xref: /openbmc/linux/fs/btrfs/free-space-cache.c (revision f7d84fa7)
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/sched/signal.h>
22 #include <linux/slab.h>
23 #include <linux/math64.h>
24 #include <linux/ratelimit.h>
25 #include "ctree.h"
26 #include "free-space-cache.h"
27 #include "transaction.h"
28 #include "disk-io.h"
29 #include "extent_io.h"
30 #include "inode-map.h"
31 #include "volumes.h"
32 
33 #define BITS_PER_BITMAP		(PAGE_SIZE * 8UL)
34 #define MAX_CACHE_BYTES_PER_GIG	SZ_32K
35 
36 struct btrfs_trim_range {
37 	u64 start;
38 	u64 bytes;
39 	struct list_head list;
40 };
41 
42 static int link_free_space(struct btrfs_free_space_ctl *ctl,
43 			   struct btrfs_free_space *info);
44 static void unlink_free_space(struct btrfs_free_space_ctl *ctl,
45 			      struct btrfs_free_space *info);
46 static int btrfs_wait_cache_io_root(struct btrfs_root *root,
47 			     struct btrfs_trans_handle *trans,
48 			     struct btrfs_io_ctl *io_ctl,
49 			     struct btrfs_path *path);
50 
51 static struct inode *__lookup_free_space_inode(struct btrfs_root *root,
52 					       struct btrfs_path *path,
53 					       u64 offset)
54 {
55 	struct btrfs_fs_info *fs_info = root->fs_info;
56 	struct btrfs_key key;
57 	struct btrfs_key location;
58 	struct btrfs_disk_key disk_key;
59 	struct btrfs_free_space_header *header;
60 	struct extent_buffer *leaf;
61 	struct inode *inode = NULL;
62 	int ret;
63 
64 	key.objectid = BTRFS_FREE_SPACE_OBJECTID;
65 	key.offset = offset;
66 	key.type = 0;
67 
68 	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
69 	if (ret < 0)
70 		return ERR_PTR(ret);
71 	if (ret > 0) {
72 		btrfs_release_path(path);
73 		return ERR_PTR(-ENOENT);
74 	}
75 
76 	leaf = path->nodes[0];
77 	header = btrfs_item_ptr(leaf, path->slots[0],
78 				struct btrfs_free_space_header);
79 	btrfs_free_space_key(leaf, header, &disk_key);
80 	btrfs_disk_key_to_cpu(&location, &disk_key);
81 	btrfs_release_path(path);
82 
83 	inode = btrfs_iget(fs_info->sb, &location, root, NULL);
84 	if (IS_ERR(inode))
85 		return inode;
86 	if (is_bad_inode(inode)) {
87 		iput(inode);
88 		return ERR_PTR(-ENOENT);
89 	}
90 
91 	mapping_set_gfp_mask(inode->i_mapping,
92 			mapping_gfp_constraint(inode->i_mapping,
93 			~(__GFP_FS | __GFP_HIGHMEM)));
94 
95 	return inode;
96 }
97 
98 struct inode *lookup_free_space_inode(struct btrfs_fs_info *fs_info,
99 				      struct btrfs_block_group_cache
100 				      *block_group, struct btrfs_path *path)
101 {
102 	struct inode *inode = NULL;
103 	u32 flags = BTRFS_INODE_NODATASUM | BTRFS_INODE_NODATACOW;
104 
105 	spin_lock(&block_group->lock);
106 	if (block_group->inode)
107 		inode = igrab(block_group->inode);
108 	spin_unlock(&block_group->lock);
109 	if (inode)
110 		return inode;
111 
112 	inode = __lookup_free_space_inode(fs_info->tree_root, path,
113 					  block_group->key.objectid);
114 	if (IS_ERR(inode))
115 		return inode;
116 
117 	spin_lock(&block_group->lock);
118 	if (!((BTRFS_I(inode)->flags & flags) == flags)) {
119 		btrfs_info(fs_info, "Old style space inode found, converting.");
120 		BTRFS_I(inode)->flags |= BTRFS_INODE_NODATASUM |
121 			BTRFS_INODE_NODATACOW;
122 		block_group->disk_cache_state = BTRFS_DC_CLEAR;
123 	}
124 
125 	if (!block_group->iref) {
126 		block_group->inode = igrab(inode);
127 		block_group->iref = 1;
128 	}
129 	spin_unlock(&block_group->lock);
130 
131 	return inode;
132 }
133 
134 static int __create_free_space_inode(struct btrfs_root *root,
135 				     struct btrfs_trans_handle *trans,
136 				     struct btrfs_path *path,
137 				     u64 ino, u64 offset)
138 {
139 	struct btrfs_key key;
140 	struct btrfs_disk_key disk_key;
141 	struct btrfs_free_space_header *header;
142 	struct btrfs_inode_item *inode_item;
143 	struct extent_buffer *leaf;
144 	u64 flags = BTRFS_INODE_NOCOMPRESS | BTRFS_INODE_PREALLOC;
145 	int ret;
146 
147 	ret = btrfs_insert_empty_inode(trans, root, path, ino);
148 	if (ret)
149 		return ret;
150 
151 	/* We inline crc's for the free disk space cache */
152 	if (ino != BTRFS_FREE_INO_OBJECTID)
153 		flags |= BTRFS_INODE_NODATASUM | BTRFS_INODE_NODATACOW;
154 
155 	leaf = path->nodes[0];
156 	inode_item = btrfs_item_ptr(leaf, path->slots[0],
157 				    struct btrfs_inode_item);
158 	btrfs_item_key(leaf, &disk_key, path->slots[0]);
159 	memzero_extent_buffer(leaf, (unsigned long)inode_item,
160 			     sizeof(*inode_item));
161 	btrfs_set_inode_generation(leaf, inode_item, trans->transid);
162 	btrfs_set_inode_size(leaf, inode_item, 0);
163 	btrfs_set_inode_nbytes(leaf, inode_item, 0);
164 	btrfs_set_inode_uid(leaf, inode_item, 0);
165 	btrfs_set_inode_gid(leaf, inode_item, 0);
166 	btrfs_set_inode_mode(leaf, inode_item, S_IFREG | 0600);
167 	btrfs_set_inode_flags(leaf, inode_item, flags);
168 	btrfs_set_inode_nlink(leaf, inode_item, 1);
169 	btrfs_set_inode_transid(leaf, inode_item, trans->transid);
170 	btrfs_set_inode_block_group(leaf, inode_item, offset);
171 	btrfs_mark_buffer_dirty(leaf);
172 	btrfs_release_path(path);
173 
174 	key.objectid = BTRFS_FREE_SPACE_OBJECTID;
175 	key.offset = offset;
176 	key.type = 0;
177 	ret = btrfs_insert_empty_item(trans, root, path, &key,
178 				      sizeof(struct btrfs_free_space_header));
179 	if (ret < 0) {
180 		btrfs_release_path(path);
181 		return ret;
182 	}
183 
184 	leaf = path->nodes[0];
185 	header = btrfs_item_ptr(leaf, path->slots[0],
186 				struct btrfs_free_space_header);
187 	memzero_extent_buffer(leaf, (unsigned long)header, sizeof(*header));
188 	btrfs_set_free_space_key(leaf, header, &disk_key);
189 	btrfs_mark_buffer_dirty(leaf);
190 	btrfs_release_path(path);
191 
192 	return 0;
193 }
194 
195 int create_free_space_inode(struct btrfs_fs_info *fs_info,
196 			    struct btrfs_trans_handle *trans,
197 			    struct btrfs_block_group_cache *block_group,
198 			    struct btrfs_path *path)
199 {
200 	int ret;
201 	u64 ino;
202 
203 	ret = btrfs_find_free_objectid(fs_info->tree_root, &ino);
204 	if (ret < 0)
205 		return ret;
206 
207 	return __create_free_space_inode(fs_info->tree_root, trans, path, ino,
208 					 block_group->key.objectid);
209 }
210 
211 int btrfs_check_trunc_cache_free_space(struct btrfs_fs_info *fs_info,
212 				       struct btrfs_block_rsv *rsv)
213 {
214 	u64 needed_bytes;
215 	int ret;
216 
217 	/* 1 for slack space, 1 for updating the inode */
218 	needed_bytes = btrfs_calc_trunc_metadata_size(fs_info, 1) +
219 		btrfs_calc_trans_metadata_size(fs_info, 1);
220 
221 	spin_lock(&rsv->lock);
222 	if (rsv->reserved < needed_bytes)
223 		ret = -ENOSPC;
224 	else
225 		ret = 0;
226 	spin_unlock(&rsv->lock);
227 	return ret;
228 }
229 
230 int btrfs_truncate_free_space_cache(struct btrfs_trans_handle *trans,
231 				    struct btrfs_block_group_cache *block_group,
232 				    struct inode *inode)
233 {
234 	struct btrfs_root *root = BTRFS_I(inode)->root;
235 	int ret = 0;
236 	bool locked = false;
237 
238 	if (block_group) {
239 		struct btrfs_path *path = btrfs_alloc_path();
240 
241 		if (!path) {
242 			ret = -ENOMEM;
243 			goto fail;
244 		}
245 		locked = true;
246 		mutex_lock(&trans->transaction->cache_write_mutex);
247 		if (!list_empty(&block_group->io_list)) {
248 			list_del_init(&block_group->io_list);
249 
250 			btrfs_wait_cache_io(trans, block_group, path);
251 			btrfs_put_block_group(block_group);
252 		}
253 
254 		/*
255 		 * now that we've truncated the cache away, its no longer
256 		 * setup or written
257 		 */
258 		spin_lock(&block_group->lock);
259 		block_group->disk_cache_state = BTRFS_DC_CLEAR;
260 		spin_unlock(&block_group->lock);
261 		btrfs_free_path(path);
262 	}
263 
264 	btrfs_i_size_write(BTRFS_I(inode), 0);
265 	truncate_pagecache(inode, 0);
266 
267 	/*
268 	 * We don't need an orphan item because truncating the free space cache
269 	 * will never be split across transactions.
270 	 * We don't need to check for -EAGAIN because we're a free space
271 	 * cache inode
272 	 */
273 	ret = btrfs_truncate_inode_items(trans, root, inode,
274 					 0, BTRFS_EXTENT_DATA_KEY);
275 	if (ret)
276 		goto fail;
277 
278 	ret = btrfs_update_inode(trans, root, inode);
279 
280 fail:
281 	if (locked)
282 		mutex_unlock(&trans->transaction->cache_write_mutex);
283 	if (ret)
284 		btrfs_abort_transaction(trans, ret);
285 
286 	return ret;
287 }
288 
289 static void readahead_cache(struct inode *inode)
290 {
291 	struct file_ra_state *ra;
292 	unsigned long last_index;
293 
294 	ra = kzalloc(sizeof(*ra), GFP_NOFS);
295 	if (!ra)
296 		return;
297 
298 	file_ra_state_init(ra, inode->i_mapping);
299 	last_index = (i_size_read(inode) - 1) >> PAGE_SHIFT;
300 
301 	page_cache_sync_readahead(inode->i_mapping, ra, NULL, 0, last_index);
302 
303 	kfree(ra);
304 }
305 
306 static int io_ctl_init(struct btrfs_io_ctl *io_ctl, struct inode *inode,
307 		       int write)
308 {
309 	int num_pages;
310 	int check_crcs = 0;
311 
312 	num_pages = DIV_ROUND_UP(i_size_read(inode), PAGE_SIZE);
313 
314 	if (btrfs_ino(BTRFS_I(inode)) != BTRFS_FREE_INO_OBJECTID)
315 		check_crcs = 1;
316 
317 	/* Make sure we can fit our crcs into the first page */
318 	if (write && check_crcs &&
319 	    (num_pages * sizeof(u32)) >= PAGE_SIZE)
320 		return -ENOSPC;
321 
322 	memset(io_ctl, 0, sizeof(struct btrfs_io_ctl));
323 
324 	io_ctl->pages = kcalloc(num_pages, sizeof(struct page *), GFP_NOFS);
325 	if (!io_ctl->pages)
326 		return -ENOMEM;
327 
328 	io_ctl->num_pages = num_pages;
329 	io_ctl->fs_info = btrfs_sb(inode->i_sb);
330 	io_ctl->check_crcs = check_crcs;
331 	io_ctl->inode = inode;
332 
333 	return 0;
334 }
335 
336 static void io_ctl_free(struct btrfs_io_ctl *io_ctl)
337 {
338 	kfree(io_ctl->pages);
339 	io_ctl->pages = NULL;
340 }
341 
342 static void io_ctl_unmap_page(struct btrfs_io_ctl *io_ctl)
343 {
344 	if (io_ctl->cur) {
345 		io_ctl->cur = NULL;
346 		io_ctl->orig = NULL;
347 	}
348 }
349 
350 static void io_ctl_map_page(struct btrfs_io_ctl *io_ctl, int clear)
351 {
352 	ASSERT(io_ctl->index < io_ctl->num_pages);
353 	io_ctl->page = io_ctl->pages[io_ctl->index++];
354 	io_ctl->cur = page_address(io_ctl->page);
355 	io_ctl->orig = io_ctl->cur;
356 	io_ctl->size = PAGE_SIZE;
357 	if (clear)
358 		clear_page(io_ctl->cur);
359 }
360 
361 static void io_ctl_drop_pages(struct btrfs_io_ctl *io_ctl)
362 {
363 	int i;
364 
365 	io_ctl_unmap_page(io_ctl);
366 
367 	for (i = 0; i < io_ctl->num_pages; i++) {
368 		if (io_ctl->pages[i]) {
369 			ClearPageChecked(io_ctl->pages[i]);
370 			unlock_page(io_ctl->pages[i]);
371 			put_page(io_ctl->pages[i]);
372 		}
373 	}
374 }
375 
376 static int io_ctl_prepare_pages(struct btrfs_io_ctl *io_ctl, struct inode *inode,
377 				int uptodate)
378 {
379 	struct page *page;
380 	gfp_t mask = btrfs_alloc_write_mask(inode->i_mapping);
381 	int i;
382 
383 	for (i = 0; i < io_ctl->num_pages; i++) {
384 		page = find_or_create_page(inode->i_mapping, i, mask);
385 		if (!page) {
386 			io_ctl_drop_pages(io_ctl);
387 			return -ENOMEM;
388 		}
389 		io_ctl->pages[i] = page;
390 		if (uptodate && !PageUptodate(page)) {
391 			btrfs_readpage(NULL, page);
392 			lock_page(page);
393 			if (!PageUptodate(page)) {
394 				btrfs_err(BTRFS_I(inode)->root->fs_info,
395 					   "error reading free space cache");
396 				io_ctl_drop_pages(io_ctl);
397 				return -EIO;
398 			}
399 		}
400 	}
401 
402 	for (i = 0; i < io_ctl->num_pages; i++) {
403 		clear_page_dirty_for_io(io_ctl->pages[i]);
404 		set_page_extent_mapped(io_ctl->pages[i]);
405 	}
406 
407 	return 0;
408 }
409 
410 static void io_ctl_set_generation(struct btrfs_io_ctl *io_ctl, u64 generation)
411 {
412 	__le64 *val;
413 
414 	io_ctl_map_page(io_ctl, 1);
415 
416 	/*
417 	 * Skip the csum areas.  If we don't check crcs then we just have a
418 	 * 64bit chunk at the front of the first page.
419 	 */
420 	if (io_ctl->check_crcs) {
421 		io_ctl->cur += (sizeof(u32) * io_ctl->num_pages);
422 		io_ctl->size -= sizeof(u64) + (sizeof(u32) * io_ctl->num_pages);
423 	} else {
424 		io_ctl->cur += sizeof(u64);
425 		io_ctl->size -= sizeof(u64) * 2;
426 	}
427 
428 	val = io_ctl->cur;
429 	*val = cpu_to_le64(generation);
430 	io_ctl->cur += sizeof(u64);
431 }
432 
433 static int io_ctl_check_generation(struct btrfs_io_ctl *io_ctl, u64 generation)
434 {
435 	__le64 *gen;
436 
437 	/*
438 	 * Skip the crc area.  If we don't check crcs then we just have a 64bit
439 	 * chunk at the front of the first page.
440 	 */
441 	if (io_ctl->check_crcs) {
442 		io_ctl->cur += sizeof(u32) * io_ctl->num_pages;
443 		io_ctl->size -= sizeof(u64) +
444 			(sizeof(u32) * io_ctl->num_pages);
445 	} else {
446 		io_ctl->cur += sizeof(u64);
447 		io_ctl->size -= sizeof(u64) * 2;
448 	}
449 
450 	gen = io_ctl->cur;
451 	if (le64_to_cpu(*gen) != generation) {
452 		btrfs_err_rl(io_ctl->fs_info,
453 			"space cache generation (%llu) does not match inode (%llu)",
454 				*gen, generation);
455 		io_ctl_unmap_page(io_ctl);
456 		return -EIO;
457 	}
458 	io_ctl->cur += sizeof(u64);
459 	return 0;
460 }
461 
462 static void io_ctl_set_crc(struct btrfs_io_ctl *io_ctl, int index)
463 {
464 	u32 *tmp;
465 	u32 crc = ~(u32)0;
466 	unsigned offset = 0;
467 
468 	if (!io_ctl->check_crcs) {
469 		io_ctl_unmap_page(io_ctl);
470 		return;
471 	}
472 
473 	if (index == 0)
474 		offset = sizeof(u32) * io_ctl->num_pages;
475 
476 	crc = btrfs_csum_data(io_ctl->orig + offset, crc,
477 			      PAGE_SIZE - offset);
478 	btrfs_csum_final(crc, (u8 *)&crc);
479 	io_ctl_unmap_page(io_ctl);
480 	tmp = page_address(io_ctl->pages[0]);
481 	tmp += index;
482 	*tmp = crc;
483 }
484 
485 static int io_ctl_check_crc(struct btrfs_io_ctl *io_ctl, int index)
486 {
487 	u32 *tmp, val;
488 	u32 crc = ~(u32)0;
489 	unsigned offset = 0;
490 
491 	if (!io_ctl->check_crcs) {
492 		io_ctl_map_page(io_ctl, 0);
493 		return 0;
494 	}
495 
496 	if (index == 0)
497 		offset = sizeof(u32) * io_ctl->num_pages;
498 
499 	tmp = page_address(io_ctl->pages[0]);
500 	tmp += index;
501 	val = *tmp;
502 
503 	io_ctl_map_page(io_ctl, 0);
504 	crc = btrfs_csum_data(io_ctl->orig + offset, crc,
505 			      PAGE_SIZE - offset);
506 	btrfs_csum_final(crc, (u8 *)&crc);
507 	if (val != crc) {
508 		btrfs_err_rl(io_ctl->fs_info,
509 			"csum mismatch on free space cache");
510 		io_ctl_unmap_page(io_ctl);
511 		return -EIO;
512 	}
513 
514 	return 0;
515 }
516 
517 static int io_ctl_add_entry(struct btrfs_io_ctl *io_ctl, u64 offset, u64 bytes,
518 			    void *bitmap)
519 {
520 	struct btrfs_free_space_entry *entry;
521 
522 	if (!io_ctl->cur)
523 		return -ENOSPC;
524 
525 	entry = io_ctl->cur;
526 	entry->offset = cpu_to_le64(offset);
527 	entry->bytes = cpu_to_le64(bytes);
528 	entry->type = (bitmap) ? BTRFS_FREE_SPACE_BITMAP :
529 		BTRFS_FREE_SPACE_EXTENT;
530 	io_ctl->cur += sizeof(struct btrfs_free_space_entry);
531 	io_ctl->size -= sizeof(struct btrfs_free_space_entry);
532 
533 	if (io_ctl->size >= sizeof(struct btrfs_free_space_entry))
534 		return 0;
535 
536 	io_ctl_set_crc(io_ctl, io_ctl->index - 1);
537 
538 	/* No more pages to map */
539 	if (io_ctl->index >= io_ctl->num_pages)
540 		return 0;
541 
542 	/* map the next page */
543 	io_ctl_map_page(io_ctl, 1);
544 	return 0;
545 }
546 
547 static int io_ctl_add_bitmap(struct btrfs_io_ctl *io_ctl, void *bitmap)
548 {
549 	if (!io_ctl->cur)
550 		return -ENOSPC;
551 
552 	/*
553 	 * If we aren't at the start of the current page, unmap this one and
554 	 * map the next one if there is any left.
555 	 */
556 	if (io_ctl->cur != io_ctl->orig) {
557 		io_ctl_set_crc(io_ctl, io_ctl->index - 1);
558 		if (io_ctl->index >= io_ctl->num_pages)
559 			return -ENOSPC;
560 		io_ctl_map_page(io_ctl, 0);
561 	}
562 
563 	memcpy(io_ctl->cur, bitmap, PAGE_SIZE);
564 	io_ctl_set_crc(io_ctl, io_ctl->index - 1);
565 	if (io_ctl->index < io_ctl->num_pages)
566 		io_ctl_map_page(io_ctl, 0);
567 	return 0;
568 }
569 
570 static void io_ctl_zero_remaining_pages(struct btrfs_io_ctl *io_ctl)
571 {
572 	/*
573 	 * If we're not on the boundary we know we've modified the page and we
574 	 * need to crc the page.
575 	 */
576 	if (io_ctl->cur != io_ctl->orig)
577 		io_ctl_set_crc(io_ctl, io_ctl->index - 1);
578 	else
579 		io_ctl_unmap_page(io_ctl);
580 
581 	while (io_ctl->index < io_ctl->num_pages) {
582 		io_ctl_map_page(io_ctl, 1);
583 		io_ctl_set_crc(io_ctl, io_ctl->index - 1);
584 	}
585 }
586 
587 static int io_ctl_read_entry(struct btrfs_io_ctl *io_ctl,
588 			    struct btrfs_free_space *entry, u8 *type)
589 {
590 	struct btrfs_free_space_entry *e;
591 	int ret;
592 
593 	if (!io_ctl->cur) {
594 		ret = io_ctl_check_crc(io_ctl, io_ctl->index);
595 		if (ret)
596 			return ret;
597 	}
598 
599 	e = io_ctl->cur;
600 	entry->offset = le64_to_cpu(e->offset);
601 	entry->bytes = le64_to_cpu(e->bytes);
602 	*type = e->type;
603 	io_ctl->cur += sizeof(struct btrfs_free_space_entry);
604 	io_ctl->size -= sizeof(struct btrfs_free_space_entry);
605 
606 	if (io_ctl->size >= sizeof(struct btrfs_free_space_entry))
607 		return 0;
608 
609 	io_ctl_unmap_page(io_ctl);
610 
611 	return 0;
612 }
613 
614 static int io_ctl_read_bitmap(struct btrfs_io_ctl *io_ctl,
615 			      struct btrfs_free_space *entry)
616 {
617 	int ret;
618 
619 	ret = io_ctl_check_crc(io_ctl, io_ctl->index);
620 	if (ret)
621 		return ret;
622 
623 	memcpy(entry->bitmap, io_ctl->cur, PAGE_SIZE);
624 	io_ctl_unmap_page(io_ctl);
625 
626 	return 0;
627 }
628 
629 /*
630  * Since we attach pinned extents after the fact we can have contiguous sections
631  * of free space that are split up in entries.  This poses a problem with the
632  * tree logging stuff since it could have allocated across what appears to be 2
633  * entries since we would have merged the entries when adding the pinned extents
634  * back to the free space cache.  So run through the space cache that we just
635  * loaded and merge contiguous entries.  This will make the log replay stuff not
636  * blow up and it will make for nicer allocator behavior.
637  */
638 static void merge_space_tree(struct btrfs_free_space_ctl *ctl)
639 {
640 	struct btrfs_free_space *e, *prev = NULL;
641 	struct rb_node *n;
642 
643 again:
644 	spin_lock(&ctl->tree_lock);
645 	for (n = rb_first(&ctl->free_space_offset); n; n = rb_next(n)) {
646 		e = rb_entry(n, struct btrfs_free_space, offset_index);
647 		if (!prev)
648 			goto next;
649 		if (e->bitmap || prev->bitmap)
650 			goto next;
651 		if (prev->offset + prev->bytes == e->offset) {
652 			unlink_free_space(ctl, prev);
653 			unlink_free_space(ctl, e);
654 			prev->bytes += e->bytes;
655 			kmem_cache_free(btrfs_free_space_cachep, e);
656 			link_free_space(ctl, prev);
657 			prev = NULL;
658 			spin_unlock(&ctl->tree_lock);
659 			goto again;
660 		}
661 next:
662 		prev = e;
663 	}
664 	spin_unlock(&ctl->tree_lock);
665 }
666 
667 static int __load_free_space_cache(struct btrfs_root *root, struct inode *inode,
668 				   struct btrfs_free_space_ctl *ctl,
669 				   struct btrfs_path *path, u64 offset)
670 {
671 	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
672 	struct btrfs_free_space_header *header;
673 	struct extent_buffer *leaf;
674 	struct btrfs_io_ctl io_ctl;
675 	struct btrfs_key key;
676 	struct btrfs_free_space *e, *n;
677 	LIST_HEAD(bitmaps);
678 	u64 num_entries;
679 	u64 num_bitmaps;
680 	u64 generation;
681 	u8 type;
682 	int ret = 0;
683 
684 	/* Nothing in the space cache, goodbye */
685 	if (!i_size_read(inode))
686 		return 0;
687 
688 	key.objectid = BTRFS_FREE_SPACE_OBJECTID;
689 	key.offset = offset;
690 	key.type = 0;
691 
692 	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
693 	if (ret < 0)
694 		return 0;
695 	else if (ret > 0) {
696 		btrfs_release_path(path);
697 		return 0;
698 	}
699 
700 	ret = -1;
701 
702 	leaf = path->nodes[0];
703 	header = btrfs_item_ptr(leaf, path->slots[0],
704 				struct btrfs_free_space_header);
705 	num_entries = btrfs_free_space_entries(leaf, header);
706 	num_bitmaps = btrfs_free_space_bitmaps(leaf, header);
707 	generation = btrfs_free_space_generation(leaf, header);
708 	btrfs_release_path(path);
709 
710 	if (!BTRFS_I(inode)->generation) {
711 		btrfs_info(fs_info,
712 			   "The free space cache file (%llu) is invalid. skip it\n",
713 			   offset);
714 		return 0;
715 	}
716 
717 	if (BTRFS_I(inode)->generation != generation) {
718 		btrfs_err(fs_info,
719 			  "free space inode generation (%llu) did not match free space cache generation (%llu)",
720 			  BTRFS_I(inode)->generation, generation);
721 		return 0;
722 	}
723 
724 	if (!num_entries)
725 		return 0;
726 
727 	ret = io_ctl_init(&io_ctl, inode, 0);
728 	if (ret)
729 		return ret;
730 
731 	readahead_cache(inode);
732 
733 	ret = io_ctl_prepare_pages(&io_ctl, inode, 1);
734 	if (ret)
735 		goto out;
736 
737 	ret = io_ctl_check_crc(&io_ctl, 0);
738 	if (ret)
739 		goto free_cache;
740 
741 	ret = io_ctl_check_generation(&io_ctl, generation);
742 	if (ret)
743 		goto free_cache;
744 
745 	while (num_entries) {
746 		e = kmem_cache_zalloc(btrfs_free_space_cachep,
747 				      GFP_NOFS);
748 		if (!e)
749 			goto free_cache;
750 
751 		ret = io_ctl_read_entry(&io_ctl, e, &type);
752 		if (ret) {
753 			kmem_cache_free(btrfs_free_space_cachep, e);
754 			goto free_cache;
755 		}
756 
757 		if (!e->bytes) {
758 			kmem_cache_free(btrfs_free_space_cachep, e);
759 			goto free_cache;
760 		}
761 
762 		if (type == BTRFS_FREE_SPACE_EXTENT) {
763 			spin_lock(&ctl->tree_lock);
764 			ret = link_free_space(ctl, e);
765 			spin_unlock(&ctl->tree_lock);
766 			if (ret) {
767 				btrfs_err(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 		} else {
773 			ASSERT(num_bitmaps);
774 			num_bitmaps--;
775 			e->bitmap = kzalloc(PAGE_SIZE, GFP_NOFS);
776 			if (!e->bitmap) {
777 				kmem_cache_free(
778 					btrfs_free_space_cachep, e);
779 				goto free_cache;
780 			}
781 			spin_lock(&ctl->tree_lock);
782 			ret = link_free_space(ctl, e);
783 			ctl->total_bitmaps++;
784 			ctl->op->recalc_thresholds(ctl);
785 			spin_unlock(&ctl->tree_lock);
786 			if (ret) {
787 				btrfs_err(fs_info,
788 					"Duplicate entries in free space cache, dumping");
789 				kmem_cache_free(btrfs_free_space_cachep, e);
790 				goto free_cache;
791 			}
792 			list_add_tail(&e->list, &bitmaps);
793 		}
794 
795 		num_entries--;
796 	}
797 
798 	io_ctl_unmap_page(&io_ctl);
799 
800 	/*
801 	 * We add the bitmaps at the end of the entries in order that
802 	 * the bitmap entries are added to the cache.
803 	 */
804 	list_for_each_entry_safe(e, n, &bitmaps, list) {
805 		list_del_init(&e->list);
806 		ret = io_ctl_read_bitmap(&io_ctl, e);
807 		if (ret)
808 			goto free_cache;
809 	}
810 
811 	io_ctl_drop_pages(&io_ctl);
812 	merge_space_tree(ctl);
813 	ret = 1;
814 out:
815 	io_ctl_free(&io_ctl);
816 	return ret;
817 free_cache:
818 	io_ctl_drop_pages(&io_ctl);
819 	__btrfs_remove_free_space_cache(ctl);
820 	goto out;
821 }
822 
823 int load_free_space_cache(struct btrfs_fs_info *fs_info,
824 			  struct btrfs_block_group_cache *block_group)
825 {
826 	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
827 	struct inode *inode;
828 	struct btrfs_path *path;
829 	int ret = 0;
830 	bool matched;
831 	u64 used = btrfs_block_group_used(&block_group->item);
832 
833 	/*
834 	 * If this block group has been marked to be cleared for one reason or
835 	 * another then we can't trust the on disk cache, so just return.
836 	 */
837 	spin_lock(&block_group->lock);
838 	if (block_group->disk_cache_state != BTRFS_DC_WRITTEN) {
839 		spin_unlock(&block_group->lock);
840 		return 0;
841 	}
842 	spin_unlock(&block_group->lock);
843 
844 	path = btrfs_alloc_path();
845 	if (!path)
846 		return 0;
847 	path->search_commit_root = 1;
848 	path->skip_locking = 1;
849 
850 	inode = lookup_free_space_inode(fs_info, block_group, path);
851 	if (IS_ERR(inode)) {
852 		btrfs_free_path(path);
853 		return 0;
854 	}
855 
856 	/* We may have converted the inode and made the cache invalid. */
857 	spin_lock(&block_group->lock);
858 	if (block_group->disk_cache_state != BTRFS_DC_WRITTEN) {
859 		spin_unlock(&block_group->lock);
860 		btrfs_free_path(path);
861 		goto out;
862 	}
863 	spin_unlock(&block_group->lock);
864 
865 	ret = __load_free_space_cache(fs_info->tree_root, inode, ctl,
866 				      path, block_group->key.objectid);
867 	btrfs_free_path(path);
868 	if (ret <= 0)
869 		goto out;
870 
871 	spin_lock(&ctl->tree_lock);
872 	matched = (ctl->free_space == (block_group->key.offset - used -
873 				       block_group->bytes_super));
874 	spin_unlock(&ctl->tree_lock);
875 
876 	if (!matched) {
877 		__btrfs_remove_free_space_cache(ctl);
878 		btrfs_warn(fs_info,
879 			   "block group %llu has wrong amount of free space",
880 			   block_group->key.objectid);
881 		ret = -1;
882 	}
883 out:
884 	if (ret < 0) {
885 		/* This cache is bogus, make sure it gets cleared */
886 		spin_lock(&block_group->lock);
887 		block_group->disk_cache_state = BTRFS_DC_CLEAR;
888 		spin_unlock(&block_group->lock);
889 		ret = 0;
890 
891 		btrfs_warn(fs_info,
892 			   "failed to load free space cache for block group %llu, rebuilding it now",
893 			   block_group->key.objectid);
894 	}
895 
896 	iput(inode);
897 	return ret;
898 }
899 
900 static noinline_for_stack
901 int write_cache_extent_entries(struct btrfs_io_ctl *io_ctl,
902 			      struct btrfs_free_space_ctl *ctl,
903 			      struct btrfs_block_group_cache *block_group,
904 			      int *entries, int *bitmaps,
905 			      struct list_head *bitmap_list)
906 {
907 	int ret;
908 	struct btrfs_free_cluster *cluster = NULL;
909 	struct btrfs_free_cluster *cluster_locked = NULL;
910 	struct rb_node *node = rb_first(&ctl->free_space_offset);
911 	struct btrfs_trim_range *trim_entry;
912 
913 	/* Get the cluster for this block_group if it exists */
914 	if (block_group && !list_empty(&block_group->cluster_list)) {
915 		cluster = list_entry(block_group->cluster_list.next,
916 				     struct btrfs_free_cluster,
917 				     block_group_list);
918 	}
919 
920 	if (!node && cluster) {
921 		cluster_locked = cluster;
922 		spin_lock(&cluster_locked->lock);
923 		node = rb_first(&cluster->root);
924 		cluster = NULL;
925 	}
926 
927 	/* Write out the extent entries */
928 	while (node) {
929 		struct btrfs_free_space *e;
930 
931 		e = rb_entry(node, struct btrfs_free_space, offset_index);
932 		*entries += 1;
933 
934 		ret = io_ctl_add_entry(io_ctl, e->offset, e->bytes,
935 				       e->bitmap);
936 		if (ret)
937 			goto fail;
938 
939 		if (e->bitmap) {
940 			list_add_tail(&e->list, bitmap_list);
941 			*bitmaps += 1;
942 		}
943 		node = rb_next(node);
944 		if (!node && cluster) {
945 			node = rb_first(&cluster->root);
946 			cluster_locked = cluster;
947 			spin_lock(&cluster_locked->lock);
948 			cluster = NULL;
949 		}
950 	}
951 	if (cluster_locked) {
952 		spin_unlock(&cluster_locked->lock);
953 		cluster_locked = NULL;
954 	}
955 
956 	/*
957 	 * Make sure we don't miss any range that was removed from our rbtree
958 	 * because trimming is running. Otherwise after a umount+mount (or crash
959 	 * after committing the transaction) we would leak free space and get
960 	 * an inconsistent free space cache report from fsck.
961 	 */
962 	list_for_each_entry(trim_entry, &ctl->trimming_ranges, list) {
963 		ret = io_ctl_add_entry(io_ctl, trim_entry->start,
964 				       trim_entry->bytes, NULL);
965 		if (ret)
966 			goto fail;
967 		*entries += 1;
968 	}
969 
970 	return 0;
971 fail:
972 	if (cluster_locked)
973 		spin_unlock(&cluster_locked->lock);
974 	return -ENOSPC;
975 }
976 
977 static noinline_for_stack int
978 update_cache_item(struct btrfs_trans_handle *trans,
979 		  struct btrfs_root *root,
980 		  struct inode *inode,
981 		  struct btrfs_path *path, u64 offset,
982 		  int entries, int bitmaps)
983 {
984 	struct btrfs_key key;
985 	struct btrfs_free_space_header *header;
986 	struct extent_buffer *leaf;
987 	int ret;
988 
989 	key.objectid = BTRFS_FREE_SPACE_OBJECTID;
990 	key.offset = offset;
991 	key.type = 0;
992 
993 	ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
994 	if (ret < 0) {
995 		clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, inode->i_size - 1,
996 				 EXTENT_DIRTY | EXTENT_DELALLOC, 0, 0, NULL,
997 				 GFP_NOFS);
998 		goto fail;
999 	}
1000 	leaf = path->nodes[0];
1001 	if (ret > 0) {
1002 		struct btrfs_key found_key;
1003 		ASSERT(path->slots[0]);
1004 		path->slots[0]--;
1005 		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
1006 		if (found_key.objectid != BTRFS_FREE_SPACE_OBJECTID ||
1007 		    found_key.offset != offset) {
1008 			clear_extent_bit(&BTRFS_I(inode)->io_tree, 0,
1009 					 inode->i_size - 1,
1010 					 EXTENT_DIRTY | EXTENT_DELALLOC, 0, 0,
1011 					 NULL, GFP_NOFS);
1012 			btrfs_release_path(path);
1013 			goto fail;
1014 		}
1015 	}
1016 
1017 	BTRFS_I(inode)->generation = trans->transid;
1018 	header = btrfs_item_ptr(leaf, path->slots[0],
1019 				struct btrfs_free_space_header);
1020 	btrfs_set_free_space_entries(leaf, header, entries);
1021 	btrfs_set_free_space_bitmaps(leaf, header, bitmaps);
1022 	btrfs_set_free_space_generation(leaf, header, trans->transid);
1023 	btrfs_mark_buffer_dirty(leaf);
1024 	btrfs_release_path(path);
1025 
1026 	return 0;
1027 
1028 fail:
1029 	return -1;
1030 }
1031 
1032 static noinline_for_stack int
1033 write_pinned_extent_entries(struct btrfs_fs_info *fs_info,
1034 			    struct btrfs_block_group_cache *block_group,
1035 			    struct btrfs_io_ctl *io_ctl,
1036 			    int *entries)
1037 {
1038 	u64 start, extent_start, extent_end, len;
1039 	struct extent_io_tree *unpin = NULL;
1040 	int ret;
1041 
1042 	if (!block_group)
1043 		return 0;
1044 
1045 	/*
1046 	 * We want to add any pinned extents to our free space cache
1047 	 * so we don't leak the space
1048 	 *
1049 	 * We shouldn't have switched the pinned extents yet so this is the
1050 	 * right one
1051 	 */
1052 	unpin = fs_info->pinned_extents;
1053 
1054 	start = block_group->key.objectid;
1055 
1056 	while (start < block_group->key.objectid + block_group->key.offset) {
1057 		ret = find_first_extent_bit(unpin, start,
1058 					    &extent_start, &extent_end,
1059 					    EXTENT_DIRTY, NULL);
1060 		if (ret)
1061 			return 0;
1062 
1063 		/* This pinned extent is out of our range */
1064 		if (extent_start >= block_group->key.objectid +
1065 		    block_group->key.offset)
1066 			return 0;
1067 
1068 		extent_start = max(extent_start, start);
1069 		extent_end = min(block_group->key.objectid +
1070 				 block_group->key.offset, extent_end + 1);
1071 		len = extent_end - extent_start;
1072 
1073 		*entries += 1;
1074 		ret = io_ctl_add_entry(io_ctl, extent_start, len, NULL);
1075 		if (ret)
1076 			return -ENOSPC;
1077 
1078 		start = extent_end;
1079 	}
1080 
1081 	return 0;
1082 }
1083 
1084 static noinline_for_stack int
1085 write_bitmap_entries(struct btrfs_io_ctl *io_ctl, struct list_head *bitmap_list)
1086 {
1087 	struct btrfs_free_space *entry, *next;
1088 	int ret;
1089 
1090 	/* Write out the bitmaps */
1091 	list_for_each_entry_safe(entry, next, bitmap_list, list) {
1092 		ret = io_ctl_add_bitmap(io_ctl, entry->bitmap);
1093 		if (ret)
1094 			return -ENOSPC;
1095 		list_del_init(&entry->list);
1096 	}
1097 
1098 	return 0;
1099 }
1100 
1101 static int flush_dirty_cache(struct inode *inode)
1102 {
1103 	int ret;
1104 
1105 	ret = btrfs_wait_ordered_range(inode, 0, (u64)-1);
1106 	if (ret)
1107 		clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, inode->i_size - 1,
1108 				 EXTENT_DIRTY | EXTENT_DELALLOC, 0, 0, NULL,
1109 				 GFP_NOFS);
1110 
1111 	return ret;
1112 }
1113 
1114 static void noinline_for_stack
1115 cleanup_bitmap_list(struct list_head *bitmap_list)
1116 {
1117 	struct btrfs_free_space *entry, *next;
1118 
1119 	list_for_each_entry_safe(entry, next, bitmap_list, list)
1120 		list_del_init(&entry->list);
1121 }
1122 
1123 static void noinline_for_stack
1124 cleanup_write_cache_enospc(struct inode *inode,
1125 			   struct btrfs_io_ctl *io_ctl,
1126 			   struct extent_state **cached_state)
1127 {
1128 	io_ctl_drop_pages(io_ctl);
1129 	unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0,
1130 			     i_size_read(inode) - 1, cached_state,
1131 			     GFP_NOFS);
1132 }
1133 
1134 static int __btrfs_wait_cache_io(struct btrfs_root *root,
1135 				 struct btrfs_trans_handle *trans,
1136 				 struct btrfs_block_group_cache *block_group,
1137 				 struct btrfs_io_ctl *io_ctl,
1138 				 struct btrfs_path *path, u64 offset)
1139 {
1140 	int ret;
1141 	struct inode *inode = io_ctl->inode;
1142 	struct btrfs_fs_info *fs_info;
1143 
1144 	if (!inode)
1145 		return 0;
1146 
1147 	fs_info = btrfs_sb(inode->i_sb);
1148 
1149 	/* Flush the dirty pages in the cache file. */
1150 	ret = flush_dirty_cache(inode);
1151 	if (ret)
1152 		goto out;
1153 
1154 	/* Update the cache item to tell everyone this cache file is valid. */
1155 	ret = update_cache_item(trans, root, inode, path, offset,
1156 				io_ctl->entries, io_ctl->bitmaps);
1157 out:
1158 	io_ctl_free(io_ctl);
1159 	if (ret) {
1160 		invalidate_inode_pages2(inode->i_mapping);
1161 		BTRFS_I(inode)->generation = 0;
1162 		if (block_group) {
1163 #ifdef DEBUG
1164 			btrfs_err(fs_info,
1165 				  "failed to write free space cache for block group %llu",
1166 				  block_group->key.objectid);
1167 #endif
1168 		}
1169 	}
1170 	btrfs_update_inode(trans, root, inode);
1171 
1172 	if (block_group) {
1173 		/* the dirty list is protected by the dirty_bgs_lock */
1174 		spin_lock(&trans->transaction->dirty_bgs_lock);
1175 
1176 		/* the disk_cache_state is protected by the block group lock */
1177 		spin_lock(&block_group->lock);
1178 
1179 		/*
1180 		 * only mark this as written if we didn't get put back on
1181 		 * the dirty list while waiting for IO.   Otherwise our
1182 		 * cache state won't be right, and we won't get written again
1183 		 */
1184 		if (!ret && list_empty(&block_group->dirty_list))
1185 			block_group->disk_cache_state = BTRFS_DC_WRITTEN;
1186 		else if (ret)
1187 			block_group->disk_cache_state = BTRFS_DC_ERROR;
1188 
1189 		spin_unlock(&block_group->lock);
1190 		spin_unlock(&trans->transaction->dirty_bgs_lock);
1191 		io_ctl->inode = NULL;
1192 		iput(inode);
1193 	}
1194 
1195 	return ret;
1196 
1197 }
1198 
1199 static int btrfs_wait_cache_io_root(struct btrfs_root *root,
1200 				    struct btrfs_trans_handle *trans,
1201 				    struct btrfs_io_ctl *io_ctl,
1202 				    struct btrfs_path *path)
1203 {
1204 	return __btrfs_wait_cache_io(root, trans, NULL, io_ctl, path, 0);
1205 }
1206 
1207 int btrfs_wait_cache_io(struct btrfs_trans_handle *trans,
1208 			struct btrfs_block_group_cache *block_group,
1209 			struct btrfs_path *path)
1210 {
1211 	return __btrfs_wait_cache_io(block_group->fs_info->tree_root, trans,
1212 				     block_group, &block_group->io_ctl,
1213 				     path, block_group->key.objectid);
1214 }
1215 
1216 /**
1217  * __btrfs_write_out_cache - write out cached info to an inode
1218  * @root - the root the inode belongs to
1219  * @ctl - the free space cache we are going to write out
1220  * @block_group - the block_group for this cache if it belongs to a block_group
1221  * @trans - the trans handle
1222  *
1223  * This function writes out a free space cache struct to disk for quick recovery
1224  * on mount.  This will return 0 if it was successful in writing the cache out,
1225  * or an errno if it was not.
1226  */
1227 static int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
1228 				   struct btrfs_free_space_ctl *ctl,
1229 				   struct btrfs_block_group_cache *block_group,
1230 				   struct btrfs_io_ctl *io_ctl,
1231 				   struct btrfs_trans_handle *trans)
1232 {
1233 	struct btrfs_fs_info *fs_info = root->fs_info;
1234 	struct extent_state *cached_state = NULL;
1235 	LIST_HEAD(bitmap_list);
1236 	int entries = 0;
1237 	int bitmaps = 0;
1238 	int ret;
1239 	int must_iput = 0;
1240 
1241 	if (!i_size_read(inode))
1242 		return -EIO;
1243 
1244 	WARN_ON(io_ctl->pages);
1245 	ret = io_ctl_init(io_ctl, inode, 1);
1246 	if (ret)
1247 		return ret;
1248 
1249 	if (block_group && (block_group->flags & BTRFS_BLOCK_GROUP_DATA)) {
1250 		down_write(&block_group->data_rwsem);
1251 		spin_lock(&block_group->lock);
1252 		if (block_group->delalloc_bytes) {
1253 			block_group->disk_cache_state = BTRFS_DC_WRITTEN;
1254 			spin_unlock(&block_group->lock);
1255 			up_write(&block_group->data_rwsem);
1256 			BTRFS_I(inode)->generation = 0;
1257 			ret = 0;
1258 			must_iput = 1;
1259 			goto out;
1260 		}
1261 		spin_unlock(&block_group->lock);
1262 	}
1263 
1264 	/* Lock all pages first so we can lock the extent safely. */
1265 	ret = io_ctl_prepare_pages(io_ctl, inode, 0);
1266 	if (ret)
1267 		goto out;
1268 
1269 	lock_extent_bits(&BTRFS_I(inode)->io_tree, 0, i_size_read(inode) - 1,
1270 			 &cached_state);
1271 
1272 	io_ctl_set_generation(io_ctl, trans->transid);
1273 
1274 	mutex_lock(&ctl->cache_writeout_mutex);
1275 	/* Write out the extent entries in the free space cache */
1276 	spin_lock(&ctl->tree_lock);
1277 	ret = write_cache_extent_entries(io_ctl, ctl,
1278 					 block_group, &entries, &bitmaps,
1279 					 &bitmap_list);
1280 	if (ret)
1281 		goto out_nospc_locked;
1282 
1283 	/*
1284 	 * Some spaces that are freed in the current transaction are pinned,
1285 	 * they will be added into free space cache after the transaction is
1286 	 * committed, we shouldn't lose them.
1287 	 *
1288 	 * If this changes while we are working we'll get added back to
1289 	 * the dirty list and redo it.  No locking needed
1290 	 */
1291 	ret = write_pinned_extent_entries(fs_info, block_group,
1292 					  io_ctl, &entries);
1293 	if (ret)
1294 		goto out_nospc_locked;
1295 
1296 	/*
1297 	 * At last, we write out all the bitmaps and keep cache_writeout_mutex
1298 	 * locked while doing it because a concurrent trim can be manipulating
1299 	 * or freeing the bitmap.
1300 	 */
1301 	ret = write_bitmap_entries(io_ctl, &bitmap_list);
1302 	spin_unlock(&ctl->tree_lock);
1303 	mutex_unlock(&ctl->cache_writeout_mutex);
1304 	if (ret)
1305 		goto out_nospc;
1306 
1307 	/* Zero out the rest of the pages just to make sure */
1308 	io_ctl_zero_remaining_pages(io_ctl);
1309 
1310 	/* Everything is written out, now we dirty the pages in the file. */
1311 	ret = btrfs_dirty_pages(inode, io_ctl->pages, io_ctl->num_pages, 0,
1312 				i_size_read(inode), &cached_state);
1313 	if (ret)
1314 		goto out_nospc;
1315 
1316 	if (block_group && (block_group->flags & BTRFS_BLOCK_GROUP_DATA))
1317 		up_write(&block_group->data_rwsem);
1318 	/*
1319 	 * Release the pages and unlock the extent, we will flush
1320 	 * them out later
1321 	 */
1322 	io_ctl_drop_pages(io_ctl);
1323 
1324 	unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0,
1325 			     i_size_read(inode) - 1, &cached_state, GFP_NOFS);
1326 
1327 	/*
1328 	 * at this point the pages are under IO and we're happy,
1329 	 * The caller is responsible for waiting on them and updating the
1330 	 * the cache and the inode
1331 	 */
1332 	io_ctl->entries = entries;
1333 	io_ctl->bitmaps = bitmaps;
1334 
1335 	ret = btrfs_fdatawrite_range(inode, 0, (u64)-1);
1336 	if (ret)
1337 		goto out;
1338 
1339 	return 0;
1340 
1341 out:
1342 	io_ctl->inode = NULL;
1343 	io_ctl_free(io_ctl);
1344 	if (ret) {
1345 		invalidate_inode_pages2(inode->i_mapping);
1346 		BTRFS_I(inode)->generation = 0;
1347 	}
1348 	btrfs_update_inode(trans, root, inode);
1349 	if (must_iput)
1350 		iput(inode);
1351 	return ret;
1352 
1353 out_nospc_locked:
1354 	cleanup_bitmap_list(&bitmap_list);
1355 	spin_unlock(&ctl->tree_lock);
1356 	mutex_unlock(&ctl->cache_writeout_mutex);
1357 
1358 out_nospc:
1359 	cleanup_write_cache_enospc(inode, io_ctl, &cached_state);
1360 
1361 	if (block_group && (block_group->flags & BTRFS_BLOCK_GROUP_DATA))
1362 		up_write(&block_group->data_rwsem);
1363 
1364 	goto out;
1365 }
1366 
1367 int btrfs_write_out_cache(struct btrfs_fs_info *fs_info,
1368 			  struct btrfs_trans_handle *trans,
1369 			  struct btrfs_block_group_cache *block_group,
1370 			  struct btrfs_path *path)
1371 {
1372 	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
1373 	struct inode *inode;
1374 	int ret = 0;
1375 
1376 	spin_lock(&block_group->lock);
1377 	if (block_group->disk_cache_state < BTRFS_DC_SETUP) {
1378 		spin_unlock(&block_group->lock);
1379 		return 0;
1380 	}
1381 	spin_unlock(&block_group->lock);
1382 
1383 	inode = lookup_free_space_inode(fs_info, block_group, path);
1384 	if (IS_ERR(inode))
1385 		return 0;
1386 
1387 	ret = __btrfs_write_out_cache(fs_info->tree_root, inode, ctl,
1388 				block_group, &block_group->io_ctl, trans);
1389 	if (ret) {
1390 #ifdef DEBUG
1391 		btrfs_err(fs_info,
1392 			  "failed to write free space cache for block group %llu",
1393 			  block_group->key.objectid);
1394 #endif
1395 		spin_lock(&block_group->lock);
1396 		block_group->disk_cache_state = BTRFS_DC_ERROR;
1397 		spin_unlock(&block_group->lock);
1398 
1399 		block_group->io_ctl.inode = NULL;
1400 		iput(inode);
1401 	}
1402 
1403 	/*
1404 	 * if ret == 0 the caller is expected to call btrfs_wait_cache_io
1405 	 * to wait for IO and put the inode
1406 	 */
1407 
1408 	return ret;
1409 }
1410 
1411 static inline unsigned long offset_to_bit(u64 bitmap_start, u32 unit,
1412 					  u64 offset)
1413 {
1414 	ASSERT(offset >= bitmap_start);
1415 	offset -= bitmap_start;
1416 	return (unsigned long)(div_u64(offset, unit));
1417 }
1418 
1419 static inline unsigned long bytes_to_bits(u64 bytes, u32 unit)
1420 {
1421 	return (unsigned long)(div_u64(bytes, unit));
1422 }
1423 
1424 static inline u64 offset_to_bitmap(struct btrfs_free_space_ctl *ctl,
1425 				   u64 offset)
1426 {
1427 	u64 bitmap_start;
1428 	u64 bytes_per_bitmap;
1429 
1430 	bytes_per_bitmap = BITS_PER_BITMAP * ctl->unit;
1431 	bitmap_start = offset - ctl->start;
1432 	bitmap_start = div64_u64(bitmap_start, bytes_per_bitmap);
1433 	bitmap_start *= bytes_per_bitmap;
1434 	bitmap_start += ctl->start;
1435 
1436 	return bitmap_start;
1437 }
1438 
1439 static int tree_insert_offset(struct rb_root *root, u64 offset,
1440 			      struct rb_node *node, int bitmap)
1441 {
1442 	struct rb_node **p = &root->rb_node;
1443 	struct rb_node *parent = NULL;
1444 	struct btrfs_free_space *info;
1445 
1446 	while (*p) {
1447 		parent = *p;
1448 		info = rb_entry(parent, struct btrfs_free_space, offset_index);
1449 
1450 		if (offset < info->offset) {
1451 			p = &(*p)->rb_left;
1452 		} else if (offset > info->offset) {
1453 			p = &(*p)->rb_right;
1454 		} else {
1455 			/*
1456 			 * we could have a bitmap entry and an extent entry
1457 			 * share the same offset.  If this is the case, we want
1458 			 * the extent entry to always be found first if we do a
1459 			 * linear search through the tree, since we want to have
1460 			 * the quickest allocation time, and allocating from an
1461 			 * extent is faster than allocating from a bitmap.  So
1462 			 * if we're inserting a bitmap and we find an entry at
1463 			 * this offset, we want to go right, or after this entry
1464 			 * logically.  If we are inserting an extent and we've
1465 			 * found a bitmap, we want to go left, or before
1466 			 * logically.
1467 			 */
1468 			if (bitmap) {
1469 				if (info->bitmap) {
1470 					WARN_ON_ONCE(1);
1471 					return -EEXIST;
1472 				}
1473 				p = &(*p)->rb_right;
1474 			} else {
1475 				if (!info->bitmap) {
1476 					WARN_ON_ONCE(1);
1477 					return -EEXIST;
1478 				}
1479 				p = &(*p)->rb_left;
1480 			}
1481 		}
1482 	}
1483 
1484 	rb_link_node(node, parent, p);
1485 	rb_insert_color(node, root);
1486 
1487 	return 0;
1488 }
1489 
1490 /*
1491  * searches the tree for the given offset.
1492  *
1493  * fuzzy - If this is set, then we are trying to make an allocation, and we just
1494  * want a section that has at least bytes size and comes at or after the given
1495  * offset.
1496  */
1497 static struct btrfs_free_space *
1498 tree_search_offset(struct btrfs_free_space_ctl *ctl,
1499 		   u64 offset, int bitmap_only, int fuzzy)
1500 {
1501 	struct rb_node *n = ctl->free_space_offset.rb_node;
1502 	struct btrfs_free_space *entry, *prev = NULL;
1503 
1504 	/* find entry that is closest to the 'offset' */
1505 	while (1) {
1506 		if (!n) {
1507 			entry = NULL;
1508 			break;
1509 		}
1510 
1511 		entry = rb_entry(n, struct btrfs_free_space, offset_index);
1512 		prev = entry;
1513 
1514 		if (offset < entry->offset)
1515 			n = n->rb_left;
1516 		else if (offset > entry->offset)
1517 			n = n->rb_right;
1518 		else
1519 			break;
1520 	}
1521 
1522 	if (bitmap_only) {
1523 		if (!entry)
1524 			return NULL;
1525 		if (entry->bitmap)
1526 			return entry;
1527 
1528 		/*
1529 		 * bitmap entry and extent entry may share same offset,
1530 		 * in that case, bitmap entry comes after extent entry.
1531 		 */
1532 		n = rb_next(n);
1533 		if (!n)
1534 			return NULL;
1535 		entry = rb_entry(n, struct btrfs_free_space, offset_index);
1536 		if (entry->offset != offset)
1537 			return NULL;
1538 
1539 		WARN_ON(!entry->bitmap);
1540 		return entry;
1541 	} else if (entry) {
1542 		if (entry->bitmap) {
1543 			/*
1544 			 * if previous extent entry covers the offset,
1545 			 * we should return it instead of the bitmap entry
1546 			 */
1547 			n = rb_prev(&entry->offset_index);
1548 			if (n) {
1549 				prev = rb_entry(n, struct btrfs_free_space,
1550 						offset_index);
1551 				if (!prev->bitmap &&
1552 				    prev->offset + prev->bytes > offset)
1553 					entry = prev;
1554 			}
1555 		}
1556 		return entry;
1557 	}
1558 
1559 	if (!prev)
1560 		return NULL;
1561 
1562 	/* find last entry before the 'offset' */
1563 	entry = prev;
1564 	if (entry->offset > offset) {
1565 		n = rb_prev(&entry->offset_index);
1566 		if (n) {
1567 			entry = rb_entry(n, struct btrfs_free_space,
1568 					offset_index);
1569 			ASSERT(entry->offset <= offset);
1570 		} else {
1571 			if (fuzzy)
1572 				return entry;
1573 			else
1574 				return NULL;
1575 		}
1576 	}
1577 
1578 	if (entry->bitmap) {
1579 		n = rb_prev(&entry->offset_index);
1580 		if (n) {
1581 			prev = rb_entry(n, struct btrfs_free_space,
1582 					offset_index);
1583 			if (!prev->bitmap &&
1584 			    prev->offset + prev->bytes > offset)
1585 				return prev;
1586 		}
1587 		if (entry->offset + BITS_PER_BITMAP * ctl->unit > offset)
1588 			return entry;
1589 	} else if (entry->offset + entry->bytes > offset)
1590 		return entry;
1591 
1592 	if (!fuzzy)
1593 		return NULL;
1594 
1595 	while (1) {
1596 		if (entry->bitmap) {
1597 			if (entry->offset + BITS_PER_BITMAP *
1598 			    ctl->unit > offset)
1599 				break;
1600 		} else {
1601 			if (entry->offset + entry->bytes > offset)
1602 				break;
1603 		}
1604 
1605 		n = rb_next(&entry->offset_index);
1606 		if (!n)
1607 			return NULL;
1608 		entry = rb_entry(n, struct btrfs_free_space, offset_index);
1609 	}
1610 	return entry;
1611 }
1612 
1613 static inline void
1614 __unlink_free_space(struct btrfs_free_space_ctl *ctl,
1615 		    struct btrfs_free_space *info)
1616 {
1617 	rb_erase(&info->offset_index, &ctl->free_space_offset);
1618 	ctl->free_extents--;
1619 }
1620 
1621 static void unlink_free_space(struct btrfs_free_space_ctl *ctl,
1622 			      struct btrfs_free_space *info)
1623 {
1624 	__unlink_free_space(ctl, info);
1625 	ctl->free_space -= info->bytes;
1626 }
1627 
1628 static int link_free_space(struct btrfs_free_space_ctl *ctl,
1629 			   struct btrfs_free_space *info)
1630 {
1631 	int ret = 0;
1632 
1633 	ASSERT(info->bytes || info->bitmap);
1634 	ret = tree_insert_offset(&ctl->free_space_offset, info->offset,
1635 				 &info->offset_index, (info->bitmap != NULL));
1636 	if (ret)
1637 		return ret;
1638 
1639 	ctl->free_space += info->bytes;
1640 	ctl->free_extents++;
1641 	return ret;
1642 }
1643 
1644 static void recalculate_thresholds(struct btrfs_free_space_ctl *ctl)
1645 {
1646 	struct btrfs_block_group_cache *block_group = ctl->private;
1647 	u64 max_bytes;
1648 	u64 bitmap_bytes;
1649 	u64 extent_bytes;
1650 	u64 size = block_group->key.offset;
1651 	u64 bytes_per_bg = BITS_PER_BITMAP * ctl->unit;
1652 	u64 max_bitmaps = div64_u64(size + bytes_per_bg - 1, bytes_per_bg);
1653 
1654 	max_bitmaps = max_t(u64, max_bitmaps, 1);
1655 
1656 	ASSERT(ctl->total_bitmaps <= max_bitmaps);
1657 
1658 	/*
1659 	 * The goal is to keep the total amount of memory used per 1gb of space
1660 	 * at or below 32k, so we need to adjust how much memory we allow to be
1661 	 * used by extent based free space tracking
1662 	 */
1663 	if (size < SZ_1G)
1664 		max_bytes = MAX_CACHE_BYTES_PER_GIG;
1665 	else
1666 		max_bytes = MAX_CACHE_BYTES_PER_GIG * div_u64(size, SZ_1G);
1667 
1668 	/*
1669 	 * we want to account for 1 more bitmap than what we have so we can make
1670 	 * sure we don't go over our overall goal of MAX_CACHE_BYTES_PER_GIG as
1671 	 * we add more bitmaps.
1672 	 */
1673 	bitmap_bytes = (ctl->total_bitmaps + 1) * ctl->unit;
1674 
1675 	if (bitmap_bytes >= max_bytes) {
1676 		ctl->extents_thresh = 0;
1677 		return;
1678 	}
1679 
1680 	/*
1681 	 * we want the extent entry threshold to always be at most 1/2 the max
1682 	 * bytes we can have, or whatever is less than that.
1683 	 */
1684 	extent_bytes = max_bytes - bitmap_bytes;
1685 	extent_bytes = min_t(u64, extent_bytes, max_bytes >> 1);
1686 
1687 	ctl->extents_thresh =
1688 		div_u64(extent_bytes, sizeof(struct btrfs_free_space));
1689 }
1690 
1691 static inline void __bitmap_clear_bits(struct btrfs_free_space_ctl *ctl,
1692 				       struct btrfs_free_space *info,
1693 				       u64 offset, u64 bytes)
1694 {
1695 	unsigned long start, count;
1696 
1697 	start = offset_to_bit(info->offset, ctl->unit, offset);
1698 	count = bytes_to_bits(bytes, ctl->unit);
1699 	ASSERT(start + count <= BITS_PER_BITMAP);
1700 
1701 	bitmap_clear(info->bitmap, start, count);
1702 
1703 	info->bytes -= bytes;
1704 }
1705 
1706 static void bitmap_clear_bits(struct btrfs_free_space_ctl *ctl,
1707 			      struct btrfs_free_space *info, u64 offset,
1708 			      u64 bytes)
1709 {
1710 	__bitmap_clear_bits(ctl, info, offset, bytes);
1711 	ctl->free_space -= bytes;
1712 }
1713 
1714 static void bitmap_set_bits(struct btrfs_free_space_ctl *ctl,
1715 			    struct btrfs_free_space *info, u64 offset,
1716 			    u64 bytes)
1717 {
1718 	unsigned long start, count;
1719 
1720 	start = offset_to_bit(info->offset, ctl->unit, offset);
1721 	count = bytes_to_bits(bytes, ctl->unit);
1722 	ASSERT(start + count <= BITS_PER_BITMAP);
1723 
1724 	bitmap_set(info->bitmap, start, count);
1725 
1726 	info->bytes += bytes;
1727 	ctl->free_space += bytes;
1728 }
1729 
1730 /*
1731  * If we can not find suitable extent, we will use bytes to record
1732  * the size of the max extent.
1733  */
1734 static int search_bitmap(struct btrfs_free_space_ctl *ctl,
1735 			 struct btrfs_free_space *bitmap_info, u64 *offset,
1736 			 u64 *bytes, bool for_alloc)
1737 {
1738 	unsigned long found_bits = 0;
1739 	unsigned long max_bits = 0;
1740 	unsigned long bits, i;
1741 	unsigned long next_zero;
1742 	unsigned long extent_bits;
1743 
1744 	/*
1745 	 * Skip searching the bitmap if we don't have a contiguous section that
1746 	 * is large enough for this allocation.
1747 	 */
1748 	if (for_alloc &&
1749 	    bitmap_info->max_extent_size &&
1750 	    bitmap_info->max_extent_size < *bytes) {
1751 		*bytes = bitmap_info->max_extent_size;
1752 		return -1;
1753 	}
1754 
1755 	i = offset_to_bit(bitmap_info->offset, ctl->unit,
1756 			  max_t(u64, *offset, bitmap_info->offset));
1757 	bits = bytes_to_bits(*bytes, ctl->unit);
1758 
1759 	for_each_set_bit_from(i, bitmap_info->bitmap, BITS_PER_BITMAP) {
1760 		if (for_alloc && bits == 1) {
1761 			found_bits = 1;
1762 			break;
1763 		}
1764 		next_zero = find_next_zero_bit(bitmap_info->bitmap,
1765 					       BITS_PER_BITMAP, i);
1766 		extent_bits = next_zero - i;
1767 		if (extent_bits >= bits) {
1768 			found_bits = extent_bits;
1769 			break;
1770 		} else if (extent_bits > max_bits) {
1771 			max_bits = extent_bits;
1772 		}
1773 		i = next_zero;
1774 	}
1775 
1776 	if (found_bits) {
1777 		*offset = (u64)(i * ctl->unit) + bitmap_info->offset;
1778 		*bytes = (u64)(found_bits) * ctl->unit;
1779 		return 0;
1780 	}
1781 
1782 	*bytes = (u64)(max_bits) * ctl->unit;
1783 	bitmap_info->max_extent_size = *bytes;
1784 	return -1;
1785 }
1786 
1787 /* Cache the size of the max extent in bytes */
1788 static struct btrfs_free_space *
1789 find_free_space(struct btrfs_free_space_ctl *ctl, u64 *offset, u64 *bytes,
1790 		unsigned long align, u64 *max_extent_size)
1791 {
1792 	struct btrfs_free_space *entry;
1793 	struct rb_node *node;
1794 	u64 tmp;
1795 	u64 align_off;
1796 	int ret;
1797 
1798 	if (!ctl->free_space_offset.rb_node)
1799 		goto out;
1800 
1801 	entry = tree_search_offset(ctl, offset_to_bitmap(ctl, *offset), 0, 1);
1802 	if (!entry)
1803 		goto out;
1804 
1805 	for (node = &entry->offset_index; node; node = rb_next(node)) {
1806 		entry = rb_entry(node, struct btrfs_free_space, offset_index);
1807 		if (entry->bytes < *bytes) {
1808 			if (entry->bytes > *max_extent_size)
1809 				*max_extent_size = entry->bytes;
1810 			continue;
1811 		}
1812 
1813 		/* make sure the space returned is big enough
1814 		 * to match our requested alignment
1815 		 */
1816 		if (*bytes >= align) {
1817 			tmp = entry->offset - ctl->start + align - 1;
1818 			tmp = div64_u64(tmp, align);
1819 			tmp = tmp * align + ctl->start;
1820 			align_off = tmp - entry->offset;
1821 		} else {
1822 			align_off = 0;
1823 			tmp = entry->offset;
1824 		}
1825 
1826 		if (entry->bytes < *bytes + align_off) {
1827 			if (entry->bytes > *max_extent_size)
1828 				*max_extent_size = entry->bytes;
1829 			continue;
1830 		}
1831 
1832 		if (entry->bitmap) {
1833 			u64 size = *bytes;
1834 
1835 			ret = search_bitmap(ctl, entry, &tmp, &size, true);
1836 			if (!ret) {
1837 				*offset = tmp;
1838 				*bytes = size;
1839 				return entry;
1840 			} else if (size > *max_extent_size) {
1841 				*max_extent_size = size;
1842 			}
1843 			continue;
1844 		}
1845 
1846 		*offset = tmp;
1847 		*bytes = entry->bytes - align_off;
1848 		return entry;
1849 	}
1850 out:
1851 	return NULL;
1852 }
1853 
1854 static void add_new_bitmap(struct btrfs_free_space_ctl *ctl,
1855 			   struct btrfs_free_space *info, u64 offset)
1856 {
1857 	info->offset = offset_to_bitmap(ctl, offset);
1858 	info->bytes = 0;
1859 	INIT_LIST_HEAD(&info->list);
1860 	link_free_space(ctl, info);
1861 	ctl->total_bitmaps++;
1862 
1863 	ctl->op->recalc_thresholds(ctl);
1864 }
1865 
1866 static void free_bitmap(struct btrfs_free_space_ctl *ctl,
1867 			struct btrfs_free_space *bitmap_info)
1868 {
1869 	unlink_free_space(ctl, bitmap_info);
1870 	kfree(bitmap_info->bitmap);
1871 	kmem_cache_free(btrfs_free_space_cachep, bitmap_info);
1872 	ctl->total_bitmaps--;
1873 	ctl->op->recalc_thresholds(ctl);
1874 }
1875 
1876 static noinline int remove_from_bitmap(struct btrfs_free_space_ctl *ctl,
1877 			      struct btrfs_free_space *bitmap_info,
1878 			      u64 *offset, u64 *bytes)
1879 {
1880 	u64 end;
1881 	u64 search_start, search_bytes;
1882 	int ret;
1883 
1884 again:
1885 	end = bitmap_info->offset + (u64)(BITS_PER_BITMAP * ctl->unit) - 1;
1886 
1887 	/*
1888 	 * We need to search for bits in this bitmap.  We could only cover some
1889 	 * of the extent in this bitmap thanks to how we add space, so we need
1890 	 * to search for as much as it as we can and clear that amount, and then
1891 	 * go searching for the next bit.
1892 	 */
1893 	search_start = *offset;
1894 	search_bytes = ctl->unit;
1895 	search_bytes = min(search_bytes, end - search_start + 1);
1896 	ret = search_bitmap(ctl, bitmap_info, &search_start, &search_bytes,
1897 			    false);
1898 	if (ret < 0 || search_start != *offset)
1899 		return -EINVAL;
1900 
1901 	/* We may have found more bits than what we need */
1902 	search_bytes = min(search_bytes, *bytes);
1903 
1904 	/* Cannot clear past the end of the bitmap */
1905 	search_bytes = min(search_bytes, end - search_start + 1);
1906 
1907 	bitmap_clear_bits(ctl, bitmap_info, search_start, search_bytes);
1908 	*offset += search_bytes;
1909 	*bytes -= search_bytes;
1910 
1911 	if (*bytes) {
1912 		struct rb_node *next = rb_next(&bitmap_info->offset_index);
1913 		if (!bitmap_info->bytes)
1914 			free_bitmap(ctl, bitmap_info);
1915 
1916 		/*
1917 		 * no entry after this bitmap, but we still have bytes to
1918 		 * remove, so something has gone wrong.
1919 		 */
1920 		if (!next)
1921 			return -EINVAL;
1922 
1923 		bitmap_info = rb_entry(next, struct btrfs_free_space,
1924 				       offset_index);
1925 
1926 		/*
1927 		 * if the next entry isn't a bitmap we need to return to let the
1928 		 * extent stuff do its work.
1929 		 */
1930 		if (!bitmap_info->bitmap)
1931 			return -EAGAIN;
1932 
1933 		/*
1934 		 * Ok the next item is a bitmap, but it may not actually hold
1935 		 * the information for the rest of this free space stuff, so
1936 		 * look for it, and if we don't find it return so we can try
1937 		 * everything over again.
1938 		 */
1939 		search_start = *offset;
1940 		search_bytes = ctl->unit;
1941 		ret = search_bitmap(ctl, bitmap_info, &search_start,
1942 				    &search_bytes, false);
1943 		if (ret < 0 || search_start != *offset)
1944 			return -EAGAIN;
1945 
1946 		goto again;
1947 	} else if (!bitmap_info->bytes)
1948 		free_bitmap(ctl, bitmap_info);
1949 
1950 	return 0;
1951 }
1952 
1953 static u64 add_bytes_to_bitmap(struct btrfs_free_space_ctl *ctl,
1954 			       struct btrfs_free_space *info, u64 offset,
1955 			       u64 bytes)
1956 {
1957 	u64 bytes_to_set = 0;
1958 	u64 end;
1959 
1960 	end = info->offset + (u64)(BITS_PER_BITMAP * ctl->unit);
1961 
1962 	bytes_to_set = min(end - offset, bytes);
1963 
1964 	bitmap_set_bits(ctl, info, offset, bytes_to_set);
1965 
1966 	/*
1967 	 * We set some bytes, we have no idea what the max extent size is
1968 	 * anymore.
1969 	 */
1970 	info->max_extent_size = 0;
1971 
1972 	return bytes_to_set;
1973 
1974 }
1975 
1976 static bool use_bitmap(struct btrfs_free_space_ctl *ctl,
1977 		      struct btrfs_free_space *info)
1978 {
1979 	struct btrfs_block_group_cache *block_group = ctl->private;
1980 	struct btrfs_fs_info *fs_info = block_group->fs_info;
1981 	bool forced = false;
1982 
1983 #ifdef CONFIG_BTRFS_DEBUG
1984 	if (btrfs_should_fragment_free_space(block_group))
1985 		forced = true;
1986 #endif
1987 
1988 	/*
1989 	 * If we are below the extents threshold then we can add this as an
1990 	 * extent, and don't have to deal with the bitmap
1991 	 */
1992 	if (!forced && ctl->free_extents < ctl->extents_thresh) {
1993 		/*
1994 		 * If this block group has some small extents we don't want to
1995 		 * use up all of our free slots in the cache with them, we want
1996 		 * to reserve them to larger extents, however if we have plenty
1997 		 * of cache left then go ahead an dadd them, no sense in adding
1998 		 * the overhead of a bitmap if we don't have to.
1999 		 */
2000 		if (info->bytes <= fs_info->sectorsize * 4) {
2001 			if (ctl->free_extents * 2 <= ctl->extents_thresh)
2002 				return false;
2003 		} else {
2004 			return false;
2005 		}
2006 	}
2007 
2008 	/*
2009 	 * The original block groups from mkfs can be really small, like 8
2010 	 * megabytes, so don't bother with a bitmap for those entries.  However
2011 	 * some block groups can be smaller than what a bitmap would cover but
2012 	 * are still large enough that they could overflow the 32k memory limit,
2013 	 * so allow those block groups to still be allowed to have a bitmap
2014 	 * entry.
2015 	 */
2016 	if (((BITS_PER_BITMAP * ctl->unit) >> 1) > block_group->key.offset)
2017 		return false;
2018 
2019 	return true;
2020 }
2021 
2022 static const struct btrfs_free_space_op free_space_op = {
2023 	.recalc_thresholds	= recalculate_thresholds,
2024 	.use_bitmap		= use_bitmap,
2025 };
2026 
2027 static int insert_into_bitmap(struct btrfs_free_space_ctl *ctl,
2028 			      struct btrfs_free_space *info)
2029 {
2030 	struct btrfs_free_space *bitmap_info;
2031 	struct btrfs_block_group_cache *block_group = NULL;
2032 	int added = 0;
2033 	u64 bytes, offset, bytes_added;
2034 	int ret;
2035 
2036 	bytes = info->bytes;
2037 	offset = info->offset;
2038 
2039 	if (!ctl->op->use_bitmap(ctl, info))
2040 		return 0;
2041 
2042 	if (ctl->op == &free_space_op)
2043 		block_group = ctl->private;
2044 again:
2045 	/*
2046 	 * Since we link bitmaps right into the cluster we need to see if we
2047 	 * have a cluster here, and if so and it has our bitmap we need to add
2048 	 * the free space to that bitmap.
2049 	 */
2050 	if (block_group && !list_empty(&block_group->cluster_list)) {
2051 		struct btrfs_free_cluster *cluster;
2052 		struct rb_node *node;
2053 		struct btrfs_free_space *entry;
2054 
2055 		cluster = list_entry(block_group->cluster_list.next,
2056 				     struct btrfs_free_cluster,
2057 				     block_group_list);
2058 		spin_lock(&cluster->lock);
2059 		node = rb_first(&cluster->root);
2060 		if (!node) {
2061 			spin_unlock(&cluster->lock);
2062 			goto no_cluster_bitmap;
2063 		}
2064 
2065 		entry = rb_entry(node, struct btrfs_free_space, offset_index);
2066 		if (!entry->bitmap) {
2067 			spin_unlock(&cluster->lock);
2068 			goto no_cluster_bitmap;
2069 		}
2070 
2071 		if (entry->offset == offset_to_bitmap(ctl, offset)) {
2072 			bytes_added = add_bytes_to_bitmap(ctl, entry,
2073 							  offset, bytes);
2074 			bytes -= bytes_added;
2075 			offset += bytes_added;
2076 		}
2077 		spin_unlock(&cluster->lock);
2078 		if (!bytes) {
2079 			ret = 1;
2080 			goto out;
2081 		}
2082 	}
2083 
2084 no_cluster_bitmap:
2085 	bitmap_info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset),
2086 					 1, 0);
2087 	if (!bitmap_info) {
2088 		ASSERT(added == 0);
2089 		goto new_bitmap;
2090 	}
2091 
2092 	bytes_added = add_bytes_to_bitmap(ctl, bitmap_info, offset, bytes);
2093 	bytes -= bytes_added;
2094 	offset += bytes_added;
2095 	added = 0;
2096 
2097 	if (!bytes) {
2098 		ret = 1;
2099 		goto out;
2100 	} else
2101 		goto again;
2102 
2103 new_bitmap:
2104 	if (info && info->bitmap) {
2105 		add_new_bitmap(ctl, info, offset);
2106 		added = 1;
2107 		info = NULL;
2108 		goto again;
2109 	} else {
2110 		spin_unlock(&ctl->tree_lock);
2111 
2112 		/* no pre-allocated info, allocate a new one */
2113 		if (!info) {
2114 			info = kmem_cache_zalloc(btrfs_free_space_cachep,
2115 						 GFP_NOFS);
2116 			if (!info) {
2117 				spin_lock(&ctl->tree_lock);
2118 				ret = -ENOMEM;
2119 				goto out;
2120 			}
2121 		}
2122 
2123 		/* allocate the bitmap */
2124 		info->bitmap = kzalloc(PAGE_SIZE, GFP_NOFS);
2125 		spin_lock(&ctl->tree_lock);
2126 		if (!info->bitmap) {
2127 			ret = -ENOMEM;
2128 			goto out;
2129 		}
2130 		goto again;
2131 	}
2132 
2133 out:
2134 	if (info) {
2135 		if (info->bitmap)
2136 			kfree(info->bitmap);
2137 		kmem_cache_free(btrfs_free_space_cachep, info);
2138 	}
2139 
2140 	return ret;
2141 }
2142 
2143 static bool try_merge_free_space(struct btrfs_free_space_ctl *ctl,
2144 			  struct btrfs_free_space *info, bool update_stat)
2145 {
2146 	struct btrfs_free_space *left_info;
2147 	struct btrfs_free_space *right_info;
2148 	bool merged = false;
2149 	u64 offset = info->offset;
2150 	u64 bytes = info->bytes;
2151 
2152 	/*
2153 	 * first we want to see if there is free space adjacent to the range we
2154 	 * are adding, if there is remove that struct and add a new one to
2155 	 * cover the entire range
2156 	 */
2157 	right_info = tree_search_offset(ctl, offset + bytes, 0, 0);
2158 	if (right_info && rb_prev(&right_info->offset_index))
2159 		left_info = rb_entry(rb_prev(&right_info->offset_index),
2160 				     struct btrfs_free_space, offset_index);
2161 	else
2162 		left_info = tree_search_offset(ctl, offset - 1, 0, 0);
2163 
2164 	if (right_info && !right_info->bitmap) {
2165 		if (update_stat)
2166 			unlink_free_space(ctl, right_info);
2167 		else
2168 			__unlink_free_space(ctl, right_info);
2169 		info->bytes += right_info->bytes;
2170 		kmem_cache_free(btrfs_free_space_cachep, right_info);
2171 		merged = true;
2172 	}
2173 
2174 	if (left_info && !left_info->bitmap &&
2175 	    left_info->offset + left_info->bytes == offset) {
2176 		if (update_stat)
2177 			unlink_free_space(ctl, left_info);
2178 		else
2179 			__unlink_free_space(ctl, left_info);
2180 		info->offset = left_info->offset;
2181 		info->bytes += left_info->bytes;
2182 		kmem_cache_free(btrfs_free_space_cachep, left_info);
2183 		merged = true;
2184 	}
2185 
2186 	return merged;
2187 }
2188 
2189 static bool steal_from_bitmap_to_end(struct btrfs_free_space_ctl *ctl,
2190 				     struct btrfs_free_space *info,
2191 				     bool update_stat)
2192 {
2193 	struct btrfs_free_space *bitmap;
2194 	unsigned long i;
2195 	unsigned long j;
2196 	const u64 end = info->offset + info->bytes;
2197 	const u64 bitmap_offset = offset_to_bitmap(ctl, end);
2198 	u64 bytes;
2199 
2200 	bitmap = tree_search_offset(ctl, bitmap_offset, 1, 0);
2201 	if (!bitmap)
2202 		return false;
2203 
2204 	i = offset_to_bit(bitmap->offset, ctl->unit, end);
2205 	j = find_next_zero_bit(bitmap->bitmap, BITS_PER_BITMAP, i);
2206 	if (j == i)
2207 		return false;
2208 	bytes = (j - i) * ctl->unit;
2209 	info->bytes += bytes;
2210 
2211 	if (update_stat)
2212 		bitmap_clear_bits(ctl, bitmap, end, bytes);
2213 	else
2214 		__bitmap_clear_bits(ctl, bitmap, end, bytes);
2215 
2216 	if (!bitmap->bytes)
2217 		free_bitmap(ctl, bitmap);
2218 
2219 	return true;
2220 }
2221 
2222 static bool steal_from_bitmap_to_front(struct btrfs_free_space_ctl *ctl,
2223 				       struct btrfs_free_space *info,
2224 				       bool update_stat)
2225 {
2226 	struct btrfs_free_space *bitmap;
2227 	u64 bitmap_offset;
2228 	unsigned long i;
2229 	unsigned long j;
2230 	unsigned long prev_j;
2231 	u64 bytes;
2232 
2233 	bitmap_offset = offset_to_bitmap(ctl, info->offset);
2234 	/* If we're on a boundary, try the previous logical bitmap. */
2235 	if (bitmap_offset == info->offset) {
2236 		if (info->offset == 0)
2237 			return false;
2238 		bitmap_offset = offset_to_bitmap(ctl, info->offset - 1);
2239 	}
2240 
2241 	bitmap = tree_search_offset(ctl, bitmap_offset, 1, 0);
2242 	if (!bitmap)
2243 		return false;
2244 
2245 	i = offset_to_bit(bitmap->offset, ctl->unit, info->offset) - 1;
2246 	j = 0;
2247 	prev_j = (unsigned long)-1;
2248 	for_each_clear_bit_from(j, bitmap->bitmap, BITS_PER_BITMAP) {
2249 		if (j > i)
2250 			break;
2251 		prev_j = j;
2252 	}
2253 	if (prev_j == i)
2254 		return false;
2255 
2256 	if (prev_j == (unsigned long)-1)
2257 		bytes = (i + 1) * ctl->unit;
2258 	else
2259 		bytes = (i - prev_j) * ctl->unit;
2260 
2261 	info->offset -= bytes;
2262 	info->bytes += bytes;
2263 
2264 	if (update_stat)
2265 		bitmap_clear_bits(ctl, bitmap, info->offset, bytes);
2266 	else
2267 		__bitmap_clear_bits(ctl, bitmap, info->offset, bytes);
2268 
2269 	if (!bitmap->bytes)
2270 		free_bitmap(ctl, bitmap);
2271 
2272 	return true;
2273 }
2274 
2275 /*
2276  * We prefer always to allocate from extent entries, both for clustered and
2277  * non-clustered allocation requests. So when attempting to add a new extent
2278  * entry, try to see if there's adjacent free space in bitmap entries, and if
2279  * there is, migrate that space from the bitmaps to the extent.
2280  * Like this we get better chances of satisfying space allocation requests
2281  * because we attempt to satisfy them based on a single cache entry, and never
2282  * on 2 or more entries - even if the entries represent a contiguous free space
2283  * region (e.g. 1 extent entry + 1 bitmap entry starting where the extent entry
2284  * ends).
2285  */
2286 static void steal_from_bitmap(struct btrfs_free_space_ctl *ctl,
2287 			      struct btrfs_free_space *info,
2288 			      bool update_stat)
2289 {
2290 	/*
2291 	 * Only work with disconnected entries, as we can change their offset,
2292 	 * and must be extent entries.
2293 	 */
2294 	ASSERT(!info->bitmap);
2295 	ASSERT(RB_EMPTY_NODE(&info->offset_index));
2296 
2297 	if (ctl->total_bitmaps > 0) {
2298 		bool stole_end;
2299 		bool stole_front = false;
2300 
2301 		stole_end = steal_from_bitmap_to_end(ctl, info, update_stat);
2302 		if (ctl->total_bitmaps > 0)
2303 			stole_front = steal_from_bitmap_to_front(ctl, info,
2304 								 update_stat);
2305 
2306 		if (stole_end || stole_front)
2307 			try_merge_free_space(ctl, info, update_stat);
2308 	}
2309 }
2310 
2311 int __btrfs_add_free_space(struct btrfs_fs_info *fs_info,
2312 			   struct btrfs_free_space_ctl *ctl,
2313 			   u64 offset, u64 bytes)
2314 {
2315 	struct btrfs_free_space *info;
2316 	int ret = 0;
2317 
2318 	info = kmem_cache_zalloc(btrfs_free_space_cachep, GFP_NOFS);
2319 	if (!info)
2320 		return -ENOMEM;
2321 
2322 	info->offset = offset;
2323 	info->bytes = bytes;
2324 	RB_CLEAR_NODE(&info->offset_index);
2325 
2326 	spin_lock(&ctl->tree_lock);
2327 
2328 	if (try_merge_free_space(ctl, info, true))
2329 		goto link;
2330 
2331 	/*
2332 	 * There was no extent directly to the left or right of this new
2333 	 * extent then we know we're going to have to allocate a new extent, so
2334 	 * before we do that see if we need to drop this into a bitmap
2335 	 */
2336 	ret = insert_into_bitmap(ctl, info);
2337 	if (ret < 0) {
2338 		goto out;
2339 	} else if (ret) {
2340 		ret = 0;
2341 		goto out;
2342 	}
2343 link:
2344 	/*
2345 	 * Only steal free space from adjacent bitmaps if we're sure we're not
2346 	 * going to add the new free space to existing bitmap entries - because
2347 	 * that would mean unnecessary work that would be reverted. Therefore
2348 	 * attempt to steal space from bitmaps if we're adding an extent entry.
2349 	 */
2350 	steal_from_bitmap(ctl, info, true);
2351 
2352 	ret = link_free_space(ctl, info);
2353 	if (ret)
2354 		kmem_cache_free(btrfs_free_space_cachep, info);
2355 out:
2356 	spin_unlock(&ctl->tree_lock);
2357 
2358 	if (ret) {
2359 		btrfs_crit(fs_info, "unable to add free space :%d", ret);
2360 		ASSERT(ret != -EEXIST);
2361 	}
2362 
2363 	return ret;
2364 }
2365 
2366 int btrfs_remove_free_space(struct btrfs_block_group_cache *block_group,
2367 			    u64 offset, u64 bytes)
2368 {
2369 	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2370 	struct btrfs_free_space *info;
2371 	int ret;
2372 	bool re_search = false;
2373 
2374 	spin_lock(&ctl->tree_lock);
2375 
2376 again:
2377 	ret = 0;
2378 	if (!bytes)
2379 		goto out_lock;
2380 
2381 	info = tree_search_offset(ctl, offset, 0, 0);
2382 	if (!info) {
2383 		/*
2384 		 * oops didn't find an extent that matched the space we wanted
2385 		 * to remove, look for a bitmap instead
2386 		 */
2387 		info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset),
2388 					  1, 0);
2389 		if (!info) {
2390 			/*
2391 			 * If we found a partial bit of our free space in a
2392 			 * bitmap but then couldn't find the other part this may
2393 			 * be a problem, so WARN about it.
2394 			 */
2395 			WARN_ON(re_search);
2396 			goto out_lock;
2397 		}
2398 	}
2399 
2400 	re_search = false;
2401 	if (!info->bitmap) {
2402 		unlink_free_space(ctl, info);
2403 		if (offset == info->offset) {
2404 			u64 to_free = min(bytes, info->bytes);
2405 
2406 			info->bytes -= to_free;
2407 			info->offset += to_free;
2408 			if (info->bytes) {
2409 				ret = link_free_space(ctl, info);
2410 				WARN_ON(ret);
2411 			} else {
2412 				kmem_cache_free(btrfs_free_space_cachep, info);
2413 			}
2414 
2415 			offset += to_free;
2416 			bytes -= to_free;
2417 			goto again;
2418 		} else {
2419 			u64 old_end = info->bytes + info->offset;
2420 
2421 			info->bytes = offset - info->offset;
2422 			ret = link_free_space(ctl, info);
2423 			WARN_ON(ret);
2424 			if (ret)
2425 				goto out_lock;
2426 
2427 			/* Not enough bytes in this entry to satisfy us */
2428 			if (old_end < offset + bytes) {
2429 				bytes -= old_end - offset;
2430 				offset = old_end;
2431 				goto again;
2432 			} else if (old_end == offset + bytes) {
2433 				/* all done */
2434 				goto out_lock;
2435 			}
2436 			spin_unlock(&ctl->tree_lock);
2437 
2438 			ret = btrfs_add_free_space(block_group, offset + bytes,
2439 						   old_end - (offset + bytes));
2440 			WARN_ON(ret);
2441 			goto out;
2442 		}
2443 	}
2444 
2445 	ret = remove_from_bitmap(ctl, info, &offset, &bytes);
2446 	if (ret == -EAGAIN) {
2447 		re_search = true;
2448 		goto again;
2449 	}
2450 out_lock:
2451 	spin_unlock(&ctl->tree_lock);
2452 out:
2453 	return ret;
2454 }
2455 
2456 void btrfs_dump_free_space(struct btrfs_block_group_cache *block_group,
2457 			   u64 bytes)
2458 {
2459 	struct btrfs_fs_info *fs_info = block_group->fs_info;
2460 	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2461 	struct btrfs_free_space *info;
2462 	struct rb_node *n;
2463 	int count = 0;
2464 
2465 	for (n = rb_first(&ctl->free_space_offset); n; n = rb_next(n)) {
2466 		info = rb_entry(n, struct btrfs_free_space, offset_index);
2467 		if (info->bytes >= bytes && !block_group->ro)
2468 			count++;
2469 		btrfs_crit(fs_info, "entry offset %llu, bytes %llu, bitmap %s",
2470 			   info->offset, info->bytes,
2471 		       (info->bitmap) ? "yes" : "no");
2472 	}
2473 	btrfs_info(fs_info, "block group has cluster?: %s",
2474 	       list_empty(&block_group->cluster_list) ? "no" : "yes");
2475 	btrfs_info(fs_info,
2476 		   "%d blocks of free space at or bigger than bytes is", count);
2477 }
2478 
2479 void btrfs_init_free_space_ctl(struct btrfs_block_group_cache *block_group)
2480 {
2481 	struct btrfs_fs_info *fs_info = block_group->fs_info;
2482 	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2483 
2484 	spin_lock_init(&ctl->tree_lock);
2485 	ctl->unit = fs_info->sectorsize;
2486 	ctl->start = block_group->key.objectid;
2487 	ctl->private = block_group;
2488 	ctl->op = &free_space_op;
2489 	INIT_LIST_HEAD(&ctl->trimming_ranges);
2490 	mutex_init(&ctl->cache_writeout_mutex);
2491 
2492 	/*
2493 	 * we only want to have 32k of ram per block group for keeping
2494 	 * track of free space, and if we pass 1/2 of that we want to
2495 	 * start converting things over to using bitmaps
2496 	 */
2497 	ctl->extents_thresh = (SZ_32K / 2) / sizeof(struct btrfs_free_space);
2498 }
2499 
2500 /*
2501  * for a given cluster, put all of its extents back into the free
2502  * space cache.  If the block group passed doesn't match the block group
2503  * pointed to by the cluster, someone else raced in and freed the
2504  * cluster already.  In that case, we just return without changing anything
2505  */
2506 static int
2507 __btrfs_return_cluster_to_free_space(
2508 			     struct btrfs_block_group_cache *block_group,
2509 			     struct btrfs_free_cluster *cluster)
2510 {
2511 	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2512 	struct btrfs_free_space *entry;
2513 	struct rb_node *node;
2514 
2515 	spin_lock(&cluster->lock);
2516 	if (cluster->block_group != block_group)
2517 		goto out;
2518 
2519 	cluster->block_group = NULL;
2520 	cluster->window_start = 0;
2521 	list_del_init(&cluster->block_group_list);
2522 
2523 	node = rb_first(&cluster->root);
2524 	while (node) {
2525 		bool bitmap;
2526 
2527 		entry = rb_entry(node, struct btrfs_free_space, offset_index);
2528 		node = rb_next(&entry->offset_index);
2529 		rb_erase(&entry->offset_index, &cluster->root);
2530 		RB_CLEAR_NODE(&entry->offset_index);
2531 
2532 		bitmap = (entry->bitmap != NULL);
2533 		if (!bitmap) {
2534 			try_merge_free_space(ctl, entry, false);
2535 			steal_from_bitmap(ctl, entry, false);
2536 		}
2537 		tree_insert_offset(&ctl->free_space_offset,
2538 				   entry->offset, &entry->offset_index, bitmap);
2539 	}
2540 	cluster->root = RB_ROOT;
2541 
2542 out:
2543 	spin_unlock(&cluster->lock);
2544 	btrfs_put_block_group(block_group);
2545 	return 0;
2546 }
2547 
2548 static void __btrfs_remove_free_space_cache_locked(
2549 				struct btrfs_free_space_ctl *ctl)
2550 {
2551 	struct btrfs_free_space *info;
2552 	struct rb_node *node;
2553 
2554 	while ((node = rb_last(&ctl->free_space_offset)) != NULL) {
2555 		info = rb_entry(node, struct btrfs_free_space, offset_index);
2556 		if (!info->bitmap) {
2557 			unlink_free_space(ctl, info);
2558 			kmem_cache_free(btrfs_free_space_cachep, info);
2559 		} else {
2560 			free_bitmap(ctl, info);
2561 		}
2562 
2563 		cond_resched_lock(&ctl->tree_lock);
2564 	}
2565 }
2566 
2567 void __btrfs_remove_free_space_cache(struct btrfs_free_space_ctl *ctl)
2568 {
2569 	spin_lock(&ctl->tree_lock);
2570 	__btrfs_remove_free_space_cache_locked(ctl);
2571 	spin_unlock(&ctl->tree_lock);
2572 }
2573 
2574 void btrfs_remove_free_space_cache(struct btrfs_block_group_cache *block_group)
2575 {
2576 	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2577 	struct btrfs_free_cluster *cluster;
2578 	struct list_head *head;
2579 
2580 	spin_lock(&ctl->tree_lock);
2581 	while ((head = block_group->cluster_list.next) !=
2582 	       &block_group->cluster_list) {
2583 		cluster = list_entry(head, struct btrfs_free_cluster,
2584 				     block_group_list);
2585 
2586 		WARN_ON(cluster->block_group != block_group);
2587 		__btrfs_return_cluster_to_free_space(block_group, cluster);
2588 
2589 		cond_resched_lock(&ctl->tree_lock);
2590 	}
2591 	__btrfs_remove_free_space_cache_locked(ctl);
2592 	spin_unlock(&ctl->tree_lock);
2593 
2594 }
2595 
2596 u64 btrfs_find_space_for_alloc(struct btrfs_block_group_cache *block_group,
2597 			       u64 offset, u64 bytes, u64 empty_size,
2598 			       u64 *max_extent_size)
2599 {
2600 	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2601 	struct btrfs_free_space *entry = NULL;
2602 	u64 bytes_search = bytes + empty_size;
2603 	u64 ret = 0;
2604 	u64 align_gap = 0;
2605 	u64 align_gap_len = 0;
2606 
2607 	spin_lock(&ctl->tree_lock);
2608 	entry = find_free_space(ctl, &offset, &bytes_search,
2609 				block_group->full_stripe_len, max_extent_size);
2610 	if (!entry)
2611 		goto out;
2612 
2613 	ret = offset;
2614 	if (entry->bitmap) {
2615 		bitmap_clear_bits(ctl, entry, offset, bytes);
2616 		if (!entry->bytes)
2617 			free_bitmap(ctl, entry);
2618 	} else {
2619 		unlink_free_space(ctl, entry);
2620 		align_gap_len = offset - entry->offset;
2621 		align_gap = entry->offset;
2622 
2623 		entry->offset = offset + bytes;
2624 		WARN_ON(entry->bytes < bytes + align_gap_len);
2625 
2626 		entry->bytes -= bytes + align_gap_len;
2627 		if (!entry->bytes)
2628 			kmem_cache_free(btrfs_free_space_cachep, entry);
2629 		else
2630 			link_free_space(ctl, entry);
2631 	}
2632 out:
2633 	spin_unlock(&ctl->tree_lock);
2634 
2635 	if (align_gap_len)
2636 		__btrfs_add_free_space(block_group->fs_info, ctl,
2637 				       align_gap, align_gap_len);
2638 	return ret;
2639 }
2640 
2641 /*
2642  * given a cluster, put all of its extents back into the free space
2643  * cache.  If a block group is passed, this function will only free
2644  * a cluster that belongs to the passed block group.
2645  *
2646  * Otherwise, it'll get a reference on the block group pointed to by the
2647  * cluster and remove the cluster from it.
2648  */
2649 int btrfs_return_cluster_to_free_space(
2650 			       struct btrfs_block_group_cache *block_group,
2651 			       struct btrfs_free_cluster *cluster)
2652 {
2653 	struct btrfs_free_space_ctl *ctl;
2654 	int ret;
2655 
2656 	/* first, get a safe pointer to the block group */
2657 	spin_lock(&cluster->lock);
2658 	if (!block_group) {
2659 		block_group = cluster->block_group;
2660 		if (!block_group) {
2661 			spin_unlock(&cluster->lock);
2662 			return 0;
2663 		}
2664 	} else if (cluster->block_group != block_group) {
2665 		/* someone else has already freed it don't redo their work */
2666 		spin_unlock(&cluster->lock);
2667 		return 0;
2668 	}
2669 	atomic_inc(&block_group->count);
2670 	spin_unlock(&cluster->lock);
2671 
2672 	ctl = block_group->free_space_ctl;
2673 
2674 	/* now return any extents the cluster had on it */
2675 	spin_lock(&ctl->tree_lock);
2676 	ret = __btrfs_return_cluster_to_free_space(block_group, cluster);
2677 	spin_unlock(&ctl->tree_lock);
2678 
2679 	/* finally drop our ref */
2680 	btrfs_put_block_group(block_group);
2681 	return ret;
2682 }
2683 
2684 static u64 btrfs_alloc_from_bitmap(struct btrfs_block_group_cache *block_group,
2685 				   struct btrfs_free_cluster *cluster,
2686 				   struct btrfs_free_space *entry,
2687 				   u64 bytes, u64 min_start,
2688 				   u64 *max_extent_size)
2689 {
2690 	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2691 	int err;
2692 	u64 search_start = cluster->window_start;
2693 	u64 search_bytes = bytes;
2694 	u64 ret = 0;
2695 
2696 	search_start = min_start;
2697 	search_bytes = bytes;
2698 
2699 	err = search_bitmap(ctl, entry, &search_start, &search_bytes, true);
2700 	if (err) {
2701 		if (search_bytes > *max_extent_size)
2702 			*max_extent_size = search_bytes;
2703 		return 0;
2704 	}
2705 
2706 	ret = search_start;
2707 	__bitmap_clear_bits(ctl, entry, ret, bytes);
2708 
2709 	return ret;
2710 }
2711 
2712 /*
2713  * given a cluster, try to allocate 'bytes' from it, returns 0
2714  * if it couldn't find anything suitably large, or a logical disk offset
2715  * if things worked out
2716  */
2717 u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group,
2718 			     struct btrfs_free_cluster *cluster, u64 bytes,
2719 			     u64 min_start, u64 *max_extent_size)
2720 {
2721 	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2722 	struct btrfs_free_space *entry = NULL;
2723 	struct rb_node *node;
2724 	u64 ret = 0;
2725 
2726 	spin_lock(&cluster->lock);
2727 	if (bytes > cluster->max_size)
2728 		goto out;
2729 
2730 	if (cluster->block_group != block_group)
2731 		goto out;
2732 
2733 	node = rb_first(&cluster->root);
2734 	if (!node)
2735 		goto out;
2736 
2737 	entry = rb_entry(node, struct btrfs_free_space, offset_index);
2738 	while (1) {
2739 		if (entry->bytes < bytes && entry->bytes > *max_extent_size)
2740 			*max_extent_size = entry->bytes;
2741 
2742 		if (entry->bytes < bytes ||
2743 		    (!entry->bitmap && entry->offset < min_start)) {
2744 			node = rb_next(&entry->offset_index);
2745 			if (!node)
2746 				break;
2747 			entry = rb_entry(node, struct btrfs_free_space,
2748 					 offset_index);
2749 			continue;
2750 		}
2751 
2752 		if (entry->bitmap) {
2753 			ret = btrfs_alloc_from_bitmap(block_group,
2754 						      cluster, entry, bytes,
2755 						      cluster->window_start,
2756 						      max_extent_size);
2757 			if (ret == 0) {
2758 				node = rb_next(&entry->offset_index);
2759 				if (!node)
2760 					break;
2761 				entry = rb_entry(node, struct btrfs_free_space,
2762 						 offset_index);
2763 				continue;
2764 			}
2765 			cluster->window_start += bytes;
2766 		} else {
2767 			ret = entry->offset;
2768 
2769 			entry->offset += bytes;
2770 			entry->bytes -= bytes;
2771 		}
2772 
2773 		if (entry->bytes == 0)
2774 			rb_erase(&entry->offset_index, &cluster->root);
2775 		break;
2776 	}
2777 out:
2778 	spin_unlock(&cluster->lock);
2779 
2780 	if (!ret)
2781 		return 0;
2782 
2783 	spin_lock(&ctl->tree_lock);
2784 
2785 	ctl->free_space -= bytes;
2786 	if (entry->bytes == 0) {
2787 		ctl->free_extents--;
2788 		if (entry->bitmap) {
2789 			kfree(entry->bitmap);
2790 			ctl->total_bitmaps--;
2791 			ctl->op->recalc_thresholds(ctl);
2792 		}
2793 		kmem_cache_free(btrfs_free_space_cachep, entry);
2794 	}
2795 
2796 	spin_unlock(&ctl->tree_lock);
2797 
2798 	return ret;
2799 }
2800 
2801 static int btrfs_bitmap_cluster(struct btrfs_block_group_cache *block_group,
2802 				struct btrfs_free_space *entry,
2803 				struct btrfs_free_cluster *cluster,
2804 				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 	unsigned long next_zero;
2809 	unsigned long i;
2810 	unsigned long want_bits;
2811 	unsigned long min_bits;
2812 	unsigned long found_bits;
2813 	unsigned long max_bits = 0;
2814 	unsigned long start = 0;
2815 	unsigned long total_found = 0;
2816 	int ret;
2817 
2818 	i = offset_to_bit(entry->offset, ctl->unit,
2819 			  max_t(u64, offset, entry->offset));
2820 	want_bits = bytes_to_bits(bytes, ctl->unit);
2821 	min_bits = bytes_to_bits(min_bytes, ctl->unit);
2822 
2823 	/*
2824 	 * Don't bother looking for a cluster in this bitmap if it's heavily
2825 	 * fragmented.
2826 	 */
2827 	if (entry->max_extent_size &&
2828 	    entry->max_extent_size < cont1_bytes)
2829 		return -ENOSPC;
2830 again:
2831 	found_bits = 0;
2832 	for_each_set_bit_from(i, entry->bitmap, BITS_PER_BITMAP) {
2833 		next_zero = find_next_zero_bit(entry->bitmap,
2834 					       BITS_PER_BITMAP, i);
2835 		if (next_zero - i >= min_bits) {
2836 			found_bits = next_zero - i;
2837 			if (found_bits > max_bits)
2838 				max_bits = found_bits;
2839 			break;
2840 		}
2841 		if (next_zero - i > max_bits)
2842 			max_bits = next_zero - i;
2843 		i = next_zero;
2844 	}
2845 
2846 	if (!found_bits) {
2847 		entry->max_extent_size = (u64)max_bits * ctl->unit;
2848 		return -ENOSPC;
2849 	}
2850 
2851 	if (!total_found) {
2852 		start = i;
2853 		cluster->max_size = 0;
2854 	}
2855 
2856 	total_found += found_bits;
2857 
2858 	if (cluster->max_size < found_bits * ctl->unit)
2859 		cluster->max_size = found_bits * ctl->unit;
2860 
2861 	if (total_found < want_bits || cluster->max_size < cont1_bytes) {
2862 		i = next_zero + 1;
2863 		goto again;
2864 	}
2865 
2866 	cluster->window_start = start * ctl->unit + entry->offset;
2867 	rb_erase(&entry->offset_index, &ctl->free_space_offset);
2868 	ret = tree_insert_offset(&cluster->root, entry->offset,
2869 				 &entry->offset_index, 1);
2870 	ASSERT(!ret); /* -EEXIST; Logic error */
2871 
2872 	trace_btrfs_setup_cluster(block_group, cluster,
2873 				  total_found * ctl->unit, 1);
2874 	return 0;
2875 }
2876 
2877 /*
2878  * This searches the block group for just extents to fill the cluster with.
2879  * Try to find a cluster with at least bytes total bytes, at least one
2880  * extent of cont1_bytes, and other clusters of at least min_bytes.
2881  */
2882 static noinline int
2883 setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group,
2884 			struct btrfs_free_cluster *cluster,
2885 			struct list_head *bitmaps, u64 offset, u64 bytes,
2886 			u64 cont1_bytes, u64 min_bytes)
2887 {
2888 	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2889 	struct btrfs_free_space *first = NULL;
2890 	struct btrfs_free_space *entry = NULL;
2891 	struct btrfs_free_space *last;
2892 	struct rb_node *node;
2893 	u64 window_free;
2894 	u64 max_extent;
2895 	u64 total_size = 0;
2896 
2897 	entry = tree_search_offset(ctl, offset, 0, 1);
2898 	if (!entry)
2899 		return -ENOSPC;
2900 
2901 	/*
2902 	 * We don't want bitmaps, so just move along until we find a normal
2903 	 * extent entry.
2904 	 */
2905 	while (entry->bitmap || entry->bytes < min_bytes) {
2906 		if (entry->bitmap && list_empty(&entry->list))
2907 			list_add_tail(&entry->list, bitmaps);
2908 		node = rb_next(&entry->offset_index);
2909 		if (!node)
2910 			return -ENOSPC;
2911 		entry = rb_entry(node, struct btrfs_free_space, offset_index);
2912 	}
2913 
2914 	window_free = entry->bytes;
2915 	max_extent = entry->bytes;
2916 	first = entry;
2917 	last = entry;
2918 
2919 	for (node = rb_next(&entry->offset_index); node;
2920 	     node = rb_next(&entry->offset_index)) {
2921 		entry = rb_entry(node, struct btrfs_free_space, offset_index);
2922 
2923 		if (entry->bitmap) {
2924 			if (list_empty(&entry->list))
2925 				list_add_tail(&entry->list, bitmaps);
2926 			continue;
2927 		}
2928 
2929 		if (entry->bytes < min_bytes)
2930 			continue;
2931 
2932 		last = entry;
2933 		window_free += entry->bytes;
2934 		if (entry->bytes > max_extent)
2935 			max_extent = entry->bytes;
2936 	}
2937 
2938 	if (window_free < bytes || max_extent < cont1_bytes)
2939 		return -ENOSPC;
2940 
2941 	cluster->window_start = first->offset;
2942 
2943 	node = &first->offset_index;
2944 
2945 	/*
2946 	 * now we've found our entries, pull them out of the free space
2947 	 * cache and put them into the cluster rbtree
2948 	 */
2949 	do {
2950 		int ret;
2951 
2952 		entry = rb_entry(node, struct btrfs_free_space, offset_index);
2953 		node = rb_next(&entry->offset_index);
2954 		if (entry->bitmap || entry->bytes < min_bytes)
2955 			continue;
2956 
2957 		rb_erase(&entry->offset_index, &ctl->free_space_offset);
2958 		ret = tree_insert_offset(&cluster->root, entry->offset,
2959 					 &entry->offset_index, 0);
2960 		total_size += entry->bytes;
2961 		ASSERT(!ret); /* -EEXIST; Logic error */
2962 	} while (node && entry != last);
2963 
2964 	cluster->max_size = max_extent;
2965 	trace_btrfs_setup_cluster(block_group, cluster, total_size, 0);
2966 	return 0;
2967 }
2968 
2969 /*
2970  * This specifically looks for bitmaps that may work in the cluster, we assume
2971  * that we have already failed to find extents that will work.
2972  */
2973 static noinline int
2974 setup_cluster_bitmap(struct btrfs_block_group_cache *block_group,
2975 		     struct btrfs_free_cluster *cluster,
2976 		     struct list_head *bitmaps, u64 offset, u64 bytes,
2977 		     u64 cont1_bytes, u64 min_bytes)
2978 {
2979 	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
2980 	struct btrfs_free_space *entry = NULL;
2981 	int ret = -ENOSPC;
2982 	u64 bitmap_offset = offset_to_bitmap(ctl, offset);
2983 
2984 	if (ctl->total_bitmaps == 0)
2985 		return -ENOSPC;
2986 
2987 	/*
2988 	 * The bitmap that covers offset won't be in the list unless offset
2989 	 * is just its start offset.
2990 	 */
2991 	if (!list_empty(bitmaps))
2992 		entry = list_first_entry(bitmaps, struct btrfs_free_space, list);
2993 
2994 	if (!entry || entry->offset != bitmap_offset) {
2995 		entry = tree_search_offset(ctl, bitmap_offset, 1, 0);
2996 		if (entry && list_empty(&entry->list))
2997 			list_add(&entry->list, bitmaps);
2998 	}
2999 
3000 	list_for_each_entry(entry, bitmaps, list) {
3001 		if (entry->bytes < bytes)
3002 			continue;
3003 		ret = btrfs_bitmap_cluster(block_group, entry, cluster, offset,
3004 					   bytes, cont1_bytes, min_bytes);
3005 		if (!ret)
3006 			return 0;
3007 	}
3008 
3009 	/*
3010 	 * The bitmaps list has all the bitmaps that record free space
3011 	 * starting after offset, so no more search is required.
3012 	 */
3013 	return -ENOSPC;
3014 }
3015 
3016 /*
3017  * here we try to find a cluster of blocks in a block group.  The goal
3018  * is to find at least bytes+empty_size.
3019  * We might not find them all in one contiguous area.
3020  *
3021  * returns zero and sets up cluster if things worked out, otherwise
3022  * it returns -enospc
3023  */
3024 int btrfs_find_space_cluster(struct btrfs_fs_info *fs_info,
3025 			     struct btrfs_block_group_cache *block_group,
3026 			     struct btrfs_free_cluster *cluster,
3027 			     u64 offset, u64 bytes, u64 empty_size)
3028 {
3029 	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
3030 	struct btrfs_free_space *entry, *tmp;
3031 	LIST_HEAD(bitmaps);
3032 	u64 min_bytes;
3033 	u64 cont1_bytes;
3034 	int ret;
3035 
3036 	/*
3037 	 * Choose the minimum extent size we'll require for this
3038 	 * cluster.  For SSD_SPREAD, don't allow any fragmentation.
3039 	 * For metadata, allow allocates with smaller extents.  For
3040 	 * data, keep it dense.
3041 	 */
3042 	if (btrfs_test_opt(fs_info, SSD_SPREAD)) {
3043 		cont1_bytes = min_bytes = bytes + empty_size;
3044 	} else if (block_group->flags & BTRFS_BLOCK_GROUP_METADATA) {
3045 		cont1_bytes = bytes;
3046 		min_bytes = fs_info->sectorsize;
3047 	} else {
3048 		cont1_bytes = max(bytes, (bytes + empty_size) >> 2);
3049 		min_bytes = fs_info->sectorsize;
3050 	}
3051 
3052 	spin_lock(&ctl->tree_lock);
3053 
3054 	/*
3055 	 * If we know we don't have enough space to make a cluster don't even
3056 	 * bother doing all the work to try and find one.
3057 	 */
3058 	if (ctl->free_space < bytes) {
3059 		spin_unlock(&ctl->tree_lock);
3060 		return -ENOSPC;
3061 	}
3062 
3063 	spin_lock(&cluster->lock);
3064 
3065 	/* someone already found a cluster, hooray */
3066 	if (cluster->block_group) {
3067 		ret = 0;
3068 		goto out;
3069 	}
3070 
3071 	trace_btrfs_find_cluster(block_group, offset, bytes, empty_size,
3072 				 min_bytes);
3073 
3074 	ret = setup_cluster_no_bitmap(block_group, cluster, &bitmaps, offset,
3075 				      bytes + empty_size,
3076 				      cont1_bytes, min_bytes);
3077 	if (ret)
3078 		ret = setup_cluster_bitmap(block_group, cluster, &bitmaps,
3079 					   offset, bytes + empty_size,
3080 					   cont1_bytes, min_bytes);
3081 
3082 	/* Clear our temporary list */
3083 	list_for_each_entry_safe(entry, tmp, &bitmaps, list)
3084 		list_del_init(&entry->list);
3085 
3086 	if (!ret) {
3087 		atomic_inc(&block_group->count);
3088 		list_add_tail(&cluster->block_group_list,
3089 			      &block_group->cluster_list);
3090 		cluster->block_group = block_group;
3091 	} else {
3092 		trace_btrfs_failed_cluster_setup(block_group);
3093 	}
3094 out:
3095 	spin_unlock(&cluster->lock);
3096 	spin_unlock(&ctl->tree_lock);
3097 
3098 	return ret;
3099 }
3100 
3101 /*
3102  * simple code to zero out a cluster
3103  */
3104 void btrfs_init_free_cluster(struct btrfs_free_cluster *cluster)
3105 {
3106 	spin_lock_init(&cluster->lock);
3107 	spin_lock_init(&cluster->refill_lock);
3108 	cluster->root = RB_ROOT;
3109 	cluster->max_size = 0;
3110 	cluster->fragmented = false;
3111 	INIT_LIST_HEAD(&cluster->block_group_list);
3112 	cluster->block_group = NULL;
3113 }
3114 
3115 static int do_trimming(struct btrfs_block_group_cache *block_group,
3116 		       u64 *total_trimmed, u64 start, u64 bytes,
3117 		       u64 reserved_start, u64 reserved_bytes,
3118 		       struct btrfs_trim_range *trim_entry)
3119 {
3120 	struct btrfs_space_info *space_info = block_group->space_info;
3121 	struct btrfs_fs_info *fs_info = block_group->fs_info;
3122 	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
3123 	int ret;
3124 	int update = 0;
3125 	u64 trimmed = 0;
3126 
3127 	spin_lock(&space_info->lock);
3128 	spin_lock(&block_group->lock);
3129 	if (!block_group->ro) {
3130 		block_group->reserved += reserved_bytes;
3131 		space_info->bytes_reserved += reserved_bytes;
3132 		update = 1;
3133 	}
3134 	spin_unlock(&block_group->lock);
3135 	spin_unlock(&space_info->lock);
3136 
3137 	ret = btrfs_discard_extent(fs_info, start, bytes, &trimmed);
3138 	if (!ret)
3139 		*total_trimmed += trimmed;
3140 
3141 	mutex_lock(&ctl->cache_writeout_mutex);
3142 	btrfs_add_free_space(block_group, reserved_start, reserved_bytes);
3143 	list_del(&trim_entry->list);
3144 	mutex_unlock(&ctl->cache_writeout_mutex);
3145 
3146 	if (update) {
3147 		spin_lock(&space_info->lock);
3148 		spin_lock(&block_group->lock);
3149 		if (block_group->ro)
3150 			space_info->bytes_readonly += reserved_bytes;
3151 		block_group->reserved -= reserved_bytes;
3152 		space_info->bytes_reserved -= reserved_bytes;
3153 		spin_unlock(&space_info->lock);
3154 		spin_unlock(&block_group->lock);
3155 	}
3156 
3157 	return ret;
3158 }
3159 
3160 static int trim_no_bitmap(struct btrfs_block_group_cache *block_group,
3161 			  u64 *total_trimmed, u64 start, u64 end, u64 minlen)
3162 {
3163 	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
3164 	struct btrfs_free_space *entry;
3165 	struct rb_node *node;
3166 	int ret = 0;
3167 	u64 extent_start;
3168 	u64 extent_bytes;
3169 	u64 bytes;
3170 
3171 	while (start < end) {
3172 		struct btrfs_trim_range trim_entry;
3173 
3174 		mutex_lock(&ctl->cache_writeout_mutex);
3175 		spin_lock(&ctl->tree_lock);
3176 
3177 		if (ctl->free_space < minlen) {
3178 			spin_unlock(&ctl->tree_lock);
3179 			mutex_unlock(&ctl->cache_writeout_mutex);
3180 			break;
3181 		}
3182 
3183 		entry = tree_search_offset(ctl, start, 0, 1);
3184 		if (!entry) {
3185 			spin_unlock(&ctl->tree_lock);
3186 			mutex_unlock(&ctl->cache_writeout_mutex);
3187 			break;
3188 		}
3189 
3190 		/* skip bitmaps */
3191 		while (entry->bitmap) {
3192 			node = rb_next(&entry->offset_index);
3193 			if (!node) {
3194 				spin_unlock(&ctl->tree_lock);
3195 				mutex_unlock(&ctl->cache_writeout_mutex);
3196 				goto out;
3197 			}
3198 			entry = rb_entry(node, struct btrfs_free_space,
3199 					 offset_index);
3200 		}
3201 
3202 		if (entry->offset >= end) {
3203 			spin_unlock(&ctl->tree_lock);
3204 			mutex_unlock(&ctl->cache_writeout_mutex);
3205 			break;
3206 		}
3207 
3208 		extent_start = entry->offset;
3209 		extent_bytes = entry->bytes;
3210 		start = max(start, extent_start);
3211 		bytes = min(extent_start + extent_bytes, end) - start;
3212 		if (bytes < minlen) {
3213 			spin_unlock(&ctl->tree_lock);
3214 			mutex_unlock(&ctl->cache_writeout_mutex);
3215 			goto next;
3216 		}
3217 
3218 		unlink_free_space(ctl, entry);
3219 		kmem_cache_free(btrfs_free_space_cachep, entry);
3220 
3221 		spin_unlock(&ctl->tree_lock);
3222 		trim_entry.start = extent_start;
3223 		trim_entry.bytes = extent_bytes;
3224 		list_add_tail(&trim_entry.list, &ctl->trimming_ranges);
3225 		mutex_unlock(&ctl->cache_writeout_mutex);
3226 
3227 		ret = do_trimming(block_group, total_trimmed, start, bytes,
3228 				  extent_start, extent_bytes, &trim_entry);
3229 		if (ret)
3230 			break;
3231 next:
3232 		start += bytes;
3233 
3234 		if (fatal_signal_pending(current)) {
3235 			ret = -ERESTARTSYS;
3236 			break;
3237 		}
3238 
3239 		cond_resched();
3240 	}
3241 out:
3242 	return ret;
3243 }
3244 
3245 static int trim_bitmaps(struct btrfs_block_group_cache *block_group,
3246 			u64 *total_trimmed, u64 start, u64 end, u64 minlen)
3247 {
3248 	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
3249 	struct btrfs_free_space *entry;
3250 	int ret = 0;
3251 	int ret2;
3252 	u64 bytes;
3253 	u64 offset = offset_to_bitmap(ctl, start);
3254 
3255 	while (offset < end) {
3256 		bool next_bitmap = false;
3257 		struct btrfs_trim_range trim_entry;
3258 
3259 		mutex_lock(&ctl->cache_writeout_mutex);
3260 		spin_lock(&ctl->tree_lock);
3261 
3262 		if (ctl->free_space < minlen) {
3263 			spin_unlock(&ctl->tree_lock);
3264 			mutex_unlock(&ctl->cache_writeout_mutex);
3265 			break;
3266 		}
3267 
3268 		entry = tree_search_offset(ctl, offset, 1, 0);
3269 		if (!entry) {
3270 			spin_unlock(&ctl->tree_lock);
3271 			mutex_unlock(&ctl->cache_writeout_mutex);
3272 			next_bitmap = true;
3273 			goto next;
3274 		}
3275 
3276 		bytes = minlen;
3277 		ret2 = search_bitmap(ctl, entry, &start, &bytes, false);
3278 		if (ret2 || start >= end) {
3279 			spin_unlock(&ctl->tree_lock);
3280 			mutex_unlock(&ctl->cache_writeout_mutex);
3281 			next_bitmap = true;
3282 			goto next;
3283 		}
3284 
3285 		bytes = min(bytes, end - start);
3286 		if (bytes < minlen) {
3287 			spin_unlock(&ctl->tree_lock);
3288 			mutex_unlock(&ctl->cache_writeout_mutex);
3289 			goto next;
3290 		}
3291 
3292 		bitmap_clear_bits(ctl, entry, start, bytes);
3293 		if (entry->bytes == 0)
3294 			free_bitmap(ctl, entry);
3295 
3296 		spin_unlock(&ctl->tree_lock);
3297 		trim_entry.start = start;
3298 		trim_entry.bytes = bytes;
3299 		list_add_tail(&trim_entry.list, &ctl->trimming_ranges);
3300 		mutex_unlock(&ctl->cache_writeout_mutex);
3301 
3302 		ret = do_trimming(block_group, total_trimmed, start, bytes,
3303 				  start, bytes, &trim_entry);
3304 		if (ret)
3305 			break;
3306 next:
3307 		if (next_bitmap) {
3308 			offset += BITS_PER_BITMAP * ctl->unit;
3309 		} else {
3310 			start += bytes;
3311 			if (start >= offset + BITS_PER_BITMAP * ctl->unit)
3312 				offset += BITS_PER_BITMAP * ctl->unit;
3313 		}
3314 
3315 		if (fatal_signal_pending(current)) {
3316 			ret = -ERESTARTSYS;
3317 			break;
3318 		}
3319 
3320 		cond_resched();
3321 	}
3322 
3323 	return ret;
3324 }
3325 
3326 void btrfs_get_block_group_trimming(struct btrfs_block_group_cache *cache)
3327 {
3328 	atomic_inc(&cache->trimming);
3329 }
3330 
3331 void btrfs_put_block_group_trimming(struct btrfs_block_group_cache *block_group)
3332 {
3333 	struct btrfs_fs_info *fs_info = block_group->fs_info;
3334 	struct extent_map_tree *em_tree;
3335 	struct extent_map *em;
3336 	bool cleanup;
3337 
3338 	spin_lock(&block_group->lock);
3339 	cleanup = (atomic_dec_and_test(&block_group->trimming) &&
3340 		   block_group->removed);
3341 	spin_unlock(&block_group->lock);
3342 
3343 	if (cleanup) {
3344 		mutex_lock(&fs_info->chunk_mutex);
3345 		em_tree = &fs_info->mapping_tree.map_tree;
3346 		write_lock(&em_tree->lock);
3347 		em = lookup_extent_mapping(em_tree, block_group->key.objectid,
3348 					   1);
3349 		BUG_ON(!em); /* logic error, can't happen */
3350 		/*
3351 		 * remove_extent_mapping() will delete us from the pinned_chunks
3352 		 * list, which is protected by the chunk mutex.
3353 		 */
3354 		remove_extent_mapping(em_tree, em);
3355 		write_unlock(&em_tree->lock);
3356 		mutex_unlock(&fs_info->chunk_mutex);
3357 
3358 		/* once for us and once for the tree */
3359 		free_extent_map(em);
3360 		free_extent_map(em);
3361 
3362 		/*
3363 		 * We've left one free space entry and other tasks trimming
3364 		 * this block group have left 1 entry each one. Free them.
3365 		 */
3366 		__btrfs_remove_free_space_cache(block_group->free_space_ctl);
3367 	}
3368 }
3369 
3370 int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group,
3371 			   u64 *trimmed, u64 start, u64 end, u64 minlen)
3372 {
3373 	int ret;
3374 
3375 	*trimmed = 0;
3376 
3377 	spin_lock(&block_group->lock);
3378 	if (block_group->removed) {
3379 		spin_unlock(&block_group->lock);
3380 		return 0;
3381 	}
3382 	btrfs_get_block_group_trimming(block_group);
3383 	spin_unlock(&block_group->lock);
3384 
3385 	ret = trim_no_bitmap(block_group, trimmed, start, end, minlen);
3386 	if (ret)
3387 		goto out;
3388 
3389 	ret = trim_bitmaps(block_group, trimmed, start, end, minlen);
3390 out:
3391 	btrfs_put_block_group_trimming(block_group);
3392 	return ret;
3393 }
3394 
3395 /*
3396  * Find the left-most item in the cache tree, and then return the
3397  * smallest inode number in the item.
3398  *
3399  * Note: the returned inode number may not be the smallest one in
3400  * the tree, if the left-most item is a bitmap.
3401  */
3402 u64 btrfs_find_ino_for_alloc(struct btrfs_root *fs_root)
3403 {
3404 	struct btrfs_free_space_ctl *ctl = fs_root->free_ino_ctl;
3405 	struct btrfs_free_space *entry = NULL;
3406 	u64 ino = 0;
3407 
3408 	spin_lock(&ctl->tree_lock);
3409 
3410 	if (RB_EMPTY_ROOT(&ctl->free_space_offset))
3411 		goto out;
3412 
3413 	entry = rb_entry(rb_first(&ctl->free_space_offset),
3414 			 struct btrfs_free_space, offset_index);
3415 
3416 	if (!entry->bitmap) {
3417 		ino = entry->offset;
3418 
3419 		unlink_free_space(ctl, entry);
3420 		entry->offset++;
3421 		entry->bytes--;
3422 		if (!entry->bytes)
3423 			kmem_cache_free(btrfs_free_space_cachep, entry);
3424 		else
3425 			link_free_space(ctl, entry);
3426 	} else {
3427 		u64 offset = 0;
3428 		u64 count = 1;
3429 		int ret;
3430 
3431 		ret = search_bitmap(ctl, entry, &offset, &count, true);
3432 		/* Logic error; Should be empty if it can't find anything */
3433 		ASSERT(!ret);
3434 
3435 		ino = offset;
3436 		bitmap_clear_bits(ctl, entry, offset, 1);
3437 		if (entry->bytes == 0)
3438 			free_bitmap(ctl, entry);
3439 	}
3440 out:
3441 	spin_unlock(&ctl->tree_lock);
3442 
3443 	return ino;
3444 }
3445 
3446 struct inode *lookup_free_ino_inode(struct btrfs_root *root,
3447 				    struct btrfs_path *path)
3448 {
3449 	struct inode *inode = NULL;
3450 
3451 	spin_lock(&root->ino_cache_lock);
3452 	if (root->ino_cache_inode)
3453 		inode = igrab(root->ino_cache_inode);
3454 	spin_unlock(&root->ino_cache_lock);
3455 	if (inode)
3456 		return inode;
3457 
3458 	inode = __lookup_free_space_inode(root, path, 0);
3459 	if (IS_ERR(inode))
3460 		return inode;
3461 
3462 	spin_lock(&root->ino_cache_lock);
3463 	if (!btrfs_fs_closing(root->fs_info))
3464 		root->ino_cache_inode = igrab(inode);
3465 	spin_unlock(&root->ino_cache_lock);
3466 
3467 	return inode;
3468 }
3469 
3470 int create_free_ino_inode(struct btrfs_root *root,
3471 			  struct btrfs_trans_handle *trans,
3472 			  struct btrfs_path *path)
3473 {
3474 	return __create_free_space_inode(root, trans, path,
3475 					 BTRFS_FREE_INO_OBJECTID, 0);
3476 }
3477 
3478 int load_free_ino_cache(struct btrfs_fs_info *fs_info, struct btrfs_root *root)
3479 {
3480 	struct btrfs_free_space_ctl *ctl = root->free_ino_ctl;
3481 	struct btrfs_path *path;
3482 	struct inode *inode;
3483 	int ret = 0;
3484 	u64 root_gen = btrfs_root_generation(&root->root_item);
3485 
3486 	if (!btrfs_test_opt(fs_info, INODE_MAP_CACHE))
3487 		return 0;
3488 
3489 	/*
3490 	 * If we're unmounting then just return, since this does a search on the
3491 	 * normal root and not the commit root and we could deadlock.
3492 	 */
3493 	if (btrfs_fs_closing(fs_info))
3494 		return 0;
3495 
3496 	path = btrfs_alloc_path();
3497 	if (!path)
3498 		return 0;
3499 
3500 	inode = lookup_free_ino_inode(root, path);
3501 	if (IS_ERR(inode))
3502 		goto out;
3503 
3504 	if (root_gen != BTRFS_I(inode)->generation)
3505 		goto out_put;
3506 
3507 	ret = __load_free_space_cache(root, inode, ctl, path, 0);
3508 
3509 	if (ret < 0)
3510 		btrfs_err(fs_info,
3511 			"failed to load free ino cache for root %llu",
3512 			root->root_key.objectid);
3513 out_put:
3514 	iput(inode);
3515 out:
3516 	btrfs_free_path(path);
3517 	return ret;
3518 }
3519 
3520 int btrfs_write_out_ino_cache(struct btrfs_root *root,
3521 			      struct btrfs_trans_handle *trans,
3522 			      struct btrfs_path *path,
3523 			      struct inode *inode)
3524 {
3525 	struct btrfs_fs_info *fs_info = root->fs_info;
3526 	struct btrfs_free_space_ctl *ctl = root->free_ino_ctl;
3527 	int ret;
3528 	struct btrfs_io_ctl io_ctl;
3529 	bool release_metadata = true;
3530 
3531 	if (!btrfs_test_opt(fs_info, INODE_MAP_CACHE))
3532 		return 0;
3533 
3534 	memset(&io_ctl, 0, sizeof(io_ctl));
3535 	ret = __btrfs_write_out_cache(root, inode, ctl, NULL, &io_ctl, trans);
3536 	if (!ret) {
3537 		/*
3538 		 * At this point writepages() didn't error out, so our metadata
3539 		 * reservation is released when the writeback finishes, at
3540 		 * inode.c:btrfs_finish_ordered_io(), regardless of it finishing
3541 		 * with or without an error.
3542 		 */
3543 		release_metadata = false;
3544 		ret = btrfs_wait_cache_io_root(root, trans, &io_ctl, path);
3545 	}
3546 
3547 	if (ret) {
3548 		if (release_metadata)
3549 			btrfs_delalloc_release_metadata(BTRFS_I(inode),
3550 					inode->i_size);
3551 #ifdef DEBUG
3552 		btrfs_err(fs_info,
3553 			  "failed to write free ino cache for root %llu",
3554 			  root->root_key.objectid);
3555 #endif
3556 	}
3557 
3558 	return ret;
3559 }
3560 
3561 #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS
3562 /*
3563  * Use this if you need to make a bitmap or extent entry specifically, it
3564  * doesn't do any of the merging that add_free_space does, this acts a lot like
3565  * how the free space cache loading stuff works, so you can get really weird
3566  * configurations.
3567  */
3568 int test_add_free_space_entry(struct btrfs_block_group_cache *cache,
3569 			      u64 offset, u64 bytes, bool bitmap)
3570 {
3571 	struct btrfs_free_space_ctl *ctl = cache->free_space_ctl;
3572 	struct btrfs_free_space *info = NULL, *bitmap_info;
3573 	void *map = NULL;
3574 	u64 bytes_added;
3575 	int ret;
3576 
3577 again:
3578 	if (!info) {
3579 		info = kmem_cache_zalloc(btrfs_free_space_cachep, GFP_NOFS);
3580 		if (!info)
3581 			return -ENOMEM;
3582 	}
3583 
3584 	if (!bitmap) {
3585 		spin_lock(&ctl->tree_lock);
3586 		info->offset = offset;
3587 		info->bytes = bytes;
3588 		info->max_extent_size = 0;
3589 		ret = link_free_space(ctl, info);
3590 		spin_unlock(&ctl->tree_lock);
3591 		if (ret)
3592 			kmem_cache_free(btrfs_free_space_cachep, info);
3593 		return ret;
3594 	}
3595 
3596 	if (!map) {
3597 		map = kzalloc(PAGE_SIZE, GFP_NOFS);
3598 		if (!map) {
3599 			kmem_cache_free(btrfs_free_space_cachep, info);
3600 			return -ENOMEM;
3601 		}
3602 	}
3603 
3604 	spin_lock(&ctl->tree_lock);
3605 	bitmap_info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset),
3606 					 1, 0);
3607 	if (!bitmap_info) {
3608 		info->bitmap = map;
3609 		map = NULL;
3610 		add_new_bitmap(ctl, info, offset);
3611 		bitmap_info = info;
3612 		info = NULL;
3613 	}
3614 
3615 	bytes_added = add_bytes_to_bitmap(ctl, bitmap_info, offset, bytes);
3616 
3617 	bytes -= bytes_added;
3618 	offset += bytes_added;
3619 	spin_unlock(&ctl->tree_lock);
3620 
3621 	if (bytes)
3622 		goto again;
3623 
3624 	if (info)
3625 		kmem_cache_free(btrfs_free_space_cachep, info);
3626 	if (map)
3627 		kfree(map);
3628 	return 0;
3629 }
3630 
3631 /*
3632  * Checks to see if the given range is in the free space cache.  This is really
3633  * just used to check the absence of space, so if there is free space in the
3634  * range at all we will return 1.
3635  */
3636 int test_check_exists(struct btrfs_block_group_cache *cache,
3637 		      u64 offset, u64 bytes)
3638 {
3639 	struct btrfs_free_space_ctl *ctl = cache->free_space_ctl;
3640 	struct btrfs_free_space *info;
3641 	int ret = 0;
3642 
3643 	spin_lock(&ctl->tree_lock);
3644 	info = tree_search_offset(ctl, offset, 0, 0);
3645 	if (!info) {
3646 		info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset),
3647 					  1, 0);
3648 		if (!info)
3649 			goto out;
3650 	}
3651 
3652 have_info:
3653 	if (info->bitmap) {
3654 		u64 bit_off, bit_bytes;
3655 		struct rb_node *n;
3656 		struct btrfs_free_space *tmp;
3657 
3658 		bit_off = offset;
3659 		bit_bytes = ctl->unit;
3660 		ret = search_bitmap(ctl, info, &bit_off, &bit_bytes, false);
3661 		if (!ret) {
3662 			if (bit_off == offset) {
3663 				ret = 1;
3664 				goto out;
3665 			} else if (bit_off > offset &&
3666 				   offset + bytes > bit_off) {
3667 				ret = 1;
3668 				goto out;
3669 			}
3670 		}
3671 
3672 		n = rb_prev(&info->offset_index);
3673 		while (n) {
3674 			tmp = rb_entry(n, struct btrfs_free_space,
3675 				       offset_index);
3676 			if (tmp->offset + tmp->bytes < offset)
3677 				break;
3678 			if (offset + bytes < tmp->offset) {
3679 				n = rb_prev(&tmp->offset_index);
3680 				continue;
3681 			}
3682 			info = tmp;
3683 			goto have_info;
3684 		}
3685 
3686 		n = rb_next(&info->offset_index);
3687 		while (n) {
3688 			tmp = rb_entry(n, struct btrfs_free_space,
3689 				       offset_index);
3690 			if (offset + bytes < tmp->offset)
3691 				break;
3692 			if (tmp->offset + tmp->bytes < offset) {
3693 				n = rb_next(&tmp->offset_index);
3694 				continue;
3695 			}
3696 			info = tmp;
3697 			goto have_info;
3698 		}
3699 
3700 		ret = 0;
3701 		goto out;
3702 	}
3703 
3704 	if (info->offset == offset) {
3705 		ret = 1;
3706 		goto out;
3707 	}
3708 
3709 	if (offset > info->offset && offset < info->offset + info->bytes)
3710 		ret = 1;
3711 out:
3712 	spin_unlock(&ctl->tree_lock);
3713 	return ret;
3714 }
3715 #endif /* CONFIG_BTRFS_FS_RUN_SANITY_TESTS */
3716