Blame view

fs/btrfs/extent_io.c 109 KB
d1310b2e0   Chris Mason   Btrfs: Split the ...
1
2
3
4
  #include <linux/bitops.h>
  #include <linux/slab.h>
  #include <linux/bio.h>
  #include <linux/mm.h>
d1310b2e0   Chris Mason   Btrfs: Split the ...
5
6
7
8
9
10
  #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>
d1310b2e0   Chris Mason   Btrfs: Split the ...
11
12
  #include <linux/writeback.h>
  #include <linux/pagevec.h>
268bb0ce3   Linus Torvalds   sanitize <linux/p...
13
  #include <linux/prefetch.h>
90a887c9a   Dan Magenheimer   btrfs: add cleanc...
14
  #include <linux/cleancache.h>
d1310b2e0   Chris Mason   Btrfs: Split the ...
15
16
  #include "extent_io.h"
  #include "extent_map.h"
2db04966a   David Woodhouse   Btrfs: Change Tes...
17
  #include "compat.h"
902b22f34   David Woodhouse   Btrfs: Remove bro...
18
19
  #include "ctree.h"
  #include "btrfs_inode.h"
4a54c8c16   Jan Schmidt   btrfs: Moved repa...
20
  #include "volumes.h"
d1310b2e0   Chris Mason   Btrfs: Split the ...
21

d1310b2e0   Chris Mason   Btrfs: Split the ...
22
23
24
25
26
  static struct kmem_cache *extent_state_cache;
  static struct kmem_cache *extent_buffer_cache;
  
  static LIST_HEAD(buffers);
  static LIST_HEAD(states);
4bef08485   Chris Mason   Btrfs: Tree loggi...
27

b47eda869   Chris Mason   Btrfs: Turn off e...
28
  #define LEAK_DEBUG 0
3935127c5   Chris Mason   Btrfs: disable le...
29
  #if LEAK_DEBUG
d397712bc   Chris Mason   Btrfs: Fix checkp...
30
  static DEFINE_SPINLOCK(leak_lock);
4bef08485   Chris Mason   Btrfs: Tree loggi...
31
  #endif
d1310b2e0   Chris Mason   Btrfs: Split the ...
32

d1310b2e0   Chris Mason   Btrfs: Split the ...
33
34
35
36
37
  #define BUFFER_LRU_MAX 64
  
  struct tree_entry {
  	u64 start;
  	u64 end;
d1310b2e0   Chris Mason   Btrfs: Split the ...
38
39
40
41
42
43
44
  	struct rb_node rb_node;
  };
  
  struct extent_page_data {
  	struct bio *bio;
  	struct extent_io_tree *tree;
  	get_extent_t *get_extent;
771ed689d   Chris Mason   Btrfs: Optimize c...
45
46
47
48
  
  	/* tells writepage not to lock the state bits for this range
  	 * it still does the unlocking
  	 */
ffbd517d5   Chris Mason   Btrfs: use WRITE_...
49
50
51
52
  	unsigned int extent_locked:1;
  
  	/* tells the submit_bio code to use a WRITE_SYNC */
  	unsigned int sync_io:1;
d1310b2e0   Chris Mason   Btrfs: Split the ...
53
54
55
56
  };
  
  int __init extent_io_init(void)
  {
9601e3f63   Christoph Hellwig   Btrfs: kill btrfs...
57
58
59
  	extent_state_cache = kmem_cache_create("extent_state",
  			sizeof(struct extent_state), 0,
  			SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
d1310b2e0   Chris Mason   Btrfs: Split the ...
60
61
  	if (!extent_state_cache)
  		return -ENOMEM;
9601e3f63   Christoph Hellwig   Btrfs: kill btrfs...
62
63
64
  	extent_buffer_cache = kmem_cache_create("extent_buffers",
  			sizeof(struct extent_buffer), 0,
  			SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
d1310b2e0   Chris Mason   Btrfs: Split the ...
65
66
67
68
69
70
71
72
73
74
75
76
  	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;
2d2ae5479   Chris Mason   Btrfs: Add leak d...
77
  	struct extent_buffer *eb;
d1310b2e0   Chris Mason   Btrfs: Split the ...
78
79
  
  	while (!list_empty(&states)) {
2d2ae5479   Chris Mason   Btrfs: Add leak d...
80
  		state = list_entry(states.next, struct extent_state, leak_list);
d397712bc   Chris Mason   Btrfs: Fix checkp...
81
82
83
84
85
86
  		printk(KERN_ERR "btrfs state leak: start %llu end %llu "
  		       "state %lu in tree %p refs %d
  ",
  		       (unsigned long long)state->start,
  		       (unsigned long long)state->end,
  		       state->state, state->tree, atomic_read(&state->refs));
2d2ae5479   Chris Mason   Btrfs: Add leak d...
87
  		list_del(&state->leak_list);
d1310b2e0   Chris Mason   Btrfs: Split the ...
88
89
90
  		kmem_cache_free(extent_state_cache, state);
  
  	}
2d2ae5479   Chris Mason   Btrfs: Add leak d...
91
92
  	while (!list_empty(&buffers)) {
  		eb = list_entry(buffers.next, struct extent_buffer, leak_list);
d397712bc   Chris Mason   Btrfs: Fix checkp...
93
94
95
96
  		printk(KERN_ERR "btrfs buffer leak start %llu len %lu "
  		       "refs %d
  ", (unsigned long long)eb->start,
  		       eb->len, atomic_read(&eb->refs));
2d2ae5479   Chris Mason   Btrfs: Add leak d...
97
98
99
  		list_del(&eb->leak_list);
  		kmem_cache_free(extent_buffer_cache, eb);
  	}
d1310b2e0   Chris Mason   Btrfs: Split the ...
100
101
102
103
104
105
106
  	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,
f993c883a   David Sterba   btrfs: drop unuse...
107
  			 struct address_space *mapping)
d1310b2e0   Chris Mason   Btrfs: Split the ...
108
  {
6bef4d317   Eric Paris   Btrfs: use RB_ROO...
109
  	tree->state = RB_ROOT;
19fe0a8b7   Miao Xie   Btrfs: Switch the...
110
  	INIT_RADIX_TREE(&tree->buffer, GFP_ATOMIC);
d1310b2e0   Chris Mason   Btrfs: Split the ...
111
112
  	tree->ops = NULL;
  	tree->dirty_bytes = 0;
70dec8079   Chris Mason   Btrfs: extent_io ...
113
  	spin_lock_init(&tree->lock);
6af118ce5   Chris Mason   Btrfs: Index exte...
114
  	spin_lock_init(&tree->buffer_lock);
d1310b2e0   Chris Mason   Btrfs: Split the ...
115
  	tree->mapping = mapping;
d1310b2e0   Chris Mason   Btrfs: Split the ...
116
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
117

b2950863c   Christoph Hellwig   Btrfs: make thing...
118
  static struct extent_state *alloc_extent_state(gfp_t mask)
d1310b2e0   Chris Mason   Btrfs: Split the ...
119
120
  {
  	struct extent_state *state;
3935127c5   Chris Mason   Btrfs: disable le...
121
  #if LEAK_DEBUG
2d2ae5479   Chris Mason   Btrfs: Add leak d...
122
  	unsigned long flags;
4bef08485   Chris Mason   Btrfs: Tree loggi...
123
  #endif
d1310b2e0   Chris Mason   Btrfs: Split the ...
124
125
  
  	state = kmem_cache_alloc(extent_state_cache, mask);
2b114d1d3   Peter   Btrfs: Correct us...
126
  	if (!state)
d1310b2e0   Chris Mason   Btrfs: Split the ...
127
128
  		return state;
  	state->state = 0;
d1310b2e0   Chris Mason   Btrfs: Split the ...
129
  	state->private = 0;
70dec8079   Chris Mason   Btrfs: extent_io ...
130
  	state->tree = NULL;
3935127c5   Chris Mason   Btrfs: disable le...
131
  #if LEAK_DEBUG
2d2ae5479   Chris Mason   Btrfs: Add leak d...
132
133
134
  	spin_lock_irqsave(&leak_lock, flags);
  	list_add(&state->leak_list, &states);
  	spin_unlock_irqrestore(&leak_lock, flags);
4bef08485   Chris Mason   Btrfs: Tree loggi...
135
  #endif
d1310b2e0   Chris Mason   Btrfs: Split the ...
136
137
138
139
  	atomic_set(&state->refs, 1);
  	init_waitqueue_head(&state->wq);
  	return state;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
140

4845e44ff   Chris Mason   Btrfs: rework O_D...
141
  void free_extent_state(struct extent_state *state)
d1310b2e0   Chris Mason   Btrfs: Split the ...
142
  {
d1310b2e0   Chris Mason   Btrfs: Split the ...
143
144
145
  	if (!state)
  		return;
  	if (atomic_dec_and_test(&state->refs)) {
3935127c5   Chris Mason   Btrfs: disable le...
146
  #if LEAK_DEBUG
2d2ae5479   Chris Mason   Btrfs: Add leak d...
147
  		unsigned long flags;
4bef08485   Chris Mason   Btrfs: Tree loggi...
148
  #endif
70dec8079   Chris Mason   Btrfs: extent_io ...
149
  		WARN_ON(state->tree);
3935127c5   Chris Mason   Btrfs: disable le...
150
  #if LEAK_DEBUG
2d2ae5479   Chris Mason   Btrfs: Add leak d...
151
152
153
  		spin_lock_irqsave(&leak_lock, flags);
  		list_del(&state->leak_list);
  		spin_unlock_irqrestore(&leak_lock, flags);
4bef08485   Chris Mason   Btrfs: Tree loggi...
154
  #endif
d1310b2e0   Chris Mason   Btrfs: Split the ...
155
156
157
  		kmem_cache_free(extent_state_cache, state);
  	}
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
158
159
160
161
  
  static struct rb_node *tree_insert(struct rb_root *root, u64 offset,
  				   struct rb_node *node)
  {
d397712bc   Chris Mason   Btrfs: Fix checkp...
162
163
  	struct rb_node **p = &root->rb_node;
  	struct rb_node *parent = NULL;
d1310b2e0   Chris Mason   Btrfs: Split the ...
164
  	struct tree_entry *entry;
d397712bc   Chris Mason   Btrfs: Fix checkp...
165
  	while (*p) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
166
167
168
169
170
171
172
173
174
175
176
177
  		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);
d1310b2e0   Chris Mason   Btrfs: Split the ...
178
179
180
181
  	rb_link_node(node, parent, p);
  	rb_insert_color(node, root);
  	return NULL;
  }
80ea96b1f   Chris Mason   Btrfs: Add a look...
182
  static struct rb_node *__etree_search(struct extent_io_tree *tree, u64 offset,
d1310b2e0   Chris Mason   Btrfs: Split the ...
183
184
185
  				     struct rb_node **prev_ret,
  				     struct rb_node **next_ret)
  {
80ea96b1f   Chris Mason   Btrfs: Add a look...
186
  	struct rb_root *root = &tree->state;
d397712bc   Chris Mason   Btrfs: Fix checkp...
187
  	struct rb_node *n = root->rb_node;
d1310b2e0   Chris Mason   Btrfs: Split the ...
188
189
190
191
  	struct rb_node *prev = NULL;
  	struct rb_node *orig_prev = NULL;
  	struct tree_entry *entry;
  	struct tree_entry *prev_entry = NULL;
d397712bc   Chris Mason   Btrfs: Fix checkp...
192
  	while (n) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
193
194
195
196
197
198
199
200
  		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;
d397712bc   Chris Mason   Btrfs: Fix checkp...
201
  		else
d1310b2e0   Chris Mason   Btrfs: Split the ...
202
203
204
205
206
  			return n;
  	}
  
  	if (prev_ret) {
  		orig_prev = prev;
d397712bc   Chris Mason   Btrfs: Fix checkp...
207
  		while (prev && offset > prev_entry->end) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
208
209
210
211
212
213
214
215
216
  			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);
d397712bc   Chris Mason   Btrfs: Fix checkp...
217
  		while (prev && offset < prev_entry->start) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
218
219
220
221
222
223
224
  			prev = rb_prev(prev);
  			prev_entry = rb_entry(prev, struct tree_entry, rb_node);
  		}
  		*next_ret = prev;
  	}
  	return NULL;
  }
80ea96b1f   Chris Mason   Btrfs: Add a look...
225
226
  static inline struct rb_node *tree_search(struct extent_io_tree *tree,
  					  u64 offset)
d1310b2e0   Chris Mason   Btrfs: Split the ...
227
  {
70dec8079   Chris Mason   Btrfs: extent_io ...
228
  	struct rb_node *prev = NULL;
d1310b2e0   Chris Mason   Btrfs: Split the ...
229
  	struct rb_node *ret;
70dec8079   Chris Mason   Btrfs: extent_io ...
230

80ea96b1f   Chris Mason   Btrfs: Add a look...
231
  	ret = __etree_search(tree, offset, &prev, NULL);
d397712bc   Chris Mason   Btrfs: Fix checkp...
232
  	if (!ret)
d1310b2e0   Chris Mason   Btrfs: Split the ...
233
234
235
  		return prev;
  	return ret;
  }
9ed74f2db   Josef Bacik   Btrfs: proper -EN...
236
237
238
239
240
241
242
  static void merge_cb(struct extent_io_tree *tree, struct extent_state *new,
  		     struct extent_state *other)
  {
  	if (tree->ops && tree->ops->merge_extent_hook)
  		tree->ops->merge_extent_hook(tree->mapping->host, new,
  					     other);
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
243
244
245
246
247
248
249
250
251
  /*
   * 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.
   */
1bf85046e   Jeff Mahoney   btrfs: Make exten...
252
253
  static void merge_state(struct extent_io_tree *tree,
  		        struct extent_state *state)
d1310b2e0   Chris Mason   Btrfs: Split the ...
254
255
256
  {
  	struct extent_state *other;
  	struct rb_node *other_node;
5b21f2ed3   Zheng Yan   Btrfs: extent_map...
257
  	if (state->state & (EXTENT_IOBITS | EXTENT_BOUNDARY))
1bf85046e   Jeff Mahoney   btrfs: Make exten...
258
  		return;
d1310b2e0   Chris Mason   Btrfs: Split the ...
259
260
261
262
263
264
  
  	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) {
9ed74f2db   Josef Bacik   Btrfs: proper -EN...
265
  			merge_cb(tree, state, other);
d1310b2e0   Chris Mason   Btrfs: Split the ...
266
  			state->start = other->start;
70dec8079   Chris Mason   Btrfs: extent_io ...
267
  			other->tree = NULL;
d1310b2e0   Chris Mason   Btrfs: Split the ...
268
269
270
271
272
273
274
275
276
  			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) {
9ed74f2db   Josef Bacik   Btrfs: proper -EN...
277
  			merge_cb(tree, state, other);
df98b6e2c   Josef Bacik   Btrfs: fix how we...
278
279
280
281
  			state->end = other->end;
  			other->tree = NULL;
  			rb_erase(&other->rb_node, &tree->state);
  			free_extent_state(other);
d1310b2e0   Chris Mason   Btrfs: Split the ...
282
283
  		}
  	}
d1310b2e0   Chris Mason   Btrfs: Split the ...
284
  }
1bf85046e   Jeff Mahoney   btrfs: Make exten...
285
  static void set_state_cb(struct extent_io_tree *tree,
0ca1f7ceb   Yan, Zheng   Btrfs: Update met...
286
  			 struct extent_state *state, int *bits)
291d673e6   Chris Mason   Btrfs: Do delallo...
287
  {
1bf85046e   Jeff Mahoney   btrfs: Make exten...
288
289
  	if (tree->ops && tree->ops->set_bit_hook)
  		tree->ops->set_bit_hook(tree->mapping->host, state, bits);
291d673e6   Chris Mason   Btrfs: Do delallo...
290
291
292
  }
  
  static void clear_state_cb(struct extent_io_tree *tree,
0ca1f7ceb   Yan, Zheng   Btrfs: Update met...
293
  			   struct extent_state *state, int *bits)
291d673e6   Chris Mason   Btrfs: Do delallo...
294
  {
9ed74f2db   Josef Bacik   Btrfs: proper -EN...
295
296
  	if (tree->ops && tree->ops->clear_bit_hook)
  		tree->ops->clear_bit_hook(tree->mapping->host, state, bits);
291d673e6   Chris Mason   Btrfs: Do delallo...
297
  }
3150b6996   Xiao Guangrong   Btrfs: clean up f...
298
299
  static void set_state_bits(struct extent_io_tree *tree,
  			   struct extent_state *state, int *bits);
d1310b2e0   Chris Mason   Btrfs: Split the ...
300
301
302
303
304
305
306
307
308
309
310
311
  /*
   * 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,
0ca1f7ceb   Yan, Zheng   Btrfs: Update met...
312
  			int *bits)
d1310b2e0   Chris Mason   Btrfs: Split the ...
313
314
315
316
  {
  	struct rb_node *node;
  
  	if (end < start) {
d397712bc   Chris Mason   Btrfs: Fix checkp...
317
318
319
320
  		printk(KERN_ERR "btrfs end < start %llu %llu
  ",
  		       (unsigned long long)end,
  		       (unsigned long long)start);
d1310b2e0   Chris Mason   Btrfs: Split the ...
321
322
  		WARN_ON(1);
  	}
d1310b2e0   Chris Mason   Btrfs: Split the ...
323
324
  	state->start = start;
  	state->end = end;
9ed74f2db   Josef Bacik   Btrfs: proper -EN...
325

3150b6996   Xiao Guangrong   Btrfs: clean up f...
326
  	set_state_bits(tree, state, bits);
d1310b2e0   Chris Mason   Btrfs: Split the ...
327
328
329
330
  	node = tree_insert(&tree->state, end, &state->rb_node);
  	if (node) {
  		struct extent_state *found;
  		found = rb_entry(node, struct extent_state, rb_node);
d397712bc   Chris Mason   Btrfs: Fix checkp...
331
332
333
334
335
  		printk(KERN_ERR "btrfs found node %llu %llu on insert of "
  		       "%llu %llu
  ", (unsigned long long)found->start,
  		       (unsigned long long)found->end,
  		       (unsigned long long)start, (unsigned long long)end);
d1310b2e0   Chris Mason   Btrfs: Split the ...
336
337
  		return -EEXIST;
  	}
70dec8079   Chris Mason   Btrfs: extent_io ...
338
  	state->tree = tree;
d1310b2e0   Chris Mason   Btrfs: Split the ...
339
340
341
  	merge_state(tree, state);
  	return 0;
  }
1bf85046e   Jeff Mahoney   btrfs: Make exten...
342
  static void split_cb(struct extent_io_tree *tree, struct extent_state *orig,
9ed74f2db   Josef Bacik   Btrfs: proper -EN...
343
344
345
  		     u64 split)
  {
  	if (tree->ops && tree->ops->split_extent_hook)
1bf85046e   Jeff Mahoney   btrfs: Make exten...
346
  		tree->ops->split_extent_hook(tree->mapping->host, orig, split);
9ed74f2db   Josef Bacik   Btrfs: proper -EN...
347
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
  /*
   * 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;
9ed74f2db   Josef Bacik   Btrfs: proper -EN...
366
367
  
  	split_cb(tree, orig, split);
d1310b2e0   Chris Mason   Btrfs: Split the ...
368
369
370
371
372
373
374
  	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) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
375
376
377
  		free_extent_state(prealloc);
  		return -EEXIST;
  	}
70dec8079   Chris Mason   Btrfs: extent_io ...
378
  	prealloc->tree = tree;
d1310b2e0   Chris Mason   Btrfs: Split the ...
379
380
381
382
383
384
385
386
387
388
389
390
  	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,
0ca1f7ceb   Yan, Zheng   Btrfs: Update met...
391
392
  			    struct extent_state *state,
  			    int *bits, int wake)
d1310b2e0   Chris Mason   Btrfs: Split the ...
393
  {
0ca1f7ceb   Yan, Zheng   Btrfs: Update met...
394
  	int bits_to_clear = *bits & ~EXTENT_CTLBITS;
32c00aff7   Josef Bacik   Btrfs: release de...
395
  	int ret = state->state & bits_to_clear;
d1310b2e0   Chris Mason   Btrfs: Split the ...
396

0ca1f7ceb   Yan, Zheng   Btrfs: Update met...
397
  	if ((bits_to_clear & EXTENT_DIRTY) && (state->state & EXTENT_DIRTY)) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
398
399
400
401
  		u64 range = state->end - state->start + 1;
  		WARN_ON(range > tree->dirty_bytes);
  		tree->dirty_bytes -= range;
  	}
291d673e6   Chris Mason   Btrfs: Do delallo...
402
  	clear_state_cb(tree, state, bits);
32c00aff7   Josef Bacik   Btrfs: release de...
403
  	state->state &= ~bits_to_clear;
d1310b2e0   Chris Mason   Btrfs: Split the ...
404
405
  	if (wake)
  		wake_up(&state->wq);
0ca1f7ceb   Yan, Zheng   Btrfs: Update met...
406
  	if (state->state == 0) {
70dec8079   Chris Mason   Btrfs: extent_io ...
407
  		if (state->tree) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
408
  			rb_erase(&state->rb_node, &tree->state);
70dec8079   Chris Mason   Btrfs: extent_io ...
409
  			state->tree = NULL;
d1310b2e0   Chris Mason   Btrfs: Split the ...
410
411
412
413
414
415
416
417
418
  			free_extent_state(state);
  		} else {
  			WARN_ON(1);
  		}
  	} else {
  		merge_state(tree, state);
  	}
  	return ret;
  }
8233767a2   Xiao Guangrong   Btrfs: allocate e...
419
420
421
422
423
424
425
426
  static struct extent_state *
  alloc_extent_state_atomic(struct extent_state *prealloc)
  {
  	if (!prealloc)
  		prealloc = alloc_extent_state(GFP_ATOMIC);
  
  	return prealloc;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
427
428
429
430
431
432
433
434
435
436
437
438
439
440
  /*
   * 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,
2c64c53d8   Chris Mason   Btrfs: cache valu...
441
442
443
  		     int bits, int wake, int delete,
  		     struct extent_state **cached_state,
  		     gfp_t mask)
d1310b2e0   Chris Mason   Btrfs: Split the ...
444
445
  {
  	struct extent_state *state;
2c64c53d8   Chris Mason   Btrfs: cache valu...
446
  	struct extent_state *cached;
d1310b2e0   Chris Mason   Btrfs: Split the ...
447
  	struct extent_state *prealloc = NULL;
2c64c53d8   Chris Mason   Btrfs: cache valu...
448
  	struct rb_node *next_node;
d1310b2e0   Chris Mason   Btrfs: Split the ...
449
  	struct rb_node *node;
5c939df56   Yan Zheng   btrfs: Fix set/cl...
450
  	u64 last_end;
d1310b2e0   Chris Mason   Btrfs: Split the ...
451
452
  	int err;
  	int set = 0;
2ac55d41b   Josef Bacik   Btrfs: cache the ...
453
  	int clear = 0;
d1310b2e0   Chris Mason   Btrfs: Split the ...
454

0ca1f7ceb   Yan, Zheng   Btrfs: Update met...
455
456
457
  	if (delete)
  		bits |= ~EXTENT_CTLBITS;
  	bits |= EXTENT_FIRST_DELALLOC;
2ac55d41b   Josef Bacik   Btrfs: cache the ...
458
459
  	if (bits & (EXTENT_IOBITS | EXTENT_BOUNDARY))
  		clear = 1;
d1310b2e0   Chris Mason   Btrfs: Split the ...
460
461
462
463
464
465
  again:
  	if (!prealloc && (mask & __GFP_WAIT)) {
  		prealloc = alloc_extent_state(mask);
  		if (!prealloc)
  			return -ENOMEM;
  	}
cad321ad5   Chris Mason   Btrfs: shift all ...
466
  	spin_lock(&tree->lock);
2c64c53d8   Chris Mason   Btrfs: cache valu...
467
468
  	if (cached_state) {
  		cached = *cached_state;
2ac55d41b   Josef Bacik   Btrfs: cache the ...
469
470
471
472
473
  
  		if (clear) {
  			*cached_state = NULL;
  			cached_state = NULL;
  		}
df98b6e2c   Josef Bacik   Btrfs: fix how we...
474
475
  		if (cached && cached->tree && cached->start <= start &&
  		    cached->end > start) {
2ac55d41b   Josef Bacik   Btrfs: cache the ...
476
477
  			if (clear)
  				atomic_dec(&cached->refs);
2c64c53d8   Chris Mason   Btrfs: cache valu...
478
  			state = cached;
42daec299   Chris Mason   Btrfs: fix errors...
479
  			goto hit_next;
2c64c53d8   Chris Mason   Btrfs: cache valu...
480
  		}
2ac55d41b   Josef Bacik   Btrfs: cache the ...
481
482
  		if (clear)
  			free_extent_state(cached);
2c64c53d8   Chris Mason   Btrfs: cache valu...
483
  	}
d1310b2e0   Chris Mason   Btrfs: Split the ...
484
485
486
487
  	/*
  	 * this search will find the extents that end after
  	 * our range starts
  	 */
80ea96b1f   Chris Mason   Btrfs: Add a look...
488
  	node = tree_search(tree, start);
d1310b2e0   Chris Mason   Btrfs: Split the ...
489
490
491
  	if (!node)
  		goto out;
  	state = rb_entry(node, struct extent_state, rb_node);
2c64c53d8   Chris Mason   Btrfs: cache valu...
492
  hit_next:
d1310b2e0   Chris Mason   Btrfs: Split the ...
493
494
495
  	if (state->start > end)
  		goto out;
  	WARN_ON(state->end < start);
5c939df56   Yan Zheng   btrfs: Fix set/cl...
496
  	last_end = state->end;
d1310b2e0   Chris Mason   Btrfs: Split the ...
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
  
  	/*
  	 *     | ---- 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) {
8233767a2   Xiao Guangrong   Btrfs: allocate e...
515
516
  		prealloc = alloc_extent_state_atomic(prealloc);
  		BUG_ON(!prealloc);
d1310b2e0   Chris Mason   Btrfs: Split the ...
517
518
519
520
521
522
  		err = split_state(tree, state, prealloc, start);
  		BUG_ON(err == -EEXIST);
  		prealloc = NULL;
  		if (err)
  			goto out;
  		if (state->end <= end) {
0ca1f7ceb   Yan, Zheng   Btrfs: Update met...
523
  			set |= clear_state_bit(tree, state, &bits, wake);
5c939df56   Yan Zheng   btrfs: Fix set/cl...
524
525
526
  			if (last_end == (u64)-1)
  				goto out;
  			start = last_end + 1;
d1310b2e0   Chris Mason   Btrfs: Split the ...
527
528
529
530
531
532
533
534
535
536
  		}
  		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) {
8233767a2   Xiao Guangrong   Btrfs: allocate e...
537
538
  		prealloc = alloc_extent_state_atomic(prealloc);
  		BUG_ON(!prealloc);
d1310b2e0   Chris Mason   Btrfs: Split the ...
539
540
  		err = split_state(tree, state, prealloc, end + 1);
  		BUG_ON(err == -EEXIST);
d1310b2e0   Chris Mason   Btrfs: Split the ...
541
542
  		if (wake)
  			wake_up(&state->wq);
42daec299   Chris Mason   Btrfs: fix errors...
543

0ca1f7ceb   Yan, Zheng   Btrfs: Update met...
544
  		set |= clear_state_bit(tree, prealloc, &bits, wake);
9ed74f2db   Josef Bacik   Btrfs: proper -EN...
545

d1310b2e0   Chris Mason   Btrfs: Split the ...
546
547
548
  		prealloc = NULL;
  		goto out;
  	}
42daec299   Chris Mason   Btrfs: fix errors...
549

2c64c53d8   Chris Mason   Btrfs: cache valu...
550
551
552
553
  	if (state->end < end && prealloc && !need_resched())
  		next_node = rb_next(&state->rb_node);
  	else
  		next_node = NULL;
42daec299   Chris Mason   Btrfs: fix errors...
554

0ca1f7ceb   Yan, Zheng   Btrfs: Update met...
555
  	set |= clear_state_bit(tree, state, &bits, wake);
5c939df56   Yan Zheng   btrfs: Fix set/cl...
556
557
558
  	if (last_end == (u64)-1)
  		goto out;
  	start = last_end + 1;
2c64c53d8   Chris Mason   Btrfs: cache valu...
559
560
561
562
563
564
  	if (start <= end && next_node) {
  		state = rb_entry(next_node, struct extent_state,
  				 rb_node);
  		if (state->start == start)
  			goto hit_next;
  	}
d1310b2e0   Chris Mason   Btrfs: Split the ...
565
566
567
  	goto search_again;
  
  out:
cad321ad5   Chris Mason   Btrfs: shift all ...
568
  	spin_unlock(&tree->lock);
d1310b2e0   Chris Mason   Btrfs: Split the ...
569
570
571
572
573
574
575
576
  	if (prealloc)
  		free_extent_state(prealloc);
  
  	return set;
  
  search_again:
  	if (start > end)
  		goto out;
cad321ad5   Chris Mason   Btrfs: shift all ...
577
  	spin_unlock(&tree->lock);
d1310b2e0   Chris Mason   Btrfs: Split the ...
578
579
580
581
  	if (mask & __GFP_WAIT)
  		cond_resched();
  	goto again;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
582
583
584
  
  static int wait_on_state(struct extent_io_tree *tree,
  			 struct extent_state *state)
641f5219f   Christoph Hellwig   Btrfs: sparse loc...
585
586
  		__releases(tree->lock)
  		__acquires(tree->lock)
d1310b2e0   Chris Mason   Btrfs: Split the ...
587
588
589
  {
  	DEFINE_WAIT(wait);
  	prepare_to_wait(&state->wq, &wait, TASK_UNINTERRUPTIBLE);
cad321ad5   Chris Mason   Btrfs: shift all ...
590
  	spin_unlock(&tree->lock);
d1310b2e0   Chris Mason   Btrfs: Split the ...
591
  	schedule();
cad321ad5   Chris Mason   Btrfs: shift all ...
592
  	spin_lock(&tree->lock);
d1310b2e0   Chris Mason   Btrfs: Split the ...
593
594
595
596
597
598
599
600
601
602
603
604
605
  	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;
cad321ad5   Chris Mason   Btrfs: shift all ...
606
  	spin_lock(&tree->lock);
d1310b2e0   Chris Mason   Btrfs: Split the ...
607
608
609
610
611
612
  again:
  	while (1) {
  		/*
  		 * this search will find all the extents that end after
  		 * our range starts
  		 */
80ea96b1f   Chris Mason   Btrfs: Add a look...
613
  		node = tree_search(tree, start);
d1310b2e0   Chris Mason   Btrfs: Split the ...
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
  		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;
ded91f081   Xiao Guangrong   Btrfs: clean up f...
633
  		cond_resched_lock(&tree->lock);
d1310b2e0   Chris Mason   Btrfs: Split the ...
634
635
  	}
  out:
cad321ad5   Chris Mason   Btrfs: shift all ...
636
  	spin_unlock(&tree->lock);
d1310b2e0   Chris Mason   Btrfs: Split the ...
637
638
  	return 0;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
639

1bf85046e   Jeff Mahoney   btrfs: Make exten...
640
  static void set_state_bits(struct extent_io_tree *tree,
d1310b2e0   Chris Mason   Btrfs: Split the ...
641
  			   struct extent_state *state,
0ca1f7ceb   Yan, Zheng   Btrfs: Update met...
642
  			   int *bits)
d1310b2e0   Chris Mason   Btrfs: Split the ...
643
  {
0ca1f7ceb   Yan, Zheng   Btrfs: Update met...
644
  	int bits_to_set = *bits & ~EXTENT_CTLBITS;
9ed74f2db   Josef Bacik   Btrfs: proper -EN...
645

1bf85046e   Jeff Mahoney   btrfs: Make exten...
646
  	set_state_cb(tree, state, bits);
0ca1f7ceb   Yan, Zheng   Btrfs: Update met...
647
  	if ((bits_to_set & EXTENT_DIRTY) && !(state->state & EXTENT_DIRTY)) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
648
649
650
  		u64 range = state->end - state->start + 1;
  		tree->dirty_bytes += range;
  	}
0ca1f7ceb   Yan, Zheng   Btrfs: Update met...
651
  	state->state |= bits_to_set;
d1310b2e0   Chris Mason   Btrfs: Split the ...
652
  }
2c64c53d8   Chris Mason   Btrfs: cache valu...
653
654
655
656
657
658
659
660
661
662
  static void cache_state(struct extent_state *state,
  			struct extent_state **cached_ptr)
  {
  	if (cached_ptr && !(*cached_ptr)) {
  		if (state->state & (EXTENT_IOBITS | EXTENT_BOUNDARY)) {
  			*cached_ptr = state;
  			atomic_inc(&state->refs);
  		}
  	}
  }
507903b81   Arne Jansen   btrfs: using cach...
663
664
665
666
  static void uncache_state(struct extent_state **cached_ptr)
  {
  	if (cached_ptr && (*cached_ptr)) {
  		struct extent_state *state = *cached_ptr;
109b36a2b   Chris Mason   Btrfs: make uncac...
667
668
  		*cached_ptr = NULL;
  		free_extent_state(state);
507903b81   Arne Jansen   btrfs: using cach...
669
670
  	}
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
671
  /*
1edbb734b   Chris Mason   Btrfs: reduce CPU...
672
673
   * 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.
d1310b2e0   Chris Mason   Btrfs: Split the ...
674
   *
1edbb734b   Chris Mason   Btrfs: reduce CPU...
675
676
677
   * If any of the exclusive bits are set, 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.
d1310b2e0   Chris Mason   Btrfs: Split the ...
678
   *
1edbb734b   Chris Mason   Btrfs: reduce CPU...
679
   * [start, end] is inclusive This takes the tree lock.
d1310b2e0   Chris Mason   Btrfs: Split the ...
680
   */
1edbb734b   Chris Mason   Btrfs: reduce CPU...
681

4845e44ff   Chris Mason   Btrfs: rework O_D...
682
683
684
  int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
  		   int bits, int exclusive_bits, u64 *failed_start,
  		   struct extent_state **cached_state, gfp_t mask)
d1310b2e0   Chris Mason   Btrfs: Split the ...
685
686
687
688
  {
  	struct extent_state *state;
  	struct extent_state *prealloc = NULL;
  	struct rb_node *node;
d1310b2e0   Chris Mason   Btrfs: Split the ...
689
  	int err = 0;
d1310b2e0   Chris Mason   Btrfs: Split the ...
690
691
  	u64 last_start;
  	u64 last_end;
42daec299   Chris Mason   Btrfs: fix errors...
692

0ca1f7ceb   Yan, Zheng   Btrfs: Update met...
693
  	bits |= EXTENT_FIRST_DELALLOC;
d1310b2e0   Chris Mason   Btrfs: Split the ...
694
695
696
  again:
  	if (!prealloc && (mask & __GFP_WAIT)) {
  		prealloc = alloc_extent_state(mask);
8233767a2   Xiao Guangrong   Btrfs: allocate e...
697
  		BUG_ON(!prealloc);
d1310b2e0   Chris Mason   Btrfs: Split the ...
698
  	}
cad321ad5   Chris Mason   Btrfs: shift all ...
699
  	spin_lock(&tree->lock);
9655d2982   Chris Mason   Btrfs: use a cach...
700
701
  	if (cached_state && *cached_state) {
  		state = *cached_state;
df98b6e2c   Josef Bacik   Btrfs: fix how we...
702
703
  		if (state->start <= start && state->end > start &&
  		    state->tree) {
9655d2982   Chris Mason   Btrfs: use a cach...
704
705
706
707
  			node = &state->rb_node;
  			goto hit_next;
  		}
  	}
d1310b2e0   Chris Mason   Btrfs: Split the ...
708
709
710
711
  	/*
  	 * this search will find all the extents that end after
  	 * our range starts.
  	 */
80ea96b1f   Chris Mason   Btrfs: Add a look...
712
  	node = tree_search(tree, start);
d1310b2e0   Chris Mason   Btrfs: Split the ...
713
  	if (!node) {
8233767a2   Xiao Guangrong   Btrfs: allocate e...
714
715
  		prealloc = alloc_extent_state_atomic(prealloc);
  		BUG_ON(!prealloc);
0ca1f7ceb   Yan, Zheng   Btrfs: Update met...
716
  		err = insert_state(tree, prealloc, start, end, &bits);
d1310b2e0   Chris Mason   Btrfs: Split the ...
717
718
719
720
  		prealloc = NULL;
  		BUG_ON(err == -EEXIST);
  		goto out;
  	}
d1310b2e0   Chris Mason   Btrfs: Split the ...
721
  	state = rb_entry(node, struct extent_state, rb_node);
40431d6c1   Chris Mason   Btrfs: optimize s...
722
  hit_next:
d1310b2e0   Chris Mason   Btrfs: Split the ...
723
724
725
726
727
728
729
730
731
732
  	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) {
40431d6c1   Chris Mason   Btrfs: optimize s...
733
  		struct rb_node *next_node;
1edbb734b   Chris Mason   Btrfs: reduce CPU...
734
  		if (state->state & exclusive_bits) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
735
736
737
738
  			*failed_start = state->start;
  			err = -EEXIST;
  			goto out;
  		}
42daec299   Chris Mason   Btrfs: fix errors...
739

1bf85046e   Jeff Mahoney   btrfs: Make exten...
740
  		set_state_bits(tree, state, &bits);
9ed74f2db   Josef Bacik   Btrfs: proper -EN...
741

2c64c53d8   Chris Mason   Btrfs: cache valu...
742
  		cache_state(state, cached_state);
d1310b2e0   Chris Mason   Btrfs: Split the ...
743
  		merge_state(tree, state);
5c939df56   Yan Zheng   btrfs: Fix set/cl...
744
745
  		if (last_end == (u64)-1)
  			goto out;
40431d6c1   Chris Mason   Btrfs: optimize s...
746

5c939df56   Yan Zheng   btrfs: Fix set/cl...
747
  		start = last_end + 1;
df98b6e2c   Josef Bacik   Btrfs: fix how we...
748
  		next_node = rb_next(&state->rb_node);
c7f895a2b   Xiao Guangrong   Btrfs: fix unsafe...
749
750
751
752
753
  		if (next_node && start < end && prealloc && !need_resched()) {
  			state = rb_entry(next_node, struct extent_state,
  					 rb_node);
  			if (state->start == start)
  				goto hit_next;
40431d6c1   Chris Mason   Btrfs: optimize s...
754
  		}
d1310b2e0   Chris Mason   Btrfs: Split the ...
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
  		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) {
1edbb734b   Chris Mason   Btrfs: reduce CPU...
775
  		if (state->state & exclusive_bits) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
776
777
778
779
  			*failed_start = start;
  			err = -EEXIST;
  			goto out;
  		}
8233767a2   Xiao Guangrong   Btrfs: allocate e...
780
781
782
  
  		prealloc = alloc_extent_state_atomic(prealloc);
  		BUG_ON(!prealloc);
d1310b2e0   Chris Mason   Btrfs: Split the ...
783
784
785
786
787
788
  		err = split_state(tree, state, prealloc, start);
  		BUG_ON(err == -EEXIST);
  		prealloc = NULL;
  		if (err)
  			goto out;
  		if (state->end <= end) {
1bf85046e   Jeff Mahoney   btrfs: Make exten...
789
  			set_state_bits(tree, state, &bits);
2c64c53d8   Chris Mason   Btrfs: cache valu...
790
  			cache_state(state, cached_state);
d1310b2e0   Chris Mason   Btrfs: Split the ...
791
  			merge_state(tree, state);
5c939df56   Yan Zheng   btrfs: Fix set/cl...
792
793
794
  			if (last_end == (u64)-1)
  				goto out;
  			start = last_end + 1;
d1310b2e0   Chris Mason   Btrfs: Split the ...
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
  		}
  		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
d397712bc   Chris Mason   Btrfs: Fix checkp...
810
  			this_end = last_start - 1;
8233767a2   Xiao Guangrong   Btrfs: allocate e...
811
812
813
  
  		prealloc = alloc_extent_state_atomic(prealloc);
  		BUG_ON(!prealloc);
c7f895a2b   Xiao Guangrong   Btrfs: fix unsafe...
814
815
816
817
818
  
  		/*
  		 * Avoid to free 'prealloc' if it can be merged with
  		 * the later extent.
  		 */
d1310b2e0   Chris Mason   Btrfs: Split the ...
819
  		err = insert_state(tree, prealloc, start, this_end,
0ca1f7ceb   Yan, Zheng   Btrfs: Update met...
820
  				   &bits);
d1310b2e0   Chris Mason   Btrfs: Split the ...
821
  		BUG_ON(err == -EEXIST);
9ed74f2db   Josef Bacik   Btrfs: proper -EN...
822
  		if (err) {
c7f895a2b   Xiao Guangrong   Btrfs: fix unsafe...
823
  			free_extent_state(prealloc);
9ed74f2db   Josef Bacik   Btrfs: proper -EN...
824
  			prealloc = NULL;
d1310b2e0   Chris Mason   Btrfs: Split the ...
825
  			goto out;
9ed74f2db   Josef Bacik   Btrfs: proper -EN...
826
827
828
  		}
  		cache_state(prealloc, cached_state);
  		prealloc = NULL;
d1310b2e0   Chris Mason   Btrfs: Split the ...
829
830
831
832
833
834
835
836
837
838
  		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) {
1edbb734b   Chris Mason   Btrfs: reduce CPU...
839
  		if (state->state & exclusive_bits) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
840
841
842
843
  			*failed_start = start;
  			err = -EEXIST;
  			goto out;
  		}
8233767a2   Xiao Guangrong   Btrfs: allocate e...
844
845
846
  
  		prealloc = alloc_extent_state_atomic(prealloc);
  		BUG_ON(!prealloc);
d1310b2e0   Chris Mason   Btrfs: Split the ...
847
848
  		err = split_state(tree, state, prealloc, end + 1);
  		BUG_ON(err == -EEXIST);
1bf85046e   Jeff Mahoney   btrfs: Make exten...
849
  		set_state_bits(tree, prealloc, &bits);
2c64c53d8   Chris Mason   Btrfs: cache valu...
850
  		cache_state(prealloc, cached_state);
d1310b2e0   Chris Mason   Btrfs: Split the ...
851
852
853
854
855
856
857
858
  		merge_state(tree, prealloc);
  		prealloc = NULL;
  		goto out;
  	}
  
  	goto search_again;
  
  out:
cad321ad5   Chris Mason   Btrfs: shift all ...
859
  	spin_unlock(&tree->lock);
d1310b2e0   Chris Mason   Btrfs: Split the ...
860
861
862
863
864
865
866
867
  	if (prealloc)
  		free_extent_state(prealloc);
  
  	return err;
  
  search_again:
  	if (start > end)
  		goto out;
cad321ad5   Chris Mason   Btrfs: shift all ...
868
  	spin_unlock(&tree->lock);
d1310b2e0   Chris Mason   Btrfs: Split the ...
869
870
871
872
  	if (mask & __GFP_WAIT)
  		cond_resched();
  	goto again;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
873

462d6fac8   Josef Bacik   Btrfs: introduce ...
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
  /**
   * convert_extent - convert all bits in a given range from one bit to another
   * @tree:	the io tree to search
   * @start:	the start offset in bytes
   * @end:	the end offset in bytes (inclusive)
   * @bits:	the bits to set in this range
   * @clear_bits:	the bits to clear in this range
   * @mask:	the allocation mask
   *
   * This will go through and set bits for the given range.  If any states exist
   * already in this range they are set with the given bit and cleared of the
   * clear_bits.  This is only meant to be used by things that are mergeable, ie
   * converting from say DELALLOC to DIRTY.  This is not meant to be used with
   * boundary bits like LOCK.
   */
  int convert_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
  		       int bits, int clear_bits, gfp_t mask)
  {
  	struct extent_state *state;
  	struct extent_state *prealloc = NULL;
  	struct rb_node *node;
  	int err = 0;
  	u64 last_start;
  	u64 last_end;
  
  again:
  	if (!prealloc && (mask & __GFP_WAIT)) {
  		prealloc = alloc_extent_state(mask);
  		if (!prealloc)
  			return -ENOMEM;
  	}
  
  	spin_lock(&tree->lock);
  	/*
  	 * this search will find all the extents that end after
  	 * our range starts.
  	 */
  	node = tree_search(tree, start);
  	if (!node) {
  		prealloc = alloc_extent_state_atomic(prealloc);
1cf4ffdb3   Liu Bo   Btrfs: drop spin ...
914
915
916
917
  		if (!prealloc) {
  			err = -ENOMEM;
  			goto out;
  		}
462d6fac8   Josef Bacik   Btrfs: introduce ...
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
  		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);
  hit_next:
  	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) {
  		struct rb_node *next_node;
  
  		set_state_bits(tree, state, &bits);
  		clear_state_bit(tree, state, &clear_bits, 0);
  
  		merge_state(tree, state);
  		if (last_end == (u64)-1)
  			goto out;
  
  		start = last_end + 1;
  		next_node = rb_next(&state->rb_node);
  		if (next_node && start < end && prealloc && !need_resched()) {
  			state = rb_entry(next_node, struct extent_state,
  					 rb_node);
  			if (state->start == start)
  				goto hit_next;
  		}
  		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) {
  		prealloc = alloc_extent_state_atomic(prealloc);
1cf4ffdb3   Liu Bo   Btrfs: drop spin ...
973
974
975
976
  		if (!prealloc) {
  			err = -ENOMEM;
  			goto out;
  		}
462d6fac8   Josef Bacik   Btrfs: introduce ...
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
  		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);
  			clear_state_bit(tree, state, &clear_bits, 0);
  			merge_state(tree, state);
  			if (last_end == (u64)-1)
  				goto out;
  			start = last_end + 1;
  		}
  		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;
  
  		prealloc = alloc_extent_state_atomic(prealloc);
1cf4ffdb3   Liu Bo   Btrfs: drop spin ...
1007
1008
1009
1010
  		if (!prealloc) {
  			err = -ENOMEM;
  			goto out;
  		}
462d6fac8   Josef Bacik   Btrfs: introduce ...
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
  
  		/*
  		 * Avoid to free 'prealloc' if it can be merged with
  		 * the later extent.
  		 */
  		err = insert_state(tree, prealloc, start, this_end,
  				   &bits);
  		BUG_ON(err == -EEXIST);
  		if (err) {
  			free_extent_state(prealloc);
  			prealloc = NULL;
  			goto out;
  		}
  		prealloc = NULL;
  		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) {
  		prealloc = alloc_extent_state_atomic(prealloc);
1cf4ffdb3   Liu Bo   Btrfs: drop spin ...
1036
1037
1038
1039
  		if (!prealloc) {
  			err = -ENOMEM;
  			goto out;
  		}
462d6fac8   Josef Bacik   Btrfs: introduce ...
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
  
  		err = split_state(tree, state, prealloc, end + 1);
  		BUG_ON(err == -EEXIST);
  
  		set_state_bits(tree, prealloc, &bits);
  		clear_state_bit(tree, prealloc, &clear_bits, 0);
  
  		merge_state(tree, prealloc);
  		prealloc = NULL;
  		goto out;
  	}
  
  	goto search_again;
  
  out:
  	spin_unlock(&tree->lock);
  	if (prealloc)
  		free_extent_state(prealloc);
  
  	return err;
  
  search_again:
  	if (start > end)
  		goto out;
  	spin_unlock(&tree->lock);
  	if (mask & __GFP_WAIT)
  		cond_resched();
  	goto again;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
1069
1070
1071
1072
1073
  /* 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,
2c64c53d8   Chris Mason   Btrfs: cache valu...
1074
  			      NULL, mask);
d1310b2e0   Chris Mason   Btrfs: Split the ...
1075
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
1076
1077
1078
1079
1080
  
  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,
2c64c53d8   Chris Mason   Btrfs: cache valu...
1081
  			      NULL, mask);
d1310b2e0   Chris Mason   Btrfs: Split the ...
1082
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
1083
1084
1085
1086
  
  int clear_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
  		      int bits, gfp_t mask)
  {
2c64c53d8   Chris Mason   Btrfs: cache valu...
1087
  	return clear_extent_bit(tree, start, end, bits, 0, 0, NULL, mask);
d1310b2e0   Chris Mason   Btrfs: Split the ...
1088
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
1089
1090
  
  int set_extent_delalloc(struct extent_io_tree *tree, u64 start, u64 end,
2ac55d41b   Josef Bacik   Btrfs: cache the ...
1091
  			struct extent_state **cached_state, gfp_t mask)
d1310b2e0   Chris Mason   Btrfs: Split the ...
1092
1093
  {
  	return set_extent_bit(tree, start, end,
fee187d9d   Liu Bo   Btrfs: do not set...
1094
  			      EXTENT_DELALLOC | EXTENT_UPTODATE,
2ac55d41b   Josef Bacik   Btrfs: cache the ...
1095
  			      0, NULL, cached_state, mask);
d1310b2e0   Chris Mason   Btrfs: Split the ...
1096
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
1097
1098
1099
1100
1101
  
  int clear_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end,
  		       gfp_t mask)
  {
  	return clear_extent_bit(tree, start, end,
32c00aff7   Josef Bacik   Btrfs: release de...
1102
  				EXTENT_DIRTY | EXTENT_DELALLOC |
0ca1f7ceb   Yan, Zheng   Btrfs: Update met...
1103
  				EXTENT_DO_ACCOUNTING, 0, 0, NULL, mask);
d1310b2e0   Chris Mason   Btrfs: Split the ...
1104
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
1105
1106
1107
1108
1109
  
  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,
2c64c53d8   Chris Mason   Btrfs: cache valu...
1110
  			      NULL, mask);
d1310b2e0   Chris Mason   Btrfs: Split the ...
1111
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
1112

d1310b2e0   Chris Mason   Btrfs: Split the ...
1113
  int set_extent_uptodate(struct extent_io_tree *tree, u64 start, u64 end,
507903b81   Arne Jansen   btrfs: using cach...
1114
  			struct extent_state **cached_state, gfp_t mask)
d1310b2e0   Chris Mason   Btrfs: Split the ...
1115
  {
507903b81   Arne Jansen   btrfs: using cach...
1116
1117
  	return set_extent_bit(tree, start, end, EXTENT_UPTODATE, 0,
  			      NULL, cached_state, mask);
d1310b2e0   Chris Mason   Btrfs: Split the ...
1118
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
1119

d397712bc   Chris Mason   Btrfs: Fix checkp...
1120
  static int clear_extent_uptodate(struct extent_io_tree *tree, u64 start,
2ac55d41b   Josef Bacik   Btrfs: cache the ...
1121
1122
  				 u64 end, struct extent_state **cached_state,
  				 gfp_t mask)
d1310b2e0   Chris Mason   Btrfs: Split the ...
1123
  {
2c64c53d8   Chris Mason   Btrfs: cache valu...
1124
  	return clear_extent_bit(tree, start, end, EXTENT_UPTODATE, 0, 0,
2ac55d41b   Josef Bacik   Btrfs: cache the ...
1125
  				cached_state, mask);
d1310b2e0   Chris Mason   Btrfs: Split the ...
1126
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
1127

d352ac681   Chris Mason   Btrfs: add and im...
1128
1129
1130
1131
  /*
   * either insert or lock state struct between start and end use mask to tell
   * us if waiting is desired.
   */
1edbb734b   Chris Mason   Btrfs: reduce CPU...
1132
  int lock_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
2c64c53d8   Chris Mason   Btrfs: cache valu...
1133
  		     int bits, struct extent_state **cached_state, gfp_t mask)
d1310b2e0   Chris Mason   Btrfs: Split the ...
1134
1135
1136
1137
  {
  	int err;
  	u64 failed_start;
  	while (1) {
1edbb734b   Chris Mason   Btrfs: reduce CPU...
1138
  		err = set_extent_bit(tree, start, end, EXTENT_LOCKED | bits,
2c64c53d8   Chris Mason   Btrfs: cache valu...
1139
1140
  				     EXTENT_LOCKED, &failed_start,
  				     cached_state, mask);
d1310b2e0   Chris Mason   Btrfs: Split the ...
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
  		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;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
1151

1edbb734b   Chris Mason   Btrfs: reduce CPU...
1152
1153
  int lock_extent(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask)
  {
2c64c53d8   Chris Mason   Btrfs: cache valu...
1154
  	return lock_extent_bits(tree, start, end, 0, NULL, mask);
1edbb734b   Chris Mason   Btrfs: reduce CPU...
1155
  }
251792013   Josef Bacik   Btrfs: nuke fs wi...
1156
1157
1158
1159
1160
  int try_lock_extent(struct extent_io_tree *tree, u64 start, u64 end,
  		    gfp_t mask)
  {
  	int err;
  	u64 failed_start;
2c64c53d8   Chris Mason   Btrfs: cache valu...
1161
1162
  	err = set_extent_bit(tree, start, end, EXTENT_LOCKED, EXTENT_LOCKED,
  			     &failed_start, NULL, mask);
6643558db   Yan Zheng   Btrfs: Fix booken...
1163
1164
1165
  	if (err == -EEXIST) {
  		if (failed_start > start)
  			clear_extent_bit(tree, start, failed_start - 1,
2c64c53d8   Chris Mason   Btrfs: cache valu...
1166
  					 EXTENT_LOCKED, 1, 0, NULL, mask);
251792013   Josef Bacik   Btrfs: nuke fs wi...
1167
  		return 0;
6643558db   Yan Zheng   Btrfs: Fix booken...
1168
  	}
251792013   Josef Bacik   Btrfs: nuke fs wi...
1169
1170
  	return 1;
  }
251792013   Josef Bacik   Btrfs: nuke fs wi...
1171

2c64c53d8   Chris Mason   Btrfs: cache valu...
1172
1173
1174
1175
1176
1177
  int unlock_extent_cached(struct extent_io_tree *tree, u64 start, u64 end,
  			 struct extent_state **cached, gfp_t mask)
  {
  	return clear_extent_bit(tree, start, end, EXTENT_LOCKED, 1, 0, cached,
  				mask);
  }
507903b81   Arne Jansen   btrfs: using cach...
1178
  int unlock_extent(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask)
d1310b2e0   Chris Mason   Btrfs: Split the ...
1179
  {
2c64c53d8   Chris Mason   Btrfs: cache valu...
1180
1181
  	return clear_extent_bit(tree, start, end, EXTENT_LOCKED, 1, 0, NULL,
  				mask);
d1310b2e0   Chris Mason   Btrfs: Split the ...
1182
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
1183
1184
  
  /*
d1310b2e0   Chris Mason   Btrfs: Split the ...
1185
1186
   * helper function to set both pages and extents in the tree writeback
   */
b2950863c   Christoph Hellwig   Btrfs: make thing...
1187
  static int set_range_writeback(struct extent_io_tree *tree, u64 start, u64 end)
d1310b2e0   Chris Mason   Btrfs: Split the ...
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
  {
  	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++;
  	}
d1310b2e0   Chris Mason   Btrfs: Split the ...
1200
1201
  	return 0;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
1202

d352ac681   Chris Mason   Btrfs: add and im...
1203
1204
1205
1206
  /* find the first state struct with 'bits' set after 'start', and
   * return it.  tree->lock must be held.  NULL will returned if
   * nothing was found after 'start'
   */
d7fc640e6   Chris Mason   Btrfs: Allocator ...
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
  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);
d397712bc   Chris Mason   Btrfs: Fix checkp...
1218
  	if (!node)
d7fc640e6   Chris Mason   Btrfs: Allocator ...
1219
  		goto out;
d7fc640e6   Chris Mason   Btrfs: Allocator ...
1220

d397712bc   Chris Mason   Btrfs: Fix checkp...
1221
  	while (1) {
d7fc640e6   Chris Mason   Btrfs: Allocator ...
1222
  		state = rb_entry(node, struct extent_state, rb_node);
d397712bc   Chris Mason   Btrfs: Fix checkp...
1223
  		if (state->end >= start && (state->state & bits))
d7fc640e6   Chris Mason   Btrfs: Allocator ...
1224
  			return state;
d397712bc   Chris Mason   Btrfs: Fix checkp...
1225

d7fc640e6   Chris Mason   Btrfs: Allocator ...
1226
1227
1228
1229
1230
1231
1232
  		node = rb_next(node);
  		if (!node)
  			break;
  	}
  out:
  	return NULL;
  }
d7fc640e6   Chris Mason   Btrfs: Allocator ...
1233

d352ac681   Chris Mason   Btrfs: add and im...
1234
  /*
69261c4b6   Xiao Guangrong   Btrfs: clean up f...
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
   * find the first offset in the io tree with 'bits' set. zero is
   * returned if we find something, and *start_ret and *end_ret are
   * set to reflect the state struct that was found.
   *
   * If nothing was found, 1 is returned, < 0 on error
   */
  int find_first_extent_bit(struct extent_io_tree *tree, u64 start,
  			  u64 *start_ret, u64 *end_ret, int bits)
  {
  	struct extent_state *state;
  	int ret = 1;
  
  	spin_lock(&tree->lock);
  	state = find_first_extent_bit_state(tree, start, bits);
  	if (state) {
  		*start_ret = state->start;
  		*end_ret = state->end;
  		ret = 0;
  	}
  	spin_unlock(&tree->lock);
  	return ret;
  }
  
  /*
d352ac681   Chris Mason   Btrfs: add and im...
1259
1260
1261
1262
1263
   * find a contiguous range of bytes in the file marked as delalloc, not
   * more than 'max_bytes'.  start and end are used to return the range,
   *
   * 1 is returned if we find something, 0 if nothing was in the tree
   */
c8b978188   Chris Mason   Btrfs: Add zlib c...
1264
  static noinline u64 find_delalloc_range(struct extent_io_tree *tree,
c2a128d28   Josef Bacik   Btrfs: cache exte...
1265
1266
  					u64 *start, u64 *end, u64 max_bytes,
  					struct extent_state **cached_state)
d1310b2e0   Chris Mason   Btrfs: Split the ...
1267
1268
1269
1270
1271
1272
  {
  	struct rb_node *node;
  	struct extent_state *state;
  	u64 cur_start = *start;
  	u64 found = 0;
  	u64 total_bytes = 0;
cad321ad5   Chris Mason   Btrfs: shift all ...
1273
  	spin_lock(&tree->lock);
c8b978188   Chris Mason   Btrfs: Add zlib c...
1274

d1310b2e0   Chris Mason   Btrfs: Split the ...
1275
1276
1277
1278
  	/*
  	 * this search will find all the extents that end after
  	 * our range starts.
  	 */
80ea96b1f   Chris Mason   Btrfs: Add a look...
1279
  	node = tree_search(tree, cur_start);
2b114d1d3   Peter   Btrfs: Correct us...
1280
  	if (!node) {
3b951516e   Chris Mason   Btrfs: Use the ex...
1281
1282
  		if (!found)
  			*end = (u64)-1;
d1310b2e0   Chris Mason   Btrfs: Split the ...
1283
1284
  		goto out;
  	}
d397712bc   Chris Mason   Btrfs: Fix checkp...
1285
  	while (1) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
1286
  		state = rb_entry(node, struct extent_state, rb_node);
5b21f2ed3   Zheng Yan   Btrfs: extent_map...
1287
1288
  		if (found && (state->start != cur_start ||
  			      (state->state & EXTENT_BOUNDARY))) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
1289
1290
1291
1292
1293
1294
1295
  			goto out;
  		}
  		if (!(state->state & EXTENT_DELALLOC)) {
  			if (!found)
  				*end = state->end;
  			goto out;
  		}
c2a128d28   Josef Bacik   Btrfs: cache exte...
1296
  		if (!found) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
1297
  			*start = state->start;
c2a128d28   Josef Bacik   Btrfs: cache exte...
1298
1299
1300
  			*cached_state = state;
  			atomic_inc(&state->refs);
  		}
d1310b2e0   Chris Mason   Btrfs: Split the ...
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
  		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:
cad321ad5   Chris Mason   Btrfs: shift all ...
1312
  	spin_unlock(&tree->lock);
d1310b2e0   Chris Mason   Btrfs: Split the ...
1313
1314
  	return found;
  }
c8b978188   Chris Mason   Btrfs: Add zlib c...
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
  static noinline int __unlock_for_delalloc(struct inode *inode,
  					  struct page *locked_page,
  					  u64 start, u64 end)
  {
  	int ret;
  	struct page *pages[16];
  	unsigned long index = start >> PAGE_CACHE_SHIFT;
  	unsigned long end_index = end >> PAGE_CACHE_SHIFT;
  	unsigned long nr_pages = end_index - index + 1;
  	int i;
  
  	if (index == locked_page->index && end_index == index)
  		return 0;
d397712bc   Chris Mason   Btrfs: Fix checkp...
1328
  	while (nr_pages > 0) {
c8b978188   Chris Mason   Btrfs: Add zlib c...
1329
  		ret = find_get_pages_contig(inode->i_mapping, index,
5b050f04c   Chris Mason   Btrfs: Fix compil...
1330
1331
  				     min_t(unsigned long, nr_pages,
  				     ARRAY_SIZE(pages)), pages);
c8b978188   Chris Mason   Btrfs: Add zlib c...
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
  		for (i = 0; i < ret; i++) {
  			if (pages[i] != locked_page)
  				unlock_page(pages[i]);
  			page_cache_release(pages[i]);
  		}
  		nr_pages -= ret;
  		index += ret;
  		cond_resched();
  	}
  	return 0;
  }
  
  static noinline int lock_delalloc_pages(struct inode *inode,
  					struct page *locked_page,
  					u64 delalloc_start,
  					u64 delalloc_end)
  {
  	unsigned long index = delalloc_start >> PAGE_CACHE_SHIFT;
  	unsigned long start_index = index;
  	unsigned long end_index = delalloc_end >> PAGE_CACHE_SHIFT;
  	unsigned long pages_locked = 0;
  	struct page *pages[16];
  	unsigned long nrpages;
  	int ret;
  	int i;
  
  	/* the caller is responsible for locking the start index */
  	if (index == locked_page->index && index == end_index)
  		return 0;
  
  	/* skip the page at the start index */
  	nrpages = end_index - index + 1;
d397712bc   Chris Mason   Btrfs: Fix checkp...
1364
  	while (nrpages > 0) {
c8b978188   Chris Mason   Btrfs: Add zlib c...
1365
  		ret = find_get_pages_contig(inode->i_mapping, index,
5b050f04c   Chris Mason   Btrfs: Fix compil...
1366
1367
  				     min_t(unsigned long,
  				     nrpages, ARRAY_SIZE(pages)), pages);
c8b978188   Chris Mason   Btrfs: Add zlib c...
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
  		if (ret == 0) {
  			ret = -EAGAIN;
  			goto done;
  		}
  		/* now we have an array of pages, lock them all */
  		for (i = 0; i < ret; i++) {
  			/*
  			 * the caller is taking responsibility for
  			 * locked_page
  			 */
771ed689d   Chris Mason   Btrfs: Optimize c...
1378
  			if (pages[i] != locked_page) {
c8b978188   Chris Mason   Btrfs: Add zlib c...
1379
  				lock_page(pages[i]);
f2b1c41cf   Chris Mason   Btrfs: Make sure ...
1380
1381
  				if (!PageDirty(pages[i]) ||
  				    pages[i]->mapping != inode->i_mapping) {
771ed689d   Chris Mason   Btrfs: Optimize c...
1382
1383
1384
1385
1386
1387
  					ret = -EAGAIN;
  					unlock_page(pages[i]);
  					page_cache_release(pages[i]);
  					goto done;
  				}
  			}
c8b978188   Chris Mason   Btrfs: Add zlib c...
1388
  			page_cache_release(pages[i]);
771ed689d   Chris Mason   Btrfs: Optimize c...
1389
  			pages_locked++;
c8b978188   Chris Mason   Btrfs: Add zlib c...
1390
  		}
c8b978188   Chris Mason   Btrfs: Add zlib c...
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
  		nrpages -= ret;
  		index += ret;
  		cond_resched();
  	}
  	ret = 0;
  done:
  	if (ret && pages_locked) {
  		__unlock_for_delalloc(inode, locked_page,
  			      delalloc_start,
  			      ((u64)(start_index + pages_locked - 1)) <<
  			      PAGE_CACHE_SHIFT);
  	}
  	return ret;
  }
  
  /*
   * find a contiguous range of bytes in the file marked as delalloc, not
   * more than 'max_bytes'.  start and end are used to return the range,
   *
   * 1 is returned if we find something, 0 if nothing was in the tree
   */
  static noinline u64 find_lock_delalloc_range(struct inode *inode,
  					     struct extent_io_tree *tree,
  					     struct page *locked_page,
  					     u64 *start, u64 *end,
  					     u64 max_bytes)
  {
  	u64 delalloc_start;
  	u64 delalloc_end;
  	u64 found;
9655d2982   Chris Mason   Btrfs: use a cach...
1421
  	struct extent_state *cached_state = NULL;
c8b978188   Chris Mason   Btrfs: Add zlib c...
1422
1423
1424
1425
1426
1427
1428
1429
  	int ret;
  	int loops = 0;
  
  again:
  	/* step one, find a bunch of delalloc bytes starting at start */
  	delalloc_start = *start;
  	delalloc_end = 0;
  	found = find_delalloc_range(tree, &delalloc_start, &delalloc_end,
c2a128d28   Josef Bacik   Btrfs: cache exte...
1430
  				    max_bytes, &cached_state);
70b99e695   Chris Mason   Btrfs: Compressio...
1431
  	if (!found || delalloc_end <= *start) {
c8b978188   Chris Mason   Btrfs: Add zlib c...
1432
1433
  		*start = delalloc_start;
  		*end = delalloc_end;
c2a128d28   Josef Bacik   Btrfs: cache exte...
1434
  		free_extent_state(cached_state);
c8b978188   Chris Mason   Btrfs: Add zlib c...
1435
1436
1437
1438
  		return found;
  	}
  
  	/*
70b99e695   Chris Mason   Btrfs: Compressio...
1439
1440
1441
1442
  	 * start comes from the offset of locked_page.  We have to lock
  	 * pages in order, so we can't process delalloc bytes before
  	 * locked_page
  	 */
d397712bc   Chris Mason   Btrfs: Fix checkp...
1443
  	if (delalloc_start < *start)
70b99e695   Chris Mason   Btrfs: Compressio...
1444
  		delalloc_start = *start;
70b99e695   Chris Mason   Btrfs: Compressio...
1445
1446
  
  	/*
c8b978188   Chris Mason   Btrfs: Add zlib c...
1447
1448
1449
  	 * make sure to limit the number of pages we try to lock down
  	 * if we're looping.
  	 */
d397712bc   Chris Mason   Btrfs: Fix checkp...
1450
  	if (delalloc_end + 1 - delalloc_start > max_bytes && loops)
771ed689d   Chris Mason   Btrfs: Optimize c...
1451
  		delalloc_end = delalloc_start + PAGE_CACHE_SIZE - 1;
d397712bc   Chris Mason   Btrfs: Fix checkp...
1452

c8b978188   Chris Mason   Btrfs: Add zlib c...
1453
1454
1455
1456
1457
1458
1459
  	/* step two, lock all the pages after the page that has start */
  	ret = lock_delalloc_pages(inode, locked_page,
  				  delalloc_start, delalloc_end);
  	if (ret == -EAGAIN) {
  		/* some of the pages are gone, lets avoid looping by
  		 * shortening the size of the delalloc range we're searching
  		 */
9655d2982   Chris Mason   Btrfs: use a cach...
1460
  		free_extent_state(cached_state);
c8b978188   Chris Mason   Btrfs: Add zlib c...
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
  		if (!loops) {
  			unsigned long offset = (*start) & (PAGE_CACHE_SIZE - 1);
  			max_bytes = PAGE_CACHE_SIZE - offset;
  			loops = 1;
  			goto again;
  		} else {
  			found = 0;
  			goto out_failed;
  		}
  	}
  	BUG_ON(ret);
  
  	/* step three, lock the state bits for the whole range */
9655d2982   Chris Mason   Btrfs: use a cach...
1474
1475
  	lock_extent_bits(tree, delalloc_start, delalloc_end,
  			 0, &cached_state, GFP_NOFS);
c8b978188   Chris Mason   Btrfs: Add zlib c...
1476
1477
1478
  
  	/* then test to make sure it is all still delalloc */
  	ret = test_range_bit(tree, delalloc_start, delalloc_end,
9655d2982   Chris Mason   Btrfs: use a cach...
1479
  			     EXTENT_DELALLOC, 1, cached_state);
c8b978188   Chris Mason   Btrfs: Add zlib c...
1480
  	if (!ret) {
9655d2982   Chris Mason   Btrfs: use a cach...
1481
1482
  		unlock_extent_cached(tree, delalloc_start, delalloc_end,
  				     &cached_state, GFP_NOFS);
c8b978188   Chris Mason   Btrfs: Add zlib c...
1483
1484
1485
1486
1487
  		__unlock_for_delalloc(inode, locked_page,
  			      delalloc_start, delalloc_end);
  		cond_resched();
  		goto again;
  	}
9655d2982   Chris Mason   Btrfs: use a cach...
1488
  	free_extent_state(cached_state);
c8b978188   Chris Mason   Btrfs: Add zlib c...
1489
1490
1491
1492
1493
1494
1495
1496
1497
  	*start = delalloc_start;
  	*end = delalloc_end;
  out_failed:
  	return found;
  }
  
  int extent_clear_unlock_delalloc(struct inode *inode,
  				struct extent_io_tree *tree,
  				u64 start, u64 end, struct page *locked_page,
a791e35e1   Chris Mason   Btrfs: cleanup ex...
1498
  				unsigned long op)
c8b978188   Chris Mason   Btrfs: Add zlib c...
1499
1500
1501
1502
1503
1504
1505
  {
  	int ret;
  	struct page *pages[16];
  	unsigned long index = start >> PAGE_CACHE_SHIFT;
  	unsigned long end_index = end >> PAGE_CACHE_SHIFT;
  	unsigned long nr_pages = end_index - index + 1;
  	int i;
771ed689d   Chris Mason   Btrfs: Optimize c...
1506
  	int clear_bits = 0;
c8b978188   Chris Mason   Btrfs: Add zlib c...
1507

a791e35e1   Chris Mason   Btrfs: cleanup ex...
1508
  	if (op & EXTENT_CLEAR_UNLOCK)
771ed689d   Chris Mason   Btrfs: Optimize c...
1509
  		clear_bits |= EXTENT_LOCKED;
a791e35e1   Chris Mason   Btrfs: cleanup ex...
1510
  	if (op & EXTENT_CLEAR_DIRTY)
c8b978188   Chris Mason   Btrfs: Add zlib c...
1511
  		clear_bits |= EXTENT_DIRTY;
a791e35e1   Chris Mason   Btrfs: cleanup ex...
1512
  	if (op & EXTENT_CLEAR_DELALLOC)
771ed689d   Chris Mason   Btrfs: Optimize c...
1513
  		clear_bits |= EXTENT_DELALLOC;
2c64c53d8   Chris Mason   Btrfs: cache valu...
1514
  	clear_extent_bit(tree, start, end, clear_bits, 1, 0, NULL, GFP_NOFS);
32c00aff7   Josef Bacik   Btrfs: release de...
1515
1516
1517
  	if (!(op & (EXTENT_CLEAR_UNLOCK_PAGE | EXTENT_CLEAR_DIRTY |
  		    EXTENT_SET_WRITEBACK | EXTENT_END_WRITEBACK |
  		    EXTENT_SET_PRIVATE2)))
771ed689d   Chris Mason   Btrfs: Optimize c...
1518
  		return 0;
c8b978188   Chris Mason   Btrfs: Add zlib c...
1519

d397712bc   Chris Mason   Btrfs: Fix checkp...
1520
  	while (nr_pages > 0) {
c8b978188   Chris Mason   Btrfs: Add zlib c...
1521
  		ret = find_get_pages_contig(inode->i_mapping, index,
5b050f04c   Chris Mason   Btrfs: Fix compil...
1522
1523
  				     min_t(unsigned long,
  				     nr_pages, ARRAY_SIZE(pages)), pages);
c8b978188   Chris Mason   Btrfs: Add zlib c...
1524
  		for (i = 0; i < ret; i++) {
8b62b72b2   Chris Mason   Btrfs: Use PagePr...
1525

a791e35e1   Chris Mason   Btrfs: cleanup ex...
1526
  			if (op & EXTENT_SET_PRIVATE2)
8b62b72b2   Chris Mason   Btrfs: Use PagePr...
1527
  				SetPagePrivate2(pages[i]);
c8b978188   Chris Mason   Btrfs: Add zlib c...
1528
1529
1530
1531
  			if (pages[i] == locked_page) {
  				page_cache_release(pages[i]);
  				continue;
  			}
a791e35e1   Chris Mason   Btrfs: cleanup ex...
1532
  			if (op & EXTENT_CLEAR_DIRTY)
c8b978188   Chris Mason   Btrfs: Add zlib c...
1533
  				clear_page_dirty_for_io(pages[i]);
a791e35e1   Chris Mason   Btrfs: cleanup ex...
1534
  			if (op & EXTENT_SET_WRITEBACK)
c8b978188   Chris Mason   Btrfs: Add zlib c...
1535
  				set_page_writeback(pages[i]);
a791e35e1   Chris Mason   Btrfs: cleanup ex...
1536
  			if (op & EXTENT_END_WRITEBACK)
c8b978188   Chris Mason   Btrfs: Add zlib c...
1537
  				end_page_writeback(pages[i]);
a791e35e1   Chris Mason   Btrfs: cleanup ex...
1538
  			if (op & EXTENT_CLEAR_UNLOCK_PAGE)
771ed689d   Chris Mason   Btrfs: Optimize c...
1539
  				unlock_page(pages[i]);
c8b978188   Chris Mason   Btrfs: Add zlib c...
1540
1541
1542
1543
1544
1545
1546
1547
  			page_cache_release(pages[i]);
  		}
  		nr_pages -= ret;
  		index += ret;
  		cond_resched();
  	}
  	return 0;
  }
c8b978188   Chris Mason   Btrfs: Add zlib c...
1548

d352ac681   Chris Mason   Btrfs: add and im...
1549
1550
1551
1552
1553
  /*
   * count the number of bytes in the tree that have a given bit(s)
   * set.  This can be fairly slow, except for EXTENT_DIRTY which is
   * cached.  The total number found is returned.
   */
d1310b2e0   Chris Mason   Btrfs: Split the ...
1554
1555
  u64 count_range_bits(struct extent_io_tree *tree,
  		     u64 *start, u64 search_end, u64 max_bytes,
ec29ed5b4   Chris Mason   Btrfs: fix fiemap...
1556
  		     unsigned long bits, int contig)
d1310b2e0   Chris Mason   Btrfs: Split the ...
1557
1558
1559
1560
1561
  {
  	struct rb_node *node;
  	struct extent_state *state;
  	u64 cur_start = *start;
  	u64 total_bytes = 0;
ec29ed5b4   Chris Mason   Btrfs: fix fiemap...
1562
  	u64 last = 0;
d1310b2e0   Chris Mason   Btrfs: Split the ...
1563
1564
1565
  	int found = 0;
  
  	if (search_end <= cur_start) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
1566
1567
1568
  		WARN_ON(1);
  		return 0;
  	}
cad321ad5   Chris Mason   Btrfs: shift all ...
1569
  	spin_lock(&tree->lock);
d1310b2e0   Chris Mason   Btrfs: Split the ...
1570
1571
1572
1573
1574
1575
1576
1577
  	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.
  	 */
80ea96b1f   Chris Mason   Btrfs: Add a look...
1578
  	node = tree_search(tree, cur_start);
d397712bc   Chris Mason   Btrfs: Fix checkp...
1579
  	if (!node)
d1310b2e0   Chris Mason   Btrfs: Split the ...
1580
  		goto out;
d1310b2e0   Chris Mason   Btrfs: Split the ...
1581

d397712bc   Chris Mason   Btrfs: Fix checkp...
1582
  	while (1) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
1583
1584
1585
  		state = rb_entry(node, struct extent_state, rb_node);
  		if (state->start > search_end)
  			break;
ec29ed5b4   Chris Mason   Btrfs: fix fiemap...
1586
1587
1588
  		if (contig && found && state->start > last + 1)
  			break;
  		if (state->end >= cur_start && (state->state & bits) == bits) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
1589
1590
1591
1592
1593
  			total_bytes += min(search_end, state->end) + 1 -
  				       max(cur_start, state->start);
  			if (total_bytes >= max_bytes)
  				break;
  			if (!found) {
af60bed24   Josef Bacik   Btrfs: set range_...
1594
  				*start = max(cur_start, state->start);
d1310b2e0   Chris Mason   Btrfs: Split the ...
1595
1596
  				found = 1;
  			}
ec29ed5b4   Chris Mason   Btrfs: fix fiemap...
1597
1598
1599
  			last = state->end;
  		} else if (contig && found) {
  			break;
d1310b2e0   Chris Mason   Btrfs: Split the ...
1600
1601
1602
1603
1604
1605
  		}
  		node = rb_next(node);
  		if (!node)
  			break;
  	}
  out:
cad321ad5   Chris Mason   Btrfs: shift all ...
1606
  	spin_unlock(&tree->lock);
d1310b2e0   Chris Mason   Btrfs: Split the ...
1607
1608
  	return total_bytes;
  }
b2950863c   Christoph Hellwig   Btrfs: make thing...
1609

d352ac681   Chris Mason   Btrfs: add and im...
1610
1611
1612
1613
  /*
   * set the private field for a given byte offset in the tree.  If there isn't
   * an extent_state there already, this does nothing.
   */
d1310b2e0   Chris Mason   Btrfs: Split the ...
1614
1615
1616
1617
1618
  int set_state_private(struct extent_io_tree *tree, u64 start, u64 private)
  {
  	struct rb_node *node;
  	struct extent_state *state;
  	int ret = 0;
cad321ad5   Chris Mason   Btrfs: shift all ...
1619
  	spin_lock(&tree->lock);
d1310b2e0   Chris Mason   Btrfs: Split the ...
1620
1621
1622
1623
  	/*
  	 * this search will find all the extents that end after
  	 * our range starts.
  	 */
80ea96b1f   Chris Mason   Btrfs: Add a look...
1624
  	node = tree_search(tree, start);
2b114d1d3   Peter   Btrfs: Correct us...
1625
  	if (!node) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
  		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:
cad321ad5   Chris Mason   Btrfs: shift all ...
1636
  	spin_unlock(&tree->lock);
d1310b2e0   Chris Mason   Btrfs: Split the ...
1637
1638
1639
1640
1641
1642
1643
1644
  	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;
cad321ad5   Chris Mason   Btrfs: shift all ...
1645
  	spin_lock(&tree->lock);
d1310b2e0   Chris Mason   Btrfs: Split the ...
1646
1647
1648
1649
  	/*
  	 * this search will find all the extents that end after
  	 * our range starts.
  	 */
80ea96b1f   Chris Mason   Btrfs: Add a look...
1650
  	node = tree_search(tree, start);
2b114d1d3   Peter   Btrfs: Correct us...
1651
  	if (!node) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
  		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:
cad321ad5   Chris Mason   Btrfs: shift all ...
1662
  	spin_unlock(&tree->lock);
d1310b2e0   Chris Mason   Btrfs: Split the ...
1663
1664
1665
1666
1667
  	return ret;
  }
  
  /*
   * searches a range in the state tree for a given mask.
70dec8079   Chris Mason   Btrfs: extent_io ...
1668
   * If 'filled' == 1, this returns 1 only if every extent in the tree
d1310b2e0   Chris Mason   Btrfs: Split the ...
1669
1670
1671
1672
   * 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,
9655d2982   Chris Mason   Btrfs: use a cach...
1673
  		   int bits, int filled, struct extent_state *cached)
d1310b2e0   Chris Mason   Btrfs: Split the ...
1674
1675
1676
1677
  {
  	struct extent_state *state = NULL;
  	struct rb_node *node;
  	int bitset = 0;
d1310b2e0   Chris Mason   Btrfs: Split the ...
1678

cad321ad5   Chris Mason   Btrfs: shift all ...
1679
  	spin_lock(&tree->lock);
df98b6e2c   Josef Bacik   Btrfs: fix how we...
1680
1681
  	if (cached && cached->tree && cached->start <= start &&
  	    cached->end > start)
9655d2982   Chris Mason   Btrfs: use a cach...
1682
1683
1684
  		node = &cached->rb_node;
  	else
  		node = tree_search(tree, start);
d1310b2e0   Chris Mason   Btrfs: Split the ...
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
  	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;
  		}
46562cec9   Chris Mason   Btrfs: Fix test_r...
1704
1705
1706
  
  		if (state->end == (u64)-1)
  			break;
d1310b2e0   Chris Mason   Btrfs: Split the ...
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
  		start = state->end + 1;
  		if (start > end)
  			break;
  		node = rb_next(node);
  		if (!node) {
  			if (filled)
  				bitset = 0;
  			break;
  		}
  	}
cad321ad5   Chris Mason   Btrfs: shift all ...
1717
  	spin_unlock(&tree->lock);
d1310b2e0   Chris Mason   Btrfs: Split the ...
1718
1719
  	return bitset;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
  
  /*
   * 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;
9655d2982   Chris Mason   Btrfs: use a cach...
1730
  	if (test_range_bit(tree, start, end, EXTENT_UPTODATE, 1, NULL))
d1310b2e0   Chris Mason   Btrfs: Split the ...
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
  		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;
9655d2982   Chris Mason   Btrfs: use a cach...
1744
  	if (!test_range_bit(tree, start, end, EXTENT_LOCKED, 0, NULL))
d1310b2e0   Chris Mason   Btrfs: Split the ...
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
  		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)
  {
1edbb734b   Chris Mason   Btrfs: reduce CPU...
1756
  	end_page_writeback(page);
d1310b2e0   Chris Mason   Btrfs: Split the ...
1757
1758
  	return 0;
  }
4a54c8c16   Jan Schmidt   btrfs: Moved repa...
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
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
1896
1897
1898
1899
1900
1901
1902
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
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
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
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
  /*
   * When IO fails, either with EIO or csum verification fails, we
   * try other mirrors that might have a good copy of the data.  This
   * io_failure_record is used to record state as we go through all the
   * mirrors.  If another mirror has good data, the page is set up to date
   * and things continue.  If a good mirror can't be found, the original
   * bio end_io callback is called to indicate things have failed.
   */
  struct io_failure_record {
  	struct page *page;
  	u64 start;
  	u64 len;
  	u64 logical;
  	unsigned long bio_flags;
  	int this_mirror;
  	int failed_mirror;
  	int in_validation;
  };
  
  static int free_io_failure(struct inode *inode, struct io_failure_record *rec,
  				int did_repair)
  {
  	int ret;
  	int err = 0;
  	struct extent_io_tree *failure_tree = &BTRFS_I(inode)->io_failure_tree;
  
  	set_state_private(failure_tree, rec->start, 0);
  	ret = clear_extent_bits(failure_tree, rec->start,
  				rec->start + rec->len - 1,
  				EXTENT_LOCKED | EXTENT_DIRTY, GFP_NOFS);
  	if (ret)
  		err = ret;
  
  	if (did_repair) {
  		ret = clear_extent_bits(&BTRFS_I(inode)->io_tree, rec->start,
  					rec->start + rec->len - 1,
  					EXTENT_DAMAGED, GFP_NOFS);
  		if (ret && !err)
  			err = ret;
  	}
  
  	kfree(rec);
  	return err;
  }
  
  static void repair_io_failure_callback(struct bio *bio, int err)
  {
  	complete(bio->bi_private);
  }
  
  /*
   * this bypasses the standard btrfs submit functions deliberately, as
   * the standard behavior is to write all copies in a raid setup. here we only
   * want to write the one bad copy. so we do the mapping for ourselves and issue
   * submit_bio directly.
   * to avoid any synchonization issues, wait for the data after writing, which
   * actually prevents the read that triggered the error from finishing.
   * currently, there can be no more than two copies of every data bit. thus,
   * exactly one rewrite is required.
   */
  int repair_io_failure(struct btrfs_mapping_tree *map_tree, u64 start,
  			u64 length, u64 logical, struct page *page,
  			int mirror_num)
  {
  	struct bio *bio;
  	struct btrfs_device *dev;
  	DECLARE_COMPLETION_ONSTACK(compl);
  	u64 map_length = 0;
  	u64 sector;
  	struct btrfs_bio *bbio = NULL;
  	int ret;
  
  	BUG_ON(!mirror_num);
  
  	bio = bio_alloc(GFP_NOFS, 1);
  	if (!bio)
  		return -EIO;
  	bio->bi_private = &compl;
  	bio->bi_end_io = repair_io_failure_callback;
  	bio->bi_size = 0;
  	map_length = length;
  
  	ret = btrfs_map_block(map_tree, WRITE, logical,
  			      &map_length, &bbio, mirror_num);
  	if (ret) {
  		bio_put(bio);
  		return -EIO;
  	}
  	BUG_ON(mirror_num != bbio->mirror_num);
  	sector = bbio->stripes[mirror_num-1].physical >> 9;
  	bio->bi_sector = sector;
  	dev = bbio->stripes[mirror_num-1].dev;
  	kfree(bbio);
  	if (!dev || !dev->bdev || !dev->writeable) {
  		bio_put(bio);
  		return -EIO;
  	}
  	bio->bi_bdev = dev->bdev;
  	bio_add_page(bio, page, length, start-page_offset(page));
  	submit_bio(WRITE_SYNC, bio);
  	wait_for_completion(&compl);
  
  	if (!test_bit(BIO_UPTODATE, &bio->bi_flags)) {
  		/* try to remap that extent elsewhere? */
  		bio_put(bio);
  		return -EIO;
  	}
  
  	printk(KERN_INFO "btrfs read error corrected: ino %lu off %llu (dev %s "
  			"sector %llu)
  ", page->mapping->host->i_ino, start,
  			dev->name, sector);
  
  	bio_put(bio);
  	return 0;
  }
  
  /*
   * each time an IO finishes, we do a fast check in the IO failure tree
   * to see if we need to process or clean up an io_failure_record
   */
  static int clean_io_failure(u64 start, struct page *page)
  {
  	u64 private;
  	u64 private_failure;
  	struct io_failure_record *failrec;
  	struct btrfs_mapping_tree *map_tree;
  	struct extent_state *state;
  	int num_copies;
  	int did_repair = 0;
  	int ret;
  	struct inode *inode = page->mapping->host;
  
  	private = 0;
  	ret = count_range_bits(&BTRFS_I(inode)->io_failure_tree, &private,
  				(u64)-1, 1, EXTENT_DIRTY, 0);
  	if (!ret)
  		return 0;
  
  	ret = get_state_private(&BTRFS_I(inode)->io_failure_tree, start,
  				&private_failure);
  	if (ret)
  		return 0;
  
  	failrec = (struct io_failure_record *)(unsigned long) private_failure;
  	BUG_ON(!failrec->this_mirror);
  
  	if (failrec->in_validation) {
  		/* there was no real error, just free the record */
  		pr_debug("clean_io_failure: freeing dummy error at %llu
  ",
  			 failrec->start);
  		did_repair = 1;
  		goto out;
  	}
  
  	spin_lock(&BTRFS_I(inode)->io_tree.lock);
  	state = find_first_extent_bit_state(&BTRFS_I(inode)->io_tree,
  					    failrec->start,
  					    EXTENT_LOCKED);
  	spin_unlock(&BTRFS_I(inode)->io_tree.lock);
  
  	if (state && state->start == failrec->start) {
  		map_tree = &BTRFS_I(inode)->root->fs_info->mapping_tree;
  		num_copies = btrfs_num_copies(map_tree, failrec->logical,
  						failrec->len);
  		if (num_copies > 1)  {
  			ret = repair_io_failure(map_tree, start, failrec->len,
  						failrec->logical, page,
  						failrec->failed_mirror);
  			did_repair = !ret;
  		}
  	}
  
  out:
  	if (!ret)
  		ret = free_io_failure(inode, failrec, did_repair);
  
  	return ret;
  }
  
  /*
   * this is a generic handler for readpage errors (default
   * readpage_io_failed_hook). if other copies exist, read those and write back
   * good data to the failed position. does not investigate in remapping the
   * failed extent elsewhere, hoping the device will be smart enough to do this as
   * needed
   */
  
  static int bio_readpage_error(struct bio *failed_bio, struct page *page,
  				u64 start, u64 end, int failed_mirror,
  				struct extent_state *state)
  {
  	struct io_failure_record *failrec = NULL;
  	u64 private;
  	struct extent_map *em;
  	struct inode *inode = page->mapping->host;
  	struct extent_io_tree *failure_tree = &BTRFS_I(inode)->io_failure_tree;
  	struct extent_io_tree *tree = &BTRFS_I(inode)->io_tree;
  	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
  	struct bio *bio;
  	int num_copies;
  	int ret;
  	int read_mode;
  	u64 logical;
  
  	BUG_ON(failed_bio->bi_rw & REQ_WRITE);
  
  	ret = get_state_private(failure_tree, start, &private);
  	if (ret) {
  		failrec = kzalloc(sizeof(*failrec), GFP_NOFS);
  		if (!failrec)
  			return -ENOMEM;
  		failrec->start = start;
  		failrec->len = end - start + 1;
  		failrec->this_mirror = 0;
  		failrec->bio_flags = 0;
  		failrec->in_validation = 0;
  
  		read_lock(&em_tree->lock);
  		em = lookup_extent_mapping(em_tree, start, failrec->len);
  		if (!em) {
  			read_unlock(&em_tree->lock);
  			kfree(failrec);
  			return -EIO;
  		}
  
  		if (em->start > start || em->start + em->len < start) {
  			free_extent_map(em);
  			em = NULL;
  		}
  		read_unlock(&em_tree->lock);
  
  		if (!em || IS_ERR(em)) {
  			kfree(failrec);
  			return -EIO;
  		}
  		logical = start - em->start;
  		logical = em->block_start + logical;
  		if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags)) {
  			logical = em->block_start;
  			failrec->bio_flags = EXTENT_BIO_COMPRESSED;
  			extent_set_compress_type(&failrec->bio_flags,
  						 em->compress_type);
  		}
  		pr_debug("bio_readpage_error: (new) logical=%llu, start=%llu, "
  			 "len=%llu
  ", logical, start, failrec->len);
  		failrec->logical = logical;
  		free_extent_map(em);
  
  		/* set the bits in the private failure tree */
  		ret = set_extent_bits(failure_tree, start, end,
  					EXTENT_LOCKED | EXTENT_DIRTY, GFP_NOFS);
  		if (ret >= 0)
  			ret = set_state_private(failure_tree, start,
  						(u64)(unsigned long)failrec);
  		/* set the bits in the inode's tree */
  		if (ret >= 0)
  			ret = set_extent_bits(tree, start, end, EXTENT_DAMAGED,
  						GFP_NOFS);
  		if (ret < 0) {
  			kfree(failrec);
  			return ret;
  		}
  	} else {
  		failrec = (struct io_failure_record *)(unsigned long)private;
  		pr_debug("bio_readpage_error: (found) logical=%llu, "
  			 "start=%llu, len=%llu, validation=%d
  ",
  			 failrec->logical, failrec->start, failrec->len,
  			 failrec->in_validation);
  		/*
  		 * when data can be on disk more than twice, add to failrec here
  		 * (e.g. with a list for failed_mirror) to make
  		 * clean_io_failure() clean all those errors at once.
  		 */
  	}
  	num_copies = btrfs_num_copies(
  			      &BTRFS_I(inode)->root->fs_info->mapping_tree,
  			      failrec->logical, failrec->len);
  	if (num_copies == 1) {
  		/*
  		 * we only have a single copy of the data, so don't bother with
  		 * all the retry and error correction code that follows. no
  		 * matter what the error is, it is very likely to persist.
  		 */
  		pr_debug("bio_readpage_error: cannot repair, num_copies == 1. "
  			 "state=%p, num_copies=%d, next_mirror %d, "
  			 "failed_mirror %d
  ", state, num_copies,
  			 failrec->this_mirror, failed_mirror);
  		free_io_failure(inode, failrec, 0);
  		return -EIO;
  	}
  
  	if (!state) {
  		spin_lock(&tree->lock);
  		state = find_first_extent_bit_state(tree, failrec->start,
  						    EXTENT_LOCKED);
  		if (state && state->start != failrec->start)
  			state = NULL;
  		spin_unlock(&tree->lock);
  	}
  
  	/*
  	 * there are two premises:
  	 *	a) deliver good data to the caller
  	 *	b) correct the bad sectors on disk
  	 */
  	if (failed_bio->bi_vcnt > 1) {
  		/*
  		 * to fulfill b), we need to know the exact failing sectors, as
  		 * we don't want to rewrite any more than the failed ones. thus,
  		 * we need separate read requests for the failed bio
  		 *
  		 * if the following BUG_ON triggers, our validation request got
  		 * merged. we need separate requests for our algorithm to work.
  		 */
  		BUG_ON(failrec->in_validation);
  		failrec->in_validation = 1;
  		failrec->this_mirror = failed_mirror;
  		read_mode = READ_SYNC | REQ_FAILFAST_DEV;
  	} else {
  		/*
  		 * we're ready to fulfill a) and b) alongside. get a good copy
  		 * of the failed sector and if we succeed, we have setup
  		 * everything for repair_io_failure to do the rest for us.
  		 */
  		if (failrec->in_validation) {
  			BUG_ON(failrec->this_mirror != failed_mirror);
  			failrec->in_validation = 0;
  			failrec->this_mirror = 0;
  		}
  		failrec->failed_mirror = failed_mirror;
  		failrec->this_mirror++;
  		if (failrec->this_mirror == failed_mirror)
  			failrec->this_mirror++;
  		read_mode = READ_SYNC;
  	}
  
  	if (!state || failrec->this_mirror > num_copies) {
  		pr_debug("bio_readpage_error: (fail) state=%p, num_copies=%d, "
  			 "next_mirror %d, failed_mirror %d
  ", state,
  			 num_copies, failrec->this_mirror, failed_mirror);
  		free_io_failure(inode, failrec, 0);
  		return -EIO;
  	}
  
  	bio = bio_alloc(GFP_NOFS, 1);
  	bio->bi_private = state;
  	bio->bi_end_io = failed_bio->bi_end_io;
  	bio->bi_sector = failrec->logical >> 9;
  	bio->bi_bdev = BTRFS_I(inode)->root->fs_info->fs_devices->latest_bdev;
  	bio->bi_size = 0;
  
  	bio_add_page(bio, page, failrec->len, start - page_offset(page));
  
  	pr_debug("bio_readpage_error: submitting new read[%#x] to "
  		 "this_mirror=%d, num_copies=%d, in_validation=%d
  ", read_mode,
  		 failrec->this_mirror, num_copies, failrec->in_validation);
  
  	tree->ops->submit_bio_hook(inode, read_mode, bio, failrec->this_mirror,
  					failrec->bio_flags, 0);
  	return 0;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
  /* 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.
   */
d1310b2e0   Chris Mason   Btrfs: Split the ...
2138
  static void end_bio_extent_writepage(struct bio *bio, int err)
d1310b2e0   Chris Mason   Btrfs: Split the ...
2139
  {
1259ab75c   Chris Mason   Btrfs: Handle wri...
2140
  	int uptodate = err == 0;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2141
  	struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
902b22f34   David Woodhouse   Btrfs: Remove bro...
2142
  	struct extent_io_tree *tree;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2143
2144
2145
  	u64 start;
  	u64 end;
  	int whole_page;
1259ab75c   Chris Mason   Btrfs: Handle wri...
2146
  	int ret;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2147

d1310b2e0   Chris Mason   Btrfs: Split the ...
2148
2149
  	do {
  		struct page *page = bvec->bv_page;
902b22f34   David Woodhouse   Btrfs: Remove bro...
2150
  		tree = &BTRFS_I(page->mapping->host)->io_tree;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
  		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);
1259ab75c   Chris Mason   Btrfs: Handle wri...
2162
2163
  		if (tree->ops && tree->ops->writepage_end_io_hook) {
  			ret = tree->ops->writepage_end_io_hook(page, start,
902b22f34   David Woodhouse   Btrfs: Remove bro...
2164
  						       end, NULL, uptodate);
1259ab75c   Chris Mason   Btrfs: Handle wri...
2165
2166
2167
2168
2169
2170
2171
  			if (ret)
  				uptodate = 0;
  		}
  
  		if (!uptodate && tree->ops &&
  		    tree->ops->writepage_io_failed_hook) {
  			ret = tree->ops->writepage_io_failed_hook(bio, page,
902b22f34   David Woodhouse   Btrfs: Remove bro...
2172
  							 start, end, NULL);
1259ab75c   Chris Mason   Btrfs: Handle wri...
2173
  			if (ret == 0) {
1259ab75c   Chris Mason   Btrfs: Handle wri...
2174
2175
2176
2177
  				uptodate = (err == 0);
  				continue;
  			}
  		}
d1310b2e0   Chris Mason   Btrfs: Split the ...
2178
  		if (!uptodate) {
2ac55d41b   Josef Bacik   Btrfs: cache the ...
2179
  			clear_extent_uptodate(tree, start, end, NULL, GFP_NOFS);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2180
2181
2182
  			ClearPageUptodate(page);
  			SetPageError(page);
  		}
70dec8079   Chris Mason   Btrfs: extent_io ...
2183

d1310b2e0   Chris Mason   Btrfs: Split the ...
2184
2185
2186
2187
  		if (whole_page)
  			end_page_writeback(page);
  		else
  			check_page_writeback(tree, page);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2188
  	} while (bvec >= bio->bi_io_vec);
2b1f55b0f   Chris Mason   Remove Btrfs comp...
2189

d1310b2e0   Chris Mason   Btrfs: Split the ...
2190
  	bio_put(bio);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
  }
  
  /*
   * 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.
   */
d1310b2e0   Chris Mason   Btrfs: Split the ...
2204
  static void end_bio_extent_readpage(struct bio *bio, int err)
d1310b2e0   Chris Mason   Btrfs: Split the ...
2205
2206
  {
  	int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
4125bf761   Chris Mason   Btrfs: finish rea...
2207
2208
  	struct bio_vec *bvec_end = bio->bi_io_vec + bio->bi_vcnt - 1;
  	struct bio_vec *bvec = bio->bi_io_vec;
902b22f34   David Woodhouse   Btrfs: Remove bro...
2209
  	struct extent_io_tree *tree;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2210
2211
2212
2213
  	u64 start;
  	u64 end;
  	int whole_page;
  	int ret;
d20f7043f   Chris Mason   Btrfs: move data ...
2214
2215
  	if (err)
  		uptodate = 0;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2216
2217
  	do {
  		struct page *page = bvec->bv_page;
507903b81   Arne Jansen   btrfs: using cach...
2218
2219
  		struct extent_state *cached = NULL;
  		struct extent_state *state;
4a54c8c16   Jan Schmidt   btrfs: Moved repa...
2220
2221
2222
2223
  		pr_debug("end_bio_extent_readpage: bi_vcnt=%d, idx=%d, err=%d, "
  			 "mirror=%ld
  ", bio->bi_vcnt, bio->bi_idx, err,
  			 (long int)bio->bi_bdev);
902b22f34   David Woodhouse   Btrfs: Remove bro...
2224
  		tree = &BTRFS_I(page->mapping->host)->io_tree;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2225
2226
2227
2228
2229
2230
2231
2232
  		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;
4125bf761   Chris Mason   Btrfs: finish rea...
2233
  		if (++bvec <= bvec_end)
d1310b2e0   Chris Mason   Btrfs: Split the ...
2234
  			prefetchw(&bvec->bv_page->flags);
507903b81   Arne Jansen   btrfs: using cach...
2235
  		spin_lock(&tree->lock);
0d399205e   Chris Mason   Btrfs end_bio_ext...
2236
  		state = find_first_extent_bit_state(tree, start, EXTENT_LOCKED);
109b36a2b   Chris Mason   Btrfs: make uncac...
2237
  		if (state && state->start == start) {
507903b81   Arne Jansen   btrfs: using cach...
2238
2239
2240
2241
2242
2243
2244
  			/*
  			 * take a reference on the state, unlock will drop
  			 * the ref
  			 */
  			cache_state(state, &cached);
  		}
  		spin_unlock(&tree->lock);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2245
  		if (uptodate && tree->ops && tree->ops->readpage_end_io_hook) {
70dec8079   Chris Mason   Btrfs: extent_io ...
2246
  			ret = tree->ops->readpage_end_io_hook(page, start, end,
507903b81   Arne Jansen   btrfs: using cach...
2247
  							      state);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2248
2249
  			if (ret)
  				uptodate = 0;
4a54c8c16   Jan Schmidt   btrfs: Moved repa...
2250
2251
  			else
  				clean_io_failure(start, page);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2252
  		}
4a54c8c16   Jan Schmidt   btrfs: Moved repa...
2253
  		if (!uptodate) {
32240a913   Jan Schmidt   btrfs: mirror_num...
2254
2255
  			int failed_mirror;
  			failed_mirror = (int)(unsigned long)bio->bi_bdev;
f4a8e6563   Jan Schmidt   Btrfs: fix meta d...
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
  			/*
  			 * The generic bio_readpage_error handles errors the
  			 * following way: If possible, new read requests are
  			 * created and submitted and will end up in
  			 * end_bio_extent_readpage as well (if we're lucky, not
  			 * in the !uptodate case). In that case it returns 0 and
  			 * we just go on with the next page in our bio. If it
  			 * can't handle the error it will return -EIO and we
  			 * remain responsible for that page.
  			 */
  			ret = bio_readpage_error(bio, page, start, end,
  							failed_mirror, NULL);
7e38326f5   Chris Mason   Btrfs: Handle che...
2268
  			if (ret == 0) {
f4a8e6563   Jan Schmidt   Btrfs: fix meta d...
2269
  error_handled:
3b951516e   Chris Mason   Btrfs: Use the ex...
2270
2271
  				uptodate =
  					test_bit(BIO_UPTODATE, &bio->bi_flags);
d20f7043f   Chris Mason   Btrfs: move data ...
2272
2273
  				if (err)
  					uptodate = 0;
507903b81   Arne Jansen   btrfs: using cach...
2274
  				uncache_state(&cached);
7e38326f5   Chris Mason   Btrfs: Handle che...
2275
2276
  				continue;
  			}
f4a8e6563   Jan Schmidt   Btrfs: fix meta d...
2277
2278
2279
2280
2281
2282
2283
  			if (tree->ops && tree->ops->readpage_io_failed_hook) {
  				ret = tree->ops->readpage_io_failed_hook(
  							bio, page, start, end,
  							failed_mirror, state);
  				if (ret == 0)
  					goto error_handled;
  			}
7e38326f5   Chris Mason   Btrfs: Handle che...
2284
  		}
d1310b2e0   Chris Mason   Btrfs: Split the ...
2285

771ed689d   Chris Mason   Btrfs: Optimize c...
2286
  		if (uptodate) {
507903b81   Arne Jansen   btrfs: using cach...
2287
  			set_extent_uptodate(tree, start, end, &cached,
902b22f34   David Woodhouse   Btrfs: Remove bro...
2288
  					    GFP_ATOMIC);
771ed689d   Chris Mason   Btrfs: Optimize c...
2289
  		}
507903b81   Arne Jansen   btrfs: using cach...
2290
  		unlock_extent_cached(tree, start, end, &cached, GFP_ATOMIC);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2291

70dec8079   Chris Mason   Btrfs: extent_io ...
2292
2293
2294
2295
2296
2297
2298
  		if (whole_page) {
  			if (uptodate) {
  				SetPageUptodate(page);
  			} else {
  				ClearPageUptodate(page);
  				SetPageError(page);
  			}
d1310b2e0   Chris Mason   Btrfs: Split the ...
2299
  			unlock_page(page);
70dec8079   Chris Mason   Btrfs: extent_io ...
2300
2301
2302
2303
2304
2305
2306
  		} else {
  			if (uptodate) {
  				check_page_uptodate(tree, page);
  			} else {
  				ClearPageUptodate(page);
  				SetPageError(page);
  			}
d1310b2e0   Chris Mason   Btrfs: Split the ...
2307
  			check_page_locked(tree, page);
70dec8079   Chris Mason   Btrfs: extent_io ...
2308
  		}
4125bf761   Chris Mason   Btrfs: finish rea...
2309
  	} while (bvec <= bvec_end);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2310
2311
  
  	bio_put(bio);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2312
  }
88f794ede   Miao Xie   btrfs: cleanup du...
2313
2314
2315
  struct bio *
  btrfs_bio_alloc(struct block_device *bdev, u64 first_sector, int nr_vecs,
  		gfp_t gfp_flags)
d1310b2e0   Chris Mason   Btrfs: Split the ...
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
  {
  	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) {
e1c4b7451   Chris Mason   Fix btrfs_get_ext...
2327
  		bio->bi_size = 0;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2328
2329
2330
2331
2332
  		bio->bi_bdev = bdev;
  		bio->bi_sector = first_sector;
  	}
  	return bio;
  }
c8b978188   Chris Mason   Btrfs: Add zlib c...
2333
2334
  static int submit_one_bio(int rw, struct bio *bio, int mirror_num,
  			  unsigned long bio_flags)
d1310b2e0   Chris Mason   Btrfs: Split the ...
2335
  {
d1310b2e0   Chris Mason   Btrfs: Split the ...
2336
  	int ret = 0;
70dec8079   Chris Mason   Btrfs: extent_io ...
2337
2338
2339
  	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;
70dec8079   Chris Mason   Btrfs: extent_io ...
2340
  	u64 start;
70dec8079   Chris Mason   Btrfs: extent_io ...
2341
2342
  
  	start = ((u64)page->index << PAGE_CACHE_SHIFT) + bvec->bv_offset;
70dec8079   Chris Mason   Btrfs: extent_io ...
2343

902b22f34   David Woodhouse   Btrfs: Remove bro...
2344
  	bio->bi_private = NULL;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2345
2346
  
  	bio_get(bio);
065631f6d   Chris Mason   Btrfs: checksum f...
2347
  	if (tree->ops && tree->ops->submit_bio_hook)
6b82ce8d8   liubo   btrfs: fix unchec...
2348
  		ret = tree->ops->submit_bio_hook(page->mapping->host, rw, bio,
eaf25d933   Chris Mason   Btrfs: use async ...
2349
  					   mirror_num, bio_flags, start);
0b86a832a   Chris Mason   Btrfs: Add suppor...
2350
2351
  	else
  		submit_bio(rw, bio);
4a54c8c16   Jan Schmidt   btrfs: Moved repa...
2352

d1310b2e0   Chris Mason   Btrfs: Split the ...
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
  	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,
f188591e9   Chris Mason   Btrfs: Retry meta...
2365
  			      bio_end_io_t end_io_func,
c8b978188   Chris Mason   Btrfs: Add zlib c...
2366
2367
2368
  			      int mirror_num,
  			      unsigned long prev_bio_flags,
  			      unsigned long bio_flags)
d1310b2e0   Chris Mason   Btrfs: Split the ...
2369
2370
2371
2372
  {
  	int ret = 0;
  	struct bio *bio;
  	int nr;
c8b978188   Chris Mason   Btrfs: Add zlib c...
2373
2374
2375
  	int contig = 0;
  	int this_compressed = bio_flags & EXTENT_BIO_COMPRESSED;
  	int old_compressed = prev_bio_flags & EXTENT_BIO_COMPRESSED;
5b050f04c   Chris Mason   Btrfs: Fix compil...
2376
  	size_t page_size = min_t(size_t, size, PAGE_CACHE_SIZE);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2377
2378
2379
  
  	if (bio_ret && *bio_ret) {
  		bio = *bio_ret;
c8b978188   Chris Mason   Btrfs: Add zlib c...
2380
2381
2382
2383
2384
2385
2386
  		if (old_compressed)
  			contig = bio->bi_sector == sector;
  		else
  			contig = bio->bi_sector + (bio->bi_size >> 9) ==
  				sector;
  
  		if (prev_bio_flags != bio_flags || !contig ||
239b14b32   Chris Mason   Btrfs: Bring back...
2387
  		    (tree->ops && tree->ops->merge_bio_hook &&
c8b978188   Chris Mason   Btrfs: Add zlib c...
2388
2389
2390
2391
2392
  		     tree->ops->merge_bio_hook(page, offset, page_size, bio,
  					       bio_flags)) ||
  		    bio_add_page(bio, page, page_size, offset) < page_size) {
  			ret = submit_one_bio(rw, bio, mirror_num,
  					     prev_bio_flags);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2393
2394
2395
2396
2397
  			bio = NULL;
  		} else {
  			return 0;
  		}
  	}
c8b978188   Chris Mason   Btrfs: Add zlib c...
2398
2399
2400
2401
  	if (this_compressed)
  		nr = BIO_MAX_PAGES;
  	else
  		nr = bio_get_nr_vecs(bdev);
88f794ede   Miao Xie   btrfs: cleanup du...
2402
  	bio = btrfs_bio_alloc(bdev, sector, nr, GFP_NOFS | __GFP_HIGH);
5df670834   Tsutomu Itoh   btrfs: checking N...
2403
2404
  	if (!bio)
  		return -ENOMEM;
70dec8079   Chris Mason   Btrfs: extent_io ...
2405

c8b978188   Chris Mason   Btrfs: Add zlib c...
2406
  	bio_add_page(bio, page, page_size, offset);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2407
2408
  	bio->bi_end_io = end_io_func;
  	bio->bi_private = tree;
70dec8079   Chris Mason   Btrfs: extent_io ...
2409

d397712bc   Chris Mason   Btrfs: Fix checkp...
2410
  	if (bio_ret)
d1310b2e0   Chris Mason   Btrfs: Split the ...
2411
  		*bio_ret = bio;
d397712bc   Chris Mason   Btrfs: Fix checkp...
2412
  	else
c8b978188   Chris Mason   Btrfs: Add zlib c...
2413
  		ret = submit_one_bio(rw, bio, mirror_num, bio_flags);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2414
2415
2416
2417
2418
2419
2420
2421
  
  	return ret;
  }
  
  void set_page_extent_mapped(struct page *page)
  {
  	if (!PagePrivate(page)) {
  		SetPagePrivate(page);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2422
  		page_cache_get(page);
6af118ce5   Chris Mason   Btrfs: Index exte...
2423
  		set_page_private(page, EXTENT_PAGE_PRIVATE);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2424
2425
  	}
  }
b2950863c   Christoph Hellwig   Btrfs: make thing...
2426
  static void set_page_extent_head(struct page *page, unsigned long len)
d1310b2e0   Chris Mason   Btrfs: Split the ...
2427
  {
eb14ab8ed   Chris Mason   Btrfs: fix page->...
2428
  	WARN_ON(!PagePrivate(page));
d1310b2e0   Chris Mason   Btrfs: Split the ...
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
  	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,
c8b978188   Chris Mason   Btrfs: Add zlib c...
2440
2441
  				   struct bio **bio, int mirror_num,
  				   unsigned long *bio_flags)
d1310b2e0   Chris Mason   Btrfs: Split the ...
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
  {
  	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;
11c65dccf   Josef Bacik   Btrfs: do aio_wri...
2455
  	struct btrfs_ordered_extent *ordered;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2456
2457
  	int ret;
  	int nr = 0;
306e16ce1   David Sterba   btrfs: rename var...
2458
  	size_t pg_offset = 0;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2459
  	size_t iosize;
c8b978188   Chris Mason   Btrfs: Add zlib c...
2460
  	size_t disk_io_size;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2461
  	size_t blocksize = inode->i_sb->s_blocksize;
c8b978188   Chris Mason   Btrfs: Add zlib c...
2462
  	unsigned long this_bio_flag = 0;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2463
2464
  
  	set_page_extent_mapped(page);
90a887c9a   Dan Magenheimer   btrfs: add cleanc...
2465
2466
2467
2468
2469
2470
  	if (!PageUptodate(page)) {
  		if (cleancache_get_page(page) == 0) {
  			BUG_ON(blocksize != PAGE_SIZE);
  			goto out;
  		}
  	}
d1310b2e0   Chris Mason   Btrfs: Split the ...
2471
  	end = page_end;
11c65dccf   Josef Bacik   Btrfs: do aio_wri...
2472
2473
2474
2475
2476
2477
2478
2479
2480
  	while (1) {
  		lock_extent(tree, start, end, GFP_NOFS);
  		ordered = btrfs_lookup_ordered_extent(inode, start);
  		if (!ordered)
  			break;
  		unlock_extent(tree, start, end, GFP_NOFS);
  		btrfs_start_ordered_extent(inode, ordered, 1);
  		btrfs_put_ordered_extent(ordered);
  	}
d1310b2e0   Chris Mason   Btrfs: Split the ...
2481

c8b978188   Chris Mason   Btrfs: Add zlib c...
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
  	if (page->index == last_byte >> PAGE_CACHE_SHIFT) {
  		char *userpage;
  		size_t zero_offset = last_byte & (PAGE_CACHE_SIZE - 1);
  
  		if (zero_offset) {
  			iosize = PAGE_CACHE_SIZE - zero_offset;
  			userpage = kmap_atomic(page, KM_USER0);
  			memset(userpage + zero_offset, 0, iosize);
  			flush_dcache_page(page);
  			kunmap_atomic(userpage, KM_USER0);
  		}
  	}
d1310b2e0   Chris Mason   Btrfs: Split the ...
2494
2495
2496
  	while (cur <= end) {
  		if (cur >= last_byte) {
  			char *userpage;
507903b81   Arne Jansen   btrfs: using cach...
2497
  			struct extent_state *cached = NULL;
306e16ce1   David Sterba   btrfs: rename var...
2498
  			iosize = PAGE_CACHE_SIZE - pg_offset;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2499
  			userpage = kmap_atomic(page, KM_USER0);
306e16ce1   David Sterba   btrfs: rename var...
2500
  			memset(userpage + pg_offset, 0, iosize);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2501
2502
2503
  			flush_dcache_page(page);
  			kunmap_atomic(userpage, KM_USER0);
  			set_extent_uptodate(tree, cur, cur + iosize - 1,
507903b81   Arne Jansen   btrfs: using cach...
2504
2505
2506
  					    &cached, GFP_NOFS);
  			unlock_extent_cached(tree, cur, cur + iosize - 1,
  					     &cached, GFP_NOFS);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2507
2508
  			break;
  		}
306e16ce1   David Sterba   btrfs: rename var...
2509
  		em = get_extent(inode, page, pg_offset, cur,
d1310b2e0   Chris Mason   Btrfs: Split the ...
2510
  				end - cur + 1, 0);
c704005d8   David Sterba   btrfs: unify chec...
2511
  		if (IS_ERR_OR_NULL(em)) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
2512
2513
2514
2515
  			SetPageError(page);
  			unlock_extent(tree, cur, end, GFP_NOFS);
  			break;
  		}
d1310b2e0   Chris Mason   Btrfs: Split the ...
2516
2517
2518
  		extent_offset = cur - em->start;
  		BUG_ON(extent_map_end(em) <= cur);
  		BUG_ON(end < cur);
261507a02   Li Zefan   btrfs: Allow to a...
2519
  		if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags)) {
c8b978188   Chris Mason   Btrfs: Add zlib c...
2520
  			this_bio_flag = EXTENT_BIO_COMPRESSED;
261507a02   Li Zefan   btrfs: Allow to a...
2521
2522
2523
  			extent_set_compress_type(&this_bio_flag,
  						 em->compress_type);
  		}
c8b978188   Chris Mason   Btrfs: Add zlib c...
2524

d1310b2e0   Chris Mason   Btrfs: Split the ...
2525
2526
2527
  		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);
c8b978188   Chris Mason   Btrfs: Add zlib c...
2528
2529
2530
2531
2532
2533
2534
  		if (this_bio_flag & EXTENT_BIO_COMPRESSED) {
  			disk_io_size = em->block_len;
  			sector = em->block_start >> 9;
  		} else {
  			sector = (em->block_start + extent_offset) >> 9;
  			disk_io_size = iosize;
  		}
d1310b2e0   Chris Mason   Btrfs: Split the ...
2535
2536
  		bdev = em->bdev;
  		block_start = em->block_start;
d899e0521   Yan Zheng   Btrfs: Add falloc...
2537
2538
  		if (test_bit(EXTENT_FLAG_PREALLOC, &em->flags))
  			block_start = EXTENT_MAP_HOLE;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2539
2540
2541
2542
2543
2544
  		free_extent_map(em);
  		em = NULL;
  
  		/* we've found a hole, just zero and go on */
  		if (block_start == EXTENT_MAP_HOLE) {
  			char *userpage;
507903b81   Arne Jansen   btrfs: using cach...
2545
  			struct extent_state *cached = NULL;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2546
  			userpage = kmap_atomic(page, KM_USER0);
306e16ce1   David Sterba   btrfs: rename var...
2547
  			memset(userpage + pg_offset, 0, iosize);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2548
2549
2550
2551
  			flush_dcache_page(page);
  			kunmap_atomic(userpage, KM_USER0);
  
  			set_extent_uptodate(tree, cur, cur + iosize - 1,
507903b81   Arne Jansen   btrfs: using cach...
2552
2553
2554
  					    &cached, GFP_NOFS);
  			unlock_extent_cached(tree, cur, cur + iosize - 1,
  			                     &cached, GFP_NOFS);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2555
  			cur = cur + iosize;
306e16ce1   David Sterba   btrfs: rename var...
2556
  			pg_offset += iosize;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2557
2558
2559
  			continue;
  		}
  		/* the get_extent function already copied into the page */
9655d2982   Chris Mason   Btrfs: use a cach...
2560
2561
  		if (test_range_bit(tree, cur, cur_end,
  				   EXTENT_UPTODATE, 1, NULL)) {
a1b32a593   Chris Mason   Btrfs: Add debugg...
2562
  			check_page_uptodate(tree, page);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2563
2564
  			unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS);
  			cur = cur + iosize;
306e16ce1   David Sterba   btrfs: rename var...
2565
  			pg_offset += iosize;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2566
2567
  			continue;
  		}
70dec8079   Chris Mason   Btrfs: extent_io ...
2568
2569
2570
2571
2572
2573
2574
  		/* 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;
306e16ce1   David Sterba   btrfs: rename var...
2575
  			pg_offset += iosize;
70dec8079   Chris Mason   Btrfs: extent_io ...
2576
2577
  			continue;
  		}
d1310b2e0   Chris Mason   Btrfs: Split the ...
2578
2579
2580
2581
2582
2583
2584
  
  		ret = 0;
  		if (tree->ops && tree->ops->readpage_io_hook) {
  			ret = tree->ops->readpage_io_hook(page, cur,
  							  cur + iosize - 1);
  		}
  		if (!ret) {
89642229a   Chris Mason   Btrfs: Search dat...
2585
2586
  			unsigned long pnr = (last_byte >> PAGE_CACHE_SHIFT) + 1;
  			pnr -= page->index;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2587
  			ret = submit_extent_page(READ, tree, page,
306e16ce1   David Sterba   btrfs: rename var...
2588
  					 sector, disk_io_size, pg_offset,
89642229a   Chris Mason   Btrfs: Search dat...
2589
  					 bdev, bio, pnr,
c8b978188   Chris Mason   Btrfs: Add zlib c...
2590
2591
2592
  					 end_bio_extent_readpage, mirror_num,
  					 *bio_flags,
  					 this_bio_flag);
89642229a   Chris Mason   Btrfs: Search dat...
2593
  			nr++;
c8b978188   Chris Mason   Btrfs: Add zlib c...
2594
  			*bio_flags = this_bio_flag;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2595
2596
2597
2598
  		}
  		if (ret)
  			SetPageError(page);
  		cur = cur + iosize;
306e16ce1   David Sterba   btrfs: rename var...
2599
  		pg_offset += iosize;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2600
  	}
90a887c9a   Dan Magenheimer   btrfs: add cleanc...
2601
  out:
d1310b2e0   Chris Mason   Btrfs: Split the ...
2602
2603
2604
2605
2606
2607
2608
2609
2610
  	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,
8ddc7d9cd   Jan Schmidt   btrfs: add mirror...
2611
  			    get_extent_t *get_extent, int mirror_num)
d1310b2e0   Chris Mason   Btrfs: Split the ...
2612
2613
  {
  	struct bio *bio = NULL;
c8b978188   Chris Mason   Btrfs: Add zlib c...
2614
  	unsigned long bio_flags = 0;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2615
  	int ret;
8ddc7d9cd   Jan Schmidt   btrfs: add mirror...
2616
  	ret = __extent_read_full_page(tree, page, get_extent, &bio, mirror_num,
c8b978188   Chris Mason   Btrfs: Add zlib c...
2617
  				      &bio_flags);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2618
  	if (bio)
8ddc7d9cd   Jan Schmidt   btrfs: add mirror...
2619
  		ret = submit_one_bio(READ, bio, mirror_num, bio_flags);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2620
2621
  	return ret;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
2622

11c8349b4   Chris Mason   Btrfs: fix oops o...
2623
2624
2625
2626
2627
2628
2629
2630
2631
  static noinline void update_nr_written(struct page *page,
  				      struct writeback_control *wbc,
  				      unsigned long nr_written)
  {
  	wbc->nr_to_write -= nr_written;
  	if (wbc->range_cyclic || (wbc->nr_to_write > 0 &&
  	    wbc->range_start == 0 && wbc->range_end == LLONG_MAX))
  		page->mapping->writeback_index = page->index + nr_written;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
  /*
   * 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;
  	sector_t sector;
2c64c53d8   Chris Mason   Btrfs: cache valu...
2654
  	struct extent_state *cached_state = NULL;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2655
2656
2657
2658
  	struct extent_map *em;
  	struct block_device *bdev;
  	int ret;
  	int nr = 0;
7f3c74fb8   Chris Mason   Btrfs: Keep exten...
2659
  	size_t pg_offset = 0;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2660
2661
2662
2663
2664
  	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;
c8b978188   Chris Mason   Btrfs: Add zlib c...
2665
2666
  	int page_started;
  	int compressed;
ffbd517d5   Chris Mason   Btrfs: use WRITE_...
2667
  	int write_flags;
771ed689d   Chris Mason   Btrfs: Optimize c...
2668
  	unsigned long nr_written = 0;
9e4871070   Josef Bacik   Btrfs: skip looki...
2669
  	bool fill_delalloc = true;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2670

ffbd517d5   Chris Mason   Btrfs: use WRITE_...
2671
  	if (wbc->sync_mode == WB_SYNC_ALL)
721a9602e   Jens Axboe   block: kill off R...
2672
  		write_flags = WRITE_SYNC;
ffbd517d5   Chris Mason   Btrfs: use WRITE_...
2673
2674
  	else
  		write_flags = WRITE;
1abe9b8a1   liubo   Btrfs: add initia...
2675
  	trace___extent_writepage(page, inode, wbc);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2676
  	WARN_ON(!PageLocked(page));
bf0da8c18   Chris Mason   Btrfs: ClearPageE...
2677
2678
  
  	ClearPageError(page);
7f3c74fb8   Chris Mason   Btrfs: Keep exten...
2679
  	pg_offset = i_size & (PAGE_CACHE_SIZE - 1);
211c17f51   Chris Mason   Fix corners in wr...
2680
  	if (page->index > end_index ||
7f3c74fb8   Chris Mason   Btrfs: Keep exten...
2681
  	   (page->index == end_index && !pg_offset)) {
39be25cd8   Chris Mason   Btrfs: Use invali...
2682
  		page->mapping->a_ops->invalidatepage(page, 0);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2683
2684
2685
2686
2687
2688
  		unlock_page(page);
  		return 0;
  	}
  
  	if (page->index == end_index) {
  		char *userpage;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2689
  		userpage = kmap_atomic(page, KM_USER0);
7f3c74fb8   Chris Mason   Btrfs: Keep exten...
2690
2691
  		memset(userpage + pg_offset, 0,
  		       PAGE_CACHE_SIZE - pg_offset);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2692
  		kunmap_atomic(userpage, KM_USER0);
211c17f51   Chris Mason   Fix corners in wr...
2693
  		flush_dcache_page(page);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2694
  	}
7f3c74fb8   Chris Mason   Btrfs: Keep exten...
2695
  	pg_offset = 0;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2696
2697
  
  	set_page_extent_mapped(page);
9e4871070   Josef Bacik   Btrfs: skip looki...
2698
2699
  	if (!tree->ops || !tree->ops->fill_delalloc)
  		fill_delalloc = false;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2700
2701
  	delalloc_start = start;
  	delalloc_end = 0;
c8b978188   Chris Mason   Btrfs: Add zlib c...
2702
  	page_started = 0;
9e4871070   Josef Bacik   Btrfs: skip looki...
2703
  	if (!epd->extent_locked && fill_delalloc) {
f85d7d6c8   Chris Mason   Btrfs: properly h...
2704
  		u64 delalloc_to_write = 0;
11c8349b4   Chris Mason   Btrfs: fix oops o...
2705
2706
2707
2708
2709
  		/*
  		 * make sure the wbc mapping index is at least updated
  		 * to this page.
  		 */
  		update_nr_written(page, wbc, 0);
d397712bc   Chris Mason   Btrfs: Fix checkp...
2710
  		while (delalloc_end < page_end) {
771ed689d   Chris Mason   Btrfs: Optimize c...
2711
  			nr_delalloc = find_lock_delalloc_range(inode, tree,
c8b978188   Chris Mason   Btrfs: Add zlib c...
2712
2713
  						       page,
  						       &delalloc_start,
d1310b2e0   Chris Mason   Btrfs: Split the ...
2714
2715
  						       &delalloc_end,
  						       128 * 1024 * 1024);
771ed689d   Chris Mason   Btrfs: Optimize c...
2716
2717
2718
2719
2720
2721
2722
  			if (nr_delalloc == 0) {
  				delalloc_start = delalloc_end + 1;
  				continue;
  			}
  			tree->ops->fill_delalloc(inode, page, delalloc_start,
  						 delalloc_end, &page_started,
  						 &nr_written);
f85d7d6c8   Chris Mason   Btrfs: properly h...
2723
2724
2725
2726
2727
2728
2729
2730
  			/*
  			 * delalloc_end is already one less than the total
  			 * length, so we don't subtract one from
  			 * PAGE_CACHE_SIZE
  			 */
  			delalloc_to_write += (delalloc_end - delalloc_start +
  					      PAGE_CACHE_SIZE) >>
  					      PAGE_CACHE_SHIFT;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2731
  			delalloc_start = delalloc_end + 1;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2732
  		}
f85d7d6c8   Chris Mason   Btrfs: properly h...
2733
2734
2735
2736
2737
2738
2739
2740
  		if (wbc->nr_to_write < delalloc_to_write) {
  			int thresh = 8192;
  
  			if (delalloc_to_write < thresh * 2)
  				thresh = delalloc_to_write;
  			wbc->nr_to_write = min_t(u64, delalloc_to_write,
  						 thresh);
  		}
c8b978188   Chris Mason   Btrfs: Add zlib c...
2741

771ed689d   Chris Mason   Btrfs: Optimize c...
2742
2743
2744
2745
2746
  		/* did the fill delalloc function already unlock and start
  		 * the IO?
  		 */
  		if (page_started) {
  			ret = 0;
11c8349b4   Chris Mason   Btrfs: fix oops o...
2747
2748
2749
2750
2751
2752
2753
  			/*
  			 * we've unlocked the page, so we can't update
  			 * the mapping's writeback index, just update
  			 * nr_to_write.
  			 */
  			wbc->nr_to_write -= nr_written;
  			goto done_unlocked;
771ed689d   Chris Mason   Btrfs: Optimize c...
2754
  		}
c8b978188   Chris Mason   Btrfs: Add zlib c...
2755
  	}
247e743cb   Chris Mason   Btrfs: Use async ...
2756
  	if (tree->ops && tree->ops->writepage_start_hook) {
c8b978188   Chris Mason   Btrfs: Add zlib c...
2757
2758
  		ret = tree->ops->writepage_start_hook(page, start,
  						      page_end);
247e743cb   Chris Mason   Btrfs: Use async ...
2759
  		if (ret == -EAGAIN) {
247e743cb   Chris Mason   Btrfs: Use async ...
2760
  			redirty_page_for_writepage(wbc, page);
11c8349b4   Chris Mason   Btrfs: fix oops o...
2761
  			update_nr_written(page, wbc, nr_written);
247e743cb   Chris Mason   Btrfs: Use async ...
2762
  			unlock_page(page);
771ed689d   Chris Mason   Btrfs: Optimize c...
2763
  			ret = 0;
11c8349b4   Chris Mason   Btrfs: fix oops o...
2764
  			goto done_unlocked;
247e743cb   Chris Mason   Btrfs: Use async ...
2765
2766
  		}
  	}
11c8349b4   Chris Mason   Btrfs: fix oops o...
2767
2768
2769
2770
2771
  	/*
  	 * we don't want to touch the inode after unlocking the page,
  	 * so we update the mapping writeback index now
  	 */
  	update_nr_written(page, wbc, nr_written + 1);
771ed689d   Chris Mason   Btrfs: Optimize c...
2772

d1310b2e0   Chris Mason   Btrfs: Split the ...
2773
  	end = page_end;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2774
  	if (last_byte <= start) {
e6dcd2dc9   Chris Mason   Btrfs: New data=o...
2775
2776
2777
  		if (tree->ops && tree->ops->writepage_end_io_hook)
  			tree->ops->writepage_end_io_hook(page, start,
  							 page_end, NULL, 1);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2778
2779
  		goto done;
  	}
d1310b2e0   Chris Mason   Btrfs: Split the ...
2780
2781
2782
2783
  	blocksize = inode->i_sb->s_blocksize;
  
  	while (cur <= end) {
  		if (cur >= last_byte) {
e6dcd2dc9   Chris Mason   Btrfs: New data=o...
2784
2785
2786
  			if (tree->ops && tree->ops->writepage_end_io_hook)
  				tree->ops->writepage_end_io_hook(page, cur,
  							 page_end, NULL, 1);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2787
2788
  			break;
  		}
7f3c74fb8   Chris Mason   Btrfs: Keep exten...
2789
  		em = epd->get_extent(inode, page, pg_offset, cur,
d1310b2e0   Chris Mason   Btrfs: Split the ...
2790
  				     end - cur + 1, 1);
c704005d8   David Sterba   btrfs: unify chec...
2791
  		if (IS_ERR_OR_NULL(em)) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
2792
2793
2794
2795
2796
2797
2798
2799
2800
2801
2802
2803
  			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;
c8b978188   Chris Mason   Btrfs: Add zlib c...
2804
  		compressed = test_bit(EXTENT_FLAG_COMPRESSED, &em->flags);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2805
2806
  		free_extent_map(em);
  		em = NULL;
c8b978188   Chris Mason   Btrfs: Add zlib c...
2807
2808
2809
2810
2811
  		/*
  		 * compressed and inline extents are written through other
  		 * paths in the FS
  		 */
  		if (compressed || block_start == EXTENT_MAP_HOLE ||
d1310b2e0   Chris Mason   Btrfs: Split the ...
2812
  		    block_start == EXTENT_MAP_INLINE) {
c8b978188   Chris Mason   Btrfs: Add zlib c...
2813
2814
2815
2816
2817
2818
  			/*
  			 * end_io notification does not happen here for
  			 * compressed extents
  			 */
  			if (!compressed && tree->ops &&
  			    tree->ops->writepage_end_io_hook)
e6dcd2dc9   Chris Mason   Btrfs: New data=o...
2819
2820
2821
  				tree->ops->writepage_end_io_hook(page, cur,
  							 cur + iosize - 1,
  							 NULL, 1);
c8b978188   Chris Mason   Btrfs: Add zlib c...
2822
2823
2824
2825
2826
2827
2828
2829
2830
  			else if (compressed) {
  				/* we don't want to end_page_writeback on
  				 * a compressed extent.  this happens
  				 * elsewhere
  				 */
  				nr++;
  			}
  
  			cur += iosize;
7f3c74fb8   Chris Mason   Btrfs: Keep exten...
2831
  			pg_offset += iosize;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2832
2833
  			continue;
  		}
d1310b2e0   Chris Mason   Btrfs: Split the ...
2834
2835
  		/* leave this out until we have a page_mkwrite call */
  		if (0 && !test_range_bit(tree, cur, cur + iosize - 1,
9655d2982   Chris Mason   Btrfs: use a cach...
2836
  				   EXTENT_DIRTY, 0, NULL)) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
2837
  			cur = cur + iosize;
7f3c74fb8   Chris Mason   Btrfs: Keep exten...
2838
  			pg_offset += iosize;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2839
2840
  			continue;
  		}
c8b978188   Chris Mason   Btrfs: Add zlib c...
2841

d1310b2e0   Chris Mason   Btrfs: Split the ...
2842
2843
2844
2845
2846
2847
  		if (tree->ops && tree->ops->writepage_io_hook) {
  			ret = tree->ops->writepage_io_hook(page, cur,
  						cur + iosize - 1);
  		} else {
  			ret = 0;
  		}
1259ab75c   Chris Mason   Btrfs: Handle wri...
2848
  		if (ret) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
2849
  			SetPageError(page);
1259ab75c   Chris Mason   Btrfs: Handle wri...
2850
  		} else {
d1310b2e0   Chris Mason   Btrfs: Split the ...
2851
  			unsigned long max_nr = end_index + 1;
7f3c74fb8   Chris Mason   Btrfs: Keep exten...
2852

d1310b2e0   Chris Mason   Btrfs: Split the ...
2853
2854
  			set_range_writeback(tree, cur, cur + iosize - 1);
  			if (!PageWriteback(page)) {
d397712bc   Chris Mason   Btrfs: Fix checkp...
2855
2856
2857
2858
  				printk(KERN_ERR "btrfs warning page %lu not "
  				       "writeback, cur %llu end %llu
  ",
  				       page->index, (unsigned long long)cur,
d1310b2e0   Chris Mason   Btrfs: Split the ...
2859
2860
  				       (unsigned long long)end);
  			}
ffbd517d5   Chris Mason   Btrfs: use WRITE_...
2861
2862
2863
  			ret = submit_extent_page(write_flags, tree, page,
  						 sector, iosize, pg_offset,
  						 bdev, &epd->bio, max_nr,
c8b978188   Chris Mason   Btrfs: Add zlib c...
2864
2865
  						 end_bio_extent_writepage,
  						 0, 0, 0);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2866
2867
2868
2869
  			if (ret)
  				SetPageError(page);
  		}
  		cur = cur + iosize;
7f3c74fb8   Chris Mason   Btrfs: Keep exten...
2870
  		pg_offset += iosize;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2871
2872
2873
2874
2875
2876
2877
2878
  		nr++;
  	}
  done:
  	if (nr == 0) {
  		/* make sure the mapping tag for page dirty gets cleared */
  		set_page_writeback(page);
  		end_page_writeback(page);
  	}
d1310b2e0   Chris Mason   Btrfs: Split the ...
2879
  	unlock_page(page);
771ed689d   Chris Mason   Btrfs: Optimize c...
2880

11c8349b4   Chris Mason   Btrfs: fix oops o...
2881
  done_unlocked:
2c64c53d8   Chris Mason   Btrfs: cache valu...
2882
2883
  	/* drop our reference on any cached states */
  	free_extent_state(cached_state);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2884
2885
  	return 0;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
2886
  /**
4bef08485   Chris Mason   Btrfs: Tree loggi...
2887
   * write_cache_pages - walk the list of dirty pages of the given address space and write all of them.
d1310b2e0   Chris Mason   Btrfs: Split the ...
2888
2889
2890
2891
2892
2893
2894
2895
2896
2897
2898
2899
2900
   * @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.
   */
b2950863c   Christoph Hellwig   Btrfs: make thing...
2901
  static int extent_write_cache_pages(struct extent_io_tree *tree,
4bef08485   Chris Mason   Btrfs: Tree loggi...
2902
2903
  			     struct address_space *mapping,
  			     struct writeback_control *wbc,
d2c3f4f69   Chris Mason   Btrfs: Avoid writ...
2904
2905
  			     writepage_t writepage, void *data,
  			     void (*flush_fn)(void *))
d1310b2e0   Chris Mason   Btrfs: Split the ...
2906
  {
d1310b2e0   Chris Mason   Btrfs: Split the ...
2907
2908
  	int ret = 0;
  	int done = 0;
f85d7d6c8   Chris Mason   Btrfs: properly h...
2909
  	int nr_to_write_done = 0;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2910
2911
2912
2913
2914
  	struct pagevec pvec;
  	int nr_pages;
  	pgoff_t index;
  	pgoff_t end;		/* Inclusive */
  	int scanned = 0;
f7aaa06bf   Josef Bacik   Btrfs: tag pages ...
2915
  	int tag;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2916

d1310b2e0   Chris Mason   Btrfs: Split the ...
2917
2918
2919
2920
2921
2922
2923
  	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;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2924
2925
  		scanned = 1;
  	}
f7aaa06bf   Josef Bacik   Btrfs: tag pages ...
2926
2927
2928
2929
  	if (wbc->sync_mode == WB_SYNC_ALL)
  		tag = PAGECACHE_TAG_TOWRITE;
  	else
  		tag = PAGECACHE_TAG_DIRTY;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2930
  retry:
f7aaa06bf   Josef Bacik   Btrfs: tag pages ...
2931
2932
  	if (wbc->sync_mode == WB_SYNC_ALL)
  		tag_pages_for_writeback(mapping, index, end);
f85d7d6c8   Chris Mason   Btrfs: properly h...
2933
  	while (!done && !nr_to_write_done && (index <= end) &&
f7aaa06bf   Josef Bacik   Btrfs: tag pages ...
2934
2935
  	       (nr_pages = pagevec_lookup_tag(&pvec, mapping, &index, tag,
  			min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1))) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
2936
2937
2938
2939
2940
2941
2942
2943
2944
2945
2946
2947
2948
  		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
  			 */
01d658f2c   Chris Mason   Btrfs: make sure ...
2949
2950
2951
2952
2953
2954
2955
2956
2957
2958
  			if (tree->ops &&
  			    tree->ops->write_cache_pages_lock_hook) {
  				tree->ops->write_cache_pages_lock_hook(page,
  							       data, flush_fn);
  			} else {
  				if (!trylock_page(page)) {
  					flush_fn(data);
  					lock_page(page);
  				}
  			}
d1310b2e0   Chris Mason   Btrfs: Split the ...
2959
2960
2961
2962
2963
2964
2965
2966
2967
2968
2969
  
  			if (unlikely(page->mapping != mapping)) {
  				unlock_page(page);
  				continue;
  			}
  
  			if (!wbc->range_cyclic && page->index > end) {
  				done = 1;
  				unlock_page(page);
  				continue;
  			}
d2c3f4f69   Chris Mason   Btrfs: Avoid writ...
2970
  			if (wbc->sync_mode != WB_SYNC_NONE) {
0e6bd956e   Chris Mason   Btrfs: only flush...
2971
2972
  				if (PageWriteback(page))
  					flush_fn(data);
d1310b2e0   Chris Mason   Btrfs: Split the ...
2973
  				wait_on_page_writeback(page);
d2c3f4f69   Chris Mason   Btrfs: Avoid writ...
2974
  			}
d1310b2e0   Chris Mason   Btrfs: Split the ...
2975
2976
2977
2978
2979
2980
2981
2982
2983
2984
2985
2986
2987
  
  			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;
  			}
f85d7d6c8   Chris Mason   Btrfs: properly h...
2988
  			if (ret)
d1310b2e0   Chris Mason   Btrfs: Split the ...
2989
  				done = 1;
f85d7d6c8   Chris Mason   Btrfs: properly h...
2990
2991
2992
2993
2994
2995
2996
  
  			/*
  			 * the filesystem may choose to bump up nr_to_write.
  			 * We have to make sure to honor the new nr_to_write
  			 * at any time
  			 */
  			nr_to_write_done = wbc->nr_to_write <= 0;
d1310b2e0   Chris Mason   Btrfs: Split the ...
2997
2998
2999
3000
3001
3002
3003
3004
3005
3006
3007
3008
3009
  		}
  		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;
  	}
d1310b2e0   Chris Mason   Btrfs: Split the ...
3010
3011
  	return ret;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
3012

ffbd517d5   Chris Mason   Btrfs: use WRITE_...
3013
  static void flush_epd_write_bio(struct extent_page_data *epd)
d2c3f4f69   Chris Mason   Btrfs: Avoid writ...
3014
  {
d2c3f4f69   Chris Mason   Btrfs: Avoid writ...
3015
  	if (epd->bio) {
ffbd517d5   Chris Mason   Btrfs: use WRITE_...
3016
3017
3018
3019
  		if (epd->sync_io)
  			submit_one_bio(WRITE_SYNC, epd->bio, 0, 0);
  		else
  			submit_one_bio(WRITE, epd->bio, 0, 0);
d2c3f4f69   Chris Mason   Btrfs: Avoid writ...
3020
3021
3022
  		epd->bio = NULL;
  	}
  }
ffbd517d5   Chris Mason   Btrfs: use WRITE_...
3023
3024
3025
3026
3027
  static noinline void flush_write_bio(void *data)
  {
  	struct extent_page_data *epd = data;
  	flush_epd_write_bio(epd);
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
3028
3029
3030
3031
3032
  int extent_write_full_page(struct extent_io_tree *tree, struct page *page,
  			  get_extent_t *get_extent,
  			  struct writeback_control *wbc)
  {
  	int ret;
d1310b2e0   Chris Mason   Btrfs: Split the ...
3033
3034
3035
3036
  	struct extent_page_data epd = {
  		.bio = NULL,
  		.tree = tree,
  		.get_extent = get_extent,
771ed689d   Chris Mason   Btrfs: Optimize c...
3037
  		.extent_locked = 0,
ffbd517d5   Chris Mason   Btrfs: use WRITE_...
3038
  		.sync_io = wbc->sync_mode == WB_SYNC_ALL,
d1310b2e0   Chris Mason   Btrfs: Split the ...
3039
  	};
d1310b2e0   Chris Mason   Btrfs: Split the ...
3040

d1310b2e0   Chris Mason   Btrfs: Split the ...
3041
  	ret = __extent_writepage(page, wbc, &epd);
ffbd517d5   Chris Mason   Btrfs: use WRITE_...
3042
  	flush_epd_write_bio(&epd);
d1310b2e0   Chris Mason   Btrfs: Split the ...
3043
3044
  	return ret;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
3045

771ed689d   Chris Mason   Btrfs: Optimize c...
3046
3047
3048
3049
3050
3051
3052
3053
3054
3055
3056
3057
3058
3059
3060
  int extent_write_locked_range(struct extent_io_tree *tree, struct inode *inode,
  			      u64 start, u64 end, get_extent_t *get_extent,
  			      int mode)
  {
  	int ret = 0;
  	struct address_space *mapping = inode->i_mapping;
  	struct page *page;
  	unsigned long nr_pages = (end - start + PAGE_CACHE_SIZE) >>
  		PAGE_CACHE_SHIFT;
  
  	struct extent_page_data epd = {
  		.bio = NULL,
  		.tree = tree,
  		.get_extent = get_extent,
  		.extent_locked = 1,
ffbd517d5   Chris Mason   Btrfs: use WRITE_...
3061
  		.sync_io = mode == WB_SYNC_ALL,
771ed689d   Chris Mason   Btrfs: Optimize c...
3062
3063
  	};
  	struct writeback_control wbc_writepages = {
771ed689d   Chris Mason   Btrfs: Optimize c...
3064
  		.sync_mode	= mode,
771ed689d   Chris Mason   Btrfs: Optimize c...
3065
3066
3067
3068
  		.nr_to_write	= nr_pages * 2,
  		.range_start	= start,
  		.range_end	= end + 1,
  	};
d397712bc   Chris Mason   Btrfs: Fix checkp...
3069
  	while (start <= end) {
771ed689d   Chris Mason   Btrfs: Optimize c...
3070
3071
3072
3073
3074
3075
3076
3077
3078
3079
3080
3081
3082
  		page = find_get_page(mapping, start >> PAGE_CACHE_SHIFT);
  		if (clear_page_dirty_for_io(page))
  			ret = __extent_writepage(page, &wbc_writepages, &epd);
  		else {
  			if (tree->ops && tree->ops->writepage_end_io_hook)
  				tree->ops->writepage_end_io_hook(page, start,
  						 start + PAGE_CACHE_SIZE - 1,
  						 NULL, 1);
  			unlock_page(page);
  		}
  		page_cache_release(page);
  		start += PAGE_CACHE_SIZE;
  	}
ffbd517d5   Chris Mason   Btrfs: use WRITE_...
3083
  	flush_epd_write_bio(&epd);
771ed689d   Chris Mason   Btrfs: Optimize c...
3084
3085
  	return ret;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
3086
3087
3088
3089
3090
3091
3092
3093
3094
3095
3096
  
  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,
771ed689d   Chris Mason   Btrfs: Optimize c...
3097
  		.extent_locked = 0,
ffbd517d5   Chris Mason   Btrfs: use WRITE_...
3098
  		.sync_io = wbc->sync_mode == WB_SYNC_ALL,
d1310b2e0   Chris Mason   Btrfs: Split the ...
3099
  	};
4bef08485   Chris Mason   Btrfs: Tree loggi...
3100
  	ret = extent_write_cache_pages(tree, mapping, wbc,
d2c3f4f69   Chris Mason   Btrfs: Avoid writ...
3101
3102
  				       __extent_writepage, &epd,
  				       flush_write_bio);
ffbd517d5   Chris Mason   Btrfs: use WRITE_...
3103
  	flush_epd_write_bio(&epd);
d1310b2e0   Chris Mason   Btrfs: Split the ...
3104
3105
  	return ret;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
3106
3107
3108
3109
3110
3111
3112
3113
  
  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;
c8b978188   Chris Mason   Btrfs: Add zlib c...
3114
  	unsigned long bio_flags = 0;
d1310b2e0   Chris Mason   Btrfs: Split the ...
3115

d1310b2e0   Chris Mason   Btrfs: Split the ...
3116
3117
3118
3119
3120
  	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);
28ecb6090   Nick Piggin   Btrfs: use add_to...
3121
  		if (!add_to_page_cache_lru(page, mapping,
43e817a1f   Itaru Kitayama   btrfs: fix wrong ...
3122
  					page->index, GFP_NOFS)) {
f188591e9   Chris Mason   Btrfs: Retry meta...
3123
  			__extent_read_full_page(tree, page, get_extent,
c8b978188   Chris Mason   Btrfs: Add zlib c...
3124
  						&bio, 0, &bio_flags);
d1310b2e0   Chris Mason   Btrfs: Split the ...
3125
3126
3127
  		}
  		page_cache_release(page);
  	}
d1310b2e0   Chris Mason   Btrfs: Split the ...
3128
3129
  	BUG_ON(!list_empty(pages));
  	if (bio)
c8b978188   Chris Mason   Btrfs: Add zlib c...
3130
  		submit_one_bio(READ, bio, 0, bio_flags);
d1310b2e0   Chris Mason   Btrfs: Split the ...
3131
3132
  	return 0;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
3133
3134
3135
3136
3137
3138
3139
3140
3141
  
  /*
   * 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)
  {
2ac55d41b   Josef Bacik   Btrfs: cache the ...
3142
  	struct extent_state *cached_state = NULL;
d1310b2e0   Chris Mason   Btrfs: Split the ...
3143
3144
3145
  	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;
d397712bc   Chris Mason   Btrfs: Fix checkp...
3146
  	start += (offset + blocksize - 1) & ~(blocksize - 1);
d1310b2e0   Chris Mason   Btrfs: Split the ...
3147
3148
  	if (start > end)
  		return 0;
2ac55d41b   Josef Bacik   Btrfs: cache the ...
3149
  	lock_extent_bits(tree, start, end, 0, &cached_state, GFP_NOFS);
1edbb734b   Chris Mason   Btrfs: reduce CPU...
3150
  	wait_on_page_writeback(page);
d1310b2e0   Chris Mason   Btrfs: Split the ...
3151
  	clear_extent_bit(tree, start, end,
32c00aff7   Josef Bacik   Btrfs: release de...
3152
3153
  			 EXTENT_LOCKED | EXTENT_DIRTY | EXTENT_DELALLOC |
  			 EXTENT_DO_ACCOUNTING,
2ac55d41b   Josef Bacik   Btrfs: cache the ...
3154
  			 1, 1, &cached_state, GFP_NOFS);
d1310b2e0   Chris Mason   Btrfs: Split the ...
3155
3156
  	return 0;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
3157
3158
  
  /*
7b13b7b11   Chris Mason   Btrfs: Don't drop...
3159
3160
3161
3162
3163
3164
3165
3166
3167
3168
3169
   * 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;
211f90e68   Chris Mason   Btrfs: Don't allo...
3170
  	if (test_range_bit(tree, start, end,
8b62b72b2   Chris Mason   Btrfs: Use PagePr...
3171
  			   EXTENT_IOBITS, 0, NULL))
7b13b7b11   Chris Mason   Btrfs: Don't drop...
3172
3173
3174
3175
  		ret = 0;
  	else {
  		if ((mask & GFP_NOFS) == GFP_NOFS)
  			mask = GFP_NOFS;
11ef160fd   Chris Mason   Btrfs: fix releas...
3176
3177
3178
3179
  		/*
  		 * at this point we can safely clear everything except the
  		 * locked bit and the nodatasum bit
  		 */
e3f24cc52   Chris Mason   Btrfs: don't rele...
3180
  		ret = clear_extent_bit(tree, start, end,
11ef160fd   Chris Mason   Btrfs: fix releas...
3181
3182
  				 ~(EXTENT_LOCKED | EXTENT_NODATASUM),
  				 0, 0, NULL, mask);
e3f24cc52   Chris Mason   Btrfs: don't rele...
3183
3184
3185
3186
3187
3188
3189
3190
  
  		/* if clear_extent_bit failed for enomem reasons,
  		 * we can't allow the release to continue.
  		 */
  		if (ret < 0)
  			ret = 0;
  		else
  			ret = 1;
7b13b7b11   Chris Mason   Btrfs: Don't drop...
3191
3192
3193
  	}
  	return ret;
  }
7b13b7b11   Chris Mason   Btrfs: Don't drop...
3194
3195
  
  /*
d1310b2e0   Chris Mason   Btrfs: Split the ...
3196
3197
3198
3199
3200
   * 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,
70dec8079   Chris Mason   Btrfs: extent_io ...
3201
3202
  			       struct extent_io_tree *tree, struct page *page,
  			       gfp_t mask)
d1310b2e0   Chris Mason   Btrfs: Split the ...
3203
3204
3205
3206
  {
  	struct extent_map *em;
  	u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
  	u64 end = start + PAGE_CACHE_SIZE - 1;
7b13b7b11   Chris Mason   Btrfs: Don't drop...
3207

70dec8079   Chris Mason   Btrfs: extent_io ...
3208
3209
  	if ((mask & __GFP_WAIT) &&
  	    page->mapping->host->i_size > 16 * 1024 * 1024) {
39b5637f6   Yan   Btrfs: Fix "no cs...
3210
  		u64 len;
70dec8079   Chris Mason   Btrfs: extent_io ...
3211
  		while (start <= end) {
39b5637f6   Yan   Btrfs: Fix "no cs...
3212
  			len = end - start + 1;
890871be8   Chris Mason   Btrfs: switch ext...
3213
  			write_lock(&map->lock);
39b5637f6   Yan   Btrfs: Fix "no cs...
3214
  			em = lookup_extent_mapping(map, start, len);
c704005d8   David Sterba   btrfs: unify chec...
3215
  			if (IS_ERR_OR_NULL(em)) {
890871be8   Chris Mason   Btrfs: switch ext...
3216
  				write_unlock(&map->lock);
70dec8079   Chris Mason   Btrfs: extent_io ...
3217
3218
  				break;
  			}
7f3c74fb8   Chris Mason   Btrfs: Keep exten...
3219
3220
  			if (test_bit(EXTENT_FLAG_PINNED, &em->flags) ||
  			    em->start != start) {
890871be8   Chris Mason   Btrfs: switch ext...
3221
  				write_unlock(&map->lock);
70dec8079   Chris Mason   Btrfs: extent_io ...
3222
3223
3224
3225
3226
  				free_extent_map(em);
  				break;
  			}
  			if (!test_range_bit(tree, em->start,
  					    extent_map_end(em) - 1,
8b62b72b2   Chris Mason   Btrfs: Use PagePr...
3227
  					    EXTENT_LOCKED | EXTENT_WRITEBACK,
9655d2982   Chris Mason   Btrfs: use a cach...
3228
  					    0, NULL)) {
70dec8079   Chris Mason   Btrfs: extent_io ...
3229
3230
3231
3232
3233
  				remove_extent_mapping(map, em);
  				/* once for the rb tree */
  				free_extent_map(em);
  			}
  			start = extent_map_end(em);
890871be8   Chris Mason   Btrfs: switch ext...
3234
  			write_unlock(&map->lock);
70dec8079   Chris Mason   Btrfs: extent_io ...
3235
3236
  
  			/* once for us */
d1310b2e0   Chris Mason   Btrfs: Split the ...
3237
3238
  			free_extent_map(em);
  		}
d1310b2e0   Chris Mason   Btrfs: Split the ...
3239
  	}
7b13b7b11   Chris Mason   Btrfs: Don't drop...
3240
  	return try_release_extent_state(map, tree, page, mask);
d1310b2e0   Chris Mason   Btrfs: Split the ...
3241
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
3242

ec29ed5b4   Chris Mason   Btrfs: fix fiemap...
3243
3244
3245
3246
3247
3248
3249
3250
3251
3252
3253
3254
3255
3256
3257
3258
3259
3260
3261
3262
3263
3264
  /*
   * helper function for fiemap, which doesn't want to see any holes.
   * This maps until we find something past 'last'
   */
  static struct extent_map *get_extent_skip_holes(struct inode *inode,
  						u64 offset,
  						u64 last,
  						get_extent_t *get_extent)
  {
  	u64 sectorsize = BTRFS_I(inode)->root->sectorsize;
  	struct extent_map *em;
  	u64 len;
  
  	if (offset >= last)
  		return NULL;
  
  	while(1) {
  		len = last - offset;
  		if (len == 0)
  			break;
  		len = (len + sectorsize - 1) & ~(sectorsize - 1);
  		em = get_extent(inode, NULL, 0, offset, len, 0);
c704005d8   David Sterba   btrfs: unify chec...
3265
  		if (IS_ERR_OR_NULL(em))
ec29ed5b4   Chris Mason   Btrfs: fix fiemap...
3266
3267
3268
3269
3270
3271
3272
3273
3274
3275
3276
3277
3278
3279
3280
3281
  			return em;
  
  		/* if this isn't a hole return it */
  		if (!test_bit(EXTENT_FLAG_VACANCY, &em->flags) &&
  		    em->block_start != EXTENT_MAP_HOLE) {
  			return em;
  		}
  
  		/* this is a hole, advance to the next extent */
  		offset = extent_map_end(em);
  		free_extent_map(em);
  		if (offset >= last)
  			break;
  	}
  	return NULL;
  }
1506fcc81   Yehuda Sadeh   Btrfs: fiemap sup...
3282
3283
3284
  int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
  		__u64 start, __u64 len, get_extent_t *get_extent)
  {
975f84fee   Josef Bacik   Btrfs: fix fiemap
3285
  	int ret = 0;
1506fcc81   Yehuda Sadeh   Btrfs: fiemap sup...
3286
3287
3288
  	u64 off = start;
  	u64 max = start + len;
  	u32 flags = 0;
975f84fee   Josef Bacik   Btrfs: fix fiemap
3289
3290
  	u32 found_type;
  	u64 last;
ec29ed5b4   Chris Mason   Btrfs: fix fiemap...
3291
  	u64 last_for_get_extent = 0;
1506fcc81   Yehuda Sadeh   Btrfs: fiemap sup...
3292
  	u64 disko = 0;
ec29ed5b4   Chris Mason   Btrfs: fix fiemap...
3293
  	u64 isize = i_size_read(inode);
975f84fee   Josef Bacik   Btrfs: fix fiemap
3294
  	struct btrfs_key found_key;
1506fcc81   Yehuda Sadeh   Btrfs: fiemap sup...
3295
  	struct extent_map *em = NULL;
2ac55d41b   Josef Bacik   Btrfs: cache the ...
3296
  	struct extent_state *cached_state = NULL;
975f84fee   Josef Bacik   Btrfs: fix fiemap
3297
3298
  	struct btrfs_path *path;
  	struct btrfs_file_extent_item *item;
1506fcc81   Yehuda Sadeh   Btrfs: fiemap sup...
3299
  	int end = 0;
ec29ed5b4   Chris Mason   Btrfs: fix fiemap...
3300
3301
3302
  	u64 em_start = 0;
  	u64 em_len = 0;
  	u64 em_end = 0;
1506fcc81   Yehuda Sadeh   Btrfs: fiemap sup...
3303
  	unsigned long emflags;
1506fcc81   Yehuda Sadeh   Btrfs: fiemap sup...
3304
3305
3306
  
  	if (len == 0)
  		return -EINVAL;
975f84fee   Josef Bacik   Btrfs: fix fiemap
3307
3308
3309
3310
  	path = btrfs_alloc_path();
  	if (!path)
  		return -ENOMEM;
  	path->leave_spinning = 1;
4d479cf01   Josef Bacik   Btrfs: sectorsize...
3311
3312
  	start = ALIGN(start, BTRFS_I(inode)->root->sectorsize);
  	len = ALIGN(len, BTRFS_I(inode)->root->sectorsize);
ec29ed5b4   Chris Mason   Btrfs: fix fiemap...
3313
3314
3315
3316
  	/*
  	 * lookup the last file extent.  We're not using i_size here
  	 * because there might be preallocation past i_size
  	 */
975f84fee   Josef Bacik   Btrfs: fix fiemap
3317
  	ret = btrfs_lookup_file_extent(NULL, BTRFS_I(inode)->root,
33345d015   Li Zefan   Btrfs: Always use...
3318
  				       path, btrfs_ino(inode), -1, 0);
975f84fee   Josef Bacik   Btrfs: fix fiemap
3319
3320
3321
3322
3323
3324
3325
3326
3327
3328
  	if (ret < 0) {
  		btrfs_free_path(path);
  		return ret;
  	}
  	WARN_ON(!ret);
  	path->slots[0]--;
  	item = btrfs_item_ptr(path->nodes[0], path->slots[0],
  			      struct btrfs_file_extent_item);
  	btrfs_item_key_to_cpu(path->nodes[0], &found_key, path->slots[0]);
  	found_type = btrfs_key_type(&found_key);
ec29ed5b4   Chris Mason   Btrfs: fix fiemap...
3329
  	/* No extents, but there might be delalloc bits */
33345d015   Li Zefan   Btrfs: Always use...
3330
  	if (found_key.objectid != btrfs_ino(inode) ||
975f84fee   Josef Bacik   Btrfs: fix fiemap
3331
  	    found_type != BTRFS_EXTENT_DATA_KEY) {
ec29ed5b4   Chris Mason   Btrfs: fix fiemap...
3332
3333
3334
3335
3336
3337
3338
3339
3340
3341
3342
  		/* have to trust i_size as the end */
  		last = (u64)-1;
  		last_for_get_extent = isize;
  	} else {
  		/*
  		 * remember the start of the last extent.  There are a
  		 * bunch of different factors that go into the length of the
  		 * extent, so its much less complex to remember where it started
  		 */
  		last = found_key.offset;
  		last_for_get_extent = last + 1;
975f84fee   Josef Bacik   Btrfs: fix fiemap
3343
  	}
975f84fee   Josef Bacik   Btrfs: fix fiemap
3344
  	btrfs_free_path(path);
ec29ed5b4   Chris Mason   Btrfs: fix fiemap...
3345
3346
3347
3348
3349
3350
3351
3352
3353
  	/*
  	 * we might have some extents allocated but more delalloc past those
  	 * extents.  so, we trust isize unless the start of the last extent is
  	 * beyond isize
  	 */
  	if (last < isize) {
  		last = (u64)-1;
  		last_for_get_extent = isize;
  	}
2ac55d41b   Josef Bacik   Btrfs: cache the ...
3354
3355
  	lock_extent_bits(&BTRFS_I(inode)->io_tree, start, start + len, 0,
  			 &cached_state, GFP_NOFS);
ec29ed5b4   Chris Mason   Btrfs: fix fiemap...
3356

4d479cf01   Josef Bacik   Btrfs: sectorsize...
3357
  	em = get_extent_skip_holes(inode, start, last_for_get_extent,
ec29ed5b4   Chris Mason   Btrfs: fix fiemap...
3358
  				   get_extent);
1506fcc81   Yehuda Sadeh   Btrfs: fiemap sup...
3359
3360
3361
3362
3363
3364
  	if (!em)
  		goto out;
  	if (IS_ERR(em)) {
  		ret = PTR_ERR(em);
  		goto out;
  	}
975f84fee   Josef Bacik   Btrfs: fix fiemap
3365

1506fcc81   Yehuda Sadeh   Btrfs: fiemap sup...
3366
  	while (!end) {
ea8efc74b   Chris Mason   Btrfs: make sure ...
3367
3368
3369
3370
3371
3372
3373
3374
3375
3376
3377
3378
3379
  		u64 offset_in_extent;
  
  		/* break if the extent we found is outside the range */
  		if (em->start >= max || extent_map_end(em) < off)
  			break;
  
  		/*
  		 * get_extent may return an extent that starts before our
  		 * requested range.  We have to make sure the ranges
  		 * we return to fiemap always move forward and don't
  		 * overlap, so adjust the offsets here
  		 */
  		em_start = max(em->start, off);
1506fcc81   Yehuda Sadeh   Btrfs: fiemap sup...
3380

ea8efc74b   Chris Mason   Btrfs: make sure ...
3381
3382
3383
3384
3385
  		/*
  		 * record the offset from the start of the extent
  		 * for adjusting the disk offset below
  		 */
  		offset_in_extent = em_start - em->start;
ec29ed5b4   Chris Mason   Btrfs: fix fiemap...
3386
  		em_end = extent_map_end(em);
ea8efc74b   Chris Mason   Btrfs: make sure ...
3387
  		em_len = em_end - em_start;
ec29ed5b4   Chris Mason   Btrfs: fix fiemap...
3388
  		emflags = em->flags;
1506fcc81   Yehuda Sadeh   Btrfs: fiemap sup...
3389
3390
  		disko = 0;
  		flags = 0;
ea8efc74b   Chris Mason   Btrfs: make sure ...
3391
3392
3393
3394
3395
3396
  		/*
  		 * bump off for our next call to get_extent
  		 */
  		off = extent_map_end(em);
  		if (off >= max)
  			end = 1;
93dbfad7a   Heiko Carstens   Btrfs: fix __ucmp...
3397
  		if (em->block_start == EXTENT_MAP_LAST_BYTE) {
1506fcc81   Yehuda Sadeh   Btrfs: fiemap sup...
3398
3399
  			end = 1;
  			flags |= FIEMAP_EXTENT_LAST;
93dbfad7a   Heiko Carstens   Btrfs: fix __ucmp...
3400
  		} else if (em->block_start == EXTENT_MAP_INLINE) {
1506fcc81   Yehuda Sadeh   Btrfs: fiemap sup...
3401
3402
  			flags |= (FIEMAP_EXTENT_DATA_INLINE |
  				  FIEMAP_EXTENT_NOT_ALIGNED);
93dbfad7a   Heiko Carstens   Btrfs: fix __ucmp...
3403
  		} else if (em->block_start == EXTENT_MAP_DELALLOC) {
1506fcc81   Yehuda Sadeh   Btrfs: fiemap sup...
3404
3405
  			flags |= (FIEMAP_EXTENT_DELALLOC |
  				  FIEMAP_EXTENT_UNKNOWN);
93dbfad7a   Heiko Carstens   Btrfs: fix __ucmp...
3406
  		} else {
ea8efc74b   Chris Mason   Btrfs: make sure ...
3407
  			disko = em->block_start + offset_in_extent;
1506fcc81   Yehuda Sadeh   Btrfs: fiemap sup...
3408
3409
3410
  		}
  		if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags))
  			flags |= FIEMAP_EXTENT_ENCODED;
1506fcc81   Yehuda Sadeh   Btrfs: fiemap sup...
3411
3412
  		free_extent_map(em);
  		em = NULL;
ec29ed5b4   Chris Mason   Btrfs: fix fiemap...
3413
3414
  		if ((em_start >= last) || em_len == (u64)-1 ||
  		   (last == (u64)-1 && isize <= em_end)) {
1506fcc81   Yehuda Sadeh   Btrfs: fiemap sup...
3415
3416
3417
  			flags |= FIEMAP_EXTENT_LAST;
  			end = 1;
  		}
ec29ed5b4   Chris Mason   Btrfs: fix fiemap...
3418
3419
3420
3421
3422
3423
3424
3425
  		/* now scan forward to see if this is really the last extent. */
  		em = get_extent_skip_holes(inode, off, last_for_get_extent,
  					   get_extent);
  		if (IS_ERR(em)) {
  			ret = PTR_ERR(em);
  			goto out;
  		}
  		if (!em) {
975f84fee   Josef Bacik   Btrfs: fix fiemap
3426
3427
3428
  			flags |= FIEMAP_EXTENT_LAST;
  			end = 1;
  		}
ec29ed5b4   Chris Mason   Btrfs: fix fiemap...
3429
3430
3431
3432
  		ret = fiemap_fill_next_extent(fieinfo, em_start, disko,
  					      em_len, flags);
  		if (ret)
  			goto out_free;
1506fcc81   Yehuda Sadeh   Btrfs: fiemap sup...
3433
3434
3435
3436
  	}
  out_free:
  	free_extent_map(em);
  out:
2ac55d41b   Josef Bacik   Btrfs: cache the ...
3437
3438
  	unlock_extent_cached(&BTRFS_I(inode)->io_tree, start, start + len,
  			     &cached_state, GFP_NOFS);
1506fcc81   Yehuda Sadeh   Btrfs: fiemap sup...
3439
3440
  	return ret;
  }
4a54c8c16   Jan Schmidt   btrfs: Moved repa...
3441
  inline struct page *extent_buffer_page(struct extent_buffer *eb,
d1310b2e0   Chris Mason   Btrfs: Split the ...
3442
3443
3444
3445
3446
3447
3448
3449
3450
  					      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;
33958dc6d   Chris Mason   Btrfs: Fix verify...
3451
3452
  	if (!mapping)
  		return NULL;
0ee0fda06   Sven Wegener   Btrfs: Add compat...
3453
3454
3455
3456
3457
3458
  
  	/*
  	 * 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.
  	 */
0ee0fda06   Sven Wegener   Btrfs: Add compat...
3459
  	rcu_read_lock();
d1310b2e0   Chris Mason   Btrfs: Split the ...
3460
  	p = radix_tree_lookup(&mapping->page_tree, i);
0ee0fda06   Sven Wegener   Btrfs: Add compat...
3461
  	rcu_read_unlock();
2b1f55b0f   Chris Mason   Remove Btrfs comp...
3462

d1310b2e0   Chris Mason   Btrfs: Split the ...
3463
3464
  	return p;
  }
4a54c8c16   Jan Schmidt   btrfs: Moved repa...
3465
  inline unsigned long num_extent_pages(u64 start, u64 len)
728131d8e   Chris Mason   Btrfs: Add additi...
3466
  {
6af118ce5   Chris Mason   Btrfs: Index exte...
3467
3468
  	return ((start + len + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT) -
  		(start >> PAGE_CACHE_SHIFT);
728131d8e   Chris Mason   Btrfs: Add additi...
3469
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
3470
3471
3472
3473
3474
3475
  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;
3935127c5   Chris Mason   Btrfs: disable le...
3476
  #if LEAK_DEBUG
2d2ae5479   Chris Mason   Btrfs: Add leak d...
3477
  	unsigned long flags;
4bef08485   Chris Mason   Btrfs: Tree loggi...
3478
  #endif
d1310b2e0   Chris Mason   Btrfs: Split the ...
3479

d1310b2e0   Chris Mason   Btrfs: Split the ...
3480
  	eb = kmem_cache_zalloc(extent_buffer_cache, mask);
91ca338d7   Tsutomu Itoh   btrfs: check NULL...
3481
3482
  	if (eb == NULL)
  		return NULL;
d1310b2e0   Chris Mason   Btrfs: Split the ...
3483
3484
  	eb->start = start;
  	eb->len = len;
bd681513f   Chris Mason   Btrfs: switch the...
3485
3486
3487
3488
3489
3490
3491
3492
3493
  	rwlock_init(&eb->lock);
  	atomic_set(&eb->write_locks, 0);
  	atomic_set(&eb->read_locks, 0);
  	atomic_set(&eb->blocking_readers, 0);
  	atomic_set(&eb->blocking_writers, 0);
  	atomic_set(&eb->spinning_readers, 0);
  	atomic_set(&eb->spinning_writers, 0);
  	init_waitqueue_head(&eb->write_lock_wq);
  	init_waitqueue_head(&eb->read_lock_wq);
b4ce94de9   Chris Mason   Btrfs: Change btr...
3494

3935127c5   Chris Mason   Btrfs: disable le...
3495
  #if LEAK_DEBUG
2d2ae5479   Chris Mason   Btrfs: Add leak d...
3496
3497
3498
  	spin_lock_irqsave(&leak_lock, flags);
  	list_add(&eb->leak_list, &buffers);
  	spin_unlock_irqrestore(&leak_lock, flags);
4bef08485   Chris Mason   Btrfs: Tree loggi...
3499
  #endif
d1310b2e0   Chris Mason   Btrfs: Split the ...
3500
3501
3502
3503
3504
3505
3506
  	atomic_set(&eb->refs, 1);
  
  	return eb;
  }
  
  static void __free_extent_buffer(struct extent_buffer *eb)
  {
3935127c5   Chris Mason   Btrfs: disable le...
3507
  #if LEAK_DEBUG
2d2ae5479   Chris Mason   Btrfs: Add leak d...
3508
3509
3510
3511
  	unsigned long flags;
  	spin_lock_irqsave(&leak_lock, flags);
  	list_del(&eb->leak_list);
  	spin_unlock_irqrestore(&leak_lock, flags);
4bef08485   Chris Mason   Btrfs: Tree loggi...
3512
  #endif
d1310b2e0   Chris Mason   Btrfs: Split the ...
3513
3514
  	kmem_cache_free(extent_buffer_cache, eb);
  }
897ca6e9b   Miao Xie   Btrfs: restructur...
3515
3516
3517
3518
3519
3520
3521
3522
3523
3524
3525
3526
3527
3528
3529
3530
3531
3532
3533
3534
3535
3536
3537
3538
3539
3540
3541
3542
3543
3544
3545
3546
  /*
   * Helper for releasing extent buffer page.
   */
  static void btrfs_release_extent_buffer_page(struct extent_buffer *eb,
  						unsigned long start_idx)
  {
  	unsigned long index;
  	struct page *page;
  
  	if (!eb->first_page)
  		return;
  
  	index = num_extent_pages(eb->start, eb->len);
  	if (start_idx >= index)
  		return;
  
  	do {
  		index--;
  		page = extent_buffer_page(eb, index);
  		if (page)
  			page_cache_release(page);
  	} while (index != start_idx);
  }
  
  /*
   * Helper for releasing the extent buffer.
   */
  static inline void btrfs_release_extent_buffer(struct extent_buffer *eb)
  {
  	btrfs_release_extent_buffer_page(eb, 0);
  	__free_extent_buffer(eb);
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
3547
3548
  struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree,
  					  u64 start, unsigned long len,
ba1441926   David Sterba   btrfs: drop gfp p...
3549
  					  struct page *page0)
d1310b2e0   Chris Mason   Btrfs: Split the ...
3550
3551
3552
3553
3554
  {
  	unsigned long num_pages = num_extent_pages(start, len);
  	unsigned long i;
  	unsigned long index = start >> PAGE_CACHE_SHIFT;
  	struct extent_buffer *eb;
6af118ce5   Chris Mason   Btrfs: Index exte...
3555
  	struct extent_buffer *exists = NULL;
d1310b2e0   Chris Mason   Btrfs: Split the ...
3556
3557
3558
  	struct page *p;
  	struct address_space *mapping = tree->mapping;
  	int uptodate = 1;
19fe0a8b7   Miao Xie   Btrfs: Switch the...
3559
  	int ret;
d1310b2e0   Chris Mason   Btrfs: Split the ...
3560

19fe0a8b7   Miao Xie   Btrfs: Switch the...
3561
3562
3563
3564
  	rcu_read_lock();
  	eb = radix_tree_lookup(&tree->buffer, start >> PAGE_CACHE_SHIFT);
  	if (eb && atomic_inc_not_zero(&eb->refs)) {
  		rcu_read_unlock();
0f9dd46cd   Josef Bacik   Btrfs: free space...
3565
  		mark_page_accessed(eb->first_page);
6af118ce5   Chris Mason   Btrfs: Index exte...
3566
3567
  		return eb;
  	}
19fe0a8b7   Miao Xie   Btrfs: Switch the...
3568
  	rcu_read_unlock();
6af118ce5   Chris Mason   Btrfs: Index exte...
3569

ba1441926   David Sterba   btrfs: drop gfp p...
3570
  	eb = __alloc_extent_buffer(tree, start, len, GFP_NOFS);
2b114d1d3   Peter   Btrfs: Correct us...
3571
  	if (!eb)
d1310b2e0   Chris Mason   Btrfs: Split the ...
3572
  		return NULL;
d1310b2e0   Chris Mason   Btrfs: Split the ...
3573
3574
3575
3576
3577
3578
3579
  	if (page0) {
  		eb->first_page = page0;
  		i = 1;
  		index++;
  		page_cache_get(page0);
  		mark_page_accessed(page0);
  		set_page_extent_mapped(page0);
d1310b2e0   Chris Mason   Btrfs: Split the ...
3580
  		set_page_extent_head(page0, len);
f188591e9   Chris Mason   Btrfs: Retry meta...
3581
  		uptodate = PageUptodate(page0);
d1310b2e0   Chris Mason   Btrfs: Split the ...
3582
3583
3584
3585
  	} else {
  		i = 0;
  	}
  	for (; i < num_pages; i++, index++) {
a65917156   Chris Mason   Btrfs: stop using...
3586
  		p = find_or_create_page(mapping, index, GFP_NOFS);
d1310b2e0   Chris Mason   Btrfs: Split the ...
3587
3588
  		if (!p) {
  			WARN_ON(1);
6af118ce5   Chris Mason   Btrfs: Index exte...
3589
  			goto free_eb;
d1310b2e0   Chris Mason   Btrfs: Split the ...
3590
3591
3592
3593
3594
3595
3596
3597
3598
3599
3600
  		}
  		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;
eb14ab8ed   Chris Mason   Btrfs: fix page->...
3601
3602
3603
3604
3605
3606
3607
  
  		/*
  		 * see below about how we avoid a nasty race with release page
  		 * and why we unlock later
  		 */
  		if (i != 0)
  			unlock_page(p);
d1310b2e0   Chris Mason   Btrfs: Split the ...
3608
3609
  	}
  	if (uptodate)
b4ce94de9   Chris Mason   Btrfs: Change btr...
3610
  		set_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
d1310b2e0   Chris Mason   Btrfs: Split the ...
3611

19fe0a8b7   Miao Xie   Btrfs: Switch the...
3612
3613
3614
  	ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM);
  	if (ret)
  		goto free_eb;
6af118ce5   Chris Mason   Btrfs: Index exte...
3615
  	spin_lock(&tree->buffer_lock);
19fe0a8b7   Miao Xie   Btrfs: Switch the...
3616
3617
3618
3619
  	ret = radix_tree_insert(&tree->buffer, start >> PAGE_CACHE_SHIFT, eb);
  	if (ret == -EEXIST) {
  		exists = radix_tree_lookup(&tree->buffer,
  						start >> PAGE_CACHE_SHIFT);
6af118ce5   Chris Mason   Btrfs: Index exte...
3620
3621
3622
  		/* add one reference for the caller */
  		atomic_inc(&exists->refs);
  		spin_unlock(&tree->buffer_lock);
19fe0a8b7   Miao Xie   Btrfs: Switch the...
3623
  		radix_tree_preload_end();
6af118ce5   Chris Mason   Btrfs: Index exte...
3624
3625
  		goto free_eb;
  	}
6af118ce5   Chris Mason   Btrfs: Index exte...
3626
3627
  	/* add one reference for the tree */
  	atomic_inc(&eb->refs);
f044ba783   Yan, Zheng   Btrfs: fix race b...
3628
  	spin_unlock(&tree->buffer_lock);
19fe0a8b7   Miao Xie   Btrfs: Switch the...
3629
  	radix_tree_preload_end();
eb14ab8ed   Chris Mason   Btrfs: fix page->...
3630
3631
3632
3633
3634
3635
3636
3637
3638
3639
3640
3641
3642
3643
  
  	/*
  	 * there is a race where release page may have
  	 * tried to find this extent buffer in the radix
  	 * but failed.  It will tell the VM it is safe to
  	 * reclaim the, and it will clear the page private bit.
  	 * We must make sure to set the page private bit properly
  	 * after the extent buffer is in the radix tree so
  	 * it doesn't get lost
  	 */
  	set_page_extent_mapped(eb->first_page);
  	set_page_extent_head(eb->first_page, eb->len);
  	if (!page0)
  		unlock_page(eb->first_page);
d1310b2e0   Chris Mason   Btrfs: Split the ...
3644
  	return eb;
6af118ce5   Chris Mason   Btrfs: Index exte...
3645
  free_eb:
eb14ab8ed   Chris Mason   Btrfs: fix page->...
3646
3647
  	if (eb->first_page && !page0)
  		unlock_page(eb->first_page);
d1310b2e0   Chris Mason   Btrfs: Split the ...
3648
  	if (!atomic_dec_and_test(&eb->refs))
6af118ce5   Chris Mason   Btrfs: Index exte...
3649
  		return exists;
897ca6e9b   Miao Xie   Btrfs: restructur...
3650
  	btrfs_release_extent_buffer(eb);
6af118ce5   Chris Mason   Btrfs: Index exte...
3651
  	return exists;
d1310b2e0   Chris Mason   Btrfs: Split the ...
3652
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
3653
3654
  
  struct extent_buffer *find_extent_buffer(struct extent_io_tree *tree,
f09d1f60e   David Sterba   btrfs: drop gfp p...
3655
  					 u64 start, unsigned long len)
d1310b2e0   Chris Mason   Btrfs: Split the ...
3656
  {
d1310b2e0   Chris Mason   Btrfs: Split the ...
3657
  	struct extent_buffer *eb;
d1310b2e0   Chris Mason   Btrfs: Split the ...
3658

19fe0a8b7   Miao Xie   Btrfs: Switch the...
3659
3660
3661
3662
  	rcu_read_lock();
  	eb = radix_tree_lookup(&tree->buffer, start >> PAGE_CACHE_SHIFT);
  	if (eb && atomic_inc_not_zero(&eb->refs)) {
  		rcu_read_unlock();
0f9dd46cd   Josef Bacik   Btrfs: free space...
3663
  		mark_page_accessed(eb->first_page);
19fe0a8b7   Miao Xie   Btrfs: Switch the...
3664
3665
3666
  		return eb;
  	}
  	rcu_read_unlock();
0f9dd46cd   Josef Bacik   Btrfs: free space...
3667

19fe0a8b7   Miao Xie   Btrfs: Switch the...
3668
  	return NULL;
d1310b2e0   Chris Mason   Btrfs: Split the ...
3669
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
3670
3671
3672
  
  void free_extent_buffer(struct extent_buffer *eb)
  {
d1310b2e0   Chris Mason   Btrfs: Split the ...
3673
3674
3675
3676
3677
  	if (!eb)
  		return;
  
  	if (!atomic_dec_and_test(&eb->refs))
  		return;
6af118ce5   Chris Mason   Btrfs: Index exte...
3678
  	WARN_ON(1);
d1310b2e0   Chris Mason   Btrfs: Split the ...
3679
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
3680
3681
3682
3683
  
  int clear_extent_buffer_dirty(struct extent_io_tree *tree,
  			      struct extent_buffer *eb)
  {
d1310b2e0   Chris Mason   Btrfs: Split the ...
3684
3685
3686
  	unsigned long i;
  	unsigned long num_pages;
  	struct page *page;
d1310b2e0   Chris Mason   Btrfs: Split the ...
3687
3688
3689
3690
  	num_pages = num_extent_pages(eb->start, eb->len);
  
  	for (i = 0; i < num_pages; i++) {
  		page = extent_buffer_page(eb, i);
b9473439d   Chris Mason   Btrfs: leave btre...
3691
  		if (!PageDirty(page))
d2c3f4f69   Chris Mason   Btrfs: Avoid writ...
3692
  			continue;
a61e6f29d   Chris Mason   Btrfs: Use a mute...
3693
  		lock_page(page);
eb14ab8ed   Chris Mason   Btrfs: fix page->...
3694
3695
3696
  		WARN_ON(!PagePrivate(page));
  
  		set_page_extent_mapped(page);
d1310b2e0   Chris Mason   Btrfs: Split the ...
3697
3698
  		if (i == 0)
  			set_page_extent_head(page, eb->len);
d1310b2e0   Chris Mason   Btrfs: Split the ...
3699

d1310b2e0   Chris Mason   Btrfs: Split the ...
3700
  		clear_page_dirty_for_io(page);
0ee0fda06   Sven Wegener   Btrfs: Add compat...
3701
  		spin_lock_irq(&page->mapping->tree_lock);
d1310b2e0   Chris Mason   Btrfs: Split the ...
3702
3703
3704
3705
3706
  		if (!PageDirty(page)) {
  			radix_tree_tag_clear(&page->mapping->page_tree,
  						page_index(page),
  						PAGECACHE_TAG_DIRTY);
  		}
0ee0fda06   Sven Wegener   Btrfs: Add compat...
3707
  		spin_unlock_irq(&page->mapping->tree_lock);
bf0da8c18   Chris Mason   Btrfs: ClearPageE...
3708
  		ClearPageError(page);
a61e6f29d   Chris Mason   Btrfs: Use a mute...
3709
  		unlock_page(page);
d1310b2e0   Chris Mason   Btrfs: Split the ...
3710
3711
3712
  	}
  	return 0;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
3713

d1310b2e0   Chris Mason   Btrfs: Split the ...
3714
3715
3716
3717
3718
  int set_extent_buffer_dirty(struct extent_io_tree *tree,
  			     struct extent_buffer *eb)
  {
  	unsigned long i;
  	unsigned long num_pages;
b9473439d   Chris Mason   Btrfs: leave btre...
3719
  	int was_dirty = 0;
d1310b2e0   Chris Mason   Btrfs: Split the ...
3720

b9473439d   Chris Mason   Btrfs: leave btre...
3721
  	was_dirty = test_and_set_bit(EXTENT_BUFFER_DIRTY, &eb->bflags);
d1310b2e0   Chris Mason   Btrfs: Split the ...
3722
  	num_pages = num_extent_pages(eb->start, eb->len);
b9473439d   Chris Mason   Btrfs: leave btre...
3723
  	for (i = 0; i < num_pages; i++)
d1310b2e0   Chris Mason   Btrfs: Split the ...
3724
  		__set_page_dirty_nobuffers(extent_buffer_page(eb, i));
b9473439d   Chris Mason   Btrfs: leave btre...
3725
  	return was_dirty;
d1310b2e0   Chris Mason   Btrfs: Split the ...
3726
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
3727

19b6caf4a   Chris Mason   Btrfs: reduce ext...
3728
3729
3730
3731
3732
3733
3734
3735
3736
3737
3738
3739
3740
3741
3742
  static int __eb_straddles_pages(u64 start, u64 len)
  {
  	if (len < PAGE_CACHE_SIZE)
  		return 1;
  	if (start & (PAGE_CACHE_SIZE - 1))
  		return 1;
  	if ((start + len) & (PAGE_CACHE_SIZE - 1))
  		return 1;
  	return 0;
  }
  
  static int eb_straddles_pages(struct extent_buffer *eb)
  {
  	return __eb_straddles_pages(eb->start, eb->len);
  }
1259ab75c   Chris Mason   Btrfs: Handle wri...
3743
  int clear_extent_buffer_uptodate(struct extent_io_tree *tree,
2ac55d41b   Josef Bacik   Btrfs: cache the ...
3744
3745
  				struct extent_buffer *eb,
  				struct extent_state **cached_state)
1259ab75c   Chris Mason   Btrfs: Handle wri...
3746
3747
3748
3749
3750
3751
  {
  	unsigned long i;
  	struct page *page;
  	unsigned long num_pages;
  
  	num_pages = num_extent_pages(eb->start, eb->len);
b4ce94de9   Chris Mason   Btrfs: Change btr...
3752
  	clear_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
1259ab75c   Chris Mason   Btrfs: Handle wri...
3753

19b6caf4a   Chris Mason   Btrfs: reduce ext...
3754
3755
3756
3757
  	if (eb_straddles_pages(eb)) {
  		clear_extent_uptodate(tree, eb->start, eb->start + eb->len - 1,
  				      cached_state, GFP_NOFS);
  	}
1259ab75c   Chris Mason   Btrfs: Handle wri...
3758
3759
  	for (i = 0; i < num_pages; i++) {
  		page = extent_buffer_page(eb, i);
33958dc6d   Chris Mason   Btrfs: Fix verify...
3760
3761
  		if (page)
  			ClearPageUptodate(page);
1259ab75c   Chris Mason   Btrfs: Handle wri...
3762
3763
3764
  	}
  	return 0;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
3765
3766
3767
3768
3769
3770
3771
3772
  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);
19b6caf4a   Chris Mason   Btrfs: reduce ext...
3773
3774
3775
3776
  	if (eb_straddles_pages(eb)) {
  		set_extent_uptodate(tree, eb->start, eb->start + eb->len - 1,
  				    NULL, GFP_NOFS);
  	}
d1310b2e0   Chris Mason   Btrfs: Split the ...
3777
3778
3779
3780
3781
3782
3783
3784
3785
3786
3787
3788
  	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;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
3789

ce9adaa5a   Chris Mason   Btrfs: Do metadat...
3790
3791
3792
3793
3794
3795
3796
3797
  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;
19b6caf4a   Chris Mason   Btrfs: reduce ext...
3798
3799
3800
3801
3802
3803
  	if (__eb_straddles_pages(start, end - start + 1)) {
  		ret = test_range_bit(tree, start, end,
  				     EXTENT_UPTODATE, 1, NULL);
  		if (ret)
  			return 1;
  	}
d397712bc   Chris Mason   Btrfs: Fix checkp...
3804
  	while (start <= end) {
ce9adaa5a   Chris Mason   Btrfs: Do metadat...
3805
3806
3807
3808
3809
3810
3811
3812
3813
3814
3815
3816
  		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;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
3817
  int extent_buffer_uptodate(struct extent_io_tree *tree,
2ac55d41b   Josef Bacik   Btrfs: cache the ...
3818
3819
  			   struct extent_buffer *eb,
  			   struct extent_state *cached_state)
d1310b2e0   Chris Mason   Btrfs: Split the ...
3820
  {
728131d8e   Chris Mason   Btrfs: Add additi...
3821
  	int ret = 0;
ce9adaa5a   Chris Mason   Btrfs: Do metadat...
3822
3823
  	unsigned long num_pages;
  	unsigned long i;
728131d8e   Chris Mason   Btrfs: Add additi...
3824
3825
  	struct page *page;
  	int pg_uptodate = 1;
b4ce94de9   Chris Mason   Btrfs: Change btr...
3826
  	if (test_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags))
4235298e4   Chris Mason   Btrfs: Drop some ...
3827
  		return 1;
728131d8e   Chris Mason   Btrfs: Add additi...
3828

19b6caf4a   Chris Mason   Btrfs: reduce ext...
3829
3830
3831
3832
3833
3834
  	if (eb_straddles_pages(eb)) {
  		ret = test_range_bit(tree, eb->start, eb->start + eb->len - 1,
  				   EXTENT_UPTODATE, 1, cached_state);
  		if (ret)
  			return ret;
  	}
728131d8e   Chris Mason   Btrfs: Add additi...
3835
3836
3837
3838
3839
3840
3841
3842
3843
  
  	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;
  		}
  	}
4235298e4   Chris Mason   Btrfs: Drop some ...
3844
  	return pg_uptodate;
d1310b2e0   Chris Mason   Btrfs: Split the ...
3845
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
3846
3847
  
  int read_extent_buffer_pages(struct extent_io_tree *tree,
bb82ab88d   Arne Jansen   btrfs: add an ext...
3848
  			     struct extent_buffer *eb, u64 start, int wait,
f188591e9   Chris Mason   Btrfs: Retry meta...
3849
  			     get_extent_t *get_extent, int mirror_num)
d1310b2e0   Chris Mason   Btrfs: Split the ...
3850
3851
3852
3853
3854
3855
  {
  	unsigned long i;
  	unsigned long start_i;
  	struct page *page;
  	int err;
  	int ret = 0;
ce9adaa5a   Chris Mason   Btrfs: Do metadat...
3856
3857
3858
  	int locked_pages = 0;
  	int all_uptodate = 1;
  	int inc_all_pages = 0;
d1310b2e0   Chris Mason   Btrfs: Split the ...
3859
  	unsigned long num_pages;
a86c12c73   Chris Mason   Btrfs: Create lar...
3860
  	struct bio *bio = NULL;
c8b978188   Chris Mason   Btrfs: Add zlib c...
3861
  	unsigned long bio_flags = 0;
a86c12c73   Chris Mason   Btrfs: Create lar...
3862

b4ce94de9   Chris Mason   Btrfs: Change btr...
3863
  	if (test_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags))
d1310b2e0   Chris Mason   Btrfs: Split the ...
3864
  		return 0;
19b6caf4a   Chris Mason   Btrfs: reduce ext...
3865
3866
3867
3868
3869
  	if (eb_straddles_pages(eb)) {
  		if (test_range_bit(tree, eb->start, eb->start + eb->len - 1,
  				   EXTENT_UPTODATE, 1, NULL)) {
  			return 0;
  		}
d1310b2e0   Chris Mason   Btrfs: Split the ...
3870
3871
3872
3873
3874
3875
3876
3877
3878
3879
3880
3881
3882
  	}
  
  	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);
bb82ab88d   Arne Jansen   btrfs: add an ext...
3883
  		if (wait == WAIT_NONE) {
2db04966a   David Woodhouse   Btrfs: Change Tes...
3884
  			if (!trylock_page(page))
ce9adaa5a   Chris Mason   Btrfs: Do metadat...
3885
  				goto unlock_exit;
d1310b2e0   Chris Mason   Btrfs: Split the ...
3886
3887
3888
  		} else {
  			lock_page(page);
  		}
ce9adaa5a   Chris Mason   Btrfs: Do metadat...
3889
  		locked_pages++;
d397712bc   Chris Mason   Btrfs: Fix checkp...
3890
  		if (!PageUptodate(page))
ce9adaa5a   Chris Mason   Btrfs: Do metadat...
3891
  			all_uptodate = 0;
ce9adaa5a   Chris Mason   Btrfs: Do metadat...
3892
3893
3894
  	}
  	if (all_uptodate) {
  		if (start_i == 0)
b4ce94de9   Chris Mason   Btrfs: Change btr...
3895
  			set_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
ce9adaa5a   Chris Mason   Btrfs: Do metadat...
3896
3897
3898
3899
3900
  		goto unlock_exit;
  	}
  
  	for (i = start_i; i < num_pages; i++) {
  		page = extent_buffer_page(eb, i);
eb14ab8ed   Chris Mason   Btrfs: fix page->...
3901
3902
3903
3904
3905
3906
  
  		WARN_ON(!PagePrivate(page));
  
  		set_page_extent_mapped(page);
  		if (i == 0)
  			set_page_extent_head(page, eb->len);
ce9adaa5a   Chris Mason   Btrfs: Do metadat...
3907
3908
3909
3910
3911
  		if (inc_all_pages)
  			page_cache_get(page);
  		if (!PageUptodate(page)) {
  			if (start_i == 0)
  				inc_all_pages = 1;
f188591e9   Chris Mason   Btrfs: Retry meta...
3912
  			ClearPageError(page);
a86c12c73   Chris Mason   Btrfs: Create lar...
3913
  			err = __extent_read_full_page(tree, page,
f188591e9   Chris Mason   Btrfs: Retry meta...
3914
  						      get_extent, &bio,
c8b978188   Chris Mason   Btrfs: Add zlib c...
3915
  						      mirror_num, &bio_flags);
d397712bc   Chris Mason   Btrfs: Fix checkp...
3916
  			if (err)
d1310b2e0   Chris Mason   Btrfs: Split the ...
3917
  				ret = err;
d1310b2e0   Chris Mason   Btrfs: Split the ...
3918
3919
3920
3921
  		} else {
  			unlock_page(page);
  		}
  	}
a86c12c73   Chris Mason   Btrfs: Create lar...
3922
  	if (bio)
c8b978188   Chris Mason   Btrfs: Add zlib c...
3923
  		submit_one_bio(READ, bio, mirror_num, bio_flags);
a86c12c73   Chris Mason   Btrfs: Create lar...
3924

bb82ab88d   Arne Jansen   btrfs: add an ext...
3925
  	if (ret || wait != WAIT_COMPLETE)
d1310b2e0   Chris Mason   Btrfs: Split the ...
3926
  		return ret;
d397712bc   Chris Mason   Btrfs: Fix checkp...
3927

d1310b2e0   Chris Mason   Btrfs: Split the ...
3928
3929
3930
  	for (i = start_i; i < num_pages; i++) {
  		page = extent_buffer_page(eb, i);
  		wait_on_page_locked(page);
d397712bc   Chris Mason   Btrfs: Fix checkp...
3931
  		if (!PageUptodate(page))
d1310b2e0   Chris Mason   Btrfs: Split the ...
3932
  			ret = -EIO;
d1310b2e0   Chris Mason   Btrfs: Split the ...
3933
  	}
d397712bc   Chris Mason   Btrfs: Fix checkp...
3934

d1310b2e0   Chris Mason   Btrfs: Split the ...
3935
  	if (!ret)
b4ce94de9   Chris Mason   Btrfs: Change btr...
3936
  		set_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
d1310b2e0   Chris Mason   Btrfs: Split the ...
3937
  	return ret;
ce9adaa5a   Chris Mason   Btrfs: Do metadat...
3938
3939
3940
  
  unlock_exit:
  	i = start_i;
d397712bc   Chris Mason   Btrfs: Fix checkp...
3941
  	while (locked_pages > 0) {
ce9adaa5a   Chris Mason   Btrfs: Do metadat...
3942
3943
3944
3945
3946
3947
  		page = extent_buffer_page(eb, i);
  		i++;
  		unlock_page(page);
  		locked_pages--;
  	}
  	return ret;
d1310b2e0   Chris Mason   Btrfs: Split the ...
3948
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
3949
3950
3951
3952
3953
3954
3955
3956
3957
3958
3959
3960
  
  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;
d1310b2e0   Chris Mason   Btrfs: Split the ...
3961
3962
3963
3964
3965
  
  	WARN_ON(start > eb->len);
  	WARN_ON(start + len > eb->start + eb->len);
  
  	offset = (start_offset + start) & ((unsigned long)PAGE_CACHE_SIZE - 1);
d397712bc   Chris Mason   Btrfs: Fix checkp...
3966
  	while (len > 0) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
3967
  		page = extent_buffer_page(eb, i);
d1310b2e0   Chris Mason   Btrfs: Split the ...
3968
3969
  
  		cur = min(len, (PAGE_CACHE_SIZE - offset));
a65917156   Chris Mason   Btrfs: stop using...
3970
  		kaddr = page_address(page);
d1310b2e0   Chris Mason   Btrfs: Split the ...
3971
  		memcpy(dst, kaddr + offset, cur);
d1310b2e0   Chris Mason   Btrfs: Split the ...
3972
3973
3974
3975
3976
3977
3978
  
  		dst += cur;
  		len -= cur;
  		offset = 0;
  		i++;
  	}
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
3979
3980
  
  int map_private_extent_buffer(struct extent_buffer *eb, unsigned long start,
a65917156   Chris Mason   Btrfs: stop using...
3981
  			       unsigned long min_len, char **map,
d1310b2e0   Chris Mason   Btrfs: Split the ...
3982
  			       unsigned long *map_start,
a65917156   Chris Mason   Btrfs: stop using...
3983
  			       unsigned long *map_len)
d1310b2e0   Chris Mason   Btrfs: Split the ...
3984
3985
3986
3987
3988
3989
3990
3991
3992
3993
3994
3995
3996
3997
3998
3999
4000
4001
4002
  {
  	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;
  	}
d397712bc   Chris Mason   Btrfs: Fix checkp...
4003

d1310b2e0   Chris Mason   Btrfs: Split the ...
4004
  	if (start + min_len > eb->len) {
d397712bc   Chris Mason   Btrfs: Fix checkp...
4005
4006
4007
4008
  		printk(KERN_ERR "btrfs bad mapping eb start %llu len %lu, "
  		       "wanted %lu %lu
  ", (unsigned long long)eb->start,
  		       eb->len, start, min_len);
d1310b2e0   Chris Mason   Btrfs: Split the ...
4009
  		WARN_ON(1);
850265335   Josef Bacik   Btrfs: return err...
4010
  		return -EINVAL;
d1310b2e0   Chris Mason   Btrfs: Split the ...
4011
4012
4013
  	}
  
  	p = extent_buffer_page(eb, i);
a65917156   Chris Mason   Btrfs: stop using...
4014
  	kaddr = page_address(p);
d1310b2e0   Chris Mason   Btrfs: Split the ...
4015
4016
4017
4018
  	*map = kaddr + offset;
  	*map_len = PAGE_CACHE_SIZE - offset;
  	return 0;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
4019

d1310b2e0   Chris Mason   Btrfs: Split the ...
4020
4021
4022
4023
4024
4025
4026
4027
4028
4029
4030
4031
4032
4033
4034
4035
4036
  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);
d397712bc   Chris Mason   Btrfs: Fix checkp...
4037
  	while (len > 0) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
4038
  		page = extent_buffer_page(eb, i);
d1310b2e0   Chris Mason   Btrfs: Split the ...
4039
4040
  
  		cur = min(len, (PAGE_CACHE_SIZE - offset));
a65917156   Chris Mason   Btrfs: stop using...
4041
  		kaddr = page_address(page);
d1310b2e0   Chris Mason   Btrfs: Split the ...
4042
  		ret = memcmp(ptr, kaddr + offset, cur);
d1310b2e0   Chris Mason   Btrfs: Split the ...
4043
4044
4045
4046
4047
4048
4049
4050
4051
4052
  		if (ret)
  			break;
  
  		ptr += cur;
  		len -= cur;
  		offset = 0;
  		i++;
  	}
  	return ret;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
4053
4054
4055
4056
4057
4058
4059
4060
4061
4062
4063
4064
4065
4066
4067
4068
  
  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);
d397712bc   Chris Mason   Btrfs: Fix checkp...
4069
  	while (len > 0) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
4070
4071
4072
4073
  		page = extent_buffer_page(eb, i);
  		WARN_ON(!PageUptodate(page));
  
  		cur = min(len, PAGE_CACHE_SIZE - offset);
a65917156   Chris Mason   Btrfs: stop using...
4074
  		kaddr = page_address(page);
d1310b2e0   Chris Mason   Btrfs: Split the ...
4075
  		memcpy(kaddr + offset, src, cur);
d1310b2e0   Chris Mason   Btrfs: Split the ...
4076
4077
4078
4079
4080
4081
4082
  
  		src += cur;
  		len -= cur;
  		offset = 0;
  		i++;
  	}
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
4083
4084
4085
4086
4087
4088
4089
4090
4091
4092
4093
4094
4095
4096
4097
  
  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);
d397712bc   Chris Mason   Btrfs: Fix checkp...
4098
  	while (len > 0) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
4099
4100
4101
4102
  		page = extent_buffer_page(eb, i);
  		WARN_ON(!PageUptodate(page));
  
  		cur = min(len, PAGE_CACHE_SIZE - offset);
a65917156   Chris Mason   Btrfs: stop using...
4103
  		kaddr = page_address(page);
d1310b2e0   Chris Mason   Btrfs: Split the ...
4104
  		memset(kaddr + offset, c, cur);
d1310b2e0   Chris Mason   Btrfs: Split the ...
4105
4106
4107
4108
4109
4110
  
  		len -= cur;
  		offset = 0;
  		i++;
  	}
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
4111
4112
4113
4114
4115
4116
4117
4118
4119
4120
4121
4122
4123
4124
4125
4126
4127
  
  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);
d397712bc   Chris Mason   Btrfs: Fix checkp...
4128
  	while (len > 0) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
4129
4130
4131
4132
  		page = extent_buffer_page(dst, i);
  		WARN_ON(!PageUptodate(page));
  
  		cur = min(len, (unsigned long)(PAGE_CACHE_SIZE - offset));
a65917156   Chris Mason   Btrfs: stop using...
4133
  		kaddr = page_address(page);
d1310b2e0   Chris Mason   Btrfs: Split the ...
4134
  		read_extent_buffer(src, kaddr + offset, src_offset, cur);
d1310b2e0   Chris Mason   Btrfs: Split the ...
4135
4136
4137
4138
4139
4140
4141
  
  		src_offset += cur;
  		len -= cur;
  		offset = 0;
  		i++;
  	}
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
4142
4143
4144
4145
4146
  
  static void move_pages(struct page *dst_page, struct page *src_page,
  		       unsigned long dst_off, unsigned long src_off,
  		       unsigned long len)
  {
a65917156   Chris Mason   Btrfs: stop using...
4147
  	char *dst_kaddr = page_address(dst_page);
d1310b2e0   Chris Mason   Btrfs: Split the ...
4148
4149
4150
  	if (dst_page == src_page) {
  		memmove(dst_kaddr + dst_off, dst_kaddr + src_off, len);
  	} else {
a65917156   Chris Mason   Btrfs: stop using...
4151
  		char *src_kaddr = page_address(src_page);
d1310b2e0   Chris Mason   Btrfs: Split the ...
4152
4153
4154
4155
4156
  		char *p = dst_kaddr + dst_off + len;
  		char *s = src_kaddr + src_off + len;
  
  		while (len--)
  			*--p = *--s;
d1310b2e0   Chris Mason   Btrfs: Split the ...
4157
  	}
d1310b2e0   Chris Mason   Btrfs: Split the ...
4158
  }
3387206f2   Sergei Trofimovich   btrfs: properly h...
4159
4160
4161
4162
4163
  static inline bool areas_overlap(unsigned long src, unsigned long dst, unsigned long len)
  {
  	unsigned long distance = (src > dst) ? src - dst : dst - src;
  	return distance < len;
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
4164
4165
4166
4167
  static void copy_pages(struct page *dst_page, struct page *src_page,
  		       unsigned long dst_off, unsigned long src_off,
  		       unsigned long len)
  {
a65917156   Chris Mason   Btrfs: stop using...
4168
  	char *dst_kaddr = page_address(dst_page);
d1310b2e0   Chris Mason   Btrfs: Split the ...
4169
  	char *src_kaddr;
3387206f2   Sergei Trofimovich   btrfs: properly h...
4170
  	if (dst_page != src_page) {
a65917156   Chris Mason   Btrfs: stop using...
4171
  		src_kaddr = page_address(src_page);
3387206f2   Sergei Trofimovich   btrfs: properly h...
4172
  	} else {
d1310b2e0   Chris Mason   Btrfs: Split the ...
4173
  		src_kaddr = dst_kaddr;
3387206f2   Sergei Trofimovich   btrfs: properly h...
4174
4175
  		BUG_ON(areas_overlap(src_off, dst_off, len));
  	}
d1310b2e0   Chris Mason   Btrfs: Split the ...
4176
4177
  
  	memcpy(dst_kaddr + dst_off, src_kaddr + src_off, len);
d1310b2e0   Chris Mason   Btrfs: Split the ...
4178
4179
4180
4181
4182
4183
4184
4185
4186
4187
4188
4189
4190
  }
  
  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) {
d397712bc   Chris Mason   Btrfs: Fix checkp...
4191
4192
4193
  		printk(KERN_ERR "btrfs memmove bogus src_offset %lu move "
  		       "len %lu dst len %lu
  ", src_offset, len, dst->len);
d1310b2e0   Chris Mason   Btrfs: Split the ...
4194
4195
4196
  		BUG_ON(1);
  	}
  	if (dst_offset + len > dst->len) {
d397712bc   Chris Mason   Btrfs: Fix checkp...
4197
4198
4199
  		printk(KERN_ERR "btrfs memmove bogus dst_offset %lu move "
  		       "len %lu dst len %lu
  ", dst_offset, len, dst->len);
d1310b2e0   Chris Mason   Btrfs: Split the ...
4200
4201
  		BUG_ON(1);
  	}
d397712bc   Chris Mason   Btrfs: Fix checkp...
4202
  	while (len > 0) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
4203
4204
4205
4206
4207
4208
4209
4210
4211
4212
4213
4214
4215
4216
4217
4218
4219
4220
4221
4222
4223
4224
  		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;
  	}
  }
d1310b2e0   Chris Mason   Btrfs: Split the ...
4225
4226
4227
4228
4229
4230
4231
4232
4233
4234
4235
4236
4237
4238
  
  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) {
d397712bc   Chris Mason   Btrfs: Fix checkp...
4239
4240
4241
  		printk(KERN_ERR "btrfs memmove bogus src_offset %lu move "
  		       "len %lu len %lu
  ", src_offset, len, dst->len);
d1310b2e0   Chris Mason   Btrfs: Split the ...
4242
4243
4244
  		BUG_ON(1);
  	}
  	if (dst_offset + len > dst->len) {
d397712bc   Chris Mason   Btrfs: Fix checkp...
4245
4246
4247
  		printk(KERN_ERR "btrfs memmove bogus dst_offset %lu move "
  		       "len %lu len %lu
  ", dst_offset, len, dst->len);
d1310b2e0   Chris Mason   Btrfs: Split the ...
4248
4249
  		BUG_ON(1);
  	}
3387206f2   Sergei Trofimovich   btrfs: properly h...
4250
  	if (!areas_overlap(src_offset, dst_offset, len)) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
4251
4252
4253
  		memcpy_extent_buffer(dst, dst_offset, src_offset, len);
  		return;
  	}
d397712bc   Chris Mason   Btrfs: Fix checkp...
4254
  	while (len > 0) {
d1310b2e0   Chris Mason   Btrfs: Split the ...
4255
4256
4257
4258
4259
4260
4261
4262
4263
4264
4265
4266
4267
4268
4269
4270
4271
4272
4273
4274
  		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;
  	}
  }
6af118ce5   Chris Mason   Btrfs: Index exte...
4275

19fe0a8b7   Miao Xie   Btrfs: Switch the...
4276
4277
4278
4279
4280
4281
4282
  static inline void btrfs_release_extent_buffer_rcu(struct rcu_head *head)
  {
  	struct extent_buffer *eb =
  			container_of(head, struct extent_buffer, rcu_head);
  
  	btrfs_release_extent_buffer(eb);
  }
6af118ce5   Chris Mason   Btrfs: Index exte...
4283
4284
4285
4286
4287
  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;
6af118ce5   Chris Mason   Btrfs: Index exte...
4288
4289
  
  	spin_lock(&tree->buffer_lock);
19fe0a8b7   Miao Xie   Btrfs: Switch the...
4290
  	eb = radix_tree_lookup(&tree->buffer, start >> PAGE_CACHE_SHIFT);
45f49bce9   Chris Mason   Btrfs: avoid NULL...
4291
4292
4293
4294
  	if (!eb) {
  		spin_unlock(&tree->buffer_lock);
  		return ret;
  	}
6af118ce5   Chris Mason   Btrfs: Index exte...
4295

19fe0a8b7   Miao Xie   Btrfs: Switch the...
4296
  	if (test_bit(EXTENT_BUFFER_DIRTY, &eb->bflags)) {
6af118ce5   Chris Mason   Btrfs: Index exte...
4297
4298
4299
  		ret = 0;
  		goto out;
  	}
19fe0a8b7   Miao Xie   Btrfs: Switch the...
4300
4301
4302
4303
4304
4305
  
  	/*
  	 * set @eb->refs to 0 if it is already 1, and then release the @eb.
  	 * Or go back.
  	 */
  	if (atomic_cmpxchg(&eb->refs, 1, 0) != 1) {
b9473439d   Chris Mason   Btrfs: leave btre...
4306
4307
4308
  		ret = 0;
  		goto out;
  	}
897ca6e9b   Miao Xie   Btrfs: restructur...
4309

19fe0a8b7   Miao Xie   Btrfs: Switch the...
4310
  	radix_tree_delete(&tree->buffer, start >> PAGE_CACHE_SHIFT);
6af118ce5   Chris Mason   Btrfs: Index exte...
4311
4312
  out:
  	spin_unlock(&tree->buffer_lock);
19fe0a8b7   Miao Xie   Btrfs: Switch the...
4313
4314
4315
4316
  
  	/* at this point we can safely release the extent buffer */
  	if (atomic_read(&eb->refs) == 0)
  		call_rcu(&eb->rcu_head, btrfs_release_extent_buffer_rcu);
6af118ce5   Chris Mason   Btrfs: Index exte...
4317
4318
  	return ret;
  }