btrfs: cleanup 64bit/32bit divs, provably bounded values
[deliverable/linux.git] / fs / btrfs / free-space-cache.c
... / ...
CommitLineData
1/*
2 * Copyright (C) 2008 Red Hat. All rights reserved.
3 *
4 * This program is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU General Public
6 * License v2 as published by the Free Software Foundation.
7 *
8 * This program is distributed in the hope that it will be useful,
9 * but WITHOUT ANY WARRANTY; without even the implied warranty of
10 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
11 * General Public License for more details.
12 *
13 * You should have received a copy of the GNU General Public
14 * License along with this program; if not, write to the
15 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
16 * Boston, MA 021110-1307, USA.
17 */
18
19#include <linux/pagemap.h>
20#include <linux/sched.h>
21#include <linux/slab.h>
22#include <linux/math64.h>
23#include <linux/ratelimit.h>
24#include "ctree.h"
25#include "free-space-cache.h"
26#include "transaction.h"
27#include "disk-io.h"
28#include "extent_io.h"
29#include "inode-map.h"
30#include "volumes.h"
31
32#define BITS_PER_BITMAP (PAGE_CACHE_SIZE * 8)
33#define MAX_CACHE_BYTES_PER_GIG (32 * 1024)
34
35struct btrfs_trim_range {
36 u64 start;
37 u64 bytes;
38 struct list_head list;
39};
40
41static int link_free_space(struct btrfs_free_space_ctl *ctl,
42 struct btrfs_free_space *info);
43static void unlink_free_space(struct btrfs_free_space_ctl *ctl,
44 struct btrfs_free_space *info);
45
46static struct inode *__lookup_free_space_inode(struct btrfs_root *root,
47 struct btrfs_path *path,
48 u64 offset)
49{
50 struct btrfs_key key;
51 struct btrfs_key location;
52 struct btrfs_disk_key disk_key;
53 struct btrfs_free_space_header *header;
54 struct extent_buffer *leaf;
55 struct inode *inode = NULL;
56 int ret;
57
58 key.objectid = BTRFS_FREE_SPACE_OBJECTID;
59 key.offset = offset;
60 key.type = 0;
61
62 ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
63 if (ret < 0)
64 return ERR_PTR(ret);
65 if (ret > 0) {
66 btrfs_release_path(path);
67 return ERR_PTR(-ENOENT);
68 }
69
70 leaf = path->nodes[0];
71 header = btrfs_item_ptr(leaf, path->slots[0],
72 struct btrfs_free_space_header);
73 btrfs_free_space_key(leaf, header, &disk_key);
74 btrfs_disk_key_to_cpu(&location, &disk_key);
75 btrfs_release_path(path);
76
77 inode = btrfs_iget(root->fs_info->sb, &location, root, NULL);
78 if (!inode)
79 return ERR_PTR(-ENOENT);
80 if (IS_ERR(inode))
81 return inode;
82 if (is_bad_inode(inode)) {
83 iput(inode);
84 return ERR_PTR(-ENOENT);
85 }
86
87 mapping_set_gfp_mask(inode->i_mapping,
88 mapping_gfp_mask(inode->i_mapping) & ~__GFP_FS);
89
90 return inode;
91}
92
93struct inode *lookup_free_space_inode(struct btrfs_root *root,
94 struct btrfs_block_group_cache
95 *block_group, struct btrfs_path *path)
96{
97 struct inode *inode = NULL;
98 u32 flags = BTRFS_INODE_NODATASUM | BTRFS_INODE_NODATACOW;
99
100 spin_lock(&block_group->lock);
101 if (block_group->inode)
102 inode = igrab(block_group->inode);
103 spin_unlock(&block_group->lock);
104 if (inode)
105 return inode;
106
107 inode = __lookup_free_space_inode(root, path,
108 block_group->key.objectid);
109 if (IS_ERR(inode))
110 return inode;
111
112 spin_lock(&block_group->lock);
113 if (!((BTRFS_I(inode)->flags & flags) == flags)) {
114 btrfs_info(root->fs_info,
115 "Old style space inode found, converting.");
116 BTRFS_I(inode)->flags |= BTRFS_INODE_NODATASUM |
117 BTRFS_INODE_NODATACOW;
118 block_group->disk_cache_state = BTRFS_DC_CLEAR;
119 }
120
121 if (!block_group->iref) {
122 block_group->inode = igrab(inode);
123 block_group->iref = 1;
124 }
125 spin_unlock(&block_group->lock);
126
127 return inode;
128}
129
130static int __create_free_space_inode(struct btrfs_root *root,
131 struct btrfs_trans_handle *trans,
132 struct btrfs_path *path,
133 u64 ino, u64 offset)
134{
135 struct btrfs_key key;
136 struct btrfs_disk_key disk_key;
137 struct btrfs_free_space_header *header;
138 struct btrfs_inode_item *inode_item;
139 struct extent_buffer *leaf;
140 u64 flags = BTRFS_INODE_NOCOMPRESS | BTRFS_INODE_PREALLOC;
141 int ret;
142
143 ret = btrfs_insert_empty_inode(trans, root, path, ino);
144 if (ret)
145 return ret;
146
147 /* We inline crc's for the free disk space cache */
148 if (ino != BTRFS_FREE_INO_OBJECTID)
149 flags |= BTRFS_INODE_NODATASUM | BTRFS_INODE_NODATACOW;
150
151 leaf = path->nodes[0];
152 inode_item = btrfs_item_ptr(leaf, path->slots[0],
153 struct btrfs_inode_item);
154 btrfs_item_key(leaf, &disk_key, path->slots[0]);
155 memset_extent_buffer(leaf, 0, (unsigned long)inode_item,
156 sizeof(*inode_item));
157 btrfs_set_inode_generation(leaf, inode_item, trans->transid);
158 btrfs_set_inode_size(leaf, inode_item, 0);
159 btrfs_set_inode_nbytes(leaf, inode_item, 0);
160 btrfs_set_inode_uid(leaf, inode_item, 0);
161 btrfs_set_inode_gid(leaf, inode_item, 0);
162 btrfs_set_inode_mode(leaf, inode_item, S_IFREG | 0600);
163 btrfs_set_inode_flags(leaf, inode_item, flags);
164 btrfs_set_inode_nlink(leaf, inode_item, 1);
165 btrfs_set_inode_transid(leaf, inode_item, trans->transid);
166 btrfs_set_inode_block_group(leaf, inode_item, offset);
167 btrfs_mark_buffer_dirty(leaf);
168 btrfs_release_path(path);
169
170 key.objectid = BTRFS_FREE_SPACE_OBJECTID;
171 key.offset = offset;
172 key.type = 0;
173
174 ret = btrfs_insert_empty_item(trans, root, path, &key,
175 sizeof(struct btrfs_free_space_header));
176 if (ret < 0) {
177 btrfs_release_path(path);
178 return ret;
179 }
180 leaf = path->nodes[0];
181 header = btrfs_item_ptr(leaf, path->slots[0],
182 struct btrfs_free_space_header);
183 memset_extent_buffer(leaf, 0, (unsigned long)header, sizeof(*header));
184 btrfs_set_free_space_key(leaf, header, &disk_key);
185 btrfs_mark_buffer_dirty(leaf);
186 btrfs_release_path(path);
187
188 return 0;
189}
190
191int create_free_space_inode(struct btrfs_root *root,
192 struct btrfs_trans_handle *trans,
193 struct btrfs_block_group_cache *block_group,
194 struct btrfs_path *path)
195{
196 int ret;
197 u64 ino;
198
199 ret = btrfs_find_free_objectid(root, &ino);
200 if (ret < 0)
201 return ret;
202
203 return __create_free_space_inode(root, trans, path, ino,
204 block_group->key.objectid);
205}
206
207int btrfs_check_trunc_cache_free_space(struct btrfs_root *root,
208 struct btrfs_block_rsv *rsv)
209{
210 u64 needed_bytes;
211 int ret;
212
213 /* 1 for slack space, 1 for updating the inode */
214 needed_bytes = btrfs_calc_trunc_metadata_size(root, 1) +
215 btrfs_calc_trans_metadata_size(root, 1);
216
217 spin_lock(&rsv->lock);
218 if (rsv->reserved < needed_bytes)
219 ret = -ENOSPC;
220 else
221 ret = 0;
222 spin_unlock(&rsv->lock);
223 return ret;
224}
225
226int btrfs_truncate_free_space_cache(struct btrfs_root *root,
227 struct btrfs_trans_handle *trans,
228 struct inode *inode)
229{
230 int ret = 0;
231
232 btrfs_i_size_write(inode, 0);
233 truncate_pagecache(inode, 0);
234
235 /*
236 * We don't need an orphan item because truncating the free space cache
237 * will never be split across transactions.
238 */
239 ret = btrfs_truncate_inode_items(trans, root, inode,
240 0, BTRFS_EXTENT_DATA_KEY);
241 if (ret) {
242 btrfs_abort_transaction(trans, root, ret);
243 return ret;
244 }
245
246 ret = btrfs_update_inode(trans, root, inode);
247 if (ret)
248 btrfs_abort_transaction(trans, root, ret);
249
250 return ret;
251}
252
253static int readahead_cache(struct inode *inode)
254{
255 struct file_ra_state *ra;
256 unsigned long last_index;
257
258 ra = kzalloc(sizeof(*ra), GFP_NOFS);
259 if (!ra)
260 return -ENOMEM;
261
262 file_ra_state_init(ra, inode->i_mapping);
263 last_index = (i_size_read(inode) - 1) >> PAGE_CACHE_SHIFT;
264
265 page_cache_sync_readahead(inode->i_mapping, ra, NULL, 0, last_index);
266
267 kfree(ra);
268
269 return 0;
270}
271
272struct io_ctl {
273 void *cur, *orig;
274 struct page *page;
275 struct page **pages;
276 struct btrfs_root *root;
277 unsigned long size;
278 int index;
279 int num_pages;
280 unsigned check_crcs:1;
281};
282
283static int io_ctl_init(struct io_ctl *io_ctl, struct inode *inode,
284 struct btrfs_root *root, int write)
285{
286 int num_pages;
287 int check_crcs = 0;
288
289 num_pages = DIV_ROUND_UP(i_size_read(inode), PAGE_CACHE_SIZE);
290
291 if (btrfs_ino(inode) != BTRFS_FREE_INO_OBJECTID)
292 check_crcs = 1;
293
294 /* Make sure we can fit our crcs into the first page */
295 if (write && check_crcs &&
296 (num_pages * sizeof(u32)) >= PAGE_CACHE_SIZE)
297 return -ENOSPC;
298
299 memset(io_ctl, 0, sizeof(struct io_ctl));
300
301 io_ctl->pages = kcalloc(num_pages, sizeof(struct page *), GFP_NOFS);
302 if (!io_ctl->pages)
303 return -ENOMEM;
304
305 io_ctl->num_pages = num_pages;
306 io_ctl->root = root;
307 io_ctl->check_crcs = check_crcs;
308
309 return 0;
310}
311
312static void io_ctl_free(struct io_ctl *io_ctl)
313{
314 kfree(io_ctl->pages);
315}
316
317static void io_ctl_unmap_page(struct io_ctl *io_ctl)
318{
319 if (io_ctl->cur) {
320 kunmap(io_ctl->page);
321 io_ctl->cur = NULL;
322 io_ctl->orig = NULL;
323 }
324}
325
326static void io_ctl_map_page(struct io_ctl *io_ctl, int clear)
327{
328 ASSERT(io_ctl->index < io_ctl->num_pages);
329 io_ctl->page = io_ctl->pages[io_ctl->index++];
330 io_ctl->cur = kmap(io_ctl->page);
331 io_ctl->orig = io_ctl->cur;
332 io_ctl->size = PAGE_CACHE_SIZE;
333 if (clear)
334 memset(io_ctl->cur, 0, PAGE_CACHE_SIZE);
335}
336
337static void io_ctl_drop_pages(struct io_ctl *io_ctl)
338{
339 int i;
340
341 io_ctl_unmap_page(io_ctl);
342
343 for (i = 0; i < io_ctl->num_pages; i++) {
344 if (io_ctl->pages[i]) {
345 ClearPageChecked(io_ctl->pages[i]);
346 unlock_page(io_ctl->pages[i]);
347 page_cache_release(io_ctl->pages[i]);
348 }
349 }
350}
351
352static int io_ctl_prepare_pages(struct io_ctl *io_ctl, struct inode *inode,
353 int uptodate)
354{
355 struct page *page;
356 gfp_t mask = btrfs_alloc_write_mask(inode->i_mapping);
357 int i;
358
359 for (i = 0; i < io_ctl->num_pages; i++) {
360 page = find_or_create_page(inode->i_mapping, i, mask);
361 if (!page) {
362 io_ctl_drop_pages(io_ctl);
363 return -ENOMEM;
364 }
365 io_ctl->pages[i] = page;
366 if (uptodate && !PageUptodate(page)) {
367 btrfs_readpage(NULL, page);
368 lock_page(page);
369 if (!PageUptodate(page)) {
370 btrfs_err(BTRFS_I(inode)->root->fs_info,
371 "error reading free space cache");
372 io_ctl_drop_pages(io_ctl);
373 return -EIO;
374 }
375 }
376 }
377
378 for (i = 0; i < io_ctl->num_pages; i++) {
379 clear_page_dirty_for_io(io_ctl->pages[i]);
380 set_page_extent_mapped(io_ctl->pages[i]);
381 }
382
383 return 0;
384}
385
386static void io_ctl_set_generation(struct io_ctl *io_ctl, u64 generation)
387{
388 __le64 *val;
389
390 io_ctl_map_page(io_ctl, 1);
391
392 /*
393 * Skip the csum areas. If we don't check crcs then we just have a
394 * 64bit chunk at the front of the first page.
395 */
396 if (io_ctl->check_crcs) {
397 io_ctl->cur += (sizeof(u32) * io_ctl->num_pages);
398 io_ctl->size -= sizeof(u64) + (sizeof(u32) * io_ctl->num_pages);
399 } else {
400 io_ctl->cur += sizeof(u64);
401 io_ctl->size -= sizeof(u64) * 2;
402 }
403
404 val = io_ctl->cur;
405 *val = cpu_to_le64(generation);
406 io_ctl->cur += sizeof(u64);
407}
408
409static int io_ctl_check_generation(struct io_ctl *io_ctl, u64 generation)
410{
411 __le64 *gen;
412
413 /*
414 * Skip the crc area. If we don't check crcs then we just have a 64bit
415 * chunk at the front of the first page.
416 */
417 if (io_ctl->check_crcs) {
418 io_ctl->cur += sizeof(u32) * io_ctl->num_pages;
419 io_ctl->size -= sizeof(u64) +
420 (sizeof(u32) * io_ctl->num_pages);
421 } else {
422 io_ctl->cur += sizeof(u64);
423 io_ctl->size -= sizeof(u64) * 2;
424 }
425
426 gen = io_ctl->cur;
427 if (le64_to_cpu(*gen) != generation) {
428 printk_ratelimited(KERN_ERR "BTRFS: space cache generation "
429 "(%Lu) does not match inode (%Lu)\n", *gen,
430 generation);
431 io_ctl_unmap_page(io_ctl);
432 return -EIO;
433 }
434 io_ctl->cur += sizeof(u64);
435 return 0;
436}
437
438static void io_ctl_set_crc(struct io_ctl *io_ctl, int index)
439{
440 u32 *tmp;
441 u32 crc = ~(u32)0;
442 unsigned offset = 0;
443
444 if (!io_ctl->check_crcs) {
445 io_ctl_unmap_page(io_ctl);
446 return;
447 }
448
449 if (index == 0)
450 offset = sizeof(u32) * io_ctl->num_pages;
451
452 crc = btrfs_csum_data(io_ctl->orig + offset, crc,
453 PAGE_CACHE_SIZE - offset);
454 btrfs_csum_final(crc, (char *)&crc);
455 io_ctl_unmap_page(io_ctl);
456 tmp = kmap(io_ctl->pages[0]);
457 tmp += index;
458 *tmp = crc;
459 kunmap(io_ctl->pages[0]);
460}
461
462static int io_ctl_check_crc(struct io_ctl *io_ctl, int index)
463{
464 u32 *tmp, val;
465 u32 crc = ~(u32)0;
466 unsigned offset = 0;
467
468 if (!io_ctl->check_crcs) {
469 io_ctl_map_page(io_ctl, 0);
470 return 0;
471 }
472
473 if (index == 0)
474 offset = sizeof(u32) * io_ctl->num_pages;
475
476 tmp = kmap(io_ctl->pages[0]);
477 tmp += index;
478 val = *tmp;
479 kunmap(io_ctl->pages[0]);
480
481 io_ctl_map_page(io_ctl, 0);
482 crc = btrfs_csum_data(io_ctl->orig + offset, crc,
483 PAGE_CACHE_SIZE - offset);
484 btrfs_csum_final(crc, (char *)&crc);
485 if (val != crc) {
486 printk_ratelimited(KERN_ERR "BTRFS: csum mismatch on free "
487 "space cache\n");
488 io_ctl_unmap_page(io_ctl);
489 return -EIO;
490 }
491
492 return 0;
493}
494
495static int io_ctl_add_entry(struct io_ctl *io_ctl, u64 offset, u64 bytes,
496 void *bitmap)
497{
498 struct btrfs_free_space_entry *entry;
499
500 if (!io_ctl->cur)
501 return -ENOSPC;
502
503 entry = io_ctl->cur;
504 entry->offset = cpu_to_le64(offset);
505 entry->bytes = cpu_to_le64(bytes);
506 entry->type = (bitmap) ? BTRFS_FREE_SPACE_BITMAP :
507 BTRFS_FREE_SPACE_EXTENT;
508 io_ctl->cur += sizeof(struct btrfs_free_space_entry);
509 io_ctl->size -= sizeof(struct btrfs_free_space_entry);
510
511 if (io_ctl->size >= sizeof(struct btrfs_free_space_entry))
512 return 0;
513
514 io_ctl_set_crc(io_ctl, io_ctl->index - 1);
515
516 /* No more pages to map */
517 if (io_ctl->index >= io_ctl->num_pages)
518 return 0;
519
520 /* map the next page */
521 io_ctl_map_page(io_ctl, 1);
522 return 0;
523}
524
525static int io_ctl_add_bitmap(struct io_ctl *io_ctl, void *bitmap)
526{
527 if (!io_ctl->cur)
528 return -ENOSPC;
529
530 /*
531 * If we aren't at the start of the current page, unmap this one and
532 * map the next one if there is any left.
533 */
534 if (io_ctl->cur != io_ctl->orig) {
535 io_ctl_set_crc(io_ctl, io_ctl->index - 1);
536 if (io_ctl->index >= io_ctl->num_pages)
537 return -ENOSPC;
538 io_ctl_map_page(io_ctl, 0);
539 }
540
541 memcpy(io_ctl->cur, bitmap, PAGE_CACHE_SIZE);
542 io_ctl_set_crc(io_ctl, io_ctl->index - 1);
543 if (io_ctl->index < io_ctl->num_pages)
544 io_ctl_map_page(io_ctl, 0);
545 return 0;
546}
547
548static void io_ctl_zero_remaining_pages(struct io_ctl *io_ctl)
549{
550 /*
551 * If we're not on the boundary we know we've modified the page and we
552 * need to crc the page.
553 */
554 if (io_ctl->cur != io_ctl->orig)
555 io_ctl_set_crc(io_ctl, io_ctl->index - 1);
556 else
557 io_ctl_unmap_page(io_ctl);
558
559 while (io_ctl->index < io_ctl->num_pages) {
560 io_ctl_map_page(io_ctl, 1);
561 io_ctl_set_crc(io_ctl, io_ctl->index - 1);
562 }
563}
564
565static int io_ctl_read_entry(struct io_ctl *io_ctl,
566 struct btrfs_free_space *entry, u8 *type)
567{
568 struct btrfs_free_space_entry *e;
569 int ret;
570
571 if (!io_ctl->cur) {
572 ret = io_ctl_check_crc(io_ctl, io_ctl->index);
573 if (ret)
574 return ret;
575 }
576
577 e = io_ctl->cur;
578 entry->offset = le64_to_cpu(e->offset);
579 entry->bytes = le64_to_cpu(e->bytes);
580 *type = e->type;
581 io_ctl->cur += sizeof(struct btrfs_free_space_entry);
582 io_ctl->size -= sizeof(struct btrfs_free_space_entry);
583
584 if (io_ctl->size >= sizeof(struct btrfs_free_space_entry))
585 return 0;
586
587 io_ctl_unmap_page(io_ctl);
588
589 return 0;
590}
591
592static int io_ctl_read_bitmap(struct io_ctl *io_ctl,
593 struct btrfs_free_space *entry)
594{
595 int ret;
596
597 ret = io_ctl_check_crc(io_ctl, io_ctl->index);
598 if (ret)
599 return ret;
600
601 memcpy(entry->bitmap, io_ctl->cur, PAGE_CACHE_SIZE);
602 io_ctl_unmap_page(io_ctl);
603
604 return 0;
605}
606
607/*
608 * Since we attach pinned extents after the fact we can have contiguous sections
609 * of free space that are split up in entries. This poses a problem with the
610 * tree logging stuff since it could have allocated across what appears to be 2
611 * entries since we would have merged the entries when adding the pinned extents
612 * back to the free space cache. So run through the space cache that we just
613 * loaded and merge contiguous entries. This will make the log replay stuff not
614 * blow up and it will make for nicer allocator behavior.
615 */
616static void merge_space_tree(struct btrfs_free_space_ctl *ctl)
617{
618 struct btrfs_free_space *e, *prev = NULL;
619 struct rb_node *n;
620
621again:
622 spin_lock(&ctl->tree_lock);
623 for (n = rb_first(&ctl->free_space_offset); n; n = rb_next(n)) {
624 e = rb_entry(n, struct btrfs_free_space, offset_index);
625 if (!prev)
626 goto next;
627 if (e->bitmap || prev->bitmap)
628 goto next;
629 if (prev->offset + prev->bytes == e->offset) {
630 unlink_free_space(ctl, prev);
631 unlink_free_space(ctl, e);
632 prev->bytes += e->bytes;
633 kmem_cache_free(btrfs_free_space_cachep, e);
634 link_free_space(ctl, prev);
635 prev = NULL;
636 spin_unlock(&ctl->tree_lock);
637 goto again;
638 }
639next:
640 prev = e;
641 }
642 spin_unlock(&ctl->tree_lock);
643}
644
645static int __load_free_space_cache(struct btrfs_root *root, struct inode *inode,
646 struct btrfs_free_space_ctl *ctl,
647 struct btrfs_path *path, u64 offset)
648{
649 struct btrfs_free_space_header *header;
650 struct extent_buffer *leaf;
651 struct io_ctl io_ctl;
652 struct btrfs_key key;
653 struct btrfs_free_space *e, *n;
654 LIST_HEAD(bitmaps);
655 u64 num_entries;
656 u64 num_bitmaps;
657 u64 generation;
658 u8 type;
659 int ret = 0;
660
661 /* Nothing in the space cache, goodbye */
662 if (!i_size_read(inode))
663 return 0;
664
665 key.objectid = BTRFS_FREE_SPACE_OBJECTID;
666 key.offset = offset;
667 key.type = 0;
668
669 ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
670 if (ret < 0)
671 return 0;
672 else if (ret > 0) {
673 btrfs_release_path(path);
674 return 0;
675 }
676
677 ret = -1;
678
679 leaf = path->nodes[0];
680 header = btrfs_item_ptr(leaf, path->slots[0],
681 struct btrfs_free_space_header);
682 num_entries = btrfs_free_space_entries(leaf, header);
683 num_bitmaps = btrfs_free_space_bitmaps(leaf, header);
684 generation = btrfs_free_space_generation(leaf, header);
685 btrfs_release_path(path);
686
687 if (!BTRFS_I(inode)->generation) {
688 btrfs_info(root->fs_info,
689 "The free space cache file (%llu) is invalid. skip it\n",
690 offset);
691 return 0;
692 }
693
694 if (BTRFS_I(inode)->generation != generation) {
695 btrfs_err(root->fs_info,
696 "free space inode generation (%llu) "
697 "did not match free space cache generation (%llu)",
698 BTRFS_I(inode)->generation, generation);
699 return 0;
700 }
701
702 if (!num_entries)
703 return 0;
704
705 ret = io_ctl_init(&io_ctl, inode, root, 0);
706 if (ret)
707 return ret;
708
709 ret = readahead_cache(inode);
710 if (ret)
711 goto out;
712
713 ret = io_ctl_prepare_pages(&io_ctl, inode, 1);
714 if (ret)
715 goto out;
716
717 ret = io_ctl_check_crc(&io_ctl, 0);
718 if (ret)
719 goto free_cache;
720
721 ret = io_ctl_check_generation(&io_ctl, generation);
722 if (ret)
723 goto free_cache;
724
725 while (num_entries) {
726 e = kmem_cache_zalloc(btrfs_free_space_cachep,
727 GFP_NOFS);
728 if (!e)
729 goto free_cache;
730
731 ret = io_ctl_read_entry(&io_ctl, e, &type);
732 if (ret) {
733 kmem_cache_free(btrfs_free_space_cachep, e);
734 goto free_cache;
735 }
736
737 if (!e->bytes) {
738 kmem_cache_free(btrfs_free_space_cachep, e);
739 goto free_cache;
740 }
741
742 if (type == BTRFS_FREE_SPACE_EXTENT) {
743 spin_lock(&ctl->tree_lock);
744 ret = link_free_space(ctl, e);
745 spin_unlock(&ctl->tree_lock);
746 if (ret) {
747 btrfs_err(root->fs_info,
748 "Duplicate entries in free space cache, dumping");
749 kmem_cache_free(btrfs_free_space_cachep, e);
750 goto free_cache;
751 }
752 } else {
753 ASSERT(num_bitmaps);
754 num_bitmaps--;
755 e->bitmap = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS);
756 if (!e->bitmap) {
757 kmem_cache_free(
758 btrfs_free_space_cachep, e);
759 goto free_cache;
760 }
761 spin_lock(&ctl->tree_lock);
762 ret = link_free_space(ctl, e);
763 ctl->total_bitmaps++;
764 ctl->op->recalc_thresholds(ctl);
765 spin_unlock(&ctl->tree_lock);
766 if (ret) {
767 btrfs_err(root->fs_info,
768 "Duplicate entries in free space cache, dumping");
769 kmem_cache_free(btrfs_free_space_cachep, e);
770 goto free_cache;
771 }
772 list_add_tail(&e->list, &bitmaps);
773 }
774
775 num_entries--;
776 }
777
778 io_ctl_unmap_page(&io_ctl);
779
780 /*
781 * We add the bitmaps at the end of the entries in order that
782 * the bitmap entries are added to the cache.
783 */
784 list_for_each_entry_safe(e, n, &bitmaps, list) {
785 list_del_init(&e->list);
786 ret = io_ctl_read_bitmap(&io_ctl, e);
787 if (ret)
788 goto free_cache;
789 }
790
791 io_ctl_drop_pages(&io_ctl);
792 merge_space_tree(ctl);
793 ret = 1;
794out:
795 io_ctl_free(&io_ctl);
796 return ret;
797free_cache:
798 io_ctl_drop_pages(&io_ctl);
799 __btrfs_remove_free_space_cache(ctl);
800 goto out;
801}
802
803int load_free_space_cache(struct btrfs_fs_info *fs_info,
804 struct btrfs_block_group_cache *block_group)
805{
806 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
807 struct btrfs_root *root = fs_info->tree_root;
808 struct inode *inode;
809 struct btrfs_path *path;
810 int ret = 0;
811 bool matched;
812 u64 used = btrfs_block_group_used(&block_group->item);
813
814 /*
815 * If this block group has been marked to be cleared for one reason or
816 * another then we can't trust the on disk cache, so just return.
817 */
818 spin_lock(&block_group->lock);
819 if (block_group->disk_cache_state != BTRFS_DC_WRITTEN) {
820 spin_unlock(&block_group->lock);
821 return 0;
822 }
823 spin_unlock(&block_group->lock);
824
825 path = btrfs_alloc_path();
826 if (!path)
827 return 0;
828 path->search_commit_root = 1;
829 path->skip_locking = 1;
830
831 inode = lookup_free_space_inode(root, block_group, path);
832 if (IS_ERR(inode)) {
833 btrfs_free_path(path);
834 return 0;
835 }
836
837 /* We may have converted the inode and made the cache invalid. */
838 spin_lock(&block_group->lock);
839 if (block_group->disk_cache_state != BTRFS_DC_WRITTEN) {
840 spin_unlock(&block_group->lock);
841 btrfs_free_path(path);
842 goto out;
843 }
844 spin_unlock(&block_group->lock);
845
846 ret = __load_free_space_cache(fs_info->tree_root, inode, ctl,
847 path, block_group->key.objectid);
848 btrfs_free_path(path);
849 if (ret <= 0)
850 goto out;
851
852 spin_lock(&ctl->tree_lock);
853 matched = (ctl->free_space == (block_group->key.offset - used -
854 block_group->bytes_super));
855 spin_unlock(&ctl->tree_lock);
856
857 if (!matched) {
858 __btrfs_remove_free_space_cache(ctl);
859 btrfs_warn(fs_info, "block group %llu has wrong amount of free space",
860 block_group->key.objectid);
861 ret = -1;
862 }
863out:
864 if (ret < 0) {
865 /* This cache is bogus, make sure it gets cleared */
866 spin_lock(&block_group->lock);
867 block_group->disk_cache_state = BTRFS_DC_CLEAR;
868 spin_unlock(&block_group->lock);
869 ret = 0;
870
871 btrfs_warn(fs_info, "failed to load free space cache for block group %llu, rebuild it now",
872 block_group->key.objectid);
873 }
874
875 iput(inode);
876 return ret;
877}
878
879static noinline_for_stack
880int write_cache_extent_entries(struct io_ctl *io_ctl,
881 struct btrfs_free_space_ctl *ctl,
882 struct btrfs_block_group_cache *block_group,
883 int *entries, int *bitmaps,
884 struct list_head *bitmap_list)
885{
886 int ret;
887 struct btrfs_free_cluster *cluster = NULL;
888 struct rb_node *node = rb_first(&ctl->free_space_offset);
889 struct btrfs_trim_range *trim_entry;
890
891 /* Get the cluster for this block_group if it exists */
892 if (block_group && !list_empty(&block_group->cluster_list)) {
893 cluster = list_entry(block_group->cluster_list.next,
894 struct btrfs_free_cluster,
895 block_group_list);
896 }
897
898 if (!node && cluster) {
899 node = rb_first(&cluster->root);
900 cluster = NULL;
901 }
902
903 /* Write out the extent entries */
904 while (node) {
905 struct btrfs_free_space *e;
906
907 e = rb_entry(node, struct btrfs_free_space, offset_index);
908 *entries += 1;
909
910 ret = io_ctl_add_entry(io_ctl, e->offset, e->bytes,
911 e->bitmap);
912 if (ret)
913 goto fail;
914
915 if (e->bitmap) {
916 list_add_tail(&e->list, bitmap_list);
917 *bitmaps += 1;
918 }
919 node = rb_next(node);
920 if (!node && cluster) {
921 node = rb_first(&cluster->root);
922 cluster = NULL;
923 }
924 }
925
926 /*
927 * Make sure we don't miss any range that was removed from our rbtree
928 * because trimming is running. Otherwise after a umount+mount (or crash
929 * after committing the transaction) we would leak free space and get
930 * an inconsistent free space cache report from fsck.
931 */
932 list_for_each_entry(trim_entry, &ctl->trimming_ranges, list) {
933 ret = io_ctl_add_entry(io_ctl, trim_entry->start,
934 trim_entry->bytes, NULL);
935 if (ret)
936 goto fail;
937 *entries += 1;
938 }
939
940 return 0;
941fail:
942 return -ENOSPC;
943}
944
945static noinline_for_stack int
946update_cache_item(struct btrfs_trans_handle *trans,
947 struct btrfs_root *root,
948 struct inode *inode,
949 struct btrfs_path *path, u64 offset,
950 int entries, int bitmaps)
951{
952 struct btrfs_key key;
953 struct btrfs_free_space_header *header;
954 struct extent_buffer *leaf;
955 int ret;
956
957 key.objectid = BTRFS_FREE_SPACE_OBJECTID;
958 key.offset = offset;
959 key.type = 0;
960
961 ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
962 if (ret < 0) {
963 clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, inode->i_size - 1,
964 EXTENT_DIRTY | EXTENT_DELALLOC, 0, 0, NULL,
965 GFP_NOFS);
966 goto fail;
967 }
968 leaf = path->nodes[0];
969 if (ret > 0) {
970 struct btrfs_key found_key;
971 ASSERT(path->slots[0]);
972 path->slots[0]--;
973 btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
974 if (found_key.objectid != BTRFS_FREE_SPACE_OBJECTID ||
975 found_key.offset != offset) {
976 clear_extent_bit(&BTRFS_I(inode)->io_tree, 0,
977 inode->i_size - 1,
978 EXTENT_DIRTY | EXTENT_DELALLOC, 0, 0,
979 NULL, GFP_NOFS);
980 btrfs_release_path(path);
981 goto fail;
982 }
983 }
984
985 BTRFS_I(inode)->generation = trans->transid;
986 header = btrfs_item_ptr(leaf, path->slots[0],
987 struct btrfs_free_space_header);
988 btrfs_set_free_space_entries(leaf, header, entries);
989 btrfs_set_free_space_bitmaps(leaf, header, bitmaps);
990 btrfs_set_free_space_generation(leaf, header, trans->transid);
991 btrfs_mark_buffer_dirty(leaf);
992 btrfs_release_path(path);
993
994 return 0;
995
996fail:
997 return -1;
998}
999
1000static noinline_for_stack int
1001write_pinned_extent_entries(struct btrfs_root *root,
1002 struct btrfs_block_group_cache *block_group,
1003 struct io_ctl *io_ctl,
1004 int *entries)
1005{
1006 u64 start, extent_start, extent_end, len;
1007 struct extent_io_tree *unpin = NULL;
1008 int ret;
1009
1010 if (!block_group)
1011 return 0;
1012
1013 /*
1014 * We want to add any pinned extents to our free space cache
1015 * so we don't leak the space
1016 *
1017 * We shouldn't have switched the pinned extents yet so this is the
1018 * right one
1019 */
1020 unpin = root->fs_info->pinned_extents;
1021
1022 start = block_group->key.objectid;
1023
1024 while (start < block_group->key.objectid + block_group->key.offset) {
1025 ret = find_first_extent_bit(unpin, start,
1026 &extent_start, &extent_end,
1027 EXTENT_DIRTY, NULL);
1028 if (ret)
1029 return 0;
1030
1031 /* This pinned extent is out of our range */
1032 if (extent_start >= block_group->key.objectid +
1033 block_group->key.offset)
1034 return 0;
1035
1036 extent_start = max(extent_start, start);
1037 extent_end = min(block_group->key.objectid +
1038 block_group->key.offset, extent_end + 1);
1039 len = extent_end - extent_start;
1040
1041 *entries += 1;
1042 ret = io_ctl_add_entry(io_ctl, extent_start, len, NULL);
1043 if (ret)
1044 return -ENOSPC;
1045
1046 start = extent_end;
1047 }
1048
1049 return 0;
1050}
1051
1052static noinline_for_stack int
1053write_bitmap_entries(struct io_ctl *io_ctl, struct list_head *bitmap_list)
1054{
1055 struct list_head *pos, *n;
1056 int ret;
1057
1058 /* Write out the bitmaps */
1059 list_for_each_safe(pos, n, bitmap_list) {
1060 struct btrfs_free_space *entry =
1061 list_entry(pos, struct btrfs_free_space, list);
1062
1063 ret = io_ctl_add_bitmap(io_ctl, entry->bitmap);
1064 if (ret)
1065 return -ENOSPC;
1066 list_del_init(&entry->list);
1067 }
1068
1069 return 0;
1070}
1071
1072static int flush_dirty_cache(struct inode *inode)
1073{
1074 int ret;
1075
1076 ret = btrfs_wait_ordered_range(inode, 0, (u64)-1);
1077 if (ret)
1078 clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, inode->i_size - 1,
1079 EXTENT_DIRTY | EXTENT_DELALLOC, 0, 0, NULL,
1080 GFP_NOFS);
1081
1082 return ret;
1083}
1084
1085static void noinline_for_stack
1086cleanup_write_cache_enospc(struct inode *inode,
1087 struct io_ctl *io_ctl,
1088 struct extent_state **cached_state,
1089 struct list_head *bitmap_list)
1090{
1091 struct list_head *pos, *n;
1092
1093 list_for_each_safe(pos, n, bitmap_list) {
1094 struct btrfs_free_space *entry =
1095 list_entry(pos, struct btrfs_free_space, list);
1096 list_del_init(&entry->list);
1097 }
1098 io_ctl_drop_pages(io_ctl);
1099 unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0,
1100 i_size_read(inode) - 1, cached_state,
1101 GFP_NOFS);
1102}
1103
1104/**
1105 * __btrfs_write_out_cache - write out cached info to an inode
1106 * @root - the root the inode belongs to
1107 * @ctl - the free space cache we are going to write out
1108 * @block_group - the block_group for this cache if it belongs to a block_group
1109 * @trans - the trans handle
1110 * @path - the path to use
1111 * @offset - the offset for the key we'll insert
1112 *
1113 * This function writes out a free space cache struct to disk for quick recovery
1114 * on mount. This will return 0 if it was successfull in writing the cache out,
1115 * and -1 if it was not.
1116 */
1117static int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
1118 struct btrfs_free_space_ctl *ctl,
1119 struct btrfs_block_group_cache *block_group,
1120 struct btrfs_trans_handle *trans,
1121 struct btrfs_path *path, u64 offset)
1122{
1123 struct extent_state *cached_state = NULL;
1124 struct io_ctl io_ctl;
1125 LIST_HEAD(bitmap_list);
1126 int entries = 0;
1127 int bitmaps = 0;
1128 int ret;
1129
1130 if (!i_size_read(inode))
1131 return -1;
1132
1133 ret = io_ctl_init(&io_ctl, inode, root, 1);
1134 if (ret)
1135 return -1;
1136
1137 if (block_group && (block_group->flags & BTRFS_BLOCK_GROUP_DATA)) {
1138 down_write(&block_group->data_rwsem);
1139 spin_lock(&block_group->lock);
1140 if (block_group->delalloc_bytes) {
1141 block_group->disk_cache_state = BTRFS_DC_WRITTEN;
1142 spin_unlock(&block_group->lock);
1143 up_write(&block_group->data_rwsem);
1144 BTRFS_I(inode)->generation = 0;
1145 ret = 0;
1146 goto out;
1147 }
1148 spin_unlock(&block_group->lock);
1149 }
1150
1151 /* Lock all pages first so we can lock the extent safely. */
1152 io_ctl_prepare_pages(&io_ctl, inode, 0);
1153
1154 lock_extent_bits(&BTRFS_I(inode)->io_tree, 0, i_size_read(inode) - 1,
1155 0, &cached_state);
1156
1157 io_ctl_set_generation(&io_ctl, trans->transid);
1158
1159 mutex_lock(&ctl->cache_writeout_mutex);
1160 /* Write out the extent entries in the free space cache */
1161 ret = write_cache_extent_entries(&io_ctl, ctl,
1162 block_group, &entries, &bitmaps,
1163 &bitmap_list);
1164 if (ret) {
1165 mutex_unlock(&ctl->cache_writeout_mutex);
1166 goto out_nospc;
1167 }
1168
1169 /*
1170 * Some spaces that are freed in the current transaction are pinned,
1171 * they will be added into free space cache after the transaction is
1172 * committed, we shouldn't lose them.
1173 */
1174 ret = write_pinned_extent_entries(root, block_group, &io_ctl, &entries);
1175 if (ret) {
1176 mutex_unlock(&ctl->cache_writeout_mutex);
1177 goto out_nospc;
1178 }
1179
1180 /*
1181 * At last, we write out all the bitmaps and keep cache_writeout_mutex
1182 * locked while doing it because a concurrent trim can be manipulating
1183 * or freeing the bitmap.
1184 */
1185 ret = write_bitmap_entries(&io_ctl, &bitmap_list);
1186 mutex_unlock(&ctl->cache_writeout_mutex);
1187 if (ret)
1188 goto out_nospc;
1189
1190 /* Zero out the rest of the pages just to make sure */
1191 io_ctl_zero_remaining_pages(&io_ctl);
1192
1193 /* Everything is written out, now we dirty the pages in the file. */
1194 ret = btrfs_dirty_pages(root, inode, io_ctl.pages, io_ctl.num_pages,
1195 0, i_size_read(inode), &cached_state);
1196 if (ret)
1197 goto out_nospc;
1198
1199 if (block_group && (block_group->flags & BTRFS_BLOCK_GROUP_DATA))
1200 up_write(&block_group->data_rwsem);
1201 /*
1202 * Release the pages and unlock the extent, we will flush
1203 * them out later
1204 */
1205 io_ctl_drop_pages(&io_ctl);
1206
1207 unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0,
1208 i_size_read(inode) - 1, &cached_state, GFP_NOFS);
1209
1210 /* Flush the dirty pages in the cache file. */
1211 ret = flush_dirty_cache(inode);
1212 if (ret)
1213 goto out;
1214
1215 /* Update the cache item to tell everyone this cache file is valid. */
1216 ret = update_cache_item(trans, root, inode, path, offset,
1217 entries, bitmaps);
1218out:
1219 io_ctl_free(&io_ctl);
1220 if (ret) {
1221 invalidate_inode_pages2(inode->i_mapping);
1222 BTRFS_I(inode)->generation = 0;
1223 }
1224 btrfs_update_inode(trans, root, inode);
1225 return ret;
1226
1227out_nospc:
1228 cleanup_write_cache_enospc(inode, &io_ctl, &cached_state, &bitmap_list);
1229
1230 if (block_group && (block_group->flags & BTRFS_BLOCK_GROUP_DATA))
1231 up_write(&block_group->data_rwsem);
1232
1233 goto out;
1234}
1235
1236int btrfs_write_out_cache(struct btrfs_root *root,
1237 struct btrfs_trans_handle *trans,
1238 struct btrfs_block_group_cache *block_group,
1239 struct btrfs_path *path)
1240{
1241 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
1242 struct inode *inode;
1243 int ret = 0;
1244 enum btrfs_disk_cache_state dcs = BTRFS_DC_WRITTEN;
1245
1246 root = root->fs_info->tree_root;
1247
1248 spin_lock(&block_group->lock);
1249 if (block_group->disk_cache_state < BTRFS_DC_SETUP) {
1250 spin_unlock(&block_group->lock);
1251 return 0;
1252 }
1253
1254 if (block_group->delalloc_bytes) {
1255 block_group->disk_cache_state = BTRFS_DC_WRITTEN;
1256 spin_unlock(&block_group->lock);
1257 return 0;
1258 }
1259 spin_unlock(&block_group->lock);
1260
1261 inode = lookup_free_space_inode(root, block_group, path);
1262 if (IS_ERR(inode))
1263 return 0;
1264
1265 ret = __btrfs_write_out_cache(root, inode, ctl, block_group, trans,
1266 path, block_group->key.objectid);
1267 if (ret) {
1268 dcs = BTRFS_DC_ERROR;
1269 ret = 0;
1270#ifdef DEBUG
1271 btrfs_err(root->fs_info,
1272 "failed to write free space cache for block group %llu",
1273 block_group->key.objectid);
1274#endif
1275 }
1276
1277 spin_lock(&block_group->lock);
1278 block_group->disk_cache_state = dcs;
1279 spin_unlock(&block_group->lock);
1280 iput(inode);
1281 return ret;
1282}
1283
1284static inline unsigned long offset_to_bit(u64 bitmap_start, u32 unit,
1285 u64 offset)
1286{
1287 ASSERT(offset >= bitmap_start);
1288 offset -= bitmap_start;
1289 return (unsigned long)(div_u64(offset, unit));
1290}
1291
1292static inline unsigned long bytes_to_bits(u64 bytes, u32 unit)
1293{
1294 return (unsigned long)(div_u64(bytes, unit));
1295}
1296
1297static inline u64 offset_to_bitmap(struct btrfs_free_space_ctl *ctl,
1298 u64 offset)
1299{
1300 u64 bitmap_start;
1301 u32 bytes_per_bitmap;
1302
1303 bytes_per_bitmap = BITS_PER_BITMAP * ctl->unit;
1304 bitmap_start = offset - ctl->start;
1305 bitmap_start = div_u64(bitmap_start, bytes_per_bitmap);
1306 bitmap_start *= bytes_per_bitmap;
1307 bitmap_start += ctl->start;
1308
1309 return bitmap_start;
1310}
1311
1312static int tree_insert_offset(struct rb_root *root, u64 offset,
1313 struct rb_node *node, int bitmap)
1314{
1315 struct rb_node **p = &root->rb_node;
1316 struct rb_node *parent = NULL;
1317 struct btrfs_free_space *info;
1318
1319 while (*p) {
1320 parent = *p;
1321 info = rb_entry(parent, struct btrfs_free_space, offset_index);
1322
1323 if (offset < info->offset) {
1324 p = &(*p)->rb_left;
1325 } else if (offset > info->offset) {
1326 p = &(*p)->rb_right;
1327 } else {
1328 /*
1329 * we could have a bitmap entry and an extent entry
1330 * share the same offset. If this is the case, we want
1331 * the extent entry to always be found first if we do a
1332 * linear search through the tree, since we want to have
1333 * the quickest allocation time, and allocating from an
1334 * extent is faster than allocating from a bitmap. So
1335 * if we're inserting a bitmap and we find an entry at
1336 * this offset, we want to go right, or after this entry
1337 * logically. If we are inserting an extent and we've
1338 * found a bitmap, we want to go left, or before
1339 * logically.
1340 */
1341 if (bitmap) {
1342 if (info->bitmap) {
1343 WARN_ON_ONCE(1);
1344 return -EEXIST;
1345 }
1346 p = &(*p)->rb_right;
1347 } else {
1348 if (!info->bitmap) {
1349 WARN_ON_ONCE(1);
1350 return -EEXIST;
1351 }
1352 p = &(*p)->rb_left;
1353 }
1354 }
1355 }
1356
1357 rb_link_node(node, parent, p);
1358 rb_insert_color(node, root);
1359
1360 return 0;
1361}
1362
1363/*
1364 * searches the tree for the given offset.
1365 *
1366 * fuzzy - If this is set, then we are trying to make an allocation, and we just
1367 * want a section that has at least bytes size and comes at or after the given
1368 * offset.
1369 */
1370static struct btrfs_free_space *
1371tree_search_offset(struct btrfs_free_space_ctl *ctl,
1372 u64 offset, int bitmap_only, int fuzzy)
1373{
1374 struct rb_node *n = ctl->free_space_offset.rb_node;
1375 struct btrfs_free_space *entry, *prev = NULL;
1376
1377 /* find entry that is closest to the 'offset' */
1378 while (1) {
1379 if (!n) {
1380 entry = NULL;
1381 break;
1382 }
1383
1384 entry = rb_entry(n, struct btrfs_free_space, offset_index);
1385 prev = entry;
1386
1387 if (offset < entry->offset)
1388 n = n->rb_left;
1389 else if (offset > entry->offset)
1390 n = n->rb_right;
1391 else
1392 break;
1393 }
1394
1395 if (bitmap_only) {
1396 if (!entry)
1397 return NULL;
1398 if (entry->bitmap)
1399 return entry;
1400
1401 /*
1402 * bitmap entry and extent entry may share same offset,
1403 * in that case, bitmap entry comes after extent entry.
1404 */
1405 n = rb_next(n);
1406 if (!n)
1407 return NULL;
1408 entry = rb_entry(n, struct btrfs_free_space, offset_index);
1409 if (entry->offset != offset)
1410 return NULL;
1411
1412 WARN_ON(!entry->bitmap);
1413 return entry;
1414 } else if (entry) {
1415 if (entry->bitmap) {
1416 /*
1417 * if previous extent entry covers the offset,
1418 * we should return it instead of the bitmap entry
1419 */
1420 n = rb_prev(&entry->offset_index);
1421 if (n) {
1422 prev = rb_entry(n, struct btrfs_free_space,
1423 offset_index);
1424 if (!prev->bitmap &&
1425 prev->offset + prev->bytes > offset)
1426 entry = prev;
1427 }
1428 }
1429 return entry;
1430 }
1431
1432 if (!prev)
1433 return NULL;
1434
1435 /* find last entry before the 'offset' */
1436 entry = prev;
1437 if (entry->offset > offset) {
1438 n = rb_prev(&entry->offset_index);
1439 if (n) {
1440 entry = rb_entry(n, struct btrfs_free_space,
1441 offset_index);
1442 ASSERT(entry->offset <= offset);
1443 } else {
1444 if (fuzzy)
1445 return entry;
1446 else
1447 return NULL;
1448 }
1449 }
1450
1451 if (entry->bitmap) {
1452 n = rb_prev(&entry->offset_index);
1453 if (n) {
1454 prev = rb_entry(n, struct btrfs_free_space,
1455 offset_index);
1456 if (!prev->bitmap &&
1457 prev->offset + prev->bytes > offset)
1458 return prev;
1459 }
1460 if (entry->offset + BITS_PER_BITMAP * ctl->unit > offset)
1461 return entry;
1462 } else if (entry->offset + entry->bytes > offset)
1463 return entry;
1464
1465 if (!fuzzy)
1466 return NULL;
1467
1468 while (1) {
1469 if (entry->bitmap) {
1470 if (entry->offset + BITS_PER_BITMAP *
1471 ctl->unit > offset)
1472 break;
1473 } else {
1474 if (entry->offset + entry->bytes > offset)
1475 break;
1476 }
1477
1478 n = rb_next(&entry->offset_index);
1479 if (!n)
1480 return NULL;
1481 entry = rb_entry(n, struct btrfs_free_space, offset_index);
1482 }
1483 return entry;
1484}
1485
1486static inline void
1487__unlink_free_space(struct btrfs_free_space_ctl *ctl,
1488 struct btrfs_free_space *info)
1489{
1490 rb_erase(&info->offset_index, &ctl->free_space_offset);
1491 ctl->free_extents--;
1492}
1493
1494static void unlink_free_space(struct btrfs_free_space_ctl *ctl,
1495 struct btrfs_free_space *info)
1496{
1497 __unlink_free_space(ctl, info);
1498 ctl->free_space -= info->bytes;
1499}
1500
1501static int link_free_space(struct btrfs_free_space_ctl *ctl,
1502 struct btrfs_free_space *info)
1503{
1504 int ret = 0;
1505
1506 ASSERT(info->bytes || info->bitmap);
1507 ret = tree_insert_offset(&ctl->free_space_offset, info->offset,
1508 &info->offset_index, (info->bitmap != NULL));
1509 if (ret)
1510 return ret;
1511
1512 ctl->free_space += info->bytes;
1513 ctl->free_extents++;
1514 return ret;
1515}
1516
1517static void recalculate_thresholds(struct btrfs_free_space_ctl *ctl)
1518{
1519 struct btrfs_block_group_cache *block_group = ctl->private;
1520 u64 max_bytes;
1521 u64 bitmap_bytes;
1522 u64 extent_bytes;
1523 u64 size = block_group->key.offset;
1524 u32 bytes_per_bg = BITS_PER_BITMAP * ctl->unit;
1525 u32 max_bitmaps = div_u64(size + bytes_per_bg - 1, bytes_per_bg);
1526
1527 max_bitmaps = max_t(u32, max_bitmaps, 1);
1528
1529 ASSERT(ctl->total_bitmaps <= max_bitmaps);
1530
1531 /*
1532 * The goal is to keep the total amount of memory used per 1gb of space
1533 * at or below 32k, so we need to adjust how much memory we allow to be
1534 * used by extent based free space tracking
1535 */
1536 if (size < 1024 * 1024 * 1024)
1537 max_bytes = MAX_CACHE_BYTES_PER_GIG;
1538 else
1539 max_bytes = MAX_CACHE_BYTES_PER_GIG *
1540 div_u64(size, 1024 * 1024 * 1024);
1541
1542 /*
1543 * we want to account for 1 more bitmap than what we have so we can make
1544 * sure we don't go over our overall goal of MAX_CACHE_BYTES_PER_GIG as
1545 * we add more bitmaps.
1546 */
1547 bitmap_bytes = (ctl->total_bitmaps + 1) * PAGE_CACHE_SIZE;
1548
1549 if (bitmap_bytes >= max_bytes) {
1550 ctl->extents_thresh = 0;
1551 return;
1552 }
1553
1554 /*
1555 * we want the extent entry threshold to always be at most 1/2 the max
1556 * bytes we can have, or whatever is less than that.
1557 */
1558 extent_bytes = max_bytes - bitmap_bytes;
1559 extent_bytes = min_t(u64, extent_bytes, max_bytes >> 1);
1560
1561 ctl->extents_thresh =
1562 div_u64(extent_bytes, sizeof(struct btrfs_free_space));
1563}
1564
1565static inline void __bitmap_clear_bits(struct btrfs_free_space_ctl *ctl,
1566 struct btrfs_free_space *info,
1567 u64 offset, u64 bytes)
1568{
1569 unsigned long start, count;
1570
1571 start = offset_to_bit(info->offset, ctl->unit, offset);
1572 count = bytes_to_bits(bytes, ctl->unit);
1573 ASSERT(start + count <= BITS_PER_BITMAP);
1574
1575 bitmap_clear(info->bitmap, start, count);
1576
1577 info->bytes -= bytes;
1578}
1579
1580static void bitmap_clear_bits(struct btrfs_free_space_ctl *ctl,
1581 struct btrfs_free_space *info, u64 offset,
1582 u64 bytes)
1583{
1584 __bitmap_clear_bits(ctl, info, offset, bytes);
1585 ctl->free_space -= bytes;
1586}
1587
1588static void bitmap_set_bits(struct btrfs_free_space_ctl *ctl,
1589 struct btrfs_free_space *info, u64 offset,
1590 u64 bytes)
1591{
1592 unsigned long start, count;
1593
1594 start = offset_to_bit(info->offset, ctl->unit, offset);
1595 count = bytes_to_bits(bytes, ctl->unit);
1596 ASSERT(start + count <= BITS_PER_BITMAP);
1597
1598 bitmap_set(info->bitmap, start, count);
1599
1600 info->bytes += bytes;
1601 ctl->free_space += bytes;
1602}
1603
1604/*
1605 * If we can not find suitable extent, we will use bytes to record
1606 * the size of the max extent.
1607 */
1608static int search_bitmap(struct btrfs_free_space_ctl *ctl,
1609 struct btrfs_free_space *bitmap_info, u64 *offset,
1610 u64 *bytes)
1611{
1612 unsigned long found_bits = 0;
1613 unsigned long max_bits = 0;
1614 unsigned long bits, i;
1615 unsigned long next_zero;
1616 unsigned long extent_bits;
1617
1618 i = offset_to_bit(bitmap_info->offset, ctl->unit,
1619 max_t(u64, *offset, bitmap_info->offset));
1620 bits = bytes_to_bits(*bytes, ctl->unit);
1621
1622 for_each_set_bit_from(i, bitmap_info->bitmap, BITS_PER_BITMAP) {
1623 next_zero = find_next_zero_bit(bitmap_info->bitmap,
1624 BITS_PER_BITMAP, i);
1625 extent_bits = next_zero - i;
1626 if (extent_bits >= bits) {
1627 found_bits = extent_bits;
1628 break;
1629 } else if (extent_bits > max_bits) {
1630 max_bits = extent_bits;
1631 }
1632 i = next_zero;
1633 }
1634
1635 if (found_bits) {
1636 *offset = (u64)(i * ctl->unit) + bitmap_info->offset;
1637 *bytes = (u64)(found_bits) * ctl->unit;
1638 return 0;
1639 }
1640
1641 *bytes = (u64)(max_bits) * ctl->unit;
1642 return -1;
1643}
1644
1645/* Cache the size of the max extent in bytes */
1646static struct btrfs_free_space *
1647find_free_space(struct btrfs_free_space_ctl *ctl, u64 *offset, u64 *bytes,
1648 unsigned long align, u64 *max_extent_size)
1649{
1650 struct btrfs_free_space *entry;
1651 struct rb_node *node;
1652 u64 tmp;
1653 u64 align_off;
1654 int ret;
1655
1656 if (!ctl->free_space_offset.rb_node)
1657 goto out;
1658
1659 entry = tree_search_offset(ctl, offset_to_bitmap(ctl, *offset), 0, 1);
1660 if (!entry)
1661 goto out;
1662
1663 for (node = &entry->offset_index; node; node = rb_next(node)) {
1664 entry = rb_entry(node, struct btrfs_free_space, offset_index);
1665 if (entry->bytes < *bytes) {
1666 if (entry->bytes > *max_extent_size)
1667 *max_extent_size = entry->bytes;
1668 continue;
1669 }
1670
1671 /* make sure the space returned is big enough
1672 * to match our requested alignment
1673 */
1674 if (*bytes >= align) {
1675 tmp = entry->offset - ctl->start + align - 1;
1676 do_div(tmp, align);
1677 tmp = tmp * align + ctl->start;
1678 align_off = tmp - entry->offset;
1679 } else {
1680 align_off = 0;
1681 tmp = entry->offset;
1682 }
1683
1684 if (entry->bytes < *bytes + align_off) {
1685 if (entry->bytes > *max_extent_size)
1686 *max_extent_size = entry->bytes;
1687 continue;
1688 }
1689
1690 if (entry->bitmap) {
1691 u64 size = *bytes;
1692
1693 ret = search_bitmap(ctl, entry, &tmp, &size);
1694 if (!ret) {
1695 *offset = tmp;
1696 *bytes = size;
1697 return entry;
1698 } else if (size > *max_extent_size) {
1699 *max_extent_size = size;
1700 }
1701 continue;
1702 }
1703
1704 *offset = tmp;
1705 *bytes = entry->bytes - align_off;
1706 return entry;
1707 }
1708out:
1709 return NULL;
1710}
1711
1712static void add_new_bitmap(struct btrfs_free_space_ctl *ctl,
1713 struct btrfs_free_space *info, u64 offset)
1714{
1715 info->offset = offset_to_bitmap(ctl, offset);
1716 info->bytes = 0;
1717 INIT_LIST_HEAD(&info->list);
1718 link_free_space(ctl, info);
1719 ctl->total_bitmaps++;
1720
1721 ctl->op->recalc_thresholds(ctl);
1722}
1723
1724static void free_bitmap(struct btrfs_free_space_ctl *ctl,
1725 struct btrfs_free_space *bitmap_info)
1726{
1727 unlink_free_space(ctl, bitmap_info);
1728 kfree(bitmap_info->bitmap);
1729 kmem_cache_free(btrfs_free_space_cachep, bitmap_info);
1730 ctl->total_bitmaps--;
1731 ctl->op->recalc_thresholds(ctl);
1732}
1733
1734static noinline int remove_from_bitmap(struct btrfs_free_space_ctl *ctl,
1735 struct btrfs_free_space *bitmap_info,
1736 u64 *offset, u64 *bytes)
1737{
1738 u64 end;
1739 u64 search_start, search_bytes;
1740 int ret;
1741
1742again:
1743 end = bitmap_info->offset + (u64)(BITS_PER_BITMAP * ctl->unit) - 1;
1744
1745 /*
1746 * We need to search for bits in this bitmap. We could only cover some
1747 * of the extent in this bitmap thanks to how we add space, so we need
1748 * to search for as much as it as we can and clear that amount, and then
1749 * go searching for the next bit.
1750 */
1751 search_start = *offset;
1752 search_bytes = ctl->unit;
1753 search_bytes = min(search_bytes, end - search_start + 1);
1754 ret = search_bitmap(ctl, bitmap_info, &search_start, &search_bytes);
1755 if (ret < 0 || search_start != *offset)
1756 return -EINVAL;
1757
1758 /* We may have found more bits than what we need */
1759 search_bytes = min(search_bytes, *bytes);
1760
1761 /* Cannot clear past the end of the bitmap */
1762 search_bytes = min(search_bytes, end - search_start + 1);
1763
1764 bitmap_clear_bits(ctl, bitmap_info, search_start, search_bytes);
1765 *offset += search_bytes;
1766 *bytes -= search_bytes;
1767
1768 if (*bytes) {
1769 struct rb_node *next = rb_next(&bitmap_info->offset_index);
1770 if (!bitmap_info->bytes)
1771 free_bitmap(ctl, bitmap_info);
1772
1773 /*
1774 * no entry after this bitmap, but we still have bytes to
1775 * remove, so something has gone wrong.
1776 */
1777 if (!next)
1778 return -EINVAL;
1779
1780 bitmap_info = rb_entry(next, struct btrfs_free_space,
1781 offset_index);
1782
1783 /*
1784 * if the next entry isn't a bitmap we need to return to let the
1785 * extent stuff do its work.
1786 */
1787 if (!bitmap_info->bitmap)
1788 return -EAGAIN;
1789
1790 /*
1791 * Ok the next item is a bitmap, but it may not actually hold
1792 * the information for the rest of this free space stuff, so
1793 * look for it, and if we don't find it return so we can try
1794 * everything over again.
1795 */
1796 search_start = *offset;
1797 search_bytes = ctl->unit;
1798 ret = search_bitmap(ctl, bitmap_info, &search_start,
1799 &search_bytes);
1800 if (ret < 0 || search_start != *offset)
1801 return -EAGAIN;
1802
1803 goto again;
1804 } else if (!bitmap_info->bytes)
1805 free_bitmap(ctl, bitmap_info);
1806
1807 return 0;
1808}
1809
1810static u64 add_bytes_to_bitmap(struct btrfs_free_space_ctl *ctl,
1811 struct btrfs_free_space *info, u64 offset,
1812 u64 bytes)
1813{
1814 u64 bytes_to_set = 0;
1815 u64 end;
1816
1817 end = info->offset + (u64)(BITS_PER_BITMAP * ctl->unit);
1818
1819 bytes_to_set = min(end - offset, bytes);
1820
1821 bitmap_set_bits(ctl, info, offset, bytes_to_set);
1822
1823 return bytes_to_set;
1824
1825}
1826
1827static bool use_bitmap(struct btrfs_free_space_ctl *ctl,
1828 struct btrfs_free_space *info)
1829{
1830 struct btrfs_block_group_cache *block_group = ctl->private;
1831
1832 /*
1833 * If we are below the extents threshold then we can add this as an
1834 * extent, and don't have to deal with the bitmap
1835 */
1836 if (ctl->free_extents < ctl->extents_thresh) {
1837 /*
1838 * If this block group has some small extents we don't want to
1839 * use up all of our free slots in the cache with them, we want
1840 * to reserve them to larger extents, however if we have plent
1841 * of cache left then go ahead an dadd them, no sense in adding
1842 * the overhead of a bitmap if we don't have to.
1843 */
1844 if (info->bytes <= block_group->sectorsize * 4) {
1845 if (ctl->free_extents * 2 <= ctl->extents_thresh)
1846 return false;
1847 } else {
1848 return false;
1849 }
1850 }
1851
1852 /*
1853 * The original block groups from mkfs can be really small, like 8
1854 * megabytes, so don't bother with a bitmap for those entries. However
1855 * some block groups can be smaller than what a bitmap would cover but
1856 * are still large enough that they could overflow the 32k memory limit,
1857 * so allow those block groups to still be allowed to have a bitmap
1858 * entry.
1859 */
1860 if (((BITS_PER_BITMAP * ctl->unit) >> 1) > block_group->key.offset)
1861 return false;
1862
1863 return true;
1864}
1865
1866static struct btrfs_free_space_op free_space_op = {
1867 .recalc_thresholds = recalculate_thresholds,
1868 .use_bitmap = use_bitmap,
1869};
1870
1871static int insert_into_bitmap(struct btrfs_free_space_ctl *ctl,
1872 struct btrfs_free_space *info)
1873{
1874 struct btrfs_free_space *bitmap_info;
1875 struct btrfs_block_group_cache *block_group = NULL;
1876 int added = 0;
1877 u64 bytes, offset, bytes_added;
1878 int ret;
1879
1880 bytes = info->bytes;
1881 offset = info->offset;
1882
1883 if (!ctl->op->use_bitmap(ctl, info))
1884 return 0;
1885
1886 if (ctl->op == &free_space_op)
1887 block_group = ctl->private;
1888again:
1889 /*
1890 * Since we link bitmaps right into the cluster we need to see if we
1891 * have a cluster here, and if so and it has our bitmap we need to add
1892 * the free space to that bitmap.
1893 */
1894 if (block_group && !list_empty(&block_group->cluster_list)) {
1895 struct btrfs_free_cluster *cluster;
1896 struct rb_node *node;
1897 struct btrfs_free_space *entry;
1898
1899 cluster = list_entry(block_group->cluster_list.next,
1900 struct btrfs_free_cluster,
1901 block_group_list);
1902 spin_lock(&cluster->lock);
1903 node = rb_first(&cluster->root);
1904 if (!node) {
1905 spin_unlock(&cluster->lock);
1906 goto no_cluster_bitmap;
1907 }
1908
1909 entry = rb_entry(node, struct btrfs_free_space, offset_index);
1910 if (!entry->bitmap) {
1911 spin_unlock(&cluster->lock);
1912 goto no_cluster_bitmap;
1913 }
1914
1915 if (entry->offset == offset_to_bitmap(ctl, offset)) {
1916 bytes_added = add_bytes_to_bitmap(ctl, entry,
1917 offset, bytes);
1918 bytes -= bytes_added;
1919 offset += bytes_added;
1920 }
1921 spin_unlock(&cluster->lock);
1922 if (!bytes) {
1923 ret = 1;
1924 goto out;
1925 }
1926 }
1927
1928no_cluster_bitmap:
1929 bitmap_info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset),
1930 1, 0);
1931 if (!bitmap_info) {
1932 ASSERT(added == 0);
1933 goto new_bitmap;
1934 }
1935
1936 bytes_added = add_bytes_to_bitmap(ctl, bitmap_info, offset, bytes);
1937 bytes -= bytes_added;
1938 offset += bytes_added;
1939 added = 0;
1940
1941 if (!bytes) {
1942 ret = 1;
1943 goto out;
1944 } else
1945 goto again;
1946
1947new_bitmap:
1948 if (info && info->bitmap) {
1949 add_new_bitmap(ctl, info, offset);
1950 added = 1;
1951 info = NULL;
1952 goto again;
1953 } else {
1954 spin_unlock(&ctl->tree_lock);
1955
1956 /* no pre-allocated info, allocate a new one */
1957 if (!info) {
1958 info = kmem_cache_zalloc(btrfs_free_space_cachep,
1959 GFP_NOFS);
1960 if (!info) {
1961 spin_lock(&ctl->tree_lock);
1962 ret = -ENOMEM;
1963 goto out;
1964 }
1965 }
1966
1967 /* allocate the bitmap */
1968 info->bitmap = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS);
1969 spin_lock(&ctl->tree_lock);
1970 if (!info->bitmap) {
1971 ret = -ENOMEM;
1972 goto out;
1973 }
1974 goto again;
1975 }
1976
1977out:
1978 if (info) {
1979 if (info->bitmap)
1980 kfree(info->bitmap);
1981 kmem_cache_free(btrfs_free_space_cachep, info);
1982 }
1983
1984 return ret;
1985}
1986
1987static bool try_merge_free_space(struct btrfs_free_space_ctl *ctl,
1988 struct btrfs_free_space *info, bool update_stat)
1989{
1990 struct btrfs_free_space *left_info;
1991 struct btrfs_free_space *right_info;
1992 bool merged = false;
1993 u64 offset = info->offset;
1994 u64 bytes = info->bytes;
1995
1996 /*
1997 * first we want to see if there is free space adjacent to the range we
1998 * are adding, if there is remove that struct and add a new one to
1999 * cover the entire range
2000 */
2001 right_info = tree_search_offset(ctl, offset + bytes, 0, 0);
2002 if (right_info && rb_prev(&right_info->offset_index))
2003 left_info = rb_entry(rb_prev(&right_info->offset_index),
2004 struct btrfs_free_space, offset_index);
2005 else
2006 left_info = tree_search_offset(ctl, offset - 1, 0, 0);
2007
2008 if (right_info && !right_info->bitmap) {
2009 if (update_stat)
2010 unlink_free_space(ctl, right_info);
2011 else
2012 __unlink_free_space(ctl, right_info);
2013 info->bytes += right_info->bytes;
2014 kmem_cache_free(btrfs_free_space_cachep, right_info);
2015 merged = true;
2016 }
2017
2018 if (left_info && !left_info->bitmap &&
2019 left_info->offset + left_info->bytes == offset) {
2020 if (update_stat)
2021 unlink_free_space(ctl, left_info);
2022 else
2023 __unlink_free_space(ctl, left_info);
2024 info->offset = left_info->offset;
2025 info->bytes += left_info->bytes;
2026 kmem_cache_free(btrfs_free_space_cachep, left_info);
2027 merged = true;
2028 }
2029
2030 return merged;
2031}
2032
2033static bool steal_from_bitmap_to_end(struct btrfs_free_space_ctl *ctl,
2034 struct btrfs_free_space *info,
2035 bool update_stat)
2036{
2037 struct btrfs_free_space *bitmap;
2038 unsigned long i;
2039 unsigned long j;
2040 const u64 end = info->offset + info->bytes;
2041 const u64 bitmap_offset = offset_to_bitmap(ctl, end);
2042 u64 bytes;
2043
2044 bitmap = tree_search_offset(ctl, bitmap_offset, 1, 0);
2045 if (!bitmap)
2046 return false;
2047
2048 i = offset_to_bit(bitmap->offset, ctl->unit, end);
2049 j = find_next_zero_bit(bitmap->bitmap, BITS_PER_BITMAP, i);
2050 if (j == i)
2051 return false;
2052 bytes = (j - i) * ctl->unit;
2053 info->bytes += bytes;
2054
2055 if (update_stat)
2056 bitmap_clear_bits(ctl, bitmap, end, bytes);
2057 else
2058 __bitmap_clear_bits(ctl, bitmap, end, bytes);
2059
2060 if (!bitmap->bytes)
2061 free_bitmap(ctl, bitmap);
2062
2063 return true;
2064}
2065
2066static bool steal_from_bitmap_to_front(struct btrfs_free_space_ctl *ctl,
2067 struct btrfs_free_space *info,
2068 bool update_stat)
2069{
2070 struct btrfs_free_space *bitmap;
2071 u64 bitmap_offset;
2072 unsigned long i;
2073 unsigned long j;
2074 unsigned long prev_j;
2075 u64 bytes;
2076
2077 bitmap_offset = offset_to_bitmap(ctl, info->offset);
2078 /* If we're on a boundary, try the previous logical bitmap. */
2079 if (bitmap_offset == info->offset) {
2080 if (info->offset == 0)
2081 return false;
2082 bitmap_offset = offset_to_bitmap(ctl, info->offset - 1);
2083 }
2084
2085 bitmap = tree_search_offset(ctl, bitmap_offset, 1, 0);
2086 if (!bitmap)
2087 return false;
2088
2089 i = offset_to_bit(bitmap->offset, ctl->unit, info->offset) - 1;
2090 j = 0;
2091 prev_j = (unsigned long)-1;
2092 for_each_clear_bit_from(j, bitmap->bitmap, BITS_PER_BITMAP) {
2093 if (j > i)
2094 break;
2095 prev_j = j;
2096 }
2097 if (prev_j == i)
2098 return false;
2099
2100 if (prev_j == (unsigned long)-1)
2101 bytes = (i + 1) * ctl->unit;
2102 else
2103 bytes = (i - prev_j) * ctl->unit;
2104
2105 info->offset -= bytes;
2106 info->bytes += bytes;
2107
2108 if (update_stat)
2109 bitmap_clear_bits(ctl, bitmap, info->offset, bytes);
2110 else
2111 __bitmap_clear_bits(ctl, bitmap, info->offset, bytes);
2112
2113 if (!bitmap->bytes)
2114 free_bitmap(ctl, bitmap);
2115
2116 return true;
2117}
2118
2119/*
2120 * We prefer always to allocate from extent entries, both for clustered and
2121 * non-clustered allocation requests. So when attempting to add a new extent
2122 * entry, try to see if there's adjacent free space in bitmap entries, and if
2123 * there is, migrate that space from the bitmaps to the extent.
2124 * Like this we get better chances of satisfying space allocation requests
2125 * because we attempt to satisfy them based on a single cache entry, and never
2126 * on 2 or more entries - even if the entries represent a contiguous free space
2127 * region (e.g. 1 extent entry + 1 bitmap entry starting where the extent entry
2128 * ends).
2129 */
2130static void steal_from_bitmap(struct btrfs_free_space_ctl *ctl,
2131 struct btrfs_free_space *info,
2132 bool update_stat)
2133{
2134 /*
2135 * Only work with disconnected entries, as we can change their offset,
2136 * and must be extent entries.
2137 */
2138 ASSERT(!info->bitmap);
2139 ASSERT(RB_EMPTY_NODE(&info->offset_index));
2140
2141 if (ctl->total_bitmaps > 0) {
2142 bool stole_end;
2143 bool stole_front = false;
2144
2145 stole_end = steal_from_bitmap_to_end(ctl, info, update_stat);
2146 if (ctl->total_bitmaps > 0)
2147 stole_front = steal_from_bitmap_to_front(ctl, info,
2148 update_stat);
2149
2150 if (stole_end || stole_front)
2151 try_merge_free_space(ctl, info, update_stat);
2152 }
2153}
2154
2155int __btrfs_add_free_space(struct btrfs_free_space_ctl *ctl,
2156 u64 offset, u64 bytes)
2157{
2158 struct btrfs_free_space *info;
2159 int ret = 0;
2160
2161 info = kmem_cache_zalloc(btrfs_free_space_cachep, GFP_NOFS);
2162 if (!info)
2163 return -ENOMEM;
2164
2165 info->offset = offset;
2166 info->bytes = bytes;
2167 RB_CLEAR_NODE(&info->offset_index);
2168
2169 spin_lock(&ctl->tree_lock);
2170
2171 if (try_merge_free_space(ctl, info, true))
2172 goto link;
2173
2174 /*
2175 * There was no extent directly to the left or right of this new
2176 * extent then we know we're going to have to allocate a new extent, so
2177 * before we do that see if we need to drop this into a bitmap
2178 */
2179 ret = insert_into_bitmap(ctl, info);
2180 if (ret < 0) {
2181 goto out;
2182 } else if (ret) {
2183 ret = 0;
2184 goto out;
2185 }
2186link:
2187 /*
2188 * Only steal free space from adjacent bitmaps if we're sure we're not
2189 * going to add the new free space to existing bitmap entries - because
2190 * that would mean unnecessary work that would be reverted. Therefore
2191 * attempt to steal space from bitmaps if we're adding an extent entry.
2192 */
2193 steal_from_bitmap(ctl, info, true);
2194
2195 ret = link_free_space(ctl, info);
2196 if (ret)
2197 kmem_cache_free(btrfs_free_space_cachep, info);
2198out:
2199 spin_unlock(&ctl->tree_lock);
2200
2201 if (ret) {
2202 printk(KERN_CRIT "BTRFS: unable to add free space :%d\n", ret);
2203 ASSERT(ret != -EEXIST);
2204 }
2205
2206 return ret;
2207}
2208
2209int btrfs_remove_free_space(struct btrfs_block_group_cache *block_group,
2210 u64 offset, u64 bytes)
2211{
2212 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2213 struct btrfs_free_space *info;
2214 int ret;
2215 bool re_search = false;
2216
2217 spin_lock(&ctl->tree_lock);
2218
2219again:
2220 ret = 0;
2221 if (!bytes)
2222 goto out_lock;
2223
2224 info = tree_search_offset(ctl, offset, 0, 0);
2225 if (!info) {
2226 /*
2227 * oops didn't find an extent that matched the space we wanted
2228 * to remove, look for a bitmap instead
2229 */
2230 info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset),
2231 1, 0);
2232 if (!info) {
2233 /*
2234 * If we found a partial bit of our free space in a
2235 * bitmap but then couldn't find the other part this may
2236 * be a problem, so WARN about it.
2237 */
2238 WARN_ON(re_search);
2239 goto out_lock;
2240 }
2241 }
2242
2243 re_search = false;
2244 if (!info->bitmap) {
2245 unlink_free_space(ctl, info);
2246 if (offset == info->offset) {
2247 u64 to_free = min(bytes, info->bytes);
2248
2249 info->bytes -= to_free;
2250 info->offset += to_free;
2251 if (info->bytes) {
2252 ret = link_free_space(ctl, info);
2253 WARN_ON(ret);
2254 } else {
2255 kmem_cache_free(btrfs_free_space_cachep, info);
2256 }
2257
2258 offset += to_free;
2259 bytes -= to_free;
2260 goto again;
2261 } else {
2262 u64 old_end = info->bytes + info->offset;
2263
2264 info->bytes = offset - info->offset;
2265 ret = link_free_space(ctl, info);
2266 WARN_ON(ret);
2267 if (ret)
2268 goto out_lock;
2269
2270 /* Not enough bytes in this entry to satisfy us */
2271 if (old_end < offset + bytes) {
2272 bytes -= old_end - offset;
2273 offset = old_end;
2274 goto again;
2275 } else if (old_end == offset + bytes) {
2276 /* all done */
2277 goto out_lock;
2278 }
2279 spin_unlock(&ctl->tree_lock);
2280
2281 ret = btrfs_add_free_space(block_group, offset + bytes,
2282 old_end - (offset + bytes));
2283 WARN_ON(ret);
2284 goto out;
2285 }
2286 }
2287
2288 ret = remove_from_bitmap(ctl, info, &offset, &bytes);
2289 if (ret == -EAGAIN) {
2290 re_search = true;
2291 goto again;
2292 }
2293out_lock:
2294 spin_unlock(&ctl->tree_lock);
2295out:
2296 return ret;
2297}
2298
2299void btrfs_dump_free_space(struct btrfs_block_group_cache *block_group,
2300 u64 bytes)
2301{
2302 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2303 struct btrfs_free_space *info;
2304 struct rb_node *n;
2305 int count = 0;
2306
2307 for (n = rb_first(&ctl->free_space_offset); n; n = rb_next(n)) {
2308 info = rb_entry(n, struct btrfs_free_space, offset_index);
2309 if (info->bytes >= bytes && !block_group->ro)
2310 count++;
2311 btrfs_crit(block_group->fs_info,
2312 "entry offset %llu, bytes %llu, bitmap %s",
2313 info->offset, info->bytes,
2314 (info->bitmap) ? "yes" : "no");
2315 }
2316 btrfs_info(block_group->fs_info, "block group has cluster?: %s",
2317 list_empty(&block_group->cluster_list) ? "no" : "yes");
2318 btrfs_info(block_group->fs_info,
2319 "%d blocks of free space at or bigger than bytes is", count);
2320}
2321
2322void btrfs_init_free_space_ctl(struct btrfs_block_group_cache *block_group)
2323{
2324 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2325
2326 spin_lock_init(&ctl->tree_lock);
2327 ctl->unit = block_group->sectorsize;
2328 ctl->start = block_group->key.objectid;
2329 ctl->private = block_group;
2330 ctl->op = &free_space_op;
2331 INIT_LIST_HEAD(&ctl->trimming_ranges);
2332 mutex_init(&ctl->cache_writeout_mutex);
2333
2334 /*
2335 * we only want to have 32k of ram per block group for keeping
2336 * track of free space, and if we pass 1/2 of that we want to
2337 * start converting things over to using bitmaps
2338 */
2339 ctl->extents_thresh = ((1024 * 32) / 2) /
2340 sizeof(struct btrfs_free_space);
2341}
2342
2343/*
2344 * for a given cluster, put all of its extents back into the free
2345 * space cache. If the block group passed doesn't match the block group
2346 * pointed to by the cluster, someone else raced in and freed the
2347 * cluster already. In that case, we just return without changing anything
2348 */
2349static int
2350__btrfs_return_cluster_to_free_space(
2351 struct btrfs_block_group_cache *block_group,
2352 struct btrfs_free_cluster *cluster)
2353{
2354 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2355 struct btrfs_free_space *entry;
2356 struct rb_node *node;
2357
2358 spin_lock(&cluster->lock);
2359 if (cluster->block_group != block_group)
2360 goto out;
2361
2362 cluster->block_group = NULL;
2363 cluster->window_start = 0;
2364 list_del_init(&cluster->block_group_list);
2365
2366 node = rb_first(&cluster->root);
2367 while (node) {
2368 bool bitmap;
2369
2370 entry = rb_entry(node, struct btrfs_free_space, offset_index);
2371 node = rb_next(&entry->offset_index);
2372 rb_erase(&entry->offset_index, &cluster->root);
2373 RB_CLEAR_NODE(&entry->offset_index);
2374
2375 bitmap = (entry->bitmap != NULL);
2376 if (!bitmap) {
2377 try_merge_free_space(ctl, entry, false);
2378 steal_from_bitmap(ctl, entry, false);
2379 }
2380 tree_insert_offset(&ctl->free_space_offset,
2381 entry->offset, &entry->offset_index, bitmap);
2382 }
2383 cluster->root = RB_ROOT;
2384
2385out:
2386 spin_unlock(&cluster->lock);
2387 btrfs_put_block_group(block_group);
2388 return 0;
2389}
2390
2391static void __btrfs_remove_free_space_cache_locked(
2392 struct btrfs_free_space_ctl *ctl)
2393{
2394 struct btrfs_free_space *info;
2395 struct rb_node *node;
2396
2397 while ((node = rb_last(&ctl->free_space_offset)) != NULL) {
2398 info = rb_entry(node, struct btrfs_free_space, offset_index);
2399 if (!info->bitmap) {
2400 unlink_free_space(ctl, info);
2401 kmem_cache_free(btrfs_free_space_cachep, info);
2402 } else {
2403 free_bitmap(ctl, info);
2404 }
2405
2406 cond_resched_lock(&ctl->tree_lock);
2407 }
2408}
2409
2410void __btrfs_remove_free_space_cache(struct btrfs_free_space_ctl *ctl)
2411{
2412 spin_lock(&ctl->tree_lock);
2413 __btrfs_remove_free_space_cache_locked(ctl);
2414 spin_unlock(&ctl->tree_lock);
2415}
2416
2417void btrfs_remove_free_space_cache(struct btrfs_block_group_cache *block_group)
2418{
2419 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2420 struct btrfs_free_cluster *cluster;
2421 struct list_head *head;
2422
2423 spin_lock(&ctl->tree_lock);
2424 while ((head = block_group->cluster_list.next) !=
2425 &block_group->cluster_list) {
2426 cluster = list_entry(head, struct btrfs_free_cluster,
2427 block_group_list);
2428
2429 WARN_ON(cluster->block_group != block_group);
2430 __btrfs_return_cluster_to_free_space(block_group, cluster);
2431
2432 cond_resched_lock(&ctl->tree_lock);
2433 }
2434 __btrfs_remove_free_space_cache_locked(ctl);
2435 spin_unlock(&ctl->tree_lock);
2436
2437}
2438
2439u64 btrfs_find_space_for_alloc(struct btrfs_block_group_cache *block_group,
2440 u64 offset, u64 bytes, u64 empty_size,
2441 u64 *max_extent_size)
2442{
2443 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2444 struct btrfs_free_space *entry = NULL;
2445 u64 bytes_search = bytes + empty_size;
2446 u64 ret = 0;
2447 u64 align_gap = 0;
2448 u64 align_gap_len = 0;
2449
2450 spin_lock(&ctl->tree_lock);
2451 entry = find_free_space(ctl, &offset, &bytes_search,
2452 block_group->full_stripe_len, max_extent_size);
2453 if (!entry)
2454 goto out;
2455
2456 ret = offset;
2457 if (entry->bitmap) {
2458 bitmap_clear_bits(ctl, entry, offset, bytes);
2459 if (!entry->bytes)
2460 free_bitmap(ctl, entry);
2461 } else {
2462 unlink_free_space(ctl, entry);
2463 align_gap_len = offset - entry->offset;
2464 align_gap = entry->offset;
2465
2466 entry->offset = offset + bytes;
2467 WARN_ON(entry->bytes < bytes + align_gap_len);
2468
2469 entry->bytes -= bytes + align_gap_len;
2470 if (!entry->bytes)
2471 kmem_cache_free(btrfs_free_space_cachep, entry);
2472 else
2473 link_free_space(ctl, entry);
2474 }
2475out:
2476 spin_unlock(&ctl->tree_lock);
2477
2478 if (align_gap_len)
2479 __btrfs_add_free_space(ctl, align_gap, align_gap_len);
2480 return ret;
2481}
2482
2483/*
2484 * given a cluster, put all of its extents back into the free space
2485 * cache. If a block group is passed, this function will only free
2486 * a cluster that belongs to the passed block group.
2487 *
2488 * Otherwise, it'll get a reference on the block group pointed to by the
2489 * cluster and remove the cluster from it.
2490 */
2491int btrfs_return_cluster_to_free_space(
2492 struct btrfs_block_group_cache *block_group,
2493 struct btrfs_free_cluster *cluster)
2494{
2495 struct btrfs_free_space_ctl *ctl;
2496 int ret;
2497
2498 /* first, get a safe pointer to the block group */
2499 spin_lock(&cluster->lock);
2500 if (!block_group) {
2501 block_group = cluster->block_group;
2502 if (!block_group) {
2503 spin_unlock(&cluster->lock);
2504 return 0;
2505 }
2506 } else if (cluster->block_group != block_group) {
2507 /* someone else has already freed it don't redo their work */
2508 spin_unlock(&cluster->lock);
2509 return 0;
2510 }
2511 atomic_inc(&block_group->count);
2512 spin_unlock(&cluster->lock);
2513
2514 ctl = block_group->free_space_ctl;
2515
2516 /* now return any extents the cluster had on it */
2517 spin_lock(&ctl->tree_lock);
2518 ret = __btrfs_return_cluster_to_free_space(block_group, cluster);
2519 spin_unlock(&ctl->tree_lock);
2520
2521 /* finally drop our ref */
2522 btrfs_put_block_group(block_group);
2523 return ret;
2524}
2525
2526static u64 btrfs_alloc_from_bitmap(struct btrfs_block_group_cache *block_group,
2527 struct btrfs_free_cluster *cluster,
2528 struct btrfs_free_space *entry,
2529 u64 bytes, u64 min_start,
2530 u64 *max_extent_size)
2531{
2532 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2533 int err;
2534 u64 search_start = cluster->window_start;
2535 u64 search_bytes = bytes;
2536 u64 ret = 0;
2537
2538 search_start = min_start;
2539 search_bytes = bytes;
2540
2541 err = search_bitmap(ctl, entry, &search_start, &search_bytes);
2542 if (err) {
2543 if (search_bytes > *max_extent_size)
2544 *max_extent_size = search_bytes;
2545 return 0;
2546 }
2547
2548 ret = search_start;
2549 __bitmap_clear_bits(ctl, entry, ret, bytes);
2550
2551 return ret;
2552}
2553
2554/*
2555 * given a cluster, try to allocate 'bytes' from it, returns 0
2556 * if it couldn't find anything suitably large, or a logical disk offset
2557 * if things worked out
2558 */
2559u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group,
2560 struct btrfs_free_cluster *cluster, u64 bytes,
2561 u64 min_start, u64 *max_extent_size)
2562{
2563 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2564 struct btrfs_free_space *entry = NULL;
2565 struct rb_node *node;
2566 u64 ret = 0;
2567
2568 spin_lock(&cluster->lock);
2569 if (bytes > cluster->max_size)
2570 goto out;
2571
2572 if (cluster->block_group != block_group)
2573 goto out;
2574
2575 node = rb_first(&cluster->root);
2576 if (!node)
2577 goto out;
2578
2579 entry = rb_entry(node, struct btrfs_free_space, offset_index);
2580 while (1) {
2581 if (entry->bytes < bytes && entry->bytes > *max_extent_size)
2582 *max_extent_size = entry->bytes;
2583
2584 if (entry->bytes < bytes ||
2585 (!entry->bitmap && entry->offset < min_start)) {
2586 node = rb_next(&entry->offset_index);
2587 if (!node)
2588 break;
2589 entry = rb_entry(node, struct btrfs_free_space,
2590 offset_index);
2591 continue;
2592 }
2593
2594 if (entry->bitmap) {
2595 ret = btrfs_alloc_from_bitmap(block_group,
2596 cluster, entry, bytes,
2597 cluster->window_start,
2598 max_extent_size);
2599 if (ret == 0) {
2600 node = rb_next(&entry->offset_index);
2601 if (!node)
2602 break;
2603 entry = rb_entry(node, struct btrfs_free_space,
2604 offset_index);
2605 continue;
2606 }
2607 cluster->window_start += bytes;
2608 } else {
2609 ret = entry->offset;
2610
2611 entry->offset += bytes;
2612 entry->bytes -= bytes;
2613 }
2614
2615 if (entry->bytes == 0)
2616 rb_erase(&entry->offset_index, &cluster->root);
2617 break;
2618 }
2619out:
2620 spin_unlock(&cluster->lock);
2621
2622 if (!ret)
2623 return 0;
2624
2625 spin_lock(&ctl->tree_lock);
2626
2627 ctl->free_space -= bytes;
2628 if (entry->bytes == 0) {
2629 ctl->free_extents--;
2630 if (entry->bitmap) {
2631 kfree(entry->bitmap);
2632 ctl->total_bitmaps--;
2633 ctl->op->recalc_thresholds(ctl);
2634 }
2635 kmem_cache_free(btrfs_free_space_cachep, entry);
2636 }
2637
2638 spin_unlock(&ctl->tree_lock);
2639
2640 return ret;
2641}
2642
2643static int btrfs_bitmap_cluster(struct btrfs_block_group_cache *block_group,
2644 struct btrfs_free_space *entry,
2645 struct btrfs_free_cluster *cluster,
2646 u64 offset, u64 bytes,
2647 u64 cont1_bytes, u64 min_bytes)
2648{
2649 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2650 unsigned long next_zero;
2651 unsigned long i;
2652 unsigned long want_bits;
2653 unsigned long min_bits;
2654 unsigned long found_bits;
2655 unsigned long start = 0;
2656 unsigned long total_found = 0;
2657 int ret;
2658
2659 i = offset_to_bit(entry->offset, ctl->unit,
2660 max_t(u64, offset, entry->offset));
2661 want_bits = bytes_to_bits(bytes, ctl->unit);
2662 min_bits = bytes_to_bits(min_bytes, ctl->unit);
2663
2664again:
2665 found_bits = 0;
2666 for_each_set_bit_from(i, entry->bitmap, BITS_PER_BITMAP) {
2667 next_zero = find_next_zero_bit(entry->bitmap,
2668 BITS_PER_BITMAP, i);
2669 if (next_zero - i >= min_bits) {
2670 found_bits = next_zero - i;
2671 break;
2672 }
2673 i = next_zero;
2674 }
2675
2676 if (!found_bits)
2677 return -ENOSPC;
2678
2679 if (!total_found) {
2680 start = i;
2681 cluster->max_size = 0;
2682 }
2683
2684 total_found += found_bits;
2685
2686 if (cluster->max_size < found_bits * ctl->unit)
2687 cluster->max_size = found_bits * ctl->unit;
2688
2689 if (total_found < want_bits || cluster->max_size < cont1_bytes) {
2690 i = next_zero + 1;
2691 goto again;
2692 }
2693
2694 cluster->window_start = start * ctl->unit + entry->offset;
2695 rb_erase(&entry->offset_index, &ctl->free_space_offset);
2696 ret = tree_insert_offset(&cluster->root, entry->offset,
2697 &entry->offset_index, 1);
2698 ASSERT(!ret); /* -EEXIST; Logic error */
2699
2700 trace_btrfs_setup_cluster(block_group, cluster,
2701 total_found * ctl->unit, 1);
2702 return 0;
2703}
2704
2705/*
2706 * This searches the block group for just extents to fill the cluster with.
2707 * Try to find a cluster with at least bytes total bytes, at least one
2708 * extent of cont1_bytes, and other clusters of at least min_bytes.
2709 */
2710static noinline int
2711setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group,
2712 struct btrfs_free_cluster *cluster,
2713 struct list_head *bitmaps, u64 offset, u64 bytes,
2714 u64 cont1_bytes, u64 min_bytes)
2715{
2716 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2717 struct btrfs_free_space *first = NULL;
2718 struct btrfs_free_space *entry = NULL;
2719 struct btrfs_free_space *last;
2720 struct rb_node *node;
2721 u64 window_free;
2722 u64 max_extent;
2723 u64 total_size = 0;
2724
2725 entry = tree_search_offset(ctl, offset, 0, 1);
2726 if (!entry)
2727 return -ENOSPC;
2728
2729 /*
2730 * We don't want bitmaps, so just move along until we find a normal
2731 * extent entry.
2732 */
2733 while (entry->bitmap || entry->bytes < min_bytes) {
2734 if (entry->bitmap && list_empty(&entry->list))
2735 list_add_tail(&entry->list, bitmaps);
2736 node = rb_next(&entry->offset_index);
2737 if (!node)
2738 return -ENOSPC;
2739 entry = rb_entry(node, struct btrfs_free_space, offset_index);
2740 }
2741
2742 window_free = entry->bytes;
2743 max_extent = entry->bytes;
2744 first = entry;
2745 last = entry;
2746
2747 for (node = rb_next(&entry->offset_index); node;
2748 node = rb_next(&entry->offset_index)) {
2749 entry = rb_entry(node, struct btrfs_free_space, offset_index);
2750
2751 if (entry->bitmap) {
2752 if (list_empty(&entry->list))
2753 list_add_tail(&entry->list, bitmaps);
2754 continue;
2755 }
2756
2757 if (entry->bytes < min_bytes)
2758 continue;
2759
2760 last = entry;
2761 window_free += entry->bytes;
2762 if (entry->bytes > max_extent)
2763 max_extent = entry->bytes;
2764 }
2765
2766 if (window_free < bytes || max_extent < cont1_bytes)
2767 return -ENOSPC;
2768
2769 cluster->window_start = first->offset;
2770
2771 node = &first->offset_index;
2772
2773 /*
2774 * now we've found our entries, pull them out of the free space
2775 * cache and put them into the cluster rbtree
2776 */
2777 do {
2778 int ret;
2779
2780 entry = rb_entry(node, struct btrfs_free_space, offset_index);
2781 node = rb_next(&entry->offset_index);
2782 if (entry->bitmap || entry->bytes < min_bytes)
2783 continue;
2784
2785 rb_erase(&entry->offset_index, &ctl->free_space_offset);
2786 ret = tree_insert_offset(&cluster->root, entry->offset,
2787 &entry->offset_index, 0);
2788 total_size += entry->bytes;
2789 ASSERT(!ret); /* -EEXIST; Logic error */
2790 } while (node && entry != last);
2791
2792 cluster->max_size = max_extent;
2793 trace_btrfs_setup_cluster(block_group, cluster, total_size, 0);
2794 return 0;
2795}
2796
2797/*
2798 * This specifically looks for bitmaps that may work in the cluster, we assume
2799 * that we have already failed to find extents that will work.
2800 */
2801static noinline int
2802setup_cluster_bitmap(struct btrfs_block_group_cache *block_group,
2803 struct btrfs_free_cluster *cluster,
2804 struct list_head *bitmaps, u64 offset, u64 bytes,
2805 u64 cont1_bytes, u64 min_bytes)
2806{
2807 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2808 struct btrfs_free_space *entry;
2809 int ret = -ENOSPC;
2810 u64 bitmap_offset = offset_to_bitmap(ctl, offset);
2811
2812 if (ctl->total_bitmaps == 0)
2813 return -ENOSPC;
2814
2815 /*
2816 * The bitmap that covers offset won't be in the list unless offset
2817 * is just its start offset.
2818 */
2819 entry = list_first_entry(bitmaps, struct btrfs_free_space, list);
2820 if (entry->offset != bitmap_offset) {
2821 entry = tree_search_offset(ctl, bitmap_offset, 1, 0);
2822 if (entry && list_empty(&entry->list))
2823 list_add(&entry->list, bitmaps);
2824 }
2825
2826 list_for_each_entry(entry, bitmaps, list) {
2827 if (entry->bytes < bytes)
2828 continue;
2829 ret = btrfs_bitmap_cluster(block_group, entry, cluster, offset,
2830 bytes, cont1_bytes, min_bytes);
2831 if (!ret)
2832 return 0;
2833 }
2834
2835 /*
2836 * The bitmaps list has all the bitmaps that record free space
2837 * starting after offset, so no more search is required.
2838 */
2839 return -ENOSPC;
2840}
2841
2842/*
2843 * here we try to find a cluster of blocks in a block group. The goal
2844 * is to find at least bytes+empty_size.
2845 * We might not find them all in one contiguous area.
2846 *
2847 * returns zero and sets up cluster if things worked out, otherwise
2848 * it returns -enospc
2849 */
2850int btrfs_find_space_cluster(struct btrfs_root *root,
2851 struct btrfs_block_group_cache *block_group,
2852 struct btrfs_free_cluster *cluster,
2853 u64 offset, u64 bytes, u64 empty_size)
2854{
2855 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2856 struct btrfs_free_space *entry, *tmp;
2857 LIST_HEAD(bitmaps);
2858 u64 min_bytes;
2859 u64 cont1_bytes;
2860 int ret;
2861
2862 /*
2863 * Choose the minimum extent size we'll require for this
2864 * cluster. For SSD_SPREAD, don't allow any fragmentation.
2865 * For metadata, allow allocates with smaller extents. For
2866 * data, keep it dense.
2867 */
2868 if (btrfs_test_opt(root, SSD_SPREAD)) {
2869 cont1_bytes = min_bytes = bytes + empty_size;
2870 } else if (block_group->flags & BTRFS_BLOCK_GROUP_METADATA) {
2871 cont1_bytes = bytes;
2872 min_bytes = block_group->sectorsize;
2873 } else {
2874 cont1_bytes = max(bytes, (bytes + empty_size) >> 2);
2875 min_bytes = block_group->sectorsize;
2876 }
2877
2878 spin_lock(&ctl->tree_lock);
2879
2880 /*
2881 * If we know we don't have enough space to make a cluster don't even
2882 * bother doing all the work to try and find one.
2883 */
2884 if (ctl->free_space < bytes) {
2885 spin_unlock(&ctl->tree_lock);
2886 return -ENOSPC;
2887 }
2888
2889 spin_lock(&cluster->lock);
2890
2891 /* someone already found a cluster, hooray */
2892 if (cluster->block_group) {
2893 ret = 0;
2894 goto out;
2895 }
2896
2897 trace_btrfs_find_cluster(block_group, offset, bytes, empty_size,
2898 min_bytes);
2899
2900 ret = setup_cluster_no_bitmap(block_group, cluster, &bitmaps, offset,
2901 bytes + empty_size,
2902 cont1_bytes, min_bytes);
2903 if (ret)
2904 ret = setup_cluster_bitmap(block_group, cluster, &bitmaps,
2905 offset, bytes + empty_size,
2906 cont1_bytes, min_bytes);
2907
2908 /* Clear our temporary list */
2909 list_for_each_entry_safe(entry, tmp, &bitmaps, list)
2910 list_del_init(&entry->list);
2911
2912 if (!ret) {
2913 atomic_inc(&block_group->count);
2914 list_add_tail(&cluster->block_group_list,
2915 &block_group->cluster_list);
2916 cluster->block_group = block_group;
2917 } else {
2918 trace_btrfs_failed_cluster_setup(block_group);
2919 }
2920out:
2921 spin_unlock(&cluster->lock);
2922 spin_unlock(&ctl->tree_lock);
2923
2924 return ret;
2925}
2926
2927/*
2928 * simple code to zero out a cluster
2929 */
2930void btrfs_init_free_cluster(struct btrfs_free_cluster *cluster)
2931{
2932 spin_lock_init(&cluster->lock);
2933 spin_lock_init(&cluster->refill_lock);
2934 cluster->root = RB_ROOT;
2935 cluster->max_size = 0;
2936 INIT_LIST_HEAD(&cluster->block_group_list);
2937 cluster->block_group = NULL;
2938}
2939
2940static int do_trimming(struct btrfs_block_group_cache *block_group,
2941 u64 *total_trimmed, u64 start, u64 bytes,
2942 u64 reserved_start, u64 reserved_bytes,
2943 struct btrfs_trim_range *trim_entry)
2944{
2945 struct btrfs_space_info *space_info = block_group->space_info;
2946 struct btrfs_fs_info *fs_info = block_group->fs_info;
2947 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2948 int ret;
2949 int update = 0;
2950 u64 trimmed = 0;
2951
2952 spin_lock(&space_info->lock);
2953 spin_lock(&block_group->lock);
2954 if (!block_group->ro) {
2955 block_group->reserved += reserved_bytes;
2956 space_info->bytes_reserved += reserved_bytes;
2957 update = 1;
2958 }
2959 spin_unlock(&block_group->lock);
2960 spin_unlock(&space_info->lock);
2961
2962 ret = btrfs_discard_extent(fs_info->extent_root,
2963 start, bytes, &trimmed);
2964 if (!ret)
2965 *total_trimmed += trimmed;
2966
2967 mutex_lock(&ctl->cache_writeout_mutex);
2968 btrfs_add_free_space(block_group, reserved_start, reserved_bytes);
2969 list_del(&trim_entry->list);
2970 mutex_unlock(&ctl->cache_writeout_mutex);
2971
2972 if (update) {
2973 spin_lock(&space_info->lock);
2974 spin_lock(&block_group->lock);
2975 if (block_group->ro)
2976 space_info->bytes_readonly += reserved_bytes;
2977 block_group->reserved -= reserved_bytes;
2978 space_info->bytes_reserved -= reserved_bytes;
2979 spin_unlock(&space_info->lock);
2980 spin_unlock(&block_group->lock);
2981 }
2982
2983 return ret;
2984}
2985
2986static int trim_no_bitmap(struct btrfs_block_group_cache *block_group,
2987 u64 *total_trimmed, u64 start, u64 end, u64 minlen)
2988{
2989 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2990 struct btrfs_free_space *entry;
2991 struct rb_node *node;
2992 int ret = 0;
2993 u64 extent_start;
2994 u64 extent_bytes;
2995 u64 bytes;
2996
2997 while (start < end) {
2998 struct btrfs_trim_range trim_entry;
2999
3000 mutex_lock(&ctl->cache_writeout_mutex);
3001 spin_lock(&ctl->tree_lock);
3002
3003 if (ctl->free_space < minlen) {
3004 spin_unlock(&ctl->tree_lock);
3005 mutex_unlock(&ctl->cache_writeout_mutex);
3006 break;
3007 }
3008
3009 entry = tree_search_offset(ctl, start, 0, 1);
3010 if (!entry) {
3011 spin_unlock(&ctl->tree_lock);
3012 mutex_unlock(&ctl->cache_writeout_mutex);
3013 break;
3014 }
3015
3016 /* skip bitmaps */
3017 while (entry->bitmap) {
3018 node = rb_next(&entry->offset_index);
3019 if (!node) {
3020 spin_unlock(&ctl->tree_lock);
3021 mutex_unlock(&ctl->cache_writeout_mutex);
3022 goto out;
3023 }
3024 entry = rb_entry(node, struct btrfs_free_space,
3025 offset_index);
3026 }
3027
3028 if (entry->offset >= end) {
3029 spin_unlock(&ctl->tree_lock);
3030 mutex_unlock(&ctl->cache_writeout_mutex);
3031 break;
3032 }
3033
3034 extent_start = entry->offset;
3035 extent_bytes = entry->bytes;
3036 start = max(start, extent_start);
3037 bytes = min(extent_start + extent_bytes, end) - start;
3038 if (bytes < minlen) {
3039 spin_unlock(&ctl->tree_lock);
3040 mutex_unlock(&ctl->cache_writeout_mutex);
3041 goto next;
3042 }
3043
3044 unlink_free_space(ctl, entry);
3045 kmem_cache_free(btrfs_free_space_cachep, entry);
3046
3047 spin_unlock(&ctl->tree_lock);
3048 trim_entry.start = extent_start;
3049 trim_entry.bytes = extent_bytes;
3050 list_add_tail(&trim_entry.list, &ctl->trimming_ranges);
3051 mutex_unlock(&ctl->cache_writeout_mutex);
3052
3053 ret = do_trimming(block_group, total_trimmed, start, bytes,
3054 extent_start, extent_bytes, &trim_entry);
3055 if (ret)
3056 break;
3057next:
3058 start += bytes;
3059
3060 if (fatal_signal_pending(current)) {
3061 ret = -ERESTARTSYS;
3062 break;
3063 }
3064
3065 cond_resched();
3066 }
3067out:
3068 return ret;
3069}
3070
3071static int trim_bitmaps(struct btrfs_block_group_cache *block_group,
3072 u64 *total_trimmed, u64 start, u64 end, u64 minlen)
3073{
3074 struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
3075 struct btrfs_free_space *entry;
3076 int ret = 0;
3077 int ret2;
3078 u64 bytes;
3079 u64 offset = offset_to_bitmap(ctl, start);
3080
3081 while (offset < end) {
3082 bool next_bitmap = false;
3083 struct btrfs_trim_range trim_entry;
3084
3085 mutex_lock(&ctl->cache_writeout_mutex);
3086 spin_lock(&ctl->tree_lock);
3087
3088 if (ctl->free_space < minlen) {
3089 spin_unlock(&ctl->tree_lock);
3090 mutex_unlock(&ctl->cache_writeout_mutex);
3091 break;
3092 }
3093
3094 entry = tree_search_offset(ctl, offset, 1, 0);
3095 if (!entry) {
3096 spin_unlock(&ctl->tree_lock);
3097 mutex_unlock(&ctl->cache_writeout_mutex);
3098 next_bitmap = true;
3099 goto next;
3100 }
3101
3102 bytes = minlen;
3103 ret2 = search_bitmap(ctl, entry, &start, &bytes);
3104 if (ret2 || start >= end) {
3105 spin_unlock(&ctl->tree_lock);
3106 mutex_unlock(&ctl->cache_writeout_mutex);
3107 next_bitmap = true;
3108 goto next;
3109 }
3110
3111 bytes = min(bytes, end - start);
3112 if (bytes < minlen) {
3113 spin_unlock(&ctl->tree_lock);
3114 mutex_unlock(&ctl->cache_writeout_mutex);
3115 goto next;
3116 }
3117
3118 bitmap_clear_bits(ctl, entry, start, bytes);
3119 if (entry->bytes == 0)
3120 free_bitmap(ctl, entry);
3121
3122 spin_unlock(&ctl->tree_lock);
3123 trim_entry.start = start;
3124 trim_entry.bytes = bytes;
3125 list_add_tail(&trim_entry.list, &ctl->trimming_ranges);
3126 mutex_unlock(&ctl->cache_writeout_mutex);
3127
3128 ret = do_trimming(block_group, total_trimmed, start, bytes,
3129 start, bytes, &trim_entry);
3130 if (ret)
3131 break;
3132next:
3133 if (next_bitmap) {
3134 offset += BITS_PER_BITMAP * ctl->unit;
3135 } else {
3136 start += bytes;
3137 if (start >= offset + BITS_PER_BITMAP * ctl->unit)
3138 offset += BITS_PER_BITMAP * ctl->unit;
3139 }
3140
3141 if (fatal_signal_pending(current)) {
3142 ret = -ERESTARTSYS;
3143 break;
3144 }
3145
3146 cond_resched();
3147 }
3148
3149 return ret;
3150}
3151
3152int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group,
3153 u64 *trimmed, u64 start, u64 end, u64 minlen)
3154{
3155 int ret;
3156
3157 *trimmed = 0;
3158
3159 spin_lock(&block_group->lock);
3160 if (block_group->removed) {
3161 spin_unlock(&block_group->lock);
3162 return 0;
3163 }
3164 atomic_inc(&block_group->trimming);
3165 spin_unlock(&block_group->lock);
3166
3167 ret = trim_no_bitmap(block_group, trimmed, start, end, minlen);
3168 if (ret)
3169 goto out;
3170
3171 ret = trim_bitmaps(block_group, trimmed, start, end, minlen);
3172out:
3173 spin_lock(&block_group->lock);
3174 if (atomic_dec_and_test(&block_group->trimming) &&
3175 block_group->removed) {
3176 struct extent_map_tree *em_tree;
3177 struct extent_map *em;
3178
3179 spin_unlock(&block_group->lock);
3180
3181 lock_chunks(block_group->fs_info->chunk_root);
3182 em_tree = &block_group->fs_info->mapping_tree.map_tree;
3183 write_lock(&em_tree->lock);
3184 em = lookup_extent_mapping(em_tree, block_group->key.objectid,
3185 1);
3186 BUG_ON(!em); /* logic error, can't happen */
3187 /*
3188 * remove_extent_mapping() will delete us from the pinned_chunks
3189 * list, which is protected by the chunk mutex.
3190 */
3191 remove_extent_mapping(em_tree, em);
3192 write_unlock(&em_tree->lock);
3193 unlock_chunks(block_group->fs_info->chunk_root);
3194
3195 /* once for us and once for the tree */
3196 free_extent_map(em);
3197 free_extent_map(em);
3198
3199 /*
3200 * We've left one free space entry and other tasks trimming
3201 * this block group have left 1 entry each one. Free them.
3202 */
3203 __btrfs_remove_free_space_cache(block_group->free_space_ctl);
3204 } else {
3205 spin_unlock(&block_group->lock);
3206 }
3207
3208 return ret;
3209}
3210
3211/*
3212 * Find the left-most item in the cache tree, and then return the
3213 * smallest inode number in the item.
3214 *
3215 * Note: the returned inode number may not be the smallest one in
3216 * the tree, if the left-most item is a bitmap.
3217 */
3218u64 btrfs_find_ino_for_alloc(struct btrfs_root *fs_root)
3219{
3220 struct btrfs_free_space_ctl *ctl = fs_root->free_ino_ctl;
3221 struct btrfs_free_space *entry = NULL;
3222 u64 ino = 0;
3223
3224 spin_lock(&ctl->tree_lock);
3225
3226 if (RB_EMPTY_ROOT(&ctl->free_space_offset))
3227 goto out;
3228
3229 entry = rb_entry(rb_first(&ctl->free_space_offset),
3230 struct btrfs_free_space, offset_index);
3231
3232 if (!entry->bitmap) {
3233 ino = entry->offset;
3234
3235 unlink_free_space(ctl, entry);
3236 entry->offset++;
3237 entry->bytes--;
3238 if (!entry->bytes)
3239 kmem_cache_free(btrfs_free_space_cachep, entry);
3240 else
3241 link_free_space(ctl, entry);
3242 } else {
3243 u64 offset = 0;
3244 u64 count = 1;
3245 int ret;
3246
3247 ret = search_bitmap(ctl, entry, &offset, &count);
3248 /* Logic error; Should be empty if it can't find anything */
3249 ASSERT(!ret);
3250
3251 ino = offset;
3252 bitmap_clear_bits(ctl, entry, offset, 1);
3253 if (entry->bytes == 0)
3254 free_bitmap(ctl, entry);
3255 }
3256out:
3257 spin_unlock(&ctl->tree_lock);
3258
3259 return ino;
3260}
3261
3262struct inode *lookup_free_ino_inode(struct btrfs_root *root,
3263 struct btrfs_path *path)
3264{
3265 struct inode *inode = NULL;
3266
3267 spin_lock(&root->ino_cache_lock);
3268 if (root->ino_cache_inode)
3269 inode = igrab(root->ino_cache_inode);
3270 spin_unlock(&root->ino_cache_lock);
3271 if (inode)
3272 return inode;
3273
3274 inode = __lookup_free_space_inode(root, path, 0);
3275 if (IS_ERR(inode))
3276 return inode;
3277
3278 spin_lock(&root->ino_cache_lock);
3279 if (!btrfs_fs_closing(root->fs_info))
3280 root->ino_cache_inode = igrab(inode);
3281 spin_unlock(&root->ino_cache_lock);
3282
3283 return inode;
3284}
3285
3286int create_free_ino_inode(struct btrfs_root *root,
3287 struct btrfs_trans_handle *trans,
3288 struct btrfs_path *path)
3289{
3290 return __create_free_space_inode(root, trans, path,
3291 BTRFS_FREE_INO_OBJECTID, 0);
3292}
3293
3294int load_free_ino_cache(struct btrfs_fs_info *fs_info, struct btrfs_root *root)
3295{
3296 struct btrfs_free_space_ctl *ctl = root->free_ino_ctl;
3297 struct btrfs_path *path;
3298 struct inode *inode;
3299 int ret = 0;
3300 u64 root_gen = btrfs_root_generation(&root->root_item);
3301
3302 if (!btrfs_test_opt(root, INODE_MAP_CACHE))
3303 return 0;
3304
3305 /*
3306 * If we're unmounting then just return, since this does a search on the
3307 * normal root and not the commit root and we could deadlock.
3308 */
3309 if (btrfs_fs_closing(fs_info))
3310 return 0;
3311
3312 path = btrfs_alloc_path();
3313 if (!path)
3314 return 0;
3315
3316 inode = lookup_free_ino_inode(root, path);
3317 if (IS_ERR(inode))
3318 goto out;
3319
3320 if (root_gen != BTRFS_I(inode)->generation)
3321 goto out_put;
3322
3323 ret = __load_free_space_cache(root, inode, ctl, path, 0);
3324
3325 if (ret < 0)
3326 btrfs_err(fs_info,
3327 "failed to load free ino cache for root %llu",
3328 root->root_key.objectid);
3329out_put:
3330 iput(inode);
3331out:
3332 btrfs_free_path(path);
3333 return ret;
3334}
3335
3336int btrfs_write_out_ino_cache(struct btrfs_root *root,
3337 struct btrfs_trans_handle *trans,
3338 struct btrfs_path *path,
3339 struct inode *inode)
3340{
3341 struct btrfs_free_space_ctl *ctl = root->free_ino_ctl;
3342 int ret;
3343
3344 if (!btrfs_test_opt(root, INODE_MAP_CACHE))
3345 return 0;
3346
3347 ret = __btrfs_write_out_cache(root, inode, ctl, NULL, trans, path, 0);
3348 if (ret) {
3349 btrfs_delalloc_release_metadata(inode, inode->i_size);
3350#ifdef DEBUG
3351 btrfs_err(root->fs_info,
3352 "failed to write free ino cache for root %llu",
3353 root->root_key.objectid);
3354#endif
3355 }
3356
3357 return ret;
3358}
3359
3360#ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS
3361/*
3362 * Use this if you need to make a bitmap or extent entry specifically, it
3363 * doesn't do any of the merging that add_free_space does, this acts a lot like
3364 * how the free space cache loading stuff works, so you can get really weird
3365 * configurations.
3366 */
3367int test_add_free_space_entry(struct btrfs_block_group_cache *cache,
3368 u64 offset, u64 bytes, bool bitmap)
3369{
3370 struct btrfs_free_space_ctl *ctl = cache->free_space_ctl;
3371 struct btrfs_free_space *info = NULL, *bitmap_info;
3372 void *map = NULL;
3373 u64 bytes_added;
3374 int ret;
3375
3376again:
3377 if (!info) {
3378 info = kmem_cache_zalloc(btrfs_free_space_cachep, GFP_NOFS);
3379 if (!info)
3380 return -ENOMEM;
3381 }
3382
3383 if (!bitmap) {
3384 spin_lock(&ctl->tree_lock);
3385 info->offset = offset;
3386 info->bytes = bytes;
3387 ret = link_free_space(ctl, info);
3388 spin_unlock(&ctl->tree_lock);
3389 if (ret)
3390 kmem_cache_free(btrfs_free_space_cachep, info);
3391 return ret;
3392 }
3393
3394 if (!map) {
3395 map = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS);
3396 if (!map) {
3397 kmem_cache_free(btrfs_free_space_cachep, info);
3398 return -ENOMEM;
3399 }
3400 }
3401
3402 spin_lock(&ctl->tree_lock);
3403 bitmap_info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset),
3404 1, 0);
3405 if (!bitmap_info) {
3406 info->bitmap = map;
3407 map = NULL;
3408 add_new_bitmap(ctl, info, offset);
3409 bitmap_info = info;
3410 info = NULL;
3411 }
3412
3413 bytes_added = add_bytes_to_bitmap(ctl, bitmap_info, offset, bytes);
3414 bytes -= bytes_added;
3415 offset += bytes_added;
3416 spin_unlock(&ctl->tree_lock);
3417
3418 if (bytes)
3419 goto again;
3420
3421 if (info)
3422 kmem_cache_free(btrfs_free_space_cachep, info);
3423 if (map)
3424 kfree(map);
3425 return 0;
3426}
3427
3428/*
3429 * Checks to see if the given range is in the free space cache. This is really
3430 * just used to check the absence of space, so if there is free space in the
3431 * range at all we will return 1.
3432 */
3433int test_check_exists(struct btrfs_block_group_cache *cache,
3434 u64 offset, u64 bytes)
3435{
3436 struct btrfs_free_space_ctl *ctl = cache->free_space_ctl;
3437 struct btrfs_free_space *info;
3438 int ret = 0;
3439
3440 spin_lock(&ctl->tree_lock);
3441 info = tree_search_offset(ctl, offset, 0, 0);
3442 if (!info) {
3443 info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset),
3444 1, 0);
3445 if (!info)
3446 goto out;
3447 }
3448
3449have_info:
3450 if (info->bitmap) {
3451 u64 bit_off, bit_bytes;
3452 struct rb_node *n;
3453 struct btrfs_free_space *tmp;
3454
3455 bit_off = offset;
3456 bit_bytes = ctl->unit;
3457 ret = search_bitmap(ctl, info, &bit_off, &bit_bytes);
3458 if (!ret) {
3459 if (bit_off == offset) {
3460 ret = 1;
3461 goto out;
3462 } else if (bit_off > offset &&
3463 offset + bytes > bit_off) {
3464 ret = 1;
3465 goto out;
3466 }
3467 }
3468
3469 n = rb_prev(&info->offset_index);
3470 while (n) {
3471 tmp = rb_entry(n, struct btrfs_free_space,
3472 offset_index);
3473 if (tmp->offset + tmp->bytes < offset)
3474 break;
3475 if (offset + bytes < tmp->offset) {
3476 n = rb_prev(&info->offset_index);
3477 continue;
3478 }
3479 info = tmp;
3480 goto have_info;
3481 }
3482
3483 n = rb_next(&info->offset_index);
3484 while (n) {
3485 tmp = rb_entry(n, struct btrfs_free_space,
3486 offset_index);
3487 if (offset + bytes < tmp->offset)
3488 break;
3489 if (tmp->offset + tmp->bytes < offset) {
3490 n = rb_next(&info->offset_index);
3491 continue;
3492 }
3493 info = tmp;
3494 goto have_info;
3495 }
3496
3497 ret = 0;
3498 goto out;
3499 }
3500
3501 if (info->offset == offset) {
3502 ret = 1;
3503 goto out;
3504 }
3505
3506 if (offset > info->offset && offset < info->offset + info->bytes)
3507 ret = 1;
3508out:
3509 spin_unlock(&ctl->tree_lock);
3510 return ret;
3511}
3512#endif /* CONFIG_BTRFS_FS_RUN_SANITY_TESTS */
This page took 0.047064 seconds and 5 git commands to generate.