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 */
66 #ifdef CONFIG_F2FS_FAULT_INJECTION
67 #define F2FS_ALL_FAULT_TYPE (GENMASK(FAULT_MAX - 1, 0))
69 struct f2fs_fault_info {
71 unsigned int inject_rate;
72 unsigned int inject_type;
75 extern const char *f2fs_fault_name[FAULT_MAX];
76 #define IS_FAULT_SET(fi, type) ((fi)->inject_type & BIT(type))
78 /* maximum retry count for injected failure */
79 #define DEFAULT_FAILURE_RETRY_COUNT 8
81 #define DEFAULT_FAILURE_RETRY_COUNT 1
87 #define F2FS_MOUNT_DISABLE_ROLL_FORWARD 0x00000002
88 #define F2FS_MOUNT_DISCARD 0x00000004
89 #define F2FS_MOUNT_NOHEAP 0x00000008
90 #define F2FS_MOUNT_XATTR_USER 0x00000010
91 #define F2FS_MOUNT_POSIX_ACL 0x00000020
92 #define F2FS_MOUNT_DISABLE_EXT_IDENTIFY 0x00000040
93 #define F2FS_MOUNT_INLINE_XATTR 0x00000080
94 #define F2FS_MOUNT_INLINE_DATA 0x00000100
95 #define F2FS_MOUNT_INLINE_DENTRY 0x00000200
96 #define F2FS_MOUNT_FLUSH_MERGE 0x00000400
97 #define F2FS_MOUNT_NOBARRIER 0x00000800
98 #define F2FS_MOUNT_FASTBOOT 0x00001000
99 #define F2FS_MOUNT_READ_EXTENT_CACHE 0x00002000
100 #define F2FS_MOUNT_DATA_FLUSH 0x00008000
101 #define F2FS_MOUNT_FAULT_INJECTION 0x00010000
102 #define F2FS_MOUNT_USRQUOTA 0x00080000
103 #define F2FS_MOUNT_GRPQUOTA 0x00100000
104 #define F2FS_MOUNT_PRJQUOTA 0x00200000
105 #define F2FS_MOUNT_QUOTA 0x00400000
106 #define F2FS_MOUNT_INLINE_XATTR_SIZE 0x00800000
107 #define F2FS_MOUNT_RESERVE_ROOT 0x01000000
108 #define F2FS_MOUNT_DISABLE_CHECKPOINT 0x02000000
109 #define F2FS_MOUNT_NORECOVERY 0x04000000
110 #define F2FS_MOUNT_ATGC 0x08000000
111 #define F2FS_MOUNT_MERGE_CHECKPOINT 0x10000000
112 #define F2FS_MOUNT_GC_MERGE 0x20000000
113 #define F2FS_MOUNT_COMPRESS_CACHE 0x40000000
115 #define F2FS_OPTION(sbi) ((sbi)->mount_opt)
116 #define clear_opt(sbi, option) (F2FS_OPTION(sbi).opt &= ~F2FS_MOUNT_##option)
117 #define set_opt(sbi, option) (F2FS_OPTION(sbi).opt |= F2FS_MOUNT_##option)
118 #define test_opt(sbi, option) (F2FS_OPTION(sbi).opt & F2FS_MOUNT_##option)
120 #define ver_after(a, b) (typecheck(unsigned long long, a) && \
121 typecheck(unsigned long long, b) && \
122 ((long long)((a) - (b)) > 0))
124 typedef u32 block_t; /*
125 * should not change u32, since it is the on-disk block
126 * address format, __le32.
130 #define COMPRESS_EXT_NUM 16
133 * An implementation of an rwsem that is explicitly unfair to readers. This
134 * prevents priority inversion when a low-priority reader acquires the read lock
135 * while sleeping on the write lock but the write lock is needed by
136 * higher-priority clients.
140 struct rw_semaphore internal_rwsem;
141 #ifdef CONFIG_F2FS_UNFAIR_RWSEM
142 wait_queue_head_t read_waiters;
146 struct f2fs_mount_info {
148 int write_io_size_bits; /* Write IO size bits */
149 block_t root_reserved_blocks; /* root reserved blocks */
150 kuid_t s_resuid; /* reserved blocks for uid */
151 kgid_t s_resgid; /* reserved blocks for gid */
152 int active_logs; /* # of active logs */
153 int inline_xattr_size; /* inline xattr size */
154 #ifdef CONFIG_F2FS_FAULT_INJECTION
155 struct f2fs_fault_info fault_info; /* For fault injection */
158 /* Names of quota files with journalled quota */
159 char *s_qf_names[MAXQUOTAS];
160 int s_jquota_fmt; /* Format of quota to use */
162 /* For which write hints are passed down to block layer */
163 int alloc_mode; /* segment allocation policy */
164 int fsync_mode; /* fsync policy */
165 int fs_mode; /* fs mode: LFS or ADAPTIVE */
166 int bggc_mode; /* bggc mode: off, on or sync */
167 int memory_mode; /* memory mode */
169 * discard command's offset/size should
170 * be aligned to this unit: block,
173 struct fscrypt_dummy_policy dummy_enc_policy; /* test dummy encryption */
174 block_t unusable_cap_perc; /* percentage for cap */
175 block_t unusable_cap; /* Amount of space allowed to be
176 * unusable when disabling checkpoint
179 /* For compression */
180 unsigned char compress_algorithm; /* algorithm type */
181 unsigned char compress_log_size; /* cluster log size */
182 unsigned char compress_level; /* compress level */
183 bool compress_chksum; /* compressed data chksum */
184 unsigned char compress_ext_cnt; /* extension count */
185 unsigned char nocompress_ext_cnt; /* nocompress extension count */
186 int compress_mode; /* compression mode */
187 unsigned char extensions[COMPRESS_EXT_NUM][F2FS_EXTENSION_LEN]; /* extensions */
188 unsigned char noextensions[COMPRESS_EXT_NUM][F2FS_EXTENSION_LEN]; /* extensions */
191 #define F2FS_FEATURE_ENCRYPT 0x0001
192 #define F2FS_FEATURE_BLKZONED 0x0002
193 #define F2FS_FEATURE_ATOMIC_WRITE 0x0004
194 #define F2FS_FEATURE_EXTRA_ATTR 0x0008
195 #define F2FS_FEATURE_PRJQUOTA 0x0010
196 #define F2FS_FEATURE_INODE_CHKSUM 0x0020
197 #define F2FS_FEATURE_FLEXIBLE_INLINE_XATTR 0x0040
198 #define F2FS_FEATURE_QUOTA_INO 0x0080
199 #define F2FS_FEATURE_INODE_CRTIME 0x0100
200 #define F2FS_FEATURE_LOST_FOUND 0x0200
201 #define F2FS_FEATURE_VERITY 0x0400
202 #define F2FS_FEATURE_SB_CHKSUM 0x0800
203 #define F2FS_FEATURE_CASEFOLD 0x1000
204 #define F2FS_FEATURE_COMPRESSION 0x2000
205 #define F2FS_FEATURE_RO 0x4000
207 #define __F2FS_HAS_FEATURE(raw_super, mask) \
208 ((raw_super->feature & cpu_to_le32(mask)) != 0)
209 #define F2FS_HAS_FEATURE(sbi, mask) __F2FS_HAS_FEATURE(sbi->raw_super, mask)
210 #define F2FS_SET_FEATURE(sbi, mask) \
211 (sbi->raw_super->feature |= cpu_to_le32(mask))
212 #define F2FS_CLEAR_FEATURE(sbi, mask) \
213 (sbi->raw_super->feature &= ~cpu_to_le32(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 /* default discard granularity of inner discard thread, unit: block count */
337 #define DEFAULT_DISCARD_GRANULARITY 16
339 /* max discard pend list number */
340 #define MAX_PLIST_NUM 512
341 #define plist_idx(blk_num) ((blk_num) >= MAX_PLIST_NUM ? \
342 (MAX_PLIST_NUM - 1) : ((blk_num) - 1))
345 D_PREP, /* initial */
346 D_PARTIAL, /* partially submitted */
347 D_SUBMIT, /* all submitted */
348 D_DONE, /* finished */
351 struct discard_info {
352 block_t lstart; /* logical start address */
353 block_t len; /* length */
354 block_t start; /* actual start address in dev */
358 struct rb_node rb_node; /* rb node located in rb-tree */
361 block_t lstart; /* logical start address */
362 block_t len; /* length */
363 block_t start; /* actual start address in dev */
365 struct discard_info di; /* discard info */
368 struct list_head list; /* command list */
369 struct completion wait; /* compleation */
370 struct block_device *bdev; /* bdev */
371 unsigned short ref; /* reference count */
372 unsigned char state; /* state */
373 unsigned char queued; /* queued discard */
374 int error; /* bio error */
375 spinlock_t lock; /* for state/bio_ref updating */
376 unsigned short bio_ref; /* bio reference count */
387 struct discard_policy {
388 int type; /* type of discard */
389 unsigned int min_interval; /* used for candidates exist */
390 unsigned int mid_interval; /* used for device busy */
391 unsigned int max_interval; /* used for candidates not exist */
392 unsigned int max_requests; /* # of discards issued per round */
393 unsigned int io_aware_gran; /* minimum granularity discard not be aware of I/O */
394 bool io_aware; /* issue discard in idle time */
395 bool sync; /* submit discard with REQ_SYNC flag */
396 bool ordered; /* issue discard by lba order */
397 bool timeout; /* discard timeout for put_super */
398 unsigned int granularity; /* discard granularity */
401 struct discard_cmd_control {
402 struct task_struct *f2fs_issue_discard; /* discard thread */
403 struct list_head entry_list; /* 4KB discard entry list */
404 struct list_head pend_list[MAX_PLIST_NUM];/* store pending entries */
405 struct list_head wait_list; /* store on-flushing entries */
406 struct list_head fstrim_list; /* in-flight discard from fstrim */
407 wait_queue_head_t discard_wait_queue; /* waiting queue for wake-up */
408 unsigned int discard_wake; /* to wake up discard thread */
409 struct mutex cmd_lock;
410 unsigned int nr_discards; /* # of discards in the list */
411 unsigned int max_discards; /* max. discards to be issued */
412 unsigned int max_discard_request; /* max. discard request per round */
413 unsigned int min_discard_issue_time; /* min. interval between discard issue */
414 unsigned int mid_discard_issue_time; /* mid. interval between discard issue */
415 unsigned int max_discard_issue_time; /* max. interval between discard issue */
416 unsigned int discard_granularity; /* discard granularity */
417 unsigned int undiscard_blks; /* # of undiscard blocks */
418 unsigned int next_pos; /* next discard position */
419 atomic_t issued_discard; /* # of issued discard */
420 atomic_t queued_discard; /* # of queued discard */
421 atomic_t discard_cmd_cnt; /* # of cached cmd count */
422 struct rb_root_cached root; /* root of discard rb-tree */
423 bool rbtree_check; /* config for consistence check */
426 /* for the list of fsync inodes, used only during recovery */
427 struct fsync_inode_entry {
428 struct list_head list; /* list head */
429 struct inode *inode; /* vfs inode pointer */
430 block_t blkaddr; /* block address locating the last fsync */
431 block_t last_dentry; /* block address locating the last dentry */
434 #define nats_in_cursum(jnl) (le16_to_cpu((jnl)->n_nats))
435 #define sits_in_cursum(jnl) (le16_to_cpu((jnl)->n_sits))
437 #define nat_in_journal(jnl, i) ((jnl)->nat_j.entries[i].ne)
438 #define nid_in_journal(jnl, i) ((jnl)->nat_j.entries[i].nid)
439 #define sit_in_journal(jnl, i) ((jnl)->sit_j.entries[i].se)
440 #define segno_in_journal(jnl, i) ((jnl)->sit_j.entries[i].segno)
442 #define MAX_NAT_JENTRIES(jnl) (NAT_JOURNAL_ENTRIES - nats_in_cursum(jnl))
443 #define MAX_SIT_JENTRIES(jnl) (SIT_JOURNAL_ENTRIES - sits_in_cursum(jnl))
445 static inline int update_nats_in_cursum(struct f2fs_journal *journal, int i)
447 int before = nats_in_cursum(journal);
449 journal->n_nats = cpu_to_le16(before + i);
453 static inline int update_sits_in_cursum(struct f2fs_journal *journal, int i)
455 int before = sits_in_cursum(journal);
457 journal->n_sits = cpu_to_le16(before + i);
461 static inline bool __has_cursum_space(struct f2fs_journal *journal,
464 if (type == NAT_JOURNAL)
465 return size <= MAX_NAT_JENTRIES(journal);
466 return size <= MAX_SIT_JENTRIES(journal);
469 /* for inline stuff */
470 #define DEF_INLINE_RESERVED_SIZE 1
471 static inline int get_extra_isize(struct inode *inode);
472 static inline int get_inline_xattr_addrs(struct inode *inode);
473 #define MAX_INLINE_DATA(inode) (sizeof(__le32) * \
474 (CUR_ADDRS_PER_INODE(inode) - \
475 get_inline_xattr_addrs(inode) - \
476 DEF_INLINE_RESERVED_SIZE))
479 #define NR_INLINE_DENTRY(inode) (MAX_INLINE_DATA(inode) * BITS_PER_BYTE / \
480 ((SIZE_OF_DIR_ENTRY + F2FS_SLOT_LEN) * \
482 #define INLINE_DENTRY_BITMAP_SIZE(inode) \
483 DIV_ROUND_UP(NR_INLINE_DENTRY(inode), BITS_PER_BYTE)
484 #define INLINE_RESERVED_SIZE(inode) (MAX_INLINE_DATA(inode) - \
485 ((SIZE_OF_DIR_ENTRY + F2FS_SLOT_LEN) * \
486 NR_INLINE_DENTRY(inode) + \
487 INLINE_DENTRY_BITMAP_SIZE(inode)))
490 * For INODE and NODE manager
492 /* for directory operations */
494 struct f2fs_filename {
496 * The filename the user specified. This is NULL for some
497 * filesystem-internal operations, e.g. converting an inline directory
498 * to a non-inline one, or roll-forward recovering an encrypted dentry.
500 const struct qstr *usr_fname;
503 * The on-disk filename. For encrypted directories, this is encrypted.
504 * This may be NULL for lookups in an encrypted dir without the key.
506 struct fscrypt_str disk_name;
508 /* The dirhash of this filename */
511 #ifdef CONFIG_FS_ENCRYPTION
513 * For lookups in encrypted directories: either the buffer backing
514 * disk_name, or a buffer that holds the decoded no-key name.
516 struct fscrypt_str crypto_buf;
518 #if IS_ENABLED(CONFIG_UNICODE)
520 * For casefolded directories: the casefolded name, but it's left NULL
521 * if the original name is not valid Unicode, if the original name is
522 * "." or "..", if the directory is both casefolded and encrypted and
523 * its encryption key is unavailable, or if the filesystem is doing an
524 * internal operation where usr_fname is also NULL. In all these cases
525 * we fall back to treating the name as an opaque byte sequence.
527 struct fscrypt_str cf_name;
531 struct f2fs_dentry_ptr {
534 struct f2fs_dir_entry *dentry;
535 __u8 (*filename)[F2FS_SLOT_LEN];
540 static inline void make_dentry_ptr_block(struct inode *inode,
541 struct f2fs_dentry_ptr *d, struct f2fs_dentry_block *t)
544 d->max = NR_DENTRY_IN_BLOCK;
545 d->nr_bitmap = SIZE_OF_DENTRY_BITMAP;
546 d->bitmap = t->dentry_bitmap;
547 d->dentry = t->dentry;
548 d->filename = t->filename;
551 static inline void make_dentry_ptr_inline(struct inode *inode,
552 struct f2fs_dentry_ptr *d, void *t)
554 int entry_cnt = NR_INLINE_DENTRY(inode);
555 int bitmap_size = INLINE_DENTRY_BITMAP_SIZE(inode);
556 int reserved_size = INLINE_RESERVED_SIZE(inode);
560 d->nr_bitmap = bitmap_size;
562 d->dentry = t + bitmap_size + reserved_size;
563 d->filename = t + bitmap_size + reserved_size +
564 SIZE_OF_DIR_ENTRY * entry_cnt;
568 * XATTR_NODE_OFFSET stores xattrs to one node block per file keeping -1
569 * as its node offset to distinguish from index node blocks.
570 * But some bits are used to mark the node block.
572 #define XATTR_NODE_OFFSET ((((unsigned int)-1) << OFFSET_BIT_SHIFT) \
575 ALLOC_NODE, /* allocate a new node page if needed */
576 LOOKUP_NODE, /* look up a node without readahead */
578 * look up a node with readahead called
583 #define DEFAULT_RETRY_IO_COUNT 8 /* maximum retry read IO or flush count */
585 /* congestion wait timeout value, default: 20ms */
586 #define DEFAULT_IO_TIMEOUT (msecs_to_jiffies(20))
588 /* maximum retry quota flush count */
589 #define DEFAULT_RETRY_QUOTA_FLUSH_COUNT 8
591 /* maximum retry of EIO'ed page */
592 #define MAX_RETRY_PAGE_EIO 100
594 #define F2FS_LINK_MAX 0xffffffff /* maximum link count per file */
596 #define MAX_DIR_RA_PAGES 4 /* maximum ra pages of dir */
598 /* dirty segments threshold for triggering CP */
599 #define DEFAULT_DIRTY_THRESHOLD 4
601 #define RECOVERY_MAX_RA_BLOCKS BIO_MAX_VECS
602 #define RECOVERY_MIN_RA_BLOCKS 1
604 #define F2FS_ONSTACK_PAGES 16 /* nr of onstack pages */
606 /* for in-memory extent cache entry */
607 #define F2FS_MIN_EXTENT_LEN 64 /* minimum extent length */
609 /* number of extent info in extent cache we try to shrink */
610 #define READ_EXTENT_CACHE_SHRINK_NUMBER 128
612 /* extent cache type */
619 struct rb_node rb_node; /* rb node located in rb-tree */
620 unsigned int ofs; /* start offset of the entry */
621 unsigned int len; /* length of the entry */
625 unsigned int fofs; /* start offset in a file */
626 unsigned int len; /* length of the extent */
628 /* read extent_cache */
630 /* start block address of the extent */
632 #ifdef CONFIG_F2FS_FS_COMPRESSION
633 /* physical extent length of compressed blocks */
641 struct rb_node rb_node; /* rb node located in rb-tree */
642 struct extent_info ei; /* extent info */
643 struct list_head list; /* node in global extent list of sbi */
644 struct extent_tree *et; /* extent tree pointer */
648 nid_t ino; /* inode number */
649 enum extent_type type; /* keep the extent tree type */
650 struct rb_root_cached root; /* root of extent info rb-tree */
651 struct extent_node *cached_en; /* recently accessed extent node */
652 struct list_head list; /* to be used by sbi->zombie_list */
653 rwlock_t lock; /* protect extent info rb-tree */
654 atomic_t node_cnt; /* # of extent node in rb-tree*/
655 bool largest_updated; /* largest extent updated */
656 struct extent_info largest; /* largest cached extent for EX_READ */
659 struct extent_tree_info {
660 struct radix_tree_root extent_tree_root;/* cache extent cache entries */
661 struct mutex extent_tree_lock; /* locking extent radix tree */
662 struct list_head extent_list; /* lru list for shrinker */
663 spinlock_t extent_lock; /* locking extent lru list */
664 atomic_t total_ext_tree; /* extent tree count */
665 struct list_head zombie_list; /* extent zombie tree list */
666 atomic_t total_zombie_tree; /* extent zombie tree count */
667 atomic_t total_ext_node; /* extent info count */
671 * This structure is taken from ext4_map_blocks.
673 * Note that, however, f2fs uses NEW and MAPPED flags for f2fs_map_blocks().
675 #define F2FS_MAP_NEW (1 << BH_New)
676 #define F2FS_MAP_MAPPED (1 << BH_Mapped)
677 #define F2FS_MAP_UNWRITTEN (1 << BH_Unwritten)
678 #define F2FS_MAP_FLAGS (F2FS_MAP_NEW | F2FS_MAP_MAPPED |\
681 struct f2fs_map_blocks {
682 struct block_device *m_bdev; /* for multi-device dio */
686 unsigned int m_flags;
687 pgoff_t *m_next_pgofs; /* point next possible non-hole pgofs */
688 pgoff_t *m_next_extent; /* point to next possible extent */
690 bool m_may_create; /* indicate it is from write path */
691 bool m_multidev_dio; /* indicate it allows multi-device dio */
694 /* for flag in get_data_block */
696 F2FS_GET_BLOCK_DEFAULT,
697 F2FS_GET_BLOCK_FIEMAP,
700 F2FS_GET_BLOCK_PRE_DIO,
701 F2FS_GET_BLOCK_PRE_AIO,
702 F2FS_GET_BLOCK_PRECACHE,
706 * i_advise uses FADVISE_XXX_BIT. We can add additional hints later.
708 #define FADVISE_COLD_BIT 0x01
709 #define FADVISE_LOST_PINO_BIT 0x02
710 #define FADVISE_ENCRYPT_BIT 0x04
711 #define FADVISE_ENC_NAME_BIT 0x08
712 #define FADVISE_KEEP_SIZE_BIT 0x10
713 #define FADVISE_HOT_BIT 0x20
714 #define FADVISE_VERITY_BIT 0x40
715 #define FADVISE_TRUNC_BIT 0x80
717 #define FADVISE_MODIFIABLE_BITS (FADVISE_COLD_BIT | FADVISE_HOT_BIT)
719 #define file_is_cold(inode) is_file(inode, FADVISE_COLD_BIT)
720 #define file_set_cold(inode) set_file(inode, FADVISE_COLD_BIT)
721 #define file_clear_cold(inode) clear_file(inode, FADVISE_COLD_BIT)
723 #define file_wrong_pino(inode) is_file(inode, FADVISE_LOST_PINO_BIT)
724 #define file_lost_pino(inode) set_file(inode, FADVISE_LOST_PINO_BIT)
725 #define file_got_pino(inode) clear_file(inode, FADVISE_LOST_PINO_BIT)
727 #define file_is_encrypt(inode) is_file(inode, FADVISE_ENCRYPT_BIT)
728 #define file_set_encrypt(inode) set_file(inode, FADVISE_ENCRYPT_BIT)
730 #define file_enc_name(inode) is_file(inode, FADVISE_ENC_NAME_BIT)
731 #define file_set_enc_name(inode) set_file(inode, FADVISE_ENC_NAME_BIT)
733 #define file_keep_isize(inode) is_file(inode, FADVISE_KEEP_SIZE_BIT)
734 #define file_set_keep_isize(inode) set_file(inode, FADVISE_KEEP_SIZE_BIT)
736 #define file_is_hot(inode) is_file(inode, FADVISE_HOT_BIT)
737 #define file_set_hot(inode) set_file(inode, FADVISE_HOT_BIT)
738 #define file_clear_hot(inode) clear_file(inode, FADVISE_HOT_BIT)
740 #define file_is_verity(inode) is_file(inode, FADVISE_VERITY_BIT)
741 #define file_set_verity(inode) set_file(inode, FADVISE_VERITY_BIT)
743 #define file_should_truncate(inode) is_file(inode, FADVISE_TRUNC_BIT)
744 #define file_need_truncate(inode) set_file(inode, FADVISE_TRUNC_BIT)
745 #define file_dont_truncate(inode) clear_file(inode, FADVISE_TRUNC_BIT)
747 #define DEF_DIR_LEVEL 0
754 /* used for f2fs_inode_info->flags */
756 FI_NEW_INODE, /* indicate newly allocated inode */
757 FI_DIRTY_INODE, /* indicate inode is dirty or not */
758 FI_AUTO_RECOVER, /* indicate inode is recoverable */
759 FI_DIRTY_DIR, /* indicate directory has dirty pages */
760 FI_INC_LINK, /* need to increment i_nlink */
761 FI_ACL_MODE, /* indicate acl mode */
762 FI_NO_ALLOC, /* should not allocate any blocks */
763 FI_FREE_NID, /* free allocated nide */
764 FI_NO_EXTENT, /* not to use the extent cache */
765 FI_INLINE_XATTR, /* used for inline xattr */
766 FI_INLINE_DATA, /* used for inline data*/
767 FI_INLINE_DENTRY, /* used for inline dentry */
768 FI_APPEND_WRITE, /* inode has appended data */
769 FI_UPDATE_WRITE, /* inode has in-place-update data */
770 FI_NEED_IPU, /* used for ipu per file */
771 FI_ATOMIC_FILE, /* indicate atomic file */
772 FI_DATA_EXIST, /* indicate data exists */
773 FI_INLINE_DOTS, /* indicate inline dot dentries */
774 FI_SKIP_WRITES, /* should skip data page writeback */
775 FI_OPU_WRITE, /* used for opu per file */
776 FI_DIRTY_FILE, /* indicate regular/symlink has dirty pages */
777 FI_PREALLOCATED_ALL, /* all blocks for write were preallocated */
778 FI_HOT_DATA, /* indicate file is hot */
779 FI_EXTRA_ATTR, /* indicate file has extra attribute */
780 FI_PROJ_INHERIT, /* indicate file inherits projectid */
781 FI_PIN_FILE, /* indicate file should not be gced */
782 FI_VERITY_IN_PROGRESS, /* building fs-verity Merkle tree */
783 FI_COMPRESSED_FILE, /* indicate file's data can be compressed */
784 FI_COMPRESS_CORRUPT, /* indicate compressed cluster is corrupted */
785 FI_MMAP_FILE, /* indicate file was mmapped */
786 FI_ENABLE_COMPRESS, /* enable compression in "user" compression mode */
787 FI_COMPRESS_RELEASED, /* compressed blocks were released */
788 FI_ALIGNED_WRITE, /* enable aligned write */
789 FI_COW_FILE, /* indicate COW file */
790 FI_ATOMIC_COMMITTED, /* indicate atomic commit completed except disk sync */
791 FI_MAX, /* max flag, never be used */
794 struct f2fs_inode_info {
795 struct inode vfs_inode; /* serve a vfs inode */
796 unsigned long i_flags; /* keep an inode flags for ioctl */
797 unsigned char i_advise; /* use to give file attribute hints */
798 unsigned char i_dir_level; /* use for dentry level for large dir */
799 unsigned int i_current_depth; /* only for directory depth */
800 /* for gc failure statistic */
801 unsigned int i_gc_failures[MAX_GC_FAILURE];
802 unsigned int i_pino; /* parent inode number */
803 umode_t i_acl_mode; /* keep file acl mode temporarily */
805 /* Use below internally in f2fs*/
806 unsigned long flags[BITS_TO_LONGS(FI_MAX)]; /* use to pass per-file flags */
807 struct f2fs_rwsem i_sem; /* protect fi info */
808 atomic_t dirty_pages; /* # of dirty pages */
809 f2fs_hash_t chash; /* hash value of given file name */
810 unsigned int clevel; /* maximum level of given file name */
811 struct task_struct *task; /* lookup and create consistency */
812 struct task_struct *cp_task; /* separate cp/wb IO stats*/
813 struct task_struct *wb_task; /* indicate inode is in context of writeback */
814 nid_t i_xattr_nid; /* node id that contains xattrs */
815 loff_t last_disk_size; /* lastly written file size */
816 spinlock_t i_size_lock; /* protect last_disk_size */
819 struct dquot *i_dquot[MAXQUOTAS];
821 /* quota space reservation, managed internally by quota code */
822 qsize_t i_reserved_quota;
824 struct list_head dirty_list; /* dirty list for dirs and files */
825 struct list_head gdirty_list; /* linked in global dirty list */
826 struct task_struct *atomic_write_task; /* store atomic write task */
827 struct extent_tree *extent_tree[NR_EXTENT_CACHES];
828 /* cached extent_tree entry */
829 struct inode *cow_inode; /* copy-on-write inode for atomic write */
831 /* avoid racing between foreground op and gc */
832 struct f2fs_rwsem i_gc_rwsem[2];
833 struct f2fs_rwsem i_xattr_sem; /* avoid racing between reading and changing EAs */
835 int i_extra_isize; /* size of extra space located in i_addr */
836 kprojid_t i_projid; /* id for project quota */
837 int i_inline_xattr_size; /* inline xattr size */
838 struct timespec64 i_crtime; /* inode creation time */
839 struct timespec64 i_disk_time[4];/* inode disk times */
841 /* for file compress */
842 atomic_t i_compr_blocks; /* # of compressed blocks */
843 unsigned char i_compress_algorithm; /* algorithm type */
844 unsigned char i_log_cluster_size; /* log of cluster size */
845 unsigned char i_compress_level; /* compress level (lz4hc,zstd) */
846 unsigned char i_compress_flag; /* compress flag */
847 unsigned int i_cluster_size; /* cluster size */
849 unsigned int atomic_write_cnt;
850 loff_t original_i_size; /* original i_size before atomic write */
853 static inline void get_read_extent_info(struct extent_info *ext,
854 struct f2fs_extent *i_ext)
856 ext->fofs = le32_to_cpu(i_ext->fofs);
857 ext->blk = le32_to_cpu(i_ext->blk);
858 ext->len = le32_to_cpu(i_ext->len);
861 static inline void set_raw_read_extent(struct extent_info *ext,
862 struct f2fs_extent *i_ext)
864 i_ext->fofs = cpu_to_le32(ext->fofs);
865 i_ext->blk = cpu_to_le32(ext->blk);
866 i_ext->len = cpu_to_le32(ext->len);
869 static inline bool __is_discard_mergeable(struct discard_info *back,
870 struct discard_info *front, unsigned int max_len)
872 return (back->lstart + back->len == front->lstart) &&
873 (back->len + front->len <= max_len);
876 static inline bool __is_discard_back_mergeable(struct discard_info *cur,
877 struct discard_info *back, unsigned int max_len)
879 return __is_discard_mergeable(back, cur, max_len);
882 static inline bool __is_discard_front_mergeable(struct discard_info *cur,
883 struct discard_info *front, unsigned int max_len)
885 return __is_discard_mergeable(cur, front, max_len);
889 * For free nid management
892 FREE_NID, /* newly added to free nid list */
893 PREALLOC_NID, /* it is preallocated */
904 struct f2fs_nm_info {
905 block_t nat_blkaddr; /* base disk address of NAT */
906 nid_t max_nid; /* maximum possible node ids */
907 nid_t available_nids; /* # of available node ids */
908 nid_t next_scan_nid; /* the next nid to be scanned */
909 nid_t max_rf_node_blocks; /* max # of nodes for recovery */
910 unsigned int ram_thresh; /* control the memory footprint */
911 unsigned int ra_nid_pages; /* # of nid pages to be readaheaded */
912 unsigned int dirty_nats_ratio; /* control dirty nats ratio threshold */
914 /* NAT cache management */
915 struct radix_tree_root nat_root;/* root of the nat entry cache */
916 struct radix_tree_root nat_set_root;/* root of the nat set cache */
917 struct f2fs_rwsem nat_tree_lock; /* protect nat entry tree */
918 struct list_head nat_entries; /* cached nat entry list (clean) */
919 spinlock_t nat_list_lock; /* protect clean nat entry list */
920 unsigned int nat_cnt[MAX_NAT_STATE]; /* the # of cached nat entries */
921 unsigned int nat_blocks; /* # of nat blocks */
923 /* free node ids management */
924 struct radix_tree_root free_nid_root;/* root of the free_nid cache */
925 struct list_head free_nid_list; /* list for free nids excluding preallocated nids */
926 unsigned int nid_cnt[MAX_NID_STATE]; /* the number of free node id */
927 spinlock_t nid_list_lock; /* protect nid lists ops */
928 struct mutex build_lock; /* lock for build free nids */
929 unsigned char **free_nid_bitmap;
930 unsigned char *nat_block_bitmap;
931 unsigned short *free_nid_count; /* free nid count of NAT block */
934 char *nat_bitmap; /* NAT bitmap pointer */
936 unsigned int nat_bits_blocks; /* # of nat bits blocks */
937 unsigned char *nat_bits; /* NAT bits blocks */
938 unsigned char *full_nat_bits; /* full NAT pages */
939 unsigned char *empty_nat_bits; /* empty NAT pages */
940 #ifdef CONFIG_F2FS_CHECK_FS
941 char *nat_bitmap_mir; /* NAT bitmap mirror */
943 int bitmap_size; /* bitmap size */
947 * this structure is used as one of function parameters.
948 * all the information are dedicated to a given direct node block determined
949 * by the data offset in a file.
951 struct dnode_of_data {
952 struct inode *inode; /* vfs inode pointer */
953 struct page *inode_page; /* its inode page, NULL is possible */
954 struct page *node_page; /* cached direct node page */
955 nid_t nid; /* node id of the direct node block */
956 unsigned int ofs_in_node; /* data offset in the node page */
957 bool inode_page_locked; /* inode page is locked or not */
958 bool node_changed; /* is node block changed */
959 char cur_level; /* level of hole node page */
960 char max_level; /* level of current page located */
961 block_t data_blkaddr; /* block address of the node block */
964 static inline void set_new_dnode(struct dnode_of_data *dn, struct inode *inode,
965 struct page *ipage, struct page *npage, nid_t nid)
967 memset(dn, 0, sizeof(*dn));
969 dn->inode_page = ipage;
970 dn->node_page = npage;
977 * By default, there are 6 active log areas across the whole main area.
978 * When considering hot and cold data separation to reduce cleaning overhead,
979 * we split 3 for data logs and 3 for node logs as hot, warm, and cold types,
981 * In the current design, you should not change the numbers intentionally.
982 * Instead, as a mount option such as active_logs=x, you can use 2, 4, and 6
983 * logs individually according to the underlying devices. (default: 6)
984 * Just in case, on-disk layout covers maximum 16 logs that consist of 8 for
985 * data and 8 for node logs.
987 #define NR_CURSEG_DATA_TYPE (3)
988 #define NR_CURSEG_NODE_TYPE (3)
989 #define NR_CURSEG_INMEM_TYPE (2)
990 #define NR_CURSEG_RO_TYPE (2)
991 #define NR_CURSEG_PERSIST_TYPE (NR_CURSEG_DATA_TYPE + NR_CURSEG_NODE_TYPE)
992 #define NR_CURSEG_TYPE (NR_CURSEG_INMEM_TYPE + NR_CURSEG_PERSIST_TYPE)
995 CURSEG_HOT_DATA = 0, /* directory entry blocks */
996 CURSEG_WARM_DATA, /* data blocks */
997 CURSEG_COLD_DATA, /* multimedia or GCed data blocks */
998 CURSEG_HOT_NODE, /* direct node blocks of directory files */
999 CURSEG_WARM_NODE, /* direct node blocks of normal files */
1000 CURSEG_COLD_NODE, /* indirect node blocks */
1001 NR_PERSISTENT_LOG, /* number of persistent log */
1002 CURSEG_COLD_DATA_PINNED = NR_PERSISTENT_LOG,
1003 /* pinned file that needs consecutive block address */
1004 CURSEG_ALL_DATA_ATGC, /* SSR alloctor in hot/warm/cold data area */
1005 NO_CHECK_TYPE, /* number of persistent & inmem log */
1009 struct completion wait;
1010 struct llist_node llnode;
1015 struct flush_cmd_control {
1016 struct task_struct *f2fs_issue_flush; /* flush thread */
1017 wait_queue_head_t flush_wait_queue; /* waiting queue for wake-up */
1018 atomic_t issued_flush; /* # of issued flushes */
1019 atomic_t queued_flush; /* # of queued flushes */
1020 struct llist_head issue_list; /* list for command issue */
1021 struct llist_node *dispatch_list; /* list for command dispatch */
1024 struct f2fs_sm_info {
1025 struct sit_info *sit_info; /* whole segment information */
1026 struct free_segmap_info *free_info; /* free segment information */
1027 struct dirty_seglist_info *dirty_info; /* dirty segment information */
1028 struct curseg_info *curseg_array; /* active segment information */
1030 struct f2fs_rwsem curseg_lock; /* for preventing curseg change */
1032 block_t seg0_blkaddr; /* block address of 0'th segment */
1033 block_t main_blkaddr; /* start block address of main area */
1034 block_t ssa_blkaddr; /* start block address of SSA area */
1036 unsigned int segment_count; /* total # of segments */
1037 unsigned int main_segments; /* # of segments in main area */
1038 unsigned int reserved_segments; /* # of reserved segments */
1039 unsigned int additional_reserved_segments;/* reserved segs for IO align feature */
1040 unsigned int ovp_segments; /* # of overprovision segments */
1042 /* a threshold to reclaim prefree segments */
1043 unsigned int rec_prefree_segments;
1045 /* for batched trimming */
1046 unsigned int trim_sections; /* # of sections to trim */
1048 struct list_head sit_entry_set; /* sit entry set list */
1050 unsigned int ipu_policy; /* in-place-update policy */
1051 unsigned int min_ipu_util; /* in-place-update threshold */
1052 unsigned int min_fsync_blocks; /* threshold for fsync */
1053 unsigned int min_seq_blocks; /* threshold for sequential blocks */
1054 unsigned int min_hot_blocks; /* threshold for hot block allocation */
1055 unsigned int min_ssr_sections; /* threshold to trigger SSR allocation */
1057 /* for flush command control */
1058 struct flush_cmd_control *fcc_info;
1060 /* for discard command control */
1061 struct discard_cmd_control *dcc_info;
1068 * COUNT_TYPE for monitoring
1070 * f2fs monitors the number of several block types such as on-writeback,
1071 * dirty dentry blocks, dirty node blocks, and dirty meta blocks.
1073 #define WB_DATA_TYPE(p, f) \
1074 (f || f2fs_is_cp_guaranteed(p) ? F2FS_WB_CP_DATA : F2FS_WB_DATA)
1093 * The below are the page types of bios used in submit_bio().
1094 * The available types are:
1095 * DATA User data pages. It operates as async mode.
1096 * NODE Node pages. It operates as async mode.
1097 * META FS metadata pages such as SIT, NAT, CP.
1098 * NR_PAGE_TYPE The number of page types.
1099 * META_FLUSH Make sure the previous pages are written
1100 * with waiting the bio's completion
1101 * ... Only can be used with META.
1103 #define PAGE_TYPE_OF_BIO(type) ((type) > META ? META : (type))
1106 NODE = 1, /* should not change this */
1110 IPU, /* the below types are used by tracepoints only. */
1115 HOT = 0, /* must be zero for meta bio */
1121 enum need_lock_type {
1127 enum cp_reason_type {
1143 APP_DIRECT_IO, /* app direct write IOs */
1144 APP_BUFFERED_IO, /* app buffered write IOs */
1145 APP_WRITE_IO, /* app write IOs */
1146 APP_MAPPED_IO, /* app mapped IOs */
1147 APP_BUFFERED_CDATA_IO, /* app buffered write IOs on compressed file */
1148 APP_MAPPED_CDATA_IO, /* app mapped write IOs on compressed file */
1149 FS_DATA_IO, /* data IOs from kworker/fsync/reclaimer */
1150 FS_CDATA_IO, /* data IOs from kworker/fsync/reclaimer on compressed file */
1151 FS_NODE_IO, /* node IOs from kworker/fsync/reclaimer */
1152 FS_META_IO, /* meta IOs from kworker/reclaimer */
1153 FS_GC_DATA_IO, /* data IOs from forground gc */
1154 FS_GC_NODE_IO, /* node IOs from forground gc */
1155 FS_CP_DATA_IO, /* data IOs from checkpoint */
1156 FS_CP_NODE_IO, /* node IOs from checkpoint */
1157 FS_CP_META_IO, /* meta IOs from checkpoint */
1160 APP_DIRECT_READ_IO, /* app direct read IOs */
1161 APP_BUFFERED_READ_IO, /* app buffered read IOs */
1162 APP_READ_IO, /* app read IOs */
1163 APP_MAPPED_READ_IO, /* app mapped read IOs */
1164 APP_BUFFERED_CDATA_READ_IO, /* app buffered read IOs on compressed file */
1165 APP_MAPPED_CDATA_READ_IO, /* app mapped read IOs on compressed file */
1166 FS_DATA_READ_IO, /* data read IOs */
1167 FS_GDATA_READ_IO, /* data read IOs from background gc */
1168 FS_CDATA_READ_IO, /* compressed data read IOs */
1169 FS_NODE_READ_IO, /* node read IOs */
1170 FS_META_READ_IO, /* meta read IOs */
1173 FS_DISCARD, /* discard */
1177 struct f2fs_io_info {
1178 struct f2fs_sb_info *sbi; /* f2fs_sb_info pointer */
1179 nid_t ino; /* inode number */
1180 enum page_type type; /* contains DATA/NODE/META/META_FLUSH */
1181 enum temp_type temp; /* contains HOT/WARM/COLD */
1182 enum req_op op; /* contains REQ_OP_ */
1183 blk_opf_t op_flags; /* req_flag_bits */
1184 block_t new_blkaddr; /* new block address to be written */
1185 block_t old_blkaddr; /* old block address before Cow */
1186 struct page *page; /* page to be written */
1187 struct page *encrypted_page; /* encrypted page */
1188 struct page *compressed_page; /* compressed page */
1189 struct list_head list; /* serialize IOs */
1190 unsigned int compr_blocks; /* # of compressed block addresses */
1191 unsigned int need_lock:8; /* indicate we need to lock cp_rwsem */
1192 unsigned int version:8; /* version of the node */
1193 unsigned int submitted:1; /* indicate IO submission */
1194 unsigned int in_list:1; /* indicate fio is in io_list */
1195 unsigned int is_por:1; /* indicate IO is from recovery or not */
1196 unsigned int retry:1; /* need to reallocate block address */
1197 unsigned int encrypted:1; /* indicate file is encrypted */
1198 unsigned int post_read:1; /* require post read */
1199 enum iostat_type io_type; /* io type */
1200 struct writeback_control *io_wbc; /* writeback control */
1201 struct bio **bio; /* bio for ipu */
1202 sector_t *last_block; /* last block number in bio */
1207 struct list_head list;
1210 #define is_read_io(rw) ((rw) == READ)
1211 struct f2fs_bio_info {
1212 struct f2fs_sb_info *sbi; /* f2fs superblock */
1213 struct bio *bio; /* bios to merge */
1214 sector_t last_block_in_bio; /* last block number */
1215 struct f2fs_io_info fio; /* store buffered io info. */
1216 struct f2fs_rwsem io_rwsem; /* blocking op for bio */
1217 spinlock_t io_lock; /* serialize DATA/NODE IOs */
1218 struct list_head io_list; /* track fios */
1219 struct list_head bio_list; /* bio entry list head */
1220 struct f2fs_rwsem bio_list_lock; /* lock to protect bio entry list */
1223 #define FDEV(i) (sbi->devs[i])
1224 #define RDEV(i) (raw_super->devs[i])
1225 struct f2fs_dev_info {
1226 struct block_device *bdev;
1227 char path[MAX_PATH_LEN];
1228 unsigned int total_segments;
1231 #ifdef CONFIG_BLK_DEV_ZONED
1232 unsigned int nr_blkz; /* Total number of zones */
1233 unsigned long *blkz_seq; /* Bitmap indicating sequential zones */
1238 DIR_INODE, /* for dirty dir inode */
1239 FILE_INODE, /* for dirty regular/symlink inode */
1240 DIRTY_META, /* for all dirtied inode metadata */
1244 /* for inner inode cache management */
1245 struct inode_management {
1246 struct radix_tree_root ino_root; /* ino entry array */
1247 spinlock_t ino_lock; /* for ino entry lock */
1248 struct list_head ino_list; /* inode list head */
1249 unsigned long ino_num; /* number of entries */
1253 struct atgc_management {
1254 bool atgc_enabled; /* ATGC is enabled or not */
1255 struct rb_root_cached root; /* root of victim rb-tree */
1256 struct list_head victim_list; /* linked with all victim entries */
1257 unsigned int victim_count; /* victim count in rb-tree */
1258 unsigned int candidate_ratio; /* candidate ratio */
1259 unsigned int max_candidate_count; /* max candidate count */
1260 unsigned int age_weight; /* age weight, vblock_weight = 100 - age_weight */
1261 unsigned long long age_threshold; /* age threshold */
1264 struct f2fs_gc_control {
1265 unsigned int victim_segno; /* target victim segment number */
1266 int init_gc_type; /* FG_GC or BG_GC */
1267 bool no_bg_gc; /* check the space and stop bg_gc */
1268 bool should_migrate_blocks; /* should migrate blocks */
1269 bool err_gc_skipped; /* return EAGAIN if GC skipped */
1270 unsigned int nr_free_secs; /* # of free sections to do GC */
1273 /* For s_flag in struct f2fs_sb_info */
1275 SBI_IS_DIRTY, /* dirty flag for checkpoint */
1276 SBI_IS_CLOSE, /* specify unmounting */
1277 SBI_NEED_FSCK, /* need fsck.f2fs to fix */
1278 SBI_POR_DOING, /* recovery is doing or not */
1279 SBI_NEED_SB_WRITE, /* need to recover superblock */
1280 SBI_NEED_CP, /* need to checkpoint */
1281 SBI_IS_SHUTDOWN, /* shutdown by ioctl */
1282 SBI_IS_RECOVERED, /* recovered orphan/data */
1283 SBI_CP_DISABLED, /* CP was disabled last mount */
1284 SBI_CP_DISABLED_QUICK, /* CP was disabled quickly */
1285 SBI_QUOTA_NEED_FLUSH, /* need to flush quota info in CP */
1286 SBI_QUOTA_SKIP_FLUSH, /* skip flushing quota in current CP */
1287 SBI_QUOTA_NEED_REPAIR, /* quota file may be corrupted */
1288 SBI_IS_RESIZEFS, /* resizefs is in process */
1289 SBI_IS_FREEZING, /* freezefs is in process */
1298 UMOUNT_DISCARD_TIMEOUT,
1314 BGGC_MODE_ON, /* background gc is on */
1315 BGGC_MODE_OFF, /* background gc is off */
1317 * background gc is on, migrating blocks
1318 * like foreground gc
1323 FS_MODE_ADAPTIVE, /* use both lfs/ssr allocation */
1324 FS_MODE_LFS, /* use lfs allocation only */
1325 FS_MODE_FRAGMENT_SEG, /* segment fragmentation mode */
1326 FS_MODE_FRAGMENT_BLK, /* block fragmentation mode */
1330 ALLOC_MODE_DEFAULT, /* stay default */
1331 ALLOC_MODE_REUSE, /* reuse segments as much as possible */
1335 FSYNC_MODE_POSIX, /* fsync follows posix semantics */
1336 FSYNC_MODE_STRICT, /* fsync behaves in line with ext4 */
1337 FSYNC_MODE_NOBARRIER, /* fsync behaves nobarrier based on posix */
1342 * automatically compress compression
1346 * automatical compression is disabled.
1347 * user can control the file compression
1353 DISCARD_UNIT_BLOCK, /* basic discard unit is block */
1354 DISCARD_UNIT_SEGMENT, /* basic discard unit is segment */
1355 DISCARD_UNIT_SECTION, /* basic discard unit is section */
1359 MEMORY_MODE_NORMAL, /* memory mode for normal devices */
1360 MEMORY_MODE_LOW, /* memory mode for low memry devices */
1365 static inline int f2fs_test_bit(unsigned int nr, char *addr);
1366 static inline void f2fs_set_bit(unsigned int nr, char *addr);
1367 static inline void f2fs_clear_bit(unsigned int nr, char *addr);
1370 * Layout of f2fs page.private:
1372 * Layout A: lowest bit should be 1
1373 * | bit0 = 1 | bit1 | bit2 | ... | bit MAX | private data .... |
1374 * bit 0 PAGE_PRIVATE_NOT_POINTER
1375 * bit 1 PAGE_PRIVATE_ATOMIC_WRITE
1376 * bit 2 PAGE_PRIVATE_DUMMY_WRITE
1377 * bit 3 PAGE_PRIVATE_ONGOING_MIGRATION
1378 * bit 4 PAGE_PRIVATE_INLINE_INODE
1379 * bit 5 PAGE_PRIVATE_REF_RESOURCE
1380 * bit 6- f2fs private data
1382 * Layout B: lowest bit should be 0
1383 * page.private is a wrapped pointer.
1386 PAGE_PRIVATE_NOT_POINTER, /* private contains non-pointer data */
1387 PAGE_PRIVATE_ATOMIC_WRITE, /* data page from atomic write path */
1388 PAGE_PRIVATE_DUMMY_WRITE, /* data page for padding aligned IO */
1389 PAGE_PRIVATE_ONGOING_MIGRATION, /* data page which is on-going migrating */
1390 PAGE_PRIVATE_INLINE_INODE, /* inode page contains inline data */
1391 PAGE_PRIVATE_REF_RESOURCE, /* dirty page has referenced resources */
1395 #define PAGE_PRIVATE_GET_FUNC(name, flagname) \
1396 static inline bool page_private_##name(struct page *page) \
1398 return PagePrivate(page) && \
1399 test_bit(PAGE_PRIVATE_NOT_POINTER, &page_private(page)) && \
1400 test_bit(PAGE_PRIVATE_##flagname, &page_private(page)); \
1403 #define PAGE_PRIVATE_SET_FUNC(name, flagname) \
1404 static inline void set_page_private_##name(struct page *page) \
1406 if (!PagePrivate(page)) { \
1408 SetPagePrivate(page); \
1409 set_page_private(page, 0); \
1411 set_bit(PAGE_PRIVATE_NOT_POINTER, &page_private(page)); \
1412 set_bit(PAGE_PRIVATE_##flagname, &page_private(page)); \
1415 #define PAGE_PRIVATE_CLEAR_FUNC(name, flagname) \
1416 static inline void clear_page_private_##name(struct page *page) \
1418 clear_bit(PAGE_PRIVATE_##flagname, &page_private(page)); \
1419 if (page_private(page) == BIT(PAGE_PRIVATE_NOT_POINTER)) { \
1420 set_page_private(page, 0); \
1421 if (PagePrivate(page)) { \
1422 ClearPagePrivate(page); \
1428 PAGE_PRIVATE_GET_FUNC(nonpointer, NOT_POINTER);
1429 PAGE_PRIVATE_GET_FUNC(reference, REF_RESOURCE);
1430 PAGE_PRIVATE_GET_FUNC(inline, INLINE_INODE);
1431 PAGE_PRIVATE_GET_FUNC(gcing, ONGOING_MIGRATION);
1432 PAGE_PRIVATE_GET_FUNC(atomic, ATOMIC_WRITE);
1433 PAGE_PRIVATE_GET_FUNC(dummy, DUMMY_WRITE);
1435 PAGE_PRIVATE_SET_FUNC(reference, REF_RESOURCE);
1436 PAGE_PRIVATE_SET_FUNC(inline, INLINE_INODE);
1437 PAGE_PRIVATE_SET_FUNC(gcing, ONGOING_MIGRATION);
1438 PAGE_PRIVATE_SET_FUNC(atomic, ATOMIC_WRITE);
1439 PAGE_PRIVATE_SET_FUNC(dummy, DUMMY_WRITE);
1441 PAGE_PRIVATE_CLEAR_FUNC(reference, REF_RESOURCE);
1442 PAGE_PRIVATE_CLEAR_FUNC(inline, INLINE_INODE);
1443 PAGE_PRIVATE_CLEAR_FUNC(gcing, ONGOING_MIGRATION);
1444 PAGE_PRIVATE_CLEAR_FUNC(atomic, ATOMIC_WRITE);
1445 PAGE_PRIVATE_CLEAR_FUNC(dummy, DUMMY_WRITE);
1447 static inline unsigned long get_page_private_data(struct page *page)
1449 unsigned long data = page_private(page);
1451 if (!test_bit(PAGE_PRIVATE_NOT_POINTER, &data))
1453 return data >> PAGE_PRIVATE_MAX;
1456 static inline void set_page_private_data(struct page *page, unsigned long data)
1458 if (!PagePrivate(page)) {
1460 SetPagePrivate(page);
1461 set_page_private(page, 0);
1463 set_bit(PAGE_PRIVATE_NOT_POINTER, &page_private(page));
1464 page_private(page) |= data << PAGE_PRIVATE_MAX;
1467 static inline void clear_page_private_data(struct page *page)
1469 page_private(page) &= GENMASK(PAGE_PRIVATE_MAX - 1, 0);
1470 if (page_private(page) == BIT(PAGE_PRIVATE_NOT_POINTER)) {
1471 set_page_private(page, 0);
1472 if (PagePrivate(page)) {
1473 ClearPagePrivate(page);
1479 /* For compression */
1480 enum compress_algorithm_type {
1488 enum compress_flag {
1493 #define COMPRESS_WATERMARK 20
1494 #define COMPRESS_PERCENT 20
1496 #define COMPRESS_DATA_RESERVED_SIZE 4
1497 struct compress_data {
1498 __le32 clen; /* compressed data size */
1499 __le32 chksum; /* compressed data chksum */
1500 __le32 reserved[COMPRESS_DATA_RESERVED_SIZE]; /* reserved */
1501 u8 cdata[]; /* compressed data */
1504 #define COMPRESS_HEADER_SIZE (sizeof(struct compress_data))
1506 #define F2FS_COMPRESSED_PAGE_MAGIC 0xF5F2C000
1508 #define F2FS_ZSTD_DEFAULT_CLEVEL 1
1510 #define COMPRESS_LEVEL_OFFSET 8
1512 /* compress context */
1513 struct compress_ctx {
1514 struct inode *inode; /* inode the context belong to */
1515 pgoff_t cluster_idx; /* cluster index number */
1516 unsigned int cluster_size; /* page count in cluster */
1517 unsigned int log_cluster_size; /* log of cluster size */
1518 struct page **rpages; /* pages store raw data in cluster */
1519 unsigned int nr_rpages; /* total page number in rpages */
1520 struct page **cpages; /* pages store compressed data in cluster */
1521 unsigned int nr_cpages; /* total page number in cpages */
1522 unsigned int valid_nr_cpages; /* valid page number in cpages */
1523 void *rbuf; /* virtual mapped address on rpages */
1524 struct compress_data *cbuf; /* virtual mapped address on cpages */
1525 size_t rlen; /* valid data length in rbuf */
1526 size_t clen; /* valid data length in cbuf */
1527 void *private; /* payload buffer for specified compression algorithm */
1528 void *private2; /* extra payload buffer */
1531 /* compress context for write IO path */
1532 struct compress_io_ctx {
1533 u32 magic; /* magic number to indicate page is compressed */
1534 struct inode *inode; /* inode the context belong to */
1535 struct page **rpages; /* pages store raw data in cluster */
1536 unsigned int nr_rpages; /* total page number in rpages */
1537 atomic_t pending_pages; /* in-flight compressed page count */
1540 /* Context for decompressing one cluster on the read IO path */
1541 struct decompress_io_ctx {
1542 u32 magic; /* magic number to indicate page is compressed */
1543 struct inode *inode; /* inode the context belong to */
1544 pgoff_t cluster_idx; /* cluster index number */
1545 unsigned int cluster_size; /* page count in cluster */
1546 unsigned int log_cluster_size; /* log of cluster size */
1547 struct page **rpages; /* pages store raw data in cluster */
1548 unsigned int nr_rpages; /* total page number in rpages */
1549 struct page **cpages; /* pages store compressed data in cluster */
1550 unsigned int nr_cpages; /* total page number in cpages */
1551 struct page **tpages; /* temp pages to pad holes in cluster */
1552 void *rbuf; /* virtual mapped address on rpages */
1553 struct compress_data *cbuf; /* virtual mapped address on cpages */
1554 size_t rlen; /* valid data length in rbuf */
1555 size_t clen; /* valid data length in cbuf */
1558 * The number of compressed pages remaining to be read in this cluster.
1559 * This is initially nr_cpages. It is decremented by 1 each time a page
1560 * has been read (or failed to be read). When it reaches 0, the cluster
1561 * is decompressed (or an error is reported).
1563 * If an error occurs before all the pages have been submitted for I/O,
1564 * then this will never reach 0. In this case the I/O submitter is
1565 * responsible for calling f2fs_decompress_end_io() instead.
1567 atomic_t remaining_pages;
1570 * Number of references to this decompress_io_ctx.
1572 * One reference is held for I/O completion. This reference is dropped
1573 * after the pagecache pages are updated and unlocked -- either after
1574 * decompression (and verity if enabled), or after an error.
1576 * In addition, each compressed page holds a reference while it is in a
1577 * bio. These references are necessary prevent compressed pages from
1578 * being freed while they are still in a bio.
1582 bool failed; /* IO error occurred before decompression? */
1583 bool need_verity; /* need fs-verity verification after decompression? */
1584 void *private; /* payload buffer for specified decompression algorithm */
1585 void *private2; /* extra payload buffer */
1586 struct work_struct verity_work; /* work to verify the decompressed pages */
1587 struct work_struct free_work; /* work for late free this structure itself */
1590 #define NULL_CLUSTER ((unsigned int)(~0))
1591 #define MIN_COMPRESS_LOG_SIZE 2
1592 #define MAX_COMPRESS_LOG_SIZE 8
1593 #define MAX_COMPRESS_WINDOW_SIZE(log_size) ((PAGE_SIZE) << (log_size))
1595 struct f2fs_sb_info {
1596 struct super_block *sb; /* pointer to VFS super block */
1597 struct proc_dir_entry *s_proc; /* proc entry */
1598 struct f2fs_super_block *raw_super; /* raw super block pointer */
1599 struct f2fs_rwsem sb_lock; /* lock for raw super block */
1600 int valid_super_block; /* valid super block no */
1601 unsigned long s_flag; /* flags for sbi */
1602 struct mutex writepages; /* mutex for writepages() */
1604 #ifdef CONFIG_BLK_DEV_ZONED
1605 unsigned int blocks_per_blkz; /* F2FS blocks per zone */
1606 unsigned int log_blocks_per_blkz; /* log2 F2FS blocks per zone */
1609 /* for node-related operations */
1610 struct f2fs_nm_info *nm_info; /* node manager */
1611 struct inode *node_inode; /* cache node blocks */
1613 /* for segment-related operations */
1614 struct f2fs_sm_info *sm_info; /* segment manager */
1616 /* for bio operations */
1617 struct f2fs_bio_info *write_io[NR_PAGE_TYPE]; /* for write bios */
1618 /* keep migration IO order for LFS mode */
1619 struct f2fs_rwsem io_order_lock;
1620 mempool_t *write_io_dummy; /* Dummy pages */
1621 pgoff_t page_eio_ofs[NR_PAGE_TYPE]; /* EIO page offset */
1622 int page_eio_cnt[NR_PAGE_TYPE]; /* EIO count */
1624 /* for checkpoint */
1625 struct f2fs_checkpoint *ckpt; /* raw checkpoint pointer */
1626 int cur_cp_pack; /* remain current cp pack */
1627 spinlock_t cp_lock; /* for flag in ckpt */
1628 struct inode *meta_inode; /* cache meta blocks */
1629 struct f2fs_rwsem cp_global_sem; /* checkpoint procedure lock */
1630 struct f2fs_rwsem cp_rwsem; /* blocking FS operations */
1631 struct f2fs_rwsem node_write; /* locking node writes */
1632 struct f2fs_rwsem node_change; /* locking node change */
1633 wait_queue_head_t cp_wait;
1634 unsigned long last_time[MAX_TIME]; /* to store time in jiffies */
1635 long interval_time[MAX_TIME]; /* to store thresholds */
1636 struct ckpt_req_control cprc_info; /* for checkpoint request control */
1638 struct inode_management im[MAX_INO_ENTRY]; /* manage inode cache */
1640 spinlock_t fsync_node_lock; /* for node entry lock */
1641 struct list_head fsync_node_list; /* node list head */
1642 unsigned int fsync_seg_id; /* sequence id */
1643 unsigned int fsync_node_num; /* number of node entries */
1645 /* for orphan inode, use 0'th array */
1646 unsigned int max_orphans; /* max orphan inodes */
1648 /* for inode management */
1649 struct list_head inode_list[NR_INODE_TYPE]; /* dirty inode list */
1650 spinlock_t inode_lock[NR_INODE_TYPE]; /* for dirty inode list lock */
1651 struct mutex flush_lock; /* for flush exclusion */
1653 /* for extent tree cache */
1654 struct extent_tree_info extent_tree[NR_EXTENT_CACHES];
1656 /* basic filesystem units */
1657 unsigned int log_sectors_per_block; /* log2 sectors per block */
1658 unsigned int log_blocksize; /* log2 block size */
1659 unsigned int blocksize; /* block size */
1660 unsigned int root_ino_num; /* root inode number*/
1661 unsigned int node_ino_num; /* node inode number*/
1662 unsigned int meta_ino_num; /* meta inode number*/
1663 unsigned int log_blocks_per_seg; /* log2 blocks per segment */
1664 unsigned int blocks_per_seg; /* blocks per segment */
1665 unsigned int unusable_blocks_per_sec; /* unusable blocks per section */
1666 unsigned int segs_per_sec; /* segments per section */
1667 unsigned int secs_per_zone; /* sections per zone */
1668 unsigned int total_sections; /* total section count */
1669 unsigned int total_node_count; /* total node block count */
1670 unsigned int total_valid_node_count; /* valid node block count */
1671 int dir_level; /* directory level */
1672 int readdir_ra; /* readahead inode in readdir */
1673 u64 max_io_bytes; /* max io bytes to merge IOs */
1675 block_t user_block_count; /* # of user blocks */
1676 block_t total_valid_block_count; /* # of valid blocks */
1677 block_t discard_blks; /* discard command candidats */
1678 block_t last_valid_block_count; /* for recovery */
1679 block_t reserved_blocks; /* configurable reserved blocks */
1680 block_t current_reserved_blocks; /* current reserved blocks */
1682 /* Additional tracking for no checkpoint mode */
1683 block_t unusable_block_count; /* # of blocks saved by last cp */
1685 unsigned int nquota_files; /* # of quota sysfile */
1686 struct f2fs_rwsem quota_sem; /* blocking cp for flags */
1688 /* # of pages, see count_type */
1689 atomic_t nr_pages[NR_COUNT_TYPE];
1690 /* # of allocated blocks */
1691 struct percpu_counter alloc_valid_block_count;
1692 /* # of node block writes as roll forward recovery */
1693 struct percpu_counter rf_node_block_count;
1695 /* writeback control */
1696 atomic_t wb_sync_req[META]; /* count # of WB_SYNC threads */
1698 /* valid inode count */
1699 struct percpu_counter total_valid_inode_count;
1701 struct f2fs_mount_info mount_opt; /* mount options */
1703 /* for cleaning operations */
1704 struct f2fs_rwsem gc_lock; /*
1705 * semaphore for GC, avoid
1706 * race between GC and GC or CP
1708 struct f2fs_gc_kthread *gc_thread; /* GC thread */
1709 struct atgc_management am; /* atgc management */
1710 unsigned int cur_victim_sec; /* current victim section num */
1711 unsigned int gc_mode; /* current GC state */
1712 unsigned int next_victim_seg[2]; /* next segment in victim section */
1713 spinlock_t gc_urgent_high_lock;
1714 unsigned int gc_urgent_high_remaining; /* remaining trial count for GC_URGENT_HIGH */
1716 /* for skip statistic */
1717 unsigned long long skipped_gc_rwsem; /* FG_GC only */
1719 /* threshold for gc trials on pinned files */
1720 u64 gc_pin_file_threshold;
1721 struct f2fs_rwsem pin_sem;
1723 /* maximum # of trials to find a victim segment for SSR and GC */
1724 unsigned int max_victim_search;
1725 /* migration granularity of garbage collection, unit: segment */
1726 unsigned int migration_granularity;
1729 * for stat information.
1730 * one is for the LFS mode, and the other is for the SSR mode.
1732 #ifdef CONFIG_F2FS_STAT_FS
1733 struct f2fs_stat_info *stat_info; /* FS status information */
1734 atomic_t meta_count[META_MAX]; /* # of meta blocks */
1735 unsigned int segment_count[2]; /* # of allocated segments */
1736 unsigned int block_count[2]; /* # of allocated blocks */
1737 atomic_t inplace_count; /* # of inplace update */
1738 /* # of lookup extent cache */
1739 atomic64_t total_hit_ext[NR_EXTENT_CACHES];
1740 /* # of hit rbtree extent node */
1741 atomic64_t read_hit_rbtree[NR_EXTENT_CACHES];
1742 /* # of hit cached extent node */
1743 atomic64_t read_hit_cached[NR_EXTENT_CACHES];
1744 /* # of hit largest extent node in read extent cache */
1745 atomic64_t read_hit_largest;
1746 atomic_t inline_xattr; /* # of inline_xattr inodes */
1747 atomic_t inline_inode; /* # of inline_data inodes */
1748 atomic_t inline_dir; /* # of inline_dentry inodes */
1749 atomic_t compr_inode; /* # of compressed inodes */
1750 atomic64_t compr_blocks; /* # of compressed blocks */
1751 atomic_t swapfile_inode; /* # of swapfile inodes */
1752 atomic_t atomic_files; /* # of opened atomic file */
1753 atomic_t max_aw_cnt; /* max # of atomic writes */
1754 unsigned int io_skip_bggc; /* skip background gc for in-flight IO */
1755 unsigned int other_skip_bggc; /* skip background gc for other reasons */
1756 unsigned int ndirty_inode[NR_INODE_TYPE]; /* # of dirty inodes */
1758 spinlock_t stat_lock; /* lock for stat operations */
1760 /* to attach REQ_META|REQ_FUA flags */
1761 unsigned int data_io_flag;
1762 unsigned int node_io_flag;
1764 /* For sysfs support */
1765 struct kobject s_kobj; /* /sys/fs/f2fs/<devname> */
1766 struct completion s_kobj_unregister;
1768 struct kobject s_stat_kobj; /* /sys/fs/f2fs/<devname>/stat */
1769 struct completion s_stat_kobj_unregister;
1771 struct kobject s_feature_list_kobj; /* /sys/fs/f2fs/<devname>/feature_list */
1772 struct completion s_feature_list_kobj_unregister;
1774 /* For shrinker support */
1775 struct list_head s_list;
1776 struct mutex umount_mutex;
1777 unsigned int shrinker_run_no;
1779 /* For multi devices */
1780 int s_ndevs; /* number of devices */
1781 struct f2fs_dev_info *devs; /* for device list */
1782 unsigned int dirty_device; /* for checkpoint data flush */
1783 spinlock_t dev_lock; /* protect dirty_device */
1784 bool aligned_blksize; /* all devices has the same logical blksize */
1786 /* For write statistics */
1787 u64 sectors_written_start;
1790 /* Reference to checksum algorithm driver via cryptoapi */
1791 struct crypto_shash *s_chksum_driver;
1793 /* Precomputed FS UUID checksum for seeding other checksums */
1794 __u32 s_chksum_seed;
1796 struct workqueue_struct *post_read_wq; /* post read workqueue */
1798 unsigned char errors[MAX_F2FS_ERRORS]; /* error flags */
1799 spinlock_t error_lock; /* protect errors array */
1800 bool error_dirty; /* errors of sb is dirty */
1802 struct kmem_cache *inline_xattr_slab; /* inline xattr entry */
1803 unsigned int inline_xattr_slab_size; /* default inline xattr slab size */
1805 /* For reclaimed segs statistics per each GC mode */
1806 unsigned int gc_segment_mode; /* GC state for reclaimed segments */
1807 unsigned int gc_reclaimed_segs[MAX_GC_MODE]; /* Reclaimed segs for each mode */
1809 unsigned long seq_file_ra_mul; /* multiplier for ra_pages of seq. files in fadvise */
1811 int max_fragment_chunk; /* max chunk size for block fragmentation mode */
1812 int max_fragment_hole; /* max hole size for block fragmentation mode */
1814 /* For atomic write statistics */
1815 atomic64_t current_atomic_write;
1816 s64 peak_atomic_write;
1817 u64 committed_atomic_block;
1818 u64 revoked_atomic_block;
1820 #ifdef CONFIG_F2FS_FS_COMPRESSION
1821 struct kmem_cache *page_array_slab; /* page array entry */
1822 unsigned int page_array_slab_size; /* default page array slab size */
1824 /* For runtime compression statistics */
1825 u64 compr_written_block;
1826 u64 compr_saved_block;
1827 u32 compr_new_inode;
1829 /* For compressed block cache */
1830 struct inode *compress_inode; /* cache compressed blocks */
1831 unsigned int compress_percent; /* cache page percentage */
1832 unsigned int compress_watermark; /* cache page watermark */
1833 atomic_t compress_page_hit; /* cache hit count */
1836 #ifdef CONFIG_F2FS_IOSTAT
1837 /* For app/fs IO statistics */
1838 spinlock_t iostat_lock;
1839 unsigned long long rw_iostat[NR_IO_TYPE];
1840 unsigned long long prev_rw_iostat[NR_IO_TYPE];
1842 unsigned long iostat_next_period;
1843 unsigned int iostat_period_ms;
1845 /* For io latency related statistics info in one iostat period */
1846 spinlock_t iostat_lat_lock;
1847 struct iostat_lat_info *iostat_io_lat;
1851 #ifdef CONFIG_F2FS_FAULT_INJECTION
1852 #define f2fs_show_injection_info(sbi, type) \
1853 printk_ratelimited("%sF2FS-fs (%s) : inject %s in %s of %pS\n", \
1854 KERN_INFO, sbi->sb->s_id, \
1855 f2fs_fault_name[type], \
1856 __func__, __builtin_return_address(0))
1857 static inline bool time_to_inject(struct f2fs_sb_info *sbi, int type)
1859 struct f2fs_fault_info *ffi = &F2FS_OPTION(sbi).fault_info;
1861 if (!ffi->inject_rate)
1864 if (!IS_FAULT_SET(ffi, type))
1867 atomic_inc(&ffi->inject_ops);
1868 if (atomic_read(&ffi->inject_ops) >= ffi->inject_rate) {
1869 atomic_set(&ffi->inject_ops, 0);
1875 #define f2fs_show_injection_info(sbi, type) do { } while (0)
1876 static inline bool time_to_inject(struct f2fs_sb_info *sbi, int type)
1883 * Test if the mounted volume is a multi-device volume.
1884 * - For a single regular disk volume, sbi->s_ndevs is 0.
1885 * - For a single zoned disk volume, sbi->s_ndevs is 1.
1886 * - For a multi-device volume, sbi->s_ndevs is always 2 or more.
1888 static inline bool f2fs_is_multi_device(struct f2fs_sb_info *sbi)
1890 return sbi->s_ndevs > 1;
1893 static inline void f2fs_update_time(struct f2fs_sb_info *sbi, int type)
1895 unsigned long now = jiffies;
1897 sbi->last_time[type] = now;
1899 /* DISCARD_TIME and GC_TIME are based on REQ_TIME */
1900 if (type == REQ_TIME) {
1901 sbi->last_time[DISCARD_TIME] = now;
1902 sbi->last_time[GC_TIME] = now;
1906 static inline bool f2fs_time_over(struct f2fs_sb_info *sbi, int type)
1908 unsigned long interval = sbi->interval_time[type] * HZ;
1910 return time_after(jiffies, sbi->last_time[type] + interval);
1913 static inline unsigned int f2fs_time_to_wait(struct f2fs_sb_info *sbi,
1916 unsigned long interval = sbi->interval_time[type] * HZ;
1917 unsigned int wait_ms = 0;
1920 delta = (sbi->last_time[type] + interval) - jiffies;
1922 wait_ms = jiffies_to_msecs(delta);
1930 static inline u32 __f2fs_crc32(struct f2fs_sb_info *sbi, u32 crc,
1931 const void *address, unsigned int length)
1934 struct shash_desc shash;
1939 BUG_ON(crypto_shash_descsize(sbi->s_chksum_driver) != sizeof(desc.ctx));
1941 desc.shash.tfm = sbi->s_chksum_driver;
1942 *(u32 *)desc.ctx = crc;
1944 err = crypto_shash_update(&desc.shash, address, length);
1947 return *(u32 *)desc.ctx;
1950 static inline u32 f2fs_crc32(struct f2fs_sb_info *sbi, const void *address,
1951 unsigned int length)
1953 return __f2fs_crc32(sbi, F2FS_SUPER_MAGIC, address, length);
1956 static inline bool f2fs_crc_valid(struct f2fs_sb_info *sbi, __u32 blk_crc,
1957 void *buf, size_t buf_size)
1959 return f2fs_crc32(sbi, buf, buf_size) == blk_crc;
1962 static inline u32 f2fs_chksum(struct f2fs_sb_info *sbi, u32 crc,
1963 const void *address, unsigned int length)
1965 return __f2fs_crc32(sbi, crc, address, length);
1968 static inline struct f2fs_inode_info *F2FS_I(struct inode *inode)
1970 return container_of(inode, struct f2fs_inode_info, vfs_inode);
1973 static inline struct f2fs_sb_info *F2FS_SB(struct super_block *sb)
1975 return sb->s_fs_info;
1978 static inline struct f2fs_sb_info *F2FS_I_SB(struct inode *inode)
1980 return F2FS_SB(inode->i_sb);
1983 static inline struct f2fs_sb_info *F2FS_M_SB(struct address_space *mapping)
1985 return F2FS_I_SB(mapping->host);
1988 static inline struct f2fs_sb_info *F2FS_P_SB(struct page *page)
1990 return F2FS_M_SB(page_file_mapping(page));
1993 static inline struct f2fs_super_block *F2FS_RAW_SUPER(struct f2fs_sb_info *sbi)
1995 return (struct f2fs_super_block *)(sbi->raw_super);
1998 static inline struct f2fs_checkpoint *F2FS_CKPT(struct f2fs_sb_info *sbi)
2000 return (struct f2fs_checkpoint *)(sbi->ckpt);
2003 static inline struct f2fs_node *F2FS_NODE(struct page *page)
2005 return (struct f2fs_node *)page_address(page);
2008 static inline struct f2fs_inode *F2FS_INODE(struct page *page)
2010 return &((struct f2fs_node *)page_address(page))->i;
2013 static inline struct f2fs_nm_info *NM_I(struct f2fs_sb_info *sbi)
2015 return (struct f2fs_nm_info *)(sbi->nm_info);
2018 static inline struct f2fs_sm_info *SM_I(struct f2fs_sb_info *sbi)
2020 return (struct f2fs_sm_info *)(sbi->sm_info);
2023 static inline struct sit_info *SIT_I(struct f2fs_sb_info *sbi)
2025 return (struct sit_info *)(SM_I(sbi)->sit_info);
2028 static inline struct free_segmap_info *FREE_I(struct f2fs_sb_info *sbi)
2030 return (struct free_segmap_info *)(SM_I(sbi)->free_info);
2033 static inline struct dirty_seglist_info *DIRTY_I(struct f2fs_sb_info *sbi)
2035 return (struct dirty_seglist_info *)(SM_I(sbi)->dirty_info);
2038 static inline struct address_space *META_MAPPING(struct f2fs_sb_info *sbi)
2040 return sbi->meta_inode->i_mapping;
2043 static inline struct address_space *NODE_MAPPING(struct f2fs_sb_info *sbi)
2045 return sbi->node_inode->i_mapping;
2048 static inline bool is_sbi_flag_set(struct f2fs_sb_info *sbi, unsigned int type)
2050 return test_bit(type, &sbi->s_flag);
2053 static inline void set_sbi_flag(struct f2fs_sb_info *sbi, unsigned int type)
2055 set_bit(type, &sbi->s_flag);
2058 static inline void clear_sbi_flag(struct f2fs_sb_info *sbi, unsigned int type)
2060 clear_bit(type, &sbi->s_flag);
2063 static inline unsigned long long cur_cp_version(struct f2fs_checkpoint *cp)
2065 return le64_to_cpu(cp->checkpoint_ver);
2068 static inline unsigned long f2fs_qf_ino(struct super_block *sb, int type)
2070 if (type < F2FS_MAX_QUOTAS)
2071 return le32_to_cpu(F2FS_SB(sb)->raw_super->qf_ino[type]);
2075 static inline __u64 cur_cp_crc(struct f2fs_checkpoint *cp)
2077 size_t crc_offset = le32_to_cpu(cp->checksum_offset);
2078 return le32_to_cpu(*((__le32 *)((unsigned char *)cp + crc_offset)));
2081 static inline bool __is_set_ckpt_flags(struct f2fs_checkpoint *cp, unsigned int f)
2083 unsigned int ckpt_flags = le32_to_cpu(cp->ckpt_flags);
2085 return ckpt_flags & f;
2088 static inline bool is_set_ckpt_flags(struct f2fs_sb_info *sbi, unsigned int f)
2090 return __is_set_ckpt_flags(F2FS_CKPT(sbi), f);
2093 static inline void __set_ckpt_flags(struct f2fs_checkpoint *cp, unsigned int f)
2095 unsigned int ckpt_flags;
2097 ckpt_flags = le32_to_cpu(cp->ckpt_flags);
2099 cp->ckpt_flags = cpu_to_le32(ckpt_flags);
2102 static inline void set_ckpt_flags(struct f2fs_sb_info *sbi, unsigned int f)
2104 unsigned long flags;
2106 spin_lock_irqsave(&sbi->cp_lock, flags);
2107 __set_ckpt_flags(F2FS_CKPT(sbi), f);
2108 spin_unlock_irqrestore(&sbi->cp_lock, flags);
2111 static inline void __clear_ckpt_flags(struct f2fs_checkpoint *cp, unsigned int f)
2113 unsigned int ckpt_flags;
2115 ckpt_flags = le32_to_cpu(cp->ckpt_flags);
2117 cp->ckpt_flags = cpu_to_le32(ckpt_flags);
2120 static inline void clear_ckpt_flags(struct f2fs_sb_info *sbi, unsigned int f)
2122 unsigned long flags;
2124 spin_lock_irqsave(&sbi->cp_lock, flags);
2125 __clear_ckpt_flags(F2FS_CKPT(sbi), f);
2126 spin_unlock_irqrestore(&sbi->cp_lock, flags);
2129 #define init_f2fs_rwsem(sem) \
2131 static struct lock_class_key __key; \
2133 __init_f2fs_rwsem((sem), #sem, &__key); \
2136 static inline void __init_f2fs_rwsem(struct f2fs_rwsem *sem,
2137 const char *sem_name, struct lock_class_key *key)
2139 __init_rwsem(&sem->internal_rwsem, sem_name, key);
2140 #ifdef CONFIG_F2FS_UNFAIR_RWSEM
2141 init_waitqueue_head(&sem->read_waiters);
2145 static inline int f2fs_rwsem_is_locked(struct f2fs_rwsem *sem)
2147 return rwsem_is_locked(&sem->internal_rwsem);
2150 static inline int f2fs_rwsem_is_contended(struct f2fs_rwsem *sem)
2152 return rwsem_is_contended(&sem->internal_rwsem);
2155 static inline void f2fs_down_read(struct f2fs_rwsem *sem)
2157 #ifdef CONFIG_F2FS_UNFAIR_RWSEM
2158 wait_event(sem->read_waiters, down_read_trylock(&sem->internal_rwsem));
2160 down_read(&sem->internal_rwsem);
2164 static inline int f2fs_down_read_trylock(struct f2fs_rwsem *sem)
2166 return down_read_trylock(&sem->internal_rwsem);
2169 static inline void f2fs_up_read(struct f2fs_rwsem *sem)
2171 up_read(&sem->internal_rwsem);
2174 static inline void f2fs_down_write(struct f2fs_rwsem *sem)
2176 down_write(&sem->internal_rwsem);
2179 #ifdef CONFIG_DEBUG_LOCK_ALLOC
2180 static inline void f2fs_down_read_nested(struct f2fs_rwsem *sem, int subclass)
2182 down_read_nested(&sem->internal_rwsem, subclass);
2185 static inline void f2fs_down_write_nested(struct f2fs_rwsem *sem, int subclass)
2187 down_write_nested(&sem->internal_rwsem, subclass);
2190 #define f2fs_down_read_nested(sem, subclass) f2fs_down_read(sem)
2191 #define f2fs_down_write_nested(sem, subclass) f2fs_down_write(sem)
2194 static inline int f2fs_down_write_trylock(struct f2fs_rwsem *sem)
2196 return down_write_trylock(&sem->internal_rwsem);
2199 static inline void f2fs_up_write(struct f2fs_rwsem *sem)
2201 up_write(&sem->internal_rwsem);
2202 #ifdef CONFIG_F2FS_UNFAIR_RWSEM
2203 wake_up_all(&sem->read_waiters);
2207 static inline void f2fs_lock_op(struct f2fs_sb_info *sbi)
2209 f2fs_down_read(&sbi->cp_rwsem);
2212 static inline int f2fs_trylock_op(struct f2fs_sb_info *sbi)
2214 if (time_to_inject(sbi, FAULT_LOCK_OP)) {
2215 f2fs_show_injection_info(sbi, FAULT_LOCK_OP);
2218 return f2fs_down_read_trylock(&sbi->cp_rwsem);
2221 static inline void f2fs_unlock_op(struct f2fs_sb_info *sbi)
2223 f2fs_up_read(&sbi->cp_rwsem);
2226 static inline void f2fs_lock_all(struct f2fs_sb_info *sbi)
2228 f2fs_down_write(&sbi->cp_rwsem);
2231 static inline void f2fs_unlock_all(struct f2fs_sb_info *sbi)
2233 f2fs_up_write(&sbi->cp_rwsem);
2236 static inline int __get_cp_reason(struct f2fs_sb_info *sbi)
2238 int reason = CP_SYNC;
2240 if (test_opt(sbi, FASTBOOT))
2241 reason = CP_FASTBOOT;
2242 if (is_sbi_flag_set(sbi, SBI_IS_CLOSE))
2247 static inline bool __remain_node_summaries(int reason)
2249 return (reason & (CP_UMOUNT | CP_FASTBOOT));
2252 static inline bool __exist_node_summaries(struct f2fs_sb_info *sbi)
2254 return (is_set_ckpt_flags(sbi, CP_UMOUNT_FLAG) ||
2255 is_set_ckpt_flags(sbi, CP_FASTBOOT_FLAG));
2259 * Check whether the inode has blocks or not
2261 static inline int F2FS_HAS_BLOCKS(struct inode *inode)
2263 block_t xattr_block = F2FS_I(inode)->i_xattr_nid ? 1 : 0;
2265 return (inode->i_blocks >> F2FS_LOG_SECTORS_PER_BLOCK) > xattr_block;
2268 static inline bool f2fs_has_xattr_block(unsigned int ofs)
2270 return ofs == XATTR_NODE_OFFSET;
2273 static inline bool __allow_reserved_blocks(struct f2fs_sb_info *sbi,
2274 struct inode *inode, bool cap)
2278 if (!test_opt(sbi, RESERVE_ROOT))
2280 if (IS_NOQUOTA(inode))
2282 if (uid_eq(F2FS_OPTION(sbi).s_resuid, current_fsuid()))
2284 if (!gid_eq(F2FS_OPTION(sbi).s_resgid, GLOBAL_ROOT_GID) &&
2285 in_group_p(F2FS_OPTION(sbi).s_resgid))
2287 if (cap && capable(CAP_SYS_RESOURCE))
2292 static inline void f2fs_i_blocks_write(struct inode *, block_t, bool, bool);
2293 static inline int inc_valid_block_count(struct f2fs_sb_info *sbi,
2294 struct inode *inode, blkcnt_t *count, bool partial)
2296 blkcnt_t diff = 0, release = 0;
2297 block_t avail_user_block_count;
2300 ret = dquot_reserve_block(inode, *count);
2304 if (time_to_inject(sbi, FAULT_BLOCK)) {
2305 f2fs_show_injection_info(sbi, FAULT_BLOCK);
2311 * let's increase this in prior to actual block count change in order
2312 * for f2fs_sync_file to avoid data races when deciding checkpoint.
2314 percpu_counter_add(&sbi->alloc_valid_block_count, (*count));
2316 spin_lock(&sbi->stat_lock);
2317 sbi->total_valid_block_count += (block_t)(*count);
2318 avail_user_block_count = sbi->user_block_count -
2319 sbi->current_reserved_blocks;
2321 if (!__allow_reserved_blocks(sbi, inode, true))
2322 avail_user_block_count -= F2FS_OPTION(sbi).root_reserved_blocks;
2324 if (F2FS_IO_ALIGNED(sbi))
2325 avail_user_block_count -= sbi->blocks_per_seg *
2326 SM_I(sbi)->additional_reserved_segments;
2328 if (unlikely(is_sbi_flag_set(sbi, SBI_CP_DISABLED))) {
2329 if (avail_user_block_count > sbi->unusable_block_count)
2330 avail_user_block_count -= sbi->unusable_block_count;
2332 avail_user_block_count = 0;
2334 if (unlikely(sbi->total_valid_block_count > avail_user_block_count)) {
2336 spin_unlock(&sbi->stat_lock);
2340 diff = sbi->total_valid_block_count - avail_user_block_count;
2345 sbi->total_valid_block_count -= diff;
2347 spin_unlock(&sbi->stat_lock);
2351 spin_unlock(&sbi->stat_lock);
2353 if (unlikely(release)) {
2354 percpu_counter_sub(&sbi->alloc_valid_block_count, release);
2355 dquot_release_reservation_block(inode, release);
2357 f2fs_i_blocks_write(inode, *count, true, true);
2361 percpu_counter_sub(&sbi->alloc_valid_block_count, release);
2363 dquot_release_reservation_block(inode, release);
2368 void f2fs_printk(struct f2fs_sb_info *sbi, const char *fmt, ...);
2370 #define f2fs_err(sbi, fmt, ...) \
2371 f2fs_printk(sbi, KERN_ERR fmt, ##__VA_ARGS__)
2372 #define f2fs_warn(sbi, fmt, ...) \
2373 f2fs_printk(sbi, KERN_WARNING fmt, ##__VA_ARGS__)
2374 #define f2fs_notice(sbi, fmt, ...) \
2375 f2fs_printk(sbi, KERN_NOTICE fmt, ##__VA_ARGS__)
2376 #define f2fs_info(sbi, fmt, ...) \
2377 f2fs_printk(sbi, KERN_INFO fmt, ##__VA_ARGS__)
2378 #define f2fs_debug(sbi, fmt, ...) \
2379 f2fs_printk(sbi, KERN_DEBUG fmt, ##__VA_ARGS__)
2381 static inline void dec_valid_block_count(struct f2fs_sb_info *sbi,
2382 struct inode *inode,
2385 blkcnt_t sectors = count << F2FS_LOG_SECTORS_PER_BLOCK;
2387 spin_lock(&sbi->stat_lock);
2388 f2fs_bug_on(sbi, sbi->total_valid_block_count < (block_t) count);
2389 sbi->total_valid_block_count -= (block_t)count;
2390 if (sbi->reserved_blocks &&
2391 sbi->current_reserved_blocks < sbi->reserved_blocks)
2392 sbi->current_reserved_blocks = min(sbi->reserved_blocks,
2393 sbi->current_reserved_blocks + count);
2394 spin_unlock(&sbi->stat_lock);
2395 if (unlikely(inode->i_blocks < sectors)) {
2396 f2fs_warn(sbi, "Inconsistent i_blocks, ino:%lu, iblocks:%llu, sectors:%llu",
2398 (unsigned long long)inode->i_blocks,
2399 (unsigned long long)sectors);
2400 set_sbi_flag(sbi, SBI_NEED_FSCK);
2403 f2fs_i_blocks_write(inode, count, false, true);
2406 static inline void inc_page_count(struct f2fs_sb_info *sbi, int count_type)
2408 atomic_inc(&sbi->nr_pages[count_type]);
2410 if (count_type == F2FS_DIRTY_DENTS ||
2411 count_type == F2FS_DIRTY_NODES ||
2412 count_type == F2FS_DIRTY_META ||
2413 count_type == F2FS_DIRTY_QDATA ||
2414 count_type == F2FS_DIRTY_IMETA)
2415 set_sbi_flag(sbi, SBI_IS_DIRTY);
2418 static inline void inode_inc_dirty_pages(struct inode *inode)
2420 atomic_inc(&F2FS_I(inode)->dirty_pages);
2421 inc_page_count(F2FS_I_SB(inode), S_ISDIR(inode->i_mode) ?
2422 F2FS_DIRTY_DENTS : F2FS_DIRTY_DATA);
2423 if (IS_NOQUOTA(inode))
2424 inc_page_count(F2FS_I_SB(inode), F2FS_DIRTY_QDATA);
2427 static inline void dec_page_count(struct f2fs_sb_info *sbi, int count_type)
2429 atomic_dec(&sbi->nr_pages[count_type]);
2432 static inline void inode_dec_dirty_pages(struct inode *inode)
2434 if (!S_ISDIR(inode->i_mode) && !S_ISREG(inode->i_mode) &&
2435 !S_ISLNK(inode->i_mode))
2438 atomic_dec(&F2FS_I(inode)->dirty_pages);
2439 dec_page_count(F2FS_I_SB(inode), S_ISDIR(inode->i_mode) ?
2440 F2FS_DIRTY_DENTS : F2FS_DIRTY_DATA);
2441 if (IS_NOQUOTA(inode))
2442 dec_page_count(F2FS_I_SB(inode), F2FS_DIRTY_QDATA);
2445 static inline void inc_atomic_write_cnt(struct inode *inode)
2447 struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
2448 struct f2fs_inode_info *fi = F2FS_I(inode);
2451 fi->atomic_write_cnt++;
2452 atomic64_inc(&sbi->current_atomic_write);
2453 current_write = atomic64_read(&sbi->current_atomic_write);
2454 if (current_write > sbi->peak_atomic_write)
2455 sbi->peak_atomic_write = current_write;
2458 static inline void release_atomic_write_cnt(struct inode *inode)
2460 struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
2461 struct f2fs_inode_info *fi = F2FS_I(inode);
2463 atomic64_sub(fi->atomic_write_cnt, &sbi->current_atomic_write);
2464 fi->atomic_write_cnt = 0;
2467 static inline s64 get_pages(struct f2fs_sb_info *sbi, int count_type)
2469 return atomic_read(&sbi->nr_pages[count_type]);
2472 static inline int get_dirty_pages(struct inode *inode)
2474 return atomic_read(&F2FS_I(inode)->dirty_pages);
2477 static inline int get_blocktype_secs(struct f2fs_sb_info *sbi, int block_type)
2479 unsigned int pages_per_sec = sbi->segs_per_sec * sbi->blocks_per_seg;
2480 unsigned int segs = (get_pages(sbi, block_type) + pages_per_sec - 1) >>
2481 sbi->log_blocks_per_seg;
2483 return segs / sbi->segs_per_sec;
2486 static inline block_t valid_user_blocks(struct f2fs_sb_info *sbi)
2488 return sbi->total_valid_block_count;
2491 static inline block_t discard_blocks(struct f2fs_sb_info *sbi)
2493 return sbi->discard_blks;
2496 static inline unsigned long __bitmap_size(struct f2fs_sb_info *sbi, int flag)
2498 struct f2fs_checkpoint *ckpt = F2FS_CKPT(sbi);
2500 /* return NAT or SIT bitmap */
2501 if (flag == NAT_BITMAP)
2502 return le32_to_cpu(ckpt->nat_ver_bitmap_bytesize);
2503 else if (flag == SIT_BITMAP)
2504 return le32_to_cpu(ckpt->sit_ver_bitmap_bytesize);
2509 static inline block_t __cp_payload(struct f2fs_sb_info *sbi)
2511 return le32_to_cpu(F2FS_RAW_SUPER(sbi)->cp_payload);
2514 static inline void *__bitmap_ptr(struct f2fs_sb_info *sbi, int flag)
2516 struct f2fs_checkpoint *ckpt = F2FS_CKPT(sbi);
2517 void *tmp_ptr = &ckpt->sit_nat_version_bitmap;
2520 if (is_set_ckpt_flags(sbi, CP_LARGE_NAT_BITMAP_FLAG)) {
2521 offset = (flag == SIT_BITMAP) ?
2522 le32_to_cpu(ckpt->nat_ver_bitmap_bytesize) : 0;
2524 * if large_nat_bitmap feature is enabled, leave checksum
2525 * protection for all nat/sit bitmaps.
2527 return tmp_ptr + offset + sizeof(__le32);
2530 if (__cp_payload(sbi) > 0) {
2531 if (flag == NAT_BITMAP)
2534 return (unsigned char *)ckpt + F2FS_BLKSIZE;
2536 offset = (flag == NAT_BITMAP) ?
2537 le32_to_cpu(ckpt->sit_ver_bitmap_bytesize) : 0;
2538 return tmp_ptr + offset;
2542 static inline block_t __start_cp_addr(struct f2fs_sb_info *sbi)
2544 block_t start_addr = le32_to_cpu(F2FS_RAW_SUPER(sbi)->cp_blkaddr);
2546 if (sbi->cur_cp_pack == 2)
2547 start_addr += sbi->blocks_per_seg;
2551 static inline block_t __start_cp_next_addr(struct f2fs_sb_info *sbi)
2553 block_t start_addr = le32_to_cpu(F2FS_RAW_SUPER(sbi)->cp_blkaddr);
2555 if (sbi->cur_cp_pack == 1)
2556 start_addr += sbi->blocks_per_seg;
2560 static inline void __set_cp_next_pack(struct f2fs_sb_info *sbi)
2562 sbi->cur_cp_pack = (sbi->cur_cp_pack == 1) ? 2 : 1;
2565 static inline block_t __start_sum_addr(struct f2fs_sb_info *sbi)
2567 return le32_to_cpu(F2FS_CKPT(sbi)->cp_pack_start_sum);
2570 extern void f2fs_mark_inode_dirty_sync(struct inode *inode, bool sync);
2571 static inline int inc_valid_node_count(struct f2fs_sb_info *sbi,
2572 struct inode *inode, bool is_inode)
2574 block_t valid_block_count;
2575 unsigned int valid_node_count, user_block_count;
2580 err = dquot_alloc_inode(inode);
2585 err = dquot_reserve_block(inode, 1);
2590 if (time_to_inject(sbi, FAULT_BLOCK)) {
2591 f2fs_show_injection_info(sbi, FAULT_BLOCK);
2595 spin_lock(&sbi->stat_lock);
2597 valid_block_count = sbi->total_valid_block_count +
2598 sbi->current_reserved_blocks + 1;
2600 if (!__allow_reserved_blocks(sbi, inode, false))
2601 valid_block_count += F2FS_OPTION(sbi).root_reserved_blocks;
2603 if (F2FS_IO_ALIGNED(sbi))
2604 valid_block_count += sbi->blocks_per_seg *
2605 SM_I(sbi)->additional_reserved_segments;
2607 user_block_count = sbi->user_block_count;
2608 if (unlikely(is_sbi_flag_set(sbi, SBI_CP_DISABLED)))
2609 user_block_count -= sbi->unusable_block_count;
2611 if (unlikely(valid_block_count > user_block_count)) {
2612 spin_unlock(&sbi->stat_lock);
2616 valid_node_count = sbi->total_valid_node_count + 1;
2617 if (unlikely(valid_node_count > sbi->total_node_count)) {
2618 spin_unlock(&sbi->stat_lock);
2622 sbi->total_valid_node_count++;
2623 sbi->total_valid_block_count++;
2624 spin_unlock(&sbi->stat_lock);
2628 f2fs_mark_inode_dirty_sync(inode, true);
2630 f2fs_i_blocks_write(inode, 1, true, true);
2633 percpu_counter_inc(&sbi->alloc_valid_block_count);
2639 dquot_free_inode(inode);
2641 dquot_release_reservation_block(inode, 1);
2646 static inline void dec_valid_node_count(struct f2fs_sb_info *sbi,
2647 struct inode *inode, bool is_inode)
2649 spin_lock(&sbi->stat_lock);
2651 if (unlikely(!sbi->total_valid_block_count ||
2652 !sbi->total_valid_node_count)) {
2653 f2fs_warn(sbi, "dec_valid_node_count: inconsistent block counts, total_valid_block:%u, total_valid_node:%u",
2654 sbi->total_valid_block_count,
2655 sbi->total_valid_node_count);
2656 set_sbi_flag(sbi, SBI_NEED_FSCK);
2658 sbi->total_valid_block_count--;
2659 sbi->total_valid_node_count--;
2662 if (sbi->reserved_blocks &&
2663 sbi->current_reserved_blocks < sbi->reserved_blocks)
2664 sbi->current_reserved_blocks++;
2666 spin_unlock(&sbi->stat_lock);
2669 dquot_free_inode(inode);
2671 if (unlikely(inode->i_blocks == 0)) {
2672 f2fs_warn(sbi, "dec_valid_node_count: inconsistent i_blocks, ino:%lu, iblocks:%llu",
2674 (unsigned long long)inode->i_blocks);
2675 set_sbi_flag(sbi, SBI_NEED_FSCK);
2678 f2fs_i_blocks_write(inode, 1, false, true);
2682 static inline unsigned int valid_node_count(struct f2fs_sb_info *sbi)
2684 return sbi->total_valid_node_count;
2687 static inline void inc_valid_inode_count(struct f2fs_sb_info *sbi)
2689 percpu_counter_inc(&sbi->total_valid_inode_count);
2692 static inline void dec_valid_inode_count(struct f2fs_sb_info *sbi)
2694 percpu_counter_dec(&sbi->total_valid_inode_count);
2697 static inline s64 valid_inode_count(struct f2fs_sb_info *sbi)
2699 return percpu_counter_sum_positive(&sbi->total_valid_inode_count);
2702 static inline struct page *f2fs_grab_cache_page(struct address_space *mapping,
2703 pgoff_t index, bool for_write)
2708 if (IS_ENABLED(CONFIG_F2FS_FAULT_INJECTION)) {
2710 page = find_get_page_flags(mapping, index,
2711 FGP_LOCK | FGP_ACCESSED);
2713 page = find_lock_page(mapping, index);
2717 if (time_to_inject(F2FS_M_SB(mapping), FAULT_PAGE_ALLOC)) {
2718 f2fs_show_injection_info(F2FS_M_SB(mapping),
2725 return grab_cache_page(mapping, index);
2727 flags = memalloc_nofs_save();
2728 page = grab_cache_page_write_begin(mapping, index);
2729 memalloc_nofs_restore(flags);
2734 static inline struct page *f2fs_pagecache_get_page(
2735 struct address_space *mapping, pgoff_t index,
2736 int fgp_flags, gfp_t gfp_mask)
2738 if (time_to_inject(F2FS_M_SB(mapping), FAULT_PAGE_GET)) {
2739 f2fs_show_injection_info(F2FS_M_SB(mapping), FAULT_PAGE_GET);
2743 return pagecache_get_page(mapping, index, fgp_flags, gfp_mask);
2746 static inline void f2fs_put_page(struct page *page, int unlock)
2752 f2fs_bug_on(F2FS_P_SB(page), !PageLocked(page));
2758 static inline void f2fs_put_dnode(struct dnode_of_data *dn)
2761 f2fs_put_page(dn->node_page, 1);
2762 if (dn->inode_page && dn->node_page != dn->inode_page)
2763 f2fs_put_page(dn->inode_page, 0);
2764 dn->node_page = NULL;
2765 dn->inode_page = NULL;
2768 static inline struct kmem_cache *f2fs_kmem_cache_create(const char *name,
2771 return kmem_cache_create(name, size, 0, SLAB_RECLAIM_ACCOUNT, NULL);
2774 static inline void *f2fs_kmem_cache_alloc_nofail(struct kmem_cache *cachep,
2779 entry = kmem_cache_alloc(cachep, flags);
2781 entry = kmem_cache_alloc(cachep, flags | __GFP_NOFAIL);
2785 static inline void *f2fs_kmem_cache_alloc(struct kmem_cache *cachep,
2786 gfp_t flags, bool nofail, struct f2fs_sb_info *sbi)
2789 return f2fs_kmem_cache_alloc_nofail(cachep, flags);
2791 if (time_to_inject(sbi, FAULT_SLAB_ALLOC)) {
2792 f2fs_show_injection_info(sbi, FAULT_SLAB_ALLOC);
2796 return kmem_cache_alloc(cachep, flags);
2799 static inline bool is_inflight_io(struct f2fs_sb_info *sbi, int type)
2801 if (get_pages(sbi, F2FS_RD_DATA) || get_pages(sbi, F2FS_RD_NODE) ||
2802 get_pages(sbi, F2FS_RD_META) || get_pages(sbi, F2FS_WB_DATA) ||
2803 get_pages(sbi, F2FS_WB_CP_DATA) ||
2804 get_pages(sbi, F2FS_DIO_READ) ||
2805 get_pages(sbi, F2FS_DIO_WRITE))
2808 if (type != DISCARD_TIME && SM_I(sbi) && SM_I(sbi)->dcc_info &&
2809 atomic_read(&SM_I(sbi)->dcc_info->queued_discard))
2812 if (SM_I(sbi) && SM_I(sbi)->fcc_info &&
2813 atomic_read(&SM_I(sbi)->fcc_info->queued_flush))
2818 static inline bool is_idle(struct f2fs_sb_info *sbi, int type)
2820 if (sbi->gc_mode == GC_URGENT_HIGH)
2823 if (is_inflight_io(sbi, type))
2826 if (sbi->gc_mode == GC_URGENT_MID)
2829 if (sbi->gc_mode == GC_URGENT_LOW &&
2830 (type == DISCARD_TIME || type == GC_TIME))
2833 return f2fs_time_over(sbi, type);
2836 static inline void f2fs_radix_tree_insert(struct radix_tree_root *root,
2837 unsigned long index, void *item)
2839 while (radix_tree_insert(root, index, item))
2843 #define RAW_IS_INODE(p) ((p)->footer.nid == (p)->footer.ino)
2845 static inline bool IS_INODE(struct page *page)
2847 struct f2fs_node *p = F2FS_NODE(page);
2849 return RAW_IS_INODE(p);
2852 static inline int offset_in_addr(struct f2fs_inode *i)
2854 return (i->i_inline & F2FS_EXTRA_ATTR) ?
2855 (le16_to_cpu(i->i_extra_isize) / sizeof(__le32)) : 0;
2858 static inline __le32 *blkaddr_in_node(struct f2fs_node *node)
2860 return RAW_IS_INODE(node) ? node->i.i_addr : node->dn.addr;
2863 static inline int f2fs_has_extra_attr(struct inode *inode);
2864 static inline block_t data_blkaddr(struct inode *inode,
2865 struct page *node_page, unsigned int offset)
2867 struct f2fs_node *raw_node;
2870 bool is_inode = IS_INODE(node_page);
2872 raw_node = F2FS_NODE(node_page);
2876 /* from GC path only */
2877 base = offset_in_addr(&raw_node->i);
2878 else if (f2fs_has_extra_attr(inode))
2879 base = get_extra_isize(inode);
2882 addr_array = blkaddr_in_node(raw_node);
2883 return le32_to_cpu(addr_array[base + offset]);
2886 static inline block_t f2fs_data_blkaddr(struct dnode_of_data *dn)
2888 return data_blkaddr(dn->inode, dn->node_page, dn->ofs_in_node);
2891 static inline int f2fs_test_bit(unsigned int nr, char *addr)
2896 mask = BIT(7 - (nr & 0x07));
2897 return mask & *addr;
2900 static inline void f2fs_set_bit(unsigned int nr, char *addr)
2905 mask = BIT(7 - (nr & 0x07));
2909 static inline void f2fs_clear_bit(unsigned int nr, char *addr)
2914 mask = BIT(7 - (nr & 0x07));
2918 static inline int f2fs_test_and_set_bit(unsigned int nr, char *addr)
2924 mask = BIT(7 - (nr & 0x07));
2930 static inline int f2fs_test_and_clear_bit(unsigned int nr, char *addr)
2936 mask = BIT(7 - (nr & 0x07));
2942 static inline void f2fs_change_bit(unsigned int nr, char *addr)
2947 mask = BIT(7 - (nr & 0x07));
2952 * On-disk inode flags (f2fs_inode::i_flags)
2954 #define F2FS_COMPR_FL 0x00000004 /* Compress file */
2955 #define F2FS_SYNC_FL 0x00000008 /* Synchronous updates */
2956 #define F2FS_IMMUTABLE_FL 0x00000010 /* Immutable file */
2957 #define F2FS_APPEND_FL 0x00000020 /* writes to file may only append */
2958 #define F2FS_NODUMP_FL 0x00000040 /* do not dump file */
2959 #define F2FS_NOATIME_FL 0x00000080 /* do not update atime */
2960 #define F2FS_NOCOMP_FL 0x00000400 /* Don't compress */
2961 #define F2FS_INDEX_FL 0x00001000 /* hash-indexed directory */
2962 #define F2FS_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */
2963 #define F2FS_PROJINHERIT_FL 0x20000000 /* Create with parents projid */
2964 #define F2FS_CASEFOLD_FL 0x40000000 /* Casefolded file */
2966 /* Flags that should be inherited by new inodes from their parent. */
2967 #define F2FS_FL_INHERITED (F2FS_SYNC_FL | F2FS_NODUMP_FL | F2FS_NOATIME_FL | \
2968 F2FS_DIRSYNC_FL | F2FS_PROJINHERIT_FL | \
2971 /* Flags that are appropriate for regular files (all but dir-specific ones). */
2972 #define F2FS_REG_FLMASK (~(F2FS_DIRSYNC_FL | F2FS_PROJINHERIT_FL | \
2975 /* Flags that are appropriate for non-directories/regular files. */
2976 #define F2FS_OTHER_FLMASK (F2FS_NODUMP_FL | F2FS_NOATIME_FL)
2978 static inline __u32 f2fs_mask_flags(umode_t mode, __u32 flags)
2982 else if (S_ISREG(mode))
2983 return flags & F2FS_REG_FLMASK;
2985 return flags & F2FS_OTHER_FLMASK;
2988 static inline void __mark_inode_dirty_flag(struct inode *inode,
2992 case FI_INLINE_XATTR:
2993 case FI_INLINE_DATA:
2994 case FI_INLINE_DENTRY:
3000 case FI_INLINE_DOTS:
3002 case FI_COMPRESS_RELEASED:
3003 case FI_ATOMIC_COMMITTED:
3004 f2fs_mark_inode_dirty_sync(inode, true);
3008 static inline void set_inode_flag(struct inode *inode, int flag)
3010 set_bit(flag, F2FS_I(inode)->flags);
3011 __mark_inode_dirty_flag(inode, flag, true);
3014 static inline int is_inode_flag_set(struct inode *inode, int flag)
3016 return test_bit(flag, F2FS_I(inode)->flags);
3019 static inline void clear_inode_flag(struct inode *inode, int flag)
3021 clear_bit(flag, F2FS_I(inode)->flags);
3022 __mark_inode_dirty_flag(inode, flag, false);
3025 static inline bool f2fs_verity_in_progress(struct inode *inode)
3027 return IS_ENABLED(CONFIG_FS_VERITY) &&
3028 is_inode_flag_set(inode, FI_VERITY_IN_PROGRESS);
3031 static inline void set_acl_inode(struct inode *inode, umode_t mode)
3033 F2FS_I(inode)->i_acl_mode = mode;
3034 set_inode_flag(inode, FI_ACL_MODE);
3035 f2fs_mark_inode_dirty_sync(inode, false);
3038 static inline void f2fs_i_links_write(struct inode *inode, bool inc)
3044 f2fs_mark_inode_dirty_sync(inode, true);
3047 static inline void f2fs_i_blocks_write(struct inode *inode,
3048 block_t diff, bool add, bool claim)
3050 bool clean = !is_inode_flag_set(inode, FI_DIRTY_INODE);
3051 bool recover = is_inode_flag_set(inode, FI_AUTO_RECOVER);
3053 /* add = 1, claim = 1 should be dquot_reserve_block in pair */
3056 dquot_claim_block(inode, diff);
3058 dquot_alloc_block_nofail(inode, diff);
3060 dquot_free_block(inode, diff);
3063 f2fs_mark_inode_dirty_sync(inode, true);
3064 if (clean || recover)
3065 set_inode_flag(inode, FI_AUTO_RECOVER);
3068 static inline bool f2fs_is_atomic_file(struct inode *inode);
3070 static inline void f2fs_i_size_write(struct inode *inode, loff_t i_size)
3072 bool clean = !is_inode_flag_set(inode, FI_DIRTY_INODE);
3073 bool recover = is_inode_flag_set(inode, FI_AUTO_RECOVER);
3075 if (i_size_read(inode) == i_size)
3078 i_size_write(inode, i_size);
3080 if (f2fs_is_atomic_file(inode))
3083 f2fs_mark_inode_dirty_sync(inode, true);
3084 if (clean || recover)
3085 set_inode_flag(inode, FI_AUTO_RECOVER);
3088 static inline void f2fs_i_depth_write(struct inode *inode, unsigned int depth)
3090 F2FS_I(inode)->i_current_depth = depth;
3091 f2fs_mark_inode_dirty_sync(inode, true);
3094 static inline void f2fs_i_gc_failures_write(struct inode *inode,
3097 F2FS_I(inode)->i_gc_failures[GC_FAILURE_PIN] = count;
3098 f2fs_mark_inode_dirty_sync(inode, true);
3101 static inline void f2fs_i_xnid_write(struct inode *inode, nid_t xnid)
3103 F2FS_I(inode)->i_xattr_nid = xnid;
3104 f2fs_mark_inode_dirty_sync(inode, true);
3107 static inline void f2fs_i_pino_write(struct inode *inode, nid_t pino)
3109 F2FS_I(inode)->i_pino = pino;
3110 f2fs_mark_inode_dirty_sync(inode, true);
3113 static inline void get_inline_info(struct inode *inode, struct f2fs_inode *ri)
3115 struct f2fs_inode_info *fi = F2FS_I(inode);
3117 if (ri->i_inline & F2FS_INLINE_XATTR)
3118 set_bit(FI_INLINE_XATTR, fi->flags);
3119 if (ri->i_inline & F2FS_INLINE_DATA)
3120 set_bit(FI_INLINE_DATA, fi->flags);
3121 if (ri->i_inline & F2FS_INLINE_DENTRY)
3122 set_bit(FI_INLINE_DENTRY, fi->flags);
3123 if (ri->i_inline & F2FS_DATA_EXIST)
3124 set_bit(FI_DATA_EXIST, fi->flags);
3125 if (ri->i_inline & F2FS_INLINE_DOTS)
3126 set_bit(FI_INLINE_DOTS, fi->flags);
3127 if (ri->i_inline & F2FS_EXTRA_ATTR)
3128 set_bit(FI_EXTRA_ATTR, fi->flags);
3129 if (ri->i_inline & F2FS_PIN_FILE)
3130 set_bit(FI_PIN_FILE, fi->flags);
3131 if (ri->i_inline & F2FS_COMPRESS_RELEASED)
3132 set_bit(FI_COMPRESS_RELEASED, fi->flags);
3135 static inline void set_raw_inline(struct inode *inode, struct f2fs_inode *ri)
3139 if (is_inode_flag_set(inode, FI_INLINE_XATTR))
3140 ri->i_inline |= F2FS_INLINE_XATTR;
3141 if (is_inode_flag_set(inode, FI_INLINE_DATA))
3142 ri->i_inline |= F2FS_INLINE_DATA;
3143 if (is_inode_flag_set(inode, FI_INLINE_DENTRY))
3144 ri->i_inline |= F2FS_INLINE_DENTRY;
3145 if (is_inode_flag_set(inode, FI_DATA_EXIST))
3146 ri->i_inline |= F2FS_DATA_EXIST;
3147 if (is_inode_flag_set(inode, FI_INLINE_DOTS))
3148 ri->i_inline |= F2FS_INLINE_DOTS;
3149 if (is_inode_flag_set(inode, FI_EXTRA_ATTR))
3150 ri->i_inline |= F2FS_EXTRA_ATTR;
3151 if (is_inode_flag_set(inode, FI_PIN_FILE))
3152 ri->i_inline |= F2FS_PIN_FILE;
3153 if (is_inode_flag_set(inode, FI_COMPRESS_RELEASED))
3154 ri->i_inline |= F2FS_COMPRESS_RELEASED;
3157 static inline int f2fs_has_extra_attr(struct inode *inode)
3159 return is_inode_flag_set(inode, FI_EXTRA_ATTR);
3162 static inline int f2fs_has_inline_xattr(struct inode *inode)
3164 return is_inode_flag_set(inode, FI_INLINE_XATTR);
3167 static inline int f2fs_compressed_file(struct inode *inode)
3169 return S_ISREG(inode->i_mode) &&
3170 is_inode_flag_set(inode, FI_COMPRESSED_FILE);
3173 static inline bool f2fs_need_compress_data(struct inode *inode)
3175 int compress_mode = F2FS_OPTION(F2FS_I_SB(inode)).compress_mode;
3177 if (!f2fs_compressed_file(inode))
3180 if (compress_mode == COMPR_MODE_FS)
3182 else if (compress_mode == COMPR_MODE_USER &&
3183 is_inode_flag_set(inode, FI_ENABLE_COMPRESS))
3189 static inline unsigned int addrs_per_inode(struct inode *inode)
3191 unsigned int addrs = CUR_ADDRS_PER_INODE(inode) -
3192 get_inline_xattr_addrs(inode);
3194 if (!f2fs_compressed_file(inode))
3196 return ALIGN_DOWN(addrs, F2FS_I(inode)->i_cluster_size);
3199 static inline unsigned int addrs_per_block(struct inode *inode)
3201 if (!f2fs_compressed_file(inode))
3202 return DEF_ADDRS_PER_BLOCK;
3203 return ALIGN_DOWN(DEF_ADDRS_PER_BLOCK, F2FS_I(inode)->i_cluster_size);
3206 static inline void *inline_xattr_addr(struct inode *inode, struct page *page)
3208 struct f2fs_inode *ri = F2FS_INODE(page);
3210 return (void *)&(ri->i_addr[DEF_ADDRS_PER_INODE -
3211 get_inline_xattr_addrs(inode)]);
3214 static inline int inline_xattr_size(struct inode *inode)
3216 if (f2fs_has_inline_xattr(inode))
3217 return get_inline_xattr_addrs(inode) * sizeof(__le32);
3222 * Notice: check inline_data flag without inode page lock is unsafe.
3223 * It could change at any time by f2fs_convert_inline_page().
3225 static inline int f2fs_has_inline_data(struct inode *inode)
3227 return is_inode_flag_set(inode, FI_INLINE_DATA);
3230 static inline int f2fs_exist_data(struct inode *inode)
3232 return is_inode_flag_set(inode, FI_DATA_EXIST);
3235 static inline int f2fs_has_inline_dots(struct inode *inode)
3237 return is_inode_flag_set(inode, FI_INLINE_DOTS);
3240 static inline int f2fs_is_mmap_file(struct inode *inode)
3242 return is_inode_flag_set(inode, FI_MMAP_FILE);
3245 static inline bool f2fs_is_pinned_file(struct inode *inode)
3247 return is_inode_flag_set(inode, FI_PIN_FILE);
3250 static inline bool f2fs_is_atomic_file(struct inode *inode)
3252 return is_inode_flag_set(inode, FI_ATOMIC_FILE);
3255 static inline bool f2fs_is_cow_file(struct inode *inode)
3257 return is_inode_flag_set(inode, FI_COW_FILE);
3260 static inline __le32 *get_dnode_addr(struct inode *inode,
3261 struct page *node_page);
3262 static inline void *inline_data_addr(struct inode *inode, struct page *page)
3264 __le32 *addr = get_dnode_addr(inode, page);
3266 return (void *)(addr + DEF_INLINE_RESERVED_SIZE);
3269 static inline int f2fs_has_inline_dentry(struct inode *inode)
3271 return is_inode_flag_set(inode, FI_INLINE_DENTRY);
3274 static inline int is_file(struct inode *inode, int type)
3276 return F2FS_I(inode)->i_advise & type;
3279 static inline void set_file(struct inode *inode, int type)
3281 if (is_file(inode, type))
3283 F2FS_I(inode)->i_advise |= type;
3284 f2fs_mark_inode_dirty_sync(inode, true);
3287 static inline void clear_file(struct inode *inode, int type)
3289 if (!is_file(inode, type))
3291 F2FS_I(inode)->i_advise &= ~type;
3292 f2fs_mark_inode_dirty_sync(inode, true);
3295 static inline bool f2fs_is_time_consistent(struct inode *inode)
3297 if (!timespec64_equal(F2FS_I(inode)->i_disk_time, &inode->i_atime))
3299 if (!timespec64_equal(F2FS_I(inode)->i_disk_time + 1, &inode->i_ctime))
3301 if (!timespec64_equal(F2FS_I(inode)->i_disk_time + 2, &inode->i_mtime))
3303 if (!timespec64_equal(F2FS_I(inode)->i_disk_time + 3,
3304 &F2FS_I(inode)->i_crtime))
3309 static inline bool f2fs_skip_inode_update(struct inode *inode, int dsync)
3314 struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
3316 spin_lock(&sbi->inode_lock[DIRTY_META]);
3317 ret = list_empty(&F2FS_I(inode)->gdirty_list);
3318 spin_unlock(&sbi->inode_lock[DIRTY_META]);
3321 if (!is_inode_flag_set(inode, FI_AUTO_RECOVER) ||
3322 file_keep_isize(inode) ||
3323 i_size_read(inode) & ~PAGE_MASK)
3326 if (!f2fs_is_time_consistent(inode))
3329 spin_lock(&F2FS_I(inode)->i_size_lock);
3330 ret = F2FS_I(inode)->last_disk_size == i_size_read(inode);
3331 spin_unlock(&F2FS_I(inode)->i_size_lock);
3336 static inline bool f2fs_readonly(struct super_block *sb)
3338 return sb_rdonly(sb);
3341 static inline bool f2fs_cp_error(struct f2fs_sb_info *sbi)
3343 return is_set_ckpt_flags(sbi, CP_ERROR_FLAG);
3346 static inline bool is_dot_dotdot(const u8 *name, size_t len)
3348 if (len == 1 && name[0] == '.')
3351 if (len == 2 && name[0] == '.' && name[1] == '.')
3357 static inline void *f2fs_kmalloc(struct f2fs_sb_info *sbi,
3358 size_t size, gfp_t flags)
3360 if (time_to_inject(sbi, FAULT_KMALLOC)) {
3361 f2fs_show_injection_info(sbi, FAULT_KMALLOC);
3365 return kmalloc(size, flags);
3368 static inline void *f2fs_kzalloc(struct f2fs_sb_info *sbi,
3369 size_t size, gfp_t flags)
3371 return f2fs_kmalloc(sbi, size, flags | __GFP_ZERO);
3374 static inline void *f2fs_kvmalloc(struct f2fs_sb_info *sbi,
3375 size_t size, gfp_t flags)
3377 if (time_to_inject(sbi, FAULT_KVMALLOC)) {
3378 f2fs_show_injection_info(sbi, FAULT_KVMALLOC);
3382 return kvmalloc(size, flags);
3385 static inline void *f2fs_kvzalloc(struct f2fs_sb_info *sbi,
3386 size_t size, gfp_t flags)
3388 return f2fs_kvmalloc(sbi, size, flags | __GFP_ZERO);
3391 static inline int get_extra_isize(struct inode *inode)
3393 return F2FS_I(inode)->i_extra_isize / sizeof(__le32);
3396 static inline int get_inline_xattr_addrs(struct inode *inode)
3398 return F2FS_I(inode)->i_inline_xattr_size;
3401 static inline __le32 *get_dnode_addr(struct inode *inode,
3402 struct page *node_page)
3406 if (IS_INODE(node_page) && f2fs_has_extra_attr(inode))
3407 base = get_extra_isize(inode);
3409 return blkaddr_in_node(F2FS_NODE(node_page)) + base;
3412 #define f2fs_get_inode_mode(i) \
3413 ((is_inode_flag_set(i, FI_ACL_MODE)) ? \
3414 (F2FS_I(i)->i_acl_mode) : ((i)->i_mode))
3416 #define F2FS_TOTAL_EXTRA_ATTR_SIZE \
3417 (offsetof(struct f2fs_inode, i_extra_end) - \
3418 offsetof(struct f2fs_inode, i_extra_isize)) \
3420 #define F2FS_OLD_ATTRIBUTE_SIZE (offsetof(struct f2fs_inode, i_addr))
3421 #define F2FS_FITS_IN_INODE(f2fs_inode, extra_isize, field) \
3422 ((offsetof(typeof(*(f2fs_inode)), field) + \
3423 sizeof((f2fs_inode)->field)) \
3424 <= (F2FS_OLD_ATTRIBUTE_SIZE + (extra_isize))) \
3426 #define __is_large_section(sbi) ((sbi)->segs_per_sec > 1)
3428 #define __is_meta_io(fio) (PAGE_TYPE_OF_BIO((fio)->type) == META)
3430 bool f2fs_is_valid_blkaddr(struct f2fs_sb_info *sbi,
3431 block_t blkaddr, int type);
3432 static inline void verify_blkaddr(struct f2fs_sb_info *sbi,
3433 block_t blkaddr, int type)
3435 if (!f2fs_is_valid_blkaddr(sbi, blkaddr, type)) {
3436 f2fs_err(sbi, "invalid blkaddr: %u, type: %d, run fsck to fix.",
3438 f2fs_bug_on(sbi, 1);
3442 static inline bool __is_valid_data_blkaddr(block_t blkaddr)
3444 if (blkaddr == NEW_ADDR || blkaddr == NULL_ADDR ||
3445 blkaddr == COMPRESS_ADDR)
3453 int f2fs_sync_file(struct file *file, loff_t start, loff_t end, int datasync);
3454 void f2fs_truncate_data_blocks(struct dnode_of_data *dn);
3455 int f2fs_do_truncate_blocks(struct inode *inode, u64 from, bool lock);
3456 int f2fs_truncate_blocks(struct inode *inode, u64 from, bool lock);
3457 int f2fs_truncate(struct inode *inode);
3458 int f2fs_getattr(struct user_namespace *mnt_userns, const struct path *path,
3459 struct kstat *stat, u32 request_mask, unsigned int flags);
3460 int f2fs_setattr(struct user_namespace *mnt_userns, struct dentry *dentry,
3461 struct iattr *attr);
3462 int f2fs_truncate_hole(struct inode *inode, pgoff_t pg_start, pgoff_t pg_end);
3463 void f2fs_truncate_data_blocks_range(struct dnode_of_data *dn, int count);
3464 int f2fs_precache_extents(struct inode *inode);
3465 int f2fs_fileattr_get(struct dentry *dentry, struct fileattr *fa);
3466 int f2fs_fileattr_set(struct user_namespace *mnt_userns,
3467 struct dentry *dentry, struct fileattr *fa);
3468 long f2fs_ioctl(struct file *filp, unsigned int cmd, unsigned long arg);
3469 long f2fs_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
3470 int f2fs_transfer_project_quota(struct inode *inode, kprojid_t kprojid);
3471 int f2fs_pin_file_control(struct inode *inode, bool inc);
3476 void f2fs_set_inode_flags(struct inode *inode);
3477 bool f2fs_inode_chksum_verify(struct f2fs_sb_info *sbi, struct page *page);
3478 void f2fs_inode_chksum_set(struct f2fs_sb_info *sbi, struct page *page);
3479 struct inode *f2fs_iget(struct super_block *sb, unsigned long ino);
3480 struct inode *f2fs_iget_retry(struct super_block *sb, unsigned long ino);
3481 int f2fs_try_to_free_nats(struct f2fs_sb_info *sbi, int nr_shrink);
3482 void f2fs_update_inode(struct inode *inode, struct page *node_page);
3483 void f2fs_update_inode_page(struct inode *inode);
3484 int f2fs_write_inode(struct inode *inode, struct writeback_control *wbc);
3485 void f2fs_evict_inode(struct inode *inode);
3486 void f2fs_handle_failed_inode(struct inode *inode);
3491 int f2fs_update_extension_list(struct f2fs_sb_info *sbi, const char *name,
3492 bool hot, bool set);
3493 struct dentry *f2fs_get_parent(struct dentry *child);
3494 int f2fs_get_tmpfile(struct user_namespace *mnt_userns, struct inode *dir,
3495 struct inode **new_inode);
3500 unsigned char f2fs_get_de_type(struct f2fs_dir_entry *de);
3501 int f2fs_init_casefolded_name(const struct inode *dir,
3502 struct f2fs_filename *fname);
3503 int f2fs_setup_filename(struct inode *dir, const struct qstr *iname,
3504 int lookup, struct f2fs_filename *fname);
3505 int f2fs_prepare_lookup(struct inode *dir, struct dentry *dentry,
3506 struct f2fs_filename *fname);
3507 void f2fs_free_filename(struct f2fs_filename *fname);
3508 struct f2fs_dir_entry *f2fs_find_target_dentry(const struct f2fs_dentry_ptr *d,
3509 const struct f2fs_filename *fname, int *max_slots);
3510 int f2fs_fill_dentries(struct dir_context *ctx, struct f2fs_dentry_ptr *d,
3511 unsigned int start_pos, struct fscrypt_str *fstr);
3512 void f2fs_do_make_empty_dir(struct inode *inode, struct inode *parent,
3513 struct f2fs_dentry_ptr *d);
3514 struct page *f2fs_init_inode_metadata(struct inode *inode, struct inode *dir,
3515 const struct f2fs_filename *fname, struct page *dpage);
3516 void f2fs_update_parent_metadata(struct inode *dir, struct inode *inode,
3517 unsigned int current_depth);
3518 int f2fs_room_for_filename(const void *bitmap, int slots, int max_slots);
3519 void f2fs_drop_nlink(struct inode *dir, struct inode *inode);
3520 struct f2fs_dir_entry *__f2fs_find_entry(struct inode *dir,
3521 const struct f2fs_filename *fname,
3522 struct page **res_page);
3523 struct f2fs_dir_entry *f2fs_find_entry(struct inode *dir,
3524 const struct qstr *child, struct page **res_page);
3525 struct f2fs_dir_entry *f2fs_parent_dir(struct inode *dir, struct page **p);
3526 ino_t f2fs_inode_by_name(struct inode *dir, const struct qstr *qstr,
3527 struct page **page);
3528 void f2fs_set_link(struct inode *dir, struct f2fs_dir_entry *de,
3529 struct page *page, struct inode *inode);
3530 bool f2fs_has_enough_room(struct inode *dir, struct page *ipage,
3531 const struct f2fs_filename *fname);
3532 void f2fs_update_dentry(nid_t ino, umode_t mode, struct f2fs_dentry_ptr *d,
3533 const struct fscrypt_str *name, f2fs_hash_t name_hash,
3534 unsigned int bit_pos);
3535 int f2fs_add_regular_entry(struct inode *dir, const struct f2fs_filename *fname,
3536 struct inode *inode, nid_t ino, umode_t mode);
3537 int f2fs_add_dentry(struct inode *dir, const struct f2fs_filename *fname,
3538 struct inode *inode, nid_t ino, umode_t mode);
3539 int f2fs_do_add_link(struct inode *dir, const struct qstr *name,
3540 struct inode *inode, nid_t ino, umode_t mode);
3541 void f2fs_delete_entry(struct f2fs_dir_entry *dentry, struct page *page,
3542 struct inode *dir, struct inode *inode);
3543 int f2fs_do_tmpfile(struct inode *inode, struct inode *dir);
3544 bool f2fs_empty_dir(struct inode *dir);
3546 static inline int f2fs_add_link(struct dentry *dentry, struct inode *inode)
3548 if (fscrypt_is_nokey_name(dentry))
3550 return f2fs_do_add_link(d_inode(dentry->d_parent), &dentry->d_name,
3551 inode, inode->i_ino, inode->i_mode);
3557 int f2fs_inode_dirtied(struct inode *inode, bool sync);
3558 void f2fs_inode_synced(struct inode *inode);
3559 int f2fs_dquot_initialize(struct inode *inode);
3560 int f2fs_enable_quota_files(struct f2fs_sb_info *sbi, bool rdonly);
3561 int f2fs_quota_sync(struct super_block *sb, int type);
3562 loff_t max_file_blocks(struct inode *inode);
3563 void f2fs_quota_off_umount(struct super_block *sb);
3564 void f2fs_handle_stop(struct f2fs_sb_info *sbi, unsigned char reason);
3565 void f2fs_save_errors(struct f2fs_sb_info *sbi, unsigned char flag);
3566 void f2fs_handle_error(struct f2fs_sb_info *sbi, unsigned char error);
3567 int f2fs_commit_super(struct f2fs_sb_info *sbi, bool recover);
3568 int f2fs_sync_fs(struct super_block *sb, int sync);
3569 int f2fs_sanity_check_ckpt(struct f2fs_sb_info *sbi);
3574 void f2fs_hash_filename(const struct inode *dir, struct f2fs_filename *fname);
3581 int f2fs_check_nid_range(struct f2fs_sb_info *sbi, nid_t nid);
3582 bool f2fs_available_free_memory(struct f2fs_sb_info *sbi, int type);
3583 bool f2fs_in_warm_node_list(struct f2fs_sb_info *sbi, struct page *page);
3584 void f2fs_init_fsync_node_info(struct f2fs_sb_info *sbi);
3585 void f2fs_del_fsync_node_entry(struct f2fs_sb_info *sbi, struct page *page);
3586 void f2fs_reset_fsync_node_info(struct f2fs_sb_info *sbi);
3587 int f2fs_need_dentry_mark(struct f2fs_sb_info *sbi, nid_t nid);
3588 bool f2fs_is_checkpointed_node(struct f2fs_sb_info *sbi, nid_t nid);
3589 bool f2fs_need_inode_block_update(struct f2fs_sb_info *sbi, nid_t ino);
3590 int f2fs_get_node_info(struct f2fs_sb_info *sbi, nid_t nid,
3591 struct node_info *ni, bool checkpoint_context);
3592 pgoff_t f2fs_get_next_page_offset(struct dnode_of_data *dn, pgoff_t pgofs);
3593 int f2fs_get_dnode_of_data(struct dnode_of_data *dn, pgoff_t index, int mode);
3594 int f2fs_truncate_inode_blocks(struct inode *inode, pgoff_t from);
3595 int f2fs_truncate_xattr_node(struct inode *inode);
3596 int f2fs_wait_on_node_pages_writeback(struct f2fs_sb_info *sbi,
3597 unsigned int seq_id);
3598 bool f2fs_nat_bitmap_enabled(struct f2fs_sb_info *sbi);
3599 int f2fs_remove_inode_page(struct inode *inode);
3600 struct page *f2fs_new_inode_page(struct inode *inode);
3601 struct page *f2fs_new_node_page(struct dnode_of_data *dn, unsigned int ofs);
3602 void f2fs_ra_node_page(struct f2fs_sb_info *sbi, nid_t nid);
3603 struct page *f2fs_get_node_page(struct f2fs_sb_info *sbi, pgoff_t nid);
3604 struct page *f2fs_get_node_page_ra(struct page *parent, int start);
3605 int f2fs_move_node_page(struct page *node_page, int gc_type);
3606 void f2fs_flush_inline_data(struct f2fs_sb_info *sbi);
3607 int f2fs_fsync_node_pages(struct f2fs_sb_info *sbi, struct inode *inode,
3608 struct writeback_control *wbc, bool atomic,
3609 unsigned int *seq_id);
3610 int f2fs_sync_node_pages(struct f2fs_sb_info *sbi,
3611 struct writeback_control *wbc,
3612 bool do_balance, enum iostat_type io_type);
3613 int f2fs_build_free_nids(struct f2fs_sb_info *sbi, bool sync, bool mount);
3614 bool f2fs_alloc_nid(struct f2fs_sb_info *sbi, nid_t *nid);
3615 void f2fs_alloc_nid_done(struct f2fs_sb_info *sbi, nid_t nid);
3616 void f2fs_alloc_nid_failed(struct f2fs_sb_info *sbi, nid_t nid);
3617 int f2fs_try_to_free_nids(struct f2fs_sb_info *sbi, int nr_shrink);
3618 int f2fs_recover_inline_xattr(struct inode *inode, struct page *page);
3619 int f2fs_recover_xattr_data(struct inode *inode, struct page *page);
3620 int f2fs_recover_inode_page(struct f2fs_sb_info *sbi, struct page *page);
3621 int f2fs_restore_node_summary(struct f2fs_sb_info *sbi,
3622 unsigned int segno, struct f2fs_summary_block *sum);
3623 void f2fs_enable_nat_bits(struct f2fs_sb_info *sbi);
3624 int f2fs_flush_nat_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc);
3625 int f2fs_build_node_manager(struct f2fs_sb_info *sbi);
3626 void f2fs_destroy_node_manager(struct f2fs_sb_info *sbi);
3627 int __init f2fs_create_node_manager_caches(void);
3628 void f2fs_destroy_node_manager_caches(void);
3633 bool f2fs_need_SSR(struct f2fs_sb_info *sbi);
3634 int f2fs_commit_atomic_write(struct inode *inode);
3635 void f2fs_abort_atomic_write(struct inode *inode, bool clean);
3636 void f2fs_balance_fs(struct f2fs_sb_info *sbi, bool need);
3637 void f2fs_balance_fs_bg(struct f2fs_sb_info *sbi, bool from_bg);
3638 int f2fs_issue_flush(struct f2fs_sb_info *sbi, nid_t ino);
3639 int f2fs_create_flush_cmd_control(struct f2fs_sb_info *sbi);
3640 int f2fs_flush_device_cache(struct f2fs_sb_info *sbi);
3641 void f2fs_destroy_flush_cmd_control(struct f2fs_sb_info *sbi, bool free);
3642 void f2fs_invalidate_blocks(struct f2fs_sb_info *sbi, block_t addr);
3643 bool f2fs_is_checkpointed_data(struct f2fs_sb_info *sbi, block_t blkaddr);
3644 int f2fs_start_discard_thread(struct f2fs_sb_info *sbi);
3645 void f2fs_drop_discard_cmd(struct f2fs_sb_info *sbi);
3646 void f2fs_stop_discard_thread(struct f2fs_sb_info *sbi);
3647 bool f2fs_issue_discard_timeout(struct f2fs_sb_info *sbi);
3648 void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
3649 struct cp_control *cpc);
3650 void f2fs_dirty_to_prefree(struct f2fs_sb_info *sbi);
3651 block_t f2fs_get_unusable_blocks(struct f2fs_sb_info *sbi);
3652 int f2fs_disable_cp_again(struct f2fs_sb_info *sbi, block_t unusable);
3653 void f2fs_release_discard_addrs(struct f2fs_sb_info *sbi);
3654 int f2fs_npages_for_summary_flush(struct f2fs_sb_info *sbi, bool for_ra);
3655 bool f2fs_segment_has_free_slot(struct f2fs_sb_info *sbi, int segno);
3656 void f2fs_init_inmem_curseg(struct f2fs_sb_info *sbi);
3657 void f2fs_save_inmem_curseg(struct f2fs_sb_info *sbi);
3658 void f2fs_restore_inmem_curseg(struct f2fs_sb_info *sbi);
3659 void f2fs_get_new_segment(struct f2fs_sb_info *sbi,
3660 unsigned int *newseg, bool new_sec, int dir);
3661 void f2fs_allocate_segment_for_resize(struct f2fs_sb_info *sbi, int type,
3662 unsigned int start, unsigned int end);
3663 void f2fs_allocate_new_section(struct f2fs_sb_info *sbi, int type, bool force);
3664 void f2fs_allocate_new_segments(struct f2fs_sb_info *sbi);
3665 int f2fs_trim_fs(struct f2fs_sb_info *sbi, struct fstrim_range *range);
3666 bool f2fs_exist_trim_candidates(struct f2fs_sb_info *sbi,
3667 struct cp_control *cpc);
3668 struct page *f2fs_get_sum_page(struct f2fs_sb_info *sbi, unsigned int segno);
3669 void f2fs_update_meta_page(struct f2fs_sb_info *sbi, void *src,
3671 void f2fs_do_write_meta_page(struct f2fs_sb_info *sbi, struct page *page,
3672 enum iostat_type io_type);
3673 void f2fs_do_write_node_page(unsigned int nid, struct f2fs_io_info *fio);
3674 void f2fs_outplace_write_data(struct dnode_of_data *dn,
3675 struct f2fs_io_info *fio);
3676 int f2fs_inplace_write_data(struct f2fs_io_info *fio);
3677 void f2fs_do_replace_block(struct f2fs_sb_info *sbi, struct f2fs_summary *sum,
3678 block_t old_blkaddr, block_t new_blkaddr,
3679 bool recover_curseg, bool recover_newaddr,
3681 void f2fs_replace_block(struct f2fs_sb_info *sbi, struct dnode_of_data *dn,
3682 block_t old_addr, block_t new_addr,
3683 unsigned char version, bool recover_curseg,
3684 bool recover_newaddr);
3685 void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page,
3686 block_t old_blkaddr, block_t *new_blkaddr,
3687 struct f2fs_summary *sum, int type,
3688 struct f2fs_io_info *fio);
3689 void f2fs_update_device_state(struct f2fs_sb_info *sbi, nid_t ino,
3690 block_t blkaddr, unsigned int blkcnt);
3691 void f2fs_wait_on_page_writeback(struct page *page,
3692 enum page_type type, bool ordered, bool locked);
3693 void f2fs_wait_on_block_writeback(struct inode *inode, block_t blkaddr);
3694 void f2fs_wait_on_block_writeback_range(struct inode *inode, block_t blkaddr,
3696 void f2fs_write_data_summaries(struct f2fs_sb_info *sbi, block_t start_blk);
3697 void f2fs_write_node_summaries(struct f2fs_sb_info *sbi, block_t start_blk);
3698 int f2fs_lookup_journal_in_cursum(struct f2fs_journal *journal, int type,
3699 unsigned int val, int alloc);
3700 void f2fs_flush_sit_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc);
3701 int f2fs_fix_curseg_write_pointer(struct f2fs_sb_info *sbi);
3702 int f2fs_check_write_pointer(struct f2fs_sb_info *sbi);
3703 int f2fs_build_segment_manager(struct f2fs_sb_info *sbi);
3704 void f2fs_destroy_segment_manager(struct f2fs_sb_info *sbi);
3705 int __init f2fs_create_segment_manager_caches(void);
3706 void f2fs_destroy_segment_manager_caches(void);
3707 int f2fs_rw_hint_to_seg_type(enum rw_hint hint);
3708 unsigned int f2fs_usable_segs_in_sec(struct f2fs_sb_info *sbi,
3709 unsigned int segno);
3710 unsigned int f2fs_usable_blks_in_seg(struct f2fs_sb_info *sbi,
3711 unsigned int segno);
3713 #define DEF_FRAGMENT_SIZE 4
3714 #define MIN_FRAGMENT_SIZE 1
3715 #define MAX_FRAGMENT_SIZE 512
3717 static inline bool f2fs_need_rand_seg(struct f2fs_sb_info *sbi)
3719 return F2FS_OPTION(sbi).fs_mode == FS_MODE_FRAGMENT_SEG ||
3720 F2FS_OPTION(sbi).fs_mode == FS_MODE_FRAGMENT_BLK;
3726 void f2fs_stop_checkpoint(struct f2fs_sb_info *sbi, bool end_io,
3727 unsigned char reason);
3728 void f2fs_flush_ckpt_thread(struct f2fs_sb_info *sbi);
3729 struct page *f2fs_grab_meta_page(struct f2fs_sb_info *sbi, pgoff_t index);
3730 struct page *f2fs_get_meta_page(struct f2fs_sb_info *sbi, pgoff_t index);
3731 struct page *f2fs_get_meta_page_retry(struct f2fs_sb_info *sbi, pgoff_t index);
3732 struct page *f2fs_get_tmp_page(struct f2fs_sb_info *sbi, pgoff_t index);
3733 bool f2fs_is_valid_blkaddr(struct f2fs_sb_info *sbi,
3734 block_t blkaddr, int type);
3735 int f2fs_ra_meta_pages(struct f2fs_sb_info *sbi, block_t start, int nrpages,
3736 int type, bool sync);
3737 void f2fs_ra_meta_pages_cond(struct f2fs_sb_info *sbi, pgoff_t index,
3738 unsigned int ra_blocks);
3739 long f2fs_sync_meta_pages(struct f2fs_sb_info *sbi, enum page_type type,
3740 long nr_to_write, enum iostat_type io_type);
3741 void f2fs_add_ino_entry(struct f2fs_sb_info *sbi, nid_t ino, int type);
3742 void f2fs_remove_ino_entry(struct f2fs_sb_info *sbi, nid_t ino, int type);
3743 void f2fs_release_ino_entry(struct f2fs_sb_info *sbi, bool all);
3744 bool f2fs_exist_written_data(struct f2fs_sb_info *sbi, nid_t ino, int mode);
3745 void f2fs_set_dirty_device(struct f2fs_sb_info *sbi, nid_t ino,
3746 unsigned int devidx, int type);
3747 bool f2fs_is_dirty_device(struct f2fs_sb_info *sbi, nid_t ino,
3748 unsigned int devidx, int type);
3749 int f2fs_sync_inode_meta(struct f2fs_sb_info *sbi);
3750 int f2fs_acquire_orphan_inode(struct f2fs_sb_info *sbi);
3751 void f2fs_release_orphan_inode(struct f2fs_sb_info *sbi);
3752 void f2fs_add_orphan_inode(struct inode *inode);
3753 void f2fs_remove_orphan_inode(struct f2fs_sb_info *sbi, nid_t ino);
3754 int f2fs_recover_orphan_inodes(struct f2fs_sb_info *sbi);
3755 int f2fs_get_valid_checkpoint(struct f2fs_sb_info *sbi);
3756 void f2fs_update_dirty_folio(struct inode *inode, struct folio *folio);
3757 void f2fs_remove_dirty_inode(struct inode *inode);
3758 int f2fs_sync_dirty_inodes(struct f2fs_sb_info *sbi, enum inode_type type,
3760 void f2fs_wait_on_all_pages(struct f2fs_sb_info *sbi, int type);
3761 u64 f2fs_get_sectors_written(struct f2fs_sb_info *sbi);
3762 int f2fs_write_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc);
3763 void f2fs_init_ino_entry_info(struct f2fs_sb_info *sbi);
3764 int __init f2fs_create_checkpoint_caches(void);
3765 void f2fs_destroy_checkpoint_caches(void);
3766 int f2fs_issue_checkpoint(struct f2fs_sb_info *sbi);
3767 int f2fs_start_ckpt_thread(struct f2fs_sb_info *sbi);
3768 void f2fs_stop_ckpt_thread(struct f2fs_sb_info *sbi);
3769 void f2fs_init_ckpt_req_control(struct f2fs_sb_info *sbi);
3774 int __init f2fs_init_bioset(void);
3775 void f2fs_destroy_bioset(void);
3776 bool f2fs_is_cp_guaranteed(struct page *page);
3777 int f2fs_init_bio_entry_cache(void);
3778 void f2fs_destroy_bio_entry_cache(void);
3779 void f2fs_submit_bio(struct f2fs_sb_info *sbi,
3780 struct bio *bio, enum page_type type);
3781 int f2fs_init_write_merge_io(struct f2fs_sb_info *sbi);
3782 void f2fs_submit_merged_write(struct f2fs_sb_info *sbi, enum page_type type);
3783 void f2fs_submit_merged_write_cond(struct f2fs_sb_info *sbi,
3784 struct inode *inode, struct page *page,
3785 nid_t ino, enum page_type type);
3786 void f2fs_submit_merged_ipu_write(struct f2fs_sb_info *sbi,
3787 struct bio **bio, struct page *page);
3788 void f2fs_flush_merged_writes(struct f2fs_sb_info *sbi);
3789 int f2fs_submit_page_bio(struct f2fs_io_info *fio);
3790 int f2fs_merge_page_bio(struct f2fs_io_info *fio);
3791 void f2fs_submit_page_write(struct f2fs_io_info *fio);
3792 struct block_device *f2fs_target_device(struct f2fs_sb_info *sbi,
3793 block_t blk_addr, sector_t *sector);
3794 int f2fs_target_device_index(struct f2fs_sb_info *sbi, block_t blkaddr);
3795 void f2fs_set_data_blkaddr(struct dnode_of_data *dn, block_t blkaddr);
3796 void f2fs_update_data_blkaddr(struct dnode_of_data *dn, block_t blkaddr);
3797 int f2fs_reserve_new_blocks(struct dnode_of_data *dn, blkcnt_t count);
3798 int f2fs_reserve_new_block(struct dnode_of_data *dn);
3799 int f2fs_get_block(struct dnode_of_data *dn, pgoff_t index);
3800 int f2fs_reserve_block(struct dnode_of_data *dn, pgoff_t index);
3801 struct page *f2fs_get_read_data_page(struct inode *inode, pgoff_t index,
3802 blk_opf_t op_flags, bool for_write, pgoff_t *next_pgofs);
3803 struct page *f2fs_find_data_page(struct inode *inode, pgoff_t index,
3804 pgoff_t *next_pgofs);
3805 struct page *f2fs_get_lock_data_page(struct inode *inode, pgoff_t index,
3807 struct page *f2fs_get_new_data_page(struct inode *inode,
3808 struct page *ipage, pgoff_t index, bool new_i_size);
3809 int f2fs_do_write_data_page(struct f2fs_io_info *fio);
3810 void f2fs_do_map_lock(struct f2fs_sb_info *sbi, int flag, bool lock);
3811 int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map,
3812 int create, int flag);
3813 int f2fs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
3814 u64 start, u64 len);
3815 int f2fs_encrypt_one_page(struct f2fs_io_info *fio);
3816 bool f2fs_should_update_inplace(struct inode *inode, struct f2fs_io_info *fio);
3817 bool f2fs_should_update_outplace(struct inode *inode, struct f2fs_io_info *fio);
3818 int f2fs_write_single_data_page(struct page *page, int *submitted,
3819 struct bio **bio, sector_t *last_block,
3820 struct writeback_control *wbc,
3821 enum iostat_type io_type,
3822 int compr_blocks, bool allow_balance);
3823 void f2fs_write_failed(struct inode *inode, loff_t to);
3824 void f2fs_invalidate_folio(struct folio *folio, size_t offset, size_t length);
3825 bool f2fs_release_folio(struct folio *folio, gfp_t wait);
3826 bool f2fs_overwrite_io(struct inode *inode, loff_t pos, size_t len);
3827 void f2fs_clear_page_cache_dirty_tag(struct page *page);
3828 int f2fs_init_post_read_processing(void);
3829 void f2fs_destroy_post_read_processing(void);
3830 int f2fs_init_post_read_wq(struct f2fs_sb_info *sbi);
3831 void f2fs_destroy_post_read_wq(struct f2fs_sb_info *sbi);
3832 extern const struct iomap_ops f2fs_iomap_ops;
3837 int f2fs_start_gc_thread(struct f2fs_sb_info *sbi);
3838 void f2fs_stop_gc_thread(struct f2fs_sb_info *sbi);
3839 block_t f2fs_start_bidx_of_node(unsigned int node_ofs, struct inode *inode);
3840 int f2fs_gc(struct f2fs_sb_info *sbi, struct f2fs_gc_control *gc_control);
3841 void f2fs_build_gc_manager(struct f2fs_sb_info *sbi);
3842 int f2fs_resize_fs(struct file *filp, __u64 block_count);
3843 int __init f2fs_create_garbage_collection_cache(void);
3844 void f2fs_destroy_garbage_collection_cache(void);
3849 int f2fs_recover_fsync_data(struct f2fs_sb_info *sbi, bool check_only);
3850 bool f2fs_space_for_roll_forward(struct f2fs_sb_info *sbi);
3851 int __init f2fs_create_recovery_cache(void);
3852 void f2fs_destroy_recovery_cache(void);
3857 #ifdef CONFIG_F2FS_STAT_FS
3858 struct f2fs_stat_info {
3859 struct list_head stat_list;
3860 struct f2fs_sb_info *sbi;
3861 int all_area_segs, sit_area_segs, nat_area_segs, ssa_area_segs;
3862 int main_area_segs, main_area_sections, main_area_zones;
3863 unsigned long long hit_cached[NR_EXTENT_CACHES];
3864 unsigned long long hit_rbtree[NR_EXTENT_CACHES];
3865 unsigned long long total_ext[NR_EXTENT_CACHES];
3866 unsigned long long hit_total[NR_EXTENT_CACHES];
3867 int ext_tree[NR_EXTENT_CACHES];
3868 int zombie_tree[NR_EXTENT_CACHES];
3869 int ext_node[NR_EXTENT_CACHES];
3870 /* to count memory footprint */
3871 unsigned long long ext_mem[NR_EXTENT_CACHES];
3872 /* for read extent cache */
3873 unsigned long long hit_largest;
3874 int ndirty_node, ndirty_dent, ndirty_meta, ndirty_imeta;
3875 int ndirty_data, ndirty_qdata;
3876 unsigned int ndirty_dirs, ndirty_files, nquota_files, ndirty_all;
3877 int nats, dirty_nats, sits, dirty_sits;
3878 int free_nids, avail_nids, alloc_nids;
3879 int total_count, utilization;
3880 int bg_gc, nr_wb_cp_data, nr_wb_data;
3881 int nr_rd_data, nr_rd_node, nr_rd_meta;
3882 int nr_dio_read, nr_dio_write;
3883 unsigned int io_skip_bggc, other_skip_bggc;
3884 int nr_flushing, nr_flushed, flush_list_empty;
3885 int nr_discarding, nr_discarded;
3887 unsigned int undiscard_blks;
3888 int nr_issued_ckpt, nr_total_ckpt, nr_queued_ckpt;
3889 unsigned int cur_ckpt_time, peak_ckpt_time;
3890 int inline_xattr, inline_inode, inline_dir, append, update, orphans;
3891 int compr_inode, swapfile_inode;
3892 unsigned long long compr_blocks;
3893 int aw_cnt, max_aw_cnt;
3894 unsigned int valid_count, valid_node_count, valid_inode_count, discard_blks;
3895 unsigned int bimodal, avg_vblocks;
3896 int util_free, util_valid, util_invalid;
3897 int rsvd_segs, overp_segs;
3898 int dirty_count, node_pages, meta_pages, compress_pages;
3899 int compress_page_hit;
3900 int prefree_count, call_count, cp_count, bg_cp_count;
3901 int tot_segs, node_segs, data_segs, free_segs, free_secs;
3902 int bg_node_segs, bg_data_segs;
3903 int tot_blks, data_blks, node_blks;
3904 int bg_data_blks, bg_node_blks;
3905 int curseg[NR_CURSEG_TYPE];
3906 int cursec[NR_CURSEG_TYPE];
3907 int curzone[NR_CURSEG_TYPE];
3908 unsigned int dirty_seg[NR_CURSEG_TYPE];
3909 unsigned int full_seg[NR_CURSEG_TYPE];
3910 unsigned int valid_blks[NR_CURSEG_TYPE];
3912 unsigned int meta_count[META_MAX];
3913 unsigned int segment_count[2];
3914 unsigned int block_count[2];
3915 unsigned int inplace_count;
3916 unsigned long long base_mem, cache_mem, page_mem;
3919 static inline struct f2fs_stat_info *F2FS_STAT(struct f2fs_sb_info *sbi)
3921 return (struct f2fs_stat_info *)sbi->stat_info;
3924 #define stat_inc_cp_count(si) ((si)->cp_count++)
3925 #define stat_inc_bg_cp_count(si) ((si)->bg_cp_count++)
3926 #define stat_inc_call_count(si) ((si)->call_count++)
3927 #define stat_inc_bggc_count(si) ((si)->bg_gc++)
3928 #define stat_io_skip_bggc_count(sbi) ((sbi)->io_skip_bggc++)
3929 #define stat_other_skip_bggc_count(sbi) ((sbi)->other_skip_bggc++)
3930 #define stat_inc_dirty_inode(sbi, type) ((sbi)->ndirty_inode[type]++)
3931 #define stat_dec_dirty_inode(sbi, type) ((sbi)->ndirty_inode[type]--)
3932 #define stat_inc_total_hit(sbi, type) (atomic64_inc(&(sbi)->total_hit_ext[type]))
3933 #define stat_inc_rbtree_node_hit(sbi, type) (atomic64_inc(&(sbi)->read_hit_rbtree[type]))
3934 #define stat_inc_largest_node_hit(sbi) (atomic64_inc(&(sbi)->read_hit_largest))
3935 #define stat_inc_cached_node_hit(sbi, type) (atomic64_inc(&(sbi)->read_hit_cached[type]))
3936 #define stat_inc_inline_xattr(inode) \
3938 if (f2fs_has_inline_xattr(inode)) \
3939 (atomic_inc(&F2FS_I_SB(inode)->inline_xattr)); \
3941 #define stat_dec_inline_xattr(inode) \
3943 if (f2fs_has_inline_xattr(inode)) \
3944 (atomic_dec(&F2FS_I_SB(inode)->inline_xattr)); \
3946 #define stat_inc_inline_inode(inode) \
3948 if (f2fs_has_inline_data(inode)) \
3949 (atomic_inc(&F2FS_I_SB(inode)->inline_inode)); \
3951 #define stat_dec_inline_inode(inode) \
3953 if (f2fs_has_inline_data(inode)) \
3954 (atomic_dec(&F2FS_I_SB(inode)->inline_inode)); \
3956 #define stat_inc_inline_dir(inode) \
3958 if (f2fs_has_inline_dentry(inode)) \
3959 (atomic_inc(&F2FS_I_SB(inode)->inline_dir)); \
3961 #define stat_dec_inline_dir(inode) \
3963 if (f2fs_has_inline_dentry(inode)) \
3964 (atomic_dec(&F2FS_I_SB(inode)->inline_dir)); \
3966 #define stat_inc_compr_inode(inode) \
3968 if (f2fs_compressed_file(inode)) \
3969 (atomic_inc(&F2FS_I_SB(inode)->compr_inode)); \
3971 #define stat_dec_compr_inode(inode) \
3973 if (f2fs_compressed_file(inode)) \
3974 (atomic_dec(&F2FS_I_SB(inode)->compr_inode)); \
3976 #define stat_add_compr_blocks(inode, blocks) \
3977 (atomic64_add(blocks, &F2FS_I_SB(inode)->compr_blocks))
3978 #define stat_sub_compr_blocks(inode, blocks) \
3979 (atomic64_sub(blocks, &F2FS_I_SB(inode)->compr_blocks))
3980 #define stat_inc_swapfile_inode(inode) \
3981 (atomic_inc(&F2FS_I_SB(inode)->swapfile_inode))
3982 #define stat_dec_swapfile_inode(inode) \
3983 (atomic_dec(&F2FS_I_SB(inode)->swapfile_inode))
3984 #define stat_inc_atomic_inode(inode) \
3985 (atomic_inc(&F2FS_I_SB(inode)->atomic_files))
3986 #define stat_dec_atomic_inode(inode) \
3987 (atomic_dec(&F2FS_I_SB(inode)->atomic_files))
3988 #define stat_inc_meta_count(sbi, blkaddr) \
3990 if (blkaddr < SIT_I(sbi)->sit_base_addr) \
3991 atomic_inc(&(sbi)->meta_count[META_CP]); \
3992 else if (blkaddr < NM_I(sbi)->nat_blkaddr) \
3993 atomic_inc(&(sbi)->meta_count[META_SIT]); \
3994 else if (blkaddr < SM_I(sbi)->ssa_blkaddr) \
3995 atomic_inc(&(sbi)->meta_count[META_NAT]); \
3996 else if (blkaddr < SM_I(sbi)->main_blkaddr) \
3997 atomic_inc(&(sbi)->meta_count[META_SSA]); \
3999 #define stat_inc_seg_type(sbi, curseg) \
4000 ((sbi)->segment_count[(curseg)->alloc_type]++)
4001 #define stat_inc_block_count(sbi, curseg) \
4002 ((sbi)->block_count[(curseg)->alloc_type]++)
4003 #define stat_inc_inplace_blocks(sbi) \
4004 (atomic_inc(&(sbi)->inplace_count))
4005 #define stat_update_max_atomic_write(inode) \
4007 int cur = atomic_read(&F2FS_I_SB(inode)->atomic_files); \
4008 int max = atomic_read(&F2FS_I_SB(inode)->max_aw_cnt); \
4010 atomic_set(&F2FS_I_SB(inode)->max_aw_cnt, cur); \
4012 #define stat_inc_seg_count(sbi, type, gc_type) \
4014 struct f2fs_stat_info *si = F2FS_STAT(sbi); \
4016 if ((type) == SUM_TYPE_DATA) { \
4018 si->bg_data_segs += (gc_type == BG_GC) ? 1 : 0; \
4021 si->bg_node_segs += (gc_type == BG_GC) ? 1 : 0; \
4025 #define stat_inc_tot_blk_count(si, blks) \
4026 ((si)->tot_blks += (blks))
4028 #define stat_inc_data_blk_count(sbi, blks, gc_type) \
4030 struct f2fs_stat_info *si = F2FS_STAT(sbi); \
4031 stat_inc_tot_blk_count(si, blks); \
4032 si->data_blks += (blks); \
4033 si->bg_data_blks += ((gc_type) == BG_GC) ? (blks) : 0; \
4036 #define stat_inc_node_blk_count(sbi, blks, gc_type) \
4038 struct f2fs_stat_info *si = F2FS_STAT(sbi); \
4039 stat_inc_tot_blk_count(si, blks); \
4040 si->node_blks += (blks); \
4041 si->bg_node_blks += ((gc_type) == BG_GC) ? (blks) : 0; \
4044 int f2fs_build_stats(struct f2fs_sb_info *sbi);
4045 void f2fs_destroy_stats(struct f2fs_sb_info *sbi);
4046 void __init f2fs_create_root_stats(void);
4047 void f2fs_destroy_root_stats(void);
4048 void f2fs_update_sit_info(struct f2fs_sb_info *sbi);
4050 #define stat_inc_cp_count(si) do { } while (0)
4051 #define stat_inc_bg_cp_count(si) do { } while (0)
4052 #define stat_inc_call_count(si) do { } while (0)
4053 #define stat_inc_bggc_count(si) do { } while (0)
4054 #define stat_io_skip_bggc_count(sbi) do { } while (0)
4055 #define stat_other_skip_bggc_count(sbi) do { } while (0)
4056 #define stat_inc_dirty_inode(sbi, type) do { } while (0)
4057 #define stat_dec_dirty_inode(sbi, type) do { } while (0)
4058 #define stat_inc_total_hit(sbi, type) do { } while (0)
4059 #define stat_inc_rbtree_node_hit(sbi, type) do { } while (0)
4060 #define stat_inc_largest_node_hit(sbi) do { } while (0)
4061 #define stat_inc_cached_node_hit(sbi, type) do { } while (0)
4062 #define stat_inc_inline_xattr(inode) do { } while (0)
4063 #define stat_dec_inline_xattr(inode) do { } while (0)
4064 #define stat_inc_inline_inode(inode) do { } while (0)
4065 #define stat_dec_inline_inode(inode) do { } while (0)
4066 #define stat_inc_inline_dir(inode) do { } while (0)
4067 #define stat_dec_inline_dir(inode) do { } while (0)
4068 #define stat_inc_compr_inode(inode) do { } while (0)
4069 #define stat_dec_compr_inode(inode) do { } while (0)
4070 #define stat_add_compr_blocks(inode, blocks) do { } while (0)
4071 #define stat_sub_compr_blocks(inode, blocks) do { } while (0)
4072 #define stat_inc_swapfile_inode(inode) do { } while (0)
4073 #define stat_dec_swapfile_inode(inode) do { } while (0)
4074 #define stat_inc_atomic_inode(inode) do { } while (0)
4075 #define stat_dec_atomic_inode(inode) do { } while (0)
4076 #define stat_update_max_atomic_write(inode) do { } while (0)
4077 #define stat_inc_meta_count(sbi, blkaddr) do { } while (0)
4078 #define stat_inc_seg_type(sbi, curseg) do { } while (0)
4079 #define stat_inc_block_count(sbi, curseg) do { } while (0)
4080 #define stat_inc_inplace_blocks(sbi) do { } while (0)
4081 #define stat_inc_seg_count(sbi, type, gc_type) do { } while (0)
4082 #define stat_inc_tot_blk_count(si, blks) do { } while (0)
4083 #define stat_inc_data_blk_count(sbi, blks, gc_type) do { } while (0)
4084 #define stat_inc_node_blk_count(sbi, blks, gc_type) do { } while (0)
4086 static inline int f2fs_build_stats(struct f2fs_sb_info *sbi) { return 0; }
4087 static inline void f2fs_destroy_stats(struct f2fs_sb_info *sbi) { }
4088 static inline void __init f2fs_create_root_stats(void) { }
4089 static inline void f2fs_destroy_root_stats(void) { }
4090 static inline void f2fs_update_sit_info(struct f2fs_sb_info *sbi) {}
4093 extern const struct file_operations f2fs_dir_operations;
4094 extern const struct file_operations f2fs_file_operations;
4095 extern const struct inode_operations f2fs_file_inode_operations;
4096 extern const struct address_space_operations f2fs_dblock_aops;
4097 extern const struct address_space_operations f2fs_node_aops;
4098 extern const struct address_space_operations f2fs_meta_aops;
4099 extern const struct inode_operations f2fs_dir_inode_operations;
4100 extern const struct inode_operations f2fs_symlink_inode_operations;
4101 extern const struct inode_operations f2fs_encrypted_symlink_inode_operations;
4102 extern const struct inode_operations f2fs_special_inode_operations;
4103 extern struct kmem_cache *f2fs_inode_entry_slab;
4108 bool f2fs_may_inline_data(struct inode *inode);
4109 bool f2fs_sanity_check_inline_data(struct inode *inode);
4110 bool f2fs_may_inline_dentry(struct inode *inode);
4111 void f2fs_do_read_inline_data(struct page *page, struct page *ipage);
4112 void f2fs_truncate_inline_inode(struct inode *inode,
4113 struct page *ipage, u64 from);
4114 int f2fs_read_inline_data(struct inode *inode, struct page *page);
4115 int f2fs_convert_inline_page(struct dnode_of_data *dn, struct page *page);
4116 int f2fs_convert_inline_inode(struct inode *inode);
4117 int f2fs_try_convert_inline_dir(struct inode *dir, struct dentry *dentry);
4118 int f2fs_write_inline_data(struct inode *inode, struct page *page);
4119 int f2fs_recover_inline_data(struct inode *inode, struct page *npage);
4120 struct f2fs_dir_entry *f2fs_find_in_inline_dir(struct inode *dir,
4121 const struct f2fs_filename *fname,
4122 struct page **res_page);
4123 int f2fs_make_empty_inline_dir(struct inode *inode, struct inode *parent,
4124 struct page *ipage);
4125 int f2fs_add_inline_entry(struct inode *dir, const struct f2fs_filename *fname,
4126 struct inode *inode, nid_t ino, umode_t mode);
4127 void f2fs_delete_inline_entry(struct f2fs_dir_entry *dentry,
4128 struct page *page, struct inode *dir,
4129 struct inode *inode);
4130 bool f2fs_empty_inline_dir(struct inode *dir);
4131 int f2fs_read_inline_dir(struct file *file, struct dir_context *ctx,
4132 struct fscrypt_str *fstr);
4133 int f2fs_inline_data_fiemap(struct inode *inode,
4134 struct fiemap_extent_info *fieinfo,
4135 __u64 start, __u64 len);
4140 unsigned long f2fs_shrink_count(struct shrinker *shrink,
4141 struct shrink_control *sc);
4142 unsigned long f2fs_shrink_scan(struct shrinker *shrink,
4143 struct shrink_control *sc);
4144 void f2fs_join_shrinker(struct f2fs_sb_info *sbi);
4145 void f2fs_leave_shrinker(struct f2fs_sb_info *sbi);
4150 bool sanity_check_extent_cache(struct inode *inode);
4151 struct rb_entry *f2fs_lookup_rb_tree(struct rb_root_cached *root,
4152 struct rb_entry *cached_re, unsigned int ofs);
4153 struct rb_node **f2fs_lookup_rb_tree_for_insert(struct f2fs_sb_info *sbi,
4154 struct rb_root_cached *root,
4155 struct rb_node **parent,
4156 unsigned int ofs, bool *leftmost);
4157 struct rb_entry *f2fs_lookup_rb_tree_ret(struct rb_root_cached *root,
4158 struct rb_entry *cached_re, unsigned int ofs,
4159 struct rb_entry **prev_entry, struct rb_entry **next_entry,
4160 struct rb_node ***insert_p, struct rb_node **insert_parent,
4161 bool force, bool *leftmost);
4162 bool f2fs_check_rb_tree_consistence(struct f2fs_sb_info *sbi,
4163 struct rb_root_cached *root);
4164 void f2fs_init_extent_tree(struct inode *inode);
4165 void f2fs_drop_extent_tree(struct inode *inode);
4166 void f2fs_destroy_extent_node(struct inode *inode);
4167 void f2fs_destroy_extent_tree(struct inode *inode);
4168 void f2fs_init_extent_cache_info(struct f2fs_sb_info *sbi);
4169 int __init f2fs_create_extent_cache(void);
4170 void f2fs_destroy_extent_cache(void);
4172 /* read extent cache ops */
4173 void f2fs_init_read_extent_tree(struct inode *inode, struct page *ipage);
4174 bool f2fs_lookup_read_extent_cache(struct inode *inode, pgoff_t pgofs,
4175 struct extent_info *ei);
4176 void f2fs_update_read_extent_cache(struct dnode_of_data *dn);
4177 void f2fs_update_read_extent_cache_range(struct dnode_of_data *dn,
4178 pgoff_t fofs, block_t blkaddr, unsigned int len);
4179 unsigned int f2fs_shrink_read_extent_tree(struct f2fs_sb_info *sbi,
4185 #define MIN_RA_MUL 2
4186 #define MAX_RA_MUL 256
4188 int __init f2fs_init_sysfs(void);
4189 void f2fs_exit_sysfs(void);
4190 int f2fs_register_sysfs(struct f2fs_sb_info *sbi);
4191 void f2fs_unregister_sysfs(struct f2fs_sb_info *sbi);
4194 extern const struct fsverity_operations f2fs_verityops;
4199 static inline bool f2fs_encrypted_file(struct inode *inode)
4201 return IS_ENCRYPTED(inode) && S_ISREG(inode->i_mode);
4204 static inline void f2fs_set_encrypted_inode(struct inode *inode)
4206 #ifdef CONFIG_FS_ENCRYPTION
4207 file_set_encrypt(inode);
4208 f2fs_set_inode_flags(inode);
4213 * Returns true if the reads of the inode's data need to undergo some
4214 * postprocessing step, like decryption or authenticity verification.
4216 static inline bool f2fs_post_read_required(struct inode *inode)
4218 return f2fs_encrypted_file(inode) || fsverity_active(inode) ||
4219 f2fs_compressed_file(inode);
4225 #ifdef CONFIG_F2FS_FS_COMPRESSION
4226 bool f2fs_is_compressed_page(struct page *page);
4227 struct page *f2fs_compress_control_page(struct page *page);
4228 int f2fs_prepare_compress_overwrite(struct inode *inode,
4229 struct page **pagep, pgoff_t index, void **fsdata);
4230 bool f2fs_compress_write_end(struct inode *inode, void *fsdata,
4231 pgoff_t index, unsigned copied);
4232 int f2fs_truncate_partial_cluster(struct inode *inode, u64 from, bool lock);
4233 void f2fs_compress_write_end_io(struct bio *bio, struct page *page);
4234 bool f2fs_is_compress_backend_ready(struct inode *inode);
4235 bool f2fs_is_compress_level_valid(int alg, int lvl);
4236 int f2fs_init_compress_mempool(void);
4237 void f2fs_destroy_compress_mempool(void);
4238 void f2fs_decompress_cluster(struct decompress_io_ctx *dic, bool in_task);
4239 void f2fs_end_read_compressed_page(struct page *page, bool failed,
4240 block_t blkaddr, bool in_task);
4241 bool f2fs_cluster_is_empty(struct compress_ctx *cc);
4242 bool f2fs_cluster_can_merge_page(struct compress_ctx *cc, pgoff_t index);
4243 bool f2fs_all_cluster_page_ready(struct compress_ctx *cc, struct page **pages,
4244 int index, int nr_pages, bool uptodate);
4245 bool f2fs_sanity_check_cluster(struct dnode_of_data *dn);
4246 void f2fs_compress_ctx_add_page(struct compress_ctx *cc, struct page *page);
4247 int f2fs_write_multi_pages(struct compress_ctx *cc,
4249 struct writeback_control *wbc,
4250 enum iostat_type io_type);
4251 int f2fs_is_compressed_cluster(struct inode *inode, pgoff_t index);
4252 void f2fs_update_read_extent_tree_range_compressed(struct inode *inode,
4253 pgoff_t fofs, block_t blkaddr,
4254 unsigned int llen, unsigned int c_len);
4255 int f2fs_read_multi_pages(struct compress_ctx *cc, struct bio **bio_ret,
4256 unsigned nr_pages, sector_t *last_block_in_bio,
4257 bool is_readahead, bool for_write);
4258 struct decompress_io_ctx *f2fs_alloc_dic(struct compress_ctx *cc);
4259 void f2fs_decompress_end_io(struct decompress_io_ctx *dic, bool failed,
4261 void f2fs_put_page_dic(struct page *page, bool in_task);
4262 unsigned int f2fs_cluster_blocks_are_contiguous(struct dnode_of_data *dn,
4263 unsigned int ofs_in_node);
4264 int f2fs_init_compress_ctx(struct compress_ctx *cc);
4265 void f2fs_destroy_compress_ctx(struct compress_ctx *cc, bool reuse);
4266 void f2fs_init_compress_info(struct f2fs_sb_info *sbi);
4267 int f2fs_init_compress_inode(struct f2fs_sb_info *sbi);
4268 void f2fs_destroy_compress_inode(struct f2fs_sb_info *sbi);
4269 int f2fs_init_page_array_cache(struct f2fs_sb_info *sbi);
4270 void f2fs_destroy_page_array_cache(struct f2fs_sb_info *sbi);
4271 int __init f2fs_init_compress_cache(void);
4272 void f2fs_destroy_compress_cache(void);
4273 struct address_space *COMPRESS_MAPPING(struct f2fs_sb_info *sbi);
4274 void f2fs_invalidate_compress_page(struct f2fs_sb_info *sbi, block_t blkaddr);
4275 void f2fs_cache_compressed_page(struct f2fs_sb_info *sbi, struct page *page,
4276 nid_t ino, block_t blkaddr);
4277 bool f2fs_load_compressed_page(struct f2fs_sb_info *sbi, struct page *page,
4279 void f2fs_invalidate_compress_pages(struct f2fs_sb_info *sbi, nid_t ino);
4280 #define inc_compr_inode_stat(inode) \
4282 struct f2fs_sb_info *sbi = F2FS_I_SB(inode); \
4283 sbi->compr_new_inode++; \
4285 #define add_compr_block_stat(inode, blocks) \
4287 struct f2fs_sb_info *sbi = F2FS_I_SB(inode); \
4288 int diff = F2FS_I(inode)->i_cluster_size - blocks; \
4289 sbi->compr_written_block += blocks; \
4290 sbi->compr_saved_block += diff; \
4293 static inline bool f2fs_is_compressed_page(struct page *page) { return false; }
4294 static inline bool f2fs_is_compress_backend_ready(struct inode *inode)
4296 if (!f2fs_compressed_file(inode))
4298 /* not support compression */
4301 static inline bool f2fs_is_compress_level_valid(int alg, int lvl) { return false; }
4302 static inline struct page *f2fs_compress_control_page(struct page *page)
4305 return ERR_PTR(-EINVAL);
4307 static inline int f2fs_init_compress_mempool(void) { return 0; }
4308 static inline void f2fs_destroy_compress_mempool(void) { }
4309 static inline void f2fs_decompress_cluster(struct decompress_io_ctx *dic,
4311 static inline void f2fs_end_read_compressed_page(struct page *page,
4312 bool failed, block_t blkaddr, bool in_task)
4316 static inline void f2fs_put_page_dic(struct page *page, bool in_task)
4320 static inline unsigned int f2fs_cluster_blocks_are_contiguous(
4321 struct dnode_of_data *dn, unsigned int ofs_in_node) { return 0; }
4322 static inline bool f2fs_sanity_check_cluster(struct dnode_of_data *dn) { return false; }
4323 static inline int f2fs_init_compress_inode(struct f2fs_sb_info *sbi) { return 0; }
4324 static inline void f2fs_destroy_compress_inode(struct f2fs_sb_info *sbi) { }
4325 static inline int f2fs_init_page_array_cache(struct f2fs_sb_info *sbi) { return 0; }
4326 static inline void f2fs_destroy_page_array_cache(struct f2fs_sb_info *sbi) { }
4327 static inline int __init f2fs_init_compress_cache(void) { return 0; }
4328 static inline void f2fs_destroy_compress_cache(void) { }
4329 static inline void f2fs_invalidate_compress_page(struct f2fs_sb_info *sbi,
4330 block_t blkaddr) { }
4331 static inline void f2fs_cache_compressed_page(struct f2fs_sb_info *sbi,
4332 struct page *page, nid_t ino, block_t blkaddr) { }
4333 static inline bool f2fs_load_compressed_page(struct f2fs_sb_info *sbi,
4334 struct page *page, block_t blkaddr) { return false; }
4335 static inline void f2fs_invalidate_compress_pages(struct f2fs_sb_info *sbi,
4337 #define inc_compr_inode_stat(inode) do { } while (0)
4338 static inline void f2fs_update_read_extent_tree_range_compressed(
4339 struct inode *inode,
4340 pgoff_t fofs, block_t blkaddr,
4341 unsigned int llen, unsigned int c_len) { }
4344 static inline int set_compress_context(struct inode *inode)
4346 #ifdef CONFIG_F2FS_FS_COMPRESSION
4347 struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
4349 F2FS_I(inode)->i_compress_algorithm =
4350 F2FS_OPTION(sbi).compress_algorithm;
4351 F2FS_I(inode)->i_log_cluster_size =
4352 F2FS_OPTION(sbi).compress_log_size;
4353 F2FS_I(inode)->i_compress_flag =
4354 F2FS_OPTION(sbi).compress_chksum ?
4355 BIT(COMPRESS_CHKSUM) : 0;
4356 F2FS_I(inode)->i_cluster_size =
4357 BIT(F2FS_I(inode)->i_log_cluster_size);
4358 if ((F2FS_I(inode)->i_compress_algorithm == COMPRESS_LZ4 ||
4359 F2FS_I(inode)->i_compress_algorithm == COMPRESS_ZSTD) &&
4360 F2FS_OPTION(sbi).compress_level)
4361 F2FS_I(inode)->i_compress_level =
4362 F2FS_OPTION(sbi).compress_level;
4363 F2FS_I(inode)->i_flags |= F2FS_COMPR_FL;
4364 set_inode_flag(inode, FI_COMPRESSED_FILE);
4365 stat_inc_compr_inode(inode);
4366 inc_compr_inode_stat(inode);
4367 f2fs_mark_inode_dirty_sync(inode, true);
4374 static inline bool f2fs_disable_compressed_file(struct inode *inode)
4376 struct f2fs_inode_info *fi = F2FS_I(inode);
4378 f2fs_down_write(&F2FS_I(inode)->i_sem);
4380 if (!f2fs_compressed_file(inode)) {
4381 f2fs_up_write(&F2FS_I(inode)->i_sem);
4384 if (f2fs_is_mmap_file(inode) ||
4385 (S_ISREG(inode->i_mode) && F2FS_HAS_BLOCKS(inode))) {
4386 f2fs_up_write(&F2FS_I(inode)->i_sem);
4390 fi->i_flags &= ~F2FS_COMPR_FL;
4391 stat_dec_compr_inode(inode);
4392 clear_inode_flag(inode, FI_COMPRESSED_FILE);
4393 f2fs_mark_inode_dirty_sync(inode, true);
4395 f2fs_up_write(&F2FS_I(inode)->i_sem);
4399 #define F2FS_FEATURE_FUNCS(name, flagname) \
4400 static inline int f2fs_sb_has_##name(struct f2fs_sb_info *sbi) \
4402 return F2FS_HAS_FEATURE(sbi, F2FS_FEATURE_##flagname); \
4405 F2FS_FEATURE_FUNCS(encrypt, ENCRYPT);
4406 F2FS_FEATURE_FUNCS(blkzoned, BLKZONED);
4407 F2FS_FEATURE_FUNCS(extra_attr, EXTRA_ATTR);
4408 F2FS_FEATURE_FUNCS(project_quota, PRJQUOTA);
4409 F2FS_FEATURE_FUNCS(inode_chksum, INODE_CHKSUM);
4410 F2FS_FEATURE_FUNCS(flexible_inline_xattr, FLEXIBLE_INLINE_XATTR);
4411 F2FS_FEATURE_FUNCS(quota_ino, QUOTA_INO);
4412 F2FS_FEATURE_FUNCS(inode_crtime, INODE_CRTIME);
4413 F2FS_FEATURE_FUNCS(lost_found, LOST_FOUND);
4414 F2FS_FEATURE_FUNCS(verity, VERITY);
4415 F2FS_FEATURE_FUNCS(sb_chksum, SB_CHKSUM);
4416 F2FS_FEATURE_FUNCS(casefold, CASEFOLD);
4417 F2FS_FEATURE_FUNCS(compression, COMPRESSION);
4418 F2FS_FEATURE_FUNCS(readonly, RO);
4420 #ifdef CONFIG_BLK_DEV_ZONED
4421 static inline bool f2fs_blkz_is_seq(struct f2fs_sb_info *sbi, int devi,
4424 unsigned int zno = blkaddr >> sbi->log_blocks_per_blkz;
4426 return test_bit(zno, FDEV(devi).blkz_seq);
4430 static inline bool f2fs_hw_should_discard(struct f2fs_sb_info *sbi)
4432 return f2fs_sb_has_blkzoned(sbi);
4435 static inline bool f2fs_bdev_support_discard(struct block_device *bdev)
4437 return bdev_max_discard_sectors(bdev) || bdev_is_zoned(bdev);
4440 static inline bool f2fs_hw_support_discard(struct f2fs_sb_info *sbi)
4444 if (!f2fs_is_multi_device(sbi))
4445 return f2fs_bdev_support_discard(sbi->sb->s_bdev);
4447 for (i = 0; i < sbi->s_ndevs; i++)
4448 if (f2fs_bdev_support_discard(FDEV(i).bdev))
4453 static inline bool f2fs_realtime_discard_enable(struct f2fs_sb_info *sbi)
4455 return (test_opt(sbi, DISCARD) && f2fs_hw_support_discard(sbi)) ||
4456 f2fs_hw_should_discard(sbi);
4459 static inline bool f2fs_hw_is_readonly(struct f2fs_sb_info *sbi)
4463 if (!f2fs_is_multi_device(sbi))
4464 return bdev_read_only(sbi->sb->s_bdev);
4466 for (i = 0; i < sbi->s_ndevs; i++)
4467 if (bdev_read_only(FDEV(i).bdev))
4472 static inline bool f2fs_dev_is_readonly(struct f2fs_sb_info *sbi)
4474 return f2fs_sb_has_readonly(sbi) || f2fs_hw_is_readonly(sbi);
4477 static inline bool f2fs_lfs_mode(struct f2fs_sb_info *sbi)
4479 return F2FS_OPTION(sbi).fs_mode == FS_MODE_LFS;
4482 static inline bool f2fs_low_mem_mode(struct f2fs_sb_info *sbi)
4484 return F2FS_OPTION(sbi).memory_mode == MEMORY_MODE_LOW;
4487 static inline bool f2fs_may_compress(struct inode *inode)
4489 if (IS_SWAPFILE(inode) || f2fs_is_pinned_file(inode) ||
4490 f2fs_is_atomic_file(inode) || f2fs_has_inline_data(inode) ||
4491 f2fs_is_mmap_file(inode))
4493 return S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode);
4496 static inline void f2fs_i_compr_blocks_update(struct inode *inode,
4497 u64 blocks, bool add)
4499 struct f2fs_inode_info *fi = F2FS_I(inode);
4500 int diff = fi->i_cluster_size - blocks;
4502 /* don't update i_compr_blocks if saved blocks were released */
4503 if (!add && !atomic_read(&fi->i_compr_blocks))
4507 atomic_add(diff, &fi->i_compr_blocks);
4508 stat_add_compr_blocks(inode, diff);
4510 atomic_sub(diff, &fi->i_compr_blocks);
4511 stat_sub_compr_blocks(inode, diff);
4513 f2fs_mark_inode_dirty_sync(inode, true);
4516 static inline bool f2fs_allow_multi_device_dio(struct f2fs_sb_info *sbi,
4519 if (!f2fs_is_multi_device(sbi))
4521 if (flag != F2FS_GET_BLOCK_DIO)
4523 return sbi->aligned_blksize;
4526 static inline bool f2fs_need_verity(const struct inode *inode, pgoff_t idx)
4528 return fsverity_active(inode) &&
4529 idx < DIV_ROUND_UP(inode->i_size, PAGE_SIZE);
4532 #ifdef CONFIG_F2FS_FAULT_INJECTION
4533 extern void f2fs_build_fault_attr(struct f2fs_sb_info *sbi, unsigned int rate,
4536 #define f2fs_build_fault_attr(sbi, rate, type) do { } while (0)
4539 static inline bool is_journalled_quota(struct f2fs_sb_info *sbi)
4542 if (f2fs_sb_has_quota_ino(sbi))
4544 if (F2FS_OPTION(sbi).s_qf_names[USRQUOTA] ||
4545 F2FS_OPTION(sbi).s_qf_names[GRPQUOTA] ||
4546 F2FS_OPTION(sbi).s_qf_names[PRJQUOTA])
4552 static inline bool f2fs_block_unit_discard(struct f2fs_sb_info *sbi)
4554 return F2FS_OPTION(sbi).discard_unit == DISCARD_UNIT_BLOCK;
4557 static inline void f2fs_io_schedule_timeout(long timeout)
4559 set_current_state(TASK_UNINTERRUPTIBLE);
4560 io_schedule_timeout(timeout);
4563 static inline void f2fs_handle_page_eio(struct f2fs_sb_info *sbi, pgoff_t ofs,
4564 enum page_type type)
4566 if (unlikely(f2fs_cp_error(sbi)))
4569 if (ofs == sbi->page_eio_ofs[type]) {
4570 if (sbi->page_eio_cnt[type]++ == MAX_RETRY_PAGE_EIO)
4571 set_ckpt_flags(sbi, CP_ERROR_FLAG);
4573 sbi->page_eio_ofs[type] = ofs;
4574 sbi->page_eio_cnt[type] = 0;
4578 #define EFSBADCRC EBADMSG /* Bad CRC detected */
4579 #define EFSCORRUPTED EUCLEAN /* Filesystem is corrupted */
4581 #endif /* _LINUX_F2FS_H */