xref: /openbmc/linux/fs/btrfs/free-space-tree.c (revision 78700c0a)
1 /*
2  * Copyright (C) 2015 Facebook.  All rights reserved.
3  *
4  * This program is free software; you can redistribute it and/or
5  * modify it under the terms of the GNU General Public
6  * License v2 as published by the Free Software Foundation.
7  *
8  * This program is distributed in the hope that it will be useful,
9  * but WITHOUT ANY WARRANTY; without even the implied warranty of
10  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
11  * General Public License for more details.
12  *
13  * You should have received a copy of the GNU General Public
14  * License along with this program; if not, write to the
15  * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
16  * Boston, MA 021110-1307, USA.
17  */
18 
19 #include <linux/kernel.h>
20 #include <linux/vmalloc.h>
21 #include "ctree.h"
22 #include "disk-io.h"
23 #include "locking.h"
24 #include "free-space-tree.h"
25 #include "transaction.h"
26 
27 static int __add_block_group_free_space(struct btrfs_trans_handle *trans,
28 					struct btrfs_fs_info *fs_info,
29 					struct btrfs_block_group_cache *block_group,
30 					struct btrfs_path *path);
31 
32 void set_free_space_tree_thresholds(struct btrfs_block_group_cache *cache)
33 {
34 	u32 bitmap_range;
35 	size_t bitmap_size;
36 	u64 num_bitmaps, total_bitmap_size;
37 
38 	/*
39 	 * We convert to bitmaps when the disk space required for using extents
40 	 * exceeds that required for using bitmaps.
41 	 */
42 	bitmap_range = cache->sectorsize * BTRFS_FREE_SPACE_BITMAP_BITS;
43 	num_bitmaps = div_u64(cache->key.offset + bitmap_range - 1,
44 			      bitmap_range);
45 	bitmap_size = sizeof(struct btrfs_item) + BTRFS_FREE_SPACE_BITMAP_SIZE;
46 	total_bitmap_size = num_bitmaps * bitmap_size;
47 	cache->bitmap_high_thresh = div_u64(total_bitmap_size,
48 					    sizeof(struct btrfs_item));
49 
50 	/*
51 	 * We allow for a small buffer between the high threshold and low
52 	 * threshold to avoid thrashing back and forth between the two formats.
53 	 */
54 	if (cache->bitmap_high_thresh > 100)
55 		cache->bitmap_low_thresh = cache->bitmap_high_thresh - 100;
56 	else
57 		cache->bitmap_low_thresh = 0;
58 }
59 
60 static int add_new_free_space_info(struct btrfs_trans_handle *trans,
61 				   struct btrfs_fs_info *fs_info,
62 				   struct btrfs_block_group_cache *block_group,
63 				   struct btrfs_path *path)
64 {
65 	struct btrfs_root *root = fs_info->free_space_root;
66 	struct btrfs_free_space_info *info;
67 	struct btrfs_key key;
68 	struct extent_buffer *leaf;
69 	int ret;
70 
71 	key.objectid = block_group->key.objectid;
72 	key.type = BTRFS_FREE_SPACE_INFO_KEY;
73 	key.offset = block_group->key.offset;
74 
75 	ret = btrfs_insert_empty_item(trans, root, path, &key, sizeof(*info));
76 	if (ret)
77 		goto out;
78 
79 	leaf = path->nodes[0];
80 	info = btrfs_item_ptr(leaf, path->slots[0],
81 			      struct btrfs_free_space_info);
82 	btrfs_set_free_space_extent_count(leaf, info, 0);
83 	btrfs_set_free_space_flags(leaf, info, 0);
84 	btrfs_mark_buffer_dirty(leaf);
85 
86 	ret = 0;
87 out:
88 	btrfs_release_path(path);
89 	return ret;
90 }
91 
92 struct btrfs_free_space_info *
93 search_free_space_info(struct btrfs_trans_handle *trans,
94 		       struct btrfs_fs_info *fs_info,
95 		       struct btrfs_block_group_cache *block_group,
96 		       struct btrfs_path *path, int cow)
97 {
98 	struct btrfs_root *root = fs_info->free_space_root;
99 	struct btrfs_key key;
100 	int ret;
101 
102 	key.objectid = block_group->key.objectid;
103 	key.type = BTRFS_FREE_SPACE_INFO_KEY;
104 	key.offset = block_group->key.offset;
105 
106 	ret = btrfs_search_slot(trans, root, &key, path, 0, cow);
107 	if (ret < 0)
108 		return ERR_PTR(ret);
109 	if (ret != 0) {
110 		btrfs_warn(fs_info, "missing free space info for %llu\n",
111 			   block_group->key.objectid);
112 		ASSERT(0);
113 		return ERR_PTR(-ENOENT);
114 	}
115 
116 	return btrfs_item_ptr(path->nodes[0], path->slots[0],
117 			      struct btrfs_free_space_info);
118 }
119 
120 /*
121  * btrfs_search_slot() but we're looking for the greatest key less than the
122  * passed key.
123  */
124 static int btrfs_search_prev_slot(struct btrfs_trans_handle *trans,
125 				  struct btrfs_root *root,
126 				  struct btrfs_key *key, struct btrfs_path *p,
127 				  int ins_len, int cow)
128 {
129 	int ret;
130 
131 	ret = btrfs_search_slot(trans, root, key, p, ins_len, cow);
132 	if (ret < 0)
133 		return ret;
134 
135 	if (ret == 0) {
136 		ASSERT(0);
137 		return -EIO;
138 	}
139 
140 	if (p->slots[0] == 0) {
141 		ASSERT(0);
142 		return -EIO;
143 	}
144 	p->slots[0]--;
145 
146 	return 0;
147 }
148 
149 static inline u32 free_space_bitmap_size(u64 size, u32 sectorsize)
150 {
151 	return DIV_ROUND_UP((u32)div_u64(size, sectorsize), BITS_PER_BYTE);
152 }
153 
154 static unsigned long *alloc_bitmap(u32 bitmap_size)
155 {
156 	void *mem;
157 
158 	/*
159 	 * The allocation size varies, observed numbers were < 4K up to 16K.
160 	 * Using vmalloc unconditionally would be too heavy, we'll try
161 	 * contiguous allocations first.
162 	 */
163 	if  (bitmap_size <= PAGE_SIZE)
164 		return kzalloc(bitmap_size, GFP_NOFS);
165 
166 	mem = kzalloc(bitmap_size, GFP_NOFS | __GFP_NOWARN);
167 	if (mem)
168 		return mem;
169 
170 	return __vmalloc(bitmap_size, GFP_NOFS | __GFP_HIGHMEM | __GFP_ZERO,
171 			 PAGE_KERNEL);
172 }
173 
174 int convert_free_space_to_bitmaps(struct btrfs_trans_handle *trans,
175 				  struct btrfs_fs_info *fs_info,
176 				  struct btrfs_block_group_cache *block_group,
177 				  struct btrfs_path *path)
178 {
179 	struct btrfs_root *root = fs_info->free_space_root;
180 	struct btrfs_free_space_info *info;
181 	struct btrfs_key key, found_key;
182 	struct extent_buffer *leaf;
183 	unsigned long *bitmap;
184 	char *bitmap_cursor;
185 	u64 start, end;
186 	u64 bitmap_range, i;
187 	u32 bitmap_size, flags, expected_extent_count;
188 	u32 extent_count = 0;
189 	int done = 0, nr;
190 	int ret;
191 
192 	bitmap_size = free_space_bitmap_size(block_group->key.offset,
193 					     block_group->sectorsize);
194 	bitmap = alloc_bitmap(bitmap_size);
195 	if (!bitmap) {
196 		ret = -ENOMEM;
197 		goto out;
198 	}
199 
200 	start = block_group->key.objectid;
201 	end = block_group->key.objectid + block_group->key.offset;
202 
203 	key.objectid = end - 1;
204 	key.type = (u8)-1;
205 	key.offset = (u64)-1;
206 
207 	while (!done) {
208 		ret = btrfs_search_prev_slot(trans, root, &key, path, -1, 1);
209 		if (ret)
210 			goto out;
211 
212 		leaf = path->nodes[0];
213 		nr = 0;
214 		path->slots[0]++;
215 		while (path->slots[0] > 0) {
216 			btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0] - 1);
217 
218 			if (found_key.type == BTRFS_FREE_SPACE_INFO_KEY) {
219 				ASSERT(found_key.objectid == block_group->key.objectid);
220 				ASSERT(found_key.offset == block_group->key.offset);
221 				done = 1;
222 				break;
223 			} else if (found_key.type == BTRFS_FREE_SPACE_EXTENT_KEY) {
224 				u64 first, last;
225 
226 				ASSERT(found_key.objectid >= start);
227 				ASSERT(found_key.objectid < end);
228 				ASSERT(found_key.objectid + found_key.offset <= end);
229 
230 				first = div_u64(found_key.objectid - start,
231 						block_group->sectorsize);
232 				last = div_u64(found_key.objectid + found_key.offset - start,
233 					       block_group->sectorsize);
234 				bitmap_set(bitmap, first, last - first);
235 
236 				extent_count++;
237 				nr++;
238 				path->slots[0]--;
239 			} else {
240 				ASSERT(0);
241 			}
242 		}
243 
244 		ret = btrfs_del_items(trans, root, path, path->slots[0], nr);
245 		if (ret)
246 			goto out;
247 		btrfs_release_path(path);
248 	}
249 
250 	info = search_free_space_info(trans, fs_info, block_group, path, 1);
251 	if (IS_ERR(info)) {
252 		ret = PTR_ERR(info);
253 		goto out;
254 	}
255 	leaf = path->nodes[0];
256 	flags = btrfs_free_space_flags(leaf, info);
257 	flags |= BTRFS_FREE_SPACE_USING_BITMAPS;
258 	btrfs_set_free_space_flags(leaf, info, flags);
259 	expected_extent_count = btrfs_free_space_extent_count(leaf, info);
260 	btrfs_mark_buffer_dirty(leaf);
261 	btrfs_release_path(path);
262 
263 	if (extent_count != expected_extent_count) {
264 		btrfs_err(fs_info, "incorrect extent count for %llu; counted %u, expected %u",
265 			  block_group->key.objectid, extent_count,
266 			  expected_extent_count);
267 		ASSERT(0);
268 		ret = -EIO;
269 		goto out;
270 	}
271 
272 	bitmap_cursor = (char *)bitmap;
273 	bitmap_range = block_group->sectorsize * BTRFS_FREE_SPACE_BITMAP_BITS;
274 	i = start;
275 	while (i < end) {
276 		unsigned long ptr;
277 		u64 extent_size;
278 		u32 data_size;
279 
280 		extent_size = min(end - i, bitmap_range);
281 		data_size = free_space_bitmap_size(extent_size,
282 						   block_group->sectorsize);
283 
284 		key.objectid = i;
285 		key.type = BTRFS_FREE_SPACE_BITMAP_KEY;
286 		key.offset = extent_size;
287 
288 		ret = btrfs_insert_empty_item(trans, root, path, &key,
289 					      data_size);
290 		if (ret)
291 			goto out;
292 
293 		leaf = path->nodes[0];
294 		ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
295 		write_extent_buffer(leaf, bitmap_cursor, ptr,
296 				    data_size);
297 		btrfs_mark_buffer_dirty(leaf);
298 		btrfs_release_path(path);
299 
300 		i += extent_size;
301 		bitmap_cursor += data_size;
302 	}
303 
304 	ret = 0;
305 out:
306 	kvfree(bitmap);
307 	if (ret)
308 		btrfs_abort_transaction(trans, root, ret);
309 	return ret;
310 }
311 
312 int convert_free_space_to_extents(struct btrfs_trans_handle *trans,
313 				  struct btrfs_fs_info *fs_info,
314 				  struct btrfs_block_group_cache *block_group,
315 				  struct btrfs_path *path)
316 {
317 	struct btrfs_root *root = fs_info->free_space_root;
318 	struct btrfs_free_space_info *info;
319 	struct btrfs_key key, found_key;
320 	struct extent_buffer *leaf;
321 	unsigned long *bitmap;
322 	u64 start, end;
323 	/* Initialize to silence GCC. */
324 	u64 extent_start = 0;
325 	u64 offset;
326 	u32 bitmap_size, flags, expected_extent_count;
327 	int prev_bit = 0, bit, bitnr;
328 	u32 extent_count = 0;
329 	int done = 0, nr;
330 	int ret;
331 
332 	bitmap_size = free_space_bitmap_size(block_group->key.offset,
333 					     block_group->sectorsize);
334 	bitmap = alloc_bitmap(bitmap_size);
335 	if (!bitmap) {
336 		ret = -ENOMEM;
337 		goto out;
338 	}
339 
340 	start = block_group->key.objectid;
341 	end = block_group->key.objectid + block_group->key.offset;
342 
343 	key.objectid = end - 1;
344 	key.type = (u8)-1;
345 	key.offset = (u64)-1;
346 
347 	while (!done) {
348 		ret = btrfs_search_prev_slot(trans, root, &key, path, -1, 1);
349 		if (ret)
350 			goto out;
351 
352 		leaf = path->nodes[0];
353 		nr = 0;
354 		path->slots[0]++;
355 		while (path->slots[0] > 0) {
356 			btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0] - 1);
357 
358 			if (found_key.type == BTRFS_FREE_SPACE_INFO_KEY) {
359 				ASSERT(found_key.objectid == block_group->key.objectid);
360 				ASSERT(found_key.offset == block_group->key.offset);
361 				done = 1;
362 				break;
363 			} else if (found_key.type == BTRFS_FREE_SPACE_BITMAP_KEY) {
364 				unsigned long ptr;
365 				char *bitmap_cursor;
366 				u32 bitmap_pos, data_size;
367 
368 				ASSERT(found_key.objectid >= start);
369 				ASSERT(found_key.objectid < end);
370 				ASSERT(found_key.objectid + found_key.offset <= end);
371 
372 				bitmap_pos = div_u64(found_key.objectid - start,
373 						     block_group->sectorsize *
374 						     BITS_PER_BYTE);
375 				bitmap_cursor = ((char *)bitmap) + bitmap_pos;
376 				data_size = free_space_bitmap_size(found_key.offset,
377 								   block_group->sectorsize);
378 
379 				ptr = btrfs_item_ptr_offset(leaf, path->slots[0] - 1);
380 				read_extent_buffer(leaf, bitmap_cursor, ptr,
381 						   data_size);
382 
383 				nr++;
384 				path->slots[0]--;
385 			} else {
386 				ASSERT(0);
387 			}
388 		}
389 
390 		ret = btrfs_del_items(trans, root, path, path->slots[0], nr);
391 		if (ret)
392 			goto out;
393 		btrfs_release_path(path);
394 	}
395 
396 	info = search_free_space_info(trans, fs_info, block_group, path, 1);
397 	if (IS_ERR(info)) {
398 		ret = PTR_ERR(info);
399 		goto out;
400 	}
401 	leaf = path->nodes[0];
402 	flags = btrfs_free_space_flags(leaf, info);
403 	flags &= ~BTRFS_FREE_SPACE_USING_BITMAPS;
404 	btrfs_set_free_space_flags(leaf, info, flags);
405 	expected_extent_count = btrfs_free_space_extent_count(leaf, info);
406 	btrfs_mark_buffer_dirty(leaf);
407 	btrfs_release_path(path);
408 
409 	offset = start;
410 	bitnr = 0;
411 	while (offset < end) {
412 		bit = !!test_bit(bitnr, bitmap);
413 		if (prev_bit == 0 && bit == 1) {
414 			extent_start = offset;
415 		} else if (prev_bit == 1 && bit == 0) {
416 			key.objectid = extent_start;
417 			key.type = BTRFS_FREE_SPACE_EXTENT_KEY;
418 			key.offset = offset - extent_start;
419 
420 			ret = btrfs_insert_empty_item(trans, root, path, &key, 0);
421 			if (ret)
422 				goto out;
423 			btrfs_release_path(path);
424 
425 			extent_count++;
426 		}
427 		prev_bit = bit;
428 		offset += block_group->sectorsize;
429 		bitnr++;
430 	}
431 	if (prev_bit == 1) {
432 		key.objectid = extent_start;
433 		key.type = BTRFS_FREE_SPACE_EXTENT_KEY;
434 		key.offset = end - extent_start;
435 
436 		ret = btrfs_insert_empty_item(trans, root, path, &key, 0);
437 		if (ret)
438 			goto out;
439 		btrfs_release_path(path);
440 
441 		extent_count++;
442 	}
443 
444 	if (extent_count != expected_extent_count) {
445 		btrfs_err(fs_info, "incorrect extent count for %llu; counted %u, expected %u",
446 			  block_group->key.objectid, extent_count,
447 			  expected_extent_count);
448 		ASSERT(0);
449 		ret = -EIO;
450 		goto out;
451 	}
452 
453 	ret = 0;
454 out:
455 	kvfree(bitmap);
456 	if (ret)
457 		btrfs_abort_transaction(trans, root, ret);
458 	return ret;
459 }
460 
461 static int update_free_space_extent_count(struct btrfs_trans_handle *trans,
462 					  struct btrfs_fs_info *fs_info,
463 					  struct btrfs_block_group_cache *block_group,
464 					  struct btrfs_path *path,
465 					  int new_extents)
466 {
467 	struct btrfs_free_space_info *info;
468 	u32 flags;
469 	u32 extent_count;
470 	int ret = 0;
471 
472 	if (new_extents == 0)
473 		return 0;
474 
475 	info = search_free_space_info(trans, fs_info, block_group, path, 1);
476 	if (IS_ERR(info)) {
477 		ret = PTR_ERR(info);
478 		goto out;
479 	}
480 	flags = btrfs_free_space_flags(path->nodes[0], info);
481 	extent_count = btrfs_free_space_extent_count(path->nodes[0], info);
482 
483 	extent_count += new_extents;
484 	btrfs_set_free_space_extent_count(path->nodes[0], info, extent_count);
485 	btrfs_mark_buffer_dirty(path->nodes[0]);
486 	btrfs_release_path(path);
487 
488 	if (!(flags & BTRFS_FREE_SPACE_USING_BITMAPS) &&
489 	    extent_count > block_group->bitmap_high_thresh) {
490 		ret = convert_free_space_to_bitmaps(trans, fs_info, block_group,
491 						    path);
492 	} else if ((flags & BTRFS_FREE_SPACE_USING_BITMAPS) &&
493 		   extent_count < block_group->bitmap_low_thresh) {
494 		ret = convert_free_space_to_extents(trans, fs_info, block_group,
495 						    path);
496 	}
497 
498 out:
499 	return ret;
500 }
501 
502 int free_space_test_bit(struct btrfs_block_group_cache *block_group,
503 			struct btrfs_path *path, u64 offset)
504 {
505 	struct extent_buffer *leaf;
506 	struct btrfs_key key;
507 	u64 found_start, found_end;
508 	unsigned long ptr, i;
509 
510 	leaf = path->nodes[0];
511 	btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
512 	ASSERT(key.type == BTRFS_FREE_SPACE_BITMAP_KEY);
513 
514 	found_start = key.objectid;
515 	found_end = key.objectid + key.offset;
516 	ASSERT(offset >= found_start && offset < found_end);
517 
518 	ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
519 	i = div_u64(offset - found_start, block_group->sectorsize);
520 	return !!extent_buffer_test_bit(leaf, ptr, i);
521 }
522 
523 static void free_space_set_bits(struct btrfs_block_group_cache *block_group,
524 				struct btrfs_path *path, u64 *start, u64 *size,
525 				int bit)
526 {
527 	struct extent_buffer *leaf;
528 	struct btrfs_key key;
529 	u64 end = *start + *size;
530 	u64 found_start, found_end;
531 	unsigned long ptr, first, last;
532 
533 	leaf = path->nodes[0];
534 	btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
535 	ASSERT(key.type == BTRFS_FREE_SPACE_BITMAP_KEY);
536 
537 	found_start = key.objectid;
538 	found_end = key.objectid + key.offset;
539 	ASSERT(*start >= found_start && *start < found_end);
540 	ASSERT(end > found_start);
541 
542 	if (end > found_end)
543 		end = found_end;
544 
545 	ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
546 	first = div_u64(*start - found_start, block_group->sectorsize);
547 	last = div_u64(end - found_start, block_group->sectorsize);
548 	if (bit)
549 		extent_buffer_bitmap_set(leaf, ptr, first, last - first);
550 	else
551 		extent_buffer_bitmap_clear(leaf, ptr, first, last - first);
552 	btrfs_mark_buffer_dirty(leaf);
553 
554 	*size -= end - *start;
555 	*start = end;
556 }
557 
558 /*
559  * We can't use btrfs_next_item() in modify_free_space_bitmap() because
560  * btrfs_next_leaf() doesn't get the path for writing. We can forgo the fancy
561  * tree walking in btrfs_next_leaf() anyways because we know exactly what we're
562  * looking for.
563  */
564 static int free_space_next_bitmap(struct btrfs_trans_handle *trans,
565 				  struct btrfs_root *root, struct btrfs_path *p)
566 {
567 	struct btrfs_key key;
568 
569 	if (p->slots[0] + 1 < btrfs_header_nritems(p->nodes[0])) {
570 		p->slots[0]++;
571 		return 0;
572 	}
573 
574 	btrfs_item_key_to_cpu(p->nodes[0], &key, p->slots[0]);
575 	btrfs_release_path(p);
576 
577 	key.objectid += key.offset;
578 	key.type = (u8)-1;
579 	key.offset = (u64)-1;
580 
581 	return btrfs_search_prev_slot(trans, root, &key, p, 0, 1);
582 }
583 
584 /*
585  * If remove is 1, then we are removing free space, thus clearing bits in the
586  * bitmap. If remove is 0, then we are adding free space, thus setting bits in
587  * the bitmap.
588  */
589 static int modify_free_space_bitmap(struct btrfs_trans_handle *trans,
590 				    struct btrfs_fs_info *fs_info,
591 				    struct btrfs_block_group_cache *block_group,
592 				    struct btrfs_path *path,
593 				    u64 start, u64 size, int remove)
594 {
595 	struct btrfs_root *root = fs_info->free_space_root;
596 	struct btrfs_key key;
597 	u64 end = start + size;
598 	u64 cur_start, cur_size;
599 	int prev_bit, next_bit;
600 	int new_extents;
601 	int ret;
602 
603 	/*
604 	 * Read the bit for the block immediately before the extent of space if
605 	 * that block is within the block group.
606 	 */
607 	if (start > block_group->key.objectid) {
608 		u64 prev_block = start - block_group->sectorsize;
609 
610 		key.objectid = prev_block;
611 		key.type = (u8)-1;
612 		key.offset = (u64)-1;
613 
614 		ret = btrfs_search_prev_slot(trans, root, &key, path, 0, 1);
615 		if (ret)
616 			goto out;
617 
618 		prev_bit = free_space_test_bit(block_group, path, prev_block);
619 
620 		/* The previous block may have been in the previous bitmap. */
621 		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
622 		if (start >= key.objectid + key.offset) {
623 			ret = free_space_next_bitmap(trans, root, path);
624 			if (ret)
625 				goto out;
626 		}
627 	} else {
628 		key.objectid = start;
629 		key.type = (u8)-1;
630 		key.offset = (u64)-1;
631 
632 		ret = btrfs_search_prev_slot(trans, root, &key, path, 0, 1);
633 		if (ret)
634 			goto out;
635 
636 		prev_bit = -1;
637 	}
638 
639 	/*
640 	 * Iterate over all of the bitmaps overlapped by the extent of space,
641 	 * clearing/setting bits as required.
642 	 */
643 	cur_start = start;
644 	cur_size = size;
645 	while (1) {
646 		free_space_set_bits(block_group, path, &cur_start, &cur_size,
647 				    !remove);
648 		if (cur_size == 0)
649 			break;
650 		ret = free_space_next_bitmap(trans, root, path);
651 		if (ret)
652 			goto out;
653 	}
654 
655 	/*
656 	 * Read the bit for the block immediately after the extent of space if
657 	 * that block is within the block group.
658 	 */
659 	if (end < block_group->key.objectid + block_group->key.offset) {
660 		/* The next block may be in the next bitmap. */
661 		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
662 		if (end >= key.objectid + key.offset) {
663 			ret = free_space_next_bitmap(trans, root, path);
664 			if (ret)
665 				goto out;
666 		}
667 
668 		next_bit = free_space_test_bit(block_group, path, end);
669 	} else {
670 		next_bit = -1;
671 	}
672 
673 	if (remove) {
674 		new_extents = -1;
675 		if (prev_bit == 1) {
676 			/* Leftover on the left. */
677 			new_extents++;
678 		}
679 		if (next_bit == 1) {
680 			/* Leftover on the right. */
681 			new_extents++;
682 		}
683 	} else {
684 		new_extents = 1;
685 		if (prev_bit == 1) {
686 			/* Merging with neighbor on the left. */
687 			new_extents--;
688 		}
689 		if (next_bit == 1) {
690 			/* Merging with neighbor on the right. */
691 			new_extents--;
692 		}
693 	}
694 
695 	btrfs_release_path(path);
696 	ret = update_free_space_extent_count(trans, fs_info, block_group, path,
697 					     new_extents);
698 
699 out:
700 	return ret;
701 }
702 
703 static int remove_free_space_extent(struct btrfs_trans_handle *trans,
704 				    struct btrfs_fs_info *fs_info,
705 				    struct btrfs_block_group_cache *block_group,
706 				    struct btrfs_path *path,
707 				    u64 start, u64 size)
708 {
709 	struct btrfs_root *root = fs_info->free_space_root;
710 	struct btrfs_key key;
711 	u64 found_start, found_end;
712 	u64 end = start + size;
713 	int new_extents = -1;
714 	int ret;
715 
716 	key.objectid = start;
717 	key.type = (u8)-1;
718 	key.offset = (u64)-1;
719 
720 	ret = btrfs_search_prev_slot(trans, root, &key, path, -1, 1);
721 	if (ret)
722 		goto out;
723 
724 	btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
725 
726 	ASSERT(key.type == BTRFS_FREE_SPACE_EXTENT_KEY);
727 
728 	found_start = key.objectid;
729 	found_end = key.objectid + key.offset;
730 	ASSERT(start >= found_start && end <= found_end);
731 
732 	/*
733 	 * Okay, now that we've found the free space extent which contains the
734 	 * free space that we are removing, there are four cases:
735 	 *
736 	 * 1. We're using the whole extent: delete the key we found and
737 	 * decrement the free space extent count.
738 	 * 2. We are using part of the extent starting at the beginning: delete
739 	 * the key we found and insert a new key representing the leftover at
740 	 * the end. There is no net change in the number of extents.
741 	 * 3. We are using part of the extent ending at the end: delete the key
742 	 * we found and insert a new key representing the leftover at the
743 	 * beginning. There is no net change in the number of extents.
744 	 * 4. We are using part of the extent in the middle: delete the key we
745 	 * found and insert two new keys representing the leftovers on each
746 	 * side. Where we used to have one extent, we now have two, so increment
747 	 * the extent count. We may need to convert the block group to bitmaps
748 	 * as a result.
749 	 */
750 
751 	/* Delete the existing key (cases 1-4). */
752 	ret = btrfs_del_item(trans, root, path);
753 	if (ret)
754 		goto out;
755 
756 	/* Add a key for leftovers at the beginning (cases 3 and 4). */
757 	if (start > found_start) {
758 		key.objectid = found_start;
759 		key.type = BTRFS_FREE_SPACE_EXTENT_KEY;
760 		key.offset = start - found_start;
761 
762 		btrfs_release_path(path);
763 		ret = btrfs_insert_empty_item(trans, root, path, &key, 0);
764 		if (ret)
765 			goto out;
766 		new_extents++;
767 	}
768 
769 	/* Add a key for leftovers at the end (cases 2 and 4). */
770 	if (end < found_end) {
771 		key.objectid = end;
772 		key.type = BTRFS_FREE_SPACE_EXTENT_KEY;
773 		key.offset = found_end - end;
774 
775 		btrfs_release_path(path);
776 		ret = btrfs_insert_empty_item(trans, root, path, &key, 0);
777 		if (ret)
778 			goto out;
779 		new_extents++;
780 	}
781 
782 	btrfs_release_path(path);
783 	ret = update_free_space_extent_count(trans, fs_info, block_group, path,
784 					     new_extents);
785 
786 out:
787 	return ret;
788 }
789 
790 int __remove_from_free_space_tree(struct btrfs_trans_handle *trans,
791 				  struct btrfs_fs_info *fs_info,
792 				  struct btrfs_block_group_cache *block_group,
793 				  struct btrfs_path *path, u64 start, u64 size)
794 {
795 	struct btrfs_free_space_info *info;
796 	u32 flags;
797 	int ret;
798 
799 	if (block_group->needs_free_space) {
800 		ret = __add_block_group_free_space(trans, fs_info, block_group,
801 						   path);
802 		if (ret)
803 			return ret;
804 	}
805 
806 	info = search_free_space_info(NULL, fs_info, block_group, path, 0);
807 	if (IS_ERR(info))
808 		return PTR_ERR(info);
809 	flags = btrfs_free_space_flags(path->nodes[0], info);
810 	btrfs_release_path(path);
811 
812 	if (flags & BTRFS_FREE_SPACE_USING_BITMAPS) {
813 		return modify_free_space_bitmap(trans, fs_info, block_group,
814 						path, start, size, 1);
815 	} else {
816 		return remove_free_space_extent(trans, fs_info, block_group,
817 						path, start, size);
818 	}
819 }
820 
821 int remove_from_free_space_tree(struct btrfs_trans_handle *trans,
822 				struct btrfs_fs_info *fs_info,
823 				u64 start, u64 size)
824 {
825 	struct btrfs_block_group_cache *block_group;
826 	struct btrfs_path *path;
827 	int ret;
828 
829 	if (!btrfs_fs_compat_ro(fs_info, FREE_SPACE_TREE))
830 		return 0;
831 
832 	path = btrfs_alloc_path();
833 	if (!path) {
834 		ret = -ENOMEM;
835 		goto out;
836 	}
837 
838 	block_group = btrfs_lookup_block_group(fs_info, start);
839 	if (!block_group) {
840 		ASSERT(0);
841 		ret = -ENOENT;
842 		goto out;
843 	}
844 
845 	mutex_lock(&block_group->free_space_lock);
846 	ret = __remove_from_free_space_tree(trans, fs_info, block_group, path,
847 					    start, size);
848 	mutex_unlock(&block_group->free_space_lock);
849 
850 	btrfs_put_block_group(block_group);
851 out:
852 	btrfs_free_path(path);
853 	if (ret)
854 		btrfs_abort_transaction(trans, fs_info->free_space_root, ret);
855 	return ret;
856 }
857 
858 static int add_free_space_extent(struct btrfs_trans_handle *trans,
859 				 struct btrfs_fs_info *fs_info,
860 				 struct btrfs_block_group_cache *block_group,
861 				 struct btrfs_path *path,
862 				 u64 start, u64 size)
863 {
864 	struct btrfs_root *root = fs_info->free_space_root;
865 	struct btrfs_key key, new_key;
866 	u64 found_start, found_end;
867 	u64 end = start + size;
868 	int new_extents = 1;
869 	int ret;
870 
871 	/*
872 	 * We are adding a new extent of free space, but we need to merge
873 	 * extents. There are four cases here:
874 	 *
875 	 * 1. The new extent does not have any immediate neighbors to merge
876 	 * with: add the new key and increment the free space extent count. We
877 	 * may need to convert the block group to bitmaps as a result.
878 	 * 2. The new extent has an immediate neighbor before it: remove the
879 	 * previous key and insert a new key combining both of them. There is no
880 	 * net change in the number of extents.
881 	 * 3. The new extent has an immediate neighbor after it: remove the next
882 	 * key and insert a new key combining both of them. There is no net
883 	 * change in the number of extents.
884 	 * 4. The new extent has immediate neighbors on both sides: remove both
885 	 * of the keys and insert a new key combining all of them. Where we used
886 	 * to have two extents, we now have one, so decrement the extent count.
887 	 */
888 
889 	new_key.objectid = start;
890 	new_key.type = BTRFS_FREE_SPACE_EXTENT_KEY;
891 	new_key.offset = size;
892 
893 	/* Search for a neighbor on the left. */
894 	if (start == block_group->key.objectid)
895 		goto right;
896 	key.objectid = start - 1;
897 	key.type = (u8)-1;
898 	key.offset = (u64)-1;
899 
900 	ret = btrfs_search_prev_slot(trans, root, &key, path, -1, 1);
901 	if (ret)
902 		goto out;
903 
904 	btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
905 
906 	if (key.type != BTRFS_FREE_SPACE_EXTENT_KEY) {
907 		ASSERT(key.type == BTRFS_FREE_SPACE_INFO_KEY);
908 		btrfs_release_path(path);
909 		goto right;
910 	}
911 
912 	found_start = key.objectid;
913 	found_end = key.objectid + key.offset;
914 	ASSERT(found_start >= block_group->key.objectid &&
915 	       found_end > block_group->key.objectid);
916 	ASSERT(found_start < start && found_end <= start);
917 
918 	/*
919 	 * Delete the neighbor on the left and absorb it into the new key (cases
920 	 * 2 and 4).
921 	 */
922 	if (found_end == start) {
923 		ret = btrfs_del_item(trans, root, path);
924 		if (ret)
925 			goto out;
926 		new_key.objectid = found_start;
927 		new_key.offset += key.offset;
928 		new_extents--;
929 	}
930 	btrfs_release_path(path);
931 
932 right:
933 	/* Search for a neighbor on the right. */
934 	if (end == block_group->key.objectid + block_group->key.offset)
935 		goto insert;
936 	key.objectid = end;
937 	key.type = (u8)-1;
938 	key.offset = (u64)-1;
939 
940 	ret = btrfs_search_prev_slot(trans, root, &key, path, -1, 1);
941 	if (ret)
942 		goto out;
943 
944 	btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
945 
946 	if (key.type != BTRFS_FREE_SPACE_EXTENT_KEY) {
947 		ASSERT(key.type == BTRFS_FREE_SPACE_INFO_KEY);
948 		btrfs_release_path(path);
949 		goto insert;
950 	}
951 
952 	found_start = key.objectid;
953 	found_end = key.objectid + key.offset;
954 	ASSERT(found_start >= block_group->key.objectid &&
955 	       found_end > block_group->key.objectid);
956 	ASSERT((found_start < start && found_end <= start) ||
957 	       (found_start >= end && found_end > end));
958 
959 	/*
960 	 * Delete the neighbor on the right and absorb it into the new key
961 	 * (cases 3 and 4).
962 	 */
963 	if (found_start == end) {
964 		ret = btrfs_del_item(trans, root, path);
965 		if (ret)
966 			goto out;
967 		new_key.offset += key.offset;
968 		new_extents--;
969 	}
970 	btrfs_release_path(path);
971 
972 insert:
973 	/* Insert the new key (cases 1-4). */
974 	ret = btrfs_insert_empty_item(trans, root, path, &new_key, 0);
975 	if (ret)
976 		goto out;
977 
978 	btrfs_release_path(path);
979 	ret = update_free_space_extent_count(trans, fs_info, block_group, path,
980 					     new_extents);
981 
982 out:
983 	return ret;
984 }
985 
986 int __add_to_free_space_tree(struct btrfs_trans_handle *trans,
987 			     struct btrfs_fs_info *fs_info,
988 			     struct btrfs_block_group_cache *block_group,
989 			     struct btrfs_path *path, u64 start, u64 size)
990 {
991 	struct btrfs_free_space_info *info;
992 	u32 flags;
993 	int ret;
994 
995 	if (block_group->needs_free_space) {
996 		ret = __add_block_group_free_space(trans, fs_info, block_group,
997 						   path);
998 		if (ret)
999 			return ret;
1000 	}
1001 
1002 	info = search_free_space_info(NULL, fs_info, block_group, path, 0);
1003 	if (IS_ERR(info))
1004 		return PTR_ERR(info);
1005 	flags = btrfs_free_space_flags(path->nodes[0], info);
1006 	btrfs_release_path(path);
1007 
1008 	if (flags & BTRFS_FREE_SPACE_USING_BITMAPS) {
1009 		return modify_free_space_bitmap(trans, fs_info, block_group,
1010 						path, start, size, 0);
1011 	} else {
1012 		return add_free_space_extent(trans, fs_info, block_group, path,
1013 					     start, size);
1014 	}
1015 }
1016 
1017 int add_to_free_space_tree(struct btrfs_trans_handle *trans,
1018 			   struct btrfs_fs_info *fs_info,
1019 			   u64 start, u64 size)
1020 {
1021 	struct btrfs_block_group_cache *block_group;
1022 	struct btrfs_path *path;
1023 	int ret;
1024 
1025 	if (!btrfs_fs_compat_ro(fs_info, FREE_SPACE_TREE))
1026 		return 0;
1027 
1028 	path = btrfs_alloc_path();
1029 	if (!path) {
1030 		ret = -ENOMEM;
1031 		goto out;
1032 	}
1033 
1034 	block_group = btrfs_lookup_block_group(fs_info, start);
1035 	if (!block_group) {
1036 		ASSERT(0);
1037 		ret = -ENOENT;
1038 		goto out;
1039 	}
1040 
1041 	mutex_lock(&block_group->free_space_lock);
1042 	ret = __add_to_free_space_tree(trans, fs_info, block_group, path, start,
1043 				       size);
1044 	mutex_unlock(&block_group->free_space_lock);
1045 
1046 	btrfs_put_block_group(block_group);
1047 out:
1048 	btrfs_free_path(path);
1049 	if (ret)
1050 		btrfs_abort_transaction(trans, fs_info->free_space_root, ret);
1051 	return ret;
1052 }
1053 
1054 /*
1055  * Populate the free space tree by walking the extent tree. Operations on the
1056  * extent tree that happen as a result of writes to the free space tree will go
1057  * through the normal add/remove hooks.
1058  */
1059 static int populate_free_space_tree(struct btrfs_trans_handle *trans,
1060 				    struct btrfs_fs_info *fs_info,
1061 				    struct btrfs_block_group_cache *block_group)
1062 {
1063 	struct btrfs_root *extent_root = fs_info->extent_root;
1064 	struct btrfs_path *path, *path2;
1065 	struct btrfs_key key;
1066 	u64 start, end;
1067 	int ret;
1068 
1069 	path = btrfs_alloc_path();
1070 	if (!path)
1071 		return -ENOMEM;
1072 	path->reada = 1;
1073 
1074 	path2 = btrfs_alloc_path();
1075 	if (!path2) {
1076 		btrfs_free_path(path);
1077 		return -ENOMEM;
1078 	}
1079 
1080 	ret = add_new_free_space_info(trans, fs_info, block_group, path2);
1081 	if (ret)
1082 		goto out;
1083 
1084 	mutex_lock(&block_group->free_space_lock);
1085 
1086 	/*
1087 	 * Iterate through all of the extent and metadata items in this block
1088 	 * group, adding the free space between them and the free space at the
1089 	 * end. Note that EXTENT_ITEM and METADATA_ITEM are less than
1090 	 * BLOCK_GROUP_ITEM, so an extent may precede the block group that it's
1091 	 * contained in.
1092 	 */
1093 	key.objectid = block_group->key.objectid;
1094 	key.type = BTRFS_EXTENT_ITEM_KEY;
1095 	key.offset = 0;
1096 
1097 	ret = btrfs_search_slot_for_read(extent_root, &key, path, 1, 0);
1098 	if (ret < 0)
1099 		goto out_locked;
1100 	ASSERT(ret == 0);
1101 
1102 	start = block_group->key.objectid;
1103 	end = block_group->key.objectid + block_group->key.offset;
1104 	while (1) {
1105 		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
1106 
1107 		if (key.type == BTRFS_EXTENT_ITEM_KEY ||
1108 		    key.type == BTRFS_METADATA_ITEM_KEY) {
1109 			if (key.objectid >= end)
1110 				break;
1111 
1112 			if (start < key.objectid) {
1113 				ret = __add_to_free_space_tree(trans, fs_info,
1114 							       block_group,
1115 							       path2, start,
1116 							       key.objectid -
1117 							       start);
1118 				if (ret)
1119 					goto out_locked;
1120 			}
1121 			start = key.objectid;
1122 			if (key.type == BTRFS_METADATA_ITEM_KEY)
1123 				start += fs_info->tree_root->nodesize;
1124 			else
1125 				start += key.offset;
1126 		} else if (key.type == BTRFS_BLOCK_GROUP_ITEM_KEY) {
1127 			if (key.objectid != block_group->key.objectid)
1128 				break;
1129 		}
1130 
1131 		ret = btrfs_next_item(extent_root, path);
1132 		if (ret < 0)
1133 			goto out_locked;
1134 		if (ret)
1135 			break;
1136 	}
1137 	if (start < end) {
1138 		ret = __add_to_free_space_tree(trans, fs_info, block_group,
1139 					       path2, start, end - start);
1140 		if (ret)
1141 			goto out_locked;
1142 	}
1143 
1144 	ret = 0;
1145 out_locked:
1146 	mutex_unlock(&block_group->free_space_lock);
1147 out:
1148 	btrfs_free_path(path2);
1149 	btrfs_free_path(path);
1150 	return ret;
1151 }
1152 
1153 int btrfs_create_free_space_tree(struct btrfs_fs_info *fs_info)
1154 {
1155 	struct btrfs_trans_handle *trans;
1156 	struct btrfs_root *tree_root = fs_info->tree_root;
1157 	struct btrfs_root *free_space_root;
1158 	struct btrfs_block_group_cache *block_group;
1159 	struct rb_node *node;
1160 	int ret;
1161 
1162 	trans = btrfs_start_transaction(tree_root, 0);
1163 	if (IS_ERR(trans))
1164 		return PTR_ERR(trans);
1165 
1166 	fs_info->creating_free_space_tree = 1;
1167 	free_space_root = btrfs_create_tree(trans, fs_info,
1168 					    BTRFS_FREE_SPACE_TREE_OBJECTID);
1169 	if (IS_ERR(free_space_root)) {
1170 		ret = PTR_ERR(free_space_root);
1171 		goto abort;
1172 	}
1173 	fs_info->free_space_root = free_space_root;
1174 
1175 	node = rb_first(&fs_info->block_group_cache_tree);
1176 	while (node) {
1177 		block_group = rb_entry(node, struct btrfs_block_group_cache,
1178 				       cache_node);
1179 		ret = populate_free_space_tree(trans, fs_info, block_group);
1180 		if (ret)
1181 			goto abort;
1182 		node = rb_next(node);
1183 	}
1184 
1185 	btrfs_set_fs_compat_ro(fs_info, FREE_SPACE_TREE);
1186 	fs_info->creating_free_space_tree = 0;
1187 
1188 	ret = btrfs_commit_transaction(trans, tree_root);
1189 	if (ret)
1190 		return ret;
1191 
1192 	return 0;
1193 
1194 abort:
1195 	fs_info->creating_free_space_tree = 0;
1196 	btrfs_abort_transaction(trans, tree_root, ret);
1197 	btrfs_end_transaction(trans, tree_root);
1198 	return ret;
1199 }
1200 
1201 static int clear_free_space_tree(struct btrfs_trans_handle *trans,
1202 				 struct btrfs_root *root)
1203 {
1204 	struct btrfs_path *path;
1205 	struct btrfs_key key;
1206 	int nr;
1207 	int ret;
1208 
1209 	path = btrfs_alloc_path();
1210 	if (!path)
1211 		return -ENOMEM;
1212 
1213 	path->leave_spinning = 1;
1214 
1215 	key.objectid = 0;
1216 	key.type = 0;
1217 	key.offset = 0;
1218 
1219 	while (1) {
1220 		ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
1221 		if (ret < 0)
1222 			goto out;
1223 
1224 		nr = btrfs_header_nritems(path->nodes[0]);
1225 		if (!nr)
1226 			break;
1227 
1228 		path->slots[0] = 0;
1229 		ret = btrfs_del_items(trans, root, path, 0, nr);
1230 		if (ret)
1231 			goto out;
1232 
1233 		btrfs_release_path(path);
1234 	}
1235 
1236 	ret = 0;
1237 out:
1238 	btrfs_free_path(path);
1239 	return ret;
1240 }
1241 
1242 int btrfs_clear_free_space_tree(struct btrfs_fs_info *fs_info)
1243 {
1244 	struct btrfs_trans_handle *trans;
1245 	struct btrfs_root *tree_root = fs_info->tree_root;
1246 	struct btrfs_root *free_space_root = fs_info->free_space_root;
1247 	int ret;
1248 
1249 	trans = btrfs_start_transaction(tree_root, 0);
1250 	if (IS_ERR(trans))
1251 		return PTR_ERR(trans);
1252 
1253 	btrfs_clear_fs_compat_ro(fs_info, FREE_SPACE_TREE);
1254 	fs_info->free_space_root = NULL;
1255 
1256 	ret = clear_free_space_tree(trans, free_space_root);
1257 	if (ret)
1258 		goto abort;
1259 
1260 	ret = btrfs_del_root(trans, tree_root, &free_space_root->root_key);
1261 	if (ret)
1262 		goto abort;
1263 
1264 	list_del(&free_space_root->dirty_list);
1265 
1266 	btrfs_tree_lock(free_space_root->node);
1267 	clean_tree_block(trans, tree_root->fs_info, free_space_root->node);
1268 	btrfs_tree_unlock(free_space_root->node);
1269 	btrfs_free_tree_block(trans, free_space_root, free_space_root->node,
1270 			      0, 1);
1271 
1272 	free_extent_buffer(free_space_root->node);
1273 	free_extent_buffer(free_space_root->commit_root);
1274 	kfree(free_space_root);
1275 
1276 	ret = btrfs_commit_transaction(trans, tree_root);
1277 	if (ret)
1278 		return ret;
1279 
1280 	return 0;
1281 
1282 abort:
1283 	btrfs_abort_transaction(trans, tree_root, ret);
1284 	btrfs_end_transaction(trans, tree_root);
1285 	return ret;
1286 }
1287 
1288 static int __add_block_group_free_space(struct btrfs_trans_handle *trans,
1289 					struct btrfs_fs_info *fs_info,
1290 					struct btrfs_block_group_cache *block_group,
1291 					struct btrfs_path *path)
1292 {
1293 	u64 start, end;
1294 	int ret;
1295 
1296 	start = block_group->key.objectid;
1297 	end = block_group->key.objectid + block_group->key.offset;
1298 
1299 	block_group->needs_free_space = 0;
1300 
1301 	ret = add_new_free_space_info(trans, fs_info, block_group, path);
1302 	if (ret)
1303 		return ret;
1304 
1305 	return __add_to_free_space_tree(trans, fs_info, block_group, path,
1306 					block_group->key.objectid,
1307 					block_group->key.offset);
1308 }
1309 
1310 int add_block_group_free_space(struct btrfs_trans_handle *trans,
1311 			       struct btrfs_fs_info *fs_info,
1312 			       struct btrfs_block_group_cache *block_group)
1313 {
1314 	struct btrfs_path *path = NULL;
1315 	int ret = 0;
1316 
1317 	if (!btrfs_fs_compat_ro(fs_info, FREE_SPACE_TREE))
1318 		return 0;
1319 
1320 	mutex_lock(&block_group->free_space_lock);
1321 	if (!block_group->needs_free_space)
1322 		goto out;
1323 
1324 	path = btrfs_alloc_path();
1325 	if (!path) {
1326 		ret = -ENOMEM;
1327 		goto out;
1328 	}
1329 
1330 	ret = __add_block_group_free_space(trans, fs_info, block_group, path);
1331 
1332 out:
1333 	btrfs_free_path(path);
1334 	mutex_unlock(&block_group->free_space_lock);
1335 	if (ret)
1336 		btrfs_abort_transaction(trans, fs_info->free_space_root, ret);
1337 	return ret;
1338 }
1339 
1340 int remove_block_group_free_space(struct btrfs_trans_handle *trans,
1341 				  struct btrfs_fs_info *fs_info,
1342 				  struct btrfs_block_group_cache *block_group)
1343 {
1344 	struct btrfs_root *root = fs_info->free_space_root;
1345 	struct btrfs_path *path;
1346 	struct btrfs_key key, found_key;
1347 	struct extent_buffer *leaf;
1348 	u64 start, end;
1349 	int done = 0, nr;
1350 	int ret;
1351 
1352 	if (!btrfs_fs_compat_ro(fs_info, FREE_SPACE_TREE))
1353 		return 0;
1354 
1355 	if (block_group->needs_free_space) {
1356 		/* We never added this block group to the free space tree. */
1357 		return 0;
1358 	}
1359 
1360 	path = btrfs_alloc_path();
1361 	if (!path) {
1362 		ret = -ENOMEM;
1363 		goto out;
1364 	}
1365 
1366 	start = block_group->key.objectid;
1367 	end = block_group->key.objectid + block_group->key.offset;
1368 
1369 	key.objectid = end - 1;
1370 	key.type = (u8)-1;
1371 	key.offset = (u64)-1;
1372 
1373 	while (!done) {
1374 		ret = btrfs_search_prev_slot(trans, root, &key, path, -1, 1);
1375 		if (ret)
1376 			goto out;
1377 
1378 		leaf = path->nodes[0];
1379 		nr = 0;
1380 		path->slots[0]++;
1381 		while (path->slots[0] > 0) {
1382 			btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0] - 1);
1383 
1384 			if (found_key.type == BTRFS_FREE_SPACE_INFO_KEY) {
1385 				ASSERT(found_key.objectid == block_group->key.objectid);
1386 				ASSERT(found_key.offset == block_group->key.offset);
1387 				done = 1;
1388 				nr++;
1389 				path->slots[0]--;
1390 				break;
1391 			} else if (found_key.type == BTRFS_FREE_SPACE_EXTENT_KEY ||
1392 				   found_key.type == BTRFS_FREE_SPACE_BITMAP_KEY) {
1393 				ASSERT(found_key.objectid >= start);
1394 				ASSERT(found_key.objectid < end);
1395 				ASSERT(found_key.objectid + found_key.offset <= end);
1396 				nr++;
1397 				path->slots[0]--;
1398 			} else {
1399 				ASSERT(0);
1400 			}
1401 		}
1402 
1403 		ret = btrfs_del_items(trans, root, path, path->slots[0], nr);
1404 		if (ret)
1405 			goto out;
1406 		btrfs_release_path(path);
1407 	}
1408 
1409 	ret = 0;
1410 out:
1411 	btrfs_free_path(path);
1412 	if (ret)
1413 		btrfs_abort_transaction(trans, root, ret);
1414 	return ret;
1415 }
1416 
1417 static int load_free_space_bitmaps(struct btrfs_caching_control *caching_ctl,
1418 				   struct btrfs_path *path,
1419 				   u32 expected_extent_count)
1420 {
1421 	struct btrfs_block_group_cache *block_group;
1422 	struct btrfs_fs_info *fs_info;
1423 	struct btrfs_root *root;
1424 	struct btrfs_key key;
1425 	int prev_bit = 0, bit;
1426 	/* Initialize to silence GCC. */
1427 	u64 extent_start = 0;
1428 	u64 end, offset;
1429 	u64 total_found = 0;
1430 	u32 extent_count = 0;
1431 	int ret;
1432 
1433 	block_group = caching_ctl->block_group;
1434 	fs_info = block_group->fs_info;
1435 	root = fs_info->free_space_root;
1436 
1437 	end = block_group->key.objectid + block_group->key.offset;
1438 
1439 	while (1) {
1440 		ret = btrfs_next_item(root, path);
1441 		if (ret < 0)
1442 			goto out;
1443 		if (ret)
1444 			break;
1445 
1446 		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
1447 
1448 		if (key.type == BTRFS_FREE_SPACE_INFO_KEY)
1449 			break;
1450 
1451 		ASSERT(key.type == BTRFS_FREE_SPACE_BITMAP_KEY);
1452 		ASSERT(key.objectid < end && key.objectid + key.offset <= end);
1453 
1454 		caching_ctl->progress = key.objectid;
1455 
1456 		offset = key.objectid;
1457 		while (offset < key.objectid + key.offset) {
1458 			bit = free_space_test_bit(block_group, path, offset);
1459 			if (prev_bit == 0 && bit == 1) {
1460 				extent_start = offset;
1461 			} else if (prev_bit == 1 && bit == 0) {
1462 				total_found += add_new_free_space(block_group,
1463 								  fs_info,
1464 								  extent_start,
1465 								  offset);
1466 				if (total_found > CACHING_CTL_WAKE_UP) {
1467 					total_found = 0;
1468 					wake_up(&caching_ctl->wait);
1469 				}
1470 				extent_count++;
1471 			}
1472 			prev_bit = bit;
1473 			offset += block_group->sectorsize;
1474 		}
1475 	}
1476 	if (prev_bit == 1) {
1477 		total_found += add_new_free_space(block_group, fs_info,
1478 						  extent_start, end);
1479 		extent_count++;
1480 	}
1481 
1482 	if (extent_count != expected_extent_count) {
1483 		btrfs_err(fs_info, "incorrect extent count for %llu; counted %u, expected %u",
1484 			  block_group->key.objectid, extent_count,
1485 			  expected_extent_count);
1486 		ASSERT(0);
1487 		ret = -EIO;
1488 		goto out;
1489 	}
1490 
1491 	caching_ctl->progress = (u64)-1;
1492 
1493 	ret = 0;
1494 out:
1495 	return ret;
1496 }
1497 
1498 static int load_free_space_extents(struct btrfs_caching_control *caching_ctl,
1499 				   struct btrfs_path *path,
1500 				   u32 expected_extent_count)
1501 {
1502 	struct btrfs_block_group_cache *block_group;
1503 	struct btrfs_fs_info *fs_info;
1504 	struct btrfs_root *root;
1505 	struct btrfs_key key;
1506 	u64 end;
1507 	u64 total_found = 0;
1508 	u32 extent_count = 0;
1509 	int ret;
1510 
1511 	block_group = caching_ctl->block_group;
1512 	fs_info = block_group->fs_info;
1513 	root = fs_info->free_space_root;
1514 
1515 	end = block_group->key.objectid + block_group->key.offset;
1516 
1517 	while (1) {
1518 		ret = btrfs_next_item(root, path);
1519 		if (ret < 0)
1520 			goto out;
1521 		if (ret)
1522 			break;
1523 
1524 		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
1525 
1526 		if (key.type == BTRFS_FREE_SPACE_INFO_KEY)
1527 			break;
1528 
1529 		ASSERT(key.type == BTRFS_FREE_SPACE_EXTENT_KEY);
1530 		ASSERT(key.objectid < end && key.objectid + key.offset <= end);
1531 
1532 		caching_ctl->progress = key.objectid;
1533 
1534 		total_found += add_new_free_space(block_group, fs_info,
1535 						  key.objectid,
1536 						  key.objectid + key.offset);
1537 		if (total_found > CACHING_CTL_WAKE_UP) {
1538 			total_found = 0;
1539 			wake_up(&caching_ctl->wait);
1540 		}
1541 		extent_count++;
1542 	}
1543 
1544 	if (extent_count != expected_extent_count) {
1545 		btrfs_err(fs_info, "incorrect extent count for %llu; counted %u, expected %u",
1546 			  block_group->key.objectid, extent_count,
1547 			  expected_extent_count);
1548 		ASSERT(0);
1549 		ret = -EIO;
1550 		goto out;
1551 	}
1552 
1553 	caching_ctl->progress = (u64)-1;
1554 
1555 	ret = 0;
1556 out:
1557 	return ret;
1558 }
1559 
1560 int load_free_space_tree(struct btrfs_caching_control *caching_ctl)
1561 {
1562 	struct btrfs_block_group_cache *block_group;
1563 	struct btrfs_fs_info *fs_info;
1564 	struct btrfs_free_space_info *info;
1565 	struct btrfs_path *path;
1566 	u32 extent_count, flags;
1567 	int ret;
1568 
1569 	block_group = caching_ctl->block_group;
1570 	fs_info = block_group->fs_info;
1571 
1572 	path = btrfs_alloc_path();
1573 	if (!path)
1574 		return -ENOMEM;
1575 
1576 	/*
1577 	 * Just like caching_thread() doesn't want to deadlock on the extent
1578 	 * tree, we don't want to deadlock on the free space tree.
1579 	 */
1580 	path->skip_locking = 1;
1581 	path->search_commit_root = 1;
1582 	path->reada = 1;
1583 
1584 	info = search_free_space_info(NULL, fs_info, block_group, path, 0);
1585 	if (IS_ERR(info)) {
1586 		ret = PTR_ERR(info);
1587 		goto out;
1588 	}
1589 	extent_count = btrfs_free_space_extent_count(path->nodes[0], info);
1590 	flags = btrfs_free_space_flags(path->nodes[0], info);
1591 
1592 	/*
1593 	 * We left path pointing to the free space info item, so now
1594 	 * load_free_space_foo can just iterate through the free space tree from
1595 	 * there.
1596 	 */
1597 	if (flags & BTRFS_FREE_SPACE_USING_BITMAPS)
1598 		ret = load_free_space_bitmaps(caching_ctl, path, extent_count);
1599 	else
1600 		ret = load_free_space_extents(caching_ctl, path, extent_count);
1601 
1602 out:
1603 	btrfs_free_path(path);
1604 	return ret;
1605 }
1606