, start, (1 << inode->i_blkbits), 0);
	if (!em || IS_ERR(em))
		return 0;

	if (em->block_start == EXTENT_MAP_INLINE ||
	    em->block_start == EXTENT_MAP_HOLE)
		goto out;

	sector = (em->block_start + start - em->start) >> inode->i_blkbits;
out:
	free_extent_map(em);
	return sector;
}

static inline struct page *extent_buffer_page(struct extent_buffer *eb,
					      unsigned long i)
{
	struct page *p;
	struct address_space *mapping;

	if (i == 0)
		return eb->first_page;
	i += eb->start >> PAGE_CACHE_SHIFT;
	mapping = eb->first_page->mapping;
C
Chris Mason 已提交
2651 2652
	if (!mapping)
		return NULL;
2653 2654 2655 2656 2657 2658 2659 2660 2661

	/*
	 * extent_buffer_page is only called after pinning the page
	 * by increasing the reference count.  So we know the page must
	 * be in the radix tree.
	 */
#if LINUX_VERSION_CODE > KERNEL_VERSION(2,6,26)
	rcu_read_lock();
#else
2662
	read_lock_irq(&mapping->tree_lock);
2663
#endif
2664
	p = radix_tree_lookup(&mapping->page_tree, i);
2665 2666 2667 2668

#if LINUX_VERSION_CODE > KERNEL_VERSION(2,6,26)
	rcu_read_unlock();
#else
2669
	read_unlock_irq(&mapping->tree_lock);
2670
#endif
2671 2672 2673
	return p;
}

2674
static inline unsigned long num_extent_pages(u64 start, u64 len)
2675
{
2676 2677
	return ((start + len + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT) -
		(start >> PAGE_CACHE_SHIFT);
2678 2679
}

2680 2681 2682 2683 2684 2685
static struct extent_buffer *__alloc_extent_buffer(struct extent_io_tree *tree,
						   u64 start,
						   unsigned long len,
						   gfp_t mask)
{
	struct extent_buffer *eb = NULL;
2686
	unsigned long flags;
2687 2688 2689 2690

	eb = kmem_cache_zalloc(extent_buffer_cache, mask);
	eb->start = start;
	eb->len = len;
2691
	mutex_init(&eb->mutex);
2692 2693 2694
	spin_lock_irqsave(&leak_lock, flags);
	list_add(&eb->leak_list, &buffers);
	spin_unlock_irqrestore(&leak_lock, flags);
2695 2696 2697 2698 2699 2700 2701
	atomic_set(&eb->refs, 1);

	return eb;
}

static void __free_extent_buffer(struct extent_buffer *eb)
{
2702 2703 2704 2705
	unsigned long flags;
	spin_lock_irqsave(&leak_lock, flags);
	list_del(&eb->leak_list);
	spin_unlock_irqrestore(&leak_lock, flags);
2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717
	kmem_cache_free(extent_buffer_cache, eb);
}

struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree,
					  u64 start, unsigned long len,
					  struct page *page0,
					  gfp_t mask)
{
	unsigned long num_pages = num_extent_pages(start, len);
	unsigned long i;
	unsigned long index = start >> PAGE_CACHE_SHIFT;
	struct extent_buffer *eb;
2718
	struct extent_buffer *exists = NULL;
2719 2720 2721 2722
	struct page *p;
	struct address_space *mapping = tree->mapping;
	int uptodate = 1;

2723 2724 2725 2726 2727 2728 2729 2730 2731
	spin_lock(&tree->buffer_lock);
	eb = buffer_search(tree, start);
	if (eb) {
		atomic_inc(&eb->refs);
		spin_unlock(&tree->buffer_lock);
		return eb;
	}
	spin_unlock(&tree->buffer_lock);

2732
	eb = __alloc_extent_buffer(tree, start, len, mask);
2733
	if (!eb)
2734 2735 2736 2737 2738 2739 2740 2741 2742 2743
		return NULL;

	if (page0) {
		eb->first_page = page0;
		i = 1;
		index++;
		page_cache_get(page0);
		mark_page_accessed(page0);
		set_page_extent_mapped(page0);
		set_page_extent_head(page0, len);
2744
		uptodate = PageUptodate(page0);
2745 2746 2747 2748 2749 2750 2751
	} else {
		i = 0;
	}
	for (; i < num_pages; i++, index++) {
		p = find_or_create_page(mapping, index, mask | __GFP_HIGHMEM);
		if (!p) {
			WARN_ON(1);
2752
			goto free_eb;
2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769
		}
		set_page_extent_mapped(p);
		mark_page_accessed(p);
		if (i == 0) {
			eb->first_page = p;
			set_page_extent_head(p, len);
		} else {
			set_page_private(p, EXTENT_PAGE_PRIVATE);
		}
		if (!PageUptodate(p))
			uptodate = 0;
		unlock_page(p);
	}
	if (uptodate)
		eb->flags |= EXTENT_UPTODATE;
	eb->flags |= EXTENT_BUFFER_FILLED;

2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781
	spin_lock(&tree->buffer_lock);
	exists = buffer_tree_insert(tree, start, &eb->rb_node);
	if (exists) {
		/* add one reference for the caller */
		atomic_inc(&exists->refs);
		spin_unlock(&tree->buffer_lock);
		goto free_eb;
	}
	spin_unlock(&tree->buffer_lock);

	/* add one reference for the tree */
	atomic_inc(&eb->refs);
2782 2783
	return eb;

2784
free_eb:
2785
	if (!atomic_dec_and_test(&eb->refs))
2786 2787
		return exists;
	for (index = 1; index < i; index++)
2788
		page_cache_release(extent_buffer_page(eb, index));
2789
	page_cache_release(extent_buffer_page(eb, 0));
2790
	__free_extent_buffer(eb);
2791
	return exists;
2792 2793 2794 2795 2796 2797 2798 2799 2800
}
EXPORT_SYMBOL(alloc_extent_buffer);

struct extent_buffer *find_extent_buffer(struct extent_io_tree *tree,
					 u64 start, unsigned long len,
					  gfp_t mask)
{
	struct extent_buffer *eb;

2801 2802 2803 2804 2805
	spin_lock(&tree->buffer_lock);
	eb = buffer_search(tree, start);
	if (eb)
		atomic_inc(&eb->refs);
	spin_unlock(&tree->buffer_lock);
2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818

	return eb;
}
EXPORT_SYMBOL(find_extent_buffer);

void free_extent_buffer(struct extent_buffer *eb)
{
	if (!eb)
		return;

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

2819
	WARN_ON(1);
2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838
}
EXPORT_SYMBOL(free_extent_buffer);

int clear_extent_buffer_dirty(struct extent_io_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);
	num_pages = num_extent_pages(eb->start, eb->len);

	for (i = 0; i < num_pages; i++) {
		page = extent_buffer_page(eb, i);
2839
		lock_page(page);
2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856
		if (i == 0)
			set_page_extent_head(page, eb->len);
		else
			set_page_private(page, EXTENT_PAGE_PRIVATE);

		/*
		 * 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) & (PAGE_CACHE_SIZE - 1)))) {
			start = (u64)page->index << PAGE_CACHE_SHIFT;
			end  = start + PAGE_CACHE_SIZE - 1;
			if (test_range_bit(tree, start, end,
					   EXTENT_DIRTY, 0)) {
2857
				unlock_page(page);
2858 2859 2860 2861
				continue;
			}
		}
		clear_page_dirty_for_io(page);
2862 2863 2864
#if LINUX_VERSION_CODE > KERNEL_VERSION(2,6,26)
		spin_lock_irq(&page->mapping->tree_lock);
#else
2865
		read_lock_irq(&page->mapping->tree_lock);
2866
#endif
2867 2868 2869 2870 2871
		if (!PageDirty(page)) {
			radix_tree_tag_clear(&page->mapping->page_tree,
						page_index(page),
						PAGECACHE_TAG_DIRTY);
		}
2872 2873 2874
#if LINUX_VERSION_CODE > KERNEL_VERSION(2,6,26)
		spin_unlock_irq(&page->mapping->tree_lock);
#else
2875
		read_unlock_irq(&page->mapping->tree_lock);
2876
#endif
2877
		unlock_page(page);
2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905
	}
	return 0;
}
EXPORT_SYMBOL(clear_extent_buffer_dirty);

int wait_on_extent_buffer_writeback(struct extent_io_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_io_tree *tree,
			     struct extent_buffer *eb)
{
	unsigned long i;
	unsigned long num_pages;

	num_pages = num_extent_pages(eb->start, eb->len);
	for (i = 0; i < num_pages; i++) {
		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) {
2906
			lock_page(page);
2907 2908 2909
			set_page_extent_head(page, eb->len);
		} else if (PagePrivate(page) &&
			   page->private != EXTENT_PAGE_PRIVATE) {
2910
			lock_page(page);
2911
			set_page_extent_mapped(page);
2912
			unlock_page(page);
2913 2914
		}
		__set_page_dirty_nobuffers(extent_buffer_page(eb, i));
2915 2916
		if (i == 0)
			unlock_page(page);
2917 2918 2919 2920 2921 2922
	}
	return set_extent_dirty(tree, eb->start,
				eb->start + eb->len - 1, GFP_NOFS);
}
EXPORT_SYMBOL(set_extent_buffer_dirty);

2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936
int clear_extent_buffer_uptodate(struct extent_io_tree *tree,
				struct extent_buffer *eb)
{
	unsigned long i;
	struct page *page;
	unsigned long num_pages;

	num_pages = num_extent_pages(eb->start, eb->len);
	eb->flags &= ~EXTENT_UPTODATE;

	clear_extent_uptodate(tree, eb->start, eb->start + eb->len - 1,
			      GFP_NOFS);
	for (i = 0; i < num_pages; i++) {
		page = extent_buffer_page(eb, i);
C
Chris Mason 已提交
2937 2938
		if (page)
			ClearPageUptodate(page);
2939 2940 2941 2942
	}
	return 0;
}

2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960 2961 2962 2963 2964 2965 2966 2967
int set_extent_buffer_uptodate(struct extent_io_tree *tree,
				struct extent_buffer *eb)
{
	unsigned long i;
	struct page *page;
	unsigned long num_pages;

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

	set_extent_uptodate(tree, eb->start, eb->start + eb->len - 1,
			    GFP_NOFS);
	for (i = 0; i < num_pages; i++) {
		page = extent_buffer_page(eb, i);
		if ((i == 0 && (eb->start & (PAGE_CACHE_SIZE - 1))) ||
		    ((i == num_pages - 1) &&
		     ((eb->start + eb->len) & (PAGE_CACHE_SIZE - 1)))) {
			check_page_uptodate(tree, page);
			continue;
		}
		SetPageUptodate(page);
	}
	return 0;
}
EXPORT_SYMBOL(set_extent_buffer_uptodate);

2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993
int extent_range_uptodate(struct extent_io_tree *tree,
			  u64 start, u64 end)
{
	struct page *page;
	int ret;
	int pg_uptodate = 1;
	int uptodate;
	unsigned long index;

	ret = test_range_bit(tree, start, end, EXTENT_UPTODATE, 1);
	if (ret)
		return 1;
	while(start <= end) {
		index = start >> PAGE_CACHE_SHIFT;
		page = find_get_page(tree->mapping, index);
		uptodate = PageUptodate(page);
		page_cache_release(page);
		if (!uptodate) {
			pg_uptodate = 0;
			break;
		}
		start += PAGE_CACHE_SIZE;
	}
	return pg_uptodate;
}

2994
int extent_buffer_uptodate(struct extent_io_tree *tree,
2995
			   struct extent_buffer *eb)
2996
{
2997
	int ret = 0;
2998 2999
	unsigned long num_pages;
	unsigned long i;
3000 3001 3002
	struct page *page;
	int pg_uptodate = 1;

3003
	if (eb->flags & EXTENT_UPTODATE)
C
Chris Mason 已提交
3004
		return 1;
3005

C
Chris Mason 已提交
3006
	ret = test_range_bit(tree, eb->start, eb->start + eb->len - 1,
3007
			   EXTENT_UPTODATE, 1);
C
Chris Mason 已提交
3008 3009
	if (ret)
		return ret;
3010 3011 3012 3013 3014 3015 3016 3017 3018

	num_pages = num_extent_pages(eb->start, eb->len);
	for (i = 0; i < num_pages; i++) {
		page = extent_buffer_page(eb, i);
		if (!PageUptodate(page)) {
			pg_uptodate = 0;
			break;
		}
	}
C
Chris Mason 已提交
3019
	return pg_uptodate;
3020 3021 3022 3023 3024
}
EXPORT_SYMBOL(extent_buffer_uptodate);

int read_extent_buffer_pages(struct extent_io_tree *tree,
			     struct extent_buffer *eb,
3025
			     u64 start, int wait,
3026
			     get_extent_t *get_extent, int mirror_num)
3027 3028 3029 3030 3031 3032
{
	unsigned long i;
	unsigned long start_i;
	struct page *page;
	int err;
	int ret = 0;
3033 3034 3035
	int locked_pages = 0;
	int all_uptodate = 1;
	int inc_all_pages = 0;
3036
	unsigned long num_pages;
3037 3038
	struct bio *bio = NULL;

3039 3040 3041
	if (eb->flags & EXTENT_UPTODATE)
		return 0;

3042
	if (test_range_bit(tree, eb->start, eb->start + eb->len - 1,
3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058
			   EXTENT_UPTODATE, 1)) {
		return 0;
	}

	if (start) {
		WARN_ON(start < eb->start);
		start_i = (start >> PAGE_CACHE_SHIFT) -
			(eb->start >> PAGE_CACHE_SHIFT);
	} else {
		start_i = 0;
	}

	num_pages = num_extent_pages(eb->start, eb->len);
	for (i = start_i; i < num_pages; i++) {
		page = extent_buffer_page(eb, i);
		if (!wait) {
3059
			if (!trylock_page(page))
3060
				goto unlock_exit;
3061 3062 3063
		} else {
			lock_page(page);
		}
3064
		locked_pages++;
3065
		if (!PageUptodate(page)) {
3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081
			all_uptodate = 0;
		}
	}
	if (all_uptodate) {
		if (start_i == 0)
			eb->flags |= EXTENT_UPTODATE;
		goto unlock_exit;
	}

	for (i = start_i; i < num_pages; i++) {
		page = extent_buffer_page(eb, i);
		if (inc_all_pages)
			page_cache_get(page);
		if (!PageUptodate(page)) {
			if (start_i == 0)
				inc_all_pages = 1;
3082
			ClearPageError(page);
3083
			err = __extent_read_full_page(tree, page,
3084 3085
						      get_extent, &bio,
						      mirror_num);
3086 3087 3088 3089 3090 3091 3092 3093
			if (err) {
				ret = err;
			}
		} else {
			unlock_page(page);
		}
	}

3094
	if (bio)
3095
		submit_one_bio(READ, bio, mirror_num);
3096

3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109
	if (ret || !wait) {
		return ret;
	}
	for (i = start_i; i < num_pages; i++) {
		page = extent_buffer_page(eb, i);
		wait_on_page_locked(page);
		if (!PageUptodate(page)) {
			ret = -EIO;
		}
	}
	if (!ret)
		eb->flags |= EXTENT_UPTODATE;
	return ret;
3110 3111 3112 3113 3114 3115 3116 3117 3118 3119

unlock_exit:
	i = start_i;
	while(locked_pages > 0) {
		page = extent_buffer_page(eb, i);
		i++;
		unlock_page(page);
		locked_pages--;
	}
	return ret;
3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212 3213 3214 3215 3216 3217 3218 3219 3220 3221 3222 3223 3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 3234 3235 3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306 3307 3308 3309 3310 3311 3312 3313 3314 3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329 3330 3331 3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 3377 3378 3379 3380 3381 3382 3383 3384 3385 3386 3387 3388 3389 3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447 3448 3449 3450 3451 3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 3468 3469 3470 3471 3472 3473 3474 3475 3476 3477 3478 3479 3480 3481 3482 3483 3484 3485 3486 3487 3488 3489 3490 3491 3492
}
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;

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

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

	while(len > 0) {
		page = extent_buffer_page(eb, i);

		cur = min(len, (PAGE_CACHE_SIZE - offset));
		kaddr = kmap_atomic(page, KM_USER1);
		memcpy(dst, kaddr + offset, cur);
		kunmap_atomic(kaddr, KM_USER1);

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

int map_private_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)
{
	size_t offset = start & (PAGE_CACHE_SIZE - 1);
	char *kaddr;
	struct page *p;
	size_t start_offset = eb->start & ((u64)PAGE_CACHE_SIZE - 1);
	unsigned long i = (start_offset + start) >> PAGE_CACHE_SHIFT;
	unsigned long end_i = (start_offset + start + min_len - 1) >>
		PAGE_CACHE_SHIFT;

	if (i != end_i)
		return -EINVAL;

	if (i == 0) {
		offset = start_offset;
		*map_start = 0;
	} else {
		offset = 0;
		*map_start = ((u64)i << PAGE_CACHE_SHIFT) - start_offset;
	}
	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);
	}

	p = extent_buffer_page(eb, i);
	kaddr = kmap_atomic(p, km);
	*token = kaddr;
	*map = kaddr + offset;
	*map_len = PAGE_CACHE_SIZE - offset;
	return 0;
}
EXPORT_SYMBOL(map_private_extent_buffer);

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;
	}
	err = map_private_extent_buffer(eb, start, min_len, token, map,
				       map_start, map_len, km);
	if (!err && save) {
		eb->map_token = *token;
		eb->kaddr = *map;
		eb->map_start = *map_start;
		eb->map_len = *map_len;
	}
	return err;
}
EXPORT_SYMBOL(map_extent_buffer);

void unmap_extent_buffer(struct extent_buffer *eb, char *token, int km)
{
	kunmap_atomic(token, km);
}
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_offset + start) & ((unsigned long)PAGE_CACHE_SIZE - 1);

	while(len > 0) {
		page = extent_buffer_page(eb, i);

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

		kaddr = kmap_atomic(page, KM_USER0);
		ret = memcmp(ptr, kaddr + offset, cur);
		kunmap_atomic(kaddr, KM_USER0);
		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_offset + start) & ((unsigned long)PAGE_CACHE_SIZE - 1);

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

		cur = min(len, PAGE_CACHE_SIZE - offset);
		kaddr = kmap_atomic(page, KM_USER1);
		memcpy(kaddr + offset, src, cur);
		kunmap_atomic(kaddr, KM_USER1);

		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_offset + start) & ((unsigned long)PAGE_CACHE_SIZE - 1);

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

		cur = min(len, PAGE_CACHE_SIZE - offset);
		kaddr = kmap_atomic(page, KM_USER0);
		memset(kaddr + offset, c, cur);
		kunmap_atomic(kaddr, KM_USER0);

		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 = (start_offset + dst_offset) &
		((unsigned long)PAGE_CACHE_SIZE - 1);

	while(len > 0) {
		page = extent_buffer_page(dst, i);
		WARN_ON(!PageUptodate(page));

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

		kaddr = kmap_atomic(page, KM_USER0);
		read_extent_buffer(src, kaddr + offset, src_offset, cur);
		kunmap_atomic(kaddr, KM_USER0);

		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)
{
	char *dst_kaddr = kmap_atomic(dst_page, KM_USER0);
	if (dst_page == src_page) {
		memmove(dst_kaddr + dst_off, dst_kaddr + src_off, len);
	} else {
		char *src_kaddr = kmap_atomic(src_page, KM_USER1);
		char *p = dst_kaddr + dst_off + len;
		char *s = src_kaddr + src_off + len;

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

		kunmap_atomic(src_kaddr, KM_USER1);
	}
	kunmap_atomic(dst_kaddr, KM_USER0);
}

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

	if (dst_page != src_page)
		src_kaddr = kmap_atomic(src_page, KM_USER1);
	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 = (start_offset + dst_offset) &
			((unsigned long)PAGE_CACHE_SIZE - 1);
		src_off_in_page = (start_offset + 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;

		cur = min(len, (unsigned long)(PAGE_CACHE_SIZE -
					       src_off_in_page));
		cur = min_t(unsigned long, cur,
			(unsigned long)(PAGE_CACHE_SIZE - dst_off_in_page));

		copy_pages(extent_buffer_page(dst, dst_i),
			   extent_buffer_page(dst, src_i),
			   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 = (start_offset + dst_end) &
			((unsigned long)PAGE_CACHE_SIZE - 1);
		src_off_in_page = (start_offset + src_end) &
			((unsigned long)PAGE_CACHE_SIZE - 1);

		cur = min_t(unsigned long, len, src_off_in_page + 1);
		cur = min(cur, dst_off_in_page + 1);
		move_pages(extent_buffer_page(dst, dst_i),
			   extent_buffer_page(dst, src_i),
			   dst_off_in_page - cur + 1,
			   src_off_in_page - cur + 1, cur);

		dst_end -= cur;
		src_end -= cur;
		len -= cur;
	}
}
EXPORT_SYMBOL(memmove_extent_buffer);
3493 3494 3495 3496 3497 3498 3499 3500 3501 3502 3503 3504 3505 3506 3507 3508 3509 3510 3511 3512 3513 3514 3515 3516 3517 3518 3519 3520 3521 3522 3523

int try_release_extent_buffer(struct extent_io_tree *tree, struct page *page)
{
	u64 start = page_offset(page);
	struct extent_buffer *eb;
	int ret = 1;
	unsigned long i;
	unsigned long num_pages;

	spin_lock(&tree->buffer_lock);
	eb = buffer_search(tree, start);
	if (!eb)
		goto out;

	if (atomic_read(&eb->refs) > 1) {
		ret = 0;
		goto out;
	}
	/* at this point we can safely release the extent buffer */
	num_pages = num_extent_pages(eb->start, eb->len);
	for (i = 0; i < num_pages; i++) {
		struct page *page = extent_buffer_page(eb, i);
		page_cache_release(page);
	}
	rb_erase(&eb->rb_node, &tree->buffer);
	__free_extent_buffer(eb);
out:
	spin_unlock(&tree->buffer_lock);
	return ret;
}
EXPORT_SYMBOL(try_release_extent_buffer);