xref: /openbmc/linux/fs/btrfs/free-space-tree.c (revision 4b7ead03)
1 // SPDX-License-Identifier: GPL-2.0
2 /*
3  * Copyright (C) 2015 Facebook.  All rights reserved.
4  */
5 
6 #include <linux/kernel.h>
7 #include <linux/sched/mm.h>
8 #include "ctree.h"
9 #include "disk-io.h"
10 #include "locking.h"
11 #include "free-space-tree.h"
12 #include "transaction.h"
13 #include "block-group.h"
14 
15 static int __add_block_group_free_space(struct btrfs_trans_handle *trans,
16 					struct btrfs_block_group *block_group,
17 					struct btrfs_path *path);
18 
19 void set_free_space_tree_thresholds(struct btrfs_block_group *cache)
20 {
21 	u32 bitmap_range;
22 	size_t bitmap_size;
23 	u64 num_bitmaps, total_bitmap_size;
24 
25 	if (WARN_ON(cache->length == 0))
26 		btrfs_warn(cache->fs_info, "block group %llu length is zero",
27 			   cache->start);
28 
29 	/*
30 	 * We convert to bitmaps when the disk space required for using extents
31 	 * exceeds that required for using bitmaps.
32 	 */
33 	bitmap_range = cache->fs_info->sectorsize * BTRFS_FREE_SPACE_BITMAP_BITS;
34 	num_bitmaps = div_u64(cache->length + bitmap_range - 1, bitmap_range);
35 	bitmap_size = sizeof(struct btrfs_item) + BTRFS_FREE_SPACE_BITMAP_SIZE;
36 	total_bitmap_size = num_bitmaps * bitmap_size;
37 	cache->bitmap_high_thresh = div_u64(total_bitmap_size,
38 					    sizeof(struct btrfs_item));
39 
40 	/*
41 	 * We allow for a small buffer between the high threshold and low
42 	 * threshold to avoid thrashing back and forth between the two formats.
43 	 */
44 	if (cache->bitmap_high_thresh > 100)
45 		cache->bitmap_low_thresh = cache->bitmap_high_thresh - 100;
46 	else
47 		cache->bitmap_low_thresh = 0;
48 }
49 
50 static int add_new_free_space_info(struct btrfs_trans_handle *trans,
51 				   struct btrfs_block_group *block_group,
52 				   struct btrfs_path *path)
53 {
54 	struct btrfs_root *root = trans->fs_info->free_space_root;
55 	struct btrfs_free_space_info *info;
56 	struct btrfs_key key;
57 	struct extent_buffer *leaf;
58 	int ret;
59 
60 	key.objectid = block_group->start;
61 	key.type = BTRFS_FREE_SPACE_INFO_KEY;
62 	key.offset = block_group->length;
63 
64 	ret = btrfs_insert_empty_item(trans, root, path, &key, sizeof(*info));
65 	if (ret)
66 		goto out;
67 
68 	leaf = path->nodes[0];
69 	info = btrfs_item_ptr(leaf, path->slots[0],
70 			      struct btrfs_free_space_info);
71 	btrfs_set_free_space_extent_count(leaf, info, 0);
72 	btrfs_set_free_space_flags(leaf, info, 0);
73 	btrfs_mark_buffer_dirty(leaf);
74 
75 	ret = 0;
76 out:
77 	btrfs_release_path(path);
78 	return ret;
79 }
80 
81 EXPORT_FOR_TESTS
82 struct btrfs_free_space_info *search_free_space_info(
83 		struct btrfs_trans_handle *trans,
84 		struct btrfs_block_group *block_group,
85 		struct btrfs_path *path, int cow)
86 {
87 	struct btrfs_fs_info *fs_info = block_group->fs_info;
88 	struct btrfs_root *root = fs_info->free_space_root;
89 	struct btrfs_key key;
90 	int ret;
91 
92 	key.objectid = block_group->start;
93 	key.type = BTRFS_FREE_SPACE_INFO_KEY;
94 	key.offset = block_group->length;
95 
96 	ret = btrfs_search_slot(trans, root, &key, path, 0, cow);
97 	if (ret < 0)
98 		return ERR_PTR(ret);
99 	if (ret != 0) {
100 		btrfs_warn(fs_info, "missing free space info for %llu",
101 			   block_group->start);
102 		ASSERT(0);
103 		return ERR_PTR(-ENOENT);
104 	}
105 
106 	return btrfs_item_ptr(path->nodes[0], path->slots[0],
107 			      struct btrfs_free_space_info);
108 }
109 
110 /*
111  * btrfs_search_slot() but we're looking for the greatest key less than the
112  * passed key.
113  */
114 static int btrfs_search_prev_slot(struct btrfs_trans_handle *trans,
115 				  struct btrfs_root *root,
116 				  struct btrfs_key *key, struct btrfs_path *p,
117 				  int ins_len, int cow)
118 {
119 	int ret;
120 
121 	ret = btrfs_search_slot(trans, root, key, p, ins_len, cow);
122 	if (ret < 0)
123 		return ret;
124 
125 	if (ret == 0) {
126 		ASSERT(0);
127 		return -EIO;
128 	}
129 
130 	if (p->slots[0] == 0) {
131 		ASSERT(0);
132 		return -EIO;
133 	}
134 	p->slots[0]--;
135 
136 	return 0;
137 }
138 
139 static inline u32 free_space_bitmap_size(const struct btrfs_fs_info *fs_info,
140 					 u64 size)
141 {
142 	return DIV_ROUND_UP(size >> fs_info->sectorsize_bits, BITS_PER_BYTE);
143 }
144 
145 static unsigned long *alloc_bitmap(u32 bitmap_size)
146 {
147 	unsigned long *ret;
148 	unsigned int nofs_flag;
149 	u32 bitmap_rounded_size = round_up(bitmap_size, sizeof(unsigned long));
150 
151 	/*
152 	 * GFP_NOFS doesn't work with kvmalloc(), but we really can't recurse
153 	 * into the filesystem as the free space bitmap can be modified in the
154 	 * critical section of a transaction commit.
155 	 *
156 	 * TODO: push the memalloc_nofs_{save,restore}() to the caller where we
157 	 * know that recursion is unsafe.
158 	 */
159 	nofs_flag = memalloc_nofs_save();
160 	ret = kvzalloc(bitmap_rounded_size, GFP_KERNEL);
161 	memalloc_nofs_restore(nofs_flag);
162 	return ret;
163 }
164 
165 static void le_bitmap_set(unsigned long *map, unsigned int start, int len)
166 {
167 	u8 *p = ((u8 *)map) + BIT_BYTE(start);
168 	const unsigned int size = start + len;
169 	int bits_to_set = BITS_PER_BYTE - (start % BITS_PER_BYTE);
170 	u8 mask_to_set = BITMAP_FIRST_BYTE_MASK(start);
171 
172 	while (len - bits_to_set >= 0) {
173 		*p |= mask_to_set;
174 		len -= bits_to_set;
175 		bits_to_set = BITS_PER_BYTE;
176 		mask_to_set = ~0;
177 		p++;
178 	}
179 	if (len) {
180 		mask_to_set &= BITMAP_LAST_BYTE_MASK(size);
181 		*p |= mask_to_set;
182 	}
183 }
184 
185 EXPORT_FOR_TESTS
186 int convert_free_space_to_bitmaps(struct btrfs_trans_handle *trans,
187 				  struct btrfs_block_group *block_group,
188 				  struct btrfs_path *path)
189 {
190 	struct btrfs_fs_info *fs_info = trans->fs_info;
191 	struct btrfs_root *root = fs_info->free_space_root;
192 	struct btrfs_free_space_info *info;
193 	struct btrfs_key key, found_key;
194 	struct extent_buffer *leaf;
195 	unsigned long *bitmap;
196 	char *bitmap_cursor;
197 	u64 start, end;
198 	u64 bitmap_range, i;
199 	u32 bitmap_size, flags, expected_extent_count;
200 	u32 extent_count = 0;
201 	int done = 0, nr;
202 	int ret;
203 
204 	bitmap_size = free_space_bitmap_size(fs_info, block_group->length);
205 	bitmap = alloc_bitmap(bitmap_size);
206 	if (!bitmap) {
207 		ret = -ENOMEM;
208 		goto out;
209 	}
210 
211 	start = block_group->start;
212 	end = block_group->start + block_group->length;
213 
214 	key.objectid = end - 1;
215 	key.type = (u8)-1;
216 	key.offset = (u64)-1;
217 
218 	while (!done) {
219 		ret = btrfs_search_prev_slot(trans, root, &key, path, -1, 1);
220 		if (ret)
221 			goto out;
222 
223 		leaf = path->nodes[0];
224 		nr = 0;
225 		path->slots[0]++;
226 		while (path->slots[0] > 0) {
227 			btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0] - 1);
228 
229 			if (found_key.type == BTRFS_FREE_SPACE_INFO_KEY) {
230 				ASSERT(found_key.objectid == block_group->start);
231 				ASSERT(found_key.offset == block_group->length);
232 				done = 1;
233 				break;
234 			} else if (found_key.type == BTRFS_FREE_SPACE_EXTENT_KEY) {
235 				u64 first, last;
236 
237 				ASSERT(found_key.objectid >= start);
238 				ASSERT(found_key.objectid < end);
239 				ASSERT(found_key.objectid + found_key.offset <= end);
240 
241 				first = div_u64(found_key.objectid - start,
242 						fs_info->sectorsize);
243 				last = div_u64(found_key.objectid + found_key.offset - start,
244 					       fs_info->sectorsize);
245 				le_bitmap_set(bitmap, first, last - first);
246 
247 				extent_count++;
248 				nr++;
249 				path->slots[0]--;
250 			} else {
251 				ASSERT(0);
252 			}
253 		}
254 
255 		ret = btrfs_del_items(trans, root, path, path->slots[0], nr);
256 		if (ret)
257 			goto out;
258 		btrfs_release_path(path);
259 	}
260 
261 	info = search_free_space_info(trans, block_group, path, 1);
262 	if (IS_ERR(info)) {
263 		ret = PTR_ERR(info);
264 		goto out;
265 	}
266 	leaf = path->nodes[0];
267 	flags = btrfs_free_space_flags(leaf, info);
268 	flags |= BTRFS_FREE_SPACE_USING_BITMAPS;
269 	btrfs_set_free_space_flags(leaf, info, flags);
270 	expected_extent_count = btrfs_free_space_extent_count(leaf, info);
271 	btrfs_mark_buffer_dirty(leaf);
272 	btrfs_release_path(path);
273 
274 	if (extent_count != expected_extent_count) {
275 		btrfs_err(fs_info,
276 			  "incorrect extent count for %llu; counted %u, expected %u",
277 			  block_group->start, extent_count,
278 			  expected_extent_count);
279 		ASSERT(0);
280 		ret = -EIO;
281 		goto out;
282 	}
283 
284 	bitmap_cursor = (char *)bitmap;
285 	bitmap_range = fs_info->sectorsize * BTRFS_FREE_SPACE_BITMAP_BITS;
286 	i = start;
287 	while (i < end) {
288 		unsigned long ptr;
289 		u64 extent_size;
290 		u32 data_size;
291 
292 		extent_size = min(end - i, bitmap_range);
293 		data_size = free_space_bitmap_size(fs_info, extent_size);
294 
295 		key.objectid = i;
296 		key.type = BTRFS_FREE_SPACE_BITMAP_KEY;
297 		key.offset = extent_size;
298 
299 		ret = btrfs_insert_empty_item(trans, root, path, &key,
300 					      data_size);
301 		if (ret)
302 			goto out;
303 
304 		leaf = path->nodes[0];
305 		ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
306 		write_extent_buffer(leaf, bitmap_cursor, ptr,
307 				    data_size);
308 		btrfs_mark_buffer_dirty(leaf);
309 		btrfs_release_path(path);
310 
311 		i += extent_size;
312 		bitmap_cursor += data_size;
313 	}
314 
315 	ret = 0;
316 out:
317 	kvfree(bitmap);
318 	if (ret)
319 		btrfs_abort_transaction(trans, ret);
320 	return ret;
321 }
322 
323 EXPORT_FOR_TESTS
324 int convert_free_space_to_extents(struct btrfs_trans_handle *trans,
325 				  struct btrfs_block_group *block_group,
326 				  struct btrfs_path *path)
327 {
328 	struct btrfs_fs_info *fs_info = trans->fs_info;
329 	struct btrfs_root *root = fs_info->free_space_root;
330 	struct btrfs_free_space_info *info;
331 	struct btrfs_key key, found_key;
332 	struct extent_buffer *leaf;
333 	unsigned long *bitmap;
334 	u64 start, end;
335 	u32 bitmap_size, flags, expected_extent_count;
336 	unsigned long nrbits, start_bit, end_bit;
337 	u32 extent_count = 0;
338 	int done = 0, nr;
339 	int ret;
340 
341 	bitmap_size = free_space_bitmap_size(fs_info, block_group->length);
342 	bitmap = alloc_bitmap(bitmap_size);
343 	if (!bitmap) {
344 		ret = -ENOMEM;
345 		goto out;
346 	}
347 
348 	start = block_group->start;
349 	end = block_group->start + block_group->length;
350 
351 	key.objectid = end - 1;
352 	key.type = (u8)-1;
353 	key.offset = (u64)-1;
354 
355 	while (!done) {
356 		ret = btrfs_search_prev_slot(trans, root, &key, path, -1, 1);
357 		if (ret)
358 			goto out;
359 
360 		leaf = path->nodes[0];
361 		nr = 0;
362 		path->slots[0]++;
363 		while (path->slots[0] > 0) {
364 			btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0] - 1);
365 
366 			if (found_key.type == BTRFS_FREE_SPACE_INFO_KEY) {
367 				ASSERT(found_key.objectid == block_group->start);
368 				ASSERT(found_key.offset == block_group->length);
369 				done = 1;
370 				break;
371 			} else if (found_key.type == BTRFS_FREE_SPACE_BITMAP_KEY) {
372 				unsigned long ptr;
373 				char *bitmap_cursor;
374 				u32 bitmap_pos, data_size;
375 
376 				ASSERT(found_key.objectid >= start);
377 				ASSERT(found_key.objectid < end);
378 				ASSERT(found_key.objectid + found_key.offset <= end);
379 
380 				bitmap_pos = div_u64(found_key.objectid - start,
381 						     fs_info->sectorsize *
382 						     BITS_PER_BYTE);
383 				bitmap_cursor = ((char *)bitmap) + bitmap_pos;
384 				data_size = free_space_bitmap_size(fs_info,
385 								found_key.offset);
386 
387 				ptr = btrfs_item_ptr_offset(leaf, path->slots[0] - 1);
388 				read_extent_buffer(leaf, bitmap_cursor, ptr,
389 						   data_size);
390 
391 				nr++;
392 				path->slots[0]--;
393 			} else {
394 				ASSERT(0);
395 			}
396 		}
397 
398 		ret = btrfs_del_items(trans, root, path, path->slots[0], nr);
399 		if (ret)
400 			goto out;
401 		btrfs_release_path(path);
402 	}
403 
404 	info = search_free_space_info(trans, block_group, path, 1);
405 	if (IS_ERR(info)) {
406 		ret = PTR_ERR(info);
407 		goto out;
408 	}
409 	leaf = path->nodes[0];
410 	flags = btrfs_free_space_flags(leaf, info);
411 	flags &= ~BTRFS_FREE_SPACE_USING_BITMAPS;
412 	btrfs_set_free_space_flags(leaf, info, flags);
413 	expected_extent_count = btrfs_free_space_extent_count(leaf, info);
414 	btrfs_mark_buffer_dirty(leaf);
415 	btrfs_release_path(path);
416 
417 	nrbits = block_group->length >> block_group->fs_info->sectorsize_bits;
418 	start_bit = find_next_bit_le(bitmap, nrbits, 0);
419 
420 	while (start_bit < nrbits) {
421 		end_bit = find_next_zero_bit_le(bitmap, nrbits, start_bit);
422 		ASSERT(start_bit < end_bit);
423 
424 		key.objectid = start + start_bit * block_group->fs_info->sectorsize;
425 		key.type = BTRFS_FREE_SPACE_EXTENT_KEY;
426 		key.offset = (end_bit - start_bit) * block_group->fs_info->sectorsize;
427 
428 		ret = btrfs_insert_empty_item(trans, root, path, &key, 0);
429 		if (ret)
430 			goto out;
431 		btrfs_release_path(path);
432 
433 		extent_count++;
434 
435 		start_bit = find_next_bit_le(bitmap, nrbits, end_bit);
436 	}
437 
438 	if (extent_count != expected_extent_count) {
439 		btrfs_err(fs_info,
440 			  "incorrect extent count for %llu; counted %u, expected %u",
441 			  block_group->start, extent_count,
442 			  expected_extent_count);
443 		ASSERT(0);
444 		ret = -EIO;
445 		goto out;
446 	}
447 
448 	ret = 0;
449 out:
450 	kvfree(bitmap);
451 	if (ret)
452 		btrfs_abort_transaction(trans, ret);
453 	return ret;
454 }
455 
456 static int update_free_space_extent_count(struct btrfs_trans_handle *trans,
457 					  struct btrfs_block_group *block_group,
458 					  struct btrfs_path *path,
459 					  int new_extents)
460 {
461 	struct btrfs_free_space_info *info;
462 	u32 flags;
463 	u32 extent_count;
464 	int ret = 0;
465 
466 	if (new_extents == 0)
467 		return 0;
468 
469 	info = search_free_space_info(trans, block_group, path, 1);
470 	if (IS_ERR(info)) {
471 		ret = PTR_ERR(info);
472 		goto out;
473 	}
474 	flags = btrfs_free_space_flags(path->nodes[0], info);
475 	extent_count = btrfs_free_space_extent_count(path->nodes[0], info);
476 
477 	extent_count += new_extents;
478 	btrfs_set_free_space_extent_count(path->nodes[0], info, extent_count);
479 	btrfs_mark_buffer_dirty(path->nodes[0]);
480 	btrfs_release_path(path);
481 
482 	if (!(flags & BTRFS_FREE_SPACE_USING_BITMAPS) &&
483 	    extent_count > block_group->bitmap_high_thresh) {
484 		ret = convert_free_space_to_bitmaps(trans, block_group, path);
485 	} else if ((flags & BTRFS_FREE_SPACE_USING_BITMAPS) &&
486 		   extent_count < block_group->bitmap_low_thresh) {
487 		ret = convert_free_space_to_extents(trans, block_group, path);
488 	}
489 
490 out:
491 	return ret;
492 }
493 
494 EXPORT_FOR_TESTS
495 int free_space_test_bit(struct btrfs_block_group *block_group,
496 			struct btrfs_path *path, u64 offset)
497 {
498 	struct extent_buffer *leaf;
499 	struct btrfs_key key;
500 	u64 found_start, found_end;
501 	unsigned long ptr, i;
502 
503 	leaf = path->nodes[0];
504 	btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
505 	ASSERT(key.type == BTRFS_FREE_SPACE_BITMAP_KEY);
506 
507 	found_start = key.objectid;
508 	found_end = key.objectid + key.offset;
509 	ASSERT(offset >= found_start && offset < found_end);
510 
511 	ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
512 	i = div_u64(offset - found_start,
513 		    block_group->fs_info->sectorsize);
514 	return !!extent_buffer_test_bit(leaf, ptr, i);
515 }
516 
517 static void free_space_set_bits(struct btrfs_block_group *block_group,
518 				struct btrfs_path *path, u64 *start, u64 *size,
519 				int bit)
520 {
521 	struct btrfs_fs_info *fs_info = block_group->fs_info;
522 	struct extent_buffer *leaf;
523 	struct btrfs_key key;
524 	u64 end = *start + *size;
525 	u64 found_start, found_end;
526 	unsigned long ptr, first, last;
527 
528 	leaf = path->nodes[0];
529 	btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
530 	ASSERT(key.type == BTRFS_FREE_SPACE_BITMAP_KEY);
531 
532 	found_start = key.objectid;
533 	found_end = key.objectid + key.offset;
534 	ASSERT(*start >= found_start && *start < found_end);
535 	ASSERT(end > found_start);
536 
537 	if (end > found_end)
538 		end = found_end;
539 
540 	ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
541 	first = (*start - found_start) >> fs_info->sectorsize_bits;
542 	last = (end - found_start) >> fs_info->sectorsize_bits;
543 	if (bit)
544 		extent_buffer_bitmap_set(leaf, ptr, first, last - first);
545 	else
546 		extent_buffer_bitmap_clear(leaf, ptr, first, last - first);
547 	btrfs_mark_buffer_dirty(leaf);
548 
549 	*size -= end - *start;
550 	*start = end;
551 }
552 
553 /*
554  * We can't use btrfs_next_item() in modify_free_space_bitmap() because
555  * btrfs_next_leaf() doesn't get the path for writing. We can forgo the fancy
556  * tree walking in btrfs_next_leaf() anyways because we know exactly what we're
557  * looking for.
558  */
559 static int free_space_next_bitmap(struct btrfs_trans_handle *trans,
560 				  struct btrfs_root *root, struct btrfs_path *p)
561 {
562 	struct btrfs_key key;
563 
564 	if (p->slots[0] + 1 < btrfs_header_nritems(p->nodes[0])) {
565 		p->slots[0]++;
566 		return 0;
567 	}
568 
569 	btrfs_item_key_to_cpu(p->nodes[0], &key, p->slots[0]);
570 	btrfs_release_path(p);
571 
572 	key.objectid += key.offset;
573 	key.type = (u8)-1;
574 	key.offset = (u64)-1;
575 
576 	return btrfs_search_prev_slot(trans, root, &key, p, 0, 1);
577 }
578 
579 /*
580  * If remove is 1, then we are removing free space, thus clearing bits in the
581  * bitmap. If remove is 0, then we are adding free space, thus setting bits in
582  * the bitmap.
583  */
584 static int modify_free_space_bitmap(struct btrfs_trans_handle *trans,
585 				    struct btrfs_block_group *block_group,
586 				    struct btrfs_path *path,
587 				    u64 start, u64 size, int remove)
588 {
589 	struct btrfs_root *root = block_group->fs_info->free_space_root;
590 	struct btrfs_key key;
591 	u64 end = start + size;
592 	u64 cur_start, cur_size;
593 	int prev_bit, next_bit;
594 	int new_extents;
595 	int ret;
596 
597 	/*
598 	 * Read the bit for the block immediately before the extent of space if
599 	 * that block is within the block group.
600 	 */
601 	if (start > block_group->start) {
602 		u64 prev_block = start - block_group->fs_info->sectorsize;
603 
604 		key.objectid = prev_block;
605 		key.type = (u8)-1;
606 		key.offset = (u64)-1;
607 
608 		ret = btrfs_search_prev_slot(trans, root, &key, path, 0, 1);
609 		if (ret)
610 			goto out;
611 
612 		prev_bit = free_space_test_bit(block_group, path, prev_block);
613 
614 		/* The previous block may have been in the previous bitmap. */
615 		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
616 		if (start >= key.objectid + key.offset) {
617 			ret = free_space_next_bitmap(trans, root, path);
618 			if (ret)
619 				goto out;
620 		}
621 	} else {
622 		key.objectid = start;
623 		key.type = (u8)-1;
624 		key.offset = (u64)-1;
625 
626 		ret = btrfs_search_prev_slot(trans, root, &key, path, 0, 1);
627 		if (ret)
628 			goto out;
629 
630 		prev_bit = -1;
631 	}
632 
633 	/*
634 	 * Iterate over all of the bitmaps overlapped by the extent of space,
635 	 * clearing/setting bits as required.
636 	 */
637 	cur_start = start;
638 	cur_size = size;
639 	while (1) {
640 		free_space_set_bits(block_group, path, &cur_start, &cur_size,
641 				    !remove);
642 		if (cur_size == 0)
643 			break;
644 		ret = free_space_next_bitmap(trans, root, path);
645 		if (ret)
646 			goto out;
647 	}
648 
649 	/*
650 	 * Read the bit for the block immediately after the extent of space if
651 	 * that block is within the block group.
652 	 */
653 	if (end < block_group->start + block_group->length) {
654 		/* The next block may be in the next bitmap. */
655 		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
656 		if (end >= key.objectid + key.offset) {
657 			ret = free_space_next_bitmap(trans, root, path);
658 			if (ret)
659 				goto out;
660 		}
661 
662 		next_bit = free_space_test_bit(block_group, path, end);
663 	} else {
664 		next_bit = -1;
665 	}
666 
667 	if (remove) {
668 		new_extents = -1;
669 		if (prev_bit == 1) {
670 			/* Leftover on the left. */
671 			new_extents++;
672 		}
673 		if (next_bit == 1) {
674 			/* Leftover on the right. */
675 			new_extents++;
676 		}
677 	} else {
678 		new_extents = 1;
679 		if (prev_bit == 1) {
680 			/* Merging with neighbor on the left. */
681 			new_extents--;
682 		}
683 		if (next_bit == 1) {
684 			/* Merging with neighbor on the right. */
685 			new_extents--;
686 		}
687 	}
688 
689 	btrfs_release_path(path);
690 	ret = update_free_space_extent_count(trans, block_group, path,
691 					     new_extents);
692 
693 out:
694 	return ret;
695 }
696 
697 static int remove_free_space_extent(struct btrfs_trans_handle *trans,
698 				    struct btrfs_block_group *block_group,
699 				    struct btrfs_path *path,
700 				    u64 start, u64 size)
701 {
702 	struct btrfs_root *root = trans->fs_info->free_space_root;
703 	struct btrfs_key key;
704 	u64 found_start, found_end;
705 	u64 end = start + size;
706 	int new_extents = -1;
707 	int ret;
708 
709 	key.objectid = start;
710 	key.type = (u8)-1;
711 	key.offset = (u64)-1;
712 
713 	ret = btrfs_search_prev_slot(trans, root, &key, path, -1, 1);
714 	if (ret)
715 		goto out;
716 
717 	btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
718 
719 	ASSERT(key.type == BTRFS_FREE_SPACE_EXTENT_KEY);
720 
721 	found_start = key.objectid;
722 	found_end = key.objectid + key.offset;
723 	ASSERT(start >= found_start && end <= found_end);
724 
725 	/*
726 	 * Okay, now that we've found the free space extent which contains the
727 	 * free space that we are removing, there are four cases:
728 	 *
729 	 * 1. We're using the whole extent: delete the key we found and
730 	 * decrement the free space extent count.
731 	 * 2. We are using part of the extent starting at the beginning: delete
732 	 * the key we found and insert a new key representing the leftover at
733 	 * the end. There is no net change in the number of extents.
734 	 * 3. We are using part of the extent ending at the end: delete the key
735 	 * we found and insert a new key representing the leftover at the
736 	 * beginning. There is no net change in the number of extents.
737 	 * 4. We are using part of the extent in the middle: delete the key we
738 	 * found and insert two new keys representing the leftovers on each
739 	 * side. Where we used to have one extent, we now have two, so increment
740 	 * the extent count. We may need to convert the block group to bitmaps
741 	 * as a result.
742 	 */
743 
744 	/* Delete the existing key (cases 1-4). */
745 	ret = btrfs_del_item(trans, root, path);
746 	if (ret)
747 		goto out;
748 
749 	/* Add a key for leftovers at the beginning (cases 3 and 4). */
750 	if (start > found_start) {
751 		key.objectid = found_start;
752 		key.type = BTRFS_FREE_SPACE_EXTENT_KEY;
753 		key.offset = start - found_start;
754 
755 		btrfs_release_path(path);
756 		ret = btrfs_insert_empty_item(trans, root, path, &key, 0);
757 		if (ret)
758 			goto out;
759 		new_extents++;
760 	}
761 
762 	/* Add a key for leftovers at the end (cases 2 and 4). */
763 	if (end < found_end) {
764 		key.objectid = end;
765 		key.type = BTRFS_FREE_SPACE_EXTENT_KEY;
766 		key.offset = found_end - end;
767 
768 		btrfs_release_path(path);
769 		ret = btrfs_insert_empty_item(trans, root, path, &key, 0);
770 		if (ret)
771 			goto out;
772 		new_extents++;
773 	}
774 
775 	btrfs_release_path(path);
776 	ret = update_free_space_extent_count(trans, block_group, path,
777 					     new_extents);
778 
779 out:
780 	return ret;
781 }
782 
783 EXPORT_FOR_TESTS
784 int __remove_from_free_space_tree(struct btrfs_trans_handle *trans,
785 				  struct btrfs_block_group *block_group,
786 				  struct btrfs_path *path, u64 start, u64 size)
787 {
788 	struct btrfs_free_space_info *info;
789 	u32 flags;
790 	int ret;
791 
792 	if (block_group->needs_free_space) {
793 		ret = __add_block_group_free_space(trans, block_group, path);
794 		if (ret)
795 			return ret;
796 	}
797 
798 	info = search_free_space_info(NULL, block_group, path, 0);
799 	if (IS_ERR(info))
800 		return PTR_ERR(info);
801 	flags = btrfs_free_space_flags(path->nodes[0], info);
802 	btrfs_release_path(path);
803 
804 	if (flags & BTRFS_FREE_SPACE_USING_BITMAPS) {
805 		return modify_free_space_bitmap(trans, block_group, path,
806 						start, size, 1);
807 	} else {
808 		return remove_free_space_extent(trans, block_group, path,
809 						start, size);
810 	}
811 }
812 
813 int remove_from_free_space_tree(struct btrfs_trans_handle *trans,
814 				u64 start, u64 size)
815 {
816 	struct btrfs_block_group *block_group;
817 	struct btrfs_path *path;
818 	int ret;
819 
820 	if (!btrfs_fs_compat_ro(trans->fs_info, FREE_SPACE_TREE))
821 		return 0;
822 
823 	path = btrfs_alloc_path();
824 	if (!path) {
825 		ret = -ENOMEM;
826 		goto out;
827 	}
828 
829 	block_group = btrfs_lookup_block_group(trans->fs_info, start);
830 	if (!block_group) {
831 		ASSERT(0);
832 		ret = -ENOENT;
833 		goto out;
834 	}
835 
836 	mutex_lock(&block_group->free_space_lock);
837 	ret = __remove_from_free_space_tree(trans, block_group, path, start,
838 					    size);
839 	mutex_unlock(&block_group->free_space_lock);
840 
841 	btrfs_put_block_group(block_group);
842 out:
843 	btrfs_free_path(path);
844 	if (ret)
845 		btrfs_abort_transaction(trans, ret);
846 	return ret;
847 }
848 
849 static int add_free_space_extent(struct btrfs_trans_handle *trans,
850 				 struct btrfs_block_group *block_group,
851 				 struct btrfs_path *path,
852 				 u64 start, u64 size)
853 {
854 	struct btrfs_root *root = trans->fs_info->free_space_root;
855 	struct btrfs_key key, new_key;
856 	u64 found_start, found_end;
857 	u64 end = start + size;
858 	int new_extents = 1;
859 	int ret;
860 
861 	/*
862 	 * We are adding a new extent of free space, but we need to merge
863 	 * extents. There are four cases here:
864 	 *
865 	 * 1. The new extent does not have any immediate neighbors to merge
866 	 * with: add the new key and increment the free space extent count. We
867 	 * may need to convert the block group to bitmaps as a result.
868 	 * 2. The new extent has an immediate neighbor before it: remove the
869 	 * previous key and insert a new key combining both of them. There is no
870 	 * net change in the number of extents.
871 	 * 3. The new extent has an immediate neighbor after it: remove the next
872 	 * key and insert a new key combining both of them. There is no net
873 	 * change in the number of extents.
874 	 * 4. The new extent has immediate neighbors on both sides: remove both
875 	 * of the keys and insert a new key combining all of them. Where we used
876 	 * to have two extents, we now have one, so decrement the extent count.
877 	 */
878 
879 	new_key.objectid = start;
880 	new_key.type = BTRFS_FREE_SPACE_EXTENT_KEY;
881 	new_key.offset = size;
882 
883 	/* Search for a neighbor on the left. */
884 	if (start == block_group->start)
885 		goto right;
886 	key.objectid = start - 1;
887 	key.type = (u8)-1;
888 	key.offset = (u64)-1;
889 
890 	ret = btrfs_search_prev_slot(trans, root, &key, path, -1, 1);
891 	if (ret)
892 		goto out;
893 
894 	btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
895 
896 	if (key.type != BTRFS_FREE_SPACE_EXTENT_KEY) {
897 		ASSERT(key.type == BTRFS_FREE_SPACE_INFO_KEY);
898 		btrfs_release_path(path);
899 		goto right;
900 	}
901 
902 	found_start = key.objectid;
903 	found_end = key.objectid + key.offset;
904 	ASSERT(found_start >= block_group->start &&
905 	       found_end > block_group->start);
906 	ASSERT(found_start < start && found_end <= start);
907 
908 	/*
909 	 * Delete the neighbor on the left and absorb it into the new key (cases
910 	 * 2 and 4).
911 	 */
912 	if (found_end == start) {
913 		ret = btrfs_del_item(trans, root, path);
914 		if (ret)
915 			goto out;
916 		new_key.objectid = found_start;
917 		new_key.offset += key.offset;
918 		new_extents--;
919 	}
920 	btrfs_release_path(path);
921 
922 right:
923 	/* Search for a neighbor on the right. */
924 	if (end == block_group->start + block_group->length)
925 		goto insert;
926 	key.objectid = end;
927 	key.type = (u8)-1;
928 	key.offset = (u64)-1;
929 
930 	ret = btrfs_search_prev_slot(trans, root, &key, path, -1, 1);
931 	if (ret)
932 		goto out;
933 
934 	btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
935 
936 	if (key.type != BTRFS_FREE_SPACE_EXTENT_KEY) {
937 		ASSERT(key.type == BTRFS_FREE_SPACE_INFO_KEY);
938 		btrfs_release_path(path);
939 		goto insert;
940 	}
941 
942 	found_start = key.objectid;
943 	found_end = key.objectid + key.offset;
944 	ASSERT(found_start >= block_group->start &&
945 	       found_end > block_group->start);
946 	ASSERT((found_start < start && found_end <= start) ||
947 	       (found_start >= end && found_end > end));
948 
949 	/*
950 	 * Delete the neighbor on the right and absorb it into the new key
951 	 * (cases 3 and 4).
952 	 */
953 	if (found_start == end) {
954 		ret = btrfs_del_item(trans, root, path);
955 		if (ret)
956 			goto out;
957 		new_key.offset += key.offset;
958 		new_extents--;
959 	}
960 	btrfs_release_path(path);
961 
962 insert:
963 	/* Insert the new key (cases 1-4). */
964 	ret = btrfs_insert_empty_item(trans, root, path, &new_key, 0);
965 	if (ret)
966 		goto out;
967 
968 	btrfs_release_path(path);
969 	ret = update_free_space_extent_count(trans, block_group, path,
970 					     new_extents);
971 
972 out:
973 	return ret;
974 }
975 
976 EXPORT_FOR_TESTS
977 int __add_to_free_space_tree(struct btrfs_trans_handle *trans,
978 			     struct btrfs_block_group *block_group,
979 			     struct btrfs_path *path, u64 start, u64 size)
980 {
981 	struct btrfs_free_space_info *info;
982 	u32 flags;
983 	int ret;
984 
985 	if (block_group->needs_free_space) {
986 		ret = __add_block_group_free_space(trans, block_group, path);
987 		if (ret)
988 			return ret;
989 	}
990 
991 	info = search_free_space_info(NULL, block_group, path, 0);
992 	if (IS_ERR(info))
993 		return PTR_ERR(info);
994 	flags = btrfs_free_space_flags(path->nodes[0], info);
995 	btrfs_release_path(path);
996 
997 	if (flags & BTRFS_FREE_SPACE_USING_BITMAPS) {
998 		return modify_free_space_bitmap(trans, block_group, path,
999 						start, size, 0);
1000 	} else {
1001 		return add_free_space_extent(trans, block_group, path, start,
1002 					     size);
1003 	}
1004 }
1005 
1006 int add_to_free_space_tree(struct btrfs_trans_handle *trans,
1007 			   u64 start, u64 size)
1008 {
1009 	struct btrfs_block_group *block_group;
1010 	struct btrfs_path *path;
1011 	int ret;
1012 
1013 	if (!btrfs_fs_compat_ro(trans->fs_info, FREE_SPACE_TREE))
1014 		return 0;
1015 
1016 	path = btrfs_alloc_path();
1017 	if (!path) {
1018 		ret = -ENOMEM;
1019 		goto out;
1020 	}
1021 
1022 	block_group = btrfs_lookup_block_group(trans->fs_info, start);
1023 	if (!block_group) {
1024 		ASSERT(0);
1025 		ret = -ENOENT;
1026 		goto out;
1027 	}
1028 
1029 	mutex_lock(&block_group->free_space_lock);
1030 	ret = __add_to_free_space_tree(trans, block_group, path, start, size);
1031 	mutex_unlock(&block_group->free_space_lock);
1032 
1033 	btrfs_put_block_group(block_group);
1034 out:
1035 	btrfs_free_path(path);
1036 	if (ret)
1037 		btrfs_abort_transaction(trans, ret);
1038 	return ret;
1039 }
1040 
1041 /*
1042  * Populate the free space tree by walking the extent tree. Operations on the
1043  * extent tree that happen as a result of writes to the free space tree will go
1044  * through the normal add/remove hooks.
1045  */
1046 static int populate_free_space_tree(struct btrfs_trans_handle *trans,
1047 				    struct btrfs_block_group *block_group)
1048 {
1049 	struct btrfs_root *extent_root = trans->fs_info->extent_root;
1050 	struct btrfs_path *path, *path2;
1051 	struct btrfs_key key;
1052 	u64 start, end;
1053 	int ret;
1054 
1055 	path = btrfs_alloc_path();
1056 	if (!path)
1057 		return -ENOMEM;
1058 	path->reada = READA_FORWARD;
1059 
1060 	path2 = btrfs_alloc_path();
1061 	if (!path2) {
1062 		btrfs_free_path(path);
1063 		return -ENOMEM;
1064 	}
1065 
1066 	ret = add_new_free_space_info(trans, block_group, path2);
1067 	if (ret)
1068 		goto out;
1069 
1070 	mutex_lock(&block_group->free_space_lock);
1071 
1072 	/*
1073 	 * Iterate through all of the extent and metadata items in this block
1074 	 * group, adding the free space between them and the free space at the
1075 	 * end. Note that EXTENT_ITEM and METADATA_ITEM are less than
1076 	 * BLOCK_GROUP_ITEM, so an extent may precede the block group that it's
1077 	 * contained in.
1078 	 */
1079 	key.objectid = block_group->start;
1080 	key.type = BTRFS_EXTENT_ITEM_KEY;
1081 	key.offset = 0;
1082 
1083 	ret = btrfs_search_slot_for_read(extent_root, &key, path, 1, 0);
1084 	if (ret < 0)
1085 		goto out_locked;
1086 	ASSERT(ret == 0);
1087 
1088 	start = block_group->start;
1089 	end = block_group->start + block_group->length;
1090 	while (1) {
1091 		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
1092 
1093 		if (key.type == BTRFS_EXTENT_ITEM_KEY ||
1094 		    key.type == BTRFS_METADATA_ITEM_KEY) {
1095 			if (key.objectid >= end)
1096 				break;
1097 
1098 			if (start < key.objectid) {
1099 				ret = __add_to_free_space_tree(trans,
1100 							       block_group,
1101 							       path2, start,
1102 							       key.objectid -
1103 							       start);
1104 				if (ret)
1105 					goto out_locked;
1106 			}
1107 			start = key.objectid;
1108 			if (key.type == BTRFS_METADATA_ITEM_KEY)
1109 				start += trans->fs_info->nodesize;
1110 			else
1111 				start += key.offset;
1112 		} else if (key.type == BTRFS_BLOCK_GROUP_ITEM_KEY) {
1113 			if (key.objectid != block_group->start)
1114 				break;
1115 		}
1116 
1117 		ret = btrfs_next_item(extent_root, path);
1118 		if (ret < 0)
1119 			goto out_locked;
1120 		if (ret)
1121 			break;
1122 	}
1123 	if (start < end) {
1124 		ret = __add_to_free_space_tree(trans, block_group, path2,
1125 					       start, end - start);
1126 		if (ret)
1127 			goto out_locked;
1128 	}
1129 
1130 	ret = 0;
1131 out_locked:
1132 	mutex_unlock(&block_group->free_space_lock);
1133 out:
1134 	btrfs_free_path(path2);
1135 	btrfs_free_path(path);
1136 	return ret;
1137 }
1138 
1139 int btrfs_create_free_space_tree(struct btrfs_fs_info *fs_info)
1140 {
1141 	struct btrfs_trans_handle *trans;
1142 	struct btrfs_root *tree_root = fs_info->tree_root;
1143 	struct btrfs_root *free_space_root;
1144 	struct btrfs_block_group *block_group;
1145 	struct rb_node *node;
1146 	int ret;
1147 
1148 	trans = btrfs_start_transaction(tree_root, 0);
1149 	if (IS_ERR(trans))
1150 		return PTR_ERR(trans);
1151 
1152 	set_bit(BTRFS_FS_CREATING_FREE_SPACE_TREE, &fs_info->flags);
1153 	free_space_root = btrfs_create_tree(trans,
1154 					    BTRFS_FREE_SPACE_TREE_OBJECTID);
1155 	if (IS_ERR(free_space_root)) {
1156 		ret = PTR_ERR(free_space_root);
1157 		goto abort;
1158 	}
1159 	fs_info->free_space_root = free_space_root;
1160 
1161 	node = rb_first(&fs_info->block_group_cache_tree);
1162 	while (node) {
1163 		block_group = rb_entry(node, struct btrfs_block_group,
1164 				       cache_node);
1165 		ret = populate_free_space_tree(trans, block_group);
1166 		if (ret)
1167 			goto abort;
1168 		node = rb_next(node);
1169 	}
1170 
1171 	btrfs_set_fs_compat_ro(fs_info, FREE_SPACE_TREE);
1172 	btrfs_set_fs_compat_ro(fs_info, FREE_SPACE_TREE_VALID);
1173 	clear_bit(BTRFS_FS_CREATING_FREE_SPACE_TREE, &fs_info->flags);
1174 
1175 	return btrfs_commit_transaction(trans);
1176 
1177 abort:
1178 	clear_bit(BTRFS_FS_CREATING_FREE_SPACE_TREE, &fs_info->flags);
1179 	btrfs_abort_transaction(trans, ret);
1180 	btrfs_end_transaction(trans);
1181 	return ret;
1182 }
1183 
1184 static int clear_free_space_tree(struct btrfs_trans_handle *trans,
1185 				 struct btrfs_root *root)
1186 {
1187 	struct btrfs_path *path;
1188 	struct btrfs_key key;
1189 	int nr;
1190 	int ret;
1191 
1192 	path = btrfs_alloc_path();
1193 	if (!path)
1194 		return -ENOMEM;
1195 
1196 	key.objectid = 0;
1197 	key.type = 0;
1198 	key.offset = 0;
1199 
1200 	while (1) {
1201 		ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
1202 		if (ret < 0)
1203 			goto out;
1204 
1205 		nr = btrfs_header_nritems(path->nodes[0]);
1206 		if (!nr)
1207 			break;
1208 
1209 		path->slots[0] = 0;
1210 		ret = btrfs_del_items(trans, root, path, 0, nr);
1211 		if (ret)
1212 			goto out;
1213 
1214 		btrfs_release_path(path);
1215 	}
1216 
1217 	ret = 0;
1218 out:
1219 	btrfs_free_path(path);
1220 	return ret;
1221 }
1222 
1223 int btrfs_clear_free_space_tree(struct btrfs_fs_info *fs_info)
1224 {
1225 	struct btrfs_trans_handle *trans;
1226 	struct btrfs_root *tree_root = fs_info->tree_root;
1227 	struct btrfs_root *free_space_root = fs_info->free_space_root;
1228 	int ret;
1229 
1230 	trans = btrfs_start_transaction(tree_root, 0);
1231 	if (IS_ERR(trans))
1232 		return PTR_ERR(trans);
1233 
1234 	btrfs_clear_fs_compat_ro(fs_info, FREE_SPACE_TREE);
1235 	btrfs_clear_fs_compat_ro(fs_info, FREE_SPACE_TREE_VALID);
1236 	fs_info->free_space_root = NULL;
1237 
1238 	ret = clear_free_space_tree(trans, free_space_root);
1239 	if (ret)
1240 		goto abort;
1241 
1242 	ret = btrfs_del_root(trans, &free_space_root->root_key);
1243 	if (ret)
1244 		goto abort;
1245 
1246 	list_del(&free_space_root->dirty_list);
1247 
1248 	btrfs_tree_lock(free_space_root->node);
1249 	btrfs_clean_tree_block(free_space_root->node);
1250 	btrfs_tree_unlock(free_space_root->node);
1251 	btrfs_free_tree_block(trans, free_space_root, free_space_root->node,
1252 			      0, 1);
1253 
1254 	btrfs_put_root(free_space_root);
1255 
1256 	return btrfs_commit_transaction(trans);
1257 
1258 abort:
1259 	btrfs_abort_transaction(trans, ret);
1260 	btrfs_end_transaction(trans);
1261 	return ret;
1262 }
1263 
1264 static int __add_block_group_free_space(struct btrfs_trans_handle *trans,
1265 					struct btrfs_block_group *block_group,
1266 					struct btrfs_path *path)
1267 {
1268 	int ret;
1269 
1270 	block_group->needs_free_space = 0;
1271 
1272 	ret = add_new_free_space_info(trans, block_group, path);
1273 	if (ret)
1274 		return ret;
1275 
1276 	return __add_to_free_space_tree(trans, block_group, path,
1277 					block_group->start,
1278 					block_group->length);
1279 }
1280 
1281 int add_block_group_free_space(struct btrfs_trans_handle *trans,
1282 			       struct btrfs_block_group *block_group)
1283 {
1284 	struct btrfs_fs_info *fs_info = trans->fs_info;
1285 	struct btrfs_path *path = NULL;
1286 	int ret = 0;
1287 
1288 	if (!btrfs_fs_compat_ro(fs_info, FREE_SPACE_TREE))
1289 		return 0;
1290 
1291 	mutex_lock(&block_group->free_space_lock);
1292 	if (!block_group->needs_free_space)
1293 		goto out;
1294 
1295 	path = btrfs_alloc_path();
1296 	if (!path) {
1297 		ret = -ENOMEM;
1298 		goto out;
1299 	}
1300 
1301 	ret = __add_block_group_free_space(trans, block_group, path);
1302 
1303 out:
1304 	btrfs_free_path(path);
1305 	mutex_unlock(&block_group->free_space_lock);
1306 	if (ret)
1307 		btrfs_abort_transaction(trans, ret);
1308 	return ret;
1309 }
1310 
1311 int remove_block_group_free_space(struct btrfs_trans_handle *trans,
1312 				  struct btrfs_block_group *block_group)
1313 {
1314 	struct btrfs_root *root = trans->fs_info->free_space_root;
1315 	struct btrfs_path *path;
1316 	struct btrfs_key key, found_key;
1317 	struct extent_buffer *leaf;
1318 	u64 start, end;
1319 	int done = 0, nr;
1320 	int ret;
1321 
1322 	if (!btrfs_fs_compat_ro(trans->fs_info, FREE_SPACE_TREE))
1323 		return 0;
1324 
1325 	if (block_group->needs_free_space) {
1326 		/* We never added this block group to the free space tree. */
1327 		return 0;
1328 	}
1329 
1330 	path = btrfs_alloc_path();
1331 	if (!path) {
1332 		ret = -ENOMEM;
1333 		goto out;
1334 	}
1335 
1336 	start = block_group->start;
1337 	end = block_group->start + block_group->length;
1338 
1339 	key.objectid = end - 1;
1340 	key.type = (u8)-1;
1341 	key.offset = (u64)-1;
1342 
1343 	while (!done) {
1344 		ret = btrfs_search_prev_slot(trans, root, &key, path, -1, 1);
1345 		if (ret)
1346 			goto out;
1347 
1348 		leaf = path->nodes[0];
1349 		nr = 0;
1350 		path->slots[0]++;
1351 		while (path->slots[0] > 0) {
1352 			btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0] - 1);
1353 
1354 			if (found_key.type == BTRFS_FREE_SPACE_INFO_KEY) {
1355 				ASSERT(found_key.objectid == block_group->start);
1356 				ASSERT(found_key.offset == block_group->length);
1357 				done = 1;
1358 				nr++;
1359 				path->slots[0]--;
1360 				break;
1361 			} else if (found_key.type == BTRFS_FREE_SPACE_EXTENT_KEY ||
1362 				   found_key.type == BTRFS_FREE_SPACE_BITMAP_KEY) {
1363 				ASSERT(found_key.objectid >= start);
1364 				ASSERT(found_key.objectid < end);
1365 				ASSERT(found_key.objectid + found_key.offset <= end);
1366 				nr++;
1367 				path->slots[0]--;
1368 			} else {
1369 				ASSERT(0);
1370 			}
1371 		}
1372 
1373 		ret = btrfs_del_items(trans, root, path, path->slots[0], nr);
1374 		if (ret)
1375 			goto out;
1376 		btrfs_release_path(path);
1377 	}
1378 
1379 	ret = 0;
1380 out:
1381 	btrfs_free_path(path);
1382 	if (ret)
1383 		btrfs_abort_transaction(trans, ret);
1384 	return ret;
1385 }
1386 
1387 static int load_free_space_bitmaps(struct btrfs_caching_control *caching_ctl,
1388 				   struct btrfs_path *path,
1389 				   u32 expected_extent_count)
1390 {
1391 	struct btrfs_block_group *block_group;
1392 	struct btrfs_fs_info *fs_info;
1393 	struct btrfs_root *root;
1394 	struct btrfs_key key;
1395 	int prev_bit = 0, bit;
1396 	/* Initialize to silence GCC. */
1397 	u64 extent_start = 0;
1398 	u64 end, offset;
1399 	u64 total_found = 0;
1400 	u32 extent_count = 0;
1401 	int ret;
1402 
1403 	block_group = caching_ctl->block_group;
1404 	fs_info = block_group->fs_info;
1405 	root = fs_info->free_space_root;
1406 
1407 	end = block_group->start + block_group->length;
1408 
1409 	while (1) {
1410 		ret = btrfs_next_item(root, path);
1411 		if (ret < 0)
1412 			goto out;
1413 		if (ret)
1414 			break;
1415 
1416 		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
1417 
1418 		if (key.type == BTRFS_FREE_SPACE_INFO_KEY)
1419 			break;
1420 
1421 		ASSERT(key.type == BTRFS_FREE_SPACE_BITMAP_KEY);
1422 		ASSERT(key.objectid < end && key.objectid + key.offset <= end);
1423 
1424 		caching_ctl->progress = key.objectid;
1425 
1426 		offset = key.objectid;
1427 		while (offset < key.objectid + key.offset) {
1428 			bit = free_space_test_bit(block_group, path, offset);
1429 			if (prev_bit == 0 && bit == 1) {
1430 				extent_start = offset;
1431 			} else if (prev_bit == 1 && bit == 0) {
1432 				total_found += add_new_free_space(block_group,
1433 								  extent_start,
1434 								  offset);
1435 				if (total_found > CACHING_CTL_WAKE_UP) {
1436 					total_found = 0;
1437 					wake_up(&caching_ctl->wait);
1438 				}
1439 				extent_count++;
1440 			}
1441 			prev_bit = bit;
1442 			offset += fs_info->sectorsize;
1443 		}
1444 	}
1445 	if (prev_bit == 1) {
1446 		total_found += add_new_free_space(block_group, extent_start,
1447 						  end);
1448 		extent_count++;
1449 	}
1450 
1451 	if (extent_count != expected_extent_count) {
1452 		btrfs_err(fs_info,
1453 			  "incorrect extent count for %llu; counted %u, expected %u",
1454 			  block_group->start, extent_count,
1455 			  expected_extent_count);
1456 		ASSERT(0);
1457 		ret = -EIO;
1458 		goto out;
1459 	}
1460 
1461 	caching_ctl->progress = (u64)-1;
1462 
1463 	ret = 0;
1464 out:
1465 	return ret;
1466 }
1467 
1468 static int load_free_space_extents(struct btrfs_caching_control *caching_ctl,
1469 				   struct btrfs_path *path,
1470 				   u32 expected_extent_count)
1471 {
1472 	struct btrfs_block_group *block_group;
1473 	struct btrfs_fs_info *fs_info;
1474 	struct btrfs_root *root;
1475 	struct btrfs_key key;
1476 	u64 end;
1477 	u64 total_found = 0;
1478 	u32 extent_count = 0;
1479 	int ret;
1480 
1481 	block_group = caching_ctl->block_group;
1482 	fs_info = block_group->fs_info;
1483 	root = fs_info->free_space_root;
1484 
1485 	end = block_group->start + block_group->length;
1486 
1487 	while (1) {
1488 		ret = btrfs_next_item(root, path);
1489 		if (ret < 0)
1490 			goto out;
1491 		if (ret)
1492 			break;
1493 
1494 		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
1495 
1496 		if (key.type == BTRFS_FREE_SPACE_INFO_KEY)
1497 			break;
1498 
1499 		ASSERT(key.type == BTRFS_FREE_SPACE_EXTENT_KEY);
1500 		ASSERT(key.objectid < end && key.objectid + key.offset <= end);
1501 
1502 		caching_ctl->progress = key.objectid;
1503 
1504 		total_found += add_new_free_space(block_group, key.objectid,
1505 						  key.objectid + key.offset);
1506 		if (total_found > CACHING_CTL_WAKE_UP) {
1507 			total_found = 0;
1508 			wake_up(&caching_ctl->wait);
1509 		}
1510 		extent_count++;
1511 	}
1512 
1513 	if (extent_count != expected_extent_count) {
1514 		btrfs_err(fs_info,
1515 			  "incorrect extent count for %llu; counted %u, expected %u",
1516 			  block_group->start, extent_count,
1517 			  expected_extent_count);
1518 		ASSERT(0);
1519 		ret = -EIO;
1520 		goto out;
1521 	}
1522 
1523 	caching_ctl->progress = (u64)-1;
1524 
1525 	ret = 0;
1526 out:
1527 	return ret;
1528 }
1529 
1530 int load_free_space_tree(struct btrfs_caching_control *caching_ctl)
1531 {
1532 	struct btrfs_block_group *block_group;
1533 	struct btrfs_free_space_info *info;
1534 	struct btrfs_path *path;
1535 	u32 extent_count, flags;
1536 	int ret;
1537 
1538 	block_group = caching_ctl->block_group;
1539 
1540 	path = btrfs_alloc_path();
1541 	if (!path)
1542 		return -ENOMEM;
1543 
1544 	/*
1545 	 * Just like caching_thread() doesn't want to deadlock on the extent
1546 	 * tree, we don't want to deadlock on the free space tree.
1547 	 */
1548 	path->skip_locking = 1;
1549 	path->search_commit_root = 1;
1550 	path->reada = READA_FORWARD;
1551 
1552 	info = search_free_space_info(NULL, block_group, path, 0);
1553 	if (IS_ERR(info)) {
1554 		ret = PTR_ERR(info);
1555 		goto out;
1556 	}
1557 	extent_count = btrfs_free_space_extent_count(path->nodes[0], info);
1558 	flags = btrfs_free_space_flags(path->nodes[0], info);
1559 
1560 	/*
1561 	 * We left path pointing to the free space info item, so now
1562 	 * load_free_space_foo can just iterate through the free space tree from
1563 	 * there.
1564 	 */
1565 	if (flags & BTRFS_FREE_SPACE_USING_BITMAPS)
1566 		ret = load_free_space_bitmaps(caching_ctl, path, extent_count);
1567 	else
1568 		ret = load_free_space_extents(caching_ctl, path, extent_count);
1569 
1570 out:
1571 	btrfs_free_path(path);
1572 	return ret;
1573 }
1574