1/* SPDX-License-Identifier: GPL-2.0 */
2/*
3 * Copyright (C) 2007 Oracle. All rights reserved.
4 */
5
6#ifndef BTRFS_CTREE_H
7#define BTRFS_CTREE_H
8
9#include <linux/pagemap.h>
10#include "locking.h"
11#include "fs.h"
12#include "accessors.h"
13
14struct btrfs_trans_handle;
15struct btrfs_transaction;
16struct btrfs_pending_snapshot;
17struct btrfs_delayed_ref_root;
18struct btrfs_space_info;
19struct btrfs_block_group;
20struct btrfs_ordered_sum;
21struct btrfs_ref;
22struct btrfs_bio;
23struct btrfs_ioctl_encoded_io_args;
24struct btrfs_device;
25struct btrfs_fs_devices;
26struct btrfs_balance_control;
27struct btrfs_delayed_root;
28struct reloc_control;
29
30/* Read ahead values for struct btrfs_path.reada */
31enum {
32 READA_NONE,
33 READA_BACK,
34 READA_FORWARD,
35 /*
36 * Similar to READA_FORWARD but unlike it:
37 *
38 * 1) It will trigger readahead even for leaves that are not close to
39 * each other on disk;
40 * 2) It also triggers readahead for nodes;
41 * 3) During a search, even when a node or leaf is already in memory, it
42 * will still trigger readahead for other nodes and leaves that follow
43 * it.
44 *
45 * This is meant to be used only when we know we are iterating over the
46 * entire tree or a very large part of it.
47 */
48 READA_FORWARD_ALWAYS,
49};
50
51/*
52 * btrfs_paths remember the path taken from the root down to the leaf.
53 * level 0 is always the leaf, and nodes[1...BTRFS_MAX_LEVEL] will point
54 * to any other levels that are present.
55 *
56 * The slots array records the index of the item or block pointer
57 * used while walking the tree.
58 */
59struct btrfs_path {
60 struct extent_buffer *nodes[BTRFS_MAX_LEVEL];
61 int slots[BTRFS_MAX_LEVEL];
62 /* if there is real range locking, this locks field will change */
63 u8 locks[BTRFS_MAX_LEVEL];
64 u8 reada;
65 /* keep some upper locks as we walk down */
66 u8 lowest_level;
67
68 /*
69 * set by btrfs_split_item, tells search_slot to keep all locks
70 * and to force calls to keep space in the nodes
71 */
72 unsigned int search_for_split:1;
73 unsigned int keep_locks:1;
74 unsigned int skip_locking:1;
75 unsigned int search_commit_root:1;
76 unsigned int need_commit_sem:1;
77 unsigned int skip_release_on_error:1;
78 /*
79 * Indicate that new item (btrfs_search_slot) is extending already
80 * existing item and ins_len contains only the data size and not item
81 * header (ie. sizeof(struct btrfs_item) is not included).
82 */
83 unsigned int search_for_extension:1;
84 /* Stop search if any locks need to be taken (for read) */
85 unsigned int nowait:1;
86};
87
88/*
89 * The state of btrfs root
90 */
91enum {
92 /*
93 * btrfs_record_root_in_trans is a multi-step process, and it can race
94 * with the balancing code. But the race is very small, and only the
95 * first time the root is added to each transaction. So IN_TRANS_SETUP
96 * is used to tell us when more checks are required
97 */
98 BTRFS_ROOT_IN_TRANS_SETUP,
99
100 /*
101 * Set if tree blocks of this root can be shared by other roots.
102 * Only subvolume trees and their reloc trees have this bit set.
103 * Conflicts with TRACK_DIRTY bit.
104 *
105 * This affects two things:
106 *
107 * - How balance works
108 * For shareable roots, we need to use reloc tree and do path
109 * replacement for balance, and need various pre/post hooks for
110 * snapshot creation to handle them.
111 *
112 * While for non-shareable trees, we just simply do a tree search
113 * with COW.
114 *
115 * - How dirty roots are tracked
116 * For shareable roots, btrfs_record_root_in_trans() is needed to
117 * track them, while non-subvolume roots have TRACK_DIRTY bit, they
118 * don't need to set this manually.
119 */
120 BTRFS_ROOT_SHAREABLE,
121 BTRFS_ROOT_TRACK_DIRTY,
122 BTRFS_ROOT_IN_RADIX,
123 BTRFS_ROOT_ORPHAN_ITEM_INSERTED,
124 BTRFS_ROOT_DEFRAG_RUNNING,
125 BTRFS_ROOT_FORCE_COW,
126 BTRFS_ROOT_MULTI_LOG_TASKS,
127 BTRFS_ROOT_DIRTY,
128 BTRFS_ROOT_DELETING,
129
130 /*
131 * Reloc tree is orphan, only kept here for qgroup delayed subtree scan
132 *
133 * Set for the subvolume tree owning the reloc tree.
134 */
135 BTRFS_ROOT_DEAD_RELOC_TREE,
136 /* Mark dead root stored on device whose cleanup needs to be resumed */
137 BTRFS_ROOT_DEAD_TREE,
138 /* The root has a log tree. Used for subvolume roots and the tree root. */
139 BTRFS_ROOT_HAS_LOG_TREE,
140 /* Qgroup flushing is in progress */
141 BTRFS_ROOT_QGROUP_FLUSHING,
142 /* We started the orphan cleanup for this root. */
143 BTRFS_ROOT_ORPHAN_CLEANUP,
144 /* This root has a drop operation that was started previously. */
145 BTRFS_ROOT_UNFINISHED_DROP,
146 /* This reloc root needs to have its buffers lockdep class reset. */
147 BTRFS_ROOT_RESET_LOCKDEP_CLASS,
148};
149
150/*
151 * Record swapped tree blocks of a subvolume tree for delayed subtree trace
152 * code. For detail check comment in fs/btrfs/qgroup.c.
153 */
154struct btrfs_qgroup_swapped_blocks {
155 spinlock_t lock;
156 /* RM_EMPTY_ROOT() of above blocks[] */
157 bool swapped;
158 struct rb_root blocks[BTRFS_MAX_LEVEL];
159};
160
161/*
162 * in ram representation of the tree. extent_root is used for all allocations
163 * and for the extent tree extent_root root.
164 */
165struct btrfs_root {
166 struct rb_node rb_node;
167
168 struct extent_buffer *node;
169
170 struct extent_buffer *commit_root;
171 struct btrfs_root *log_root;
172 struct btrfs_root *reloc_root;
173
174 unsigned long state;
175 struct btrfs_root_item root_item;
176 struct btrfs_key root_key;
177 struct btrfs_fs_info *fs_info;
178 struct extent_io_tree dirty_log_pages;
179
180 struct mutex objectid_mutex;
181
182 spinlock_t accounting_lock;
183 struct btrfs_block_rsv *block_rsv;
184
185 struct mutex log_mutex;
186 wait_queue_head_t log_writer_wait;
187 wait_queue_head_t log_commit_wait[2];
188 struct list_head log_ctxs[2];
189 /* Used only for log trees of subvolumes, not for the log root tree */
190 atomic_t log_writers;
191 atomic_t log_commit[2];
192 /* Used only for log trees of subvolumes, not for the log root tree */
193 atomic_t log_batch;
194 /*
195 * Protected by the 'log_mutex' lock but can be read without holding
196 * that lock to avoid unnecessary lock contention, in which case it
197 * should be read using btrfs_get_root_log_transid() except if it's a
198 * log tree in which case it can be directly accessed. Updates to this
199 * field should always use btrfs_set_root_log_transid(), except for log
200 * trees where the field can be updated directly.
201 */
202 int log_transid;
203 /* No matter the commit succeeds or not*/
204 int log_transid_committed;
205 /*
206 * Just be updated when the commit succeeds. Use
207 * btrfs_get_root_last_log_commit() and btrfs_set_root_last_log_commit()
208 * to access this field.
209 */
210 int last_log_commit;
211 pid_t log_start_pid;
212
213 u64 last_trans;
214
215 u32 type;
216
217 u64 free_objectid;
218
219 struct btrfs_key defrag_progress;
220 struct btrfs_key defrag_max;
221
222 /* The dirty list is only used by non-shareable roots */
223 struct list_head dirty_list;
224
225 struct list_head root_list;
226
227 spinlock_t log_extents_lock[2];
228 struct list_head logged_list[2];
229
230 spinlock_t inode_lock;
231 /* red-black tree that keeps track of in-memory inodes */
232 struct rb_root inode_tree;
233
234 /*
235 * radix tree that keeps track of delayed nodes of every inode,
236 * protected by inode_lock
237 */
238 struct radix_tree_root delayed_nodes_tree;
239 /*
240 * right now this just gets used so that a root has its own devid
241 * for stat. It may be used for more later
242 */
243 dev_t anon_dev;
244
245 spinlock_t root_item_lock;
246 refcount_t refs;
247
248 struct mutex delalloc_mutex;
249 spinlock_t delalloc_lock;
250 /*
251 * all of the inodes that have delalloc bytes. It is possible for
252 * this list to be empty even when there is still dirty data=ordered
253 * extents waiting to finish IO.
254 */
255 struct list_head delalloc_inodes;
256 struct list_head delalloc_root;
257 u64 nr_delalloc_inodes;
258
259 struct mutex ordered_extent_mutex;
260 /*
261 * this is used by the balancing code to wait for all the pending
262 * ordered extents
263 */
264 spinlock_t ordered_extent_lock;
265
266 /*
267 * all of the data=ordered extents pending writeback
268 * these can span multiple transactions and basically include
269 * every dirty data page that isn't from nodatacow
270 */
271 struct list_head ordered_extents;
272 struct list_head ordered_root;
273 u64 nr_ordered_extents;
274
275 /*
276 * Not empty if this subvolume root has gone through tree block swap
277 * (relocation)
278 *
279 * Will be used by reloc_control::dirty_subvol_roots.
280 */
281 struct list_head reloc_dirty_list;
282
283 /*
284 * Number of currently running SEND ioctls to prevent
285 * manipulation with the read-only status via SUBVOL_SETFLAGS
286 */
287 int send_in_progress;
288 /*
289 * Number of currently running deduplication operations that have a
290 * destination inode belonging to this root. Protected by the lock
291 * root_item_lock.
292 */
293 int dedupe_in_progress;
294 /* For exclusion of snapshot creation and nocow writes */
295 struct btrfs_drew_lock snapshot_lock;
296
297 atomic_t snapshot_force_cow;
298
299 /* For qgroup metadata reserved space */
300 spinlock_t qgroup_meta_rsv_lock;
301 u64 qgroup_meta_rsv_pertrans;
302 u64 qgroup_meta_rsv_prealloc;
303 wait_queue_head_t qgroup_flush_wait;
304
305 /* Number of active swapfiles */
306 atomic_t nr_swapfiles;
307
308 /* Record pairs of swapped blocks for qgroup */
309 struct btrfs_qgroup_swapped_blocks swapped_blocks;
310
311 /* Used only by log trees, when logging csum items */
312 struct extent_io_tree log_csum_range;
313
314 /* Used in simple quotas, track root during relocation. */
315 u64 relocation_src_root;
316
317#ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS
318 u64 alloc_bytenr;
319#endif
320
321#ifdef CONFIG_BTRFS_DEBUG
322 struct list_head leak_list;
323#endif
324};
325
326static inline bool btrfs_root_readonly(const struct btrfs_root *root)
327{
328 /* Byte-swap the constant at compile time, root_item::flags is LE */
329 return (root->root_item.flags & cpu_to_le64(BTRFS_ROOT_SUBVOL_RDONLY)) != 0;
330}
331
332static inline bool btrfs_root_dead(const struct btrfs_root *root)
333{
334 /* Byte-swap the constant at compile time, root_item::flags is LE */
335 return (root->root_item.flags & cpu_to_le64(BTRFS_ROOT_SUBVOL_DEAD)) != 0;
336}
337
338static inline u64 btrfs_root_id(const struct btrfs_root *root)
339{
340 return root->root_key.objectid;
341}
342
343static inline int btrfs_get_root_log_transid(const struct btrfs_root *root)
344{
345 return READ_ONCE(root->log_transid);
346}
347
348static inline void btrfs_set_root_log_transid(struct btrfs_root *root, int log_transid)
349{
350 WRITE_ONCE(root->log_transid, log_transid);
351}
352
353static inline int btrfs_get_root_last_log_commit(const struct btrfs_root *root)
354{
355 return READ_ONCE(root->last_log_commit);
356}
357
358static inline void btrfs_set_root_last_log_commit(struct btrfs_root *root, int commit_id)
359{
360 WRITE_ONCE(root->last_log_commit, commit_id);
361}
362
363/*
364 * Structure that conveys information about an extent that is going to replace
365 * all the extents in a file range.
366 */
367struct btrfs_replace_extent_info {
368 u64 disk_offset;
369 u64 disk_len;
370 u64 data_offset;
371 u64 data_len;
372 u64 file_offset;
373 /* Pointer to a file extent item of type regular or prealloc. */
374 char *extent_buf;
375 /*
376 * Set to true when attempting to replace a file range with a new extent
377 * described by this structure, set to false when attempting to clone an
378 * existing extent into a file range.
379 */
380 bool is_new_extent;
381 /* Indicate if we should update the inode's mtime and ctime. */
382 bool update_times;
383 /* Meaningful only if is_new_extent is true. */
384 int qgroup_reserved;
385 /*
386 * Meaningful only if is_new_extent is true.
387 * Used to track how many extent items we have already inserted in a
388 * subvolume tree that refer to the extent described by this structure,
389 * so that we know when to create a new delayed ref or update an existing
390 * one.
391 */
392 int insertions;
393};
394
395/* Arguments for btrfs_drop_extents() */
396struct btrfs_drop_extents_args {
397 /* Input parameters */
398
399 /*
400 * If NULL, btrfs_drop_extents() will allocate and free its own path.
401 * If 'replace_extent' is true, this must not be NULL. Also the path
402 * is always released except if 'replace_extent' is true and
403 * btrfs_drop_extents() sets 'extent_inserted' to true, in which case
404 * the path is kept locked.
405 */
406 struct btrfs_path *path;
407 /* Start offset of the range to drop extents from */
408 u64 start;
409 /* End (exclusive, last byte + 1) of the range to drop extents from */
410 u64 end;
411 /* If true drop all the extent maps in the range */
412 bool drop_cache;
413 /*
414 * If true it means we want to insert a new extent after dropping all
415 * the extents in the range. If this is true, the 'extent_item_size'
416 * parameter must be set as well and the 'extent_inserted' field will
417 * be set to true by btrfs_drop_extents() if it could insert the new
418 * extent.
419 * Note: when this is set to true the path must not be NULL.
420 */
421 bool replace_extent;
422 /*
423 * Used if 'replace_extent' is true. Size of the file extent item to
424 * insert after dropping all existing extents in the range
425 */
426 u32 extent_item_size;
427
428 /* Output parameters */
429
430 /*
431 * Set to the minimum between the input parameter 'end' and the end
432 * (exclusive, last byte + 1) of the last dropped extent. This is always
433 * set even if btrfs_drop_extents() returns an error.
434 */
435 u64 drop_end;
436 /*
437 * The number of allocated bytes found in the range. This can be smaller
438 * than the range's length when there are holes in the range.
439 */
440 u64 bytes_found;
441 /*
442 * Only set if 'replace_extent' is true. Set to true if we were able
443 * to insert a replacement extent after dropping all extents in the
444 * range, otherwise set to false by btrfs_drop_extents().
445 * Also, if btrfs_drop_extents() has set this to true it means it
446 * returned with the path locked, otherwise if it has set this to
447 * false it has returned with the path released.
448 */
449 bool extent_inserted;
450};
451
452struct btrfs_file_private {
453 void *filldir_buf;
454 u64 last_index;
455 struct extent_state *llseek_cached_state;
456};
457
458static inline u32 BTRFS_LEAF_DATA_SIZE(const struct btrfs_fs_info *info)
459{
460 return info->nodesize - sizeof(struct btrfs_header);
461}
462
463static inline u32 BTRFS_MAX_ITEM_SIZE(const struct btrfs_fs_info *info)
464{
465 return BTRFS_LEAF_DATA_SIZE(info) - sizeof(struct btrfs_item);
466}
467
468static inline u32 BTRFS_NODEPTRS_PER_BLOCK(const struct btrfs_fs_info *info)
469{
470 return BTRFS_LEAF_DATA_SIZE(info) / sizeof(struct btrfs_key_ptr);
471}
472
473static inline u32 BTRFS_MAX_XATTR_SIZE(const struct btrfs_fs_info *info)
474{
475 return BTRFS_MAX_ITEM_SIZE(info) - sizeof(struct btrfs_dir_item);
476}
477
478#define BTRFS_BYTES_TO_BLKS(fs_info, bytes) \
479 ((bytes) >> (fs_info)->sectorsize_bits)
480
481static inline gfp_t btrfs_alloc_write_mask(struct address_space *mapping)
482{
483 return mapping_gfp_constraint(mapping, gfp_mask: ~__GFP_FS);
484}
485
486int btrfs_error_unpin_extent_range(struct btrfs_fs_info *fs_info,
487 u64 start, u64 end);
488int btrfs_discard_extent(struct btrfs_fs_info *fs_info, u64 bytenr,
489 u64 num_bytes, u64 *actual_bytes);
490int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range);
491
492/* ctree.c */
493int __init btrfs_ctree_init(void);
494void __cold btrfs_ctree_exit(void);
495
496int btrfs_bin_search(struct extent_buffer *eb, int first_slot,
497 const struct btrfs_key *key, int *slot);
498
499int __pure btrfs_comp_cpu_keys(const struct btrfs_key *k1, const struct btrfs_key *k2);
500
501#ifdef __LITTLE_ENDIAN
502
503/*
504 * Compare two keys, on little-endian the disk order is same as CPU order and
505 * we can avoid the conversion.
506 */
507static inline int btrfs_comp_keys(const struct btrfs_disk_key *disk_key,
508 const struct btrfs_key *k2)
509{
510 const struct btrfs_key *k1 = (const struct btrfs_key *)disk_key;
511
512 return btrfs_comp_cpu_keys(k1, k2);
513}
514
515#else
516
517/* Compare two keys in a memcmp fashion. */
518static inline int btrfs_comp_keys(const struct btrfs_disk_key *disk,
519 const struct btrfs_key *k2)
520{
521 struct btrfs_key k1;
522
523 btrfs_disk_key_to_cpu(&k1, disk);
524
525 return btrfs_comp_cpu_keys(&k1, k2);
526}
527
528#endif
529
530int btrfs_previous_item(struct btrfs_root *root,
531 struct btrfs_path *path, u64 min_objectid,
532 int type);
533int btrfs_previous_extent_item(struct btrfs_root *root,
534 struct btrfs_path *path, u64 min_objectid);
535void btrfs_set_item_key_safe(struct btrfs_trans_handle *trans,
536 struct btrfs_path *path,
537 const struct btrfs_key *new_key);
538struct extent_buffer *btrfs_root_node(struct btrfs_root *root);
539int btrfs_find_next_key(struct btrfs_root *root, struct btrfs_path *path,
540 struct btrfs_key *key, int lowest_level,
541 u64 min_trans);
542int btrfs_search_forward(struct btrfs_root *root, struct btrfs_key *min_key,
543 struct btrfs_path *path,
544 u64 min_trans);
545struct extent_buffer *btrfs_read_node_slot(struct extent_buffer *parent,
546 int slot);
547
548int btrfs_cow_block(struct btrfs_trans_handle *trans,
549 struct btrfs_root *root, struct extent_buffer *buf,
550 struct extent_buffer *parent, int parent_slot,
551 struct extent_buffer **cow_ret,
552 enum btrfs_lock_nesting nest);
553int btrfs_force_cow_block(struct btrfs_trans_handle *trans,
554 struct btrfs_root *root,
555 struct extent_buffer *buf,
556 struct extent_buffer *parent, int parent_slot,
557 struct extent_buffer **cow_ret,
558 u64 search_start, u64 empty_size,
559 enum btrfs_lock_nesting nest);
560int btrfs_copy_root(struct btrfs_trans_handle *trans,
561 struct btrfs_root *root,
562 struct extent_buffer *buf,
563 struct extent_buffer **cow_ret, u64 new_root_objectid);
564int btrfs_block_can_be_shared(struct btrfs_trans_handle *trans,
565 struct btrfs_root *root,
566 struct extent_buffer *buf);
567int btrfs_del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root,
568 struct btrfs_path *path, int level, int slot);
569void btrfs_extend_item(struct btrfs_trans_handle *trans,
570 struct btrfs_path *path, u32 data_size);
571void btrfs_truncate_item(struct btrfs_trans_handle *trans,
572 struct btrfs_path *path, u32 new_size, int from_end);
573int btrfs_split_item(struct btrfs_trans_handle *trans,
574 struct btrfs_root *root,
575 struct btrfs_path *path,
576 const struct btrfs_key *new_key,
577 unsigned long split_offset);
578int btrfs_duplicate_item(struct btrfs_trans_handle *trans,
579 struct btrfs_root *root,
580 struct btrfs_path *path,
581 const struct btrfs_key *new_key);
582int btrfs_find_item(struct btrfs_root *fs_root, struct btrfs_path *path,
583 u64 inum, u64 ioff, u8 key_type, struct btrfs_key *found_key);
584int btrfs_search_slot(struct btrfs_trans_handle *trans, struct btrfs_root *root,
585 const struct btrfs_key *key, struct btrfs_path *p,
586 int ins_len, int cow);
587int btrfs_search_old_slot(struct btrfs_root *root, const struct btrfs_key *key,
588 struct btrfs_path *p, u64 time_seq);
589int btrfs_search_slot_for_read(struct btrfs_root *root,
590 const struct btrfs_key *key,
591 struct btrfs_path *p, int find_higher,
592 int return_any);
593void btrfs_release_path(struct btrfs_path *p);
594struct btrfs_path *btrfs_alloc_path(void);
595void btrfs_free_path(struct btrfs_path *p);
596
597int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root,
598 struct btrfs_path *path, int slot, int nr);
599static inline int btrfs_del_item(struct btrfs_trans_handle *trans,
600 struct btrfs_root *root,
601 struct btrfs_path *path)
602{
603 return btrfs_del_items(trans, root, path, slot: path->slots[0], nr: 1);
604}
605
606/*
607 * Describes a batch of items to insert in a btree. This is used by
608 * btrfs_insert_empty_items().
609 */
610struct btrfs_item_batch {
611 /*
612 * Pointer to an array containing the keys of the items to insert (in
613 * sorted order).
614 */
615 const struct btrfs_key *keys;
616 /* Pointer to an array containing the data size for each item to insert. */
617 const u32 *data_sizes;
618 /*
619 * The sum of data sizes for all items. The caller can compute this while
620 * setting up the data_sizes array, so it ends up being more efficient
621 * than having btrfs_insert_empty_items() or setup_item_for_insert()
622 * doing it, as it would avoid an extra loop over a potentially large
623 * array, and in the case of setup_item_for_insert(), we would be doing
624 * it while holding a write lock on a leaf and often on upper level nodes
625 * too, unnecessarily increasing the size of a critical section.
626 */
627 u32 total_data_size;
628 /* Size of the keys and data_sizes arrays (number of items in the batch). */
629 int nr;
630};
631
632void btrfs_setup_item_for_insert(struct btrfs_trans_handle *trans,
633 struct btrfs_root *root,
634 struct btrfs_path *path,
635 const struct btrfs_key *key,
636 u32 data_size);
637int btrfs_insert_item(struct btrfs_trans_handle *trans, struct btrfs_root *root,
638 const struct btrfs_key *key, void *data, u32 data_size);
639int btrfs_insert_empty_items(struct btrfs_trans_handle *trans,
640 struct btrfs_root *root,
641 struct btrfs_path *path,
642 const struct btrfs_item_batch *batch);
643
644static inline int btrfs_insert_empty_item(struct btrfs_trans_handle *trans,
645 struct btrfs_root *root,
646 struct btrfs_path *path,
647 const struct btrfs_key *key,
648 u32 data_size)
649{
650 struct btrfs_item_batch batch;
651
652 batch.keys = key;
653 batch.data_sizes = &data_size;
654 batch.total_data_size = data_size;
655 batch.nr = 1;
656
657 return btrfs_insert_empty_items(trans, root, path, batch: &batch);
658}
659
660int btrfs_next_old_leaf(struct btrfs_root *root, struct btrfs_path *path,
661 u64 time_seq);
662
663int btrfs_search_backwards(struct btrfs_root *root, struct btrfs_key *key,
664 struct btrfs_path *path);
665
666int btrfs_get_next_valid_item(struct btrfs_root *root, struct btrfs_key *key,
667 struct btrfs_path *path);
668
669/*
670 * Search in @root for a given @key, and store the slot found in @found_key.
671 *
672 * @root: The root node of the tree.
673 * @key: The key we are looking for.
674 * @found_key: Will hold the found item.
675 * @path: Holds the current slot/leaf.
676 * @iter_ret: Contains the value returned from btrfs_search_slot or
677 * btrfs_get_next_valid_item, whichever was executed last.
678 *
679 * The @iter_ret is an output variable that will contain the return value of
680 * btrfs_search_slot, if it encountered an error, or the value returned from
681 * btrfs_get_next_valid_item otherwise. That return value can be 0, if a valid
682 * slot was found, 1 if there were no more leaves, and <0 if there was an error.
683 *
684 * It's recommended to use a separate variable for iter_ret and then use it to
685 * set the function return value so there's no confusion of the 0/1/errno
686 * values stemming from btrfs_search_slot.
687 */
688#define btrfs_for_each_slot(root, key, found_key, path, iter_ret) \
689 for (iter_ret = btrfs_search_slot(NULL, (root), (key), (path), 0, 0); \
690 (iter_ret) >= 0 && \
691 (iter_ret = btrfs_get_next_valid_item((root), (found_key), (path))) == 0; \
692 (path)->slots[0]++ \
693 )
694
695int btrfs_next_old_item(struct btrfs_root *root, struct btrfs_path *path, u64 time_seq);
696
697/*
698 * Search the tree again to find a leaf with greater keys.
699 *
700 * Returns 0 if it found something or 1 if there are no greater leaves.
701 * Returns < 0 on error.
702 */
703static inline int btrfs_next_leaf(struct btrfs_root *root, struct btrfs_path *path)
704{
705 return btrfs_next_old_leaf(root, path, time_seq: 0);
706}
707
708static inline int btrfs_next_item(struct btrfs_root *root, struct btrfs_path *p)
709{
710 return btrfs_next_old_item(root, path: p, time_seq: 0);
711}
712int btrfs_leaf_free_space(const struct extent_buffer *leaf);
713
714static inline int is_fstree(u64 rootid)
715{
716 if (rootid == BTRFS_FS_TREE_OBJECTID ||
717 ((s64)rootid >= (s64)BTRFS_FIRST_FREE_OBJECTID &&
718 !btrfs_qgroup_level(qgroupid: rootid)))
719 return 1;
720 return 0;
721}
722
723static inline bool btrfs_is_data_reloc_root(const struct btrfs_root *root)
724{
725 return root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID;
726}
727
728u16 btrfs_csum_type_size(u16 type);
729int btrfs_super_csum_size(const struct btrfs_super_block *s);
730const char *btrfs_super_csum_name(u16 csum_type);
731const char *btrfs_super_csum_driver(u16 csum_type);
732size_t __attribute_const__ btrfs_get_num_csums(void);
733
734/*
735 * We use page status Private2 to indicate there is an ordered extent with
736 * unfinished IO.
737 *
738 * Rename the Private2 accessors to Ordered, to improve readability.
739 */
740#define PageOrdered(page) PagePrivate2(page)
741#define SetPageOrdered(page) SetPagePrivate2(page)
742#define ClearPageOrdered(page) ClearPagePrivate2(page)
743#define folio_test_ordered(folio) folio_test_private_2(folio)
744#define folio_set_ordered(folio) folio_set_private_2(folio)
745#define folio_clear_ordered(folio) folio_clear_private_2(folio)
746
747#endif
748

source code of linux/fs/btrfs/ctree.h