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
btrfs_free_space_root(struct btrfs_block_group * block_group)19 static struct btrfs_root *btrfs_free_space_root(
20 struct btrfs_block_group *block_group)
21 {
22 struct btrfs_key key = {
23 .objectid = BTRFS_FREE_SPACE_TREE_OBJECTID,
24 .type = BTRFS_ROOT_ITEM_KEY,
25 .offset = 0,
26 };
27
28 if (btrfs_fs_incompat(block_group->fs_info, EXTENT_TREE_V2))
29 key.offset = block_group->global_root_id;
30 return btrfs_global_root(block_group->fs_info, &key);
31 }
32
set_free_space_tree_thresholds(struct btrfs_block_group * cache)33 void set_free_space_tree_thresholds(struct btrfs_block_group *cache)
34 {
35 u32 bitmap_range;
36 size_t bitmap_size;
37 u64 num_bitmaps, total_bitmap_size;
38
39 if (WARN_ON(cache->length == 0))
40 btrfs_warn(cache->fs_info, "block group %llu length is zero",
41 cache->start);
42
43 /*
44 * We convert to bitmaps when the disk space required for using extents
45 * exceeds that required for using bitmaps.
46 */
47 bitmap_range = cache->fs_info->sectorsize * BTRFS_FREE_SPACE_BITMAP_BITS;
48 num_bitmaps = div_u64(cache->length + bitmap_range - 1, bitmap_range);
49 bitmap_size = sizeof(struct btrfs_item) + BTRFS_FREE_SPACE_BITMAP_SIZE;
50 total_bitmap_size = num_bitmaps * bitmap_size;
51 cache->bitmap_high_thresh = div_u64(total_bitmap_size,
52 sizeof(struct btrfs_item));
53
54 /*
55 * We allow for a small buffer between the high threshold and low
56 * threshold to avoid thrashing back and forth between the two formats.
57 */
58 if (cache->bitmap_high_thresh > 100)
59 cache->bitmap_low_thresh = cache->bitmap_high_thresh - 100;
60 else
61 cache->bitmap_low_thresh = 0;
62 }
63
add_new_free_space_info(struct btrfs_trans_handle * trans,struct btrfs_block_group * block_group,struct btrfs_path * path)64 static int add_new_free_space_info(struct btrfs_trans_handle *trans,
65 struct btrfs_block_group *block_group,
66 struct btrfs_path *path)
67 {
68 struct btrfs_root *root = btrfs_free_space_root(block_group);
69 struct btrfs_free_space_info *info;
70 struct btrfs_key key;
71 struct extent_buffer *leaf;
72 int ret;
73
74 key.objectid = block_group->start;
75 key.type = BTRFS_FREE_SPACE_INFO_KEY;
76 key.offset = block_group->length;
77
78 ret = btrfs_insert_empty_item(trans, root, path, &key, sizeof(*info));
79 if (ret)
80 goto out;
81
82 leaf = path->nodes[0];
83 info = btrfs_item_ptr(leaf, path->slots[0],
84 struct btrfs_free_space_info);
85 btrfs_set_free_space_extent_count(leaf, info, 0);
86 btrfs_set_free_space_flags(leaf, info, 0);
87 btrfs_mark_buffer_dirty(leaf);
88
89 ret = 0;
90 out:
91 btrfs_release_path(path);
92 return ret;
93 }
94
95 EXPORT_FOR_TESTS
search_free_space_info(struct btrfs_trans_handle * trans,struct btrfs_block_group * block_group,struct btrfs_path * path,int cow)96 struct btrfs_free_space_info *search_free_space_info(
97 struct btrfs_trans_handle *trans,
98 struct btrfs_block_group *block_group,
99 struct btrfs_path *path, int cow)
100 {
101 struct btrfs_fs_info *fs_info = block_group->fs_info;
102 struct btrfs_root *root = btrfs_free_space_root(block_group);
103 struct btrfs_key key;
104 int ret;
105
106 key.objectid = block_group->start;
107 key.type = BTRFS_FREE_SPACE_INFO_KEY;
108 key.offset = block_group->length;
109
110 ret = btrfs_search_slot(trans, root, &key, path, 0, cow);
111 if (ret < 0)
112 return ERR_PTR(ret);
113 if (ret != 0) {
114 btrfs_warn(fs_info, "missing free space info for %llu",
115 block_group->start);
116 ASSERT(0);
117 return ERR_PTR(-ENOENT);
118 }
119
120 return btrfs_item_ptr(path->nodes[0], path->slots[0],
121 struct btrfs_free_space_info);
122 }
123
124 /*
125 * btrfs_search_slot() but we're looking for the greatest key less than the
126 * passed key.
127 */
btrfs_search_prev_slot(struct btrfs_trans_handle * trans,struct btrfs_root * root,struct btrfs_key * key,struct btrfs_path * p,int ins_len,int cow)128 static int btrfs_search_prev_slot(struct btrfs_trans_handle *trans,
129 struct btrfs_root *root,
130 struct btrfs_key *key, struct btrfs_path *p,
131 int ins_len, int cow)
132 {
133 int ret;
134
135 ret = btrfs_search_slot(trans, root, key, p, ins_len, cow);
136 if (ret < 0)
137 return ret;
138
139 if (ret == 0) {
140 ASSERT(0);
141 return -EIO;
142 }
143
144 if (p->slots[0] == 0) {
145 ASSERT(0);
146 return -EIO;
147 }
148 p->slots[0]--;
149
150 return 0;
151 }
152
free_space_bitmap_size(const struct btrfs_fs_info * fs_info,u64 size)153 static inline u32 free_space_bitmap_size(const struct btrfs_fs_info *fs_info,
154 u64 size)
155 {
156 return DIV_ROUND_UP(size >> fs_info->sectorsize_bits, BITS_PER_BYTE);
157 }
158
alloc_bitmap(u32 bitmap_size)159 static unsigned long *alloc_bitmap(u32 bitmap_size)
160 {
161 unsigned long *ret;
162 unsigned int nofs_flag;
163 u32 bitmap_rounded_size = round_up(bitmap_size, sizeof(unsigned long));
164
165 /*
166 * GFP_NOFS doesn't work with kvmalloc(), but we really can't recurse
167 * into the filesystem as the free space bitmap can be modified in the
168 * critical section of a transaction commit.
169 *
170 * TODO: push the memalloc_nofs_{save,restore}() to the caller where we
171 * know that recursion is unsafe.
172 */
173 nofs_flag = memalloc_nofs_save();
174 ret = kvzalloc(bitmap_rounded_size, GFP_KERNEL);
175 memalloc_nofs_restore(nofs_flag);
176 return ret;
177 }
178
le_bitmap_set(unsigned long * map,unsigned int start,int len)179 static void le_bitmap_set(unsigned long *map, unsigned int start, int len)
180 {
181 u8 *p = ((u8 *)map) + BIT_BYTE(start);
182 const unsigned int size = start + len;
183 int bits_to_set = BITS_PER_BYTE - (start % BITS_PER_BYTE);
184 u8 mask_to_set = BITMAP_FIRST_BYTE_MASK(start);
185
186 while (len - bits_to_set >= 0) {
187 *p |= mask_to_set;
188 len -= bits_to_set;
189 bits_to_set = BITS_PER_BYTE;
190 mask_to_set = ~0;
191 p++;
192 }
193 if (len) {
194 mask_to_set &= BITMAP_LAST_BYTE_MASK(size);
195 *p |= mask_to_set;
196 }
197 }
198
199 EXPORT_FOR_TESTS
convert_free_space_to_bitmaps(struct btrfs_trans_handle * trans,struct btrfs_block_group * block_group,struct btrfs_path * path)200 int convert_free_space_to_bitmaps(struct btrfs_trans_handle *trans,
201 struct btrfs_block_group *block_group,
202 struct btrfs_path *path)
203 {
204 struct btrfs_fs_info *fs_info = trans->fs_info;
205 struct btrfs_root *root = btrfs_free_space_root(block_group);
206 struct btrfs_free_space_info *info;
207 struct btrfs_key key, found_key;
208 struct extent_buffer *leaf;
209 unsigned long *bitmap;
210 char *bitmap_cursor;
211 u64 start, end;
212 u64 bitmap_range, i;
213 u32 bitmap_size, flags, expected_extent_count;
214 u32 extent_count = 0;
215 int done = 0, nr;
216 int ret;
217
218 bitmap_size = free_space_bitmap_size(fs_info, block_group->length);
219 bitmap = alloc_bitmap(bitmap_size);
220 if (!bitmap) {
221 ret = -ENOMEM;
222 goto out;
223 }
224
225 start = block_group->start;
226 end = block_group->start + block_group->length;
227
228 key.objectid = end - 1;
229 key.type = (u8)-1;
230 key.offset = (u64)-1;
231
232 while (!done) {
233 ret = btrfs_search_prev_slot(trans, root, &key, path, -1, 1);
234 if (ret)
235 goto out;
236
237 leaf = path->nodes[0];
238 nr = 0;
239 path->slots[0]++;
240 while (path->slots[0] > 0) {
241 btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0] - 1);
242
243 if (found_key.type == BTRFS_FREE_SPACE_INFO_KEY) {
244 ASSERT(found_key.objectid == block_group->start);
245 ASSERT(found_key.offset == block_group->length);
246 done = 1;
247 break;
248 } else if (found_key.type == BTRFS_FREE_SPACE_EXTENT_KEY) {
249 u64 first, last;
250
251 ASSERT(found_key.objectid >= start);
252 ASSERT(found_key.objectid < end);
253 ASSERT(found_key.objectid + found_key.offset <= end);
254
255 first = div_u64(found_key.objectid - start,
256 fs_info->sectorsize);
257 last = div_u64(found_key.objectid + found_key.offset - start,
258 fs_info->sectorsize);
259 le_bitmap_set(bitmap, first, last - first);
260
261 extent_count++;
262 nr++;
263 path->slots[0]--;
264 } else {
265 ASSERT(0);
266 }
267 }
268
269 ret = btrfs_del_items(trans, root, path, path->slots[0], nr);
270 if (ret)
271 goto out;
272 btrfs_release_path(path);
273 }
274
275 info = search_free_space_info(trans, block_group, path, 1);
276 if (IS_ERR(info)) {
277 ret = PTR_ERR(info);
278 goto out;
279 }
280 leaf = path->nodes[0];
281 flags = btrfs_free_space_flags(leaf, info);
282 flags |= BTRFS_FREE_SPACE_USING_BITMAPS;
283 btrfs_set_free_space_flags(leaf, info, flags);
284 expected_extent_count = btrfs_free_space_extent_count(leaf, info);
285 btrfs_mark_buffer_dirty(leaf);
286 btrfs_release_path(path);
287
288 if (extent_count != expected_extent_count) {
289 btrfs_err(fs_info,
290 "incorrect extent count for %llu; counted %u, expected %u",
291 block_group->start, extent_count,
292 expected_extent_count);
293 ASSERT(0);
294 ret = -EIO;
295 goto out;
296 }
297
298 bitmap_cursor = (char *)bitmap;
299 bitmap_range = fs_info->sectorsize * BTRFS_FREE_SPACE_BITMAP_BITS;
300 i = start;
301 while (i < end) {
302 unsigned long ptr;
303 u64 extent_size;
304 u32 data_size;
305
306 extent_size = min(end - i, bitmap_range);
307 data_size = free_space_bitmap_size(fs_info, extent_size);
308
309 key.objectid = i;
310 key.type = BTRFS_FREE_SPACE_BITMAP_KEY;
311 key.offset = extent_size;
312
313 ret = btrfs_insert_empty_item(trans, root, path, &key,
314 data_size);
315 if (ret)
316 goto out;
317
318 leaf = path->nodes[0];
319 ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
320 write_extent_buffer(leaf, bitmap_cursor, ptr,
321 data_size);
322 btrfs_mark_buffer_dirty(leaf);
323 btrfs_release_path(path);
324
325 i += extent_size;
326 bitmap_cursor += data_size;
327 }
328
329 ret = 0;
330 out:
331 kvfree(bitmap);
332 if (ret)
333 btrfs_abort_transaction(trans, ret);
334 return ret;
335 }
336
337 EXPORT_FOR_TESTS
convert_free_space_to_extents(struct btrfs_trans_handle * trans,struct btrfs_block_group * block_group,struct btrfs_path * path)338 int convert_free_space_to_extents(struct btrfs_trans_handle *trans,
339 struct btrfs_block_group *block_group,
340 struct btrfs_path *path)
341 {
342 struct btrfs_fs_info *fs_info = trans->fs_info;
343 struct btrfs_root *root = btrfs_free_space_root(block_group);
344 struct btrfs_free_space_info *info;
345 struct btrfs_key key, found_key;
346 struct extent_buffer *leaf;
347 unsigned long *bitmap;
348 u64 start, end;
349 u32 bitmap_size, flags, expected_extent_count;
350 unsigned long nrbits, start_bit, end_bit;
351 u32 extent_count = 0;
352 int done = 0, nr;
353 int ret;
354
355 bitmap_size = free_space_bitmap_size(fs_info, block_group->length);
356 bitmap = alloc_bitmap(bitmap_size);
357 if (!bitmap) {
358 ret = -ENOMEM;
359 goto out;
360 }
361
362 start = block_group->start;
363 end = block_group->start + block_group->length;
364
365 key.objectid = end - 1;
366 key.type = (u8)-1;
367 key.offset = (u64)-1;
368
369 while (!done) {
370 ret = btrfs_search_prev_slot(trans, root, &key, path, -1, 1);
371 if (ret)
372 goto out;
373
374 leaf = path->nodes[0];
375 nr = 0;
376 path->slots[0]++;
377 while (path->slots[0] > 0) {
378 btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0] - 1);
379
380 if (found_key.type == BTRFS_FREE_SPACE_INFO_KEY) {
381 ASSERT(found_key.objectid == block_group->start);
382 ASSERT(found_key.offset == block_group->length);
383 done = 1;
384 break;
385 } else if (found_key.type == BTRFS_FREE_SPACE_BITMAP_KEY) {
386 unsigned long ptr;
387 char *bitmap_cursor;
388 u32 bitmap_pos, data_size;
389
390 ASSERT(found_key.objectid >= start);
391 ASSERT(found_key.objectid < end);
392 ASSERT(found_key.objectid + found_key.offset <= end);
393
394 bitmap_pos = div_u64(found_key.objectid - start,
395 fs_info->sectorsize *
396 BITS_PER_BYTE);
397 bitmap_cursor = ((char *)bitmap) + bitmap_pos;
398 data_size = free_space_bitmap_size(fs_info,
399 found_key.offset);
400
401 ptr = btrfs_item_ptr_offset(leaf, path->slots[0] - 1);
402 read_extent_buffer(leaf, bitmap_cursor, ptr,
403 data_size);
404
405 nr++;
406 path->slots[0]--;
407 } else {
408 ASSERT(0);
409 }
410 }
411
412 ret = btrfs_del_items(trans, root, path, path->slots[0], nr);
413 if (ret)
414 goto out;
415 btrfs_release_path(path);
416 }
417
418 info = search_free_space_info(trans, block_group, path, 1);
419 if (IS_ERR(info)) {
420 ret = PTR_ERR(info);
421 goto out;
422 }
423 leaf = path->nodes[0];
424 flags = btrfs_free_space_flags(leaf, info);
425 flags &= ~BTRFS_FREE_SPACE_USING_BITMAPS;
426 btrfs_set_free_space_flags(leaf, info, flags);
427 expected_extent_count = btrfs_free_space_extent_count(leaf, info);
428 btrfs_mark_buffer_dirty(leaf);
429 btrfs_release_path(path);
430
431 nrbits = block_group->length >> block_group->fs_info->sectorsize_bits;
432 start_bit = find_next_bit_le(bitmap, nrbits, 0);
433
434 while (start_bit < nrbits) {
435 end_bit = find_next_zero_bit_le(bitmap, nrbits, start_bit);
436 ASSERT(start_bit < end_bit);
437
438 key.objectid = start + start_bit * block_group->fs_info->sectorsize;
439 key.type = BTRFS_FREE_SPACE_EXTENT_KEY;
440 key.offset = (end_bit - start_bit) * block_group->fs_info->sectorsize;
441
442 ret = btrfs_insert_empty_item(trans, root, path, &key, 0);
443 if (ret)
444 goto out;
445 btrfs_release_path(path);
446
447 extent_count++;
448
449 start_bit = find_next_bit_le(bitmap, nrbits, end_bit);
450 }
451
452 if (extent_count != expected_extent_count) {
453 btrfs_err(fs_info,
454 "incorrect extent count for %llu; counted %u, expected %u",
455 block_group->start, extent_count,
456 expected_extent_count);
457 ASSERT(0);
458 ret = -EIO;
459 goto out;
460 }
461
462 ret = 0;
463 out:
464 kvfree(bitmap);
465 if (ret)
466 btrfs_abort_transaction(trans, ret);
467 return ret;
468 }
469
update_free_space_extent_count(struct btrfs_trans_handle * trans,struct btrfs_block_group * block_group,struct btrfs_path * path,int new_extents)470 static int update_free_space_extent_count(struct btrfs_trans_handle *trans,
471 struct btrfs_block_group *block_group,
472 struct btrfs_path *path,
473 int new_extents)
474 {
475 struct btrfs_free_space_info *info;
476 u32 flags;
477 u32 extent_count;
478 int ret = 0;
479
480 if (new_extents == 0)
481 return 0;
482
483 info = search_free_space_info(trans, block_group, path, 1);
484 if (IS_ERR(info)) {
485 ret = PTR_ERR(info);
486 goto out;
487 }
488 flags = btrfs_free_space_flags(path->nodes[0], info);
489 extent_count = btrfs_free_space_extent_count(path->nodes[0], info);
490
491 extent_count += new_extents;
492 btrfs_set_free_space_extent_count(path->nodes[0], info, extent_count);
493 btrfs_mark_buffer_dirty(path->nodes[0]);
494 btrfs_release_path(path);
495
496 if (!(flags & BTRFS_FREE_SPACE_USING_BITMAPS) &&
497 extent_count > block_group->bitmap_high_thresh) {
498 ret = convert_free_space_to_bitmaps(trans, block_group, path);
499 } else if ((flags & BTRFS_FREE_SPACE_USING_BITMAPS) &&
500 extent_count < block_group->bitmap_low_thresh) {
501 ret = convert_free_space_to_extents(trans, block_group, path);
502 }
503
504 out:
505 return ret;
506 }
507
508 EXPORT_FOR_TESTS
free_space_test_bit(struct btrfs_block_group * block_group,struct btrfs_path * path,u64 offset)509 int free_space_test_bit(struct btrfs_block_group *block_group,
510 struct btrfs_path *path, u64 offset)
511 {
512 struct extent_buffer *leaf;
513 struct btrfs_key key;
514 u64 found_start, found_end;
515 unsigned long ptr, i;
516
517 leaf = path->nodes[0];
518 btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
519 ASSERT(key.type == BTRFS_FREE_SPACE_BITMAP_KEY);
520
521 found_start = key.objectid;
522 found_end = key.objectid + key.offset;
523 ASSERT(offset >= found_start && offset < found_end);
524
525 ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
526 i = div_u64(offset - found_start,
527 block_group->fs_info->sectorsize);
528 return !!extent_buffer_test_bit(leaf, ptr, i);
529 }
530
free_space_set_bits(struct btrfs_block_group * block_group,struct btrfs_path * path,u64 * start,u64 * size,int bit)531 static void free_space_set_bits(struct btrfs_block_group *block_group,
532 struct btrfs_path *path, u64 *start, u64 *size,
533 int bit)
534 {
535 struct btrfs_fs_info *fs_info = block_group->fs_info;
536 struct extent_buffer *leaf;
537 struct btrfs_key key;
538 u64 end = *start + *size;
539 u64 found_start, found_end;
540 unsigned long ptr, first, last;
541
542 leaf = path->nodes[0];
543 btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
544 ASSERT(key.type == BTRFS_FREE_SPACE_BITMAP_KEY);
545
546 found_start = key.objectid;
547 found_end = key.objectid + key.offset;
548 ASSERT(*start >= found_start && *start < found_end);
549 ASSERT(end > found_start);
550
551 if (end > found_end)
552 end = found_end;
553
554 ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
555 first = (*start - found_start) >> fs_info->sectorsize_bits;
556 last = (end - found_start) >> fs_info->sectorsize_bits;
557 if (bit)
558 extent_buffer_bitmap_set(leaf, ptr, first, last - first);
559 else
560 extent_buffer_bitmap_clear(leaf, ptr, first, last - first);
561 btrfs_mark_buffer_dirty(leaf);
562
563 *size -= end - *start;
564 *start = end;
565 }
566
567 /*
568 * We can't use btrfs_next_item() in modify_free_space_bitmap() because
569 * btrfs_next_leaf() doesn't get the path for writing. We can forgo the fancy
570 * tree walking in btrfs_next_leaf() anyways because we know exactly what we're
571 * looking for.
572 */
free_space_next_bitmap(struct btrfs_trans_handle * trans,struct btrfs_root * root,struct btrfs_path * p)573 static int free_space_next_bitmap(struct btrfs_trans_handle *trans,
574 struct btrfs_root *root, struct btrfs_path *p)
575 {
576 struct btrfs_key key;
577
578 if (p->slots[0] + 1 < btrfs_header_nritems(p->nodes[0])) {
579 p->slots[0]++;
580 return 0;
581 }
582
583 btrfs_item_key_to_cpu(p->nodes[0], &key, p->slots[0]);
584 btrfs_release_path(p);
585
586 key.objectid += key.offset;
587 key.type = (u8)-1;
588 key.offset = (u64)-1;
589
590 return btrfs_search_prev_slot(trans, root, &key, p, 0, 1);
591 }
592
593 /*
594 * If remove is 1, then we are removing free space, thus clearing bits in the
595 * bitmap. If remove is 0, then we are adding free space, thus setting bits in
596 * the bitmap.
597 */
modify_free_space_bitmap(struct btrfs_trans_handle * trans,struct btrfs_block_group * block_group,struct btrfs_path * path,u64 start,u64 size,int remove)598 static int modify_free_space_bitmap(struct btrfs_trans_handle *trans,
599 struct btrfs_block_group *block_group,
600 struct btrfs_path *path,
601 u64 start, u64 size, int remove)
602 {
603 struct btrfs_root *root = btrfs_free_space_root(block_group);
604 struct btrfs_key key;
605 u64 end = start + size;
606 u64 cur_start, cur_size;
607 int prev_bit, next_bit;
608 int new_extents;
609 int ret;
610
611 /*
612 * Read the bit for the block immediately before the extent of space if
613 * that block is within the block group.
614 */
615 if (start > block_group->start) {
616 u64 prev_block = start - block_group->fs_info->sectorsize;
617
618 key.objectid = prev_block;
619 key.type = (u8)-1;
620 key.offset = (u64)-1;
621
622 ret = btrfs_search_prev_slot(trans, root, &key, path, 0, 1);
623 if (ret)
624 goto out;
625
626 prev_bit = free_space_test_bit(block_group, path, prev_block);
627
628 /* The previous block may have been in the previous bitmap. */
629 btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
630 if (start >= key.objectid + key.offset) {
631 ret = free_space_next_bitmap(trans, root, path);
632 if (ret)
633 goto out;
634 }
635 } else {
636 key.objectid = start;
637 key.type = (u8)-1;
638 key.offset = (u64)-1;
639
640 ret = btrfs_search_prev_slot(trans, root, &key, path, 0, 1);
641 if (ret)
642 goto out;
643
644 prev_bit = -1;
645 }
646
647 /*
648 * Iterate over all of the bitmaps overlapped by the extent of space,
649 * clearing/setting bits as required.
650 */
651 cur_start = start;
652 cur_size = size;
653 while (1) {
654 free_space_set_bits(block_group, path, &cur_start, &cur_size,
655 !remove);
656 if (cur_size == 0)
657 break;
658 ret = free_space_next_bitmap(trans, root, path);
659 if (ret)
660 goto out;
661 }
662
663 /*
664 * Read the bit for the block immediately after the extent of space if
665 * that block is within the block group.
666 */
667 if (end < block_group->start + block_group->length) {
668 /* The next block may be in the next bitmap. */
669 btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
670 if (end >= key.objectid + key.offset) {
671 ret = free_space_next_bitmap(trans, root, path);
672 if (ret)
673 goto out;
674 }
675
676 next_bit = free_space_test_bit(block_group, path, end);
677 } else {
678 next_bit = -1;
679 }
680
681 if (remove) {
682 new_extents = -1;
683 if (prev_bit == 1) {
684 /* Leftover on the left. */
685 new_extents++;
686 }
687 if (next_bit == 1) {
688 /* Leftover on the right. */
689 new_extents++;
690 }
691 } else {
692 new_extents = 1;
693 if (prev_bit == 1) {
694 /* Merging with neighbor on the left. */
695 new_extents--;
696 }
697 if (next_bit == 1) {
698 /* Merging with neighbor on the right. */
699 new_extents--;
700 }
701 }
702
703 btrfs_release_path(path);
704 ret = update_free_space_extent_count(trans, block_group, path,
705 new_extents);
706
707 out:
708 return ret;
709 }
710
remove_free_space_extent(struct btrfs_trans_handle * trans,struct btrfs_block_group * block_group,struct btrfs_path * path,u64 start,u64 size)711 static int remove_free_space_extent(struct btrfs_trans_handle *trans,
712 struct btrfs_block_group *block_group,
713 struct btrfs_path *path,
714 u64 start, u64 size)
715 {
716 struct btrfs_root *root = btrfs_free_space_root(block_group);
717 struct btrfs_key key;
718 u64 found_start, found_end;
719 u64 end = start + size;
720 int new_extents = -1;
721 int ret;
722
723 key.objectid = start;
724 key.type = (u8)-1;
725 key.offset = (u64)-1;
726
727 ret = btrfs_search_prev_slot(trans, root, &key, path, -1, 1);
728 if (ret)
729 goto out;
730
731 btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
732
733 ASSERT(key.type == BTRFS_FREE_SPACE_EXTENT_KEY);
734
735 found_start = key.objectid;
736 found_end = key.objectid + key.offset;
737 ASSERT(start >= found_start && end <= found_end);
738
739 /*
740 * Okay, now that we've found the free space extent which contains the
741 * free space that we are removing, there are four cases:
742 *
743 * 1. We're using the whole extent: delete the key we found and
744 * decrement the free space extent count.
745 * 2. We are using part of the extent starting at the beginning: delete
746 * the key we found and insert a new key representing the leftover at
747 * the end. There is no net change in the number of extents.
748 * 3. We are using part of the extent ending at the end: delete the key
749 * we found and insert a new key representing the leftover at the
750 * beginning. There is no net change in the number of extents.
751 * 4. We are using part of the extent in the middle: delete the key we
752 * found and insert two new keys representing the leftovers on each
753 * side. Where we used to have one extent, we now have two, so increment
754 * the extent count. We may need to convert the block group to bitmaps
755 * as a result.
756 */
757
758 /* Delete the existing key (cases 1-4). */
759 ret = btrfs_del_item(trans, root, path);
760 if (ret)
761 goto out;
762
763 /* Add a key for leftovers at the beginning (cases 3 and 4). */
764 if (start > found_start) {
765 key.objectid = found_start;
766 key.type = BTRFS_FREE_SPACE_EXTENT_KEY;
767 key.offset = start - found_start;
768
769 btrfs_release_path(path);
770 ret = btrfs_insert_empty_item(trans, root, path, &key, 0);
771 if (ret)
772 goto out;
773 new_extents++;
774 }
775
776 /* Add a key for leftovers at the end (cases 2 and 4). */
777 if (end < found_end) {
778 key.objectid = end;
779 key.type = BTRFS_FREE_SPACE_EXTENT_KEY;
780 key.offset = found_end - end;
781
782 btrfs_release_path(path);
783 ret = btrfs_insert_empty_item(trans, root, path, &key, 0);
784 if (ret)
785 goto out;
786 new_extents++;
787 }
788
789 btrfs_release_path(path);
790 ret = update_free_space_extent_count(trans, block_group, path,
791 new_extents);
792
793 out:
794 return ret;
795 }
796
797 EXPORT_FOR_TESTS
__remove_from_free_space_tree(struct btrfs_trans_handle * trans,struct btrfs_block_group * block_group,struct btrfs_path * path,u64 start,u64 size)798 int __remove_from_free_space_tree(struct btrfs_trans_handle *trans,
799 struct btrfs_block_group *block_group,
800 struct btrfs_path *path, u64 start, u64 size)
801 {
802 struct btrfs_free_space_info *info;
803 u32 flags;
804 int ret;
805
806 if (block_group->needs_free_space) {
807 ret = __add_block_group_free_space(trans, block_group, path);
808 if (ret)
809 return ret;
810 }
811
812 info = search_free_space_info(NULL, block_group, path, 0);
813 if (IS_ERR(info))
814 return PTR_ERR(info);
815 flags = btrfs_free_space_flags(path->nodes[0], info);
816 btrfs_release_path(path);
817
818 if (flags & BTRFS_FREE_SPACE_USING_BITMAPS) {
819 return modify_free_space_bitmap(trans, block_group, path,
820 start, size, 1);
821 } else {
822 return remove_free_space_extent(trans, block_group, path,
823 start, size);
824 }
825 }
826
remove_from_free_space_tree(struct btrfs_trans_handle * trans,u64 start,u64 size)827 int remove_from_free_space_tree(struct btrfs_trans_handle *trans,
828 u64 start, u64 size)
829 {
830 struct btrfs_block_group *block_group;
831 struct btrfs_path *path;
832 int ret;
833
834 if (!btrfs_fs_compat_ro(trans->fs_info, FREE_SPACE_TREE))
835 return 0;
836
837 path = btrfs_alloc_path();
838 if (!path) {
839 ret = -ENOMEM;
840 goto out;
841 }
842
843 block_group = btrfs_lookup_block_group(trans->fs_info, start);
844 if (!block_group) {
845 ASSERT(0);
846 ret = -ENOENT;
847 goto out;
848 }
849
850 mutex_lock(&block_group->free_space_lock);
851 ret = __remove_from_free_space_tree(trans, block_group, path, start,
852 size);
853 mutex_unlock(&block_group->free_space_lock);
854
855 btrfs_put_block_group(block_group);
856 out:
857 btrfs_free_path(path);
858 if (ret)
859 btrfs_abort_transaction(trans, ret);
860 return ret;
861 }
862
add_free_space_extent(struct btrfs_trans_handle * trans,struct btrfs_block_group * block_group,struct btrfs_path * path,u64 start,u64 size)863 static int add_free_space_extent(struct btrfs_trans_handle *trans,
864 struct btrfs_block_group *block_group,
865 struct btrfs_path *path,
866 u64 start, u64 size)
867 {
868 struct btrfs_root *root = btrfs_free_space_root(block_group);
869 struct btrfs_key key, new_key;
870 u64 found_start, found_end;
871 u64 end = start + size;
872 int new_extents = 1;
873 int ret;
874
875 /*
876 * We are adding a new extent of free space, but we need to merge
877 * extents. There are four cases here:
878 *
879 * 1. The new extent does not have any immediate neighbors to merge
880 * with: add the new key and increment the free space extent count. We
881 * may need to convert the block group to bitmaps as a result.
882 * 2. The new extent has an immediate neighbor before it: remove the
883 * previous key and insert a new key combining both of them. There is no
884 * net change in the number of extents.
885 * 3. The new extent has an immediate neighbor after it: remove the next
886 * key and insert a new key combining both of them. There is no net
887 * change in the number of extents.
888 * 4. The new extent has immediate neighbors on both sides: remove both
889 * of the keys and insert a new key combining all of them. Where we used
890 * to have two extents, we now have one, so decrement the extent count.
891 */
892
893 new_key.objectid = start;
894 new_key.type = BTRFS_FREE_SPACE_EXTENT_KEY;
895 new_key.offset = size;
896
897 /* Search for a neighbor on the left. */
898 if (start == block_group->start)
899 goto right;
900 key.objectid = start - 1;
901 key.type = (u8)-1;
902 key.offset = (u64)-1;
903
904 ret = btrfs_search_prev_slot(trans, root, &key, path, -1, 1);
905 if (ret)
906 goto out;
907
908 btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
909
910 if (key.type != BTRFS_FREE_SPACE_EXTENT_KEY) {
911 ASSERT(key.type == BTRFS_FREE_SPACE_INFO_KEY);
912 btrfs_release_path(path);
913 goto right;
914 }
915
916 found_start = key.objectid;
917 found_end = key.objectid + key.offset;
918 ASSERT(found_start >= block_group->start &&
919 found_end > block_group->start);
920 ASSERT(found_start < start && found_end <= start);
921
922 /*
923 * Delete the neighbor on the left and absorb it into the new key (cases
924 * 2 and 4).
925 */
926 if (found_end == start) {
927 ret = btrfs_del_item(trans, root, path);
928 if (ret)
929 goto out;
930 new_key.objectid = found_start;
931 new_key.offset += key.offset;
932 new_extents--;
933 }
934 btrfs_release_path(path);
935
936 right:
937 /* Search for a neighbor on the right. */
938 if (end == block_group->start + block_group->length)
939 goto insert;
940 key.objectid = end;
941 key.type = (u8)-1;
942 key.offset = (u64)-1;
943
944 ret = btrfs_search_prev_slot(trans, root, &key, path, -1, 1);
945 if (ret)
946 goto out;
947
948 btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
949
950 if (key.type != BTRFS_FREE_SPACE_EXTENT_KEY) {
951 ASSERT(key.type == BTRFS_FREE_SPACE_INFO_KEY);
952 btrfs_release_path(path);
953 goto insert;
954 }
955
956 found_start = key.objectid;
957 found_end = key.objectid + key.offset;
958 ASSERT(found_start >= block_group->start &&
959 found_end > block_group->start);
960 ASSERT((found_start < start && found_end <= start) ||
961 (found_start >= end && found_end > end));
962
963 /*
964 * Delete the neighbor on the right and absorb it into the new key
965 * (cases 3 and 4).
966 */
967 if (found_start == end) {
968 ret = btrfs_del_item(trans, root, path);
969 if (ret)
970 goto out;
971 new_key.offset += key.offset;
972 new_extents--;
973 }
974 btrfs_release_path(path);
975
976 insert:
977 /* Insert the new key (cases 1-4). */
978 ret = btrfs_insert_empty_item(trans, root, path, &new_key, 0);
979 if (ret)
980 goto out;
981
982 btrfs_release_path(path);
983 ret = update_free_space_extent_count(trans, block_group, path,
984 new_extents);
985
986 out:
987 return ret;
988 }
989
990 EXPORT_FOR_TESTS
__add_to_free_space_tree(struct btrfs_trans_handle * trans,struct btrfs_block_group * block_group,struct btrfs_path * path,u64 start,u64 size)991 int __add_to_free_space_tree(struct btrfs_trans_handle *trans,
992 struct btrfs_block_group *block_group,
993 struct btrfs_path *path, u64 start, u64 size)
994 {
995 struct btrfs_free_space_info *info;
996 u32 flags;
997 int ret;
998
999 if (block_group->needs_free_space) {
1000 ret = __add_block_group_free_space(trans, block_group, path);
1001 if (ret)
1002 return ret;
1003 }
1004
1005 info = search_free_space_info(NULL, block_group, path, 0);
1006 if (IS_ERR(info))
1007 return PTR_ERR(info);
1008 flags = btrfs_free_space_flags(path->nodes[0], info);
1009 btrfs_release_path(path);
1010
1011 if (flags & BTRFS_FREE_SPACE_USING_BITMAPS) {
1012 return modify_free_space_bitmap(trans, block_group, path,
1013 start, size, 0);
1014 } else {
1015 return add_free_space_extent(trans, block_group, path, start,
1016 size);
1017 }
1018 }
1019
add_to_free_space_tree(struct btrfs_trans_handle * trans,u64 start,u64 size)1020 int add_to_free_space_tree(struct btrfs_trans_handle *trans,
1021 u64 start, u64 size)
1022 {
1023 struct btrfs_block_group *block_group;
1024 struct btrfs_path *path;
1025 int ret;
1026
1027 if (!btrfs_fs_compat_ro(trans->fs_info, FREE_SPACE_TREE))
1028 return 0;
1029
1030 path = btrfs_alloc_path();
1031 if (!path) {
1032 ret = -ENOMEM;
1033 goto out;
1034 }
1035
1036 block_group = btrfs_lookup_block_group(trans->fs_info, start);
1037 if (!block_group) {
1038 ASSERT(0);
1039 ret = -ENOENT;
1040 goto out;
1041 }
1042
1043 mutex_lock(&block_group->free_space_lock);
1044 ret = __add_to_free_space_tree(trans, block_group, path, start, size);
1045 mutex_unlock(&block_group->free_space_lock);
1046
1047 btrfs_put_block_group(block_group);
1048 out:
1049 btrfs_free_path(path);
1050 if (ret)
1051 btrfs_abort_transaction(trans, ret);
1052 return ret;
1053 }
1054
1055 /*
1056 * Populate the free space tree by walking the extent tree. Operations on the
1057 * extent tree that happen as a result of writes to the free space tree will go
1058 * through the normal add/remove hooks.
1059 */
populate_free_space_tree(struct btrfs_trans_handle * trans,struct btrfs_block_group * block_group)1060 static int populate_free_space_tree(struct btrfs_trans_handle *trans,
1061 struct btrfs_block_group *block_group)
1062 {
1063 struct btrfs_root *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 = READA_FORWARD;
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, 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->start;
1094 key.type = BTRFS_EXTENT_ITEM_KEY;
1095 key.offset = 0;
1096
1097 extent_root = btrfs_extent_root(trans->fs_info, key.objectid);
1098 ret = btrfs_search_slot_for_read(extent_root, &key, path, 1, 0);
1099 if (ret < 0)
1100 goto out_locked;
1101 ASSERT(ret == 0);
1102
1103 start = block_group->start;
1104 end = block_group->start + block_group->length;
1105 while (1) {
1106 btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
1107
1108 if (key.type == BTRFS_EXTENT_ITEM_KEY ||
1109 key.type == BTRFS_METADATA_ITEM_KEY) {
1110 if (key.objectid >= end)
1111 break;
1112
1113 if (start < key.objectid) {
1114 ret = __add_to_free_space_tree(trans,
1115 block_group,
1116 path2, start,
1117 key.objectid -
1118 start);
1119 if (ret)
1120 goto out_locked;
1121 }
1122 start = key.objectid;
1123 if (key.type == BTRFS_METADATA_ITEM_KEY)
1124 start += trans->fs_info->nodesize;
1125 else
1126 start += key.offset;
1127 } else if (key.type == BTRFS_BLOCK_GROUP_ITEM_KEY) {
1128 if (key.objectid != block_group->start)
1129 break;
1130 }
1131
1132 ret = btrfs_next_item(extent_root, path);
1133 if (ret < 0)
1134 goto out_locked;
1135 if (ret)
1136 break;
1137 }
1138 if (start < end) {
1139 ret = __add_to_free_space_tree(trans, block_group, path2,
1140 start, end - start);
1141 if (ret)
1142 goto out_locked;
1143 }
1144
1145 ret = 0;
1146 out_locked:
1147 mutex_unlock(&block_group->free_space_lock);
1148 out:
1149 btrfs_free_path(path2);
1150 btrfs_free_path(path);
1151 return ret;
1152 }
1153
btrfs_create_free_space_tree(struct btrfs_fs_info * fs_info)1154 int btrfs_create_free_space_tree(struct btrfs_fs_info *fs_info)
1155 {
1156 struct btrfs_trans_handle *trans;
1157 struct btrfs_root *tree_root = fs_info->tree_root;
1158 struct btrfs_root *free_space_root;
1159 struct btrfs_block_group *block_group;
1160 struct rb_node *node;
1161 int ret;
1162
1163 trans = btrfs_start_transaction(tree_root, 0);
1164 if (IS_ERR(trans))
1165 return PTR_ERR(trans);
1166
1167 set_bit(BTRFS_FS_CREATING_FREE_SPACE_TREE, &fs_info->flags);
1168 set_bit(BTRFS_FS_FREE_SPACE_TREE_UNTRUSTED, &fs_info->flags);
1169 free_space_root = btrfs_create_tree(trans,
1170 BTRFS_FREE_SPACE_TREE_OBJECTID);
1171 if (IS_ERR(free_space_root)) {
1172 ret = PTR_ERR(free_space_root);
1173 goto abort;
1174 }
1175 ret = btrfs_global_root_insert(free_space_root);
1176 if (ret) {
1177 btrfs_put_root(free_space_root);
1178 goto abort;
1179 }
1180
1181 node = rb_first_cached(&fs_info->block_group_cache_tree);
1182 while (node) {
1183 block_group = rb_entry(node, struct btrfs_block_group,
1184 cache_node);
1185 ret = populate_free_space_tree(trans, block_group);
1186 if (ret)
1187 goto abort;
1188 node = rb_next(node);
1189 }
1190
1191 btrfs_set_fs_compat_ro(fs_info, FREE_SPACE_TREE);
1192 btrfs_set_fs_compat_ro(fs_info, FREE_SPACE_TREE_VALID);
1193 clear_bit(BTRFS_FS_CREATING_FREE_SPACE_TREE, &fs_info->flags);
1194 ret = btrfs_commit_transaction(trans);
1195
1196 /*
1197 * Now that we've committed the transaction any reading of our commit
1198 * root will be safe, so we can cache from the free space tree now.
1199 */
1200 clear_bit(BTRFS_FS_FREE_SPACE_TREE_UNTRUSTED, &fs_info->flags);
1201 return ret;
1202
1203 abort:
1204 clear_bit(BTRFS_FS_CREATING_FREE_SPACE_TREE, &fs_info->flags);
1205 clear_bit(BTRFS_FS_FREE_SPACE_TREE_UNTRUSTED, &fs_info->flags);
1206 btrfs_abort_transaction(trans, ret);
1207 btrfs_end_transaction(trans);
1208 return ret;
1209 }
1210
clear_free_space_tree(struct btrfs_trans_handle * trans,struct btrfs_root * root)1211 static int clear_free_space_tree(struct btrfs_trans_handle *trans,
1212 struct btrfs_root *root)
1213 {
1214 struct btrfs_path *path;
1215 struct btrfs_key key;
1216 int nr;
1217 int ret;
1218
1219 path = btrfs_alloc_path();
1220 if (!path)
1221 return -ENOMEM;
1222
1223 key.objectid = 0;
1224 key.type = 0;
1225 key.offset = 0;
1226
1227 while (1) {
1228 ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
1229 if (ret < 0)
1230 goto out;
1231
1232 nr = btrfs_header_nritems(path->nodes[0]);
1233 if (!nr)
1234 break;
1235
1236 path->slots[0] = 0;
1237 ret = btrfs_del_items(trans, root, path, 0, nr);
1238 if (ret)
1239 goto out;
1240
1241 btrfs_release_path(path);
1242 }
1243
1244 ret = 0;
1245 out:
1246 btrfs_free_path(path);
1247 return ret;
1248 }
1249
btrfs_clear_free_space_tree(struct btrfs_fs_info * fs_info)1250 int btrfs_clear_free_space_tree(struct btrfs_fs_info *fs_info)
1251 {
1252 struct btrfs_trans_handle *trans;
1253 struct btrfs_root *tree_root = fs_info->tree_root;
1254 struct btrfs_key key = {
1255 .objectid = BTRFS_FREE_SPACE_TREE_OBJECTID,
1256 .type = BTRFS_ROOT_ITEM_KEY,
1257 .offset = 0,
1258 };
1259 struct btrfs_root *free_space_root = btrfs_global_root(fs_info, &key);
1260 int ret;
1261
1262 trans = btrfs_start_transaction(tree_root, 0);
1263 if (IS_ERR(trans))
1264 return PTR_ERR(trans);
1265
1266 btrfs_clear_fs_compat_ro(fs_info, FREE_SPACE_TREE);
1267 btrfs_clear_fs_compat_ro(fs_info, FREE_SPACE_TREE_VALID);
1268
1269 ret = clear_free_space_tree(trans, free_space_root);
1270 if (ret)
1271 goto abort;
1272
1273 ret = btrfs_del_root(trans, &free_space_root->root_key);
1274 if (ret)
1275 goto abort;
1276
1277 btrfs_global_root_delete(free_space_root);
1278 list_del(&free_space_root->dirty_list);
1279
1280 btrfs_tree_lock(free_space_root->node);
1281 btrfs_clean_tree_block(free_space_root->node);
1282 btrfs_tree_unlock(free_space_root->node);
1283 btrfs_free_tree_block(trans, btrfs_root_id(free_space_root),
1284 free_space_root->node, 0, 1);
1285
1286 btrfs_put_root(free_space_root);
1287
1288 return btrfs_commit_transaction(trans);
1289
1290 abort:
1291 btrfs_abort_transaction(trans, ret);
1292 btrfs_end_transaction(trans);
1293 return ret;
1294 }
1295
__add_block_group_free_space(struct btrfs_trans_handle * trans,struct btrfs_block_group * block_group,struct btrfs_path * path)1296 static int __add_block_group_free_space(struct btrfs_trans_handle *trans,
1297 struct btrfs_block_group *block_group,
1298 struct btrfs_path *path)
1299 {
1300 int ret;
1301
1302 block_group->needs_free_space = 0;
1303
1304 ret = add_new_free_space_info(trans, block_group, path);
1305 if (ret)
1306 return ret;
1307
1308 return __add_to_free_space_tree(trans, block_group, path,
1309 block_group->start,
1310 block_group->length);
1311 }
1312
add_block_group_free_space(struct btrfs_trans_handle * trans,struct btrfs_block_group * block_group)1313 int add_block_group_free_space(struct btrfs_trans_handle *trans,
1314 struct btrfs_block_group *block_group)
1315 {
1316 struct btrfs_fs_info *fs_info = trans->fs_info;
1317 struct btrfs_path *path = NULL;
1318 int ret = 0;
1319
1320 if (!btrfs_fs_compat_ro(fs_info, FREE_SPACE_TREE))
1321 return 0;
1322
1323 mutex_lock(&block_group->free_space_lock);
1324 if (!block_group->needs_free_space)
1325 goto out;
1326
1327 path = btrfs_alloc_path();
1328 if (!path) {
1329 ret = -ENOMEM;
1330 goto out;
1331 }
1332
1333 ret = __add_block_group_free_space(trans, block_group, path);
1334
1335 out:
1336 btrfs_free_path(path);
1337 mutex_unlock(&block_group->free_space_lock);
1338 if (ret)
1339 btrfs_abort_transaction(trans, ret);
1340 return ret;
1341 }
1342
remove_block_group_free_space(struct btrfs_trans_handle * trans,struct btrfs_block_group * block_group)1343 int remove_block_group_free_space(struct btrfs_trans_handle *trans,
1344 struct btrfs_block_group *block_group)
1345 {
1346 struct btrfs_root *root = btrfs_free_space_root(block_group);
1347 struct btrfs_path *path;
1348 struct btrfs_key key, found_key;
1349 struct extent_buffer *leaf;
1350 u64 start, end;
1351 int done = 0, nr;
1352 int ret;
1353
1354 if (!btrfs_fs_compat_ro(trans->fs_info, FREE_SPACE_TREE))
1355 return 0;
1356
1357 if (block_group->needs_free_space) {
1358 /* We never added this block group to the free space tree. */
1359 return 0;
1360 }
1361
1362 path = btrfs_alloc_path();
1363 if (!path) {
1364 ret = -ENOMEM;
1365 goto out;
1366 }
1367
1368 start = block_group->start;
1369 end = block_group->start + block_group->length;
1370
1371 key.objectid = end - 1;
1372 key.type = (u8)-1;
1373 key.offset = (u64)-1;
1374
1375 while (!done) {
1376 ret = btrfs_search_prev_slot(trans, root, &key, path, -1, 1);
1377 if (ret)
1378 goto out;
1379
1380 leaf = path->nodes[0];
1381 nr = 0;
1382 path->slots[0]++;
1383 while (path->slots[0] > 0) {
1384 btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0] - 1);
1385
1386 if (found_key.type == BTRFS_FREE_SPACE_INFO_KEY) {
1387 ASSERT(found_key.objectid == block_group->start);
1388 ASSERT(found_key.offset == block_group->length);
1389 done = 1;
1390 nr++;
1391 path->slots[0]--;
1392 break;
1393 } else if (found_key.type == BTRFS_FREE_SPACE_EXTENT_KEY ||
1394 found_key.type == BTRFS_FREE_SPACE_BITMAP_KEY) {
1395 ASSERT(found_key.objectid >= start);
1396 ASSERT(found_key.objectid < end);
1397 ASSERT(found_key.objectid + found_key.offset <= end);
1398 nr++;
1399 path->slots[0]--;
1400 } else {
1401 ASSERT(0);
1402 }
1403 }
1404
1405 ret = btrfs_del_items(trans, root, path, path->slots[0], nr);
1406 if (ret)
1407 goto out;
1408 btrfs_release_path(path);
1409 }
1410
1411 ret = 0;
1412 out:
1413 btrfs_free_path(path);
1414 if (ret)
1415 btrfs_abort_transaction(trans, ret);
1416 return ret;
1417 }
1418
load_free_space_bitmaps(struct btrfs_caching_control * caching_ctl,struct btrfs_path * path,u32 expected_extent_count)1419 static int load_free_space_bitmaps(struct btrfs_caching_control *caching_ctl,
1420 struct btrfs_path *path,
1421 u32 expected_extent_count)
1422 {
1423 struct btrfs_block_group *block_group;
1424 struct btrfs_fs_info *fs_info;
1425 struct btrfs_root *root;
1426 struct btrfs_key key;
1427 int prev_bit = 0, bit;
1428 /* Initialize to silence GCC. */
1429 u64 extent_start = 0;
1430 u64 end, offset;
1431 u64 total_found = 0;
1432 u32 extent_count = 0;
1433 int ret;
1434
1435 block_group = caching_ctl->block_group;
1436 fs_info = block_group->fs_info;
1437 root = btrfs_free_space_root(block_group);
1438
1439 end = block_group->start + block_group->length;
1440
1441 while (1) {
1442 ret = btrfs_next_item(root, path);
1443 if (ret < 0)
1444 goto out;
1445 if (ret)
1446 break;
1447
1448 btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
1449
1450 if (key.type == BTRFS_FREE_SPACE_INFO_KEY)
1451 break;
1452
1453 ASSERT(key.type == BTRFS_FREE_SPACE_BITMAP_KEY);
1454 ASSERT(key.objectid < end && key.objectid + key.offset <= end);
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 extent_start,
1464 offset);
1465 if (total_found > CACHING_CTL_WAKE_UP) {
1466 total_found = 0;
1467 wake_up(&caching_ctl->wait);
1468 }
1469 extent_count++;
1470 }
1471 prev_bit = bit;
1472 offset += fs_info->sectorsize;
1473 }
1474 }
1475 if (prev_bit == 1) {
1476 total_found += add_new_free_space(block_group, extent_start,
1477 end);
1478 extent_count++;
1479 }
1480
1481 if (extent_count != expected_extent_count) {
1482 btrfs_err(fs_info,
1483 "incorrect extent count for %llu; counted %u, expected %u",
1484 block_group->start, extent_count,
1485 expected_extent_count);
1486 ASSERT(0);
1487 ret = -EIO;
1488 goto out;
1489 }
1490
1491 ret = 0;
1492 out:
1493 return ret;
1494 }
1495
load_free_space_extents(struct btrfs_caching_control * caching_ctl,struct btrfs_path * path,u32 expected_extent_count)1496 static int load_free_space_extents(struct btrfs_caching_control *caching_ctl,
1497 struct btrfs_path *path,
1498 u32 expected_extent_count)
1499 {
1500 struct btrfs_block_group *block_group;
1501 struct btrfs_fs_info *fs_info;
1502 struct btrfs_root *root;
1503 struct btrfs_key key;
1504 u64 end;
1505 u64 total_found = 0;
1506 u32 extent_count = 0;
1507 int ret;
1508
1509 block_group = caching_ctl->block_group;
1510 fs_info = block_group->fs_info;
1511 root = btrfs_free_space_root(block_group);
1512
1513 end = block_group->start + block_group->length;
1514
1515 while (1) {
1516 ret = btrfs_next_item(root, path);
1517 if (ret < 0)
1518 goto out;
1519 if (ret)
1520 break;
1521
1522 btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
1523
1524 if (key.type == BTRFS_FREE_SPACE_INFO_KEY)
1525 break;
1526
1527 ASSERT(key.type == BTRFS_FREE_SPACE_EXTENT_KEY);
1528 ASSERT(key.objectid < end && key.objectid + key.offset <= end);
1529
1530 total_found += add_new_free_space(block_group, key.objectid,
1531 key.objectid + key.offset);
1532 if (total_found > CACHING_CTL_WAKE_UP) {
1533 total_found = 0;
1534 wake_up(&caching_ctl->wait);
1535 }
1536 extent_count++;
1537 }
1538
1539 if (extent_count != expected_extent_count) {
1540 btrfs_err(fs_info,
1541 "incorrect extent count for %llu; counted %u, expected %u",
1542 block_group->start, extent_count,
1543 expected_extent_count);
1544 ASSERT(0);
1545 ret = -EIO;
1546 goto out;
1547 }
1548
1549 ret = 0;
1550 out:
1551 return ret;
1552 }
1553
load_free_space_tree(struct btrfs_caching_control * caching_ctl)1554 int load_free_space_tree(struct btrfs_caching_control *caching_ctl)
1555 {
1556 struct btrfs_block_group *block_group;
1557 struct btrfs_free_space_info *info;
1558 struct btrfs_path *path;
1559 u32 extent_count, flags;
1560 int ret;
1561
1562 block_group = caching_ctl->block_group;
1563
1564 path = btrfs_alloc_path();
1565 if (!path)
1566 return -ENOMEM;
1567
1568 /*
1569 * Just like caching_thread() doesn't want to deadlock on the extent
1570 * tree, we don't want to deadlock on the free space tree.
1571 */
1572 path->skip_locking = 1;
1573 path->search_commit_root = 1;
1574 path->reada = READA_FORWARD;
1575
1576 info = search_free_space_info(NULL, block_group, path, 0);
1577 if (IS_ERR(info)) {
1578 ret = PTR_ERR(info);
1579 goto out;
1580 }
1581 extent_count = btrfs_free_space_extent_count(path->nodes[0], info);
1582 flags = btrfs_free_space_flags(path->nodes[0], info);
1583
1584 /*
1585 * We left path pointing to the free space info item, so now
1586 * load_free_space_foo can just iterate through the free space tree from
1587 * there.
1588 */
1589 if (flags & BTRFS_FREE_SPACE_USING_BITMAPS)
1590 ret = load_free_space_bitmaps(caching_ctl, path, extent_count);
1591 else
1592 ret = load_free_space_extents(caching_ctl, path, extent_count);
1593
1594 out:
1595 btrfs_free_path(path);
1596 return ret;
1597 }
1598