Commit | Line | Data |
---|---|---|
9888c340 | 1 | /* SPDX-License-Identifier: GPL-2.0 */ |
6cbd5570 CM |
2 | /* |
3 | * Copyright (C) 2007 Oracle. All rights reserved. | |
6cbd5570 CM |
4 | */ |
5 | ||
9888c340 DS |
6 | #ifndef BTRFS_CTREE_H |
7 | #define BTRFS_CTREE_H | |
eb60ceac | 8 | |
810191ff | 9 | #include <linux/mm.h> |
174cd4b1 | 10 | #include <linux/sched/signal.h> |
810191ff | 11 | #include <linux/highmem.h> |
e20d96d6 | 12 | #include <linux/fs.h> |
a2de733c | 13 | #include <linux/rwsem.h> |
803b2f54 | 14 | #include <linux/semaphore.h> |
58176a96 | 15 | #include <linux/completion.h> |
04160088 | 16 | #include <linux/backing-dev.h> |
e6dcd2dc | 17 | #include <linux/wait.h> |
5a0e3ad6 | 18 | #include <linux/slab.h> |
1abe9b8a | 19 | #include <trace/events/btrfs.h> |
65019df8 | 20 | #include <asm/unaligned.h> |
3b16a4e3 | 21 | #include <linux/pagemap.h> |
55e301fd | 22 | #include <linux/btrfs.h> |
db671160 | 23 | #include <linux/btrfs_tree.h> |
21c7e756 | 24 | #include <linux/workqueue.h> |
f667aef6 | 25 | #include <linux/security.h> |
ee22184b | 26 | #include <linux/sizes.h> |
897a41b1 | 27 | #include <linux/dynamic_debug.h> |
1e4f4714 | 28 | #include <linux/refcount.h> |
9678c543 | 29 | #include <linux/crc32c.h> |
4e4cabec | 30 | #include <linux/iomap.h> |
9c7d3a54 | 31 | #include "extent-io-tree.h" |
d1310b2e | 32 | #include "extent_io.h" |
5f39d397 | 33 | #include "extent_map.h" |
8b712842 | 34 | #include "async-thread.h" |
d12ffdd1 | 35 | #include "block-rsv.h" |
2992df73 | 36 | #include "locking.h" |
c7321b76 | 37 | #include "misc.h" |
e20d96d6 | 38 | |
e089f05c | 39 | struct btrfs_trans_handle; |
79154b1b | 40 | struct btrfs_transaction; |
a22285a6 | 41 | struct btrfs_pending_snapshot; |
31890da0 | 42 | struct btrfs_delayed_ref_root; |
8719aaae | 43 | struct btrfs_space_info; |
32da5386 | 44 | struct btrfs_block_group; |
e6dcd2dc | 45 | struct btrfs_ordered_sum; |
82fa113f | 46 | struct btrfs_ref; |
c3a3b19b | 47 | struct btrfs_bio; |
1881fba8 | 48 | struct btrfs_ioctl_encoded_io_args; |
0e75f005 JB |
49 | struct btrfs_device; |
50 | struct btrfs_fs_devices; | |
51 | struct btrfs_balance_control; | |
52 | struct btrfs_delayed_root; | |
53 | struct reloc_control; | |
e089f05c | 54 | |
7c829b72 AJ |
55 | #define BTRFS_OLDEST_GENERATION 0ULL |
56 | ||
3954401f | 57 | #define BTRFS_EMPTY_DIR_SIZE 0 |
f254e52c | 58 | |
ee22184b | 59 | #define BTRFS_DIRTY_METADATA_THRESH SZ_32M |
e2d84521 | 60 | |
ee22184b | 61 | #define BTRFS_MAX_EXTENT_SIZE SZ_128M |
dcab6a3b | 62 | |
0b86a832 CM |
63 | static inline unsigned long btrfs_chunk_item_size(int num_stripes) |
64 | { | |
65 | BUG_ON(num_stripes == 0); | |
66 | return sizeof(struct btrfs_chunk) + | |
67 | sizeof(struct btrfs_stripe) * (num_stripes - 1); | |
68 | } | |
69 | ||
38732474 QW |
70 | #define BTRFS_SUPER_INFO_OFFSET SZ_64K |
71 | #define BTRFS_SUPER_INFO_SIZE 4096 | |
4300c58f | 72 | static_assert(sizeof(struct btrfs_super_block) == BTRFS_SUPER_INFO_SIZE); |
38732474 | 73 | |
37f85ec3 QW |
74 | /* |
75 | * The reserved space at the beginning of each device. | |
76 | * It covers the primary super block and leaves space for potential use by other | |
77 | * tools like bootloaders or to lower potential damage of accidental overwrite. | |
78 | */ | |
79 | #define BTRFS_DEVICE_RANGE_RESERVED (SZ_1M) | |
80 | ||
ace75066 FM |
81 | /* Read ahead values for struct btrfs_path.reada */ |
82 | enum { | |
83 | READA_NONE, | |
84 | READA_BACK, | |
85 | READA_FORWARD, | |
86 | /* | |
87 | * Similar to READA_FORWARD but unlike it: | |
88 | * | |
89 | * 1) It will trigger readahead even for leaves that are not close to | |
90 | * each other on disk; | |
91 | * 2) It also triggers readahead for nodes; | |
92 | * 3) During a search, even when a node or leaf is already in memory, it | |
93 | * will still trigger readahead for other nodes and leaves that follow | |
94 | * it. | |
95 | * | |
96 | * This is meant to be used only when we know we are iterating over the | |
97 | * entire tree or a very large part of it. | |
98 | */ | |
99 | READA_FORWARD_ALWAYS, | |
100 | }; | |
101 | ||
fec577fb | 102 | /* |
234b63a0 CM |
103 | * btrfs_paths remember the path taken from the root down to the leaf. |
104 | * level 0 is always the leaf, and nodes[1...BTRFS_MAX_LEVEL] will point | |
fec577fb CM |
105 | * to any other levels that are present. |
106 | * | |
107 | * The slots array records the index of the item or block pointer | |
108 | * used while walking the tree. | |
109 | */ | |
234b63a0 | 110 | struct btrfs_path { |
5f39d397 | 111 | struct extent_buffer *nodes[BTRFS_MAX_LEVEL]; |
234b63a0 | 112 | int slots[BTRFS_MAX_LEVEL]; |
925baedd | 113 | /* if there is real range locking, this locks field will change */ |
4fb72bf2 | 114 | u8 locks[BTRFS_MAX_LEVEL]; |
dccabfad | 115 | u8 reada; |
925baedd | 116 | /* keep some upper locks as we walk down */ |
7853f15b | 117 | u8 lowest_level; |
459931ec CM |
118 | |
119 | /* | |
120 | * set by btrfs_split_item, tells search_slot to keep all locks | |
121 | * and to force calls to keep space in the nodes | |
122 | */ | |
b9473439 CM |
123 | unsigned int search_for_split:1; |
124 | unsigned int keep_locks:1; | |
125 | unsigned int skip_locking:1; | |
5d4f98a2 | 126 | unsigned int search_commit_root:1; |
3f8a18cc | 127 | unsigned int need_commit_sem:1; |
5f5bc6b1 | 128 | unsigned int skip_release_on_error:1; |
9a664971 | 129 | /* |
130 | * Indicate that new item (btrfs_search_slot) is extending already | |
131 | * existing item and ins_len contains only the data size and not item | |
132 | * header (ie. sizeof(struct btrfs_item) is not included). | |
133 | */ | |
134 | unsigned int search_for_extension:1; | |
857bc13f JB |
135 | /* Stop search if any locks need to be taken (for read) */ |
136 | unsigned int nowait:1; | |
eb60ceac | 137 | }; |
d9d88fde | 138 | |
e922e087 SB |
139 | struct btrfs_dev_replace { |
140 | u64 replace_state; /* see #define above */ | |
a944442c AP |
141 | time64_t time_started; /* seconds since 1-Jan-1970 */ |
142 | time64_t time_stopped; /* seconds since 1-Jan-1970 */ | |
e922e087 SB |
143 | atomic64_t num_write_errors; |
144 | atomic64_t num_uncorrectable_read_errors; | |
145 | ||
146 | u64 cursor_left; | |
147 | u64 committed_cursor_left; | |
148 | u64 cursor_left_last_write_of_item; | |
149 | u64 cursor_right; | |
150 | ||
151 | u64 cont_reading_from_srcdev_mode; /* see #define above */ | |
152 | ||
153 | int is_valid; | |
154 | int item_needs_writeback; | |
155 | struct btrfs_device *srcdev; | |
156 | struct btrfs_device *tgtdev; | |
157 | ||
e922e087 | 158 | struct mutex lock_finishing_cancel_unmount; |
129827e3 | 159 | struct rw_semaphore rwsem; |
e922e087 SB |
160 | |
161 | struct btrfs_scrub_progress scrub_progress; | |
7f8d236a DS |
162 | |
163 | struct percpu_counter bio_counter; | |
164 | wait_queue_head_t replace_wait; | |
e922e087 SB |
165 | }; |
166 | ||
fa9c0d79 CM |
167 | /* |
168 | * free clusters are used to claim free space in relatively large chunks, | |
583b7231 HK |
169 | * allowing us to do less seeky writes. They are used for all metadata |
170 | * allocations. In ssd_spread mode they are also used for data allocations. | |
fa9c0d79 CM |
171 | */ |
172 | struct btrfs_free_cluster { | |
173 | spinlock_t lock; | |
174 | spinlock_t refill_lock; | |
175 | struct rb_root root; | |
176 | ||
177 | /* largest extent in this cluster */ | |
178 | u64 max_size; | |
179 | ||
180 | /* first extent starting offset */ | |
181 | u64 window_start; | |
182 | ||
c759c4e1 JB |
183 | /* We did a full search and couldn't create a cluster */ |
184 | bool fragmented; | |
185 | ||
32da5386 | 186 | struct btrfs_block_group *block_group; |
fa9c0d79 CM |
187 | /* |
188 | * when a cluster is allocated from a block group, we put the | |
189 | * cluster onto a list in the block group so that it can | |
190 | * be freed before the block group is freed. | |
191 | */ | |
192 | struct list_head block_group_list; | |
6324fbf3 CM |
193 | }; |
194 | ||
b0643e59 DZ |
195 | /* Discard control. */ |
196 | /* | |
197 | * Async discard uses multiple lists to differentiate the discard filter | |
6e80d4f8 DZ |
198 | * parameters. Index 0 is for completely free block groups where we need to |
199 | * ensure the entire block group is trimmed without being lossy. Indices | |
200 | * afterwards represent monotonically decreasing discard filter sizes to | |
201 | * prioritize what should be discarded next. | |
b0643e59 | 202 | */ |
7fe6d45e | 203 | #define BTRFS_NR_DISCARD_LISTS 3 |
6e80d4f8 DZ |
204 | #define BTRFS_DISCARD_INDEX_UNUSED 0 |
205 | #define BTRFS_DISCARD_INDEX_START 1 | |
b0643e59 DZ |
206 | |
207 | struct btrfs_discard_ctl { | |
208 | struct workqueue_struct *discard_workers; | |
209 | struct delayed_work work; | |
210 | spinlock_t lock; | |
211 | struct btrfs_block_group *block_group; | |
212 | struct list_head discard_list[BTRFS_NR_DISCARD_LISTS]; | |
e93591bb | 213 | u64 prev_discard; |
df903e5d | 214 | u64 prev_discard_time; |
dfb79ddb | 215 | atomic_t discardable_extents; |
5dc7c10b | 216 | atomic64_t discardable_bytes; |
19b2a2c7 | 217 | u64 max_discard_size; |
6e88f116 | 218 | u64 delay_ms; |
a2309300 | 219 | u32 iops_limit; |
e93591bb | 220 | u32 kbps_limit; |
9ddf648f DZ |
221 | u64 discard_extent_bytes; |
222 | u64 discard_bitmap_bytes; | |
223 | atomic64_t discard_bytes_saved; | |
b0643e59 DZ |
224 | }; |
225 | ||
c3e1f96c GR |
226 | /* |
227 | * Exclusive operations (device replace, resize, device add/remove, balance) | |
228 | */ | |
229 | enum btrfs_exclusive_operation { | |
230 | BTRFS_EXCLOP_NONE, | |
efc0e69c | 231 | BTRFS_EXCLOP_BALANCE_PAUSED, |
c3e1f96c GR |
232 | BTRFS_EXCLOP_BALANCE, |
233 | BTRFS_EXCLOP_DEV_ADD, | |
234 | BTRFS_EXCLOP_DEV_REMOVE, | |
235 | BTRFS_EXCLOP_DEV_REPLACE, | |
236 | BTRFS_EXCLOP_RESIZE, | |
237 | BTRFS_EXCLOP_SWAP_ACTIVATE, | |
238 | }; | |
239 | ||
e55958c8 IA |
240 | /* Store data about transaction commits, exported via sysfs. */ |
241 | struct btrfs_commit_stats { | |
242 | /* Total number of commits */ | |
243 | u64 commit_count; | |
244 | /* The maximum commit duration so far in ns */ | |
245 | u64 max_commit_dur; | |
246 | /* The last commit duration in ns */ | |
247 | u64 last_commit_dur; | |
248 | /* The total commit duration in ns */ | |
249 | u64 total_commit_dur; | |
250 | }; | |
251 | ||
9f5fae2f | 252 | struct btrfs_fs_info { |
e17cade2 | 253 | u8 chunk_tree_uuid[BTRFS_UUID_SIZE]; |
afcdd129 | 254 | unsigned long flags; |
62e2749e | 255 | struct btrfs_root *tree_root; |
0b86a832 CM |
256 | struct btrfs_root *chunk_root; |
257 | struct btrfs_root *dev_root; | |
3de4586c | 258 | struct btrfs_root *fs_root; |
416ac51d | 259 | struct btrfs_root *quota_root; |
f7a81ea4 | 260 | struct btrfs_root *uuid_root; |
aeb935a4 | 261 | struct btrfs_root *data_reloc_root; |
9c54e80d | 262 | struct btrfs_root *block_group_root; |
e02119d5 CM |
263 | |
264 | /* the log root tree is a directory of all the other log roots */ | |
265 | struct btrfs_root *log_root_tree; | |
4df27c4d | 266 | |
abed4aaa JB |
267 | /* The tree that holds the global roots (csum, extent, etc) */ |
268 | rwlock_t global_root_lock; | |
269 | struct rb_root global_root_tree; | |
270 | ||
fc7cbcd4 DS |
271 | spinlock_t fs_roots_radix_lock; |
272 | struct radix_tree_root fs_roots_radix; | |
1a5bc167 | 273 | |
0f9dd46c | 274 | /* block group cache stuff */ |
16b0c258 | 275 | rwlock_t block_group_cache_lock; |
08dddb29 | 276 | struct rb_root_cached block_group_cache_tree; |
0f9dd46c | 277 | |
2bf64758 | 278 | /* keep track of unallocated space */ |
a5ed45f8 | 279 | atomic64_t free_chunk_space; |
2bf64758 | 280 | |
fe119a6e NB |
281 | /* Track ranges which are used by log trees blocks/logged data extents */ |
282 | struct extent_io_tree excluded_extents; | |
1a5bc167 | 283 | |
0b86a832 | 284 | /* logical->physical extent mapping */ |
c8bf1b67 | 285 | struct extent_map_tree mapping_tree; |
0b86a832 | 286 | |
16cdcec7 MX |
287 | /* |
288 | * block reservation for extent, checksum, root tree and | |
289 | * delayed dir index item | |
290 | */ | |
f0486c68 | 291 | struct btrfs_block_rsv global_block_rsv; |
f0486c68 YZ |
292 | /* block reservation for metadata operations */ |
293 | struct btrfs_block_rsv trans_block_rsv; | |
294 | /* block reservation for chunk tree */ | |
295 | struct btrfs_block_rsv chunk_block_rsv; | |
6d668dda JB |
296 | /* block reservation for delayed operations */ |
297 | struct btrfs_block_rsv delayed_block_rsv; | |
ba2c4d4e JB |
298 | /* block reservation for delayed refs */ |
299 | struct btrfs_block_rsv delayed_refs_rsv; | |
f0486c68 YZ |
300 | |
301 | struct btrfs_block_rsv empty_block_rsv; | |
302 | ||
293ffd5f | 303 | u64 generation; |
15ee9bc7 | 304 | u64 last_trans_committed; |
d96b3424 FM |
305 | /* |
306 | * Generation of the last transaction used for block group relocation | |
307 | * since the filesystem was last mounted (or 0 if none happened yet). | |
308 | * Must be written and read while holding btrfs_fs_info::commit_root_sem. | |
309 | */ | |
310 | u64 last_reloc_trans; | |
0a2b2a84 | 311 | u64 avg_delayed_ref_runtime; |
12fcfd22 CM |
312 | |
313 | /* | |
314 | * this is updated to the current trans every time a full commit | |
315 | * is required instead of the faster short fsync log commits | |
316 | */ | |
317 | u64 last_trans_log_full_commit; | |
25cd999e | 318 | unsigned long mount_opt; |
55e5cfd3 | 319 | |
261507a0 | 320 | unsigned long compress_type:4; |
f51d2b59 | 321 | unsigned int compress_level; |
d3740608 | 322 | u32 commit_interval; |
8c6a3ee6 MX |
323 | /* |
324 | * It is a suggestive number, the read side is safe even it gets a | |
325 | * wrong number because we will write out the data into a regular | |
326 | * extent. The write side(mount/remount) is under ->s_umount lock, | |
327 | * so it is also safe. | |
328 | */ | |
6f568d35 | 329 | u64 max_inline; |
0d0c71b3 | 330 | |
79154b1b | 331 | struct btrfs_transaction *running_transaction; |
e6dcd2dc | 332 | wait_queue_head_t transaction_throttle; |
f9295749 | 333 | wait_queue_head_t transaction_wait; |
bb9c12c9 | 334 | wait_queue_head_t transaction_blocked_wait; |
771ed689 | 335 | wait_queue_head_t async_submit_wait; |
e02119d5 | 336 | |
ceda0864 MX |
337 | /* |
338 | * Used to protect the incompat_flags, compat_flags, compat_ro_flags | |
339 | * when they are updated. | |
340 | * | |
341 | * Because we do not clear the flags for ever, so we needn't use | |
342 | * the lock on the read side. | |
343 | * | |
344 | * We also needn't use the lock when we mount the fs, because | |
345 | * there is no other task which will update the flag. | |
346 | */ | |
347 | spinlock_t super_lock; | |
6c41761f DS |
348 | struct btrfs_super_block *super_copy; |
349 | struct btrfs_super_block *super_for_commit; | |
e20d96d6 | 350 | struct super_block *sb; |
d98237b3 | 351 | struct inode *btree_inode; |
e02119d5 | 352 | struct mutex tree_log_mutex; |
a74a4b97 CM |
353 | struct mutex transaction_kthread_mutex; |
354 | struct mutex cleaner_mutex; | |
925baedd | 355 | struct mutex chunk_mutex; |
53b381b3 | 356 | |
1bbc621e CM |
357 | /* |
358 | * this is taken to make sure we don't set block groups ro after | |
359 | * the free space cache has been allocated on them | |
360 | */ | |
361 | struct mutex ro_block_group_mutex; | |
362 | ||
53b381b3 DW |
363 | /* this is used during read/modify/write to make sure |
364 | * no two ios are trying to mod the same stripe at the same | |
365 | * time | |
366 | */ | |
367 | struct btrfs_stripe_hash_table *stripe_hash_table; | |
368 | ||
5a3f23d5 CM |
369 | /* |
370 | * this protects the ordered operations list only while we are | |
371 | * processing all of the entries on it. This way we make | |
372 | * sure the commit code doesn't find the list temporarily empty | |
373 | * because another function happens to be doing non-waiting preflush | |
374 | * before jumping into the main commit. | |
375 | */ | |
376 | struct mutex ordered_operations_mutex; | |
9ffba8cd | 377 | |
9e351cc8 | 378 | struct rw_semaphore commit_root_sem; |
5a3f23d5 | 379 | |
c71bf099 | 380 | struct rw_semaphore cleanup_work_sem; |
76dda93c | 381 | |
c71bf099 | 382 | struct rw_semaphore subvol_sem; |
76dda93c | 383 | |
a4abeea4 | 384 | spinlock_t trans_lock; |
7585717f CM |
385 | /* |
386 | * the reloc mutex goes with the trans lock, it is taken | |
387 | * during commit to protect us from the relocation code | |
388 | */ | |
389 | struct mutex reloc_mutex; | |
390 | ||
8fd17795 | 391 | struct list_head trans_list; |
facda1e7 | 392 | struct list_head dead_roots; |
11833d66 | 393 | struct list_head caching_block_groups; |
e02119d5 | 394 | |
24bbcf04 YZ |
395 | spinlock_t delayed_iput_lock; |
396 | struct list_head delayed_iputs; | |
034f784d JB |
397 | atomic_t nr_delayed_iputs; |
398 | wait_queue_head_t delayed_iputs_wait; | |
24bbcf04 | 399 | |
fc36ed7e | 400 | atomic64_t tree_mod_seq; |
f29021b2 | 401 | |
7227ff4d | 402 | /* this protects tree_mod_log and tree_mod_seq_list */ |
f29021b2 JS |
403 | rwlock_t tree_mod_log_lock; |
404 | struct rb_root tree_mod_log; | |
7227ff4d | 405 | struct list_head tree_mod_seq_list; |
f29021b2 | 406 | |
771ed689 | 407 | atomic_t async_delalloc_pages; |
ce9adaa5 | 408 | |
3eaa2885 | 409 | /* |
199c2a9c | 410 | * this is used to protect the following list -- ordered_roots. |
3eaa2885 | 411 | */ |
199c2a9c | 412 | spinlock_t ordered_root_lock; |
5a3f23d5 CM |
413 | |
414 | /* | |
199c2a9c MX |
415 | * all fs/file tree roots in which there are data=ordered extents |
416 | * pending writeback are added into this list. | |
417 | * | |
5a3f23d5 CM |
418 | * these can span multiple transactions and basically include |
419 | * every dirty data page that isn't from nodatacow | |
420 | */ | |
199c2a9c | 421 | struct list_head ordered_roots; |
5a3f23d5 | 422 | |
573bfb72 | 423 | struct mutex delalloc_root_mutex; |
eb73c1b7 MX |
424 | spinlock_t delalloc_root_lock; |
425 | /* all fs/file tree roots that have delalloc inodes. */ | |
426 | struct list_head delalloc_roots; | |
3eaa2885 | 427 | |
8b712842 CM |
428 | /* |
429 | * there is a pool of worker threads for checksumming during writes | |
430 | * and a pool for checksumming after reads. This is because readers | |
431 | * can run with FS locks held, and the writers may be waiting for | |
432 | * those locks. We don't want ordering in the pending list to cause | |
433 | * deadlocks, and so the two are serviced separately. | |
1cc127b5 CM |
434 | * |
435 | * A third pool does submit_bio to avoid deadlocking with the other | |
436 | * two | |
8b712842 | 437 | */ |
d458b054 | 438 | struct btrfs_workqueue *workers; |
a31b4a43 | 439 | struct btrfs_workqueue *hipri_workers; |
d458b054 QW |
440 | struct btrfs_workqueue *delalloc_workers; |
441 | struct btrfs_workqueue *flush_workers; | |
d7b9416f CH |
442 | struct workqueue_struct *endio_workers; |
443 | struct workqueue_struct *endio_meta_workers; | |
d34e123d | 444 | struct workqueue_struct *endio_raid56_workers; |
385de0ef | 445 | struct workqueue_struct *rmw_workers; |
fed8a72d | 446 | struct workqueue_struct *compressed_write_workers; |
d458b054 QW |
447 | struct btrfs_workqueue *endio_write_workers; |
448 | struct btrfs_workqueue *endio_freespace_worker; | |
d458b054 | 449 | struct btrfs_workqueue *caching_workers; |
bab39bf9 | 450 | |
247e743c CM |
451 | /* |
452 | * fixup workers take dirty pages that didn't properly go through | |
453 | * the cow mechanism and make them safe to write. It happens | |
454 | * for the sys_munmap function call path | |
455 | */ | |
d458b054 QW |
456 | struct btrfs_workqueue *fixup_workers; |
457 | struct btrfs_workqueue *delayed_workers; | |
a79b7d4b | 458 | |
a74a4b97 CM |
459 | struct task_struct *transaction_kthread; |
460 | struct task_struct *cleaner_kthread; | |
f7b885be | 461 | u32 thread_pool_size; |
8b712842 | 462 | |
6ab0a202 | 463 | struct kobject *space_info_kobj; |
49e5fb46 | 464 | struct kobject *qgroups_kobj; |
fb731430 | 465 | struct kobject *discard_kobj; |
9f5fae2f | 466 | |
e2d84521 MX |
467 | /* used to keep from writing metadata until there is a nice batch */ |
468 | struct percpu_counter dirty_metadata_bytes; | |
963d678b | 469 | struct percpu_counter delalloc_bytes; |
5deb17e1 | 470 | struct percpu_counter ordered_bytes; |
e2d84521 | 471 | s32 dirty_metadata_batch; |
963d678b MX |
472 | s32 delalloc_batch; |
473 | ||
0b86a832 CM |
474 | struct list_head dirty_cowonly_roots; |
475 | ||
8a4b83cc | 476 | struct btrfs_fs_devices *fs_devices; |
4184ea7f CM |
477 | |
478 | /* | |
dc2d3005 JM |
479 | * The space_info list is effectively read only after initial |
480 | * setup. It is populated at mount time and cleaned up after | |
481 | * all block groups are removed. RCU is used to protect it. | |
4184ea7f | 482 | */ |
6324fbf3 | 483 | struct list_head space_info; |
4184ea7f | 484 | |
b4d7c3c9 LZ |
485 | struct btrfs_space_info *data_sinfo; |
486 | ||
5d4f98a2 YZ |
487 | struct reloc_control *reloc_ctl; |
488 | ||
583b7231 | 489 | /* data_alloc_cluster is only used in ssd_spread mode */ |
fa9c0d79 CM |
490 | struct btrfs_free_cluster data_alloc_cluster; |
491 | ||
492 | /* all metadata allocations go through this cluster */ | |
493 | struct btrfs_free_cluster meta_alloc_cluster; | |
d18a2c44 | 494 | |
4cb5300b CM |
495 | /* auto defrag inodes go here */ |
496 | spinlock_t defrag_inodes_lock; | |
497 | struct rb_root defrag_inodes; | |
498 | atomic_t defrag_running; | |
499 | ||
de98ced9 MX |
500 | /* Used to protect avail_{data, metadata, system}_alloc_bits */ |
501 | seqlock_t profiles_lock; | |
a46d11a8 ID |
502 | /* |
503 | * these three are in extended format (availability of single | |
504 | * chunks is denoted by BTRFS_AVAIL_ALLOC_BIT_SINGLE bit, other | |
505 | * types are denoted by corresponding BTRFS_BLOCK_GROUP_* bits) | |
506 | */ | |
d18a2c44 CM |
507 | u64 avail_data_alloc_bits; |
508 | u64 avail_metadata_alloc_bits; | |
509 | u64 avail_system_alloc_bits; | |
788f20eb | 510 | |
c9e9f97b ID |
511 | /* restriper state */ |
512 | spinlock_t balance_lock; | |
513 | struct mutex balance_mutex; | |
837d5b6e | 514 | atomic_t balance_pause_req; |
a7e99c69 | 515 | atomic_t balance_cancel_req; |
c9e9f97b | 516 | struct btrfs_balance_control *balance_ctl; |
837d5b6e | 517 | wait_queue_head_t balance_wait_q; |
c9e9f97b | 518 | |
907d2710 DS |
519 | /* Cancellation requests for chunk relocation */ |
520 | atomic_t reloc_cancel_req; | |
521 | ||
d612ac59 AJ |
522 | u32 data_chunk_allocations; |
523 | u32 metadata_ratio; | |
97e728d4 | 524 | |
788f20eb | 525 | void *bdev_holder; |
acce952b | 526 | |
a2de733c AJ |
527 | /* private scrub information */ |
528 | struct mutex scrub_lock; | |
529 | atomic_t scrubs_running; | |
530 | atomic_t scrub_pause_req; | |
531 | atomic_t scrubs_paused; | |
532 | atomic_t scrub_cancel_req; | |
533 | wait_queue_head_t scrub_pause_wait; | |
c8352942 DS |
534 | /* |
535 | * The worker pointers are NULL iff the refcount is 0, ie. scrub is not | |
536 | * running. | |
537 | */ | |
ff09c4ca | 538 | refcount_t scrub_workers_refcnt; |
be539518 CH |
539 | struct workqueue_struct *scrub_workers; |
540 | struct workqueue_struct *scrub_wr_completion_workers; | |
541 | struct workqueue_struct *scrub_parity_workers; | |
8481dd80 | 542 | struct btrfs_subpage_info *subpage_info; |
a2de733c | 543 | |
b0643e59 DZ |
544 | struct btrfs_discard_ctl discard_ctl; |
545 | ||
21adbd5c SB |
546 | #ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY |
547 | u32 check_integrity_print_mask; | |
548 | #endif | |
416ac51d AJ |
549 | /* is qgroup tracking in a consistent state? */ |
550 | u64 qgroup_flags; | |
551 | ||
552 | /* holds configuration and tracking. Protected by qgroup_lock */ | |
553 | struct rb_root qgroup_tree; | |
554 | spinlock_t qgroup_lock; | |
555 | ||
1e8f9158 WS |
556 | /* |
557 | * used to avoid frequently calling ulist_alloc()/ulist_free() | |
558 | * when doing qgroup accounting, it must be protected by qgroup_lock. | |
559 | */ | |
560 | struct ulist *qgroup_ulist; | |
561 | ||
a855fbe6 FM |
562 | /* |
563 | * Protect user change for quota operations. If a transaction is needed, | |
564 | * it must be started before locking this lock. | |
565 | */ | |
f2f6ed3d WS |
566 | struct mutex qgroup_ioctl_lock; |
567 | ||
416ac51d AJ |
568 | /* list of dirty qgroups to be written at next commit */ |
569 | struct list_head dirty_qgroups; | |
570 | ||
e69bcee3 | 571 | /* used by qgroup for an efficient tree traversal */ |
416ac51d | 572 | u64 qgroup_seq; |
21adbd5c | 573 | |
2f232036 JS |
574 | /* qgroup rescan items */ |
575 | struct mutex qgroup_rescan_lock; /* protects the progress item */ | |
576 | struct btrfs_key qgroup_rescan_progress; | |
d458b054 | 577 | struct btrfs_workqueue *qgroup_rescan_workers; |
57254b6e | 578 | struct completion qgroup_rescan_completion; |
b382a324 | 579 | struct btrfs_work qgroup_rescan_work; |
d2c609b8 | 580 | bool qgroup_rescan_running; /* protected by qgroup_rescan_lock */ |
011b46c3 | 581 | u8 qgroup_drop_subtree_thres; |
2f232036 | 582 | |
acce952b | 583 | /* filesystem state */ |
87533c47 | 584 | unsigned long fs_state; |
16cdcec7 MX |
585 | |
586 | struct btrfs_delayed_root *delayed_root; | |
af31f5e5 | 587 | |
01cd3909 | 588 | /* Extent buffer radix tree */ |
f28491e0 | 589 | spinlock_t buffer_lock; |
478ef886 | 590 | /* Entries are eb->start / sectorsize */ |
01cd3909 | 591 | struct radix_tree_root buffer_radix; |
f28491e0 | 592 | |
af31f5e5 CM |
593 | /* next backup root to be overwritten */ |
594 | int backup_root_index; | |
5af3e8cc | 595 | |
e922e087 SB |
596 | /* device replace state */ |
597 | struct btrfs_dev_replace dev_replace; | |
5ac00add | 598 | |
803b2f54 | 599 | struct semaphore uuid_tree_rescan_sem; |
21c7e756 MX |
600 | |
601 | /* Used to reclaim the metadata space in the background. */ | |
602 | struct work_struct async_reclaim_work; | |
57056740 | 603 | struct work_struct async_data_reclaim_work; |
576fa348 | 604 | struct work_struct preempt_reclaim_work; |
47ab2a6c | 605 | |
18bb8bbf JT |
606 | /* Reclaim partially filled block groups in the background */ |
607 | struct work_struct reclaim_bgs_work; | |
608 | struct list_head reclaim_bgs; | |
609 | int bg_reclaim_threshold; | |
610 | ||
47ab2a6c JB |
611 | spinlock_t unused_bgs_lock; |
612 | struct list_head unused_bgs; | |
d4b450cd | 613 | struct mutex unused_bg_unpin_mutex; |
f3372065 JT |
614 | /* Protect block groups that are going to be deleted */ |
615 | struct mutex reclaim_bgs_lock; | |
f667aef6 | 616 | |
da17066c JM |
617 | /* Cached block sizes */ |
618 | u32 nodesize; | |
619 | u32 sectorsize; | |
ab108d99 DS |
620 | /* ilog2 of sectorsize, use to avoid 64bit division */ |
621 | u32 sectorsize_bits; | |
22b6331d | 622 | u32 csum_size; |
fe5ecbe8 | 623 | u32 csums_per_leaf; |
da17066c | 624 | u32 stripesize; |
fd708b81 | 625 | |
f7b12a62 NA |
626 | /* |
627 | * Maximum size of an extent. BTRFS_MAX_EXTENT_SIZE on regular | |
628 | * filesystem, on zoned it depends on the device constraints. | |
629 | */ | |
630 | u64 max_extent_size; | |
631 | ||
eede2bf3 OS |
632 | /* Block groups and devices containing active swapfiles. */ |
633 | spinlock_t swapfile_pins_lock; | |
634 | struct rb_root swapfile_pins; | |
635 | ||
6d97c6e3 JT |
636 | struct crypto_shash *csum_shash; |
637 | ||
0d7ed32c DS |
638 | /* Type of exclusive operation running, protected by super_lock */ |
639 | enum btrfs_exclusive_operation exclusive_operation; | |
c3e1f96c | 640 | |
b70f5097 NA |
641 | /* |
642 | * Zone size > 0 when in ZONED mode, otherwise it's used for a check | |
643 | * if the mode is enabled | |
644 | */ | |
8e010b3d | 645 | u64 zone_size; |
b70f5097 | 646 | |
c2ae7b77 NA |
647 | /* Max size to emit ZONE_APPEND write command */ |
648 | u64 max_zone_append_size; | |
0bc09ca1 | 649 | struct mutex zoned_meta_io_lock; |
40ab3be1 NA |
650 | spinlock_t treelog_bg_lock; |
651 | u64 treelog_bg; | |
862931c7 | 652 | |
c2707a25 JT |
653 | /* |
654 | * Start of the dedicated data relocation block group, protected by | |
655 | * relocation_bg_lock. | |
656 | */ | |
657 | spinlock_t relocation_bg_lock; | |
658 | u64 data_reloc_bg; | |
5f0addf7 | 659 | struct mutex zoned_data_reloc_io_lock; |
c2707a25 | 660 | |
f7238e50 JB |
661 | u64 nr_global_roots; |
662 | ||
afba2bc0 NA |
663 | spinlock_t zone_active_bgs_lock; |
664 | struct list_head zone_active_bgs; | |
665 | ||
e55958c8 IA |
666 | /* Updates are not protected by any lock */ |
667 | struct btrfs_commit_stats commit_stats; | |
668 | ||
12a824dc FM |
669 | /* |
670 | * Last generation where we dropped a non-relocation root. | |
671 | * Use btrfs_set_last_root_drop_gen() and btrfs_get_last_root_drop_gen() | |
672 | * to change it and to read it, respectively. | |
673 | */ | |
674 | u64 last_root_drop_gen; | |
675 | ||
e1489b4f IA |
676 | /* |
677 | * Annotations for transaction events (structures are empty when | |
678 | * compiled without lockdep). | |
679 | */ | |
680 | struct lockdep_map btrfs_trans_num_writers_map; | |
5a9ba670 | 681 | struct lockdep_map btrfs_trans_num_extwriters_map; |
3e738c53 | 682 | struct lockdep_map btrfs_state_change_map[4]; |
8b53779e | 683 | struct lockdep_map btrfs_trans_pending_ordered_map; |
5f4403e1 | 684 | struct lockdep_map btrfs_ordered_extent_map; |
e1489b4f | 685 | |
fd708b81 JB |
686 | #ifdef CONFIG_BTRFS_FS_REF_VERIFY |
687 | spinlock_t ref_verify_lock; | |
688 | struct rb_root block_tree; | |
689 | #endif | |
93945cb4 DZ |
690 | |
691 | #ifdef CONFIG_BTRFS_DEBUG | |
692 | struct kobject *debug_kobj; | |
bd647ce3 | 693 | struct list_head allocated_roots; |
3fd63727 JB |
694 | |
695 | spinlock_t eb_leak_lock; | |
696 | struct list_head allocated_ebs; | |
93945cb4 | 697 | #endif |
324ae4df | 698 | }; |
0b86a832 | 699 | |
12a824dc FM |
700 | static inline void btrfs_set_last_root_drop_gen(struct btrfs_fs_info *fs_info, |
701 | u64 gen) | |
702 | { | |
703 | WRITE_ONCE(fs_info->last_root_drop_gen, gen); | |
704 | } | |
705 | ||
706 | static inline u64 btrfs_get_last_root_drop_gen(const struct btrfs_fs_info *fs_info) | |
707 | { | |
708 | return READ_ONCE(fs_info->last_root_drop_gen); | |
709 | } | |
710 | ||
da17066c JM |
711 | static inline struct btrfs_fs_info *btrfs_sb(struct super_block *sb) |
712 | { | |
713 | return sb->s_fs_info; | |
714 | } | |
715 | ||
d9d88fde JB |
716 | /* |
717 | * Take the number of bytes to be checksummed and figure out how many leaves | |
718 | * it would require to store the csums for that many bytes. | |
719 | */ | |
720 | static inline u64 btrfs_csum_bytes_to_leaves( | |
721 | const struct btrfs_fs_info *fs_info, u64 csum_bytes) | |
722 | { | |
723 | const u64 num_csums = csum_bytes >> fs_info->sectorsize_bits; | |
724 | ||
725 | return DIV_ROUND_UP_ULL(num_csums, fs_info->csums_per_leaf); | |
726 | } | |
727 | ||
728 | /* | |
729 | * Use this if we would be adding new items, as we could split nodes as we cow | |
730 | * down the tree. | |
731 | */ | |
732 | static inline u64 btrfs_calc_insert_metadata_size(struct btrfs_fs_info *fs_info, | |
733 | unsigned num_items) | |
734 | { | |
735 | return (u64)fs_info->nodesize * BTRFS_MAX_LEVEL * 2 * num_items; | |
736 | } | |
737 | ||
738 | /* | |
739 | * Doing a truncate or a modification won't result in new nodes or leaves, just | |
740 | * what we need for COW. | |
741 | */ | |
742 | static inline u64 btrfs_calc_metadata_size(struct btrfs_fs_info *fs_info, | |
743 | unsigned num_items) | |
744 | { | |
745 | return (u64)fs_info->nodesize * BTRFS_MAX_LEVEL * num_items; | |
746 | } | |
747 | ||
748 | #define BTRFS_MAX_EXTENT_ITEM_SIZE(r) ((BTRFS_LEAF_DATA_SIZE(r->fs_info) >> 4) - \ | |
749 | sizeof(struct btrfs_item)) | |
750 | ||
751 | static inline bool btrfs_is_zoned(const struct btrfs_fs_info *fs_info) | |
752 | { | |
753 | return fs_info->zone_size > 0; | |
754 | } | |
755 | ||
756 | /* | |
757 | * Count how many fs_info->max_extent_size cover the @size | |
758 | */ | |
759 | static inline u32 count_max_extents(struct btrfs_fs_info *fs_info, u64 size) | |
760 | { | |
761 | #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS | |
762 | if (!fs_info) | |
763 | return div_u64(size + BTRFS_MAX_EXTENT_SIZE - 1, BTRFS_MAX_EXTENT_SIZE); | |
764 | #endif | |
765 | ||
766 | return div_u64(size + fs_info->max_extent_size - 1, fs_info->max_extent_size); | |
767 | } | |
768 | ||
769 | bool btrfs_exclop_start(struct btrfs_fs_info *fs_info, | |
770 | enum btrfs_exclusive_operation type); | |
771 | bool btrfs_exclop_start_try_lock(struct btrfs_fs_info *fs_info, | |
772 | enum btrfs_exclusive_operation type); | |
773 | void btrfs_exclop_start_unlock(struct btrfs_fs_info *fs_info); | |
774 | void btrfs_exclop_finish(struct btrfs_fs_info *fs_info); | |
775 | void btrfs_exclop_balance(struct btrfs_fs_info *fs_info, | |
776 | enum btrfs_exclusive_operation op); | |
777 | ||
27cdeb70 MX |
778 | /* |
779 | * The state of btrfs root | |
780 | */ | |
61fa90c1 DS |
781 | enum { |
782 | /* | |
783 | * btrfs_record_root_in_trans is a multi-step process, and it can race | |
784 | * with the balancing code. But the race is very small, and only the | |
785 | * first time the root is added to each transaction. So IN_TRANS_SETUP | |
786 | * is used to tell us when more checks are required | |
787 | */ | |
788 | BTRFS_ROOT_IN_TRANS_SETUP, | |
92a7cc42 QW |
789 | |
790 | /* | |
791 | * Set if tree blocks of this root can be shared by other roots. | |
792 | * Only subvolume trees and their reloc trees have this bit set. | |
793 | * Conflicts with TRACK_DIRTY bit. | |
794 | * | |
795 | * This affects two things: | |
796 | * | |
797 | * - How balance works | |
798 | * For shareable roots, we need to use reloc tree and do path | |
799 | * replacement for balance, and need various pre/post hooks for | |
800 | * snapshot creation to handle them. | |
801 | * | |
802 | * While for non-shareable trees, we just simply do a tree search | |
803 | * with COW. | |
804 | * | |
805 | * - How dirty roots are tracked | |
806 | * For shareable roots, btrfs_record_root_in_trans() is needed to | |
807 | * track them, while non-subvolume roots have TRACK_DIRTY bit, they | |
808 | * don't need to set this manually. | |
809 | */ | |
810 | BTRFS_ROOT_SHAREABLE, | |
61fa90c1 | 811 | BTRFS_ROOT_TRACK_DIRTY, |
fc7cbcd4 | 812 | BTRFS_ROOT_IN_RADIX, |
61fa90c1 DS |
813 | BTRFS_ROOT_ORPHAN_ITEM_INSERTED, |
814 | BTRFS_ROOT_DEFRAG_RUNNING, | |
815 | BTRFS_ROOT_FORCE_COW, | |
816 | BTRFS_ROOT_MULTI_LOG_TASKS, | |
817 | BTRFS_ROOT_DIRTY, | |
83354f07 | 818 | BTRFS_ROOT_DELETING, |
d2311e69 QW |
819 | |
820 | /* | |
821 | * Reloc tree is orphan, only kept here for qgroup delayed subtree scan | |
822 | * | |
823 | * Set for the subvolume tree owning the reloc tree. | |
824 | */ | |
825 | BTRFS_ROOT_DEAD_RELOC_TREE, | |
78c52d9e JB |
826 | /* Mark dead root stored on device whose cleanup needs to be resumed */ |
827 | BTRFS_ROOT_DEAD_TREE, | |
47876f7c | 828 | /* The root has a log tree. Used for subvolume roots and the tree root. */ |
e7a79811 | 829 | BTRFS_ROOT_HAS_LOG_TREE, |
c53e9653 QW |
830 | /* Qgroup flushing is in progress */ |
831 | BTRFS_ROOT_QGROUP_FLUSHING, | |
54230013 JB |
832 | /* We started the orphan cleanup for this root. */ |
833 | BTRFS_ROOT_ORPHAN_CLEANUP, | |
b4be6aef JB |
834 | /* This root has a drop operation that was started previously. */ |
835 | BTRFS_ROOT_UNFINISHED_DROP, | |
b40130b2 JB |
836 | /* This reloc root needs to have its buffers lockdep class reset. */ |
837 | BTRFS_ROOT_RESET_LOCKDEP_CLASS, | |
61fa90c1 | 838 | }; |
27cdeb70 | 839 | |
3e738c53 IA |
840 | enum btrfs_lockdep_trans_states { |
841 | BTRFS_LOCKDEP_TRANS_COMMIT_START, | |
842 | BTRFS_LOCKDEP_TRANS_UNBLOCKED, | |
843 | BTRFS_LOCKDEP_TRANS_SUPER_COMMITTED, | |
844 | BTRFS_LOCKDEP_TRANS_COMPLETED, | |
845 | }; | |
846 | ||
ab9a323f IA |
847 | /* |
848 | * Lockdep annotation for wait events. | |
849 | * | |
850 | * @owner: The struct where the lockdep map is defined | |
851 | * @lock: The lockdep map corresponding to a wait event | |
852 | * | |
853 | * This macro is used to annotate a wait event. In this case a thread acquires | |
854 | * the lockdep map as writer (exclusive lock) because it has to block until all | |
855 | * the threads that hold the lock as readers signal the condition for the wait | |
856 | * event and release their locks. | |
857 | */ | |
858 | #define btrfs_might_wait_for_event(owner, lock) \ | |
859 | do { \ | |
860 | rwsem_acquire(&owner->lock##_map, 0, 0, _THIS_IP_); \ | |
861 | rwsem_release(&owner->lock##_map, _THIS_IP_); \ | |
862 | } while (0) | |
863 | ||
864 | /* | |
865 | * Protection for the resource/condition of a wait event. | |
866 | * | |
867 | * @owner: The struct where the lockdep map is defined | |
868 | * @lock: The lockdep map corresponding to a wait event | |
869 | * | |
870 | * Many threads can modify the condition for the wait event at the same time | |
871 | * and signal the threads that block on the wait event. The threads that modify | |
872 | * the condition and do the signaling acquire the lock as readers (shared | |
873 | * lock). | |
874 | */ | |
875 | #define btrfs_lockdep_acquire(owner, lock) \ | |
876 | rwsem_acquire_read(&owner->lock##_map, 0, 0, _THIS_IP_) | |
877 | ||
878 | /* | |
879 | * Used after signaling the condition for a wait event to release the lockdep | |
880 | * map held by a reader thread. | |
881 | */ | |
882 | #define btrfs_lockdep_release(owner, lock) \ | |
883 | rwsem_release(&owner->lock##_map, _THIS_IP_) | |
884 | ||
3e738c53 IA |
885 | /* |
886 | * Macros for the transaction states wait events, similar to the generic wait | |
887 | * event macros. | |
888 | */ | |
889 | #define btrfs_might_wait_for_state(owner, i) \ | |
890 | do { \ | |
891 | rwsem_acquire(&owner->btrfs_state_change_map[i], 0, 0, _THIS_IP_); \ | |
892 | rwsem_release(&owner->btrfs_state_change_map[i], _THIS_IP_); \ | |
893 | } while (0) | |
894 | ||
895 | #define btrfs_trans_state_lockdep_acquire(owner, i) \ | |
896 | rwsem_acquire_read(&owner->btrfs_state_change_map[i], 0, 0, _THIS_IP_) | |
897 | ||
898 | #define btrfs_trans_state_lockdep_release(owner, i) \ | |
899 | rwsem_release(&owner->btrfs_state_change_map[i], _THIS_IP_) | |
900 | ||
ab9a323f IA |
901 | /* Initialization of the lockdep map */ |
902 | #define btrfs_lockdep_init_map(owner, lock) \ | |
903 | do { \ | |
904 | static struct lock_class_key lock##_key; \ | |
905 | lockdep_init_map(&owner->lock##_map, #lock, &lock##_key, 0); \ | |
906 | } while (0) | |
907 | ||
3e738c53 IA |
908 | /* Initialization of the transaction states lockdep maps. */ |
909 | #define btrfs_state_lockdep_init_map(owner, lock, state) \ | |
910 | do { \ | |
911 | static struct lock_class_key lock##_key; \ | |
912 | lockdep_init_map(&owner->btrfs_state_change_map[state], #lock, \ | |
913 | &lock##_key, 0); \ | |
914 | } while (0) | |
915 | ||
370a11b8 QW |
916 | /* |
917 | * Record swapped tree blocks of a subvolume tree for delayed subtree trace | |
918 | * code. For detail check comment in fs/btrfs/qgroup.c. | |
919 | */ | |
920 | struct btrfs_qgroup_swapped_blocks { | |
921 | spinlock_t lock; | |
922 | /* RM_EMPTY_ROOT() of above blocks[] */ | |
923 | bool swapped; | |
924 | struct rb_root blocks[BTRFS_MAX_LEVEL]; | |
925 | }; | |
926 | ||
9f5fae2f CM |
927 | /* |
928 | * in ram representation of the tree. extent_root is used for all allocations | |
f2458e1d | 929 | * and for the extent tree extent_root root. |
9f5fae2f CM |
930 | */ |
931 | struct btrfs_root { | |
abed4aaa JB |
932 | struct rb_node rb_node; |
933 | ||
5f39d397 | 934 | struct extent_buffer *node; |
925baedd | 935 | |
5f39d397 | 936 | struct extent_buffer *commit_root; |
e02119d5 | 937 | struct btrfs_root *log_root; |
1a40e23b | 938 | struct btrfs_root *reloc_root; |
31153d81 | 939 | |
27cdeb70 | 940 | unsigned long state; |
62e2749e CM |
941 | struct btrfs_root_item root_item; |
942 | struct btrfs_key root_key; | |
9f5fae2f | 943 | struct btrfs_fs_info *fs_info; |
d0c803c4 CM |
944 | struct extent_io_tree dirty_log_pages; |
945 | ||
a2135011 | 946 | struct mutex objectid_mutex; |
7237f183 | 947 | |
f0486c68 YZ |
948 | spinlock_t accounting_lock; |
949 | struct btrfs_block_rsv *block_rsv; | |
950 | ||
e02119d5 | 951 | struct mutex log_mutex; |
7237f183 YZ |
952 | wait_queue_head_t log_writer_wait; |
953 | wait_queue_head_t log_commit_wait[2]; | |
8b050d35 | 954 | struct list_head log_ctxs[2]; |
a93e0168 | 955 | /* Used only for log trees of subvolumes, not for the log root tree */ |
7237f183 YZ |
956 | atomic_t log_writers; |
957 | atomic_t log_commit[2]; | |
28a95795 | 958 | /* Used only for log trees of subvolumes, not for the log root tree */ |
2ecb7923 | 959 | atomic_t log_batch; |
bb14a59b | 960 | int log_transid; |
d1433deb MX |
961 | /* No matter the commit succeeds or not*/ |
962 | int log_transid_committed; | |
963 | /* Just be updated when the commit succeeds. */ | |
bb14a59b | 964 | int last_log_commit; |
ff782e0a | 965 | pid_t log_start_pid; |
ea8c2819 | 966 | |
0f7d52f4 | 967 | u64 last_trans; |
5f39d397 | 968 | |
9f5fae2f | 969 | u32 type; |
13a8a7c8 | 970 | |
6b8fad57 | 971 | u64 free_objectid; |
7585717f | 972 | |
6702ed49 | 973 | struct btrfs_key defrag_progress; |
0ef3e66b | 974 | struct btrfs_key defrag_max; |
0b86a832 | 975 | |
92a7cc42 | 976 | /* The dirty list is only used by non-shareable roots */ |
0b86a832 | 977 | struct list_head dirty_list; |
7b128766 | 978 | |
5d4f98a2 YZ |
979 | struct list_head root_list; |
980 | ||
2ab28f32 JB |
981 | spinlock_t log_extents_lock[2]; |
982 | struct list_head logged_list[2]; | |
983 | ||
5d4f98a2 YZ |
984 | spinlock_t inode_lock; |
985 | /* red-black tree that keeps track of in-memory inodes */ | |
986 | struct rb_root inode_tree; | |
987 | ||
16cdcec7 | 988 | /* |
088aea3b DS |
989 | * radix tree that keeps track of delayed nodes of every inode, |
990 | * protected by inode_lock | |
16cdcec7 | 991 | */ |
088aea3b | 992 | struct radix_tree_root delayed_nodes_tree; |
3394e160 CM |
993 | /* |
994 | * right now this just gets used so that a root has its own devid | |
995 | * for stat. It may be used for more later | |
996 | */ | |
0ee5dc67 | 997 | dev_t anon_dev; |
f1ebcc74 | 998 | |
5f3ab90a | 999 | spinlock_t root_item_lock; |
0700cea7 | 1000 | refcount_t refs; |
eb73c1b7 | 1001 | |
573bfb72 | 1002 | struct mutex delalloc_mutex; |
eb73c1b7 MX |
1003 | spinlock_t delalloc_lock; |
1004 | /* | |
1005 | * all of the inodes that have delalloc bytes. It is possible for | |
1006 | * this list to be empty even when there is still dirty data=ordered | |
1007 | * extents waiting to finish IO. | |
1008 | */ | |
1009 | struct list_head delalloc_inodes; | |
1010 | struct list_head delalloc_root; | |
1011 | u64 nr_delalloc_inodes; | |
31f3d255 MX |
1012 | |
1013 | struct mutex ordered_extent_mutex; | |
199c2a9c MX |
1014 | /* |
1015 | * this is used by the balancing code to wait for all the pending | |
1016 | * ordered extents | |
1017 | */ | |
1018 | spinlock_t ordered_extent_lock; | |
1019 | ||
1020 | /* | |
1021 | * all of the data=ordered extents pending writeback | |
1022 | * these can span multiple transactions and basically include | |
1023 | * every dirty data page that isn't from nodatacow | |
1024 | */ | |
1025 | struct list_head ordered_extents; | |
1026 | struct list_head ordered_root; | |
1027 | u64 nr_ordered_extents; | |
2c686537 | 1028 | |
d2311e69 QW |
1029 | /* |
1030 | * Not empty if this subvolume root has gone through tree block swap | |
1031 | * (relocation) | |
1032 | * | |
1033 | * Will be used by reloc_control::dirty_subvol_roots. | |
1034 | */ | |
1035 | struct list_head reloc_dirty_list; | |
1036 | ||
2c686537 DS |
1037 | /* |
1038 | * Number of currently running SEND ioctls to prevent | |
1039 | * manipulation with the read-only status via SUBVOL_SETFLAGS | |
1040 | */ | |
1041 | int send_in_progress; | |
62d54f3a FM |
1042 | /* |
1043 | * Number of currently running deduplication operations that have a | |
1044 | * destination inode belonging to this root. Protected by the lock | |
1045 | * root_item_lock. | |
1046 | */ | |
1047 | int dedupe_in_progress; | |
dcc3eb96 NB |
1048 | /* For exclusion of snapshot creation and nocow writes */ |
1049 | struct btrfs_drew_lock snapshot_lock; | |
1050 | ||
8ecebf4d | 1051 | atomic_t snapshot_force_cow; |
8287475a QW |
1052 | |
1053 | /* For qgroup metadata reserved space */ | |
1054 | spinlock_t qgroup_meta_rsv_lock; | |
1055 | u64 qgroup_meta_rsv_pertrans; | |
1056 | u64 qgroup_meta_rsv_prealloc; | |
c53e9653 | 1057 | wait_queue_head_t qgroup_flush_wait; |
57ec5fb4 | 1058 | |
eede2bf3 OS |
1059 | /* Number of active swapfiles */ |
1060 | atomic_t nr_swapfiles; | |
1061 | ||
370a11b8 QW |
1062 | /* Record pairs of swapped blocks for qgroup */ |
1063 | struct btrfs_qgroup_swapped_blocks swapped_blocks; | |
1064 | ||
e289f03e FM |
1065 | /* Used only by log trees, when logging csum items */ |
1066 | struct extent_io_tree log_csum_range; | |
1067 | ||
57ec5fb4 DS |
1068 | #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS |
1069 | u64 alloc_bytenr; | |
1070 | #endif | |
bd647ce3 JB |
1071 | |
1072 | #ifdef CONFIG_BTRFS_DEBUG | |
1073 | struct list_head leak_list; | |
1074 | #endif | |
62e2749e | 1075 | }; |
118c701e | 1076 | |
bf385648 FM |
1077 | /* |
1078 | * Structure that conveys information about an extent that is going to replace | |
1079 | * all the extents in a file range. | |
1080 | */ | |
1081 | struct btrfs_replace_extent_info { | |
690a5dbf FM |
1082 | u64 disk_offset; |
1083 | u64 disk_len; | |
1084 | u64 data_offset; | |
1085 | u64 data_len; | |
1086 | u64 file_offset; | |
fb870f6c | 1087 | /* Pointer to a file extent item of type regular or prealloc. */ |
690a5dbf | 1088 | char *extent_buf; |
8fccebfa FM |
1089 | /* |
1090 | * Set to true when attempting to replace a file range with a new extent | |
1091 | * described by this structure, set to false when attempting to clone an | |
1092 | * existing extent into a file range. | |
1093 | */ | |
1094 | bool is_new_extent; | |
983d8209 FM |
1095 | /* Indicate if we should update the inode's mtime and ctime. */ |
1096 | bool update_times; | |
8fccebfa FM |
1097 | /* Meaningful only if is_new_extent is true. */ |
1098 | int qgroup_reserved; | |
1099 | /* | |
1100 | * Meaningful only if is_new_extent is true. | |
1101 | * Used to track how many extent items we have already inserted in a | |
1102 | * subvolume tree that refer to the extent described by this structure, | |
1103 | * so that we know when to create a new delayed ref or update an existing | |
1104 | * one. | |
1105 | */ | |
1106 | int insertions; | |
690a5dbf FM |
1107 | }; |
1108 | ||
5893dfb9 FM |
1109 | /* Arguments for btrfs_drop_extents() */ |
1110 | struct btrfs_drop_extents_args { | |
1111 | /* Input parameters */ | |
1112 | ||
1113 | /* | |
1114 | * If NULL, btrfs_drop_extents() will allocate and free its own path. | |
1115 | * If 'replace_extent' is true, this must not be NULL. Also the path | |
1116 | * is always released except if 'replace_extent' is true and | |
1117 | * btrfs_drop_extents() sets 'extent_inserted' to true, in which case | |
1118 | * the path is kept locked. | |
1119 | */ | |
1120 | struct btrfs_path *path; | |
1121 | /* Start offset of the range to drop extents from */ | |
1122 | u64 start; | |
1123 | /* End (exclusive, last byte + 1) of the range to drop extents from */ | |
1124 | u64 end; | |
1125 | /* If true drop all the extent maps in the range */ | |
1126 | bool drop_cache; | |
1127 | /* | |
1128 | * If true it means we want to insert a new extent after dropping all | |
1129 | * the extents in the range. If this is true, the 'extent_item_size' | |
1130 | * parameter must be set as well and the 'extent_inserted' field will | |
1131 | * be set to true by btrfs_drop_extents() if it could insert the new | |
1132 | * extent. | |
1133 | * Note: when this is set to true the path must not be NULL. | |
1134 | */ | |
1135 | bool replace_extent; | |
1136 | /* | |
1137 | * Used if 'replace_extent' is true. Size of the file extent item to | |
1138 | * insert after dropping all existing extents in the range | |
1139 | */ | |
1140 | u32 extent_item_size; | |
1141 | ||
1142 | /* Output parameters */ | |
1143 | ||
1144 | /* | |
1145 | * Set to the minimum between the input parameter 'end' and the end | |
1146 | * (exclusive, last byte + 1) of the last dropped extent. This is always | |
1147 | * set even if btrfs_drop_extents() returns an error. | |
1148 | */ | |
1149 | u64 drop_end; | |
2766ff61 FM |
1150 | /* |
1151 | * The number of allocated bytes found in the range. This can be smaller | |
1152 | * than the range's length when there are holes in the range. | |
1153 | */ | |
1154 | u64 bytes_found; | |
5893dfb9 FM |
1155 | /* |
1156 | * Only set if 'replace_extent' is true. Set to true if we were able | |
1157 | * to insert a replacement extent after dropping all extents in the | |
1158 | * range, otherwise set to false by btrfs_drop_extents(). | |
1159 | * Also, if btrfs_drop_extents() has set this to true it means it | |
1160 | * returned with the path locked, otherwise if it has set this to | |
1161 | * false it has returned with the path released. | |
1162 | */ | |
1163 | bool extent_inserted; | |
1164 | }; | |
1165 | ||
23b5ec74 | 1166 | struct btrfs_file_private { |
23b5ec74 JB |
1167 | void *filldir_buf; |
1168 | }; | |
1169 | ||
62e2749e | 1170 | |
da17066c | 1171 | static inline u32 BTRFS_LEAF_DATA_SIZE(const struct btrfs_fs_info *info) |
1db1ff92 | 1172 | { |
118c701e NB |
1173 | |
1174 | return info->nodesize - sizeof(struct btrfs_header); | |
1db1ff92 JM |
1175 | } |
1176 | ||
da17066c | 1177 | static inline u32 BTRFS_MAX_ITEM_SIZE(const struct btrfs_fs_info *info) |
1db1ff92 | 1178 | { |
da17066c | 1179 | return BTRFS_LEAF_DATA_SIZE(info) - sizeof(struct btrfs_item); |
1db1ff92 JM |
1180 | } |
1181 | ||
da17066c | 1182 | static inline u32 BTRFS_NODEPTRS_PER_BLOCK(const struct btrfs_fs_info *info) |
1db1ff92 | 1183 | { |
da17066c | 1184 | return BTRFS_LEAF_DATA_SIZE(info) / sizeof(struct btrfs_key_ptr); |
1db1ff92 JM |
1185 | } |
1186 | ||
1187 | #define BTRFS_FILE_EXTENT_INLINE_DATA_START \ | |
1188 | (offsetof(struct btrfs_file_extent_item, disk_bytenr)) | |
da17066c | 1189 | static inline u32 BTRFS_MAX_INLINE_DATA_SIZE(const struct btrfs_fs_info *info) |
1db1ff92 | 1190 | { |
da17066c | 1191 | return BTRFS_MAX_ITEM_SIZE(info) - |
1db1ff92 JM |
1192 | BTRFS_FILE_EXTENT_INLINE_DATA_START; |
1193 | } | |
1194 | ||
da17066c | 1195 | static inline u32 BTRFS_MAX_XATTR_SIZE(const struct btrfs_fs_info *info) |
1db1ff92 | 1196 | { |
da17066c | 1197 | return BTRFS_MAX_ITEM_SIZE(info) - sizeof(struct btrfs_dir_item); |
1db1ff92 JM |
1198 | } |
1199 | ||
2e78c927 | 1200 | #define BTRFS_BYTES_TO_BLKS(fs_info, bytes) \ |
265fdfa6 | 1201 | ((bytes) >> (fs_info)->sectorsize_bits) |
2e78c927 | 1202 | |
65019df8 JT |
1203 | static inline u32 btrfs_crc32c(u32 crc, const void *address, unsigned length) |
1204 | { | |
1205 | return crc32c(crc, address, length); | |
1206 | } | |
1207 | ||
1208 | static inline void btrfs_crc32c_final(u32 crc, u8 *result) | |
1209 | { | |
1210 | put_unaligned_le32(~crc, result); | |
1211 | } | |
1212 | ||
9678c543 NB |
1213 | static inline u64 btrfs_name_hash(const char *name, int len) |
1214 | { | |
1215 | return crc32c((u32)~1, name, len); | |
1216 | } | |
1217 | ||
1218 | /* | |
1219 | * Figure the key offset of an extended inode ref | |
1220 | */ | |
1221 | static inline u64 btrfs_extref_hash(u64 parent_objectid, const char *name, | |
1222 | int len) | |
1223 | { | |
1224 | return (u64) crc32c(parent_objectid, name, len); | |
1225 | } | |
1226 | ||
3b16a4e3 JB |
1227 | static inline gfp_t btrfs_alloc_write_mask(struct address_space *mapping) |
1228 | { | |
c62d2555 | 1229 | return mapping_gfp_constraint(mapping, ~__GFP_FS); |
3b16a4e3 JB |
1230 | } |
1231 | ||
b18c6685 | 1232 | /* extent-tree.c */ |
28f75a0e | 1233 | |
167ce953 | 1234 | enum btrfs_inline_ref_type { |
bbe339cc DS |
1235 | BTRFS_REF_TYPE_INVALID, |
1236 | BTRFS_REF_TYPE_BLOCK, | |
1237 | BTRFS_REF_TYPE_DATA, | |
1238 | BTRFS_REF_TYPE_ANY, | |
167ce953 LB |
1239 | }; |
1240 | ||
1241 | int btrfs_get_extent_inline_ref_type(const struct extent_buffer *eb, | |
1242 | struct btrfs_extent_inline_ref *iref, | |
1243 | enum btrfs_inline_ref_type is_data); | |
0785a9aa | 1244 | u64 hash_extent_data_ref(u64 root_objectid, u64 owner, u64 offset); |
167ce953 | 1245 | |
16cdcec7 | 1246 | |
6f410d1b JB |
1247 | int btrfs_add_excluded_extent(struct btrfs_fs_info *fs_info, |
1248 | u64 start, u64 num_bytes); | |
32da5386 | 1249 | void btrfs_free_excluded_extents(struct btrfs_block_group *cache); |
56bec294 | 1250 | int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans, |
c79a70b1 | 1251 | unsigned long count); |
31890da0 JB |
1252 | void btrfs_cleanup_ref_head_accounting(struct btrfs_fs_info *fs_info, |
1253 | struct btrfs_delayed_ref_root *delayed_refs, | |
1254 | struct btrfs_delayed_ref_head *head); | |
2ff7e61e | 1255 | int btrfs_lookup_data_extent(struct btrfs_fs_info *fs_info, u64 start, u64 len); |
a22285a6 | 1256 | int btrfs_lookup_extent_info(struct btrfs_trans_handle *trans, |
2ff7e61e | 1257 | struct btrfs_fs_info *fs_info, u64 bytenr, |
3173a18f | 1258 | u64 offset, int metadata, u64 *refs, u64 *flags); |
b25c36f8 NB |
1259 | int btrfs_pin_extent(struct btrfs_trans_handle *trans, u64 bytenr, u64 num, |
1260 | int reserved); | |
9fce5704 | 1261 | int btrfs_pin_extent_for_log_replay(struct btrfs_trans_handle *trans, |
e688b725 | 1262 | u64 bytenr, u64 num_bytes); |
bcdc428c | 1263 | int btrfs_exclude_logged_extents(struct extent_buffer *eb); |
e4c3b2dc | 1264 | int btrfs_cross_ref_exist(struct btrfs_root *root, |
1a89f173 FM |
1265 | u64 objectid, u64 offset, u64 bytenr, bool strict, |
1266 | struct btrfs_path *path); | |
4d75f8a9 | 1267 | struct extent_buffer *btrfs_alloc_tree_block(struct btrfs_trans_handle *trans, |
310712b2 OS |
1268 | struct btrfs_root *root, |
1269 | u64 parent, u64 root_objectid, | |
1270 | const struct btrfs_disk_key *key, | |
1271 | int level, u64 hint, | |
9631e4cc JB |
1272 | u64 empty_size, |
1273 | enum btrfs_lock_nesting nest); | |
f0486c68 | 1274 | void btrfs_free_tree_block(struct btrfs_trans_handle *trans, |
7a163608 | 1275 | u64 root_id, |
f0486c68 | 1276 | struct extent_buffer *buf, |
5581a51a | 1277 | u64 parent, int last_ref); |
5d4f98a2 | 1278 | int btrfs_alloc_reserved_file_extent(struct btrfs_trans_handle *trans, |
84f7d8e6 | 1279 | struct btrfs_root *root, u64 owner, |
5846a3c2 QW |
1280 | u64 offset, u64 ram_bytes, |
1281 | struct btrfs_key *ins); | |
5d4f98a2 | 1282 | int btrfs_alloc_logged_file_extent(struct btrfs_trans_handle *trans, |
5d4f98a2 YZ |
1283 | u64 root_objectid, u64 owner, u64 offset, |
1284 | struct btrfs_key *ins); | |
18513091 | 1285 | int btrfs_reserve_extent(struct btrfs_root *root, u64 ram_bytes, u64 num_bytes, |
00361589 | 1286 | u64 min_alloc_size, u64 empty_size, u64 hint_byte, |
e570fd27 | 1287 | struct btrfs_key *ins, int is_data, int delalloc); |
e089f05c | 1288 | int btrfs_inc_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root, |
e339a6b0 | 1289 | struct extent_buffer *buf, int full_backref); |
5d4f98a2 | 1290 | int btrfs_dec_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root, |
e339a6b0 | 1291 | struct extent_buffer *buf, int full_backref); |
5d4f98a2 | 1292 | int btrfs_set_disk_extent_flags(struct btrfs_trans_handle *trans, |
2fe6a5a1 | 1293 | struct extent_buffer *eb, u64 flags, int level); |
ffd4bb2a | 1294 | int btrfs_free_extent(struct btrfs_trans_handle *trans, struct btrfs_ref *ref); |
5d4f98a2 | 1295 | |
2ff7e61e JM |
1296 | int btrfs_free_reserved_extent(struct btrfs_fs_info *fs_info, |
1297 | u64 start, u64 len, int delalloc); | |
7bfc1007 | 1298 | int btrfs_pin_reserved_extent(struct btrfs_trans_handle *trans, u64 start, |
a0fbf736 | 1299 | u64 len); |
5ead2dd0 | 1300 | int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans); |
b18c6685 | 1301 | int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans, |
82fa113f | 1302 | struct btrfs_ref *generic_ref); |
5d4f98a2 | 1303 | |
4184ea7f | 1304 | void btrfs_clear_space_info_full(struct btrfs_fs_info *info); |
08e007d2 | 1305 | |
d5c12070 MX |
1306 | int btrfs_subvolume_reserve_metadata(struct btrfs_root *root, |
1307 | struct btrfs_block_rsv *rsv, | |
c4c129db | 1308 | int nitems, bool use_global_rsv); |
e85fde51 | 1309 | void btrfs_subvolume_release_metadata(struct btrfs_root *root, |
7775c818 | 1310 | struct btrfs_block_rsv *rsv); |
8702ba93 | 1311 | void btrfs_delalloc_release_extents(struct btrfs_inode *inode, u64 num_bytes); |
8b62f87b | 1312 | |
28c9b1e7 | 1313 | int btrfs_delalloc_reserve_metadata(struct btrfs_inode *inode, u64 num_bytes, |
d4135134 | 1314 | u64 disk_num_bytes, bool noflush); |
6d07bcec | 1315 | u64 btrfs_account_ro_block_groups_free_space(struct btrfs_space_info *sinfo); |
2ff7e61e | 1316 | int btrfs_error_unpin_extent_range(struct btrfs_fs_info *fs_info, |
acce952b | 1317 | u64 start, u64 end); |
2ff7e61e | 1318 | int btrfs_discard_extent(struct btrfs_fs_info *fs_info, u64 bytenr, |
1edb647b | 1319 | u64 num_bytes, u64 *actual_bytes); |
2ff7e61e | 1320 | int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range); |
acce952b | 1321 | |
c59021f8 | 1322 | int btrfs_init_space_info(struct btrfs_fs_info *fs_info); |
bed92eae AJ |
1323 | int btrfs_delayed_refs_qgroup_accounting(struct btrfs_trans_handle *trans, |
1324 | struct btrfs_fs_info *fs_info); | |
ea14b57f DS |
1325 | int btrfs_start_write_no_snapshotting(struct btrfs_root *root); |
1326 | void btrfs_end_write_no_snapshotting(struct btrfs_root *root); | |
0bc19f90 | 1327 | void btrfs_wait_for_snapshot_creation(struct btrfs_root *root); |
a5ed9182 | 1328 | |
dee26a9f | 1329 | /* ctree.c */ |
226463d7 JB |
1330 | int __init btrfs_ctree_init(void); |
1331 | void __cold btrfs_ctree_exit(void); | |
310712b2 | 1332 | int btrfs_bin_search(struct extent_buffer *eb, const struct btrfs_key *key, |
e3b83361 | 1333 | int *slot); |
e1f60a65 | 1334 | int __pure btrfs_comp_cpu_keys(const struct btrfs_key *k1, const struct btrfs_key *k2); |
0b86a832 CM |
1335 | int btrfs_previous_item(struct btrfs_root *root, |
1336 | struct btrfs_path *path, u64 min_objectid, | |
1337 | int type); | |
ade2e0b3 WS |
1338 | int btrfs_previous_extent_item(struct btrfs_root *root, |
1339 | struct btrfs_path *path, u64 min_objectid); | |
b7a0365e DD |
1340 | void btrfs_set_item_key_safe(struct btrfs_fs_info *fs_info, |
1341 | struct btrfs_path *path, | |
310712b2 | 1342 | const struct btrfs_key *new_key); |
925baedd | 1343 | struct extent_buffer *btrfs_root_node(struct btrfs_root *root); |
e7a84565 | 1344 | int btrfs_find_next_key(struct btrfs_root *root, struct btrfs_path *path, |
3f157a2f | 1345 | struct btrfs_key *key, int lowest_level, |
de78b51a | 1346 | u64 min_trans); |
3f157a2f | 1347 | int btrfs_search_forward(struct btrfs_root *root, struct btrfs_key *min_key, |
de78b51a | 1348 | struct btrfs_path *path, |
3f157a2f | 1349 | u64 min_trans); |
4b231ae4 DS |
1350 | struct extent_buffer *btrfs_read_node_slot(struct extent_buffer *parent, |
1351 | int slot); | |
1352 | ||
5f39d397 CM |
1353 | int btrfs_cow_block(struct btrfs_trans_handle *trans, |
1354 | struct btrfs_root *root, struct extent_buffer *buf, | |
1355 | struct extent_buffer *parent, int parent_slot, | |
9631e4cc JB |
1356 | struct extent_buffer **cow_ret, |
1357 | enum btrfs_lock_nesting nest); | |
be20aa9d CM |
1358 | int btrfs_copy_root(struct btrfs_trans_handle *trans, |
1359 | struct btrfs_root *root, | |
1360 | struct extent_buffer *buf, | |
1361 | struct extent_buffer **cow_ret, u64 new_root_objectid); | |
5d4f98a2 YZ |
1362 | int btrfs_block_can_be_shared(struct btrfs_root *root, |
1363 | struct extent_buffer *buf); | |
c71dd880 | 1364 | void btrfs_extend_item(struct btrfs_path *path, u32 data_size); |
78ac4f9e | 1365 | void btrfs_truncate_item(struct btrfs_path *path, u32 new_size, int from_end); |
459931ec CM |
1366 | int btrfs_split_item(struct btrfs_trans_handle *trans, |
1367 | struct btrfs_root *root, | |
1368 | struct btrfs_path *path, | |
310712b2 | 1369 | const struct btrfs_key *new_key, |
459931ec | 1370 | unsigned long split_offset); |
ad48fd75 YZ |
1371 | int btrfs_duplicate_item(struct btrfs_trans_handle *trans, |
1372 | struct btrfs_root *root, | |
1373 | struct btrfs_path *path, | |
310712b2 | 1374 | const struct btrfs_key *new_key); |
e33d5c3d KN |
1375 | int btrfs_find_item(struct btrfs_root *fs_root, struct btrfs_path *path, |
1376 | u64 inum, u64 ioff, u8 key_type, struct btrfs_key *found_key); | |
310712b2 OS |
1377 | int btrfs_search_slot(struct btrfs_trans_handle *trans, struct btrfs_root *root, |
1378 | const struct btrfs_key *key, struct btrfs_path *p, | |
1379 | int ins_len, int cow); | |
1380 | int btrfs_search_old_slot(struct btrfs_root *root, const struct btrfs_key *key, | |
5d9e75c4 | 1381 | struct btrfs_path *p, u64 time_seq); |
2f38b3e1 | 1382 | int btrfs_search_slot_for_read(struct btrfs_root *root, |
310712b2 OS |
1383 | const struct btrfs_key *key, |
1384 | struct btrfs_path *p, int find_higher, | |
1385 | int return_any); | |
6702ed49 | 1386 | int btrfs_realloc_node(struct btrfs_trans_handle *trans, |
5f39d397 | 1387 | struct btrfs_root *root, struct extent_buffer *parent, |
de78b51a | 1388 | int start_slot, u64 *last_ret, |
a6b6e75e | 1389 | struct btrfs_key *progress); |
b3b4aa74 | 1390 | void btrfs_release_path(struct btrfs_path *p); |
2c90e5d6 CM |
1391 | struct btrfs_path *btrfs_alloc_path(void); |
1392 | void btrfs_free_path(struct btrfs_path *p); | |
b4ce94de | 1393 | |
85e21bac CM |
1394 | int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root, |
1395 | struct btrfs_path *path, int slot, int nr); | |
85e21bac CM |
1396 | static inline int btrfs_del_item(struct btrfs_trans_handle *trans, |
1397 | struct btrfs_root *root, | |
1398 | struct btrfs_path *path) | |
1399 | { | |
1400 | return btrfs_del_items(trans, root, path, path->slots[0], 1); | |
1401 | } | |
1402 | ||
b7ef5f3a FM |
1403 | /* |
1404 | * Describes a batch of items to insert in a btree. This is used by | |
f0641656 | 1405 | * btrfs_insert_empty_items(). |
b7ef5f3a FM |
1406 | */ |
1407 | struct btrfs_item_batch { | |
1408 | /* | |
1409 | * Pointer to an array containing the keys of the items to insert (in | |
1410 | * sorted order). | |
1411 | */ | |
1412 | const struct btrfs_key *keys; | |
1413 | /* Pointer to an array containing the data size for each item to insert. */ | |
1414 | const u32 *data_sizes; | |
1415 | /* | |
1416 | * The sum of data sizes for all items. The caller can compute this while | |
1417 | * setting up the data_sizes array, so it ends up being more efficient | |
1418 | * than having btrfs_insert_empty_items() or setup_item_for_insert() | |
1419 | * doing it, as it would avoid an extra loop over a potentially large | |
1420 | * array, and in the case of setup_item_for_insert(), we would be doing | |
1421 | * it while holding a write lock on a leaf and often on upper level nodes | |
1422 | * too, unnecessarily increasing the size of a critical section. | |
1423 | */ | |
1424 | u32 total_data_size; | |
1425 | /* Size of the keys and data_sizes arrays (number of items in the batch). */ | |
1426 | int nr; | |
1427 | }; | |
1428 | ||
f0641656 FM |
1429 | void btrfs_setup_item_for_insert(struct btrfs_root *root, |
1430 | struct btrfs_path *path, | |
1431 | const struct btrfs_key *key, | |
1432 | u32 data_size); | |
310712b2 OS |
1433 | int btrfs_insert_item(struct btrfs_trans_handle *trans, struct btrfs_root *root, |
1434 | const struct btrfs_key *key, void *data, u32 data_size); | |
9c58309d CM |
1435 | int btrfs_insert_empty_items(struct btrfs_trans_handle *trans, |
1436 | struct btrfs_root *root, | |
1437 | struct btrfs_path *path, | |
b7ef5f3a | 1438 | const struct btrfs_item_batch *batch); |
9c58309d CM |
1439 | |
1440 | static inline int btrfs_insert_empty_item(struct btrfs_trans_handle *trans, | |
1441 | struct btrfs_root *root, | |
1442 | struct btrfs_path *path, | |
310712b2 | 1443 | const struct btrfs_key *key, |
9c58309d CM |
1444 | u32 data_size) |
1445 | { | |
b7ef5f3a FM |
1446 | struct btrfs_item_batch batch; |
1447 | ||
1448 | batch.keys = key; | |
1449 | batch.data_sizes = &data_size; | |
1450 | batch.total_data_size = data_size; | |
1451 | batch.nr = 1; | |
1452 | ||
1453 | return btrfs_insert_empty_items(trans, root, path, &batch); | |
9c58309d CM |
1454 | } |
1455 | ||
16e7549f | 1456 | int btrfs_prev_leaf(struct btrfs_root *root, struct btrfs_path *path); |
3d7806ec JS |
1457 | int btrfs_next_old_leaf(struct btrfs_root *root, struct btrfs_path *path, |
1458 | u64 time_seq); | |
0ff40a91 MPS |
1459 | |
1460 | int btrfs_search_backwards(struct btrfs_root *root, struct btrfs_key *key, | |
1461 | struct btrfs_path *path); | |
1462 | ||
62142be3 GN |
1463 | int btrfs_get_next_valid_item(struct btrfs_root *root, struct btrfs_key *key, |
1464 | struct btrfs_path *path); | |
1465 | ||
1466 | /* | |
1467 | * Search in @root for a given @key, and store the slot found in @found_key. | |
1468 | * | |
1469 | * @root: The root node of the tree. | |
1470 | * @key: The key we are looking for. | |
1471 | * @found_key: Will hold the found item. | |
1472 | * @path: Holds the current slot/leaf. | |
1473 | * @iter_ret: Contains the value returned from btrfs_search_slot or | |
1474 | * btrfs_get_next_valid_item, whichever was executed last. | |
1475 | * | |
1476 | * The @iter_ret is an output variable that will contain the return value of | |
1477 | * btrfs_search_slot, if it encountered an error, or the value returned from | |
1478 | * btrfs_get_next_valid_item otherwise. That return value can be 0, if a valid | |
1479 | * slot was found, 1 if there were no more leaves, and <0 if there was an error. | |
1480 | * | |
1481 | * It's recommended to use a separate variable for iter_ret and then use it to | |
1482 | * set the function return value so there's no confusion of the 0/1/errno | |
1483 | * values stemming from btrfs_search_slot. | |
1484 | */ | |
1485 | #define btrfs_for_each_slot(root, key, found_key, path, iter_ret) \ | |
1486 | for (iter_ret = btrfs_search_slot(NULL, (root), (key), (path), 0, 0); \ | |
1487 | (iter_ret) >= 0 && \ | |
1488 | (iter_ret = btrfs_get_next_valid_item((root), (found_key), (path))) == 0; \ | |
1489 | (path)->slots[0]++ \ | |
1490 | ) | |
1491 | ||
890d2b1a | 1492 | int btrfs_next_old_item(struct btrfs_root *root, struct btrfs_path *path, u64 time_seq); |
809d6902 DS |
1493 | |
1494 | /* | |
1495 | * Search the tree again to find a leaf with greater keys. | |
1496 | * | |
1497 | * Returns 0 if it found something or 1 if there are no greater leaves. | |
1498 | * Returns < 0 on error. | |
1499 | */ | |
1500 | static inline int btrfs_next_leaf(struct btrfs_root *root, struct btrfs_path *path) | |
1501 | { | |
1502 | return btrfs_next_old_leaf(root, path, 0); | |
1503 | } | |
1504 | ||
1c8f52a5 AB |
1505 | static inline int btrfs_next_item(struct btrfs_root *root, struct btrfs_path *p) |
1506 | { | |
1507 | return btrfs_next_old_item(root, p, 0); | |
1508 | } | |
e902baac | 1509 | int btrfs_leaf_free_space(struct extent_buffer *leaf); |
0078a9f9 NB |
1510 | int __must_check btrfs_drop_snapshot(struct btrfs_root *root, int update_ref, |
1511 | int for_reloc); | |
f82d02d9 YZ |
1512 | int btrfs_drop_subtree(struct btrfs_trans_handle *trans, |
1513 | struct btrfs_root *root, | |
1514 | struct extent_buffer *node, | |
1515 | struct extent_buffer *parent); | |
babbf170 | 1516 | |
dee26a9f | 1517 | /* root-item.c */ |
6025c19f | 1518 | int btrfs_add_root_ref(struct btrfs_trans_handle *trans, u64 root_id, |
e43eec81 STD |
1519 | u64 ref_id, u64 dirid, u64 sequence, |
1520 | const struct qstr *name); | |
3ee1c553 | 1521 | int btrfs_del_root_ref(struct btrfs_trans_handle *trans, u64 root_id, |
e43eec81 STD |
1522 | u64 ref_id, u64 dirid, u64 *sequence, |
1523 | const struct qstr *name); | |
1cd5447e | 1524 | int btrfs_del_root(struct btrfs_trans_handle *trans, |
ab9ce7d4 | 1525 | const struct btrfs_key *key); |
310712b2 OS |
1526 | int btrfs_insert_root(struct btrfs_trans_handle *trans, struct btrfs_root *root, |
1527 | const struct btrfs_key *key, | |
1528 | struct btrfs_root_item *item); | |
b45a9d8b JM |
1529 | int __must_check btrfs_update_root(struct btrfs_trans_handle *trans, |
1530 | struct btrfs_root *root, | |
1531 | struct btrfs_key *key, | |
1532 | struct btrfs_root_item *item); | |
310712b2 | 1533 | int btrfs_find_root(struct btrfs_root *root, const struct btrfs_key *search_key, |
cb517eab MX |
1534 | struct btrfs_path *path, struct btrfs_root_item *root_item, |
1535 | struct btrfs_key *root_key); | |
6bccf3ab | 1536 | int btrfs_find_orphan_roots(struct btrfs_fs_info *fs_info); |
bf5f32ec MF |
1537 | void btrfs_set_root_node(struct btrfs_root_item *item, |
1538 | struct extent_buffer *node); | |
08fe4db1 | 1539 | void btrfs_check_and_init_root_item(struct btrfs_root_item *item); |
8ea05e3a AB |
1540 | void btrfs_update_root_times(struct btrfs_trans_handle *trans, |
1541 | struct btrfs_root *root); | |
08fe4db1 | 1542 | |
07b30a49 | 1543 | /* uuid-tree.c */ |
cdb345a8 | 1544 | int btrfs_uuid_tree_add(struct btrfs_trans_handle *trans, u8 *uuid, u8 type, |
07b30a49 | 1545 | u64 subid); |
d1957791 | 1546 | int btrfs_uuid_tree_remove(struct btrfs_trans_handle *trans, u8 *uuid, u8 type, |
07b30a49 | 1547 | u64 subid); |
560b7a4a | 1548 | int btrfs_uuid_tree_iterate(struct btrfs_fs_info *fs_info); |
07b30a49 | 1549 | |
dee26a9f | 1550 | /* dir-item.c */ |
9c52057c | 1551 | int btrfs_check_dir_item_collision(struct btrfs_root *root, u64 dir, |
e43eec81 STD |
1552 | const struct qstr *name); |
1553 | int btrfs_insert_dir_item(struct btrfs_trans_handle *trans, | |
1554 | const struct qstr *name, struct btrfs_inode *dir, | |
aec7477b | 1555 | struct btrfs_key *location, u8 type, u64 index); |
7e38180e CM |
1556 | struct btrfs_dir_item *btrfs_lookup_dir_item(struct btrfs_trans_handle *trans, |
1557 | struct btrfs_root *root, | |
1558 | struct btrfs_path *path, u64 dir, | |
e43eec81 | 1559 | const struct qstr *name, int mod); |
7e38180e CM |
1560 | struct btrfs_dir_item * |
1561 | btrfs_lookup_dir_index_item(struct btrfs_trans_handle *trans, | |
1562 | struct btrfs_root *root, | |
1563 | struct btrfs_path *path, u64 dir, | |
e43eec81 | 1564 | u64 index, const struct qstr *name, int mod); |
4df27c4d YZ |
1565 | struct btrfs_dir_item * |
1566 | btrfs_search_dir_index_item(struct btrfs_root *root, | |
1567 | struct btrfs_path *path, u64 dirid, | |
e43eec81 | 1568 | const struct qstr *name); |
7e38180e CM |
1569 | int btrfs_delete_one_dir_name(struct btrfs_trans_handle *trans, |
1570 | struct btrfs_root *root, | |
1571 | struct btrfs_path *path, | |
1572 | struct btrfs_dir_item *di); | |
5103e947 | 1573 | int btrfs_insert_xattr_item(struct btrfs_trans_handle *trans, |
f34f57a3 YZ |
1574 | struct btrfs_root *root, |
1575 | struct btrfs_path *path, u64 objectid, | |
1576 | const char *name, u16 name_len, | |
1577 | const void *data, u16 data_len); | |
5103e947 JB |
1578 | struct btrfs_dir_item *btrfs_lookup_xattr(struct btrfs_trans_handle *trans, |
1579 | struct btrfs_root *root, | |
1580 | struct btrfs_path *path, u64 dir, | |
1581 | const char *name, u16 name_len, | |
1582 | int mod); | |
2ff7e61e | 1583 | struct btrfs_dir_item *btrfs_match_dir_item_name(struct btrfs_fs_info *fs_info, |
5f5bc6b1 FM |
1584 | struct btrfs_path *path, |
1585 | const char *name, | |
1586 | int name_len); | |
7b128766 JB |
1587 | |
1588 | /* orphan.c */ | |
1589 | int btrfs_insert_orphan_item(struct btrfs_trans_handle *trans, | |
1590 | struct btrfs_root *root, u64 offset); | |
1591 | int btrfs_del_orphan_item(struct btrfs_trans_handle *trans, | |
1592 | struct btrfs_root *root, u64 offset); | |
4df27c4d | 1593 | int btrfs_find_orphan_item(struct btrfs_root *root, u64 offset); |
7b128766 | 1594 | |
dee26a9f | 1595 | /* file-item.c */ |
459931ec | 1596 | int btrfs_del_csums(struct btrfs_trans_handle *trans, |
40e046ac | 1597 | struct btrfs_root *root, u64 bytenr, u64 len); |
6275193e | 1598 | blk_status_t btrfs_lookup_bio_sums(struct inode *inode, struct bio *bio, u8 *dst); |
d1f68ba0 OS |
1599 | int btrfs_insert_hole_extent(struct btrfs_trans_handle *trans, |
1600 | struct btrfs_root *root, u64 objectid, u64 pos, | |
1601 | u64 num_bytes); | |
dee26a9f CM |
1602 | int btrfs_lookup_file_extent(struct btrfs_trans_handle *trans, |
1603 | struct btrfs_root *root, | |
1604 | struct btrfs_path *path, u64 objectid, | |
db94535d | 1605 | u64 bytenr, int mod); |
065631f6 | 1606 | int btrfs_csum_file_blocks(struct btrfs_trans_handle *trans, |
d20f7043 | 1607 | struct btrfs_root *root, |
e6dcd2dc | 1608 | struct btrfs_ordered_sum *sums); |
bd242a08 | 1609 | blk_status_t btrfs_csum_one_bio(struct btrfs_inode *inode, struct bio *bio, |
e331f6b1 | 1610 | u64 offset, bool one_ordered); |
a2de733c | 1611 | int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end, |
26ce9114 JB |
1612 | struct list_head *list, int search_commit, |
1613 | bool nowait); | |
9cdc5124 | 1614 | void btrfs_extent_item_to_extent_map(struct btrfs_inode *inode, |
7ffbb598 FM |
1615 | const struct btrfs_path *path, |
1616 | struct btrfs_file_extent_item *fi, | |
1617 | const bool new_inline, | |
1618 | struct extent_map *em); | |
41a2ee75 JB |
1619 | int btrfs_inode_clear_file_extent_range(struct btrfs_inode *inode, u64 start, |
1620 | u64 len); | |
1621 | int btrfs_inode_set_file_extent_range(struct btrfs_inode *inode, u64 start, | |
1622 | u64 len); | |
76aea537 | 1623 | void btrfs_inode_safe_disk_i_size_write(struct btrfs_inode *inode, u64 new_i_size); |
a5eeb3d1 | 1624 | u64 btrfs_file_extent_end(const struct btrfs_path *path); |
7ffbb598 | 1625 | |
39279cc3 | 1626 | /* inode.c */ |
c93104e7 CH |
1627 | void btrfs_submit_data_write_bio(struct inode *inode, struct bio *bio, int mirror_num); |
1628 | void btrfs_submit_data_read_bio(struct inode *inode, struct bio *bio, | |
1629 | int mirror_num, enum btrfs_compression_type compress_type); | |
ae643a74 QW |
1630 | int btrfs_check_sector_csum(struct btrfs_fs_info *fs_info, struct page *page, |
1631 | u32 pgoff, u8 *csum, const u8 * const csum_expected); | |
81bd9328 CH |
1632 | int btrfs_check_data_csum(struct inode *inode, struct btrfs_bio *bbio, |
1633 | u32 bio_offset, struct page *page, u32 pgoff); | |
c3a3b19b QW |
1634 | unsigned int btrfs_verify_data_csum(struct btrfs_bio *bbio, |
1635 | u32 bio_offset, struct page *page, | |
1636 | u64 start, u64 end); | |
7959bd44 CH |
1637 | int btrfs_check_data_csum(struct inode *inode, struct btrfs_bio *bbio, |
1638 | u32 bio_offset, struct page *page, u32 pgoff); | |
00361589 | 1639 | noinline int can_nocow_extent(struct inode *inode, u64 offset, u64 *len, |
7ee9e440 | 1640 | u64 *orig_start, u64 *orig_block_len, |
26ce9114 | 1641 | u64 *ram_bytes, bool nowait, bool strict); |
4881ee5a | 1642 | |
2b877331 NB |
1643 | void __btrfs_del_delalloc_inode(struct btrfs_root *root, |
1644 | struct btrfs_inode *inode); | |
3de4586c | 1645 | struct inode *btrfs_lookup_dentry(struct inode *dir, struct dentry *dentry); |
70ddc553 | 1646 | int btrfs_set_inode_index(struct btrfs_inode *dir, u64 *index); |
e02119d5 | 1647 | int btrfs_unlink_inode(struct btrfs_trans_handle *trans, |
4ec5934e | 1648 | struct btrfs_inode *dir, struct btrfs_inode *inode, |
e43eec81 | 1649 | const struct qstr *name); |
e02119d5 | 1650 | int btrfs_add_link(struct btrfs_trans_handle *trans, |
db0a669f | 1651 | struct btrfs_inode *parent_inode, struct btrfs_inode *inode, |
e43eec81 | 1652 | const struct qstr *name, int add_backref, u64 index); |
f60a2364 | 1653 | int btrfs_delete_subvolume(struct inode *dir, struct dentry *dentry); |
217f42eb NB |
1654 | int btrfs_truncate_block(struct btrfs_inode *inode, loff_t from, loff_t len, |
1655 | int front); | |
e02119d5 | 1656 | |
f9baa501 | 1657 | int btrfs_start_delalloc_snapshot(struct btrfs_root *root, bool in_reclaim_context); |
9db4dc24 | 1658 | int btrfs_start_delalloc_roots(struct btrfs_fs_info *fs_info, long nr, |
3d45f221 | 1659 | bool in_reclaim_context); |
c2566f22 | 1660 | int btrfs_set_extent_delalloc(struct btrfs_inode *inode, u64 start, u64 end, |
e3b8a485 | 1661 | unsigned int extra_bits, |
330a5827 | 1662 | struct extent_state **cached_state); |
3538d68d OS |
1663 | struct btrfs_new_inode_args { |
1664 | /* Input */ | |
1665 | struct inode *dir; | |
1666 | struct dentry *dentry; | |
1667 | struct inode *inode; | |
1668 | bool orphan; | |
1669 | bool subvol; | |
1670 | ||
1671 | /* | |
1672 | * Output from btrfs_new_inode_prepare(), input to | |
1673 | * btrfs_create_new_inode(). | |
1674 | */ | |
1675 | struct posix_acl *default_acl; | |
1676 | struct posix_acl *acl; | |
1677 | }; | |
1678 | int btrfs_new_inode_prepare(struct btrfs_new_inode_args *args, | |
1679 | unsigned int *trans_num_items); | |
1680 | int btrfs_create_new_inode(struct btrfs_trans_handle *trans, | |
caae78e0 | 1681 | struct btrfs_new_inode_args *args); |
3538d68d | 1682 | void btrfs_new_inode_args_destroy(struct btrfs_new_inode_args *args); |
a1fd0c35 OS |
1683 | struct inode *btrfs_new_subvol_inode(struct user_namespace *mnt_userns, |
1684 | struct inode *dir); | |
c629732d | 1685 | void btrfs_set_delalloc_extent(struct inode *inode, struct extent_state *state, |
6d92b304 | 1686 | u32 bits); |
a36bb5f9 | 1687 | void btrfs_clear_delalloc_extent(struct inode *inode, |
6d92b304 | 1688 | struct extent_state *state, u32 bits); |
5c848198 NB |
1689 | void btrfs_merge_delalloc_extent(struct inode *inode, struct extent_state *new, |
1690 | struct extent_state *other); | |
abbb55f4 NB |
1691 | void btrfs_split_delalloc_extent(struct inode *inode, |
1692 | struct extent_state *orig, u64 split); | |
d2a91064 | 1693 | void btrfs_set_range_writeback(struct btrfs_inode *inode, u64 start, u64 end); |
a528a241 | 1694 | vm_fault_t btrfs_page_mkwrite(struct vm_fault *vmf); |
bd555975 | 1695 | void btrfs_evict_inode(struct inode *inode); |
39279cc3 CM |
1696 | struct inode *btrfs_alloc_inode(struct super_block *sb); |
1697 | void btrfs_destroy_inode(struct inode *inode); | |
26602cab | 1698 | void btrfs_free_inode(struct inode *inode); |
45321ac5 | 1699 | int btrfs_drop_inode(struct inode *inode); |
f5c29bd9 | 1700 | int __init btrfs_init_cachep(void); |
e67c718b | 1701 | void __cold btrfs_destroy_cachep(void); |
0202e83f | 1702 | struct inode *btrfs_iget_path(struct super_block *s, u64 ino, |
4c66e0d4 | 1703 | struct btrfs_root *root, struct btrfs_path *path); |
0202e83f | 1704 | struct inode *btrfs_iget(struct super_block *s, u64 ino, struct btrfs_root *root); |
fc4f21b1 | 1705 | struct extent_map *btrfs_get_extent(struct btrfs_inode *inode, |
de2c6615 | 1706 | struct page *page, size_t pg_offset, |
39b07b5d | 1707 | u64 start, u64 end); |
a52d9a80 | 1708 | int btrfs_update_inode(struct btrfs_trans_handle *trans, |
9a56fcd1 | 1709 | struct btrfs_root *root, struct btrfs_inode *inode); |
be6aef60 | 1710 | int btrfs_update_inode_fallback(struct btrfs_trans_handle *trans, |
729f7961 | 1711 | struct btrfs_root *root, struct btrfs_inode *inode); |
73f2e545 NB |
1712 | int btrfs_orphan_add(struct btrfs_trans_handle *trans, |
1713 | struct btrfs_inode *inode); | |
66b4ffd1 | 1714 | int btrfs_orphan_cleanup(struct btrfs_root *root); |
b06359a3 | 1715 | int btrfs_cont_expand(struct btrfs_inode *inode, loff_t oldsize, loff_t size); |
24bbcf04 | 1716 | void btrfs_add_delayed_iput(struct inode *inode); |
2ff7e61e | 1717 | void btrfs_run_delayed_iputs(struct btrfs_fs_info *fs_info); |
034f784d | 1718 | int btrfs_wait_on_delayed_iputs(struct btrfs_fs_info *fs_info); |
efa56464 YZ |
1719 | int btrfs_prealloc_file_range(struct inode *inode, int mode, |
1720 | u64 start, u64 num_bytes, u64 min_size, | |
1721 | loff_t actual_len, u64 *alloc_hint); | |
0af3d00b JB |
1722 | int btrfs_prealloc_file_range_trans(struct inode *inode, |
1723 | struct btrfs_trans_handle *trans, int mode, | |
1724 | u64 start, u64 num_bytes, u64 min_size, | |
1725 | loff_t actual_len, u64 *alloc_hint); | |
98456b9c | 1726 | int btrfs_run_delalloc_range(struct btrfs_inode *inode, struct page *locked_page, |
5eaad97a NB |
1727 | u64 start, u64 end, int *page_started, unsigned long *nr_written, |
1728 | struct writeback_control *wbc); | |
a129ffb8 | 1729 | int btrfs_writepage_cow_fixup(struct page *page); |
38a39ac7 QW |
1730 | void btrfs_writepage_endio_finish_ordered(struct btrfs_inode *inode, |
1731 | struct page *page, u64 start, | |
25c1252a | 1732 | u64 end, bool uptodate); |
3ea4dc5b OS |
1733 | int btrfs_encoded_io_compression_from_extent(struct btrfs_fs_info *fs_info, |
1734 | int compress_type); | |
1735 | int btrfs_encoded_read_regular_fill_pages(struct btrfs_inode *inode, | |
1736 | u64 file_offset, u64 disk_bytenr, | |
1737 | u64 disk_io_size, | |
1738 | struct page **pages); | |
1881fba8 OS |
1739 | ssize_t btrfs_encoded_read(struct kiocb *iocb, struct iov_iter *iter, |
1740 | struct btrfs_ioctl_encoded_io_args *encoded); | |
7c0c7269 OS |
1741 | ssize_t btrfs_do_encoded_write(struct kiocb *iocb, struct iov_iter *from, |
1742 | const struct btrfs_ioctl_encoded_io_args *encoded); | |
1881fba8 | 1743 | |
8184620a FM |
1744 | ssize_t btrfs_dio_read(struct kiocb *iocb, struct iov_iter *iter, |
1745 | size_t done_before); | |
1746 | struct iomap_dio *btrfs_dio_write(struct kiocb *iocb, struct iov_iter *iter, | |
1747 | size_t done_before); | |
36e8c622 | 1748 | |
82d339d9 | 1749 | extern const struct dentry_operations btrfs_dentry_operations; |
f46b5a66 | 1750 | |
a14b78ad | 1751 | /* Inode locking type flags, by default the exclusive lock is taken */ |
c7321b76 DS |
1752 | enum btrfs_ilock_type { |
1753 | ENUM_BIT(BTRFS_ILOCK_SHARED), | |
1754 | ENUM_BIT(BTRFS_ILOCK_TRY), | |
1755 | ENUM_BIT(BTRFS_ILOCK_MMAP), | |
1756 | }; | |
a14b78ad GR |
1757 | |
1758 | int btrfs_inode_lock(struct inode *inode, unsigned int ilock_flags); | |
1759 | void btrfs_inode_unlock(struct inode *inode, unsigned int ilock_flags); | |
2766ff61 FM |
1760 | void btrfs_update_inode_bytes(struct btrfs_inode *inode, |
1761 | const u64 add_bytes, | |
1762 | const u64 del_bytes); | |
63c34cb4 | 1763 | void btrfs_assert_inode_range_clean(struct btrfs_inode *inode, u64 start, u64 end); |
f46b5a66 CH |
1764 | |
1765 | /* ioctl.c */ | |
1766 | long btrfs_ioctl(struct file *file, unsigned int cmd, unsigned long arg); | |
4c63c245 | 1767 | long btrfs_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg); |
97fc2977 MS |
1768 | int btrfs_fileattr_get(struct dentry *dentry, struct fileattr *fa); |
1769 | int btrfs_fileattr_set(struct user_namespace *mnt_userns, | |
1770 | struct dentry *dentry, struct fileattr *fa); | |
d5131b65 | 1771 | int btrfs_ioctl_get_supported_features(void __user *arg); |
7b6a221e | 1772 | void btrfs_sync_inode_flags_to_i_flags(struct inode *inode); |
e1f60a65 | 1773 | int __pure btrfs_is_empty_uuid(u8 *uuid); |
1ccc2e8a | 1774 | int btrfs_defrag_file(struct inode *inode, struct file_ra_state *ra, |
4cb5300b | 1775 | struct btrfs_ioctl_defrag_range_args *range, |
1ccc2e8a | 1776 | u64 newer_than, unsigned long max_to_defrag); |
008ef096 DS |
1777 | void btrfs_get_block_group_info(struct list_head *groups_list, |
1778 | struct btrfs_ioctl_space_info *space); | |
1779 | void btrfs_update_ioctl_balance_args(struct btrfs_fs_info *fs_info, | |
35a3621b SB |
1780 | struct btrfs_ioctl_balance_args *bargs); |
1781 | ||
39279cc3 | 1782 | /* file.c */ |
f5c29bd9 | 1783 | int __init btrfs_auto_defrag_init(void); |
e67c718b | 1784 | void __cold btrfs_auto_defrag_exit(void); |
4cb5300b | 1785 | int btrfs_add_inode_defrag(struct btrfs_trans_handle *trans, |
558732df | 1786 | struct btrfs_inode *inode, u32 extent_thresh); |
4cb5300b | 1787 | int btrfs_run_defrag_inodes(struct btrfs_fs_info *fs_info); |
26176e7c | 1788 | void btrfs_cleanup_defrag_inodes(struct btrfs_fs_info *fs_info); |
02c24a82 | 1789 | int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync); |
828c0950 | 1790 | extern const struct file_operations btrfs_file_operations; |
5dc562c5 | 1791 | int btrfs_drop_extents(struct btrfs_trans_handle *trans, |
5893dfb9 FM |
1792 | struct btrfs_root *root, struct btrfs_inode *inode, |
1793 | struct btrfs_drop_extents_args *args); | |
bfc78479 NB |
1794 | int btrfs_replace_file_extents(struct btrfs_inode *inode, |
1795 | struct btrfs_path *path, const u64 start, | |
1796 | const u64 end, | |
bf385648 | 1797 | struct btrfs_replace_extent_info *extent_info, |
690a5dbf | 1798 | struct btrfs_trans_handle **trans_out); |
d899e052 | 1799 | int btrfs_mark_extent_written(struct btrfs_trans_handle *trans, |
7a6d7067 | 1800 | struct btrfs_inode *inode, u64 start, u64 end); |
7c0c7269 OS |
1801 | ssize_t btrfs_do_write_iter(struct kiocb *iocb, struct iov_iter *from, |
1802 | const struct btrfs_ioctl_encoded_io_args *encoded); | |
6bf13c0c | 1803 | int btrfs_release_file(struct inode *inode, struct file *file); |
088545f6 | 1804 | int btrfs_dirty_pages(struct btrfs_inode *inode, struct page **pages, |
2ff7e61e | 1805 | size_t num_pages, loff_t pos, size_t write_bytes, |
aa8c1a41 | 1806 | struct extent_state **cached, bool noreserve); |
728404da | 1807 | int btrfs_fdatawrite_range(struct inode *inode, loff_t start, loff_t end); |
38d37aa9 | 1808 | int btrfs_check_nocow_lock(struct btrfs_inode *inode, loff_t pos, |
80f9d241 | 1809 | size_t *write_bytes, bool nowait); |
38d37aa9 | 1810 | void btrfs_check_nocow_unlock(struct btrfs_inode *inode); |
ac3c0d36 FM |
1811 | bool btrfs_find_delalloc_in_range(struct btrfs_inode *inode, u64 start, u64 end, |
1812 | u64 *delalloc_start_ret, u64 *delalloc_end_ret); | |
6bf13c0c | 1813 | |
6702ed49 CM |
1814 | /* tree-defrag.c */ |
1815 | int btrfs_defrag_leaves(struct btrfs_trans_handle *trans, | |
de78b51a | 1816 | struct btrfs_root *root); |
58176a96 | 1817 | |
edbd8d4e | 1818 | /* super.c */ |
2ff7e61e | 1819 | int btrfs_parse_options(struct btrfs_fs_info *info, char *options, |
96da0919 | 1820 | unsigned long new_flags); |
6bf13c0c | 1821 | int btrfs_sync_fs(struct super_block *sb, int wait); |
c0c907a4 MPS |
1822 | char *btrfs_get_subvol_name_from_objectid(struct btrfs_fs_info *fs_info, |
1823 | u64 subvol_objectid); | |
533574c6 | 1824 | |
e9306ad4 QW |
1825 | #if BITS_PER_LONG == 32 |
1826 | #define BTRFS_32BIT_MAX_FILE_SIZE (((u64)ULONG_MAX + 1) << PAGE_SHIFT) | |
1827 | /* | |
1828 | * The warning threshold is 5/8th of the MAX_LFS_FILESIZE that limits the logical | |
1829 | * addresses of extents. | |
1830 | * | |
1831 | * For 4K page size it's about 10T, for 64K it's 160T. | |
1832 | */ | |
1833 | #define BTRFS_32BIT_EARLY_WARN_THRESHOLD (BTRFS_32BIT_MAX_FILE_SIZE * 5 / 8) | |
1834 | void btrfs_warn_32bit_limit(struct btrfs_fs_info *fs_info); | |
1835 | void btrfs_err_32bit_limit(struct btrfs_fs_info *fs_info); | |
1836 | #endif | |
1837 | ||
884b07d0 QW |
1838 | /* |
1839 | * Get the correct offset inside the page of extent buffer. | |
1840 | * | |
1841 | * @eb: target extent buffer | |
1842 | * @start: offset inside the extent buffer | |
1843 | * | |
1844 | * Will handle both sectorsize == PAGE_SIZE and sectorsize < PAGE_SIZE cases. | |
1845 | */ | |
1846 | static inline size_t get_eb_offset_in_page(const struct extent_buffer *eb, | |
1847 | unsigned long offset) | |
1848 | { | |
1849 | /* | |
1850 | * For sectorsize == PAGE_SIZE case, eb->start will always be aligned | |
1851 | * to PAGE_SIZE, thus adding it won't cause any difference. | |
1852 | * | |
1853 | * For sectorsize < PAGE_SIZE, we must only read the data that belongs | |
1854 | * to the eb, thus we have to take the eb->start into consideration. | |
1855 | */ | |
1856 | return offset_in_page(offset + eb->start); | |
1857 | } | |
1858 | ||
1859 | static inline unsigned long get_eb_page_index(unsigned long offset) | |
1860 | { | |
1861 | /* | |
1862 | * For sectorsize == PAGE_SIZE case, plain >> PAGE_SHIFT is enough. | |
1863 | * | |
1864 | * For sectorsize < PAGE_SIZE case, we only support 64K PAGE_SIZE, | |
1865 | * and have ensured that all tree blocks are contained in one page, | |
1866 | * thus we always get index == 0. | |
1867 | */ | |
1868 | return offset >> PAGE_SHIFT; | |
1869 | } | |
1870 | ||
f8f591df JT |
1871 | /* |
1872 | * Use that for functions that are conditionally exported for sanity tests but | |
1873 | * otherwise static | |
1874 | */ | |
1875 | #ifndef CONFIG_BTRFS_FS_RUN_SANITY_TESTS | |
1876 | #define EXPORT_FOR_TESTS static | |
1877 | #else | |
1878 | #define EXPORT_FOR_TESTS | |
1879 | #endif | |
1880 | ||
33268eaf | 1881 | /* acl.c */ |
0eda294d | 1882 | #ifdef CONFIG_BTRFS_FS_POSIX_ACL |
0cad6246 | 1883 | struct posix_acl *btrfs_get_acl(struct inode *inode, int type, bool rcu); |
549c7297 CB |
1884 | int btrfs_set_acl(struct user_namespace *mnt_userns, struct inode *inode, |
1885 | struct posix_acl *acl, int type); | |
3538d68d OS |
1886 | int __btrfs_set_acl(struct btrfs_trans_handle *trans, struct inode *inode, |
1887 | struct posix_acl *acl, int type); | |
9b89d95a | 1888 | #else |
ed8f3737 | 1889 | #define btrfs_get_acl NULL |
996a710d | 1890 | #define btrfs_set_acl NULL |
3538d68d OS |
1891 | static inline int __btrfs_set_acl(struct btrfs_trans_handle *trans, |
1892 | struct inode *inode, struct posix_acl *acl, | |
1893 | int type) | |
9b89d95a | 1894 | { |
3538d68d | 1895 | return -EOPNOTSUPP; |
9b89d95a | 1896 | } |
9b89d95a | 1897 | #endif |
0f9dd46c | 1898 | |
5d4f98a2 | 1899 | /* relocation.c */ |
6bccf3ab | 1900 | int btrfs_relocate_block_group(struct btrfs_fs_info *fs_info, u64 group_start); |
5d4f98a2 YZ |
1901 | int btrfs_init_reloc_root(struct btrfs_trans_handle *trans, |
1902 | struct btrfs_root *root); | |
1903 | int btrfs_update_reloc_root(struct btrfs_trans_handle *trans, | |
1904 | struct btrfs_root *root); | |
7eefae6b | 1905 | int btrfs_recover_relocation(struct btrfs_fs_info *fs_info); |
7bfa9535 | 1906 | int btrfs_reloc_clone_csums(struct btrfs_inode *inode, u64 file_pos, u64 len); |
83d4cfd4 JB |
1907 | int btrfs_reloc_cow_block(struct btrfs_trans_handle *trans, |
1908 | struct btrfs_root *root, struct extent_buffer *buf, | |
1909 | struct extent_buffer *cow); | |
147d256e | 1910 | void btrfs_reloc_pre_snapshot(struct btrfs_pending_snapshot *pending, |
3fd0a558 | 1911 | u64 *bytes_to_reserve); |
49b25e05 | 1912 | int btrfs_reloc_post_snapshot(struct btrfs_trans_handle *trans, |
3fd0a558 | 1913 | struct btrfs_pending_snapshot *pending); |
726a3421 | 1914 | int btrfs_should_cancel_balance(struct btrfs_fs_info *fs_info); |
2433bea5 QW |
1915 | struct btrfs_root *find_reloc_root(struct btrfs_fs_info *fs_info, |
1916 | u64 bytenr); | |
55465730 | 1917 | int btrfs_should_ignore_reloc_root(struct btrfs_root *root); |
a2de733c AJ |
1918 | |
1919 | /* scrub.c */ | |
aa1b8cd4 SB |
1920 | int btrfs_scrub_dev(struct btrfs_fs_info *fs_info, u64 devid, u64 start, |
1921 | u64 end, struct btrfs_scrub_progress *progress, | |
63a212ab | 1922 | int readonly, int is_dev_replace); |
2ff7e61e JM |
1923 | void btrfs_scrub_pause(struct btrfs_fs_info *fs_info); |
1924 | void btrfs_scrub_continue(struct btrfs_fs_info *fs_info); | |
aa1b8cd4 | 1925 | int btrfs_scrub_cancel(struct btrfs_fs_info *info); |
163e97ee | 1926 | int btrfs_scrub_cancel_dev(struct btrfs_device *dev); |
2ff7e61e | 1927 | int btrfs_scrub_progress(struct btrfs_fs_info *fs_info, u64 devid, |
a2de733c | 1928 | struct btrfs_scrub_progress *progress); |
c404e0dc MX |
1929 | |
1930 | /* dev-replace.c */ | |
1931 | void btrfs_bio_counter_inc_blocked(struct btrfs_fs_info *fs_info); | |
4245215d MX |
1932 | void btrfs_bio_counter_sub(struct btrfs_fs_info *fs_info, s64 amount); |
1933 | ||
1934 | static inline void btrfs_bio_counter_dec(struct btrfs_fs_info *fs_info) | |
1935 | { | |
1936 | btrfs_bio_counter_sub(fs_info, 1); | |
1937 | } | |
a2de733c | 1938 | |
95a06077 JS |
1939 | static inline int is_fstree(u64 rootid) |
1940 | { | |
1941 | if (rootid == BTRFS_FS_TREE_OBJECTID || | |
e09fe2d2 QW |
1942 | ((s64)rootid >= (s64)BTRFS_FIRST_FREE_OBJECTID && |
1943 | !btrfs_qgroup_level(rootid))) | |
95a06077 JS |
1944 | return 1; |
1945 | return 0; | |
1946 | } | |
210549eb DS |
1947 | |
1948 | static inline int btrfs_defrag_cancelled(struct btrfs_fs_info *fs_info) | |
1949 | { | |
1950 | return signal_pending(current); | |
1951 | } | |
1952 | ||
14605409 BB |
1953 | /* verity.c */ |
1954 | #ifdef CONFIG_FS_VERITY | |
1955 | ||
1956 | extern const struct fsverity_operations btrfs_verityops; | |
1957 | int btrfs_drop_verity_items(struct btrfs_inode *inode); | |
38622010 | 1958 | int btrfs_get_verity_descriptor(struct inode *inode, void *buf, size_t buf_size); |
14605409 | 1959 | |
14605409 BB |
1960 | #else |
1961 | ||
1962 | static inline int btrfs_drop_verity_items(struct btrfs_inode *inode) | |
1963 | { | |
1964 | return 0; | |
1965 | } | |
1966 | ||
38622010 BB |
1967 | static inline int btrfs_get_verity_descriptor(struct inode *inode, void *buf, |
1968 | size_t buf_size) | |
1969 | { | |
1970 | return -EPERM; | |
1971 | } | |
1972 | ||
14605409 BB |
1973 | #endif |
1974 | ||
aaedb55b JB |
1975 | /* Sanity test specific functions */ |
1976 | #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS | |
1977 | void btrfs_test_destroy_inode(struct inode *inode); | |
b2fa1154 | 1978 | #endif |
9888c340 | 1979 | |
37f00a6d JT |
1980 | static inline bool btrfs_is_data_reloc_root(const struct btrfs_root *root) |
1981 | { | |
1982 | return root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID; | |
1983 | } | |
1984 | ||
f57ad937 QW |
1985 | /* |
1986 | * We use page status Private2 to indicate there is an ordered extent with | |
1987 | * unfinished IO. | |
1988 | * | |
1989 | * Rename the Private2 accessors to Ordered, to improve readability. | |
1990 | */ | |
1991 | #define PageOrdered(page) PagePrivate2(page) | |
1992 | #define SetPageOrdered(page) SetPagePrivate2(page) | |
1993 | #define ClearPageOrdered(page) ClearPagePrivate2(page) | |
895586eb MWO |
1994 | #define folio_test_ordered(folio) folio_test_private_2(folio) |
1995 | #define folio_set_ordered(folio) folio_set_private_2(folio) | |
1996 | #define folio_clear_ordered(folio) folio_clear_private_2(folio) | |
f57ad937 | 1997 | |
eb60ceac | 1998 | #endif |