1 /* SPDX-License-Identifier: GPL-2.0 */
5 * Copyright (c) 2012 Samsung Electronics Co., Ltd.
6 * http://www.samsung.com/
11 #include <linux/uio.h>
12 #include <linux/types.h>
13 #include <linux/page-flags.h>
14 #include <linux/buffer_head.h>
15 #include <linux/slab.h>
16 #include <linux/crc32.h>
17 #include <linux/magic.h>
18 #include <linux/kobject.h>
19 #include <linux/sched.h>
20 #include <linux/cred.h>
21 #include <linux/sched/mm.h>
22 #include <linux/vmalloc.h>
23 #include <linux/bio.h>
24 #include <linux/blkdev.h>
25 #include <linux/quotaops.h>
26 #include <linux/part_stat.h>
27 #include <crypto/hash.h>
29 #include <linux/fscrypt.h>
30 #include <linux/fsverity.h>
34 #ifdef CONFIG_F2FS_CHECK_FS
35 #define f2fs_bug_on(sbi, condition) BUG_ON(condition)
37 #define f2fs_bug_on(sbi, condition) \
39 if (WARN_ON(condition)) \
40 set_sbi_flag(sbi, SBI_NEED_FSCK); \
49 FAULT_ALLOC_BIO, /* it's obsolete due to bio_alloc() will never fail */
63 FAULT_BLKADDR_VALIDITY,
64 FAULT_BLKADDR_CONSISTENCE,
69 #ifdef CONFIG_F2FS_FAULT_INJECTION
70 #define F2FS_ALL_FAULT_TYPE (GENMASK(FAULT_MAX - 1, 0))
72 struct f2fs_fault_info {
74 unsigned int inject_rate;
75 unsigned int inject_type;
78 extern const char *f2fs_fault_name[FAULT_MAX];
79 #define IS_FAULT_SET(fi, type) ((fi)->inject_type & BIT(type))
81 /* maximum retry count for injected failure */
82 #define DEFAULT_FAILURE_RETRY_COUNT 8
84 #define DEFAULT_FAILURE_RETRY_COUNT 1
90 #define F2FS_MOUNT_DISABLE_ROLL_FORWARD 0x00000001
91 #define F2FS_MOUNT_DISCARD 0x00000002
92 #define F2FS_MOUNT_NOHEAP 0x00000004
93 #define F2FS_MOUNT_XATTR_USER 0x00000008
94 #define F2FS_MOUNT_POSIX_ACL 0x00000010
95 #define F2FS_MOUNT_DISABLE_EXT_IDENTIFY 0x00000020
96 #define F2FS_MOUNT_INLINE_XATTR 0x00000040
97 #define F2FS_MOUNT_INLINE_DATA 0x00000080
98 #define F2FS_MOUNT_INLINE_DENTRY 0x00000100
99 #define F2FS_MOUNT_FLUSH_MERGE 0x00000200
100 #define F2FS_MOUNT_NOBARRIER 0x00000400
101 #define F2FS_MOUNT_FASTBOOT 0x00000800
102 #define F2FS_MOUNT_READ_EXTENT_CACHE 0x00001000
103 #define F2FS_MOUNT_DATA_FLUSH 0x00002000
104 #define F2FS_MOUNT_FAULT_INJECTION 0x00004000
105 #define F2FS_MOUNT_USRQUOTA 0x00008000
106 #define F2FS_MOUNT_GRPQUOTA 0x00010000
107 #define F2FS_MOUNT_PRJQUOTA 0x00020000
108 #define F2FS_MOUNT_QUOTA 0x00040000
109 #define F2FS_MOUNT_INLINE_XATTR_SIZE 0x00080000
110 #define F2FS_MOUNT_RESERVE_ROOT 0x00100000
111 #define F2FS_MOUNT_DISABLE_CHECKPOINT 0x00200000
112 #define F2FS_MOUNT_NORECOVERY 0x00400000
113 #define F2FS_MOUNT_ATGC 0x00800000
114 #define F2FS_MOUNT_MERGE_CHECKPOINT 0x01000000
115 #define F2FS_MOUNT_GC_MERGE 0x02000000
116 #define F2FS_MOUNT_COMPRESS_CACHE 0x04000000
117 #define F2FS_MOUNT_AGE_EXTENT_CACHE 0x08000000
119 #define F2FS_OPTION(sbi) ((sbi)->mount_opt)
120 #define clear_opt(sbi, option) (F2FS_OPTION(sbi).opt &= ~F2FS_MOUNT_##option)
121 #define set_opt(sbi, option) (F2FS_OPTION(sbi).opt |= F2FS_MOUNT_##option)
122 #define test_opt(sbi, option) (F2FS_OPTION(sbi).opt & F2FS_MOUNT_##option)
124 #define ver_after(a, b) (typecheck(unsigned long long, a) && \
125 typecheck(unsigned long long, b) && \
126 ((long long)((a) - (b)) > 0))
128 typedef u32 block_t; /*
129 * should not change u32, since it is the on-disk block
130 * address format, __le32.
134 #define COMPRESS_EXT_NUM 16
137 * An implementation of an rwsem that is explicitly unfair to readers. This
138 * prevents priority inversion when a low-priority reader acquires the read lock
139 * while sleeping on the write lock but the write lock is needed by
140 * higher-priority clients.
144 struct rw_semaphore internal_rwsem;
145 #ifdef CONFIG_F2FS_UNFAIR_RWSEM
146 wait_queue_head_t read_waiters;
150 struct f2fs_mount_info {
152 block_t root_reserved_blocks; /* root reserved blocks */
153 kuid_t s_resuid; /* reserved blocks for uid */
154 kgid_t s_resgid; /* reserved blocks for gid */
155 int active_logs; /* # of active logs */
156 int inline_xattr_size; /* inline xattr size */
157 #ifdef CONFIG_F2FS_FAULT_INJECTION
158 struct f2fs_fault_info fault_info; /* For fault injection */
161 /* Names of quota files with journalled quota */
162 char *s_qf_names[MAXQUOTAS];
163 int s_jquota_fmt; /* Format of quota to use */
165 /* For which write hints are passed down to block layer */
166 int alloc_mode; /* segment allocation policy */
167 int fsync_mode; /* fsync policy */
168 int fs_mode; /* fs mode: LFS or ADAPTIVE */
169 int bggc_mode; /* bggc mode: off, on or sync */
170 int memory_mode; /* memory mode */
171 int errors; /* errors parameter */
173 * discard command's offset/size should
174 * be aligned to this unit: block,
177 struct fscrypt_dummy_policy dummy_enc_policy; /* test dummy encryption */
178 block_t unusable_cap_perc; /* percentage for cap */
179 block_t unusable_cap; /* Amount of space allowed to be
180 * unusable when disabling checkpoint
183 /* For compression */
184 unsigned char compress_algorithm; /* algorithm type */
185 unsigned char compress_log_size; /* cluster log size */
186 unsigned char compress_level; /* compress level */
187 bool compress_chksum; /* compressed data chksum */
188 unsigned char compress_ext_cnt; /* extension count */
189 unsigned char nocompress_ext_cnt; /* nocompress extension count */
190 int compress_mode; /* compression mode */
191 unsigned char extensions[COMPRESS_EXT_NUM][F2FS_EXTENSION_LEN]; /* extensions */
192 unsigned char noextensions[COMPRESS_EXT_NUM][F2FS_EXTENSION_LEN]; /* extensions */
195 #define F2FS_FEATURE_ENCRYPT 0x00000001
196 #define F2FS_FEATURE_BLKZONED 0x00000002
197 #define F2FS_FEATURE_ATOMIC_WRITE 0x00000004
198 #define F2FS_FEATURE_EXTRA_ATTR 0x00000008
199 #define F2FS_FEATURE_PRJQUOTA 0x00000010
200 #define F2FS_FEATURE_INODE_CHKSUM 0x00000020
201 #define F2FS_FEATURE_FLEXIBLE_INLINE_XATTR 0x00000040
202 #define F2FS_FEATURE_QUOTA_INO 0x00000080
203 #define F2FS_FEATURE_INODE_CRTIME 0x00000100
204 #define F2FS_FEATURE_LOST_FOUND 0x00000200
205 #define F2FS_FEATURE_VERITY 0x00000400
206 #define F2FS_FEATURE_SB_CHKSUM 0x00000800
207 #define F2FS_FEATURE_CASEFOLD 0x00001000
208 #define F2FS_FEATURE_COMPRESSION 0x00002000
209 #define F2FS_FEATURE_RO 0x00004000
211 #define __F2FS_HAS_FEATURE(raw_super, mask) \
212 ((raw_super->feature & cpu_to_le32(mask)) != 0)
213 #define F2FS_HAS_FEATURE(sbi, mask) __F2FS_HAS_FEATURE(sbi->raw_super, mask)
216 * Default values for user and/or group using reserved blocks
218 #define F2FS_DEF_RESUID 0
219 #define F2FS_DEF_RESGID 0
222 * For checkpoint manager
229 #define CP_UMOUNT 0x00000001
230 #define CP_FASTBOOT 0x00000002
231 #define CP_SYNC 0x00000004
232 #define CP_RECOVERY 0x00000008
233 #define CP_DISCARD 0x00000010
234 #define CP_TRIMMED 0x00000020
235 #define CP_PAUSE 0x00000040
236 #define CP_RESIZE 0x00000080
238 #define DEF_MAX_DISCARD_REQUEST 8 /* issue 8 discards per round */
239 #define DEF_MIN_DISCARD_ISSUE_TIME 50 /* 50 ms, if exists */
240 #define DEF_MID_DISCARD_ISSUE_TIME 500 /* 500 ms, if device busy */
241 #define DEF_MAX_DISCARD_ISSUE_TIME 60000 /* 60 s, if no candidates */
242 #define DEF_DISCARD_URGENT_UTIL 80 /* do more discard over 80% */
243 #define DEF_CP_INTERVAL 60 /* 60 secs */
244 #define DEF_IDLE_INTERVAL 5 /* 5 secs */
245 #define DEF_DISABLE_INTERVAL 5 /* 5 secs */
246 #define DEF_DISABLE_QUICK_INTERVAL 1 /* 1 secs */
247 #define DEF_UMOUNT_DISCARD_TIMEOUT 5 /* 5 secs */
257 * indicate meta/data type
266 DATA_GENERIC, /* check range only */
267 DATA_GENERIC_ENHANCE, /* strong check on range and segment bitmap */
268 DATA_GENERIC_ENHANCE_READ, /*
269 * strong check on range and segment
270 * bitmap but no warning due to race
271 * condition of read on truncated area
274 DATA_GENERIC_ENHANCE_UPDATE, /*
275 * strong check on range and segment
276 * bitmap for update case
281 /* for the list of ino */
283 ORPHAN_INO, /* for orphan ino list */
284 APPEND_INO, /* for append ino list */
285 UPDATE_INO, /* for update ino list */
286 TRANS_DIR_INO, /* for transactions dir ino list */
287 FLUSH_INO, /* for multiple device flushing */
288 MAX_INO_ENTRY, /* max. list */
292 struct list_head list; /* list head */
293 nid_t ino; /* inode number */
294 unsigned int dirty_device; /* dirty device bitmap */
297 /* for the list of inodes to be GCed */
299 struct list_head list; /* list head */
300 struct inode *inode; /* vfs inode pointer */
303 struct fsync_node_entry {
304 struct list_head list; /* list head */
305 struct page *page; /* warm node page pointer */
306 unsigned int seq_id; /* sequence id */
310 struct completion wait; /* completion for checkpoint done */
311 struct llist_node llnode; /* llist_node to be linked in wait queue */
312 int ret; /* return code of checkpoint */
313 ktime_t queue_time; /* request queued time */
316 struct ckpt_req_control {
317 struct task_struct *f2fs_issue_ckpt; /* checkpoint task */
318 int ckpt_thread_ioprio; /* checkpoint merge thread ioprio */
319 wait_queue_head_t ckpt_wait_queue; /* waiting queue for wake-up */
320 atomic_t issued_ckpt; /* # of actually issued ckpts */
321 atomic_t total_ckpt; /* # of total ckpts */
322 atomic_t queued_ckpt; /* # of queued ckpts */
323 struct llist_head issue_list; /* list for command issue */
324 spinlock_t stat_lock; /* lock for below checkpoint time stats */
325 unsigned int cur_time; /* cur wait time in msec for currently issued checkpoint */
326 unsigned int peak_time; /* peak wait time in msec until now */
329 /* for the bitmap indicate blocks to be discarded */
330 struct discard_entry {
331 struct list_head list; /* list head */
332 block_t start_blkaddr; /* start blockaddr of current segment */
333 unsigned char discard_map[SIT_VBLOCK_MAP_SIZE]; /* segment discard bitmap */
336 /* minimum discard granularity, unit: block count */
337 #define MIN_DISCARD_GRANULARITY 1
338 /* default discard granularity of inner discard thread, unit: block count */
339 #define DEFAULT_DISCARD_GRANULARITY 16
340 /* default maximum discard granularity of ordered discard, unit: block count */
341 #define DEFAULT_MAX_ORDERED_DISCARD_GRANULARITY 16
343 /* max discard pend list number */
344 #define MAX_PLIST_NUM 512
345 #define plist_idx(blk_num) ((blk_num) >= MAX_PLIST_NUM ? \
346 (MAX_PLIST_NUM - 1) : ((blk_num) - 1))
349 D_PREP, /* initial */
350 D_PARTIAL, /* partially submitted */
351 D_SUBMIT, /* all submitted */
352 D_DONE, /* finished */
355 struct discard_info {
356 block_t lstart; /* logical start address */
357 block_t len; /* length */
358 block_t start; /* actual start address in dev */
362 struct rb_node rb_node; /* rb node located in rb-tree */
363 struct discard_info di; /* discard info */
364 struct list_head list; /* command list */
365 struct completion wait; /* compleation */
366 struct block_device *bdev; /* bdev */
367 unsigned short ref; /* reference count */
368 unsigned char state; /* state */
369 unsigned char queued; /* queued discard */
370 int error; /* bio error */
371 spinlock_t lock; /* for state/bio_ref updating */
372 unsigned short bio_ref; /* bio reference count */
384 DPOLICY_IO_AWARE_DISABLE, /* force to not be aware of IO */
385 DPOLICY_IO_AWARE_ENABLE, /* force to be aware of IO */
386 DPOLICY_IO_AWARE_MAX,
389 struct discard_policy {
390 int type; /* type of discard */
391 unsigned int min_interval; /* used for candidates exist */
392 unsigned int mid_interval; /* used for device busy */
393 unsigned int max_interval; /* used for candidates not exist */
394 unsigned int max_requests; /* # of discards issued per round */
395 unsigned int io_aware_gran; /* minimum granularity discard not be aware of I/O */
396 bool io_aware; /* issue discard in idle time */
397 bool sync; /* submit discard with REQ_SYNC flag */
398 bool ordered; /* issue discard by lba order */
399 bool timeout; /* discard timeout for put_super */
400 unsigned int granularity; /* discard granularity */
403 struct discard_cmd_control {
404 struct task_struct *f2fs_issue_discard; /* discard thread */
405 struct list_head entry_list; /* 4KB discard entry list */
406 struct list_head pend_list[MAX_PLIST_NUM];/* store pending entries */
407 struct list_head wait_list; /* store on-flushing entries */
408 struct list_head fstrim_list; /* in-flight discard from fstrim */
409 wait_queue_head_t discard_wait_queue; /* waiting queue for wake-up */
410 struct mutex cmd_lock;
411 unsigned int nr_discards; /* # of discards in the list */
412 unsigned int max_discards; /* max. discards to be issued */
413 unsigned int max_discard_request; /* max. discard request per round */
414 unsigned int min_discard_issue_time; /* min. interval between discard issue */
415 unsigned int mid_discard_issue_time; /* mid. interval between discard issue */
416 unsigned int max_discard_issue_time; /* max. interval between discard issue */
417 unsigned int discard_io_aware_gran; /* minimum discard granularity not be aware of I/O */
418 unsigned int discard_urgent_util; /* utilization which issue discard proactively */
419 unsigned int discard_granularity; /* discard granularity */
420 unsigned int max_ordered_discard; /* maximum discard granularity issued by lba order */
421 unsigned int discard_io_aware; /* io_aware policy */
422 unsigned int undiscard_blks; /* # of undiscard blocks */
423 unsigned int next_pos; /* next discard position */
424 atomic_t issued_discard; /* # of issued discard */
425 atomic_t queued_discard; /* # of queued discard */
426 atomic_t discard_cmd_cnt; /* # of cached cmd count */
427 struct rb_root_cached root; /* root of discard rb-tree */
428 bool rbtree_check; /* config for consistence check */
429 bool discard_wake; /* to wake up discard thread */
432 /* for the list of fsync inodes, used only during recovery */
433 struct fsync_inode_entry {
434 struct list_head list; /* list head */
435 struct inode *inode; /* vfs inode pointer */
436 block_t blkaddr; /* block address locating the last fsync */
437 block_t last_dentry; /* block address locating the last dentry */
440 #define nats_in_cursum(jnl) (le16_to_cpu((jnl)->n_nats))
441 #define sits_in_cursum(jnl) (le16_to_cpu((jnl)->n_sits))
443 #define nat_in_journal(jnl, i) ((jnl)->nat_j.entries[i].ne)
444 #define nid_in_journal(jnl, i) ((jnl)->nat_j.entries[i].nid)
445 #define sit_in_journal(jnl, i) ((jnl)->sit_j.entries[i].se)
446 #define segno_in_journal(jnl, i) ((jnl)->sit_j.entries[i].segno)
448 #define MAX_NAT_JENTRIES(jnl) (NAT_JOURNAL_ENTRIES - nats_in_cursum(jnl))
449 #define MAX_SIT_JENTRIES(jnl) (SIT_JOURNAL_ENTRIES - sits_in_cursum(jnl))
451 static inline int update_nats_in_cursum(struct f2fs_journal *journal, int i)
453 int before = nats_in_cursum(journal);
455 journal->n_nats = cpu_to_le16(before + i);
459 static inline int update_sits_in_cursum(struct f2fs_journal *journal, int i)
461 int before = sits_in_cursum(journal);
463 journal->n_sits = cpu_to_le16(before + i);
467 static inline bool __has_cursum_space(struct f2fs_journal *journal,
470 if (type == NAT_JOURNAL)
471 return size <= MAX_NAT_JENTRIES(journal);
472 return size <= MAX_SIT_JENTRIES(journal);
475 /* for inline stuff */
476 #define DEF_INLINE_RESERVED_SIZE 1
477 static inline int get_extra_isize(struct inode *inode);
478 static inline int get_inline_xattr_addrs(struct inode *inode);
479 #define MAX_INLINE_DATA(inode) (sizeof(__le32) * \
480 (CUR_ADDRS_PER_INODE(inode) - \
481 get_inline_xattr_addrs(inode) - \
482 DEF_INLINE_RESERVED_SIZE))
485 #define NR_INLINE_DENTRY(inode) (MAX_INLINE_DATA(inode) * BITS_PER_BYTE / \
486 ((SIZE_OF_DIR_ENTRY + F2FS_SLOT_LEN) * \
488 #define INLINE_DENTRY_BITMAP_SIZE(inode) \
489 DIV_ROUND_UP(NR_INLINE_DENTRY(inode), BITS_PER_BYTE)
490 #define INLINE_RESERVED_SIZE(inode) (MAX_INLINE_DATA(inode) - \
491 ((SIZE_OF_DIR_ENTRY + F2FS_SLOT_LEN) * \
492 NR_INLINE_DENTRY(inode) + \
493 INLINE_DENTRY_BITMAP_SIZE(inode)))
496 * For INODE and NODE manager
498 /* for directory operations */
500 struct f2fs_filename {
502 * The filename the user specified. This is NULL for some
503 * filesystem-internal operations, e.g. converting an inline directory
504 * to a non-inline one, or roll-forward recovering an encrypted dentry.
506 const struct qstr *usr_fname;
509 * The on-disk filename. For encrypted directories, this is encrypted.
510 * This may be NULL for lookups in an encrypted dir without the key.
512 struct fscrypt_str disk_name;
514 /* The dirhash of this filename */
517 #ifdef CONFIG_FS_ENCRYPTION
519 * For lookups in encrypted directories: either the buffer backing
520 * disk_name, or a buffer that holds the decoded no-key name.
522 struct fscrypt_str crypto_buf;
524 #if IS_ENABLED(CONFIG_UNICODE)
526 * For casefolded directories: the casefolded name, but it's left NULL
527 * if the original name is not valid Unicode, if the original name is
528 * "." or "..", if the directory is both casefolded and encrypted and
529 * its encryption key is unavailable, or if the filesystem is doing an
530 * internal operation where usr_fname is also NULL. In all these cases
531 * we fall back to treating the name as an opaque byte sequence.
533 struct fscrypt_str cf_name;
537 struct f2fs_dentry_ptr {
540 struct f2fs_dir_entry *dentry;
541 __u8 (*filename)[F2FS_SLOT_LEN];
546 static inline void make_dentry_ptr_block(struct inode *inode,
547 struct f2fs_dentry_ptr *d, struct f2fs_dentry_block *t)
550 d->max = NR_DENTRY_IN_BLOCK;
551 d->nr_bitmap = SIZE_OF_DENTRY_BITMAP;
552 d->bitmap = t->dentry_bitmap;
553 d->dentry = t->dentry;
554 d->filename = t->filename;
557 static inline void make_dentry_ptr_inline(struct inode *inode,
558 struct f2fs_dentry_ptr *d, void *t)
560 int entry_cnt = NR_INLINE_DENTRY(inode);
561 int bitmap_size = INLINE_DENTRY_BITMAP_SIZE(inode);
562 int reserved_size = INLINE_RESERVED_SIZE(inode);
566 d->nr_bitmap = bitmap_size;
568 d->dentry = t + bitmap_size + reserved_size;
569 d->filename = t + bitmap_size + reserved_size +
570 SIZE_OF_DIR_ENTRY * entry_cnt;
574 * XATTR_NODE_OFFSET stores xattrs to one node block per file keeping -1
575 * as its node offset to distinguish from index node blocks.
576 * But some bits are used to mark the node block.
578 #define XATTR_NODE_OFFSET ((((unsigned int)-1) << OFFSET_BIT_SHIFT) \
581 ALLOC_NODE, /* allocate a new node page if needed */
582 LOOKUP_NODE, /* look up a node without readahead */
584 * look up a node with readahead called
589 #define DEFAULT_RETRY_IO_COUNT 8 /* maximum retry read IO or flush count */
591 /* congestion wait timeout value, default: 20ms */
592 #define DEFAULT_IO_TIMEOUT (msecs_to_jiffies(20))
594 /* maximum retry quota flush count */
595 #define DEFAULT_RETRY_QUOTA_FLUSH_COUNT 8
597 /* maximum retry of EIO'ed page */
598 #define MAX_RETRY_PAGE_EIO 100
600 #define F2FS_LINK_MAX 0xffffffff /* maximum link count per file */
602 #define MAX_DIR_RA_PAGES 4 /* maximum ra pages of dir */
604 /* dirty segments threshold for triggering CP */
605 #define DEFAULT_DIRTY_THRESHOLD 4
607 #define RECOVERY_MAX_RA_BLOCKS BIO_MAX_VECS
608 #define RECOVERY_MIN_RA_BLOCKS 1
610 #define F2FS_ONSTACK_PAGES 16 /* nr of onstack pages */
612 /* for in-memory extent cache entry */
613 #define F2FS_MIN_EXTENT_LEN 64 /* minimum extent length */
615 /* number of extent info in extent cache we try to shrink */
616 #define READ_EXTENT_CACHE_SHRINK_NUMBER 128
618 /* number of age extent info in extent cache we try to shrink */
619 #define AGE_EXTENT_CACHE_SHRINK_NUMBER 128
620 #define LAST_AGE_WEIGHT 30
621 #define SAME_AGE_REGION 1024
624 * Define data block with age less than 1GB as hot data
625 * define data block with age less than 10GB but more than 1GB as warm data
627 #define DEF_HOT_DATA_AGE_THRESHOLD 262144
628 #define DEF_WARM_DATA_AGE_THRESHOLD 2621440
630 /* extent cache type */
638 unsigned int fofs; /* start offset in a file */
639 unsigned int len; /* length of the extent */
641 /* read extent_cache */
643 /* start block address of the extent */
645 #ifdef CONFIG_F2FS_FS_COMPRESSION
646 /* physical extent length of compressed blocks */
650 /* block age extent_cache */
652 /* block age of the extent */
653 unsigned long long age;
654 /* last total blocks allocated */
655 unsigned long long last_blocks;
661 struct rb_node rb_node; /* rb node located in rb-tree */
662 struct extent_info ei; /* extent info */
663 struct list_head list; /* node in global extent list of sbi */
664 struct extent_tree *et; /* extent tree pointer */
668 nid_t ino; /* inode number */
669 enum extent_type type; /* keep the extent tree type */
670 struct rb_root_cached root; /* root of extent info rb-tree */
671 struct extent_node *cached_en; /* recently accessed extent node */
672 struct list_head list; /* to be used by sbi->zombie_list */
673 rwlock_t lock; /* protect extent info rb-tree */
674 atomic_t node_cnt; /* # of extent node in rb-tree*/
675 bool largest_updated; /* largest extent updated */
676 struct extent_info largest; /* largest cached extent for EX_READ */
679 struct extent_tree_info {
680 struct radix_tree_root extent_tree_root;/* cache extent cache entries */
681 struct mutex extent_tree_lock; /* locking extent radix tree */
682 struct list_head extent_list; /* lru list for shrinker */
683 spinlock_t extent_lock; /* locking extent lru list */
684 atomic_t total_ext_tree; /* extent tree count */
685 struct list_head zombie_list; /* extent zombie tree list */
686 atomic_t total_zombie_tree; /* extent zombie tree count */
687 atomic_t total_ext_node; /* extent info count */
691 * State of block returned by f2fs_map_blocks.
693 #define F2FS_MAP_NEW (1U << 0)
694 #define F2FS_MAP_MAPPED (1U << 1)
695 #define F2FS_MAP_DELALLOC (1U << 2)
696 #define F2FS_MAP_FLAGS (F2FS_MAP_NEW | F2FS_MAP_MAPPED |\
699 struct f2fs_map_blocks {
700 struct block_device *m_bdev; /* for multi-device dio */
704 unsigned int m_flags;
705 pgoff_t *m_next_pgofs; /* point next possible non-hole pgofs */
706 pgoff_t *m_next_extent; /* point to next possible extent */
708 bool m_may_create; /* indicate it is from write path */
709 bool m_multidev_dio; /* indicate it allows multi-device dio */
712 /* for flag in get_data_block */
714 F2FS_GET_BLOCK_DEFAULT,
715 F2FS_GET_BLOCK_FIEMAP,
718 F2FS_GET_BLOCK_PRE_DIO,
719 F2FS_GET_BLOCK_PRE_AIO,
720 F2FS_GET_BLOCK_PRECACHE,
724 * i_advise uses FADVISE_XXX_BIT. We can add additional hints later.
726 #define FADVISE_COLD_BIT 0x01
727 #define FADVISE_LOST_PINO_BIT 0x02
728 #define FADVISE_ENCRYPT_BIT 0x04
729 #define FADVISE_ENC_NAME_BIT 0x08
730 #define FADVISE_KEEP_SIZE_BIT 0x10
731 #define FADVISE_HOT_BIT 0x20
732 #define FADVISE_VERITY_BIT 0x40
733 #define FADVISE_TRUNC_BIT 0x80
735 #define FADVISE_MODIFIABLE_BITS (FADVISE_COLD_BIT | FADVISE_HOT_BIT)
737 #define file_is_cold(inode) is_file(inode, FADVISE_COLD_BIT)
738 #define file_set_cold(inode) set_file(inode, FADVISE_COLD_BIT)
739 #define file_clear_cold(inode) clear_file(inode, FADVISE_COLD_BIT)
741 #define file_wrong_pino(inode) is_file(inode, FADVISE_LOST_PINO_BIT)
742 #define file_lost_pino(inode) set_file(inode, FADVISE_LOST_PINO_BIT)
743 #define file_got_pino(inode) clear_file(inode, FADVISE_LOST_PINO_BIT)
745 #define file_is_encrypt(inode) is_file(inode, FADVISE_ENCRYPT_BIT)
746 #define file_set_encrypt(inode) set_file(inode, FADVISE_ENCRYPT_BIT)
748 #define file_enc_name(inode) is_file(inode, FADVISE_ENC_NAME_BIT)
749 #define file_set_enc_name(inode) set_file(inode, FADVISE_ENC_NAME_BIT)
751 #define file_keep_isize(inode) is_file(inode, FADVISE_KEEP_SIZE_BIT)
752 #define file_set_keep_isize(inode) set_file(inode, FADVISE_KEEP_SIZE_BIT)
754 #define file_is_hot(inode) is_file(inode, FADVISE_HOT_BIT)
755 #define file_set_hot(inode) set_file(inode, FADVISE_HOT_BIT)
756 #define file_clear_hot(inode) clear_file(inode, FADVISE_HOT_BIT)
758 #define file_is_verity(inode) is_file(inode, FADVISE_VERITY_BIT)
759 #define file_set_verity(inode) set_file(inode, FADVISE_VERITY_BIT)
761 #define file_should_truncate(inode) is_file(inode, FADVISE_TRUNC_BIT)
762 #define file_need_truncate(inode) set_file(inode, FADVISE_TRUNC_BIT)
763 #define file_dont_truncate(inode) clear_file(inode, FADVISE_TRUNC_BIT)
765 #define DEF_DIR_LEVEL 0
772 /* used for f2fs_inode_info->flags */
774 FI_NEW_INODE, /* indicate newly allocated inode */
775 FI_DIRTY_INODE, /* indicate inode is dirty or not */
776 FI_AUTO_RECOVER, /* indicate inode is recoverable */
777 FI_DIRTY_DIR, /* indicate directory has dirty pages */
778 FI_INC_LINK, /* need to increment i_nlink */
779 FI_ACL_MODE, /* indicate acl mode */
780 FI_NO_ALLOC, /* should not allocate any blocks */
781 FI_FREE_NID, /* free allocated nide */
782 FI_NO_EXTENT, /* not to use the extent cache */
783 FI_INLINE_XATTR, /* used for inline xattr */
784 FI_INLINE_DATA, /* used for inline data*/
785 FI_INLINE_DENTRY, /* used for inline dentry */
786 FI_APPEND_WRITE, /* inode has appended data */
787 FI_UPDATE_WRITE, /* inode has in-place-update data */
788 FI_NEED_IPU, /* used for ipu per file */
789 FI_ATOMIC_FILE, /* indicate atomic file */
790 FI_DATA_EXIST, /* indicate data exists */
791 FI_INLINE_DOTS, /* indicate inline dot dentries */
792 FI_SKIP_WRITES, /* should skip data page writeback */
793 FI_OPU_WRITE, /* used for opu per file */
794 FI_DIRTY_FILE, /* indicate regular/symlink has dirty pages */
795 FI_PREALLOCATED_ALL, /* all blocks for write were preallocated */
796 FI_HOT_DATA, /* indicate file is hot */
797 FI_EXTRA_ATTR, /* indicate file has extra attribute */
798 FI_PROJ_INHERIT, /* indicate file inherits projectid */
799 FI_PIN_FILE, /* indicate file should not be gced */
800 FI_VERITY_IN_PROGRESS, /* building fs-verity Merkle tree */
801 FI_COMPRESSED_FILE, /* indicate file's data can be compressed */
802 FI_COMPRESS_CORRUPT, /* indicate compressed cluster is corrupted */
803 FI_MMAP_FILE, /* indicate file was mmapped */
804 FI_ENABLE_COMPRESS, /* enable compression in "user" compression mode */
805 FI_COMPRESS_RELEASED, /* compressed blocks were released */
806 FI_ALIGNED_WRITE, /* enable aligned write */
807 FI_COW_FILE, /* indicate COW file */
808 FI_ATOMIC_COMMITTED, /* indicate atomic commit completed except disk sync */
809 FI_ATOMIC_REPLACE, /* indicate atomic replace */
810 FI_MAX, /* max flag, never be used */
813 struct f2fs_inode_info {
814 struct inode vfs_inode; /* serve a vfs inode */
815 unsigned long i_flags; /* keep an inode flags for ioctl */
816 unsigned char i_advise; /* use to give file attribute hints */
817 unsigned char i_dir_level; /* use for dentry level for large dir */
818 unsigned int i_current_depth; /* only for directory depth */
819 /* for gc failure statistic */
820 unsigned int i_gc_failures[MAX_GC_FAILURE];
821 unsigned int i_pino; /* parent inode number */
822 umode_t i_acl_mode; /* keep file acl mode temporarily */
824 /* Use below internally in f2fs*/
825 unsigned long flags[BITS_TO_LONGS(FI_MAX)]; /* use to pass per-file flags */
826 struct f2fs_rwsem i_sem; /* protect fi info */
827 atomic_t dirty_pages; /* # of dirty pages */
828 f2fs_hash_t chash; /* hash value of given file name */
829 unsigned int clevel; /* maximum level of given file name */
830 struct task_struct *task; /* lookup and create consistency */
831 struct task_struct *cp_task; /* separate cp/wb IO stats*/
832 struct task_struct *wb_task; /* indicate inode is in context of writeback */
833 nid_t i_xattr_nid; /* node id that contains xattrs */
834 loff_t last_disk_size; /* lastly written file size */
835 spinlock_t i_size_lock; /* protect last_disk_size */
838 struct dquot *i_dquot[MAXQUOTAS];
840 /* quota space reservation, managed internally by quota code */
841 qsize_t i_reserved_quota;
843 struct list_head dirty_list; /* dirty list for dirs and files */
844 struct list_head gdirty_list; /* linked in global dirty list */
845 struct task_struct *atomic_write_task; /* store atomic write task */
846 struct extent_tree *extent_tree[NR_EXTENT_CACHES];
847 /* cached extent_tree entry */
848 struct inode *cow_inode; /* copy-on-write inode for atomic write */
850 /* avoid racing between foreground op and gc */
851 struct f2fs_rwsem i_gc_rwsem[2];
852 struct f2fs_rwsem i_xattr_sem; /* avoid racing between reading and changing EAs */
854 int i_extra_isize; /* size of extra space located in i_addr */
855 kprojid_t i_projid; /* id for project quota */
856 int i_inline_xattr_size; /* inline xattr size */
857 struct timespec64 i_crtime; /* inode creation time */
858 struct timespec64 i_disk_time[3];/* inode disk times */
860 /* for file compress */
861 atomic_t i_compr_blocks; /* # of compressed blocks */
862 unsigned char i_compress_algorithm; /* algorithm type */
863 unsigned char i_log_cluster_size; /* log of cluster size */
864 unsigned char i_compress_level; /* compress level (lz4hc,zstd) */
865 unsigned char i_compress_flag; /* compress flag */
866 unsigned int i_cluster_size; /* cluster size */
868 unsigned int atomic_write_cnt;
869 loff_t original_i_size; /* original i_size before atomic write */
872 static inline void get_read_extent_info(struct extent_info *ext,
873 struct f2fs_extent *i_ext)
875 ext->fofs = le32_to_cpu(i_ext->fofs);
876 ext->blk = le32_to_cpu(i_ext->blk);
877 ext->len = le32_to_cpu(i_ext->len);
880 static inline void set_raw_read_extent(struct extent_info *ext,
881 struct f2fs_extent *i_ext)
883 i_ext->fofs = cpu_to_le32(ext->fofs);
884 i_ext->blk = cpu_to_le32(ext->blk);
885 i_ext->len = cpu_to_le32(ext->len);
888 static inline bool __is_discard_mergeable(struct discard_info *back,
889 struct discard_info *front, unsigned int max_len)
891 return (back->lstart + back->len == front->lstart) &&
892 (back->len + front->len <= max_len);
895 static inline bool __is_discard_back_mergeable(struct discard_info *cur,
896 struct discard_info *back, unsigned int max_len)
898 return __is_discard_mergeable(back, cur, max_len);
901 static inline bool __is_discard_front_mergeable(struct discard_info *cur,
902 struct discard_info *front, unsigned int max_len)
904 return __is_discard_mergeable(cur, front, max_len);
908 * For free nid management
911 FREE_NID, /* newly added to free nid list */
912 PREALLOC_NID, /* it is preallocated */
923 struct f2fs_nm_info {
924 block_t nat_blkaddr; /* base disk address of NAT */
925 nid_t max_nid; /* maximum possible node ids */
926 nid_t available_nids; /* # of available node ids */
927 nid_t next_scan_nid; /* the next nid to be scanned */
928 nid_t max_rf_node_blocks; /* max # of nodes for recovery */
929 unsigned int ram_thresh; /* control the memory footprint */
930 unsigned int ra_nid_pages; /* # of nid pages to be readaheaded */
931 unsigned int dirty_nats_ratio; /* control dirty nats ratio threshold */
933 /* NAT cache management */
934 struct radix_tree_root nat_root;/* root of the nat entry cache */
935 struct radix_tree_root nat_set_root;/* root of the nat set cache */
936 struct f2fs_rwsem nat_tree_lock; /* protect nat entry tree */
937 struct list_head nat_entries; /* cached nat entry list (clean) */
938 spinlock_t nat_list_lock; /* protect clean nat entry list */
939 unsigned int nat_cnt[MAX_NAT_STATE]; /* the # of cached nat entries */
940 unsigned int nat_blocks; /* # of nat blocks */
942 /* free node ids management */
943 struct radix_tree_root free_nid_root;/* root of the free_nid cache */
944 struct list_head free_nid_list; /* list for free nids excluding preallocated nids */
945 unsigned int nid_cnt[MAX_NID_STATE]; /* the number of free node id */
946 spinlock_t nid_list_lock; /* protect nid lists ops */
947 struct mutex build_lock; /* lock for build free nids */
948 unsigned char **free_nid_bitmap;
949 unsigned char *nat_block_bitmap;
950 unsigned short *free_nid_count; /* free nid count of NAT block */
953 char *nat_bitmap; /* NAT bitmap pointer */
955 unsigned int nat_bits_blocks; /* # of nat bits blocks */
956 unsigned char *nat_bits; /* NAT bits blocks */
957 unsigned char *full_nat_bits; /* full NAT pages */
958 unsigned char *empty_nat_bits; /* empty NAT pages */
959 #ifdef CONFIG_F2FS_CHECK_FS
960 char *nat_bitmap_mir; /* NAT bitmap mirror */
962 int bitmap_size; /* bitmap size */
966 * this structure is used as one of function parameters.
967 * all the information are dedicated to a given direct node block determined
968 * by the data offset in a file.
970 struct dnode_of_data {
971 struct inode *inode; /* vfs inode pointer */
972 struct page *inode_page; /* its inode page, NULL is possible */
973 struct page *node_page; /* cached direct node page */
974 nid_t nid; /* node id of the direct node block */
975 unsigned int ofs_in_node; /* data offset in the node page */
976 bool inode_page_locked; /* inode page is locked or not */
977 bool node_changed; /* is node block changed */
978 char cur_level; /* level of hole node page */
979 char max_level; /* level of current page located */
980 block_t data_blkaddr; /* block address of the node block */
983 static inline void set_new_dnode(struct dnode_of_data *dn, struct inode *inode,
984 struct page *ipage, struct page *npage, nid_t nid)
986 memset(dn, 0, sizeof(*dn));
988 dn->inode_page = ipage;
989 dn->node_page = npage;
996 * By default, there are 6 active log areas across the whole main area.
997 * When considering hot and cold data separation to reduce cleaning overhead,
998 * we split 3 for data logs and 3 for node logs as hot, warm, and cold types,
1000 * In the current design, you should not change the numbers intentionally.
1001 * Instead, as a mount option such as active_logs=x, you can use 2, 4, and 6
1002 * logs individually according to the underlying devices. (default: 6)
1003 * Just in case, on-disk layout covers maximum 16 logs that consist of 8 for
1004 * data and 8 for node logs.
1006 #define NR_CURSEG_DATA_TYPE (3)
1007 #define NR_CURSEG_NODE_TYPE (3)
1008 #define NR_CURSEG_INMEM_TYPE (2)
1009 #define NR_CURSEG_RO_TYPE (2)
1010 #define NR_CURSEG_PERSIST_TYPE (NR_CURSEG_DATA_TYPE + NR_CURSEG_NODE_TYPE)
1011 #define NR_CURSEG_TYPE (NR_CURSEG_INMEM_TYPE + NR_CURSEG_PERSIST_TYPE)
1014 CURSEG_HOT_DATA = 0, /* directory entry blocks */
1015 CURSEG_WARM_DATA, /* data blocks */
1016 CURSEG_COLD_DATA, /* multimedia or GCed data blocks */
1017 CURSEG_HOT_NODE, /* direct node blocks of directory files */
1018 CURSEG_WARM_NODE, /* direct node blocks of normal files */
1019 CURSEG_COLD_NODE, /* indirect node blocks */
1020 NR_PERSISTENT_LOG, /* number of persistent log */
1021 CURSEG_COLD_DATA_PINNED = NR_PERSISTENT_LOG,
1022 /* pinned file that needs consecutive block address */
1023 CURSEG_ALL_DATA_ATGC, /* SSR alloctor in hot/warm/cold data area */
1024 NO_CHECK_TYPE, /* number of persistent & inmem log */
1028 struct completion wait;
1029 struct llist_node llnode;
1034 struct flush_cmd_control {
1035 struct task_struct *f2fs_issue_flush; /* flush thread */
1036 wait_queue_head_t flush_wait_queue; /* waiting queue for wake-up */
1037 atomic_t issued_flush; /* # of issued flushes */
1038 atomic_t queued_flush; /* # of queued flushes */
1039 struct llist_head issue_list; /* list for command issue */
1040 struct llist_node *dispatch_list; /* list for command dispatch */
1043 struct f2fs_sm_info {
1044 struct sit_info *sit_info; /* whole segment information */
1045 struct free_segmap_info *free_info; /* free segment information */
1046 struct dirty_seglist_info *dirty_info; /* dirty segment information */
1047 struct curseg_info *curseg_array; /* active segment information */
1049 struct f2fs_rwsem curseg_lock; /* for preventing curseg change */
1051 block_t seg0_blkaddr; /* block address of 0'th segment */
1052 block_t main_blkaddr; /* start block address of main area */
1053 block_t ssa_blkaddr; /* start block address of SSA area */
1055 unsigned int segment_count; /* total # of segments */
1056 unsigned int main_segments; /* # of segments in main area */
1057 unsigned int reserved_segments; /* # of reserved segments */
1058 unsigned int additional_reserved_segments;/* reserved segs for IO align feature */
1059 unsigned int ovp_segments; /* # of overprovision segments */
1061 /* a threshold to reclaim prefree segments */
1062 unsigned int rec_prefree_segments;
1064 struct list_head sit_entry_set; /* sit entry set list */
1066 unsigned int ipu_policy; /* in-place-update policy */
1067 unsigned int min_ipu_util; /* in-place-update threshold */
1068 unsigned int min_fsync_blocks; /* threshold for fsync */
1069 unsigned int min_seq_blocks; /* threshold for sequential blocks */
1070 unsigned int min_hot_blocks; /* threshold for hot block allocation */
1071 unsigned int min_ssr_sections; /* threshold to trigger SSR allocation */
1073 /* for flush command control */
1074 struct flush_cmd_control *fcc_info;
1076 /* for discard command control */
1077 struct discard_cmd_control *dcc_info;
1084 * COUNT_TYPE for monitoring
1086 * f2fs monitors the number of several block types such as on-writeback,
1087 * dirty dentry blocks, dirty node blocks, and dirty meta blocks.
1089 #define WB_DATA_TYPE(p, f) \
1090 (f || f2fs_is_cp_guaranteed(p) ? F2FS_WB_CP_DATA : F2FS_WB_DATA)
1109 * The below are the page types of bios used in submit_bio().
1110 * The available types are:
1111 * DATA User data pages. It operates as async mode.
1112 * NODE Node pages. It operates as async mode.
1113 * META FS metadata pages such as SIT, NAT, CP.
1114 * NR_PAGE_TYPE The number of page types.
1115 * META_FLUSH Make sure the previous pages are written
1116 * with waiting the bio's completion
1117 * ... Only can be used with META.
1119 #define PAGE_TYPE_OF_BIO(type) ((type) > META ? META : (type))
1120 #define PAGE_TYPE_ON_MAIN(type) ((type) == DATA || (type) == NODE)
1123 NODE = 1, /* should not change this */
1127 IPU, /* the below types are used by tracepoints only. */
1132 HOT = 0, /* must be zero for meta bio */
1138 enum need_lock_type {
1144 enum cp_reason_type {
1160 APP_DIRECT_IO, /* app direct write IOs */
1161 APP_BUFFERED_IO, /* app buffered write IOs */
1162 APP_WRITE_IO, /* app write IOs */
1163 APP_MAPPED_IO, /* app mapped IOs */
1164 APP_BUFFERED_CDATA_IO, /* app buffered write IOs on compressed file */
1165 APP_MAPPED_CDATA_IO, /* app mapped write IOs on compressed file */
1166 FS_DATA_IO, /* data IOs from kworker/fsync/reclaimer */
1167 FS_CDATA_IO, /* data IOs from kworker/fsync/reclaimer on compressed file */
1168 FS_NODE_IO, /* node IOs from kworker/fsync/reclaimer */
1169 FS_META_IO, /* meta IOs from kworker/reclaimer */
1170 FS_GC_DATA_IO, /* data IOs from forground gc */
1171 FS_GC_NODE_IO, /* node IOs from forground gc */
1172 FS_CP_DATA_IO, /* data IOs from checkpoint */
1173 FS_CP_NODE_IO, /* node IOs from checkpoint */
1174 FS_CP_META_IO, /* meta IOs from checkpoint */
1177 APP_DIRECT_READ_IO, /* app direct read IOs */
1178 APP_BUFFERED_READ_IO, /* app buffered read IOs */
1179 APP_READ_IO, /* app read IOs */
1180 APP_MAPPED_READ_IO, /* app mapped read IOs */
1181 APP_BUFFERED_CDATA_READ_IO, /* app buffered read IOs on compressed file */
1182 APP_MAPPED_CDATA_READ_IO, /* app mapped read IOs on compressed file */
1183 FS_DATA_READ_IO, /* data read IOs */
1184 FS_GDATA_READ_IO, /* data read IOs from background gc */
1185 FS_CDATA_READ_IO, /* compressed data read IOs */
1186 FS_NODE_READ_IO, /* node read IOs */
1187 FS_META_READ_IO, /* meta read IOs */
1190 FS_DISCARD_IO, /* discard */
1191 FS_FLUSH_IO, /* flush */
1192 FS_ZONE_RESET_IO, /* zone reset */
1196 struct f2fs_io_info {
1197 struct f2fs_sb_info *sbi; /* f2fs_sb_info pointer */
1198 nid_t ino; /* inode number */
1199 enum page_type type; /* contains DATA/NODE/META/META_FLUSH */
1200 enum temp_type temp; /* contains HOT/WARM/COLD */
1201 enum req_op op; /* contains REQ_OP_ */
1202 blk_opf_t op_flags; /* req_flag_bits */
1203 block_t new_blkaddr; /* new block address to be written */
1204 block_t old_blkaddr; /* old block address before Cow */
1205 struct page *page; /* page to be written */
1206 struct page *encrypted_page; /* encrypted page */
1207 struct page *compressed_page; /* compressed page */
1208 struct list_head list; /* serialize IOs */
1209 unsigned int compr_blocks; /* # of compressed block addresses */
1210 unsigned int need_lock:8; /* indicate we need to lock cp_rwsem */
1211 unsigned int version:8; /* version of the node */
1212 unsigned int submitted:1; /* indicate IO submission */
1213 unsigned int in_list:1; /* indicate fio is in io_list */
1214 unsigned int is_por:1; /* indicate IO is from recovery or not */
1215 unsigned int encrypted:1; /* indicate file is encrypted */
1216 unsigned int post_read:1; /* require post read */
1217 enum iostat_type io_type; /* io type */
1218 struct writeback_control *io_wbc; /* writeback control */
1219 struct bio **bio; /* bio for ipu */
1220 sector_t *last_block; /* last block number in bio */
1225 struct list_head list;
1228 #define is_read_io(rw) ((rw) == READ)
1229 struct f2fs_bio_info {
1230 struct f2fs_sb_info *sbi; /* f2fs superblock */
1231 struct bio *bio; /* bios to merge */
1232 sector_t last_block_in_bio; /* last block number */
1233 struct f2fs_io_info fio; /* store buffered io info. */
1234 #ifdef CONFIG_BLK_DEV_ZONED
1235 struct completion zone_wait; /* condition value for the previous open zone to close */
1236 struct bio *zone_pending_bio; /* pending bio for the previous zone */
1237 void *bi_private; /* previous bi_private for pending bio */
1239 struct f2fs_rwsem io_rwsem; /* blocking op for bio */
1240 spinlock_t io_lock; /* serialize DATA/NODE IOs */
1241 struct list_head io_list; /* track fios */
1242 struct list_head bio_list; /* bio entry list head */
1243 struct f2fs_rwsem bio_list_lock; /* lock to protect bio entry list */
1246 #define FDEV(i) (sbi->devs[i])
1247 #define RDEV(i) (raw_super->devs[i])
1248 struct f2fs_dev_info {
1249 struct bdev_handle *bdev_handle;
1250 struct block_device *bdev;
1251 char path[MAX_PATH_LEN];
1252 unsigned int total_segments;
1255 #ifdef CONFIG_BLK_DEV_ZONED
1256 unsigned int nr_blkz; /* Total number of zones */
1257 unsigned long *blkz_seq; /* Bitmap indicating sequential zones */
1262 DIR_INODE, /* for dirty dir inode */
1263 FILE_INODE, /* for dirty regular/symlink inode */
1264 DIRTY_META, /* for all dirtied inode metadata */
1268 /* for inner inode cache management */
1269 struct inode_management {
1270 struct radix_tree_root ino_root; /* ino entry array */
1271 spinlock_t ino_lock; /* for ino entry lock */
1272 struct list_head ino_list; /* inode list head */
1273 unsigned long ino_num; /* number of entries */
1277 struct atgc_management {
1278 bool atgc_enabled; /* ATGC is enabled or not */
1279 struct rb_root_cached root; /* root of victim rb-tree */
1280 struct list_head victim_list; /* linked with all victim entries */
1281 unsigned int victim_count; /* victim count in rb-tree */
1282 unsigned int candidate_ratio; /* candidate ratio */
1283 unsigned int max_candidate_count; /* max candidate count */
1284 unsigned int age_weight; /* age weight, vblock_weight = 100 - age_weight */
1285 unsigned long long age_threshold; /* age threshold */
1288 struct f2fs_gc_control {
1289 unsigned int victim_segno; /* target victim segment number */
1290 int init_gc_type; /* FG_GC or BG_GC */
1291 bool no_bg_gc; /* check the space and stop bg_gc */
1292 bool should_migrate_blocks; /* should migrate blocks */
1293 bool err_gc_skipped; /* return EAGAIN if GC skipped */
1294 unsigned int nr_free_secs; /* # of free sections to do GC */
1298 * For s_flag in struct f2fs_sb_info
1299 * Modification on enum should be synchronized with s_flag array
1302 SBI_IS_DIRTY, /* dirty flag for checkpoint */
1303 SBI_IS_CLOSE, /* specify unmounting */
1304 SBI_NEED_FSCK, /* need fsck.f2fs to fix */
1305 SBI_POR_DOING, /* recovery is doing or not */
1306 SBI_NEED_SB_WRITE, /* need to recover superblock */
1307 SBI_NEED_CP, /* need to checkpoint */
1308 SBI_IS_SHUTDOWN, /* shutdown by ioctl */
1309 SBI_IS_RECOVERED, /* recovered orphan/data */
1310 SBI_CP_DISABLED, /* CP was disabled last mount */
1311 SBI_CP_DISABLED_QUICK, /* CP was disabled quickly */
1312 SBI_QUOTA_NEED_FLUSH, /* need to flush quota info in CP */
1313 SBI_QUOTA_SKIP_FLUSH, /* skip flushing quota in current CP */
1314 SBI_QUOTA_NEED_REPAIR, /* quota file may be corrupted */
1315 SBI_IS_RESIZEFS, /* resizefs is in process */
1316 SBI_IS_FREEZING, /* freezefs is in process */
1317 SBI_IS_WRITABLE, /* remove ro mountoption transiently */
1327 UMOUNT_DISCARD_TIMEOUT,
1331 /* Note that you need to keep synchronization with this gc_mode_names array */
1344 BGGC_MODE_ON, /* background gc is on */
1345 BGGC_MODE_OFF, /* background gc is off */
1347 * background gc is on, migrating blocks
1348 * like foreground gc
1353 FS_MODE_ADAPTIVE, /* use both lfs/ssr allocation */
1354 FS_MODE_LFS, /* use lfs allocation only */
1355 FS_MODE_FRAGMENT_SEG, /* segment fragmentation mode */
1356 FS_MODE_FRAGMENT_BLK, /* block fragmentation mode */
1360 ALLOC_MODE_DEFAULT, /* stay default */
1361 ALLOC_MODE_REUSE, /* reuse segments as much as possible */
1365 FSYNC_MODE_POSIX, /* fsync follows posix semantics */
1366 FSYNC_MODE_STRICT, /* fsync behaves in line with ext4 */
1367 FSYNC_MODE_NOBARRIER, /* fsync behaves nobarrier based on posix */
1372 * automatically compress compression
1376 * automatical compression is disabled.
1377 * user can control the file compression
1383 DISCARD_UNIT_BLOCK, /* basic discard unit is block */
1384 DISCARD_UNIT_SEGMENT, /* basic discard unit is segment */
1385 DISCARD_UNIT_SECTION, /* basic discard unit is section */
1389 MEMORY_MODE_NORMAL, /* memory mode for normal devices */
1390 MEMORY_MODE_LOW, /* memory mode for low memry devices */
1393 enum errors_option {
1394 MOUNT_ERRORS_READONLY, /* remount fs ro on errors */
1395 MOUNT_ERRORS_CONTINUE, /* continue on errors */
1396 MOUNT_ERRORS_PANIC, /* panic on errors */
1403 TOTAL_CALL = FOREGROUND,
1406 static inline int f2fs_test_bit(unsigned int nr, char *addr);
1407 static inline void f2fs_set_bit(unsigned int nr, char *addr);
1408 static inline void f2fs_clear_bit(unsigned int nr, char *addr);
1411 * Layout of f2fs page.private:
1413 * Layout A: lowest bit should be 1
1414 * | bit0 = 1 | bit1 | bit2 | ... | bit MAX | private data .... |
1415 * bit 0 PAGE_PRIVATE_NOT_POINTER
1416 * bit 1 PAGE_PRIVATE_ONGOING_MIGRATION
1417 * bit 2 PAGE_PRIVATE_INLINE_INODE
1418 * bit 3 PAGE_PRIVATE_REF_RESOURCE
1419 * bit 4- f2fs private data
1421 * Layout B: lowest bit should be 0
1422 * page.private is a wrapped pointer.
1425 PAGE_PRIVATE_NOT_POINTER, /* private contains non-pointer data */
1426 PAGE_PRIVATE_ONGOING_MIGRATION, /* data page which is on-going migrating */
1427 PAGE_PRIVATE_INLINE_INODE, /* inode page contains inline data */
1428 PAGE_PRIVATE_REF_RESOURCE, /* dirty page has referenced resources */
1432 /* For compression */
1433 enum compress_algorithm_type {
1441 enum compress_flag {
1446 #define COMPRESS_WATERMARK 20
1447 #define COMPRESS_PERCENT 20
1449 #define COMPRESS_DATA_RESERVED_SIZE 4
1450 struct compress_data {
1451 __le32 clen; /* compressed data size */
1452 __le32 chksum; /* compressed data chksum */
1453 __le32 reserved[COMPRESS_DATA_RESERVED_SIZE]; /* reserved */
1454 u8 cdata[]; /* compressed data */
1457 #define COMPRESS_HEADER_SIZE (sizeof(struct compress_data))
1459 #define F2FS_COMPRESSED_PAGE_MAGIC 0xF5F2C000
1461 #define F2FS_ZSTD_DEFAULT_CLEVEL 1
1463 #define COMPRESS_LEVEL_OFFSET 8
1465 /* compress context */
1466 struct compress_ctx {
1467 struct inode *inode; /* inode the context belong to */
1468 pgoff_t cluster_idx; /* cluster index number */
1469 unsigned int cluster_size; /* page count in cluster */
1470 unsigned int log_cluster_size; /* log of cluster size */
1471 struct page **rpages; /* pages store raw data in cluster */
1472 unsigned int nr_rpages; /* total page number in rpages */
1473 struct page **cpages; /* pages store compressed data in cluster */
1474 unsigned int nr_cpages; /* total page number in cpages */
1475 unsigned int valid_nr_cpages; /* valid page number in cpages */
1476 void *rbuf; /* virtual mapped address on rpages */
1477 struct compress_data *cbuf; /* virtual mapped address on cpages */
1478 size_t rlen; /* valid data length in rbuf */
1479 size_t clen; /* valid data length in cbuf */
1480 void *private; /* payload buffer for specified compression algorithm */
1481 void *private2; /* extra payload buffer */
1484 /* compress context for write IO path */
1485 struct compress_io_ctx {
1486 u32 magic; /* magic number to indicate page is compressed */
1487 struct inode *inode; /* inode the context belong to */
1488 struct page **rpages; /* pages store raw data in cluster */
1489 unsigned int nr_rpages; /* total page number in rpages */
1490 atomic_t pending_pages; /* in-flight compressed page count */
1493 /* Context for decompressing one cluster on the read IO path */
1494 struct decompress_io_ctx {
1495 u32 magic; /* magic number to indicate page is compressed */
1496 struct inode *inode; /* inode the context belong to */
1497 pgoff_t cluster_idx; /* cluster index number */
1498 unsigned int cluster_size; /* page count in cluster */
1499 unsigned int log_cluster_size; /* log of cluster size */
1500 struct page **rpages; /* pages store raw data in cluster */
1501 unsigned int nr_rpages; /* total page number in rpages */
1502 struct page **cpages; /* pages store compressed data in cluster */
1503 unsigned int nr_cpages; /* total page number in cpages */
1504 struct page **tpages; /* temp pages to pad holes in cluster */
1505 void *rbuf; /* virtual mapped address on rpages */
1506 struct compress_data *cbuf; /* virtual mapped address on cpages */
1507 size_t rlen; /* valid data length in rbuf */
1508 size_t clen; /* valid data length in cbuf */
1511 * The number of compressed pages remaining to be read in this cluster.
1512 * This is initially nr_cpages. It is decremented by 1 each time a page
1513 * has been read (or failed to be read). When it reaches 0, the cluster
1514 * is decompressed (or an error is reported).
1516 * If an error occurs before all the pages have been submitted for I/O,
1517 * then this will never reach 0. In this case the I/O submitter is
1518 * responsible for calling f2fs_decompress_end_io() instead.
1520 atomic_t remaining_pages;
1523 * Number of references to this decompress_io_ctx.
1525 * One reference is held for I/O completion. This reference is dropped
1526 * after the pagecache pages are updated and unlocked -- either after
1527 * decompression (and verity if enabled), or after an error.
1529 * In addition, each compressed page holds a reference while it is in a
1530 * bio. These references are necessary prevent compressed pages from
1531 * being freed while they are still in a bio.
1535 bool failed; /* IO error occurred before decompression? */
1536 bool need_verity; /* need fs-verity verification after decompression? */
1537 void *private; /* payload buffer for specified decompression algorithm */
1538 void *private2; /* extra payload buffer */
1539 struct work_struct verity_work; /* work to verify the decompressed pages */
1540 struct work_struct free_work; /* work for late free this structure itself */
1543 #define NULL_CLUSTER ((unsigned int)(~0))
1544 #define MIN_COMPRESS_LOG_SIZE 2
1545 #define MAX_COMPRESS_LOG_SIZE 8
1546 #define MAX_COMPRESS_WINDOW_SIZE(log_size) ((PAGE_SIZE) << (log_size))
1548 struct f2fs_sb_info {
1549 struct super_block *sb; /* pointer to VFS super block */
1550 struct proc_dir_entry *s_proc; /* proc entry */
1551 struct f2fs_super_block *raw_super; /* raw super block pointer */
1552 struct f2fs_rwsem sb_lock; /* lock for raw super block */
1553 int valid_super_block; /* valid super block no */
1554 unsigned long s_flag; /* flags for sbi */
1555 struct mutex writepages; /* mutex for writepages() */
1557 #ifdef CONFIG_BLK_DEV_ZONED
1558 unsigned int blocks_per_blkz; /* F2FS blocks per zone */
1561 /* for node-related operations */
1562 struct f2fs_nm_info *nm_info; /* node manager */
1563 struct inode *node_inode; /* cache node blocks */
1565 /* for segment-related operations */
1566 struct f2fs_sm_info *sm_info; /* segment manager */
1568 /* for bio operations */
1569 struct f2fs_bio_info *write_io[NR_PAGE_TYPE]; /* for write bios */
1570 /* keep migration IO order for LFS mode */
1571 struct f2fs_rwsem io_order_lock;
1572 pgoff_t page_eio_ofs[NR_PAGE_TYPE]; /* EIO page offset */
1573 int page_eio_cnt[NR_PAGE_TYPE]; /* EIO count */
1575 /* for checkpoint */
1576 struct f2fs_checkpoint *ckpt; /* raw checkpoint pointer */
1577 int cur_cp_pack; /* remain current cp pack */
1578 spinlock_t cp_lock; /* for flag in ckpt */
1579 struct inode *meta_inode; /* cache meta blocks */
1580 struct f2fs_rwsem cp_global_sem; /* checkpoint procedure lock */
1581 struct f2fs_rwsem cp_rwsem; /* blocking FS operations */
1582 struct f2fs_rwsem node_write; /* locking node writes */
1583 struct f2fs_rwsem node_change; /* locking node change */
1584 wait_queue_head_t cp_wait;
1585 unsigned long last_time[MAX_TIME]; /* to store time in jiffies */
1586 long interval_time[MAX_TIME]; /* to store thresholds */
1587 struct ckpt_req_control cprc_info; /* for checkpoint request control */
1589 struct inode_management im[MAX_INO_ENTRY]; /* manage inode cache */
1591 spinlock_t fsync_node_lock; /* for node entry lock */
1592 struct list_head fsync_node_list; /* node list head */
1593 unsigned int fsync_seg_id; /* sequence id */
1594 unsigned int fsync_node_num; /* number of node entries */
1596 /* for orphan inode, use 0'th array */
1597 unsigned int max_orphans; /* max orphan inodes */
1599 /* for inode management */
1600 struct list_head inode_list[NR_INODE_TYPE]; /* dirty inode list */
1601 spinlock_t inode_lock[NR_INODE_TYPE]; /* for dirty inode list lock */
1602 struct mutex flush_lock; /* for flush exclusion */
1604 /* for extent tree cache */
1605 struct extent_tree_info extent_tree[NR_EXTENT_CACHES];
1606 atomic64_t allocated_data_blocks; /* for block age extent_cache */
1608 /* The threshold used for hot and warm data seperation*/
1609 unsigned int hot_data_age_threshold;
1610 unsigned int warm_data_age_threshold;
1611 unsigned int last_age_weight;
1613 /* basic filesystem units */
1614 unsigned int log_sectors_per_block; /* log2 sectors per block */
1615 unsigned int log_blocksize; /* log2 block size */
1616 unsigned int blocksize; /* block size */
1617 unsigned int root_ino_num; /* root inode number*/
1618 unsigned int node_ino_num; /* node inode number*/
1619 unsigned int meta_ino_num; /* meta inode number*/
1620 unsigned int log_blocks_per_seg; /* log2 blocks per segment */
1621 unsigned int blocks_per_seg; /* blocks per segment */
1622 unsigned int unusable_blocks_per_sec; /* unusable blocks per section */
1623 unsigned int segs_per_sec; /* segments per section */
1624 unsigned int secs_per_zone; /* sections per zone */
1625 unsigned int total_sections; /* total section count */
1626 unsigned int total_node_count; /* total node block count */
1627 unsigned int total_valid_node_count; /* valid node block count */
1628 int dir_level; /* directory level */
1629 bool readdir_ra; /* readahead inode in readdir */
1630 u64 max_io_bytes; /* max io bytes to merge IOs */
1632 block_t user_block_count; /* # of user blocks */
1633 block_t total_valid_block_count; /* # of valid blocks */
1634 block_t discard_blks; /* discard command candidats */
1635 block_t last_valid_block_count; /* for recovery */
1636 block_t reserved_blocks; /* configurable reserved blocks */
1637 block_t current_reserved_blocks; /* current reserved blocks */
1639 /* Additional tracking for no checkpoint mode */
1640 block_t unusable_block_count; /* # of blocks saved by last cp */
1642 unsigned int nquota_files; /* # of quota sysfile */
1643 struct f2fs_rwsem quota_sem; /* blocking cp for flags */
1645 /* # of pages, see count_type */
1646 atomic_t nr_pages[NR_COUNT_TYPE];
1647 /* # of allocated blocks */
1648 struct percpu_counter alloc_valid_block_count;
1649 /* # of node block writes as roll forward recovery */
1650 struct percpu_counter rf_node_block_count;
1652 /* writeback control */
1653 atomic_t wb_sync_req[META]; /* count # of WB_SYNC threads */
1655 /* valid inode count */
1656 struct percpu_counter total_valid_inode_count;
1658 struct f2fs_mount_info mount_opt; /* mount options */
1660 /* for cleaning operations */
1661 struct f2fs_rwsem gc_lock; /*
1662 * semaphore for GC, avoid
1663 * race between GC and GC or CP
1665 struct f2fs_gc_kthread *gc_thread; /* GC thread */
1666 struct atgc_management am; /* atgc management */
1667 unsigned int cur_victim_sec; /* current victim section num */
1668 unsigned int gc_mode; /* current GC state */
1669 unsigned int next_victim_seg[2]; /* next segment in victim section */
1670 spinlock_t gc_remaining_trials_lock;
1671 /* remaining trial count for GC_URGENT_* and GC_IDLE_* */
1672 unsigned int gc_remaining_trials;
1674 /* for skip statistic */
1675 unsigned long long skipped_gc_rwsem; /* FG_GC only */
1677 /* threshold for gc trials on pinned files */
1678 u64 gc_pin_file_threshold;
1679 struct f2fs_rwsem pin_sem;
1681 /* maximum # of trials to find a victim segment for SSR and GC */
1682 unsigned int max_victim_search;
1683 /* migration granularity of garbage collection, unit: segment */
1684 unsigned int migration_granularity;
1687 * for stat information.
1688 * one is for the LFS mode, and the other is for the SSR mode.
1690 #ifdef CONFIG_F2FS_STAT_FS
1691 struct f2fs_stat_info *stat_info; /* FS status information */
1692 atomic_t meta_count[META_MAX]; /* # of meta blocks */
1693 unsigned int segment_count[2]; /* # of allocated segments */
1694 unsigned int block_count[2]; /* # of allocated blocks */
1695 atomic_t inplace_count; /* # of inplace update */
1696 /* # of lookup extent cache */
1697 atomic64_t total_hit_ext[NR_EXTENT_CACHES];
1698 /* # of hit rbtree extent node */
1699 atomic64_t read_hit_rbtree[NR_EXTENT_CACHES];
1700 /* # of hit cached extent node */
1701 atomic64_t read_hit_cached[NR_EXTENT_CACHES];
1702 /* # of hit largest extent node in read extent cache */
1703 atomic64_t read_hit_largest;
1704 atomic_t inline_xattr; /* # of inline_xattr inodes */
1705 atomic_t inline_inode; /* # of inline_data inodes */
1706 atomic_t inline_dir; /* # of inline_dentry inodes */
1707 atomic_t compr_inode; /* # of compressed inodes */
1708 atomic64_t compr_blocks; /* # of compressed blocks */
1709 atomic_t swapfile_inode; /* # of swapfile inodes */
1710 atomic_t atomic_files; /* # of opened atomic file */
1711 atomic_t max_aw_cnt; /* max # of atomic writes */
1712 unsigned int io_skip_bggc; /* skip background gc for in-flight IO */
1713 unsigned int other_skip_bggc; /* skip background gc for other reasons */
1714 unsigned int ndirty_inode[NR_INODE_TYPE]; /* # of dirty inodes */
1715 atomic_t cp_call_count[MAX_CALL_TYPE]; /* # of cp call */
1717 spinlock_t stat_lock; /* lock for stat operations */
1719 /* to attach REQ_META|REQ_FUA flags */
1720 unsigned int data_io_flag;
1721 unsigned int node_io_flag;
1723 /* For sysfs support */
1724 struct kobject s_kobj; /* /sys/fs/f2fs/<devname> */
1725 struct completion s_kobj_unregister;
1727 struct kobject s_stat_kobj; /* /sys/fs/f2fs/<devname>/stat */
1728 struct completion s_stat_kobj_unregister;
1730 struct kobject s_feature_list_kobj; /* /sys/fs/f2fs/<devname>/feature_list */
1731 struct completion s_feature_list_kobj_unregister;
1733 /* For shrinker support */
1734 struct list_head s_list;
1735 struct mutex umount_mutex;
1736 unsigned int shrinker_run_no;
1738 /* For multi devices */
1739 int s_ndevs; /* number of devices */
1740 struct f2fs_dev_info *devs; /* for device list */
1741 unsigned int dirty_device; /* for checkpoint data flush */
1742 spinlock_t dev_lock; /* protect dirty_device */
1743 bool aligned_blksize; /* all devices has the same logical blksize */
1745 /* For write statistics */
1746 u64 sectors_written_start;
1749 /* Reference to checksum algorithm driver via cryptoapi */
1750 struct crypto_shash *s_chksum_driver;
1752 /* Precomputed FS UUID checksum for seeding other checksums */
1753 __u32 s_chksum_seed;
1755 struct workqueue_struct *post_read_wq; /* post read workqueue */
1758 * If we are in irq context, let's update error information into
1759 * on-disk superblock in the work.
1761 struct work_struct s_error_work;
1762 unsigned char errors[MAX_F2FS_ERRORS]; /* error flags */
1763 unsigned char stop_reason[MAX_STOP_REASON]; /* stop reason */
1764 spinlock_t error_lock; /* protect errors/stop_reason array */
1765 bool error_dirty; /* errors of sb is dirty */
1767 struct kmem_cache *inline_xattr_slab; /* inline xattr entry */
1768 unsigned int inline_xattr_slab_size; /* default inline xattr slab size */
1770 /* For reclaimed segs statistics per each GC mode */
1771 unsigned int gc_segment_mode; /* GC state for reclaimed segments */
1772 unsigned int gc_reclaimed_segs[MAX_GC_MODE]; /* Reclaimed segs for each mode */
1774 unsigned long seq_file_ra_mul; /* multiplier for ra_pages of seq. files in fadvise */
1776 int max_fragment_chunk; /* max chunk size for block fragmentation mode */
1777 int max_fragment_hole; /* max hole size for block fragmentation mode */
1779 /* For atomic write statistics */
1780 atomic64_t current_atomic_write;
1781 s64 peak_atomic_write;
1782 u64 committed_atomic_block;
1783 u64 revoked_atomic_block;
1785 #ifdef CONFIG_F2FS_FS_COMPRESSION
1786 struct kmem_cache *page_array_slab; /* page array entry */
1787 unsigned int page_array_slab_size; /* default page array slab size */
1789 /* For runtime compression statistics */
1790 u64 compr_written_block;
1791 u64 compr_saved_block;
1792 u32 compr_new_inode;
1794 /* For compressed block cache */
1795 struct inode *compress_inode; /* cache compressed blocks */
1796 unsigned int compress_percent; /* cache page percentage */
1797 unsigned int compress_watermark; /* cache page watermark */
1798 atomic_t compress_page_hit; /* cache hit count */
1801 #ifdef CONFIG_F2FS_IOSTAT
1802 /* For app/fs IO statistics */
1803 spinlock_t iostat_lock;
1804 unsigned long long iostat_count[NR_IO_TYPE];
1805 unsigned long long iostat_bytes[NR_IO_TYPE];
1806 unsigned long long prev_iostat_bytes[NR_IO_TYPE];
1808 unsigned long iostat_next_period;
1809 unsigned int iostat_period_ms;
1811 /* For io latency related statistics info in one iostat period */
1812 spinlock_t iostat_lat_lock;
1813 struct iostat_lat_info *iostat_io_lat;
1817 /* Definitions to access f2fs_sb_info */
1818 #define BLKS_PER_SEG(sbi) \
1819 ((sbi)->blocks_per_seg)
1820 #define BLKS_PER_SEC(sbi) \
1821 ((sbi)->segs_per_sec << (sbi)->log_blocks_per_seg)
1822 #define SEGS_PER_SEC(sbi) \
1823 ((sbi)->segs_per_sec)
1826 void f2fs_printk(struct f2fs_sb_info *sbi, bool limit_rate, const char *fmt, ...);
1828 #define f2fs_err(sbi, fmt, ...) \
1829 f2fs_printk(sbi, false, KERN_ERR fmt, ##__VA_ARGS__)
1830 #define f2fs_warn(sbi, fmt, ...) \
1831 f2fs_printk(sbi, false, KERN_WARNING fmt, ##__VA_ARGS__)
1832 #define f2fs_notice(sbi, fmt, ...) \
1833 f2fs_printk(sbi, false, KERN_NOTICE fmt, ##__VA_ARGS__)
1834 #define f2fs_info(sbi, fmt, ...) \
1835 f2fs_printk(sbi, false, KERN_INFO fmt, ##__VA_ARGS__)
1836 #define f2fs_debug(sbi, fmt, ...) \
1837 f2fs_printk(sbi, false, KERN_DEBUG fmt, ##__VA_ARGS__)
1839 #define f2fs_err_ratelimited(sbi, fmt, ...) \
1840 f2fs_printk(sbi, true, KERN_ERR fmt, ##__VA_ARGS__)
1841 #define f2fs_warn_ratelimited(sbi, fmt, ...) \
1842 f2fs_printk(sbi, true, KERN_WARNING fmt, ##__VA_ARGS__)
1843 #define f2fs_info_ratelimited(sbi, fmt, ...) \
1844 f2fs_printk(sbi, true, KERN_INFO fmt, ##__VA_ARGS__)
1846 #ifdef CONFIG_F2FS_FAULT_INJECTION
1847 #define time_to_inject(sbi, type) __time_to_inject(sbi, type, __func__, \
1848 __builtin_return_address(0))
1849 static inline bool __time_to_inject(struct f2fs_sb_info *sbi, int type,
1850 const char *func, const char *parent_func)
1852 struct f2fs_fault_info *ffi = &F2FS_OPTION(sbi).fault_info;
1854 if (!ffi->inject_rate)
1857 if (!IS_FAULT_SET(ffi, type))
1860 atomic_inc(&ffi->inject_ops);
1861 if (atomic_read(&ffi->inject_ops) >= ffi->inject_rate) {
1862 atomic_set(&ffi->inject_ops, 0);
1863 f2fs_info_ratelimited(sbi, "inject %s in %s of %pS",
1864 f2fs_fault_name[type], func, parent_func);
1870 static inline bool time_to_inject(struct f2fs_sb_info *sbi, int type)
1877 * Test if the mounted volume is a multi-device volume.
1878 * - For a single regular disk volume, sbi->s_ndevs is 0.
1879 * - For a single zoned disk volume, sbi->s_ndevs is 1.
1880 * - For a multi-device volume, sbi->s_ndevs is always 2 or more.
1882 static inline bool f2fs_is_multi_device(struct f2fs_sb_info *sbi)
1884 return sbi->s_ndevs > 1;
1887 static inline void f2fs_update_time(struct f2fs_sb_info *sbi, int type)
1889 unsigned long now = jiffies;
1891 sbi->last_time[type] = now;
1893 /* DISCARD_TIME and GC_TIME are based on REQ_TIME */
1894 if (type == REQ_TIME) {
1895 sbi->last_time[DISCARD_TIME] = now;
1896 sbi->last_time[GC_TIME] = now;
1900 static inline bool f2fs_time_over(struct f2fs_sb_info *sbi, int type)
1902 unsigned long interval = sbi->interval_time[type] * HZ;
1904 return time_after(jiffies, sbi->last_time[type] + interval);
1907 static inline unsigned int f2fs_time_to_wait(struct f2fs_sb_info *sbi,
1910 unsigned long interval = sbi->interval_time[type] * HZ;
1911 unsigned int wait_ms = 0;
1914 delta = (sbi->last_time[type] + interval) - jiffies;
1916 wait_ms = jiffies_to_msecs(delta);
1924 static inline u32 __f2fs_crc32(struct f2fs_sb_info *sbi, u32 crc,
1925 const void *address, unsigned int length)
1928 struct shash_desc shash;
1933 BUG_ON(crypto_shash_descsize(sbi->s_chksum_driver) != sizeof(desc.ctx));
1935 desc.shash.tfm = sbi->s_chksum_driver;
1936 *(u32 *)desc.ctx = crc;
1938 err = crypto_shash_update(&desc.shash, address, length);
1941 return *(u32 *)desc.ctx;
1944 static inline u32 f2fs_crc32(struct f2fs_sb_info *sbi, const void *address,
1945 unsigned int length)
1947 return __f2fs_crc32(sbi, F2FS_SUPER_MAGIC, address, length);
1950 static inline bool f2fs_crc_valid(struct f2fs_sb_info *sbi, __u32 blk_crc,
1951 void *buf, size_t buf_size)
1953 return f2fs_crc32(sbi, buf, buf_size) == blk_crc;
1956 static inline u32 f2fs_chksum(struct f2fs_sb_info *sbi, u32 crc,
1957 const void *address, unsigned int length)
1959 return __f2fs_crc32(sbi, crc, address, length);
1962 static inline struct f2fs_inode_info *F2FS_I(struct inode *inode)
1964 return container_of(inode, struct f2fs_inode_info, vfs_inode);
1967 static inline struct f2fs_sb_info *F2FS_SB(struct super_block *sb)
1969 return sb->s_fs_info;
1972 static inline struct f2fs_sb_info *F2FS_I_SB(struct inode *inode)
1974 return F2FS_SB(inode->i_sb);
1977 static inline struct f2fs_sb_info *F2FS_M_SB(struct address_space *mapping)
1979 return F2FS_I_SB(mapping->host);
1982 static inline struct f2fs_sb_info *F2FS_P_SB(struct page *page)
1984 return F2FS_M_SB(page_file_mapping(page));
1987 static inline struct f2fs_super_block *F2FS_RAW_SUPER(struct f2fs_sb_info *sbi)
1989 return (struct f2fs_super_block *)(sbi->raw_super);
1992 static inline struct f2fs_checkpoint *F2FS_CKPT(struct f2fs_sb_info *sbi)
1994 return (struct f2fs_checkpoint *)(sbi->ckpt);
1997 static inline struct f2fs_node *F2FS_NODE(struct page *page)
1999 return (struct f2fs_node *)page_address(page);
2002 static inline struct f2fs_inode *F2FS_INODE(struct page *page)
2004 return &((struct f2fs_node *)page_address(page))->i;
2007 static inline struct f2fs_nm_info *NM_I(struct f2fs_sb_info *sbi)
2009 return (struct f2fs_nm_info *)(sbi->nm_info);
2012 static inline struct f2fs_sm_info *SM_I(struct f2fs_sb_info *sbi)
2014 return (struct f2fs_sm_info *)(sbi->sm_info);
2017 static inline struct sit_info *SIT_I(struct f2fs_sb_info *sbi)
2019 return (struct sit_info *)(SM_I(sbi)->sit_info);
2022 static inline struct free_segmap_info *FREE_I(struct f2fs_sb_info *sbi)
2024 return (struct free_segmap_info *)(SM_I(sbi)->free_info);
2027 static inline struct dirty_seglist_info *DIRTY_I(struct f2fs_sb_info *sbi)
2029 return (struct dirty_seglist_info *)(SM_I(sbi)->dirty_info);
2032 static inline struct address_space *META_MAPPING(struct f2fs_sb_info *sbi)
2034 return sbi->meta_inode->i_mapping;
2037 static inline struct address_space *NODE_MAPPING(struct f2fs_sb_info *sbi)
2039 return sbi->node_inode->i_mapping;
2042 static inline bool is_sbi_flag_set(struct f2fs_sb_info *sbi, unsigned int type)
2044 return test_bit(type, &sbi->s_flag);
2047 static inline void set_sbi_flag(struct f2fs_sb_info *sbi, unsigned int type)
2049 set_bit(type, &sbi->s_flag);
2052 static inline void clear_sbi_flag(struct f2fs_sb_info *sbi, unsigned int type)
2054 clear_bit(type, &sbi->s_flag);
2057 static inline unsigned long long cur_cp_version(struct f2fs_checkpoint *cp)
2059 return le64_to_cpu(cp->checkpoint_ver);
2062 static inline unsigned long f2fs_qf_ino(struct super_block *sb, int type)
2064 if (type < F2FS_MAX_QUOTAS)
2065 return le32_to_cpu(F2FS_SB(sb)->raw_super->qf_ino[type]);
2069 static inline __u64 cur_cp_crc(struct f2fs_checkpoint *cp)
2071 size_t crc_offset = le32_to_cpu(cp->checksum_offset);
2072 return le32_to_cpu(*((__le32 *)((unsigned char *)cp + crc_offset)));
2075 static inline bool __is_set_ckpt_flags(struct f2fs_checkpoint *cp, unsigned int f)
2077 unsigned int ckpt_flags = le32_to_cpu(cp->ckpt_flags);
2079 return ckpt_flags & f;
2082 static inline bool is_set_ckpt_flags(struct f2fs_sb_info *sbi, unsigned int f)
2084 return __is_set_ckpt_flags(F2FS_CKPT(sbi), f);
2087 static inline void __set_ckpt_flags(struct f2fs_checkpoint *cp, unsigned int f)
2089 unsigned int ckpt_flags;
2091 ckpt_flags = le32_to_cpu(cp->ckpt_flags);
2093 cp->ckpt_flags = cpu_to_le32(ckpt_flags);
2096 static inline void set_ckpt_flags(struct f2fs_sb_info *sbi, unsigned int f)
2098 unsigned long flags;
2100 spin_lock_irqsave(&sbi->cp_lock, flags);
2101 __set_ckpt_flags(F2FS_CKPT(sbi), f);
2102 spin_unlock_irqrestore(&sbi->cp_lock, flags);
2105 static inline void __clear_ckpt_flags(struct f2fs_checkpoint *cp, unsigned int f)
2107 unsigned int ckpt_flags;
2109 ckpt_flags = le32_to_cpu(cp->ckpt_flags);
2111 cp->ckpt_flags = cpu_to_le32(ckpt_flags);
2114 static inline void clear_ckpt_flags(struct f2fs_sb_info *sbi, unsigned int f)
2116 unsigned long flags;
2118 spin_lock_irqsave(&sbi->cp_lock, flags);
2119 __clear_ckpt_flags(F2FS_CKPT(sbi), f);
2120 spin_unlock_irqrestore(&sbi->cp_lock, flags);
2123 #define init_f2fs_rwsem(sem) \
2125 static struct lock_class_key __key; \
2127 __init_f2fs_rwsem((sem), #sem, &__key); \
2130 static inline void __init_f2fs_rwsem(struct f2fs_rwsem *sem,
2131 const char *sem_name, struct lock_class_key *key)
2133 __init_rwsem(&sem->internal_rwsem, sem_name, key);
2134 #ifdef CONFIG_F2FS_UNFAIR_RWSEM
2135 init_waitqueue_head(&sem->read_waiters);
2139 static inline int f2fs_rwsem_is_locked(struct f2fs_rwsem *sem)
2141 return rwsem_is_locked(&sem->internal_rwsem);
2144 static inline int f2fs_rwsem_is_contended(struct f2fs_rwsem *sem)
2146 return rwsem_is_contended(&sem->internal_rwsem);
2149 static inline void f2fs_down_read(struct f2fs_rwsem *sem)
2151 #ifdef CONFIG_F2FS_UNFAIR_RWSEM
2152 wait_event(sem->read_waiters, down_read_trylock(&sem->internal_rwsem));
2154 down_read(&sem->internal_rwsem);
2158 static inline int f2fs_down_read_trylock(struct f2fs_rwsem *sem)
2160 return down_read_trylock(&sem->internal_rwsem);
2163 static inline void f2fs_up_read(struct f2fs_rwsem *sem)
2165 up_read(&sem->internal_rwsem);
2168 static inline void f2fs_down_write(struct f2fs_rwsem *sem)
2170 down_write(&sem->internal_rwsem);
2173 #ifdef CONFIG_DEBUG_LOCK_ALLOC
2174 static inline void f2fs_down_read_nested(struct f2fs_rwsem *sem, int subclass)
2176 down_read_nested(&sem->internal_rwsem, subclass);
2179 static inline void f2fs_down_write_nested(struct f2fs_rwsem *sem, int subclass)
2181 down_write_nested(&sem->internal_rwsem, subclass);
2184 #define f2fs_down_read_nested(sem, subclass) f2fs_down_read(sem)
2185 #define f2fs_down_write_nested(sem, subclass) f2fs_down_write(sem)
2188 static inline int f2fs_down_write_trylock(struct f2fs_rwsem *sem)
2190 return down_write_trylock(&sem->internal_rwsem);
2193 static inline void f2fs_up_write(struct f2fs_rwsem *sem)
2195 up_write(&sem->internal_rwsem);
2196 #ifdef CONFIG_F2FS_UNFAIR_RWSEM
2197 wake_up_all(&sem->read_waiters);
2201 static inline void f2fs_lock_op(struct f2fs_sb_info *sbi)
2203 f2fs_down_read(&sbi->cp_rwsem);
2206 static inline int f2fs_trylock_op(struct f2fs_sb_info *sbi)
2208 if (time_to_inject(sbi, FAULT_LOCK_OP))
2210 return f2fs_down_read_trylock(&sbi->cp_rwsem);
2213 static inline void f2fs_unlock_op(struct f2fs_sb_info *sbi)
2215 f2fs_up_read(&sbi->cp_rwsem);
2218 static inline void f2fs_lock_all(struct f2fs_sb_info *sbi)
2220 f2fs_down_write(&sbi->cp_rwsem);
2223 static inline void f2fs_unlock_all(struct f2fs_sb_info *sbi)
2225 f2fs_up_write(&sbi->cp_rwsem);
2228 static inline int __get_cp_reason(struct f2fs_sb_info *sbi)
2230 int reason = CP_SYNC;
2232 if (test_opt(sbi, FASTBOOT))
2233 reason = CP_FASTBOOT;
2234 if (is_sbi_flag_set(sbi, SBI_IS_CLOSE))
2239 static inline bool __remain_node_summaries(int reason)
2241 return (reason & (CP_UMOUNT | CP_FASTBOOT));
2244 static inline bool __exist_node_summaries(struct f2fs_sb_info *sbi)
2246 return (is_set_ckpt_flags(sbi, CP_UMOUNT_FLAG) ||
2247 is_set_ckpt_flags(sbi, CP_FASTBOOT_FLAG));
2251 * Check whether the inode has blocks or not
2253 static inline int F2FS_HAS_BLOCKS(struct inode *inode)
2255 block_t xattr_block = F2FS_I(inode)->i_xattr_nid ? 1 : 0;
2257 return (inode->i_blocks >> F2FS_LOG_SECTORS_PER_BLOCK) > xattr_block;
2260 static inline bool f2fs_has_xattr_block(unsigned int ofs)
2262 return ofs == XATTR_NODE_OFFSET;
2265 static inline bool __allow_reserved_blocks(struct f2fs_sb_info *sbi,
2266 struct inode *inode, bool cap)
2270 if (!test_opt(sbi, RESERVE_ROOT))
2272 if (IS_NOQUOTA(inode))
2274 if (uid_eq(F2FS_OPTION(sbi).s_resuid, current_fsuid()))
2276 if (!gid_eq(F2FS_OPTION(sbi).s_resgid, GLOBAL_ROOT_GID) &&
2277 in_group_p(F2FS_OPTION(sbi).s_resgid))
2279 if (cap && capable(CAP_SYS_RESOURCE))
2284 static inline unsigned int get_available_block_count(struct f2fs_sb_info *sbi,
2285 struct inode *inode, bool cap)
2287 block_t avail_user_block_count;
2289 avail_user_block_count = sbi->user_block_count -
2290 sbi->current_reserved_blocks;
2292 if (!__allow_reserved_blocks(sbi, inode, cap))
2293 avail_user_block_count -= F2FS_OPTION(sbi).root_reserved_blocks;
2295 if (unlikely(is_sbi_flag_set(sbi, SBI_CP_DISABLED))) {
2296 if (avail_user_block_count > sbi->unusable_block_count)
2297 avail_user_block_count -= sbi->unusable_block_count;
2299 avail_user_block_count = 0;
2302 return avail_user_block_count;
2305 static inline void f2fs_i_blocks_write(struct inode *, block_t, bool, bool);
2306 static inline int inc_valid_block_count(struct f2fs_sb_info *sbi,
2307 struct inode *inode, blkcnt_t *count, bool partial)
2309 blkcnt_t diff = 0, release = 0;
2310 block_t avail_user_block_count;
2313 ret = dquot_reserve_block(inode, *count);
2317 if (time_to_inject(sbi, FAULT_BLOCK)) {
2323 * let's increase this in prior to actual block count change in order
2324 * for f2fs_sync_file to avoid data races when deciding checkpoint.
2326 percpu_counter_add(&sbi->alloc_valid_block_count, (*count));
2328 spin_lock(&sbi->stat_lock);
2329 sbi->total_valid_block_count += (block_t)(*count);
2330 avail_user_block_count = get_available_block_count(sbi, inode, true);
2332 if (unlikely(sbi->total_valid_block_count > avail_user_block_count)) {
2334 spin_unlock(&sbi->stat_lock);
2338 diff = sbi->total_valid_block_count - avail_user_block_count;
2343 sbi->total_valid_block_count -= diff;
2345 spin_unlock(&sbi->stat_lock);
2349 spin_unlock(&sbi->stat_lock);
2351 if (unlikely(release)) {
2352 percpu_counter_sub(&sbi->alloc_valid_block_count, release);
2353 dquot_release_reservation_block(inode, release);
2355 f2fs_i_blocks_write(inode, *count, true, true);
2359 percpu_counter_sub(&sbi->alloc_valid_block_count, release);
2361 dquot_release_reservation_block(inode, release);
2365 #define PAGE_PRIVATE_GET_FUNC(name, flagname) \
2366 static inline bool page_private_##name(struct page *page) \
2368 return PagePrivate(page) && \
2369 test_bit(PAGE_PRIVATE_NOT_POINTER, &page_private(page)) && \
2370 test_bit(PAGE_PRIVATE_##flagname, &page_private(page)); \
2373 #define PAGE_PRIVATE_SET_FUNC(name, flagname) \
2374 static inline void set_page_private_##name(struct page *page) \
2376 if (!PagePrivate(page)) \
2377 attach_page_private(page, (void *)0); \
2378 set_bit(PAGE_PRIVATE_NOT_POINTER, &page_private(page)); \
2379 set_bit(PAGE_PRIVATE_##flagname, &page_private(page)); \
2382 #define PAGE_PRIVATE_CLEAR_FUNC(name, flagname) \
2383 static inline void clear_page_private_##name(struct page *page) \
2385 clear_bit(PAGE_PRIVATE_##flagname, &page_private(page)); \
2386 if (page_private(page) == BIT(PAGE_PRIVATE_NOT_POINTER)) \
2387 detach_page_private(page); \
2390 PAGE_PRIVATE_GET_FUNC(nonpointer, NOT_POINTER);
2391 PAGE_PRIVATE_GET_FUNC(inline, INLINE_INODE);
2392 PAGE_PRIVATE_GET_FUNC(gcing, ONGOING_MIGRATION);
2394 PAGE_PRIVATE_SET_FUNC(reference, REF_RESOURCE);
2395 PAGE_PRIVATE_SET_FUNC(inline, INLINE_INODE);
2396 PAGE_PRIVATE_SET_FUNC(gcing, ONGOING_MIGRATION);
2398 PAGE_PRIVATE_CLEAR_FUNC(reference, REF_RESOURCE);
2399 PAGE_PRIVATE_CLEAR_FUNC(inline, INLINE_INODE);
2400 PAGE_PRIVATE_CLEAR_FUNC(gcing, ONGOING_MIGRATION);
2402 static inline unsigned long get_page_private_data(struct page *page)
2404 unsigned long data = page_private(page);
2406 if (!test_bit(PAGE_PRIVATE_NOT_POINTER, &data))
2408 return data >> PAGE_PRIVATE_MAX;
2411 static inline void set_page_private_data(struct page *page, unsigned long data)
2413 if (!PagePrivate(page))
2414 attach_page_private(page, (void *)0);
2415 set_bit(PAGE_PRIVATE_NOT_POINTER, &page_private(page));
2416 page_private(page) |= data << PAGE_PRIVATE_MAX;
2419 static inline void clear_page_private_data(struct page *page)
2421 page_private(page) &= GENMASK(PAGE_PRIVATE_MAX - 1, 0);
2422 if (page_private(page) == BIT(PAGE_PRIVATE_NOT_POINTER))
2423 detach_page_private(page);
2426 static inline void clear_page_private_all(struct page *page)
2428 clear_page_private_data(page);
2429 clear_page_private_reference(page);
2430 clear_page_private_gcing(page);
2431 clear_page_private_inline(page);
2433 f2fs_bug_on(F2FS_P_SB(page), page_private(page));
2436 static inline void dec_valid_block_count(struct f2fs_sb_info *sbi,
2437 struct inode *inode,
2440 blkcnt_t sectors = count << F2FS_LOG_SECTORS_PER_BLOCK;
2442 spin_lock(&sbi->stat_lock);
2443 f2fs_bug_on(sbi, sbi->total_valid_block_count < (block_t) count);
2444 sbi->total_valid_block_count -= (block_t)count;
2445 if (sbi->reserved_blocks &&
2446 sbi->current_reserved_blocks < sbi->reserved_blocks)
2447 sbi->current_reserved_blocks = min(sbi->reserved_blocks,
2448 sbi->current_reserved_blocks + count);
2449 spin_unlock(&sbi->stat_lock);
2450 if (unlikely(inode->i_blocks < sectors)) {
2451 f2fs_warn(sbi, "Inconsistent i_blocks, ino:%lu, iblocks:%llu, sectors:%llu",
2453 (unsigned long long)inode->i_blocks,
2454 (unsigned long long)sectors);
2455 set_sbi_flag(sbi, SBI_NEED_FSCK);
2458 f2fs_i_blocks_write(inode, count, false, true);
2461 static inline void inc_page_count(struct f2fs_sb_info *sbi, int count_type)
2463 atomic_inc(&sbi->nr_pages[count_type]);
2465 if (count_type == F2FS_DIRTY_DENTS ||
2466 count_type == F2FS_DIRTY_NODES ||
2467 count_type == F2FS_DIRTY_META ||
2468 count_type == F2FS_DIRTY_QDATA ||
2469 count_type == F2FS_DIRTY_IMETA)
2470 set_sbi_flag(sbi, SBI_IS_DIRTY);
2473 static inline void inode_inc_dirty_pages(struct inode *inode)
2475 atomic_inc(&F2FS_I(inode)->dirty_pages);
2476 inc_page_count(F2FS_I_SB(inode), S_ISDIR(inode->i_mode) ?
2477 F2FS_DIRTY_DENTS : F2FS_DIRTY_DATA);
2478 if (IS_NOQUOTA(inode))
2479 inc_page_count(F2FS_I_SB(inode), F2FS_DIRTY_QDATA);
2482 static inline void dec_page_count(struct f2fs_sb_info *sbi, int count_type)
2484 atomic_dec(&sbi->nr_pages[count_type]);
2487 static inline void inode_dec_dirty_pages(struct inode *inode)
2489 if (!S_ISDIR(inode->i_mode) && !S_ISREG(inode->i_mode) &&
2490 !S_ISLNK(inode->i_mode))
2493 atomic_dec(&F2FS_I(inode)->dirty_pages);
2494 dec_page_count(F2FS_I_SB(inode), S_ISDIR(inode->i_mode) ?
2495 F2FS_DIRTY_DENTS : F2FS_DIRTY_DATA);
2496 if (IS_NOQUOTA(inode))
2497 dec_page_count(F2FS_I_SB(inode), F2FS_DIRTY_QDATA);
2500 static inline void inc_atomic_write_cnt(struct inode *inode)
2502 struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
2503 struct f2fs_inode_info *fi = F2FS_I(inode);
2506 fi->atomic_write_cnt++;
2507 atomic64_inc(&sbi->current_atomic_write);
2508 current_write = atomic64_read(&sbi->current_atomic_write);
2509 if (current_write > sbi->peak_atomic_write)
2510 sbi->peak_atomic_write = current_write;
2513 static inline void release_atomic_write_cnt(struct inode *inode)
2515 struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
2516 struct f2fs_inode_info *fi = F2FS_I(inode);
2518 atomic64_sub(fi->atomic_write_cnt, &sbi->current_atomic_write);
2519 fi->atomic_write_cnt = 0;
2522 static inline s64 get_pages(struct f2fs_sb_info *sbi, int count_type)
2524 return atomic_read(&sbi->nr_pages[count_type]);
2527 static inline int get_dirty_pages(struct inode *inode)
2529 return atomic_read(&F2FS_I(inode)->dirty_pages);
2532 static inline int get_blocktype_secs(struct f2fs_sb_info *sbi, int block_type)
2534 return div_u64(get_pages(sbi, block_type) + BLKS_PER_SEC(sbi) - 1,
2538 static inline block_t valid_user_blocks(struct f2fs_sb_info *sbi)
2540 return sbi->total_valid_block_count;
2543 static inline block_t discard_blocks(struct f2fs_sb_info *sbi)
2545 return sbi->discard_blks;
2548 static inline unsigned long __bitmap_size(struct f2fs_sb_info *sbi, int flag)
2550 struct f2fs_checkpoint *ckpt = F2FS_CKPT(sbi);
2552 /* return NAT or SIT bitmap */
2553 if (flag == NAT_BITMAP)
2554 return le32_to_cpu(ckpt->nat_ver_bitmap_bytesize);
2555 else if (flag == SIT_BITMAP)
2556 return le32_to_cpu(ckpt->sit_ver_bitmap_bytesize);
2561 static inline block_t __cp_payload(struct f2fs_sb_info *sbi)
2563 return le32_to_cpu(F2FS_RAW_SUPER(sbi)->cp_payload);
2566 static inline void *__bitmap_ptr(struct f2fs_sb_info *sbi, int flag)
2568 struct f2fs_checkpoint *ckpt = F2FS_CKPT(sbi);
2569 void *tmp_ptr = &ckpt->sit_nat_version_bitmap;
2572 if (is_set_ckpt_flags(sbi, CP_LARGE_NAT_BITMAP_FLAG)) {
2573 offset = (flag == SIT_BITMAP) ?
2574 le32_to_cpu(ckpt->nat_ver_bitmap_bytesize) : 0;
2576 * if large_nat_bitmap feature is enabled, leave checksum
2577 * protection for all nat/sit bitmaps.
2579 return tmp_ptr + offset + sizeof(__le32);
2582 if (__cp_payload(sbi) > 0) {
2583 if (flag == NAT_BITMAP)
2586 return (unsigned char *)ckpt + F2FS_BLKSIZE;
2588 offset = (flag == NAT_BITMAP) ?
2589 le32_to_cpu(ckpt->sit_ver_bitmap_bytesize) : 0;
2590 return tmp_ptr + offset;
2594 static inline block_t __start_cp_addr(struct f2fs_sb_info *sbi)
2596 block_t start_addr = le32_to_cpu(F2FS_RAW_SUPER(sbi)->cp_blkaddr);
2598 if (sbi->cur_cp_pack == 2)
2599 start_addr += BLKS_PER_SEG(sbi);
2603 static inline block_t __start_cp_next_addr(struct f2fs_sb_info *sbi)
2605 block_t start_addr = le32_to_cpu(F2FS_RAW_SUPER(sbi)->cp_blkaddr);
2607 if (sbi->cur_cp_pack == 1)
2608 start_addr += BLKS_PER_SEG(sbi);
2612 static inline void __set_cp_next_pack(struct f2fs_sb_info *sbi)
2614 sbi->cur_cp_pack = (sbi->cur_cp_pack == 1) ? 2 : 1;
2617 static inline block_t __start_sum_addr(struct f2fs_sb_info *sbi)
2619 return le32_to_cpu(F2FS_CKPT(sbi)->cp_pack_start_sum);
2622 extern void f2fs_mark_inode_dirty_sync(struct inode *inode, bool sync);
2623 static inline int inc_valid_node_count(struct f2fs_sb_info *sbi,
2624 struct inode *inode, bool is_inode)
2626 block_t valid_block_count;
2627 unsigned int valid_node_count;
2628 unsigned int avail_user_block_count;
2633 err = dquot_alloc_inode(inode);
2638 err = dquot_reserve_block(inode, 1);
2643 if (time_to_inject(sbi, FAULT_BLOCK))
2646 spin_lock(&sbi->stat_lock);
2648 valid_block_count = sbi->total_valid_block_count + 1;
2649 avail_user_block_count = get_available_block_count(sbi, inode, false);
2651 if (unlikely(valid_block_count > avail_user_block_count)) {
2652 spin_unlock(&sbi->stat_lock);
2656 valid_node_count = sbi->total_valid_node_count + 1;
2657 if (unlikely(valid_node_count > sbi->total_node_count)) {
2658 spin_unlock(&sbi->stat_lock);
2662 sbi->total_valid_node_count++;
2663 sbi->total_valid_block_count++;
2664 spin_unlock(&sbi->stat_lock);
2668 f2fs_mark_inode_dirty_sync(inode, true);
2670 f2fs_i_blocks_write(inode, 1, true, true);
2673 percpu_counter_inc(&sbi->alloc_valid_block_count);
2679 dquot_free_inode(inode);
2681 dquot_release_reservation_block(inode, 1);
2686 static inline void dec_valid_node_count(struct f2fs_sb_info *sbi,
2687 struct inode *inode, bool is_inode)
2689 spin_lock(&sbi->stat_lock);
2691 if (unlikely(!sbi->total_valid_block_count ||
2692 !sbi->total_valid_node_count)) {
2693 f2fs_warn(sbi, "dec_valid_node_count: inconsistent block counts, total_valid_block:%u, total_valid_node:%u",
2694 sbi->total_valid_block_count,
2695 sbi->total_valid_node_count);
2696 set_sbi_flag(sbi, SBI_NEED_FSCK);
2698 sbi->total_valid_block_count--;
2699 sbi->total_valid_node_count--;
2702 if (sbi->reserved_blocks &&
2703 sbi->current_reserved_blocks < sbi->reserved_blocks)
2704 sbi->current_reserved_blocks++;
2706 spin_unlock(&sbi->stat_lock);
2709 dquot_free_inode(inode);
2711 if (unlikely(inode->i_blocks == 0)) {
2712 f2fs_warn(sbi, "dec_valid_node_count: inconsistent i_blocks, ino:%lu, iblocks:%llu",
2714 (unsigned long long)inode->i_blocks);
2715 set_sbi_flag(sbi, SBI_NEED_FSCK);
2718 f2fs_i_blocks_write(inode, 1, false, true);
2722 static inline unsigned int valid_node_count(struct f2fs_sb_info *sbi)
2724 return sbi->total_valid_node_count;
2727 static inline void inc_valid_inode_count(struct f2fs_sb_info *sbi)
2729 percpu_counter_inc(&sbi->total_valid_inode_count);
2732 static inline void dec_valid_inode_count(struct f2fs_sb_info *sbi)
2734 percpu_counter_dec(&sbi->total_valid_inode_count);
2737 static inline s64 valid_inode_count(struct f2fs_sb_info *sbi)
2739 return percpu_counter_sum_positive(&sbi->total_valid_inode_count);
2742 static inline struct page *f2fs_grab_cache_page(struct address_space *mapping,
2743 pgoff_t index, bool for_write)
2748 if (IS_ENABLED(CONFIG_F2FS_FAULT_INJECTION)) {
2750 page = find_get_page_flags(mapping, index,
2751 FGP_LOCK | FGP_ACCESSED);
2753 page = find_lock_page(mapping, index);
2757 if (time_to_inject(F2FS_M_SB(mapping), FAULT_PAGE_ALLOC))
2762 return grab_cache_page(mapping, index);
2764 flags = memalloc_nofs_save();
2765 page = grab_cache_page_write_begin(mapping, index);
2766 memalloc_nofs_restore(flags);
2771 static inline struct page *f2fs_pagecache_get_page(
2772 struct address_space *mapping, pgoff_t index,
2773 fgf_t fgp_flags, gfp_t gfp_mask)
2775 if (time_to_inject(F2FS_M_SB(mapping), FAULT_PAGE_GET))
2778 return pagecache_get_page(mapping, index, fgp_flags, gfp_mask);
2781 static inline void f2fs_put_page(struct page *page, int unlock)
2787 f2fs_bug_on(F2FS_P_SB(page), !PageLocked(page));
2793 static inline void f2fs_put_dnode(struct dnode_of_data *dn)
2796 f2fs_put_page(dn->node_page, 1);
2797 if (dn->inode_page && dn->node_page != dn->inode_page)
2798 f2fs_put_page(dn->inode_page, 0);
2799 dn->node_page = NULL;
2800 dn->inode_page = NULL;
2803 static inline struct kmem_cache *f2fs_kmem_cache_create(const char *name,
2806 return kmem_cache_create(name, size, 0, SLAB_RECLAIM_ACCOUNT, NULL);
2809 static inline void *f2fs_kmem_cache_alloc_nofail(struct kmem_cache *cachep,
2814 entry = kmem_cache_alloc(cachep, flags);
2816 entry = kmem_cache_alloc(cachep, flags | __GFP_NOFAIL);
2820 static inline void *f2fs_kmem_cache_alloc(struct kmem_cache *cachep,
2821 gfp_t flags, bool nofail, struct f2fs_sb_info *sbi)
2824 return f2fs_kmem_cache_alloc_nofail(cachep, flags);
2826 if (time_to_inject(sbi, FAULT_SLAB_ALLOC))
2829 return kmem_cache_alloc(cachep, flags);
2832 static inline bool is_inflight_io(struct f2fs_sb_info *sbi, int type)
2834 if (get_pages(sbi, F2FS_RD_DATA) || get_pages(sbi, F2FS_RD_NODE) ||
2835 get_pages(sbi, F2FS_RD_META) || get_pages(sbi, F2FS_WB_DATA) ||
2836 get_pages(sbi, F2FS_WB_CP_DATA) ||
2837 get_pages(sbi, F2FS_DIO_READ) ||
2838 get_pages(sbi, F2FS_DIO_WRITE))
2841 if (type != DISCARD_TIME && SM_I(sbi) && SM_I(sbi)->dcc_info &&
2842 atomic_read(&SM_I(sbi)->dcc_info->queued_discard))
2845 if (SM_I(sbi) && SM_I(sbi)->fcc_info &&
2846 atomic_read(&SM_I(sbi)->fcc_info->queued_flush))
2851 static inline bool is_idle(struct f2fs_sb_info *sbi, int type)
2853 if (sbi->gc_mode == GC_URGENT_HIGH)
2856 if (is_inflight_io(sbi, type))
2859 if (sbi->gc_mode == GC_URGENT_MID)
2862 if (sbi->gc_mode == GC_URGENT_LOW &&
2863 (type == DISCARD_TIME || type == GC_TIME))
2866 return f2fs_time_over(sbi, type);
2869 static inline void f2fs_radix_tree_insert(struct radix_tree_root *root,
2870 unsigned long index, void *item)
2872 while (radix_tree_insert(root, index, item))
2876 #define RAW_IS_INODE(p) ((p)->footer.nid == (p)->footer.ino)
2878 static inline bool IS_INODE(struct page *page)
2880 struct f2fs_node *p = F2FS_NODE(page);
2882 return RAW_IS_INODE(p);
2885 static inline int offset_in_addr(struct f2fs_inode *i)
2887 return (i->i_inline & F2FS_EXTRA_ATTR) ?
2888 (le16_to_cpu(i->i_extra_isize) / sizeof(__le32)) : 0;
2891 static inline __le32 *blkaddr_in_node(struct f2fs_node *node)
2893 return RAW_IS_INODE(node) ? node->i.i_addr : node->dn.addr;
2896 static inline int f2fs_has_extra_attr(struct inode *inode);
2897 static inline block_t data_blkaddr(struct inode *inode,
2898 struct page *node_page, unsigned int offset)
2900 struct f2fs_node *raw_node;
2903 bool is_inode = IS_INODE(node_page);
2905 raw_node = F2FS_NODE(node_page);
2909 /* from GC path only */
2910 base = offset_in_addr(&raw_node->i);
2911 else if (f2fs_has_extra_attr(inode))
2912 base = get_extra_isize(inode);
2915 addr_array = blkaddr_in_node(raw_node);
2916 return le32_to_cpu(addr_array[base + offset]);
2919 static inline block_t f2fs_data_blkaddr(struct dnode_of_data *dn)
2921 return data_blkaddr(dn->inode, dn->node_page, dn->ofs_in_node);
2924 static inline int f2fs_test_bit(unsigned int nr, char *addr)
2929 mask = BIT(7 - (nr & 0x07));
2930 return mask & *addr;
2933 static inline void f2fs_set_bit(unsigned int nr, char *addr)
2938 mask = BIT(7 - (nr & 0x07));
2942 static inline void f2fs_clear_bit(unsigned int nr, char *addr)
2947 mask = BIT(7 - (nr & 0x07));
2951 static inline int f2fs_test_and_set_bit(unsigned int nr, char *addr)
2957 mask = BIT(7 - (nr & 0x07));
2963 static inline int f2fs_test_and_clear_bit(unsigned int nr, char *addr)
2969 mask = BIT(7 - (nr & 0x07));
2975 static inline void f2fs_change_bit(unsigned int nr, char *addr)
2980 mask = BIT(7 - (nr & 0x07));
2985 * On-disk inode flags (f2fs_inode::i_flags)
2987 #define F2FS_COMPR_FL 0x00000004 /* Compress file */
2988 #define F2FS_SYNC_FL 0x00000008 /* Synchronous updates */
2989 #define F2FS_IMMUTABLE_FL 0x00000010 /* Immutable file */
2990 #define F2FS_APPEND_FL 0x00000020 /* writes to file may only append */
2991 #define F2FS_NODUMP_FL 0x00000040 /* do not dump file */
2992 #define F2FS_NOATIME_FL 0x00000080 /* do not update atime */
2993 #define F2FS_NOCOMP_FL 0x00000400 /* Don't compress */
2994 #define F2FS_INDEX_FL 0x00001000 /* hash-indexed directory */
2995 #define F2FS_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */
2996 #define F2FS_PROJINHERIT_FL 0x20000000 /* Create with parents projid */
2997 #define F2FS_CASEFOLD_FL 0x40000000 /* Casefolded file */
2999 #define F2FS_QUOTA_DEFAULT_FL (F2FS_NOATIME_FL | F2FS_IMMUTABLE_FL)
3001 /* Flags that should be inherited by new inodes from their parent. */
3002 #define F2FS_FL_INHERITED (F2FS_SYNC_FL | F2FS_NODUMP_FL | F2FS_NOATIME_FL | \
3003 F2FS_DIRSYNC_FL | F2FS_PROJINHERIT_FL | \
3006 /* Flags that are appropriate for regular files (all but dir-specific ones). */
3007 #define F2FS_REG_FLMASK (~(F2FS_DIRSYNC_FL | F2FS_PROJINHERIT_FL | \
3010 /* Flags that are appropriate for non-directories/regular files. */
3011 #define F2FS_OTHER_FLMASK (F2FS_NODUMP_FL | F2FS_NOATIME_FL)
3013 static inline __u32 f2fs_mask_flags(umode_t mode, __u32 flags)
3017 else if (S_ISREG(mode))
3018 return flags & F2FS_REG_FLMASK;
3020 return flags & F2FS_OTHER_FLMASK;
3023 static inline void __mark_inode_dirty_flag(struct inode *inode,
3027 case FI_INLINE_XATTR:
3028 case FI_INLINE_DATA:
3029 case FI_INLINE_DENTRY:
3035 case FI_INLINE_DOTS:
3037 case FI_COMPRESS_RELEASED:
3038 f2fs_mark_inode_dirty_sync(inode, true);
3042 static inline void set_inode_flag(struct inode *inode, int flag)
3044 set_bit(flag, F2FS_I(inode)->flags);
3045 __mark_inode_dirty_flag(inode, flag, true);
3048 static inline int is_inode_flag_set(struct inode *inode, int flag)
3050 return test_bit(flag, F2FS_I(inode)->flags);
3053 static inline void clear_inode_flag(struct inode *inode, int flag)
3055 clear_bit(flag, F2FS_I(inode)->flags);
3056 __mark_inode_dirty_flag(inode, flag, false);
3059 static inline bool f2fs_verity_in_progress(struct inode *inode)
3061 return IS_ENABLED(CONFIG_FS_VERITY) &&
3062 is_inode_flag_set(inode, FI_VERITY_IN_PROGRESS);
3065 static inline void set_acl_inode(struct inode *inode, umode_t mode)
3067 F2FS_I(inode)->i_acl_mode = mode;
3068 set_inode_flag(inode, FI_ACL_MODE);
3069 f2fs_mark_inode_dirty_sync(inode, false);
3072 static inline void f2fs_i_links_write(struct inode *inode, bool inc)
3078 f2fs_mark_inode_dirty_sync(inode, true);
3081 static inline void f2fs_i_blocks_write(struct inode *inode,
3082 block_t diff, bool add, bool claim)
3084 bool clean = !is_inode_flag_set(inode, FI_DIRTY_INODE);
3085 bool recover = is_inode_flag_set(inode, FI_AUTO_RECOVER);
3087 /* add = 1, claim = 1 should be dquot_reserve_block in pair */
3090 dquot_claim_block(inode, diff);
3092 dquot_alloc_block_nofail(inode, diff);
3094 dquot_free_block(inode, diff);
3097 f2fs_mark_inode_dirty_sync(inode, true);
3098 if (clean || recover)
3099 set_inode_flag(inode, FI_AUTO_RECOVER);
3102 static inline bool f2fs_is_atomic_file(struct inode *inode);
3104 static inline void f2fs_i_size_write(struct inode *inode, loff_t i_size)
3106 bool clean = !is_inode_flag_set(inode, FI_DIRTY_INODE);
3107 bool recover = is_inode_flag_set(inode, FI_AUTO_RECOVER);
3109 if (i_size_read(inode) == i_size)
3112 i_size_write(inode, i_size);
3114 if (f2fs_is_atomic_file(inode))
3117 f2fs_mark_inode_dirty_sync(inode, true);
3118 if (clean || recover)
3119 set_inode_flag(inode, FI_AUTO_RECOVER);
3122 static inline void f2fs_i_depth_write(struct inode *inode, unsigned int depth)
3124 F2FS_I(inode)->i_current_depth = depth;
3125 f2fs_mark_inode_dirty_sync(inode, true);
3128 static inline void f2fs_i_gc_failures_write(struct inode *inode,
3131 F2FS_I(inode)->i_gc_failures[GC_FAILURE_PIN] = count;
3132 f2fs_mark_inode_dirty_sync(inode, true);
3135 static inline void f2fs_i_xnid_write(struct inode *inode, nid_t xnid)
3137 F2FS_I(inode)->i_xattr_nid = xnid;
3138 f2fs_mark_inode_dirty_sync(inode, true);
3141 static inline void f2fs_i_pino_write(struct inode *inode, nid_t pino)
3143 F2FS_I(inode)->i_pino = pino;
3144 f2fs_mark_inode_dirty_sync(inode, true);
3147 static inline void get_inline_info(struct inode *inode, struct f2fs_inode *ri)
3149 struct f2fs_inode_info *fi = F2FS_I(inode);
3151 if (ri->i_inline & F2FS_INLINE_XATTR)
3152 set_bit(FI_INLINE_XATTR, fi->flags);
3153 if (ri->i_inline & F2FS_INLINE_DATA)
3154 set_bit(FI_INLINE_DATA, fi->flags);
3155 if (ri->i_inline & F2FS_INLINE_DENTRY)
3156 set_bit(FI_INLINE_DENTRY, fi->flags);
3157 if (ri->i_inline & F2FS_DATA_EXIST)
3158 set_bit(FI_DATA_EXIST, fi->flags);
3159 if (ri->i_inline & F2FS_INLINE_DOTS)
3160 set_bit(FI_INLINE_DOTS, fi->flags);
3161 if (ri->i_inline & F2FS_EXTRA_ATTR)
3162 set_bit(FI_EXTRA_ATTR, fi->flags);
3163 if (ri->i_inline & F2FS_PIN_FILE)
3164 set_bit(FI_PIN_FILE, fi->flags);
3165 if (ri->i_inline & F2FS_COMPRESS_RELEASED)
3166 set_bit(FI_COMPRESS_RELEASED, fi->flags);
3169 static inline void set_raw_inline(struct inode *inode, struct f2fs_inode *ri)
3173 if (is_inode_flag_set(inode, FI_INLINE_XATTR))
3174 ri->i_inline |= F2FS_INLINE_XATTR;
3175 if (is_inode_flag_set(inode, FI_INLINE_DATA))
3176 ri->i_inline |= F2FS_INLINE_DATA;
3177 if (is_inode_flag_set(inode, FI_INLINE_DENTRY))
3178 ri->i_inline |= F2FS_INLINE_DENTRY;
3179 if (is_inode_flag_set(inode, FI_DATA_EXIST))
3180 ri->i_inline |= F2FS_DATA_EXIST;
3181 if (is_inode_flag_set(inode, FI_INLINE_DOTS))
3182 ri->i_inline |= F2FS_INLINE_DOTS;
3183 if (is_inode_flag_set(inode, FI_EXTRA_ATTR))
3184 ri->i_inline |= F2FS_EXTRA_ATTR;
3185 if (is_inode_flag_set(inode, FI_PIN_FILE))
3186 ri->i_inline |= F2FS_PIN_FILE;
3187 if (is_inode_flag_set(inode, FI_COMPRESS_RELEASED))
3188 ri->i_inline |= F2FS_COMPRESS_RELEASED;
3191 static inline int f2fs_has_extra_attr(struct inode *inode)
3193 return is_inode_flag_set(inode, FI_EXTRA_ATTR);
3196 static inline int f2fs_has_inline_xattr(struct inode *inode)
3198 return is_inode_flag_set(inode, FI_INLINE_XATTR);
3201 static inline int f2fs_compressed_file(struct inode *inode)
3203 return S_ISREG(inode->i_mode) &&
3204 is_inode_flag_set(inode, FI_COMPRESSED_FILE);
3207 static inline bool f2fs_need_compress_data(struct inode *inode)
3209 int compress_mode = F2FS_OPTION(F2FS_I_SB(inode)).compress_mode;
3211 if (!f2fs_compressed_file(inode))
3214 if (compress_mode == COMPR_MODE_FS)
3216 else if (compress_mode == COMPR_MODE_USER &&
3217 is_inode_flag_set(inode, FI_ENABLE_COMPRESS))
3223 static inline unsigned int addrs_per_inode(struct inode *inode)
3225 unsigned int addrs = CUR_ADDRS_PER_INODE(inode) -
3226 get_inline_xattr_addrs(inode);
3228 if (!f2fs_compressed_file(inode))
3230 return ALIGN_DOWN(addrs, F2FS_I(inode)->i_cluster_size);
3233 static inline unsigned int addrs_per_block(struct inode *inode)
3235 if (!f2fs_compressed_file(inode))
3236 return DEF_ADDRS_PER_BLOCK;
3237 return ALIGN_DOWN(DEF_ADDRS_PER_BLOCK, F2FS_I(inode)->i_cluster_size);
3240 static inline void *inline_xattr_addr(struct inode *inode, struct page *page)
3242 struct f2fs_inode *ri = F2FS_INODE(page);
3244 return (void *)&(ri->i_addr[DEF_ADDRS_PER_INODE -
3245 get_inline_xattr_addrs(inode)]);
3248 static inline int inline_xattr_size(struct inode *inode)
3250 if (f2fs_has_inline_xattr(inode))
3251 return get_inline_xattr_addrs(inode) * sizeof(__le32);
3256 * Notice: check inline_data flag without inode page lock is unsafe.
3257 * It could change at any time by f2fs_convert_inline_page().
3259 static inline int f2fs_has_inline_data(struct inode *inode)
3261 return is_inode_flag_set(inode, FI_INLINE_DATA);
3264 static inline int f2fs_exist_data(struct inode *inode)
3266 return is_inode_flag_set(inode, FI_DATA_EXIST);
3269 static inline int f2fs_has_inline_dots(struct inode *inode)
3271 return is_inode_flag_set(inode, FI_INLINE_DOTS);
3274 static inline int f2fs_is_mmap_file(struct inode *inode)
3276 return is_inode_flag_set(inode, FI_MMAP_FILE);
3279 static inline bool f2fs_is_pinned_file(struct inode *inode)
3281 return is_inode_flag_set(inode, FI_PIN_FILE);
3284 static inline bool f2fs_is_atomic_file(struct inode *inode)
3286 return is_inode_flag_set(inode, FI_ATOMIC_FILE);
3289 static inline bool f2fs_is_cow_file(struct inode *inode)
3291 return is_inode_flag_set(inode, FI_COW_FILE);
3294 static inline __le32 *get_dnode_addr(struct inode *inode,
3295 struct page *node_page);
3296 static inline void *inline_data_addr(struct inode *inode, struct page *page)
3298 __le32 *addr = get_dnode_addr(inode, page);
3300 return (void *)(addr + DEF_INLINE_RESERVED_SIZE);
3303 static inline int f2fs_has_inline_dentry(struct inode *inode)
3305 return is_inode_flag_set(inode, FI_INLINE_DENTRY);
3308 static inline int is_file(struct inode *inode, int type)
3310 return F2FS_I(inode)->i_advise & type;
3313 static inline void set_file(struct inode *inode, int type)
3315 if (is_file(inode, type))
3317 F2FS_I(inode)->i_advise |= type;
3318 f2fs_mark_inode_dirty_sync(inode, true);
3321 static inline void clear_file(struct inode *inode, int type)
3323 if (!is_file(inode, type))
3325 F2FS_I(inode)->i_advise &= ~type;
3326 f2fs_mark_inode_dirty_sync(inode, true);
3329 static inline bool f2fs_is_time_consistent(struct inode *inode)
3331 struct timespec64 ts = inode_get_atime(inode);
3333 if (!timespec64_equal(F2FS_I(inode)->i_disk_time, &ts))
3335 ts = inode_get_ctime(inode);
3336 if (!timespec64_equal(F2FS_I(inode)->i_disk_time + 1, &ts))
3338 ts = inode_get_mtime(inode);
3339 if (!timespec64_equal(F2FS_I(inode)->i_disk_time + 2, &ts))
3344 static inline bool f2fs_skip_inode_update(struct inode *inode, int dsync)
3349 struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
3351 spin_lock(&sbi->inode_lock[DIRTY_META]);
3352 ret = list_empty(&F2FS_I(inode)->gdirty_list);
3353 spin_unlock(&sbi->inode_lock[DIRTY_META]);
3356 if (!is_inode_flag_set(inode, FI_AUTO_RECOVER) ||
3357 file_keep_isize(inode) ||
3358 i_size_read(inode) & ~PAGE_MASK)
3361 if (!f2fs_is_time_consistent(inode))
3364 spin_lock(&F2FS_I(inode)->i_size_lock);
3365 ret = F2FS_I(inode)->last_disk_size == i_size_read(inode);
3366 spin_unlock(&F2FS_I(inode)->i_size_lock);
3371 static inline bool f2fs_readonly(struct super_block *sb)
3373 return sb_rdonly(sb);
3376 static inline bool f2fs_cp_error(struct f2fs_sb_info *sbi)
3378 return is_set_ckpt_flags(sbi, CP_ERROR_FLAG);
3381 static inline bool is_dot_dotdot(const u8 *name, size_t len)
3383 if (len == 1 && name[0] == '.')
3386 if (len == 2 && name[0] == '.' && name[1] == '.')
3392 static inline void *f2fs_kmalloc(struct f2fs_sb_info *sbi,
3393 size_t size, gfp_t flags)
3395 if (time_to_inject(sbi, FAULT_KMALLOC))
3398 return kmalloc(size, flags);
3401 static inline void *f2fs_getname(struct f2fs_sb_info *sbi)
3403 if (time_to_inject(sbi, FAULT_KMALLOC))
3409 static inline void f2fs_putname(char *buf)
3414 static inline void *f2fs_kzalloc(struct f2fs_sb_info *sbi,
3415 size_t size, gfp_t flags)
3417 return f2fs_kmalloc(sbi, size, flags | __GFP_ZERO);
3420 static inline void *f2fs_kvmalloc(struct f2fs_sb_info *sbi,
3421 size_t size, gfp_t flags)
3423 if (time_to_inject(sbi, FAULT_KVMALLOC))
3426 return kvmalloc(size, flags);
3429 static inline void *f2fs_kvzalloc(struct f2fs_sb_info *sbi,
3430 size_t size, gfp_t flags)
3432 return f2fs_kvmalloc(sbi, size, flags | __GFP_ZERO);
3435 static inline int get_extra_isize(struct inode *inode)
3437 return F2FS_I(inode)->i_extra_isize / sizeof(__le32);
3440 static inline int get_inline_xattr_addrs(struct inode *inode)
3442 return F2FS_I(inode)->i_inline_xattr_size;
3445 static inline __le32 *get_dnode_addr(struct inode *inode,
3446 struct page *node_page)
3450 if (IS_INODE(node_page) && f2fs_has_extra_attr(inode))
3451 base = get_extra_isize(inode);
3453 return blkaddr_in_node(F2FS_NODE(node_page)) + base;
3456 #define f2fs_get_inode_mode(i) \
3457 ((is_inode_flag_set(i, FI_ACL_MODE)) ? \
3458 (F2FS_I(i)->i_acl_mode) : ((i)->i_mode))
3460 #define F2FS_MIN_EXTRA_ATTR_SIZE (sizeof(__le32))
3462 #define F2FS_TOTAL_EXTRA_ATTR_SIZE \
3463 (offsetof(struct f2fs_inode, i_extra_end) - \
3464 offsetof(struct f2fs_inode, i_extra_isize)) \
3466 #define F2FS_OLD_ATTRIBUTE_SIZE (offsetof(struct f2fs_inode, i_addr))
3467 #define F2FS_FITS_IN_INODE(f2fs_inode, extra_isize, field) \
3468 ((offsetof(typeof(*(f2fs_inode)), field) + \
3469 sizeof((f2fs_inode)->field)) \
3470 <= (F2FS_OLD_ATTRIBUTE_SIZE + (extra_isize))) \
3472 #define __is_large_section(sbi) (SEGS_PER_SEC(sbi) > 1)
3474 #define __is_meta_io(fio) (PAGE_TYPE_OF_BIO((fio)->type) == META)
3476 bool f2fs_is_valid_blkaddr(struct f2fs_sb_info *sbi,
3477 block_t blkaddr, int type);
3478 static inline void verify_blkaddr(struct f2fs_sb_info *sbi,
3479 block_t blkaddr, int type)
3481 if (!f2fs_is_valid_blkaddr(sbi, blkaddr, type))
3482 f2fs_err(sbi, "invalid blkaddr: %u, type: %d, run fsck to fix.",
3486 static inline bool __is_valid_data_blkaddr(block_t blkaddr)
3488 if (blkaddr == NEW_ADDR || blkaddr == NULL_ADDR ||
3489 blkaddr == COMPRESS_ADDR)
3497 int f2fs_sync_file(struct file *file, loff_t start, loff_t end, int datasync);
3498 int f2fs_do_truncate_blocks(struct inode *inode, u64 from, bool lock);
3499 int f2fs_truncate_blocks(struct inode *inode, u64 from, bool lock);
3500 int f2fs_truncate(struct inode *inode);
3501 int f2fs_getattr(struct mnt_idmap *idmap, const struct path *path,
3502 struct kstat *stat, u32 request_mask, unsigned int flags);
3503 int f2fs_setattr(struct mnt_idmap *idmap, struct dentry *dentry,
3504 struct iattr *attr);
3505 int f2fs_truncate_hole(struct inode *inode, pgoff_t pg_start, pgoff_t pg_end);
3506 void f2fs_truncate_data_blocks_range(struct dnode_of_data *dn, int count);
3507 int f2fs_precache_extents(struct inode *inode);
3508 int f2fs_fileattr_get(struct dentry *dentry, struct fileattr *fa);
3509 int f2fs_fileattr_set(struct mnt_idmap *idmap,
3510 struct dentry *dentry, struct fileattr *fa);
3511 long f2fs_ioctl(struct file *filp, unsigned int cmd, unsigned long arg);
3512 long f2fs_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
3513 int f2fs_transfer_project_quota(struct inode *inode, kprojid_t kprojid);
3514 int f2fs_pin_file_control(struct inode *inode, bool inc);
3519 void f2fs_set_inode_flags(struct inode *inode);
3520 bool f2fs_inode_chksum_verify(struct f2fs_sb_info *sbi, struct page *page);
3521 void f2fs_inode_chksum_set(struct f2fs_sb_info *sbi, struct page *page);
3522 struct inode *f2fs_iget(struct super_block *sb, unsigned long ino);
3523 struct inode *f2fs_iget_retry(struct super_block *sb, unsigned long ino);
3524 int f2fs_try_to_free_nats(struct f2fs_sb_info *sbi, int nr_shrink);
3525 void f2fs_update_inode(struct inode *inode, struct page *node_page);
3526 void f2fs_update_inode_page(struct inode *inode);
3527 int f2fs_write_inode(struct inode *inode, struct writeback_control *wbc);
3528 void f2fs_evict_inode(struct inode *inode);
3529 void f2fs_handle_failed_inode(struct inode *inode);
3534 int f2fs_update_extension_list(struct f2fs_sb_info *sbi, const char *name,
3535 bool hot, bool set);
3536 struct dentry *f2fs_get_parent(struct dentry *child);
3537 int f2fs_get_tmpfile(struct mnt_idmap *idmap, struct inode *dir,
3538 struct inode **new_inode);
3543 int f2fs_init_casefolded_name(const struct inode *dir,
3544 struct f2fs_filename *fname);
3545 int f2fs_setup_filename(struct inode *dir, const struct qstr *iname,
3546 int lookup, struct f2fs_filename *fname);
3547 int f2fs_prepare_lookup(struct inode *dir, struct dentry *dentry,
3548 struct f2fs_filename *fname);
3549 void f2fs_free_filename(struct f2fs_filename *fname);
3550 struct f2fs_dir_entry *f2fs_find_target_dentry(const struct f2fs_dentry_ptr *d,
3551 const struct f2fs_filename *fname, int *max_slots);
3552 int f2fs_fill_dentries(struct dir_context *ctx, struct f2fs_dentry_ptr *d,
3553 unsigned int start_pos, struct fscrypt_str *fstr);
3554 void f2fs_do_make_empty_dir(struct inode *inode, struct inode *parent,
3555 struct f2fs_dentry_ptr *d);
3556 struct page *f2fs_init_inode_metadata(struct inode *inode, struct inode *dir,
3557 const struct f2fs_filename *fname, struct page *dpage);
3558 void f2fs_update_parent_metadata(struct inode *dir, struct inode *inode,
3559 unsigned int current_depth);
3560 int f2fs_room_for_filename(const void *bitmap, int slots, int max_slots);
3561 void f2fs_drop_nlink(struct inode *dir, struct inode *inode);
3562 struct f2fs_dir_entry *__f2fs_find_entry(struct inode *dir,
3563 const struct f2fs_filename *fname,
3564 struct page **res_page);
3565 struct f2fs_dir_entry *f2fs_find_entry(struct inode *dir,
3566 const struct qstr *child, struct page **res_page);
3567 struct f2fs_dir_entry *f2fs_parent_dir(struct inode *dir, struct page **p);
3568 ino_t f2fs_inode_by_name(struct inode *dir, const struct qstr *qstr,
3569 struct page **page);
3570 void f2fs_set_link(struct inode *dir, struct f2fs_dir_entry *de,
3571 struct page *page, struct inode *inode);
3572 bool f2fs_has_enough_room(struct inode *dir, struct page *ipage,
3573 const struct f2fs_filename *fname);
3574 void f2fs_update_dentry(nid_t ino, umode_t mode, struct f2fs_dentry_ptr *d,
3575 const struct fscrypt_str *name, f2fs_hash_t name_hash,
3576 unsigned int bit_pos);
3577 int f2fs_add_regular_entry(struct inode *dir, const struct f2fs_filename *fname,
3578 struct inode *inode, nid_t ino, umode_t mode);
3579 int f2fs_add_dentry(struct inode *dir, const struct f2fs_filename *fname,
3580 struct inode *inode, nid_t ino, umode_t mode);
3581 int f2fs_do_add_link(struct inode *dir, const struct qstr *name,
3582 struct inode *inode, nid_t ino, umode_t mode);
3583 void f2fs_delete_entry(struct f2fs_dir_entry *dentry, struct page *page,
3584 struct inode *dir, struct inode *inode);
3585 int f2fs_do_tmpfile(struct inode *inode, struct inode *dir,
3586 struct f2fs_filename *fname);
3587 bool f2fs_empty_dir(struct inode *dir);
3589 static inline int f2fs_add_link(struct dentry *dentry, struct inode *inode)
3591 if (fscrypt_is_nokey_name(dentry))
3593 return f2fs_do_add_link(d_inode(dentry->d_parent), &dentry->d_name,
3594 inode, inode->i_ino, inode->i_mode);
3600 int f2fs_inode_dirtied(struct inode *inode, bool sync);
3601 void f2fs_inode_synced(struct inode *inode);
3602 int f2fs_dquot_initialize(struct inode *inode);
3603 int f2fs_enable_quota_files(struct f2fs_sb_info *sbi, bool rdonly);
3604 int f2fs_quota_sync(struct super_block *sb, int type);
3605 loff_t max_file_blocks(struct inode *inode);
3606 void f2fs_quota_off_umount(struct super_block *sb);
3607 void f2fs_save_errors(struct f2fs_sb_info *sbi, unsigned char flag);
3608 void f2fs_handle_critical_error(struct f2fs_sb_info *sbi, unsigned char reason,
3610 void f2fs_handle_error(struct f2fs_sb_info *sbi, unsigned char error);
3611 void f2fs_handle_error_async(struct f2fs_sb_info *sbi, unsigned char error);
3612 int f2fs_commit_super(struct f2fs_sb_info *sbi, bool recover);
3613 int f2fs_sync_fs(struct super_block *sb, int sync);
3614 int f2fs_sanity_check_ckpt(struct f2fs_sb_info *sbi);
3619 void f2fs_hash_filename(const struct inode *dir, struct f2fs_filename *fname);
3626 int f2fs_check_nid_range(struct f2fs_sb_info *sbi, nid_t nid);
3627 bool f2fs_available_free_memory(struct f2fs_sb_info *sbi, int type);
3628 bool f2fs_in_warm_node_list(struct f2fs_sb_info *sbi, struct page *page);
3629 void f2fs_init_fsync_node_info(struct f2fs_sb_info *sbi);
3630 void f2fs_del_fsync_node_entry(struct f2fs_sb_info *sbi, struct page *page);
3631 void f2fs_reset_fsync_node_info(struct f2fs_sb_info *sbi);
3632 int f2fs_need_dentry_mark(struct f2fs_sb_info *sbi, nid_t nid);
3633 bool f2fs_is_checkpointed_node(struct f2fs_sb_info *sbi, nid_t nid);
3634 bool f2fs_need_inode_block_update(struct f2fs_sb_info *sbi, nid_t ino);
3635 int f2fs_get_node_info(struct f2fs_sb_info *sbi, nid_t nid,
3636 struct node_info *ni, bool checkpoint_context);
3637 pgoff_t f2fs_get_next_page_offset(struct dnode_of_data *dn, pgoff_t pgofs);
3638 int f2fs_get_dnode_of_data(struct dnode_of_data *dn, pgoff_t index, int mode);
3639 int f2fs_truncate_inode_blocks(struct inode *inode, pgoff_t from);
3640 int f2fs_truncate_xattr_node(struct inode *inode);
3641 int f2fs_wait_on_node_pages_writeback(struct f2fs_sb_info *sbi,
3642 unsigned int seq_id);
3643 bool f2fs_nat_bitmap_enabled(struct f2fs_sb_info *sbi);
3644 int f2fs_remove_inode_page(struct inode *inode);
3645 struct page *f2fs_new_inode_page(struct inode *inode);
3646 struct page *f2fs_new_node_page(struct dnode_of_data *dn, unsigned int ofs);
3647 void f2fs_ra_node_page(struct f2fs_sb_info *sbi, nid_t nid);
3648 struct page *f2fs_get_node_page(struct f2fs_sb_info *sbi, pgoff_t nid);
3649 struct page *f2fs_get_node_page_ra(struct page *parent, int start);
3650 int f2fs_move_node_page(struct page *node_page, int gc_type);
3651 void f2fs_flush_inline_data(struct f2fs_sb_info *sbi);
3652 int f2fs_fsync_node_pages(struct f2fs_sb_info *sbi, struct inode *inode,
3653 struct writeback_control *wbc, bool atomic,
3654 unsigned int *seq_id);
3655 int f2fs_sync_node_pages(struct f2fs_sb_info *sbi,
3656 struct writeback_control *wbc,
3657 bool do_balance, enum iostat_type io_type);
3658 int f2fs_build_free_nids(struct f2fs_sb_info *sbi, bool sync, bool mount);
3659 bool f2fs_alloc_nid(struct f2fs_sb_info *sbi, nid_t *nid);
3660 void f2fs_alloc_nid_done(struct f2fs_sb_info *sbi, nid_t nid);
3661 void f2fs_alloc_nid_failed(struct f2fs_sb_info *sbi, nid_t nid);
3662 int f2fs_try_to_free_nids(struct f2fs_sb_info *sbi, int nr_shrink);
3663 int f2fs_recover_inline_xattr(struct inode *inode, struct page *page);
3664 int f2fs_recover_xattr_data(struct inode *inode, struct page *page);
3665 int f2fs_recover_inode_page(struct f2fs_sb_info *sbi, struct page *page);
3666 int f2fs_restore_node_summary(struct f2fs_sb_info *sbi,
3667 unsigned int segno, struct f2fs_summary_block *sum);
3668 void f2fs_enable_nat_bits(struct f2fs_sb_info *sbi);
3669 int f2fs_flush_nat_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc);
3670 int f2fs_build_node_manager(struct f2fs_sb_info *sbi);
3671 void f2fs_destroy_node_manager(struct f2fs_sb_info *sbi);
3672 int __init f2fs_create_node_manager_caches(void);
3673 void f2fs_destroy_node_manager_caches(void);
3678 bool f2fs_need_SSR(struct f2fs_sb_info *sbi);
3679 int f2fs_commit_atomic_write(struct inode *inode);
3680 void f2fs_abort_atomic_write(struct inode *inode, bool clean);
3681 void f2fs_balance_fs(struct f2fs_sb_info *sbi, bool need);
3682 void f2fs_balance_fs_bg(struct f2fs_sb_info *sbi, bool from_bg);
3683 int f2fs_issue_flush(struct f2fs_sb_info *sbi, nid_t ino);
3684 int f2fs_create_flush_cmd_control(struct f2fs_sb_info *sbi);
3685 int f2fs_flush_device_cache(struct f2fs_sb_info *sbi);
3686 void f2fs_destroy_flush_cmd_control(struct f2fs_sb_info *sbi, bool free);
3687 void f2fs_invalidate_blocks(struct f2fs_sb_info *sbi, block_t addr);
3688 bool f2fs_is_checkpointed_data(struct f2fs_sb_info *sbi, block_t blkaddr);
3689 int f2fs_start_discard_thread(struct f2fs_sb_info *sbi);
3690 void f2fs_drop_discard_cmd(struct f2fs_sb_info *sbi);
3691 void f2fs_stop_discard_thread(struct f2fs_sb_info *sbi);
3692 bool f2fs_issue_discard_timeout(struct f2fs_sb_info *sbi);
3693 void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
3694 struct cp_control *cpc);
3695 void f2fs_dirty_to_prefree(struct f2fs_sb_info *sbi);
3696 block_t f2fs_get_unusable_blocks(struct f2fs_sb_info *sbi);
3697 int f2fs_disable_cp_again(struct f2fs_sb_info *sbi, block_t unusable);
3698 void f2fs_release_discard_addrs(struct f2fs_sb_info *sbi);
3699 int f2fs_npages_for_summary_flush(struct f2fs_sb_info *sbi, bool for_ra);
3700 bool f2fs_segment_has_free_slot(struct f2fs_sb_info *sbi, int segno);
3701 void f2fs_init_inmem_curseg(struct f2fs_sb_info *sbi);
3702 void f2fs_save_inmem_curseg(struct f2fs_sb_info *sbi);
3703 void f2fs_restore_inmem_curseg(struct f2fs_sb_info *sbi);
3704 void f2fs_allocate_segment_for_resize(struct f2fs_sb_info *sbi, int type,
3705 unsigned int start, unsigned int end);
3706 int f2fs_allocate_new_section(struct f2fs_sb_info *sbi, int type, bool force);
3707 int f2fs_allocate_pinning_section(struct f2fs_sb_info *sbi);
3708 void f2fs_allocate_new_segments(struct f2fs_sb_info *sbi);
3709 int f2fs_trim_fs(struct f2fs_sb_info *sbi, struct fstrim_range *range);
3710 bool f2fs_exist_trim_candidates(struct f2fs_sb_info *sbi,
3711 struct cp_control *cpc);
3712 struct page *f2fs_get_sum_page(struct f2fs_sb_info *sbi, unsigned int segno);
3713 void f2fs_update_meta_page(struct f2fs_sb_info *sbi, void *src,
3715 void f2fs_do_write_meta_page(struct f2fs_sb_info *sbi, struct page *page,
3716 enum iostat_type io_type);
3717 void f2fs_do_write_node_page(unsigned int nid, struct f2fs_io_info *fio);
3718 void f2fs_outplace_write_data(struct dnode_of_data *dn,
3719 struct f2fs_io_info *fio);
3720 int f2fs_inplace_write_data(struct f2fs_io_info *fio);
3721 void f2fs_do_replace_block(struct f2fs_sb_info *sbi, struct f2fs_summary *sum,
3722 block_t old_blkaddr, block_t new_blkaddr,
3723 bool recover_curseg, bool recover_newaddr,
3725 void f2fs_replace_block(struct f2fs_sb_info *sbi, struct dnode_of_data *dn,
3726 block_t old_addr, block_t new_addr,
3727 unsigned char version, bool recover_curseg,
3728 bool recover_newaddr);
3729 int f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page,
3730 block_t old_blkaddr, block_t *new_blkaddr,
3731 struct f2fs_summary *sum, int type,
3732 struct f2fs_io_info *fio);
3733 void f2fs_update_device_state(struct f2fs_sb_info *sbi, nid_t ino,
3734 block_t blkaddr, unsigned int blkcnt);
3735 void f2fs_wait_on_page_writeback(struct page *page,
3736 enum page_type type, bool ordered, bool locked);
3737 void f2fs_wait_on_block_writeback(struct inode *inode, block_t blkaddr);
3738 void f2fs_wait_on_block_writeback_range(struct inode *inode, block_t blkaddr,
3740 void f2fs_write_data_summaries(struct f2fs_sb_info *sbi, block_t start_blk);
3741 void f2fs_write_node_summaries(struct f2fs_sb_info *sbi, block_t start_blk);
3742 int f2fs_lookup_journal_in_cursum(struct f2fs_journal *journal, int type,
3743 unsigned int val, int alloc);
3744 void f2fs_flush_sit_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc);
3745 int f2fs_fix_curseg_write_pointer(struct f2fs_sb_info *sbi);
3746 int f2fs_check_write_pointer(struct f2fs_sb_info *sbi);
3747 int f2fs_build_segment_manager(struct f2fs_sb_info *sbi);
3748 void f2fs_destroy_segment_manager(struct f2fs_sb_info *sbi);
3749 int __init f2fs_create_segment_manager_caches(void);
3750 void f2fs_destroy_segment_manager_caches(void);
3751 int f2fs_rw_hint_to_seg_type(enum rw_hint hint);
3752 unsigned int f2fs_usable_segs_in_sec(struct f2fs_sb_info *sbi,
3753 unsigned int segno);
3754 unsigned int f2fs_usable_blks_in_seg(struct f2fs_sb_info *sbi,
3755 unsigned int segno);
3757 #define DEF_FRAGMENT_SIZE 4
3758 #define MIN_FRAGMENT_SIZE 1
3759 #define MAX_FRAGMENT_SIZE 512
3761 static inline bool f2fs_need_rand_seg(struct f2fs_sb_info *sbi)
3763 return F2FS_OPTION(sbi).fs_mode == FS_MODE_FRAGMENT_SEG ||
3764 F2FS_OPTION(sbi).fs_mode == FS_MODE_FRAGMENT_BLK;
3770 void f2fs_stop_checkpoint(struct f2fs_sb_info *sbi, bool end_io,
3771 unsigned char reason);
3772 void f2fs_flush_ckpt_thread(struct f2fs_sb_info *sbi);
3773 struct page *f2fs_grab_meta_page(struct f2fs_sb_info *sbi, pgoff_t index);
3774 struct page *f2fs_get_meta_page(struct f2fs_sb_info *sbi, pgoff_t index);
3775 struct page *f2fs_get_meta_page_retry(struct f2fs_sb_info *sbi, pgoff_t index);
3776 struct page *f2fs_get_tmp_page(struct f2fs_sb_info *sbi, pgoff_t index);
3777 bool f2fs_is_valid_blkaddr(struct f2fs_sb_info *sbi,
3778 block_t blkaddr, int type);
3779 bool f2fs_is_valid_blkaddr_raw(struct f2fs_sb_info *sbi,
3780 block_t blkaddr, int type);
3781 int f2fs_ra_meta_pages(struct f2fs_sb_info *sbi, block_t start, int nrpages,
3782 int type, bool sync);
3783 void f2fs_ra_meta_pages_cond(struct f2fs_sb_info *sbi, pgoff_t index,
3784 unsigned int ra_blocks);
3785 long f2fs_sync_meta_pages(struct f2fs_sb_info *sbi, enum page_type type,
3786 long nr_to_write, enum iostat_type io_type);
3787 void f2fs_add_ino_entry(struct f2fs_sb_info *sbi, nid_t ino, int type);
3788 void f2fs_remove_ino_entry(struct f2fs_sb_info *sbi, nid_t ino, int type);
3789 void f2fs_release_ino_entry(struct f2fs_sb_info *sbi, bool all);
3790 bool f2fs_exist_written_data(struct f2fs_sb_info *sbi, nid_t ino, int mode);
3791 void f2fs_set_dirty_device(struct f2fs_sb_info *sbi, nid_t ino,
3792 unsigned int devidx, int type);
3793 bool f2fs_is_dirty_device(struct f2fs_sb_info *sbi, nid_t ino,
3794 unsigned int devidx, int type);
3795 int f2fs_acquire_orphan_inode(struct f2fs_sb_info *sbi);
3796 void f2fs_release_orphan_inode(struct f2fs_sb_info *sbi);
3797 void f2fs_add_orphan_inode(struct inode *inode);
3798 void f2fs_remove_orphan_inode(struct f2fs_sb_info *sbi, nid_t ino);
3799 int f2fs_recover_orphan_inodes(struct f2fs_sb_info *sbi);
3800 int f2fs_get_valid_checkpoint(struct f2fs_sb_info *sbi);
3801 void f2fs_update_dirty_folio(struct inode *inode, struct folio *folio);
3802 void f2fs_remove_dirty_inode(struct inode *inode);
3803 int f2fs_sync_dirty_inodes(struct f2fs_sb_info *sbi, enum inode_type type,
3805 void f2fs_wait_on_all_pages(struct f2fs_sb_info *sbi, int type);
3806 u64 f2fs_get_sectors_written(struct f2fs_sb_info *sbi);
3807 int f2fs_write_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc);
3808 void f2fs_init_ino_entry_info(struct f2fs_sb_info *sbi);
3809 int __init f2fs_create_checkpoint_caches(void);
3810 void f2fs_destroy_checkpoint_caches(void);
3811 int f2fs_issue_checkpoint(struct f2fs_sb_info *sbi);
3812 int f2fs_start_ckpt_thread(struct f2fs_sb_info *sbi);
3813 void f2fs_stop_ckpt_thread(struct f2fs_sb_info *sbi);
3814 void f2fs_init_ckpt_req_control(struct f2fs_sb_info *sbi);
3819 int __init f2fs_init_bioset(void);
3820 void f2fs_destroy_bioset(void);
3821 bool f2fs_is_cp_guaranteed(struct page *page);
3822 int f2fs_init_bio_entry_cache(void);
3823 void f2fs_destroy_bio_entry_cache(void);
3824 void f2fs_submit_read_bio(struct f2fs_sb_info *sbi, struct bio *bio,
3825 enum page_type type);
3826 int f2fs_init_write_merge_io(struct f2fs_sb_info *sbi);
3827 void f2fs_submit_merged_write(struct f2fs_sb_info *sbi, enum page_type type);
3828 void f2fs_submit_merged_write_cond(struct f2fs_sb_info *sbi,
3829 struct inode *inode, struct page *page,
3830 nid_t ino, enum page_type type);
3831 void f2fs_submit_merged_ipu_write(struct f2fs_sb_info *sbi,
3832 struct bio **bio, struct page *page);
3833 void f2fs_flush_merged_writes(struct f2fs_sb_info *sbi);
3834 int f2fs_submit_page_bio(struct f2fs_io_info *fio);
3835 int f2fs_merge_page_bio(struct f2fs_io_info *fio);
3836 void f2fs_submit_page_write(struct f2fs_io_info *fio);
3837 struct block_device *f2fs_target_device(struct f2fs_sb_info *sbi,
3838 block_t blk_addr, sector_t *sector);
3839 int f2fs_target_device_index(struct f2fs_sb_info *sbi, block_t blkaddr);
3840 void f2fs_set_data_blkaddr(struct dnode_of_data *dn, block_t blkaddr);
3841 void f2fs_update_data_blkaddr(struct dnode_of_data *dn, block_t blkaddr);
3842 int f2fs_reserve_new_blocks(struct dnode_of_data *dn, blkcnt_t count);
3843 int f2fs_reserve_new_block(struct dnode_of_data *dn);
3844 int f2fs_get_block_locked(struct dnode_of_data *dn, pgoff_t index);
3845 int f2fs_reserve_block(struct dnode_of_data *dn, pgoff_t index);
3846 struct page *f2fs_get_read_data_page(struct inode *inode, pgoff_t index,
3847 blk_opf_t op_flags, bool for_write, pgoff_t *next_pgofs);
3848 struct page *f2fs_find_data_page(struct inode *inode, pgoff_t index,
3849 pgoff_t *next_pgofs);
3850 struct page *f2fs_get_lock_data_page(struct inode *inode, pgoff_t index,
3852 struct page *f2fs_get_new_data_page(struct inode *inode,
3853 struct page *ipage, pgoff_t index, bool new_i_size);
3854 int f2fs_do_write_data_page(struct f2fs_io_info *fio);
3855 int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map, int flag);
3856 int f2fs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
3857 u64 start, u64 len);
3858 int f2fs_encrypt_one_page(struct f2fs_io_info *fio);
3859 bool f2fs_should_update_inplace(struct inode *inode, struct f2fs_io_info *fio);
3860 bool f2fs_should_update_outplace(struct inode *inode, struct f2fs_io_info *fio);
3861 int f2fs_write_single_data_page(struct page *page, int *submitted,
3862 struct bio **bio, sector_t *last_block,
3863 struct writeback_control *wbc,
3864 enum iostat_type io_type,
3865 int compr_blocks, bool allow_balance);
3866 void f2fs_write_failed(struct inode *inode, loff_t to);
3867 void f2fs_invalidate_folio(struct folio *folio, size_t offset, size_t length);
3868 bool f2fs_release_folio(struct folio *folio, gfp_t wait);
3869 bool f2fs_overwrite_io(struct inode *inode, loff_t pos, size_t len);
3870 void f2fs_clear_page_cache_dirty_tag(struct page *page);
3871 int f2fs_init_post_read_processing(void);
3872 void f2fs_destroy_post_read_processing(void);
3873 int f2fs_init_post_read_wq(struct f2fs_sb_info *sbi);
3874 void f2fs_destroy_post_read_wq(struct f2fs_sb_info *sbi);
3875 extern const struct iomap_ops f2fs_iomap_ops;
3880 int f2fs_start_gc_thread(struct f2fs_sb_info *sbi);
3881 void f2fs_stop_gc_thread(struct f2fs_sb_info *sbi);
3882 block_t f2fs_start_bidx_of_node(unsigned int node_ofs, struct inode *inode);
3883 int f2fs_gc(struct f2fs_sb_info *sbi, struct f2fs_gc_control *gc_control);
3884 void f2fs_build_gc_manager(struct f2fs_sb_info *sbi);
3885 int f2fs_gc_range(struct f2fs_sb_info *sbi,
3886 unsigned int start_seg, unsigned int end_seg,
3887 bool dry_run, unsigned int dry_run_sections);
3888 int f2fs_resize_fs(struct file *filp, __u64 block_count);
3889 int __init f2fs_create_garbage_collection_cache(void);
3890 void f2fs_destroy_garbage_collection_cache(void);
3891 /* victim selection function for cleaning and SSR */
3892 int f2fs_get_victim(struct f2fs_sb_info *sbi, unsigned int *result,
3893 int gc_type, int type, char alloc_mode,
3894 unsigned long long age);
3899 int f2fs_recover_fsync_data(struct f2fs_sb_info *sbi, bool check_only);
3900 bool f2fs_space_for_roll_forward(struct f2fs_sb_info *sbi);
3901 int __init f2fs_create_recovery_cache(void);
3902 void f2fs_destroy_recovery_cache(void);
3907 #ifdef CONFIG_F2FS_STAT_FS
3908 struct f2fs_stat_info {
3909 struct list_head stat_list;
3910 struct f2fs_sb_info *sbi;
3911 int all_area_segs, sit_area_segs, nat_area_segs, ssa_area_segs;
3912 int main_area_segs, main_area_sections, main_area_zones;
3913 unsigned long long hit_cached[NR_EXTENT_CACHES];
3914 unsigned long long hit_rbtree[NR_EXTENT_CACHES];
3915 unsigned long long total_ext[NR_EXTENT_CACHES];
3916 unsigned long long hit_total[NR_EXTENT_CACHES];
3917 int ext_tree[NR_EXTENT_CACHES];
3918 int zombie_tree[NR_EXTENT_CACHES];
3919 int ext_node[NR_EXTENT_CACHES];
3920 /* to count memory footprint */
3921 unsigned long long ext_mem[NR_EXTENT_CACHES];
3922 /* for read extent cache */
3923 unsigned long long hit_largest;
3924 /* for block age extent cache */
3925 unsigned long long allocated_data_blocks;
3926 int ndirty_node, ndirty_dent, ndirty_meta, ndirty_imeta;
3927 int ndirty_data, ndirty_qdata;
3928 unsigned int ndirty_dirs, ndirty_files, nquota_files, ndirty_all;
3929 int nats, dirty_nats, sits, dirty_sits;
3930 int free_nids, avail_nids, alloc_nids;
3931 int total_count, utilization;
3932 int nr_wb_cp_data, nr_wb_data;
3933 int nr_rd_data, nr_rd_node, nr_rd_meta;
3934 int nr_dio_read, nr_dio_write;
3935 unsigned int io_skip_bggc, other_skip_bggc;
3936 int nr_flushing, nr_flushed, flush_list_empty;
3937 int nr_discarding, nr_discarded;
3939 unsigned int undiscard_blks;
3940 int nr_issued_ckpt, nr_total_ckpt, nr_queued_ckpt;
3941 unsigned int cur_ckpt_time, peak_ckpt_time;
3942 int inline_xattr, inline_inode, inline_dir, append, update, orphans;
3943 int compr_inode, swapfile_inode;
3944 unsigned long long compr_blocks;
3945 int aw_cnt, max_aw_cnt;
3946 unsigned int valid_count, valid_node_count, valid_inode_count, discard_blks;
3947 unsigned int bimodal, avg_vblocks;
3948 int util_free, util_valid, util_invalid;
3949 int rsvd_segs, overp_segs;
3950 int dirty_count, node_pages, meta_pages, compress_pages;
3951 int compress_page_hit;
3952 int prefree_count, free_segs, free_secs;
3953 int cp_call_count[MAX_CALL_TYPE], cp_count;
3954 int gc_call_count[MAX_CALL_TYPE];
3957 int tot_blks, data_blks, node_blks;
3958 int bg_data_blks, bg_node_blks;
3959 int curseg[NR_CURSEG_TYPE];
3960 int cursec[NR_CURSEG_TYPE];
3961 int curzone[NR_CURSEG_TYPE];
3962 unsigned int dirty_seg[NR_CURSEG_TYPE];
3963 unsigned int full_seg[NR_CURSEG_TYPE];
3964 unsigned int valid_blks[NR_CURSEG_TYPE];
3966 unsigned int meta_count[META_MAX];
3967 unsigned int segment_count[2];
3968 unsigned int block_count[2];
3969 unsigned int inplace_count;
3970 unsigned long long base_mem, cache_mem, page_mem;
3973 static inline struct f2fs_stat_info *F2FS_STAT(struct f2fs_sb_info *sbi)
3975 return (struct f2fs_stat_info *)sbi->stat_info;
3978 #define stat_inc_cp_call_count(sbi, foreground) \
3979 atomic_inc(&sbi->cp_call_count[(foreground)])
3980 #define stat_inc_cp_count(si) (F2FS_STAT(sbi)->cp_count++)
3981 #define stat_io_skip_bggc_count(sbi) ((sbi)->io_skip_bggc++)
3982 #define stat_other_skip_bggc_count(sbi) ((sbi)->other_skip_bggc++)
3983 #define stat_inc_dirty_inode(sbi, type) ((sbi)->ndirty_inode[type]++)
3984 #define stat_dec_dirty_inode(sbi, type) ((sbi)->ndirty_inode[type]--)
3985 #define stat_inc_total_hit(sbi, type) (atomic64_inc(&(sbi)->total_hit_ext[type]))
3986 #define stat_inc_rbtree_node_hit(sbi, type) (atomic64_inc(&(sbi)->read_hit_rbtree[type]))
3987 #define stat_inc_largest_node_hit(sbi) (atomic64_inc(&(sbi)->read_hit_largest))
3988 #define stat_inc_cached_node_hit(sbi, type) (atomic64_inc(&(sbi)->read_hit_cached[type]))
3989 #define stat_inc_inline_xattr(inode) \
3991 if (f2fs_has_inline_xattr(inode)) \
3992 (atomic_inc(&F2FS_I_SB(inode)->inline_xattr)); \
3994 #define stat_dec_inline_xattr(inode) \
3996 if (f2fs_has_inline_xattr(inode)) \
3997 (atomic_dec(&F2FS_I_SB(inode)->inline_xattr)); \
3999 #define stat_inc_inline_inode(inode) \
4001 if (f2fs_has_inline_data(inode)) \
4002 (atomic_inc(&F2FS_I_SB(inode)->inline_inode)); \
4004 #define stat_dec_inline_inode(inode) \
4006 if (f2fs_has_inline_data(inode)) \
4007 (atomic_dec(&F2FS_I_SB(inode)->inline_inode)); \
4009 #define stat_inc_inline_dir(inode) \
4011 if (f2fs_has_inline_dentry(inode)) \
4012 (atomic_inc(&F2FS_I_SB(inode)->inline_dir)); \
4014 #define stat_dec_inline_dir(inode) \
4016 if (f2fs_has_inline_dentry(inode)) \
4017 (atomic_dec(&F2FS_I_SB(inode)->inline_dir)); \
4019 #define stat_inc_compr_inode(inode) \
4021 if (f2fs_compressed_file(inode)) \
4022 (atomic_inc(&F2FS_I_SB(inode)->compr_inode)); \
4024 #define stat_dec_compr_inode(inode) \
4026 if (f2fs_compressed_file(inode)) \
4027 (atomic_dec(&F2FS_I_SB(inode)->compr_inode)); \
4029 #define stat_add_compr_blocks(inode, blocks) \
4030 (atomic64_add(blocks, &F2FS_I_SB(inode)->compr_blocks))
4031 #define stat_sub_compr_blocks(inode, blocks) \
4032 (atomic64_sub(blocks, &F2FS_I_SB(inode)->compr_blocks))
4033 #define stat_inc_swapfile_inode(inode) \
4034 (atomic_inc(&F2FS_I_SB(inode)->swapfile_inode))
4035 #define stat_dec_swapfile_inode(inode) \
4036 (atomic_dec(&F2FS_I_SB(inode)->swapfile_inode))
4037 #define stat_inc_atomic_inode(inode) \
4038 (atomic_inc(&F2FS_I_SB(inode)->atomic_files))
4039 #define stat_dec_atomic_inode(inode) \
4040 (atomic_dec(&F2FS_I_SB(inode)->atomic_files))
4041 #define stat_inc_meta_count(sbi, blkaddr) \
4043 if (blkaddr < SIT_I(sbi)->sit_base_addr) \
4044 atomic_inc(&(sbi)->meta_count[META_CP]); \
4045 else if (blkaddr < NM_I(sbi)->nat_blkaddr) \
4046 atomic_inc(&(sbi)->meta_count[META_SIT]); \
4047 else if (blkaddr < SM_I(sbi)->ssa_blkaddr) \
4048 atomic_inc(&(sbi)->meta_count[META_NAT]); \
4049 else if (blkaddr < SM_I(sbi)->main_blkaddr) \
4050 atomic_inc(&(sbi)->meta_count[META_SSA]); \
4052 #define stat_inc_seg_type(sbi, curseg) \
4053 ((sbi)->segment_count[(curseg)->alloc_type]++)
4054 #define stat_inc_block_count(sbi, curseg) \
4055 ((sbi)->block_count[(curseg)->alloc_type]++)
4056 #define stat_inc_inplace_blocks(sbi) \
4057 (atomic_inc(&(sbi)->inplace_count))
4058 #define stat_update_max_atomic_write(inode) \
4060 int cur = atomic_read(&F2FS_I_SB(inode)->atomic_files); \
4061 int max = atomic_read(&F2FS_I_SB(inode)->max_aw_cnt); \
4063 atomic_set(&F2FS_I_SB(inode)->max_aw_cnt, cur); \
4065 #define stat_inc_gc_call_count(sbi, foreground) \
4066 (F2FS_STAT(sbi)->gc_call_count[(foreground)]++)
4067 #define stat_inc_gc_sec_count(sbi, type, gc_type) \
4068 (F2FS_STAT(sbi)->gc_secs[(type)][(gc_type)]++)
4069 #define stat_inc_gc_seg_count(sbi, type, gc_type) \
4070 (F2FS_STAT(sbi)->gc_segs[(type)][(gc_type)]++)
4072 #define stat_inc_tot_blk_count(si, blks) \
4073 ((si)->tot_blks += (blks))
4075 #define stat_inc_data_blk_count(sbi, blks, gc_type) \
4077 struct f2fs_stat_info *si = F2FS_STAT(sbi); \
4078 stat_inc_tot_blk_count(si, blks); \
4079 si->data_blks += (blks); \
4080 si->bg_data_blks += ((gc_type) == BG_GC) ? (blks) : 0; \
4083 #define stat_inc_node_blk_count(sbi, blks, gc_type) \
4085 struct f2fs_stat_info *si = F2FS_STAT(sbi); \
4086 stat_inc_tot_blk_count(si, blks); \
4087 si->node_blks += (blks); \
4088 si->bg_node_blks += ((gc_type) == BG_GC) ? (blks) : 0; \
4091 int f2fs_build_stats(struct f2fs_sb_info *sbi);
4092 void f2fs_destroy_stats(struct f2fs_sb_info *sbi);
4093 void __init f2fs_create_root_stats(void);
4094 void f2fs_destroy_root_stats(void);
4095 void f2fs_update_sit_info(struct f2fs_sb_info *sbi);
4097 #define stat_inc_cp_call_count(sbi, foreground) do { } while (0)
4098 #define stat_inc_cp_count(sbi) do { } while (0)
4099 #define stat_io_skip_bggc_count(sbi) do { } while (0)
4100 #define stat_other_skip_bggc_count(sbi) do { } while (0)
4101 #define stat_inc_dirty_inode(sbi, type) do { } while (0)
4102 #define stat_dec_dirty_inode(sbi, type) do { } while (0)
4103 #define stat_inc_total_hit(sbi, type) do { } while (0)
4104 #define stat_inc_rbtree_node_hit(sbi, type) do { } while (0)
4105 #define stat_inc_largest_node_hit(sbi) do { } while (0)
4106 #define stat_inc_cached_node_hit(sbi, type) do { } while (0)
4107 #define stat_inc_inline_xattr(inode) do { } while (0)
4108 #define stat_dec_inline_xattr(inode) do { } while (0)
4109 #define stat_inc_inline_inode(inode) do { } while (0)
4110 #define stat_dec_inline_inode(inode) do { } while (0)
4111 #define stat_inc_inline_dir(inode) do { } while (0)
4112 #define stat_dec_inline_dir(inode) do { } while (0)
4113 #define stat_inc_compr_inode(inode) do { } while (0)
4114 #define stat_dec_compr_inode(inode) do { } while (0)
4115 #define stat_add_compr_blocks(inode, blocks) do { } while (0)
4116 #define stat_sub_compr_blocks(inode, blocks) do { } while (0)
4117 #define stat_inc_swapfile_inode(inode) do { } while (0)
4118 #define stat_dec_swapfile_inode(inode) do { } while (0)
4119 #define stat_inc_atomic_inode(inode) do { } while (0)
4120 #define stat_dec_atomic_inode(inode) do { } while (0)
4121 #define stat_update_max_atomic_write(inode) do { } while (0)
4122 #define stat_inc_meta_count(sbi, blkaddr) do { } while (0)
4123 #define stat_inc_seg_type(sbi, curseg) do { } while (0)
4124 #define stat_inc_block_count(sbi, curseg) do { } while (0)
4125 #define stat_inc_inplace_blocks(sbi) do { } while (0)
4126 #define stat_inc_gc_call_count(sbi, foreground) do { } while (0)
4127 #define stat_inc_gc_sec_count(sbi, type, gc_type) do { } while (0)
4128 #define stat_inc_gc_seg_count(sbi, type, gc_type) do { } while (0)
4129 #define stat_inc_tot_blk_count(si, blks) do { } while (0)
4130 #define stat_inc_data_blk_count(sbi, blks, gc_type) do { } while (0)
4131 #define stat_inc_node_blk_count(sbi, blks, gc_type) do { } while (0)
4133 static inline int f2fs_build_stats(struct f2fs_sb_info *sbi) { return 0; }
4134 static inline void f2fs_destroy_stats(struct f2fs_sb_info *sbi) { }
4135 static inline void __init f2fs_create_root_stats(void) { }
4136 static inline void f2fs_destroy_root_stats(void) { }
4137 static inline void f2fs_update_sit_info(struct f2fs_sb_info *sbi) {}
4140 extern const struct file_operations f2fs_dir_operations;
4141 extern const struct file_operations f2fs_file_operations;
4142 extern const struct inode_operations f2fs_file_inode_operations;
4143 extern const struct address_space_operations f2fs_dblock_aops;
4144 extern const struct address_space_operations f2fs_node_aops;
4145 extern const struct address_space_operations f2fs_meta_aops;
4146 extern const struct inode_operations f2fs_dir_inode_operations;
4147 extern const struct inode_operations f2fs_symlink_inode_operations;
4148 extern const struct inode_operations f2fs_encrypted_symlink_inode_operations;
4149 extern const struct inode_operations f2fs_special_inode_operations;
4150 extern struct kmem_cache *f2fs_inode_entry_slab;
4155 bool f2fs_may_inline_data(struct inode *inode);
4156 bool f2fs_sanity_check_inline_data(struct inode *inode);
4157 bool f2fs_may_inline_dentry(struct inode *inode);
4158 void f2fs_do_read_inline_data(struct page *page, struct page *ipage);
4159 void f2fs_truncate_inline_inode(struct inode *inode,
4160 struct page *ipage, u64 from);
4161 int f2fs_read_inline_data(struct inode *inode, struct page *page);
4162 int f2fs_convert_inline_page(struct dnode_of_data *dn, struct page *page);
4163 int f2fs_convert_inline_inode(struct inode *inode);
4164 int f2fs_try_convert_inline_dir(struct inode *dir, struct dentry *dentry);
4165 int f2fs_write_inline_data(struct inode *inode, struct page *page);
4166 int f2fs_recover_inline_data(struct inode *inode, struct page *npage);
4167 struct f2fs_dir_entry *f2fs_find_in_inline_dir(struct inode *dir,
4168 const struct f2fs_filename *fname,
4169 struct page **res_page);
4170 int f2fs_make_empty_inline_dir(struct inode *inode, struct inode *parent,
4171 struct page *ipage);
4172 int f2fs_add_inline_entry(struct inode *dir, const struct f2fs_filename *fname,
4173 struct inode *inode, nid_t ino, umode_t mode);
4174 void f2fs_delete_inline_entry(struct f2fs_dir_entry *dentry,
4175 struct page *page, struct inode *dir,
4176 struct inode *inode);
4177 bool f2fs_empty_inline_dir(struct inode *dir);
4178 int f2fs_read_inline_dir(struct file *file, struct dir_context *ctx,
4179 struct fscrypt_str *fstr);
4180 int f2fs_inline_data_fiemap(struct inode *inode,
4181 struct fiemap_extent_info *fieinfo,
4182 __u64 start, __u64 len);
4187 unsigned long f2fs_shrink_count(struct shrinker *shrink,
4188 struct shrink_control *sc);
4189 unsigned long f2fs_shrink_scan(struct shrinker *shrink,
4190 struct shrink_control *sc);
4191 void f2fs_join_shrinker(struct f2fs_sb_info *sbi);
4192 void f2fs_leave_shrinker(struct f2fs_sb_info *sbi);
4197 bool sanity_check_extent_cache(struct inode *inode);
4198 void f2fs_init_extent_tree(struct inode *inode);
4199 void f2fs_drop_extent_tree(struct inode *inode);
4200 void f2fs_destroy_extent_node(struct inode *inode);
4201 void f2fs_destroy_extent_tree(struct inode *inode);
4202 void f2fs_init_extent_cache_info(struct f2fs_sb_info *sbi);
4203 int __init f2fs_create_extent_cache(void);
4204 void f2fs_destroy_extent_cache(void);
4206 /* read extent cache ops */
4207 void f2fs_init_read_extent_tree(struct inode *inode, struct page *ipage);
4208 bool f2fs_lookup_read_extent_cache(struct inode *inode, pgoff_t pgofs,
4209 struct extent_info *ei);
4210 bool f2fs_lookup_read_extent_cache_block(struct inode *inode, pgoff_t index,
4212 void f2fs_update_read_extent_cache(struct dnode_of_data *dn);
4213 void f2fs_update_read_extent_cache_range(struct dnode_of_data *dn,
4214 pgoff_t fofs, block_t blkaddr, unsigned int len);
4215 unsigned int f2fs_shrink_read_extent_tree(struct f2fs_sb_info *sbi,
4218 /* block age extent cache ops */
4219 void f2fs_init_age_extent_tree(struct inode *inode);
4220 bool f2fs_lookup_age_extent_cache(struct inode *inode, pgoff_t pgofs,
4221 struct extent_info *ei);
4222 void f2fs_update_age_extent_cache(struct dnode_of_data *dn);
4223 void f2fs_update_age_extent_cache_range(struct dnode_of_data *dn,
4224 pgoff_t fofs, unsigned int len);
4225 unsigned int f2fs_shrink_age_extent_tree(struct f2fs_sb_info *sbi,
4231 #define MIN_RA_MUL 2
4232 #define MAX_RA_MUL 256
4234 int __init f2fs_init_sysfs(void);
4235 void f2fs_exit_sysfs(void);
4236 int f2fs_register_sysfs(struct f2fs_sb_info *sbi);
4237 void f2fs_unregister_sysfs(struct f2fs_sb_info *sbi);
4240 extern const struct fsverity_operations f2fs_verityops;
4245 static inline bool f2fs_encrypted_file(struct inode *inode)
4247 return IS_ENCRYPTED(inode) && S_ISREG(inode->i_mode);
4250 static inline void f2fs_set_encrypted_inode(struct inode *inode)
4252 #ifdef CONFIG_FS_ENCRYPTION
4253 file_set_encrypt(inode);
4254 f2fs_set_inode_flags(inode);
4259 * Returns true if the reads of the inode's data need to undergo some
4260 * postprocessing step, like decryption or authenticity verification.
4262 static inline bool f2fs_post_read_required(struct inode *inode)
4264 return f2fs_encrypted_file(inode) || fsverity_active(inode) ||
4265 f2fs_compressed_file(inode);
4271 #ifdef CONFIG_F2FS_FS_COMPRESSION
4272 bool f2fs_is_compressed_page(struct page *page);
4273 struct page *f2fs_compress_control_page(struct page *page);
4274 int f2fs_prepare_compress_overwrite(struct inode *inode,
4275 struct page **pagep, pgoff_t index, void **fsdata);
4276 bool f2fs_compress_write_end(struct inode *inode, void *fsdata,
4277 pgoff_t index, unsigned copied);
4278 int f2fs_truncate_partial_cluster(struct inode *inode, u64 from, bool lock);
4279 void f2fs_compress_write_end_io(struct bio *bio, struct page *page);
4280 bool f2fs_is_compress_backend_ready(struct inode *inode);
4281 bool f2fs_is_compress_level_valid(int alg, int lvl);
4282 int __init f2fs_init_compress_mempool(void);
4283 void f2fs_destroy_compress_mempool(void);
4284 void f2fs_decompress_cluster(struct decompress_io_ctx *dic, bool in_task);
4285 void f2fs_end_read_compressed_page(struct page *page, bool failed,
4286 block_t blkaddr, bool in_task);
4287 bool f2fs_cluster_is_empty(struct compress_ctx *cc);
4288 bool f2fs_cluster_can_merge_page(struct compress_ctx *cc, pgoff_t index);
4289 bool f2fs_all_cluster_page_ready(struct compress_ctx *cc, struct page **pages,
4290 int index, int nr_pages, bool uptodate);
4291 bool f2fs_sanity_check_cluster(struct dnode_of_data *dn);
4292 void f2fs_compress_ctx_add_page(struct compress_ctx *cc, struct page *page);
4293 int f2fs_write_multi_pages(struct compress_ctx *cc,
4295 struct writeback_control *wbc,
4296 enum iostat_type io_type);
4297 int f2fs_is_compressed_cluster(struct inode *inode, pgoff_t index);
4298 void f2fs_update_read_extent_tree_range_compressed(struct inode *inode,
4299 pgoff_t fofs, block_t blkaddr,
4300 unsigned int llen, unsigned int c_len);
4301 int f2fs_read_multi_pages(struct compress_ctx *cc, struct bio **bio_ret,
4302 unsigned nr_pages, sector_t *last_block_in_bio,
4303 bool is_readahead, bool for_write);
4304 struct decompress_io_ctx *f2fs_alloc_dic(struct compress_ctx *cc);
4305 void f2fs_decompress_end_io(struct decompress_io_ctx *dic, bool failed,
4307 void f2fs_put_page_dic(struct page *page, bool in_task);
4308 unsigned int f2fs_cluster_blocks_are_contiguous(struct dnode_of_data *dn);
4309 int f2fs_init_compress_ctx(struct compress_ctx *cc);
4310 void f2fs_destroy_compress_ctx(struct compress_ctx *cc, bool reuse);
4311 void f2fs_init_compress_info(struct f2fs_sb_info *sbi);
4312 int f2fs_init_compress_inode(struct f2fs_sb_info *sbi);
4313 void f2fs_destroy_compress_inode(struct f2fs_sb_info *sbi);
4314 int f2fs_init_page_array_cache(struct f2fs_sb_info *sbi);
4315 void f2fs_destroy_page_array_cache(struct f2fs_sb_info *sbi);
4316 int __init f2fs_init_compress_cache(void);
4317 void f2fs_destroy_compress_cache(void);
4318 struct address_space *COMPRESS_MAPPING(struct f2fs_sb_info *sbi);
4319 void f2fs_invalidate_compress_page(struct f2fs_sb_info *sbi, block_t blkaddr);
4320 void f2fs_cache_compressed_page(struct f2fs_sb_info *sbi, struct page *page,
4321 nid_t ino, block_t blkaddr);
4322 bool f2fs_load_compressed_page(struct f2fs_sb_info *sbi, struct page *page,
4324 void f2fs_invalidate_compress_pages(struct f2fs_sb_info *sbi, nid_t ino);
4325 #define inc_compr_inode_stat(inode) \
4327 struct f2fs_sb_info *sbi = F2FS_I_SB(inode); \
4328 sbi->compr_new_inode++; \
4330 #define add_compr_block_stat(inode, blocks) \
4332 struct f2fs_sb_info *sbi = F2FS_I_SB(inode); \
4333 int diff = F2FS_I(inode)->i_cluster_size - blocks; \
4334 sbi->compr_written_block += blocks; \
4335 sbi->compr_saved_block += diff; \
4338 static inline bool f2fs_is_compressed_page(struct page *page) { return false; }
4339 static inline bool f2fs_is_compress_backend_ready(struct inode *inode)
4341 if (!f2fs_compressed_file(inode))
4343 /* not support compression */
4346 static inline bool f2fs_is_compress_level_valid(int alg, int lvl) { return false; }
4347 static inline struct page *f2fs_compress_control_page(struct page *page)
4350 return ERR_PTR(-EINVAL);
4352 static inline int __init f2fs_init_compress_mempool(void) { return 0; }
4353 static inline void f2fs_destroy_compress_mempool(void) { }
4354 static inline void f2fs_decompress_cluster(struct decompress_io_ctx *dic,
4356 static inline void f2fs_end_read_compressed_page(struct page *page,
4357 bool failed, block_t blkaddr, bool in_task)
4361 static inline void f2fs_put_page_dic(struct page *page, bool in_task)
4365 static inline unsigned int f2fs_cluster_blocks_are_contiguous(struct dnode_of_data *dn) { return 0; }
4366 static inline bool f2fs_sanity_check_cluster(struct dnode_of_data *dn) { return false; }
4367 static inline int f2fs_init_compress_inode(struct f2fs_sb_info *sbi) { return 0; }
4368 static inline void f2fs_destroy_compress_inode(struct f2fs_sb_info *sbi) { }
4369 static inline int f2fs_init_page_array_cache(struct f2fs_sb_info *sbi) { return 0; }
4370 static inline void f2fs_destroy_page_array_cache(struct f2fs_sb_info *sbi) { }
4371 static inline int __init f2fs_init_compress_cache(void) { return 0; }
4372 static inline void f2fs_destroy_compress_cache(void) { }
4373 static inline void f2fs_invalidate_compress_page(struct f2fs_sb_info *sbi,
4374 block_t blkaddr) { }
4375 static inline void f2fs_cache_compressed_page(struct f2fs_sb_info *sbi,
4376 struct page *page, nid_t ino, block_t blkaddr) { }
4377 static inline bool f2fs_load_compressed_page(struct f2fs_sb_info *sbi,
4378 struct page *page, block_t blkaddr) { return false; }
4379 static inline void f2fs_invalidate_compress_pages(struct f2fs_sb_info *sbi,
4381 #define inc_compr_inode_stat(inode) do { } while (0)
4382 static inline void f2fs_update_read_extent_tree_range_compressed(
4383 struct inode *inode,
4384 pgoff_t fofs, block_t blkaddr,
4385 unsigned int llen, unsigned int c_len) { }
4388 static inline int set_compress_context(struct inode *inode)
4390 #ifdef CONFIG_F2FS_FS_COMPRESSION
4391 struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
4393 F2FS_I(inode)->i_compress_algorithm =
4394 F2FS_OPTION(sbi).compress_algorithm;
4395 F2FS_I(inode)->i_log_cluster_size =
4396 F2FS_OPTION(sbi).compress_log_size;
4397 F2FS_I(inode)->i_compress_flag =
4398 F2FS_OPTION(sbi).compress_chksum ?
4399 BIT(COMPRESS_CHKSUM) : 0;
4400 F2FS_I(inode)->i_cluster_size =
4401 BIT(F2FS_I(inode)->i_log_cluster_size);
4402 if ((F2FS_I(inode)->i_compress_algorithm == COMPRESS_LZ4 ||
4403 F2FS_I(inode)->i_compress_algorithm == COMPRESS_ZSTD) &&
4404 F2FS_OPTION(sbi).compress_level)
4405 F2FS_I(inode)->i_compress_level =
4406 F2FS_OPTION(sbi).compress_level;
4407 F2FS_I(inode)->i_flags |= F2FS_COMPR_FL;
4408 set_inode_flag(inode, FI_COMPRESSED_FILE);
4409 stat_inc_compr_inode(inode);
4410 inc_compr_inode_stat(inode);
4411 f2fs_mark_inode_dirty_sync(inode, true);
4418 static inline bool f2fs_disable_compressed_file(struct inode *inode)
4420 struct f2fs_inode_info *fi = F2FS_I(inode);
4422 f2fs_down_write(&F2FS_I(inode)->i_sem);
4424 if (!f2fs_compressed_file(inode)) {
4425 f2fs_up_write(&F2FS_I(inode)->i_sem);
4428 if (f2fs_is_mmap_file(inode) ||
4429 (S_ISREG(inode->i_mode) && F2FS_HAS_BLOCKS(inode))) {
4430 f2fs_up_write(&F2FS_I(inode)->i_sem);
4434 fi->i_flags &= ~F2FS_COMPR_FL;
4435 stat_dec_compr_inode(inode);
4436 clear_inode_flag(inode, FI_COMPRESSED_FILE);
4437 f2fs_mark_inode_dirty_sync(inode, true);
4439 f2fs_up_write(&F2FS_I(inode)->i_sem);
4443 #define F2FS_FEATURE_FUNCS(name, flagname) \
4444 static inline bool f2fs_sb_has_##name(struct f2fs_sb_info *sbi) \
4446 return F2FS_HAS_FEATURE(sbi, F2FS_FEATURE_##flagname); \
4449 F2FS_FEATURE_FUNCS(encrypt, ENCRYPT);
4450 F2FS_FEATURE_FUNCS(blkzoned, BLKZONED);
4451 F2FS_FEATURE_FUNCS(extra_attr, EXTRA_ATTR);
4452 F2FS_FEATURE_FUNCS(project_quota, PRJQUOTA);
4453 F2FS_FEATURE_FUNCS(inode_chksum, INODE_CHKSUM);
4454 F2FS_FEATURE_FUNCS(flexible_inline_xattr, FLEXIBLE_INLINE_XATTR);
4455 F2FS_FEATURE_FUNCS(quota_ino, QUOTA_INO);
4456 F2FS_FEATURE_FUNCS(inode_crtime, INODE_CRTIME);
4457 F2FS_FEATURE_FUNCS(lost_found, LOST_FOUND);
4458 F2FS_FEATURE_FUNCS(verity, VERITY);
4459 F2FS_FEATURE_FUNCS(sb_chksum, SB_CHKSUM);
4460 F2FS_FEATURE_FUNCS(casefold, CASEFOLD);
4461 F2FS_FEATURE_FUNCS(compression, COMPRESSION);
4462 F2FS_FEATURE_FUNCS(readonly, RO);
4464 #ifdef CONFIG_BLK_DEV_ZONED
4465 static inline bool f2fs_blkz_is_seq(struct f2fs_sb_info *sbi, int devi,
4468 unsigned int zno = blkaddr / sbi->blocks_per_blkz;
4470 return test_bit(zno, FDEV(devi).blkz_seq);
4474 static inline int f2fs_bdev_index(struct f2fs_sb_info *sbi,
4475 struct block_device *bdev)
4479 if (!f2fs_is_multi_device(sbi))
4482 for (i = 0; i < sbi->s_ndevs; i++)
4483 if (FDEV(i).bdev == bdev)
4490 static inline bool f2fs_hw_should_discard(struct f2fs_sb_info *sbi)
4492 return f2fs_sb_has_blkzoned(sbi);
4495 static inline bool f2fs_bdev_support_discard(struct block_device *bdev)
4497 return bdev_max_discard_sectors(bdev) || bdev_is_zoned(bdev);
4500 static inline bool f2fs_hw_support_discard(struct f2fs_sb_info *sbi)
4504 if (!f2fs_is_multi_device(sbi))
4505 return f2fs_bdev_support_discard(sbi->sb->s_bdev);
4507 for (i = 0; i < sbi->s_ndevs; i++)
4508 if (f2fs_bdev_support_discard(FDEV(i).bdev))
4513 static inline bool f2fs_realtime_discard_enable(struct f2fs_sb_info *sbi)
4515 return (test_opt(sbi, DISCARD) && f2fs_hw_support_discard(sbi)) ||
4516 f2fs_hw_should_discard(sbi);
4519 static inline bool f2fs_hw_is_readonly(struct f2fs_sb_info *sbi)
4523 if (!f2fs_is_multi_device(sbi))
4524 return bdev_read_only(sbi->sb->s_bdev);
4526 for (i = 0; i < sbi->s_ndevs; i++)
4527 if (bdev_read_only(FDEV(i).bdev))
4532 static inline bool f2fs_dev_is_readonly(struct f2fs_sb_info *sbi)
4534 return f2fs_sb_has_readonly(sbi) || f2fs_hw_is_readonly(sbi);
4537 static inline bool f2fs_lfs_mode(struct f2fs_sb_info *sbi)
4539 return F2FS_OPTION(sbi).fs_mode == FS_MODE_LFS;
4542 static inline bool f2fs_valid_pinned_area(struct f2fs_sb_info *sbi,
4545 if (f2fs_sb_has_blkzoned(sbi)) {
4546 int devi = f2fs_target_device_index(sbi, blkaddr);
4548 return !bdev_is_zoned(FDEV(devi).bdev);
4553 static inline bool f2fs_low_mem_mode(struct f2fs_sb_info *sbi)
4555 return F2FS_OPTION(sbi).memory_mode == MEMORY_MODE_LOW;
4558 static inline bool f2fs_may_compress(struct inode *inode)
4560 if (IS_SWAPFILE(inode) || f2fs_is_pinned_file(inode) ||
4561 f2fs_is_atomic_file(inode) || f2fs_has_inline_data(inode) ||
4562 f2fs_is_mmap_file(inode))
4564 return S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode);
4567 static inline void f2fs_i_compr_blocks_update(struct inode *inode,
4568 u64 blocks, bool add)
4570 struct f2fs_inode_info *fi = F2FS_I(inode);
4571 int diff = fi->i_cluster_size - blocks;
4573 /* don't update i_compr_blocks if saved blocks were released */
4574 if (!add && !atomic_read(&fi->i_compr_blocks))
4578 atomic_add(diff, &fi->i_compr_blocks);
4579 stat_add_compr_blocks(inode, diff);
4581 atomic_sub(diff, &fi->i_compr_blocks);
4582 stat_sub_compr_blocks(inode, diff);
4584 f2fs_mark_inode_dirty_sync(inode, true);
4587 static inline bool f2fs_allow_multi_device_dio(struct f2fs_sb_info *sbi,
4590 if (!f2fs_is_multi_device(sbi))
4592 if (flag != F2FS_GET_BLOCK_DIO)
4594 return sbi->aligned_blksize;
4597 static inline bool f2fs_need_verity(const struct inode *inode, pgoff_t idx)
4599 return fsverity_active(inode) &&
4600 idx < DIV_ROUND_UP(inode->i_size, PAGE_SIZE);
4603 #ifdef CONFIG_F2FS_FAULT_INJECTION
4604 extern void f2fs_build_fault_attr(struct f2fs_sb_info *sbi, unsigned int rate,
4607 #define f2fs_build_fault_attr(sbi, rate, type) do { } while (0)
4610 static inline bool is_journalled_quota(struct f2fs_sb_info *sbi)
4613 if (f2fs_sb_has_quota_ino(sbi))
4615 if (F2FS_OPTION(sbi).s_qf_names[USRQUOTA] ||
4616 F2FS_OPTION(sbi).s_qf_names[GRPQUOTA] ||
4617 F2FS_OPTION(sbi).s_qf_names[PRJQUOTA])
4623 static inline bool f2fs_block_unit_discard(struct f2fs_sb_info *sbi)
4625 return F2FS_OPTION(sbi).discard_unit == DISCARD_UNIT_BLOCK;
4628 static inline void f2fs_io_schedule_timeout(long timeout)
4630 set_current_state(TASK_UNINTERRUPTIBLE);
4631 io_schedule_timeout(timeout);
4634 static inline void f2fs_handle_page_eio(struct f2fs_sb_info *sbi, pgoff_t ofs,
4635 enum page_type type)
4637 if (unlikely(f2fs_cp_error(sbi)))
4640 if (ofs == sbi->page_eio_ofs[type]) {
4641 if (sbi->page_eio_cnt[type]++ == MAX_RETRY_PAGE_EIO)
4642 set_ckpt_flags(sbi, CP_ERROR_FLAG);
4644 sbi->page_eio_ofs[type] = ofs;
4645 sbi->page_eio_cnt[type] = 0;
4649 static inline bool f2fs_is_readonly(struct f2fs_sb_info *sbi)
4651 return f2fs_sb_has_readonly(sbi) || f2fs_readonly(sbi->sb);
4654 static inline void f2fs_invalidate_internal_cache(struct f2fs_sb_info *sbi,
4657 invalidate_mapping_pages(META_MAPPING(sbi), blkaddr, blkaddr);
4658 f2fs_invalidate_compress_page(sbi, blkaddr);
4661 #define EFSBADCRC EBADMSG /* Bad CRC detected */
4662 #define EFSCORRUPTED EUCLEAN /* Filesystem is corrupted */
4664 #endif /* _LINUX_F2FS_H */