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 */
67 #ifdef CONFIG_F2FS_FAULT_INJECTION
68 #define F2FS_ALL_FAULT_TYPE (GENMASK(FAULT_MAX - 1, 0))
70 struct f2fs_fault_info {
72 unsigned int inject_rate;
73 unsigned int inject_type;
76 extern const char *f2fs_fault_name[FAULT_MAX];
77 #define IS_FAULT_SET(fi, type) ((fi)->inject_type & BIT(type))
83 #define F2FS_MOUNT_DISABLE_ROLL_FORWARD 0x00000002
84 #define F2FS_MOUNT_DISCARD 0x00000004
85 #define F2FS_MOUNT_NOHEAP 0x00000008
86 #define F2FS_MOUNT_XATTR_USER 0x00000010
87 #define F2FS_MOUNT_POSIX_ACL 0x00000020
88 #define F2FS_MOUNT_DISABLE_EXT_IDENTIFY 0x00000040
89 #define F2FS_MOUNT_INLINE_XATTR 0x00000080
90 #define F2FS_MOUNT_INLINE_DATA 0x00000100
91 #define F2FS_MOUNT_INLINE_DENTRY 0x00000200
92 #define F2FS_MOUNT_FLUSH_MERGE 0x00000400
93 #define F2FS_MOUNT_NOBARRIER 0x00000800
94 #define F2FS_MOUNT_FASTBOOT 0x00001000
95 #define F2FS_MOUNT_READ_EXTENT_CACHE 0x00002000
96 #define F2FS_MOUNT_DATA_FLUSH 0x00008000
97 #define F2FS_MOUNT_FAULT_INJECTION 0x00010000
98 #define F2FS_MOUNT_USRQUOTA 0x00080000
99 #define F2FS_MOUNT_GRPQUOTA 0x00100000
100 #define F2FS_MOUNT_PRJQUOTA 0x00200000
101 #define F2FS_MOUNT_QUOTA 0x00400000
102 #define F2FS_MOUNT_INLINE_XATTR_SIZE 0x00800000
103 #define F2FS_MOUNT_RESERVE_ROOT 0x01000000
104 #define F2FS_MOUNT_DISABLE_CHECKPOINT 0x02000000
105 #define F2FS_MOUNT_NORECOVERY 0x04000000
106 #define F2FS_MOUNT_ATGC 0x08000000
107 #define F2FS_MOUNT_MERGE_CHECKPOINT 0x10000000
108 #define F2FS_MOUNT_GC_MERGE 0x20000000
109 #define F2FS_MOUNT_COMPRESS_CACHE 0x40000000
110 #define F2FS_MOUNT_AGE_EXTENT_CACHE 0x80000000
112 #define F2FS_OPTION(sbi) ((sbi)->mount_opt)
113 #define clear_opt(sbi, option) (F2FS_OPTION(sbi).opt &= ~F2FS_MOUNT_##option)
114 #define set_opt(sbi, option) (F2FS_OPTION(sbi).opt |= F2FS_MOUNT_##option)
115 #define test_opt(sbi, option) (F2FS_OPTION(sbi).opt & F2FS_MOUNT_##option)
117 #define ver_after(a, b) (typecheck(unsigned long long, a) && \
118 typecheck(unsigned long long, b) && \
119 ((long long)((a) - (b)) > 0))
121 typedef u32 block_t; /*
122 * should not change u32, since it is the on-disk block
123 * address format, __le32.
127 #define COMPRESS_EXT_NUM 16
130 * An implementation of an rwsem that is explicitly unfair to readers. This
131 * prevents priority inversion when a low-priority reader acquires the read lock
132 * while sleeping on the write lock but the write lock is needed by
133 * higher-priority clients.
137 struct rw_semaphore internal_rwsem;
138 #ifdef CONFIG_F2FS_UNFAIR_RWSEM
139 wait_queue_head_t read_waiters;
143 struct f2fs_mount_info {
145 int write_io_size_bits; /* Write IO size bits */
146 block_t root_reserved_blocks; /* root reserved blocks */
147 kuid_t s_resuid; /* reserved blocks for uid */
148 kgid_t s_resgid; /* reserved blocks for gid */
149 int active_logs; /* # of active logs */
150 int inline_xattr_size; /* inline xattr size */
151 #ifdef CONFIG_F2FS_FAULT_INJECTION
152 struct f2fs_fault_info fault_info; /* For fault injection */
155 /* Names of quota files with journalled quota */
156 char *s_qf_names[MAXQUOTAS];
157 int s_jquota_fmt; /* Format of quota to use */
159 /* For which write hints are passed down to block layer */
160 int alloc_mode; /* segment allocation policy */
161 int fsync_mode; /* fsync policy */
162 int fs_mode; /* fs mode: LFS or ADAPTIVE */
163 int bggc_mode; /* bggc mode: off, on or sync */
164 int memory_mode; /* memory mode */
166 * discard command's offset/size should
167 * be aligned to this unit: block,
170 struct fscrypt_dummy_policy dummy_enc_policy; /* test dummy encryption */
171 block_t unusable_cap_perc; /* percentage for cap */
172 block_t unusable_cap; /* Amount of space allowed to be
173 * unusable when disabling checkpoint
176 /* For compression */
177 unsigned char compress_algorithm; /* algorithm type */
178 unsigned char compress_log_size; /* cluster log size */
179 unsigned char compress_level; /* compress level */
180 bool compress_chksum; /* compressed data chksum */
181 unsigned char compress_ext_cnt; /* extension count */
182 unsigned char nocompress_ext_cnt; /* nocompress extension count */
183 int compress_mode; /* compression mode */
184 unsigned char extensions[COMPRESS_EXT_NUM][F2FS_EXTENSION_LEN]; /* extensions */
185 unsigned char noextensions[COMPRESS_EXT_NUM][F2FS_EXTENSION_LEN]; /* extensions */
188 #define F2FS_FEATURE_ENCRYPT 0x0001
189 #define F2FS_FEATURE_BLKZONED 0x0002
190 #define F2FS_FEATURE_ATOMIC_WRITE 0x0004
191 #define F2FS_FEATURE_EXTRA_ATTR 0x0008
192 #define F2FS_FEATURE_PRJQUOTA 0x0010
193 #define F2FS_FEATURE_INODE_CHKSUM 0x0020
194 #define F2FS_FEATURE_FLEXIBLE_INLINE_XATTR 0x0040
195 #define F2FS_FEATURE_QUOTA_INO 0x0080
196 #define F2FS_FEATURE_INODE_CRTIME 0x0100
197 #define F2FS_FEATURE_LOST_FOUND 0x0200
198 #define F2FS_FEATURE_VERITY 0x0400
199 #define F2FS_FEATURE_SB_CHKSUM 0x0800
200 #define F2FS_FEATURE_CASEFOLD 0x1000
201 #define F2FS_FEATURE_COMPRESSION 0x2000
202 #define F2FS_FEATURE_RO 0x4000
204 #define __F2FS_HAS_FEATURE(raw_super, mask) \
205 ((raw_super->feature & cpu_to_le32(mask)) != 0)
206 #define F2FS_HAS_FEATURE(sbi, mask) __F2FS_HAS_FEATURE(sbi->raw_super, mask)
209 * Default values for user and/or group using reserved blocks
211 #define F2FS_DEF_RESUID 0
212 #define F2FS_DEF_RESGID 0
215 * For checkpoint manager
222 #define CP_UMOUNT 0x00000001
223 #define CP_FASTBOOT 0x00000002
224 #define CP_SYNC 0x00000004
225 #define CP_RECOVERY 0x00000008
226 #define CP_DISCARD 0x00000010
227 #define CP_TRIMMED 0x00000020
228 #define CP_PAUSE 0x00000040
229 #define CP_RESIZE 0x00000080
231 #define DEF_MAX_DISCARD_REQUEST 8 /* issue 8 discards per round */
232 #define DEF_MIN_DISCARD_ISSUE_TIME 50 /* 50 ms, if exists */
233 #define DEF_MID_DISCARD_ISSUE_TIME 500 /* 500 ms, if device busy */
234 #define DEF_MAX_DISCARD_ISSUE_TIME 60000 /* 60 s, if no candidates */
235 #define DEF_DISCARD_URGENT_UTIL 80 /* do more discard over 80% */
236 #define DEF_CP_INTERVAL 60 /* 60 secs */
237 #define DEF_IDLE_INTERVAL 5 /* 5 secs */
238 #define DEF_DISABLE_INTERVAL 5 /* 5 secs */
239 #define DEF_DISABLE_QUICK_INTERVAL 1 /* 1 secs */
240 #define DEF_UMOUNT_DISCARD_TIMEOUT 5 /* 5 secs */
250 * indicate meta/data type
259 DATA_GENERIC, /* check range only */
260 DATA_GENERIC_ENHANCE, /* strong check on range and segment bitmap */
261 DATA_GENERIC_ENHANCE_READ, /*
262 * strong check on range and segment
263 * bitmap but no warning due to race
264 * condition of read on truncated area
267 DATA_GENERIC_ENHANCE_UPDATE, /*
268 * strong check on range and segment
269 * bitmap for update case
274 /* for the list of ino */
276 ORPHAN_INO, /* for orphan ino list */
277 APPEND_INO, /* for append ino list */
278 UPDATE_INO, /* for update ino list */
279 TRANS_DIR_INO, /* for transactions dir ino list */
280 FLUSH_INO, /* for multiple device flushing */
281 MAX_INO_ENTRY, /* max. list */
285 struct list_head list; /* list head */
286 nid_t ino; /* inode number */
287 unsigned int dirty_device; /* dirty device bitmap */
290 /* for the list of inodes to be GCed */
292 struct list_head list; /* list head */
293 struct inode *inode; /* vfs inode pointer */
296 struct fsync_node_entry {
297 struct list_head list; /* list head */
298 struct page *page; /* warm node page pointer */
299 unsigned int seq_id; /* sequence id */
303 struct completion wait; /* completion for checkpoint done */
304 struct llist_node llnode; /* llist_node to be linked in wait queue */
305 int ret; /* return code of checkpoint */
306 ktime_t queue_time; /* request queued time */
309 struct ckpt_req_control {
310 struct task_struct *f2fs_issue_ckpt; /* checkpoint task */
311 int ckpt_thread_ioprio; /* checkpoint merge thread ioprio */
312 wait_queue_head_t ckpt_wait_queue; /* waiting queue for wake-up */
313 atomic_t issued_ckpt; /* # of actually issued ckpts */
314 atomic_t total_ckpt; /* # of total ckpts */
315 atomic_t queued_ckpt; /* # of queued ckpts */
316 struct llist_head issue_list; /* list for command issue */
317 spinlock_t stat_lock; /* lock for below checkpoint time stats */
318 unsigned int cur_time; /* cur wait time in msec for currently issued checkpoint */
319 unsigned int peak_time; /* peak wait time in msec until now */
322 /* for the bitmap indicate blocks to be discarded */
323 struct discard_entry {
324 struct list_head list; /* list head */
325 block_t start_blkaddr; /* start blockaddr of current segment */
326 unsigned char discard_map[SIT_VBLOCK_MAP_SIZE]; /* segment discard bitmap */
329 /* minimum discard granularity, unit: block count */
330 #define MIN_DISCARD_GRANULARITY 1
331 /* default discard granularity of inner discard thread, unit: block count */
332 #define DEFAULT_DISCARD_GRANULARITY 16
333 /* default maximum discard granularity of ordered discard, unit: block count */
334 #define DEFAULT_MAX_ORDERED_DISCARD_GRANULARITY 16
336 /* max discard pend list number */
337 #define MAX_PLIST_NUM 512
338 #define plist_idx(blk_num) ((blk_num) >= MAX_PLIST_NUM ? \
339 (MAX_PLIST_NUM - 1) : ((blk_num) - 1))
342 D_PREP, /* initial */
343 D_PARTIAL, /* partially submitted */
344 D_SUBMIT, /* all submitted */
345 D_DONE, /* finished */
348 struct discard_info {
349 block_t lstart; /* logical start address */
350 block_t len; /* length */
351 block_t start; /* actual start address in dev */
355 struct rb_node rb_node; /* rb node located in rb-tree */
356 struct discard_info di; /* discard info */
357 struct list_head list; /* command list */
358 struct completion wait; /* compleation */
359 struct block_device *bdev; /* bdev */
360 unsigned short ref; /* reference count */
361 unsigned char state; /* state */
362 unsigned char queued; /* queued discard */
363 int error; /* bio error */
364 spinlock_t lock; /* for state/bio_ref updating */
365 unsigned short bio_ref; /* bio reference count */
376 struct discard_policy {
377 int type; /* type of discard */
378 unsigned int min_interval; /* used for candidates exist */
379 unsigned int mid_interval; /* used for device busy */
380 unsigned int max_interval; /* used for candidates not exist */
381 unsigned int max_requests; /* # of discards issued per round */
382 unsigned int io_aware_gran; /* minimum granularity discard not be aware of I/O */
383 bool io_aware; /* issue discard in idle time */
384 bool sync; /* submit discard with REQ_SYNC flag */
385 bool ordered; /* issue discard by lba order */
386 bool timeout; /* discard timeout for put_super */
387 unsigned int granularity; /* discard granularity */
390 struct discard_cmd_control {
391 struct task_struct *f2fs_issue_discard; /* discard thread */
392 struct list_head entry_list; /* 4KB discard entry list */
393 struct list_head pend_list[MAX_PLIST_NUM];/* store pending entries */
394 struct list_head wait_list; /* store on-flushing entries */
395 struct list_head fstrim_list; /* in-flight discard from fstrim */
396 wait_queue_head_t discard_wait_queue; /* waiting queue for wake-up */
397 struct mutex cmd_lock;
398 unsigned int nr_discards; /* # of discards in the list */
399 unsigned int max_discards; /* max. discards to be issued */
400 unsigned int max_discard_request; /* max. discard request per round */
401 unsigned int min_discard_issue_time; /* min. interval between discard issue */
402 unsigned int mid_discard_issue_time; /* mid. interval between discard issue */
403 unsigned int max_discard_issue_time; /* max. interval between discard issue */
404 unsigned int discard_io_aware_gran; /* minimum discard granularity not be aware of I/O */
405 unsigned int discard_urgent_util; /* utilization which issue discard proactively */
406 unsigned int discard_granularity; /* discard granularity */
407 unsigned int max_ordered_discard; /* maximum discard granularity issued by lba order */
408 unsigned int undiscard_blks; /* # of undiscard blocks */
409 unsigned int next_pos; /* next discard position */
410 atomic_t issued_discard; /* # of issued discard */
411 atomic_t queued_discard; /* # of queued discard */
412 atomic_t discard_cmd_cnt; /* # of cached cmd count */
413 struct rb_root_cached root; /* root of discard rb-tree */
414 bool rbtree_check; /* config for consistence check */
415 bool discard_wake; /* to wake up discard thread */
418 /* for the list of fsync inodes, used only during recovery */
419 struct fsync_inode_entry {
420 struct list_head list; /* list head */
421 struct inode *inode; /* vfs inode pointer */
422 block_t blkaddr; /* block address locating the last fsync */
423 block_t last_dentry; /* block address locating the last dentry */
426 #define nats_in_cursum(jnl) (le16_to_cpu((jnl)->n_nats))
427 #define sits_in_cursum(jnl) (le16_to_cpu((jnl)->n_sits))
429 #define nat_in_journal(jnl, i) ((jnl)->nat_j.entries[i].ne)
430 #define nid_in_journal(jnl, i) ((jnl)->nat_j.entries[i].nid)
431 #define sit_in_journal(jnl, i) ((jnl)->sit_j.entries[i].se)
432 #define segno_in_journal(jnl, i) ((jnl)->sit_j.entries[i].segno)
434 #define MAX_NAT_JENTRIES(jnl) (NAT_JOURNAL_ENTRIES - nats_in_cursum(jnl))
435 #define MAX_SIT_JENTRIES(jnl) (SIT_JOURNAL_ENTRIES - sits_in_cursum(jnl))
437 static inline int update_nats_in_cursum(struct f2fs_journal *journal, int i)
439 int before = nats_in_cursum(journal);
441 journal->n_nats = cpu_to_le16(before + i);
445 static inline int update_sits_in_cursum(struct f2fs_journal *journal, int i)
447 int before = sits_in_cursum(journal);
449 journal->n_sits = cpu_to_le16(before + i);
453 static inline bool __has_cursum_space(struct f2fs_journal *journal,
456 if (type == NAT_JOURNAL)
457 return size <= MAX_NAT_JENTRIES(journal);
458 return size <= MAX_SIT_JENTRIES(journal);
461 /* for inline stuff */
462 #define DEF_INLINE_RESERVED_SIZE 1
463 static inline int get_extra_isize(struct inode *inode);
464 static inline int get_inline_xattr_addrs(struct inode *inode);
465 #define MAX_INLINE_DATA(inode) (sizeof(__le32) * \
466 (CUR_ADDRS_PER_INODE(inode) - \
467 get_inline_xattr_addrs(inode) - \
468 DEF_INLINE_RESERVED_SIZE))
471 #define NR_INLINE_DENTRY(inode) (MAX_INLINE_DATA(inode) * BITS_PER_BYTE / \
472 ((SIZE_OF_DIR_ENTRY + F2FS_SLOT_LEN) * \
474 #define INLINE_DENTRY_BITMAP_SIZE(inode) \
475 DIV_ROUND_UP(NR_INLINE_DENTRY(inode), BITS_PER_BYTE)
476 #define INLINE_RESERVED_SIZE(inode) (MAX_INLINE_DATA(inode) - \
477 ((SIZE_OF_DIR_ENTRY + F2FS_SLOT_LEN) * \
478 NR_INLINE_DENTRY(inode) + \
479 INLINE_DENTRY_BITMAP_SIZE(inode)))
482 * For INODE and NODE manager
484 /* for directory operations */
486 struct f2fs_filename {
488 * The filename the user specified. This is NULL for some
489 * filesystem-internal operations, e.g. converting an inline directory
490 * to a non-inline one, or roll-forward recovering an encrypted dentry.
492 const struct qstr *usr_fname;
495 * The on-disk filename. For encrypted directories, this is encrypted.
496 * This may be NULL for lookups in an encrypted dir without the key.
498 struct fscrypt_str disk_name;
500 /* The dirhash of this filename */
503 #ifdef CONFIG_FS_ENCRYPTION
505 * For lookups in encrypted directories: either the buffer backing
506 * disk_name, or a buffer that holds the decoded no-key name.
508 struct fscrypt_str crypto_buf;
510 #if IS_ENABLED(CONFIG_UNICODE)
512 * For casefolded directories: the casefolded name, but it's left NULL
513 * if the original name is not valid Unicode, if the original name is
514 * "." or "..", if the directory is both casefolded and encrypted and
515 * its encryption key is unavailable, or if the filesystem is doing an
516 * internal operation where usr_fname is also NULL. In all these cases
517 * we fall back to treating the name as an opaque byte sequence.
519 struct fscrypt_str cf_name;
523 struct f2fs_dentry_ptr {
526 struct f2fs_dir_entry *dentry;
527 __u8 (*filename)[F2FS_SLOT_LEN];
532 static inline void make_dentry_ptr_block(struct inode *inode,
533 struct f2fs_dentry_ptr *d, struct f2fs_dentry_block *t)
536 d->max = NR_DENTRY_IN_BLOCK;
537 d->nr_bitmap = SIZE_OF_DENTRY_BITMAP;
538 d->bitmap = t->dentry_bitmap;
539 d->dentry = t->dentry;
540 d->filename = t->filename;
543 static inline void make_dentry_ptr_inline(struct inode *inode,
544 struct f2fs_dentry_ptr *d, void *t)
546 int entry_cnt = NR_INLINE_DENTRY(inode);
547 int bitmap_size = INLINE_DENTRY_BITMAP_SIZE(inode);
548 int reserved_size = INLINE_RESERVED_SIZE(inode);
552 d->nr_bitmap = bitmap_size;
554 d->dentry = t + bitmap_size + reserved_size;
555 d->filename = t + bitmap_size + reserved_size +
556 SIZE_OF_DIR_ENTRY * entry_cnt;
560 * XATTR_NODE_OFFSET stores xattrs to one node block per file keeping -1
561 * as its node offset to distinguish from index node blocks.
562 * But some bits are used to mark the node block.
564 #define XATTR_NODE_OFFSET ((((unsigned int)-1) << OFFSET_BIT_SHIFT) \
567 ALLOC_NODE, /* allocate a new node page if needed */
568 LOOKUP_NODE, /* look up a node without readahead */
570 * look up a node with readahead called
575 #define DEFAULT_RETRY_IO_COUNT 8 /* maximum retry read IO or flush count */
577 /* congestion wait timeout value, default: 20ms */
578 #define DEFAULT_IO_TIMEOUT (msecs_to_jiffies(20))
580 /* maximum retry quota flush count */
581 #define DEFAULT_RETRY_QUOTA_FLUSH_COUNT 8
583 /* maximum retry of EIO'ed page */
584 #define MAX_RETRY_PAGE_EIO 100
586 #define F2FS_LINK_MAX 0xffffffff /* maximum link count per file */
588 #define MAX_DIR_RA_PAGES 4 /* maximum ra pages of dir */
590 /* dirty segments threshold for triggering CP */
591 #define DEFAULT_DIRTY_THRESHOLD 4
593 #define RECOVERY_MAX_RA_BLOCKS BIO_MAX_VECS
594 #define RECOVERY_MIN_RA_BLOCKS 1
596 #define F2FS_ONSTACK_PAGES 16 /* nr of onstack pages */
598 /* for in-memory extent cache entry */
599 #define F2FS_MIN_EXTENT_LEN 64 /* minimum extent length */
601 /* number of extent info in extent cache we try to shrink */
602 #define READ_EXTENT_CACHE_SHRINK_NUMBER 128
604 /* number of age extent info in extent cache we try to shrink */
605 #define AGE_EXTENT_CACHE_SHRINK_NUMBER 128
606 #define LAST_AGE_WEIGHT 30
607 #define SAME_AGE_REGION 1024
610 * Define data block with age less than 1GB as hot data
611 * define data block with age less than 10GB but more than 1GB as warm data
613 #define DEF_HOT_DATA_AGE_THRESHOLD 262144
614 #define DEF_WARM_DATA_AGE_THRESHOLD 2621440
616 /* extent cache type */
624 unsigned int fofs; /* start offset in a file */
625 unsigned int len; /* length of the extent */
627 /* read extent_cache */
629 /* start block address of the extent */
631 #ifdef CONFIG_F2FS_FS_COMPRESSION
632 /* physical extent length of compressed blocks */
636 /* block age extent_cache */
638 /* block age of the extent */
639 unsigned long long age;
640 /* last total blocks allocated */
641 unsigned long long last_blocks;
647 struct rb_node rb_node; /* rb node located in rb-tree */
648 struct extent_info ei; /* extent info */
649 struct list_head list; /* node in global extent list of sbi */
650 struct extent_tree *et; /* extent tree pointer */
654 nid_t ino; /* inode number */
655 enum extent_type type; /* keep the extent tree type */
656 struct rb_root_cached root; /* root of extent info rb-tree */
657 struct extent_node *cached_en; /* recently accessed extent node */
658 struct list_head list; /* to be used by sbi->zombie_list */
659 rwlock_t lock; /* protect extent info rb-tree */
660 atomic_t node_cnt; /* # of extent node in rb-tree*/
661 bool largest_updated; /* largest extent updated */
662 struct extent_info largest; /* largest cached extent for EX_READ */
665 struct extent_tree_info {
666 struct radix_tree_root extent_tree_root;/* cache extent cache entries */
667 struct mutex extent_tree_lock; /* locking extent radix tree */
668 struct list_head extent_list; /* lru list for shrinker */
669 spinlock_t extent_lock; /* locking extent lru list */
670 atomic_t total_ext_tree; /* extent tree count */
671 struct list_head zombie_list; /* extent zombie tree list */
672 atomic_t total_zombie_tree; /* extent zombie tree count */
673 atomic_t total_ext_node; /* extent info count */
677 * State of block returned by f2fs_map_blocks.
679 #define F2FS_MAP_NEW (1U << 0)
680 #define F2FS_MAP_MAPPED (1U << 1)
681 #define F2FS_MAP_DELALLOC (1U << 2)
682 #define F2FS_MAP_FLAGS (F2FS_MAP_NEW | F2FS_MAP_MAPPED |\
685 struct f2fs_map_blocks {
686 struct block_device *m_bdev; /* for multi-device dio */
690 unsigned int m_flags;
691 pgoff_t *m_next_pgofs; /* point next possible non-hole pgofs */
692 pgoff_t *m_next_extent; /* point to next possible extent */
694 bool m_may_create; /* indicate it is from write path */
695 bool m_multidev_dio; /* indicate it allows multi-device dio */
698 /* for flag in get_data_block */
700 F2FS_GET_BLOCK_DEFAULT,
701 F2FS_GET_BLOCK_FIEMAP,
704 F2FS_GET_BLOCK_PRE_DIO,
705 F2FS_GET_BLOCK_PRE_AIO,
706 F2FS_GET_BLOCK_PRECACHE,
710 * i_advise uses FADVISE_XXX_BIT. We can add additional hints later.
712 #define FADVISE_COLD_BIT 0x01
713 #define FADVISE_LOST_PINO_BIT 0x02
714 #define FADVISE_ENCRYPT_BIT 0x04
715 #define FADVISE_ENC_NAME_BIT 0x08
716 #define FADVISE_KEEP_SIZE_BIT 0x10
717 #define FADVISE_HOT_BIT 0x20
718 #define FADVISE_VERITY_BIT 0x40
719 #define FADVISE_TRUNC_BIT 0x80
721 #define FADVISE_MODIFIABLE_BITS (FADVISE_COLD_BIT | FADVISE_HOT_BIT)
723 #define file_is_cold(inode) is_file(inode, FADVISE_COLD_BIT)
724 #define file_set_cold(inode) set_file(inode, FADVISE_COLD_BIT)
725 #define file_clear_cold(inode) clear_file(inode, FADVISE_COLD_BIT)
727 #define file_wrong_pino(inode) is_file(inode, FADVISE_LOST_PINO_BIT)
728 #define file_lost_pino(inode) set_file(inode, FADVISE_LOST_PINO_BIT)
729 #define file_got_pino(inode) clear_file(inode, FADVISE_LOST_PINO_BIT)
731 #define file_is_encrypt(inode) is_file(inode, FADVISE_ENCRYPT_BIT)
732 #define file_set_encrypt(inode) set_file(inode, FADVISE_ENCRYPT_BIT)
734 #define file_enc_name(inode) is_file(inode, FADVISE_ENC_NAME_BIT)
735 #define file_set_enc_name(inode) set_file(inode, FADVISE_ENC_NAME_BIT)
737 #define file_keep_isize(inode) is_file(inode, FADVISE_KEEP_SIZE_BIT)
738 #define file_set_keep_isize(inode) set_file(inode, FADVISE_KEEP_SIZE_BIT)
740 #define file_is_hot(inode) is_file(inode, FADVISE_HOT_BIT)
741 #define file_set_hot(inode) set_file(inode, FADVISE_HOT_BIT)
742 #define file_clear_hot(inode) clear_file(inode, FADVISE_HOT_BIT)
744 #define file_is_verity(inode) is_file(inode, FADVISE_VERITY_BIT)
745 #define file_set_verity(inode) set_file(inode, FADVISE_VERITY_BIT)
747 #define file_should_truncate(inode) is_file(inode, FADVISE_TRUNC_BIT)
748 #define file_need_truncate(inode) set_file(inode, FADVISE_TRUNC_BIT)
749 #define file_dont_truncate(inode) clear_file(inode, FADVISE_TRUNC_BIT)
751 #define DEF_DIR_LEVEL 0
758 /* used for f2fs_inode_info->flags */
760 FI_NEW_INODE, /* indicate newly allocated inode */
761 FI_DIRTY_INODE, /* indicate inode is dirty or not */
762 FI_AUTO_RECOVER, /* indicate inode is recoverable */
763 FI_DIRTY_DIR, /* indicate directory has dirty pages */
764 FI_INC_LINK, /* need to increment i_nlink */
765 FI_ACL_MODE, /* indicate acl mode */
766 FI_NO_ALLOC, /* should not allocate any blocks */
767 FI_FREE_NID, /* free allocated nide */
768 FI_NO_EXTENT, /* not to use the extent cache */
769 FI_INLINE_XATTR, /* used for inline xattr */
770 FI_INLINE_DATA, /* used for inline data*/
771 FI_INLINE_DENTRY, /* used for inline dentry */
772 FI_APPEND_WRITE, /* inode has appended data */
773 FI_UPDATE_WRITE, /* inode has in-place-update data */
774 FI_NEED_IPU, /* used for ipu per file */
775 FI_ATOMIC_FILE, /* indicate atomic file */
776 FI_FIRST_BLOCK_WRITTEN, /* indicate #0 data block was written */
777 FI_DROP_CACHE, /* drop dirty page cache */
778 FI_DATA_EXIST, /* indicate data exists */
779 FI_INLINE_DOTS, /* indicate inline dot dentries */
780 FI_SKIP_WRITES, /* should skip data page writeback */
781 FI_OPU_WRITE, /* used for opu per file */
782 FI_DIRTY_FILE, /* indicate regular/symlink has dirty pages */
783 FI_PREALLOCATED_ALL, /* all blocks for write were preallocated */
784 FI_HOT_DATA, /* indicate file is hot */
785 FI_EXTRA_ATTR, /* indicate file has extra attribute */
786 FI_PROJ_INHERIT, /* indicate file inherits projectid */
787 FI_PIN_FILE, /* indicate file should not be gced */
788 FI_VERITY_IN_PROGRESS, /* building fs-verity Merkle tree */
789 FI_COMPRESSED_FILE, /* indicate file's data can be compressed */
790 FI_COMPRESS_CORRUPT, /* indicate compressed cluster is corrupted */
791 FI_MMAP_FILE, /* indicate file was mmapped */
792 FI_ENABLE_COMPRESS, /* enable compression in "user" compression mode */
793 FI_COMPRESS_RELEASED, /* compressed blocks were released */
794 FI_ALIGNED_WRITE, /* enable aligned write */
795 FI_COW_FILE, /* indicate COW file */
796 FI_ATOMIC_COMMITTED, /* indicate atomic commit completed except disk sync */
797 FI_ATOMIC_REPLACE, /* indicate atomic replace */
798 FI_MAX, /* max flag, never be used */
801 struct f2fs_inode_info {
802 struct inode vfs_inode; /* serve a vfs inode */
803 unsigned long i_flags; /* keep an inode flags for ioctl */
804 unsigned char i_advise; /* use to give file attribute hints */
805 unsigned char i_dir_level; /* use for dentry level for large dir */
806 unsigned int i_current_depth; /* only for directory depth */
807 /* for gc failure statistic */
808 unsigned int i_gc_failures[MAX_GC_FAILURE];
809 unsigned int i_pino; /* parent inode number */
810 umode_t i_acl_mode; /* keep file acl mode temporarily */
812 /* Use below internally in f2fs*/
813 unsigned long flags[BITS_TO_LONGS(FI_MAX)]; /* use to pass per-file flags */
814 struct f2fs_rwsem i_sem; /* protect fi info */
815 atomic_t dirty_pages; /* # of dirty pages */
816 f2fs_hash_t chash; /* hash value of given file name */
817 unsigned int clevel; /* maximum level of given file name */
818 struct task_struct *task; /* lookup and create consistency */
819 struct task_struct *cp_task; /* separate cp/wb IO stats*/
820 struct task_struct *wb_task; /* indicate inode is in context of writeback */
821 nid_t i_xattr_nid; /* node id that contains xattrs */
822 loff_t last_disk_size; /* lastly written file size */
823 spinlock_t i_size_lock; /* protect last_disk_size */
826 struct dquot *i_dquot[MAXQUOTAS];
828 /* quota space reservation, managed internally by quota code */
829 qsize_t i_reserved_quota;
831 struct list_head dirty_list; /* dirty list for dirs and files */
832 struct list_head gdirty_list; /* linked in global dirty list */
833 struct task_struct *atomic_write_task; /* store atomic write task */
834 struct extent_tree *extent_tree[NR_EXTENT_CACHES];
835 /* cached extent_tree entry */
836 struct inode *cow_inode; /* copy-on-write inode for atomic write */
838 /* avoid racing between foreground op and gc */
839 struct f2fs_rwsem i_gc_rwsem[2];
840 struct f2fs_rwsem i_xattr_sem; /* avoid racing between reading and changing EAs */
842 int i_extra_isize; /* size of extra space located in i_addr */
843 kprojid_t i_projid; /* id for project quota */
844 int i_inline_xattr_size; /* inline xattr size */
845 struct timespec64 i_crtime; /* inode creation time */
846 struct timespec64 i_disk_time[3];/* inode disk times */
848 /* for file compress */
849 atomic_t i_compr_blocks; /* # of compressed blocks */
850 unsigned char i_compress_algorithm; /* algorithm type */
851 unsigned char i_log_cluster_size; /* log of cluster size */
852 unsigned char i_compress_level; /* compress level (lz4hc,zstd) */
853 unsigned char i_compress_flag; /* compress flag */
854 unsigned int i_cluster_size; /* cluster size */
856 unsigned int atomic_write_cnt;
857 loff_t original_i_size; /* original i_size before atomic write */
860 static inline void get_read_extent_info(struct extent_info *ext,
861 struct f2fs_extent *i_ext)
863 ext->fofs = le32_to_cpu(i_ext->fofs);
864 ext->blk = le32_to_cpu(i_ext->blk);
865 ext->len = le32_to_cpu(i_ext->len);
868 static inline void set_raw_read_extent(struct extent_info *ext,
869 struct f2fs_extent *i_ext)
871 i_ext->fofs = cpu_to_le32(ext->fofs);
872 i_ext->blk = cpu_to_le32(ext->blk);
873 i_ext->len = cpu_to_le32(ext->len);
876 static inline bool __is_discard_mergeable(struct discard_info *back,
877 struct discard_info *front, unsigned int max_len)
879 return (back->lstart + back->len == front->lstart) &&
880 (back->len + front->len <= max_len);
883 static inline bool __is_discard_back_mergeable(struct discard_info *cur,
884 struct discard_info *back, unsigned int max_len)
886 return __is_discard_mergeable(back, cur, max_len);
889 static inline bool __is_discard_front_mergeable(struct discard_info *cur,
890 struct discard_info *front, unsigned int max_len)
892 return __is_discard_mergeable(cur, front, max_len);
896 * For free nid management
899 FREE_NID, /* newly added to free nid list */
900 PREALLOC_NID, /* it is preallocated */
911 struct f2fs_nm_info {
912 block_t nat_blkaddr; /* base disk address of NAT */
913 nid_t max_nid; /* maximum possible node ids */
914 nid_t available_nids; /* # of available node ids */
915 nid_t next_scan_nid; /* the next nid to be scanned */
916 nid_t max_rf_node_blocks; /* max # of nodes for recovery */
917 unsigned int ram_thresh; /* control the memory footprint */
918 unsigned int ra_nid_pages; /* # of nid pages to be readaheaded */
919 unsigned int dirty_nats_ratio; /* control dirty nats ratio threshold */
921 /* NAT cache management */
922 struct radix_tree_root nat_root;/* root of the nat entry cache */
923 struct radix_tree_root nat_set_root;/* root of the nat set cache */
924 struct f2fs_rwsem nat_tree_lock; /* protect nat entry tree */
925 struct list_head nat_entries; /* cached nat entry list (clean) */
926 spinlock_t nat_list_lock; /* protect clean nat entry list */
927 unsigned int nat_cnt[MAX_NAT_STATE]; /* the # of cached nat entries */
928 unsigned int nat_blocks; /* # of nat blocks */
930 /* free node ids management */
931 struct radix_tree_root free_nid_root;/* root of the free_nid cache */
932 struct list_head free_nid_list; /* list for free nids excluding preallocated nids */
933 unsigned int nid_cnt[MAX_NID_STATE]; /* the number of free node id */
934 spinlock_t nid_list_lock; /* protect nid lists ops */
935 struct mutex build_lock; /* lock for build free nids */
936 unsigned char **free_nid_bitmap;
937 unsigned char *nat_block_bitmap;
938 unsigned short *free_nid_count; /* free nid count of NAT block */
941 char *nat_bitmap; /* NAT bitmap pointer */
943 unsigned int nat_bits_blocks; /* # of nat bits blocks */
944 unsigned char *nat_bits; /* NAT bits blocks */
945 unsigned char *full_nat_bits; /* full NAT pages */
946 unsigned char *empty_nat_bits; /* empty NAT pages */
947 #ifdef CONFIG_F2FS_CHECK_FS
948 char *nat_bitmap_mir; /* NAT bitmap mirror */
950 int bitmap_size; /* bitmap size */
954 * this structure is used as one of function parameters.
955 * all the information are dedicated to a given direct node block determined
956 * by the data offset in a file.
958 struct dnode_of_data {
959 struct inode *inode; /* vfs inode pointer */
960 struct page *inode_page; /* its inode page, NULL is possible */
961 struct page *node_page; /* cached direct node page */
962 nid_t nid; /* node id of the direct node block */
963 unsigned int ofs_in_node; /* data offset in the node page */
964 bool inode_page_locked; /* inode page is locked or not */
965 bool node_changed; /* is node block changed */
966 char cur_level; /* level of hole node page */
967 char max_level; /* level of current page located */
968 block_t data_blkaddr; /* block address of the node block */
971 static inline void set_new_dnode(struct dnode_of_data *dn, struct inode *inode,
972 struct page *ipage, struct page *npage, nid_t nid)
974 memset(dn, 0, sizeof(*dn));
976 dn->inode_page = ipage;
977 dn->node_page = npage;
984 * By default, there are 6 active log areas across the whole main area.
985 * When considering hot and cold data separation to reduce cleaning overhead,
986 * we split 3 for data logs and 3 for node logs as hot, warm, and cold types,
988 * In the current design, you should not change the numbers intentionally.
989 * Instead, as a mount option such as active_logs=x, you can use 2, 4, and 6
990 * logs individually according to the underlying devices. (default: 6)
991 * Just in case, on-disk layout covers maximum 16 logs that consist of 8 for
992 * data and 8 for node logs.
994 #define NR_CURSEG_DATA_TYPE (3)
995 #define NR_CURSEG_NODE_TYPE (3)
996 #define NR_CURSEG_INMEM_TYPE (2)
997 #define NR_CURSEG_RO_TYPE (2)
998 #define NR_CURSEG_PERSIST_TYPE (NR_CURSEG_DATA_TYPE + NR_CURSEG_NODE_TYPE)
999 #define NR_CURSEG_TYPE (NR_CURSEG_INMEM_TYPE + NR_CURSEG_PERSIST_TYPE)
1002 CURSEG_HOT_DATA = 0, /* directory entry blocks */
1003 CURSEG_WARM_DATA, /* data blocks */
1004 CURSEG_COLD_DATA, /* multimedia or GCed data blocks */
1005 CURSEG_HOT_NODE, /* direct node blocks of directory files */
1006 CURSEG_WARM_NODE, /* direct node blocks of normal files */
1007 CURSEG_COLD_NODE, /* indirect node blocks */
1008 NR_PERSISTENT_LOG, /* number of persistent log */
1009 CURSEG_COLD_DATA_PINNED = NR_PERSISTENT_LOG,
1010 /* pinned file that needs consecutive block address */
1011 CURSEG_ALL_DATA_ATGC, /* SSR alloctor in hot/warm/cold data area */
1012 NO_CHECK_TYPE, /* number of persistent & inmem log */
1016 struct completion wait;
1017 struct llist_node llnode;
1022 struct flush_cmd_control {
1023 struct task_struct *f2fs_issue_flush; /* flush thread */
1024 wait_queue_head_t flush_wait_queue; /* waiting queue for wake-up */
1025 atomic_t issued_flush; /* # of issued flushes */
1026 atomic_t queued_flush; /* # of queued flushes */
1027 struct llist_head issue_list; /* list for command issue */
1028 struct llist_node *dispatch_list; /* list for command dispatch */
1031 struct f2fs_sm_info {
1032 struct sit_info *sit_info; /* whole segment information */
1033 struct free_segmap_info *free_info; /* free segment information */
1034 struct dirty_seglist_info *dirty_info; /* dirty segment information */
1035 struct curseg_info *curseg_array; /* active segment information */
1037 struct f2fs_rwsem curseg_lock; /* for preventing curseg change */
1039 block_t seg0_blkaddr; /* block address of 0'th segment */
1040 block_t main_blkaddr; /* start block address of main area */
1041 block_t ssa_blkaddr; /* start block address of SSA area */
1043 unsigned int segment_count; /* total # of segments */
1044 unsigned int main_segments; /* # of segments in main area */
1045 unsigned int reserved_segments; /* # of reserved segments */
1046 unsigned int additional_reserved_segments;/* reserved segs for IO align feature */
1047 unsigned int ovp_segments; /* # of overprovision segments */
1049 /* a threshold to reclaim prefree segments */
1050 unsigned int rec_prefree_segments;
1052 struct list_head sit_entry_set; /* sit entry set list */
1054 unsigned int ipu_policy; /* in-place-update policy */
1055 unsigned int min_ipu_util; /* in-place-update threshold */
1056 unsigned int min_fsync_blocks; /* threshold for fsync */
1057 unsigned int min_seq_blocks; /* threshold for sequential blocks */
1058 unsigned int min_hot_blocks; /* threshold for hot block allocation */
1059 unsigned int min_ssr_sections; /* threshold to trigger SSR allocation */
1061 /* for flush command control */
1062 struct flush_cmd_control *fcc_info;
1064 /* for discard command control */
1065 struct discard_cmd_control *dcc_info;
1072 * COUNT_TYPE for monitoring
1074 * f2fs monitors the number of several block types such as on-writeback,
1075 * dirty dentry blocks, dirty node blocks, and dirty meta blocks.
1077 #define WB_DATA_TYPE(p) (__is_cp_guaranteed(p) ? F2FS_WB_CP_DATA : F2FS_WB_DATA)
1096 * The below are the page types of bios used in submit_bio().
1097 * The available types are:
1098 * DATA User data pages. It operates as async mode.
1099 * NODE Node pages. It operates as async mode.
1100 * META FS metadata pages such as SIT, NAT, CP.
1101 * NR_PAGE_TYPE The number of page types.
1102 * META_FLUSH Make sure the previous pages are written
1103 * with waiting the bio's completion
1104 * ... Only can be used with META.
1106 #define PAGE_TYPE_OF_BIO(type) ((type) > META ? META : (type))
1109 NODE = 1, /* should not change this */
1113 IPU, /* the below types are used by tracepoints only. */
1118 HOT = 0, /* must be zero for meta bio */
1124 enum need_lock_type {
1130 enum cp_reason_type {
1146 APP_DIRECT_IO, /* app direct write IOs */
1147 APP_BUFFERED_IO, /* app buffered write IOs */
1148 APP_WRITE_IO, /* app write IOs */
1149 APP_MAPPED_IO, /* app mapped IOs */
1150 APP_BUFFERED_CDATA_IO, /* app buffered write IOs on compressed file */
1151 APP_MAPPED_CDATA_IO, /* app mapped write IOs on compressed file */
1152 FS_DATA_IO, /* data IOs from kworker/fsync/reclaimer */
1153 FS_CDATA_IO, /* data IOs from kworker/fsync/reclaimer on compressed file */
1154 FS_NODE_IO, /* node IOs from kworker/fsync/reclaimer */
1155 FS_META_IO, /* meta IOs from kworker/reclaimer */
1156 FS_GC_DATA_IO, /* data IOs from forground gc */
1157 FS_GC_NODE_IO, /* node IOs from forground gc */
1158 FS_CP_DATA_IO, /* data IOs from checkpoint */
1159 FS_CP_NODE_IO, /* node IOs from checkpoint */
1160 FS_CP_META_IO, /* meta IOs from checkpoint */
1163 APP_DIRECT_READ_IO, /* app direct read IOs */
1164 APP_BUFFERED_READ_IO, /* app buffered read IOs */
1165 APP_READ_IO, /* app read IOs */
1166 APP_MAPPED_READ_IO, /* app mapped read IOs */
1167 APP_BUFFERED_CDATA_READ_IO, /* app buffered read IOs on compressed file */
1168 APP_MAPPED_CDATA_READ_IO, /* app mapped read IOs on compressed file */
1169 FS_DATA_READ_IO, /* data read IOs */
1170 FS_GDATA_READ_IO, /* data read IOs from background gc */
1171 FS_CDATA_READ_IO, /* compressed data read IOs */
1172 FS_NODE_READ_IO, /* node read IOs */
1173 FS_META_READ_IO, /* meta read IOs */
1176 FS_DISCARD_IO, /* discard */
1177 FS_FLUSH_IO, /* flush */
1181 struct f2fs_io_info {
1182 struct f2fs_sb_info *sbi; /* f2fs_sb_info pointer */
1183 nid_t ino; /* inode number */
1184 enum page_type type; /* contains DATA/NODE/META/META_FLUSH */
1185 enum temp_type temp; /* contains HOT/WARM/COLD */
1186 enum req_op op; /* contains REQ_OP_ */
1187 blk_opf_t op_flags; /* req_flag_bits */
1188 block_t new_blkaddr; /* new block address to be written */
1189 block_t old_blkaddr; /* old block address before Cow */
1190 struct page *page; /* page to be written */
1191 struct page *encrypted_page; /* encrypted page */
1192 struct page *compressed_page; /* compressed page */
1193 struct list_head list; /* serialize IOs */
1194 unsigned int compr_blocks; /* # of compressed block addresses */
1195 unsigned int need_lock:8; /* indicate we need to lock cp_rwsem */
1196 unsigned int version:8; /* version of the node */
1197 unsigned int submitted:1; /* indicate IO submission */
1198 unsigned int in_list:1; /* indicate fio is in io_list */
1199 unsigned int is_por:1; /* indicate IO is from recovery or not */
1200 unsigned int retry:1; /* need to reallocate block address */
1201 unsigned int encrypted:1; /* indicate file is encrypted */
1202 unsigned int post_read:1; /* require post read */
1203 enum iostat_type io_type; /* io type */
1204 struct writeback_control *io_wbc; /* writeback control */
1205 struct bio **bio; /* bio for ipu */
1206 sector_t *last_block; /* last block number in bio */
1211 struct list_head list;
1214 #define is_read_io(rw) ((rw) == READ)
1215 struct f2fs_bio_info {
1216 struct f2fs_sb_info *sbi; /* f2fs superblock */
1217 struct bio *bio; /* bios to merge */
1218 sector_t last_block_in_bio; /* last block number */
1219 struct f2fs_io_info fio; /* store buffered io info. */
1220 struct f2fs_rwsem io_rwsem; /* blocking op for bio */
1221 spinlock_t io_lock; /* serialize DATA/NODE IOs */
1222 struct list_head io_list; /* track fios */
1223 struct list_head bio_list; /* bio entry list head */
1224 struct f2fs_rwsem bio_list_lock; /* lock to protect bio entry list */
1227 #define FDEV(i) (sbi->devs[i])
1228 #define RDEV(i) (raw_super->devs[i])
1229 struct f2fs_dev_info {
1230 struct block_device *bdev;
1231 char path[MAX_PATH_LEN];
1232 unsigned int total_segments;
1235 #ifdef CONFIG_BLK_DEV_ZONED
1236 unsigned int nr_blkz; /* Total number of zones */
1237 unsigned long *blkz_seq; /* Bitmap indicating sequential zones */
1242 DIR_INODE, /* for dirty dir inode */
1243 FILE_INODE, /* for dirty regular/symlink inode */
1244 DIRTY_META, /* for all dirtied inode metadata */
1248 /* for inner inode cache management */
1249 struct inode_management {
1250 struct radix_tree_root ino_root; /* ino entry array */
1251 spinlock_t ino_lock; /* for ino entry lock */
1252 struct list_head ino_list; /* inode list head */
1253 unsigned long ino_num; /* number of entries */
1257 struct atgc_management {
1258 bool atgc_enabled; /* ATGC is enabled or not */
1259 struct rb_root_cached root; /* root of victim rb-tree */
1260 struct list_head victim_list; /* linked with all victim entries */
1261 unsigned int victim_count; /* victim count in rb-tree */
1262 unsigned int candidate_ratio; /* candidate ratio */
1263 unsigned int max_candidate_count; /* max candidate count */
1264 unsigned int age_weight; /* age weight, vblock_weight = 100 - age_weight */
1265 unsigned long long age_threshold; /* age threshold */
1268 struct f2fs_gc_control {
1269 unsigned int victim_segno; /* target victim segment number */
1270 int init_gc_type; /* FG_GC or BG_GC */
1271 bool no_bg_gc; /* check the space and stop bg_gc */
1272 bool should_migrate_blocks; /* should migrate blocks */
1273 bool err_gc_skipped; /* return EAGAIN if GC skipped */
1274 unsigned int nr_free_secs; /* # of free sections to do GC */
1278 * For s_flag in struct f2fs_sb_info
1279 * Modification on enum should be synchronized with s_flag array
1282 SBI_IS_DIRTY, /* dirty flag for checkpoint */
1283 SBI_IS_CLOSE, /* specify unmounting */
1284 SBI_NEED_FSCK, /* need fsck.f2fs to fix */
1285 SBI_POR_DOING, /* recovery is doing or not */
1286 SBI_NEED_SB_WRITE, /* need to recover superblock */
1287 SBI_NEED_CP, /* need to checkpoint */
1288 SBI_IS_SHUTDOWN, /* shutdown by ioctl */
1289 SBI_IS_RECOVERED, /* recovered orphan/data */
1290 SBI_CP_DISABLED, /* CP was disabled last mount */
1291 SBI_CP_DISABLED_QUICK, /* CP was disabled quickly */
1292 SBI_QUOTA_NEED_FLUSH, /* need to flush quota info in CP */
1293 SBI_QUOTA_SKIP_FLUSH, /* skip flushing quota in current CP */
1294 SBI_QUOTA_NEED_REPAIR, /* quota file may be corrupted */
1295 SBI_IS_RESIZEFS, /* resizefs is in process */
1296 SBI_IS_FREEZING, /* freezefs is in process */
1297 SBI_IS_WRITABLE, /* remove ro mountoption transiently */
1307 UMOUNT_DISCARD_TIMEOUT,
1311 /* Note that you need to keep synchronization with this gc_mode_names array */
1324 BGGC_MODE_ON, /* background gc is on */
1325 BGGC_MODE_OFF, /* background gc is off */
1327 * background gc is on, migrating blocks
1328 * like foreground gc
1333 FS_MODE_ADAPTIVE, /* use both lfs/ssr allocation */
1334 FS_MODE_LFS, /* use lfs allocation only */
1335 FS_MODE_FRAGMENT_SEG, /* segment fragmentation mode */
1336 FS_MODE_FRAGMENT_BLK, /* block fragmentation mode */
1340 ALLOC_MODE_DEFAULT, /* stay default */
1341 ALLOC_MODE_REUSE, /* reuse segments as much as possible */
1345 FSYNC_MODE_POSIX, /* fsync follows posix semantics */
1346 FSYNC_MODE_STRICT, /* fsync behaves in line with ext4 */
1347 FSYNC_MODE_NOBARRIER, /* fsync behaves nobarrier based on posix */
1352 * automatically compress compression
1356 * automatical compression is disabled.
1357 * user can control the file compression
1363 DISCARD_UNIT_BLOCK, /* basic discard unit is block */
1364 DISCARD_UNIT_SEGMENT, /* basic discard unit is segment */
1365 DISCARD_UNIT_SECTION, /* basic discard unit is section */
1369 MEMORY_MODE_NORMAL, /* memory mode for normal devices */
1370 MEMORY_MODE_LOW, /* memory mode for low memry devices */
1373 static inline int f2fs_test_bit(unsigned int nr, char *addr);
1374 static inline void f2fs_set_bit(unsigned int nr, char *addr);
1375 static inline void f2fs_clear_bit(unsigned int nr, char *addr);
1378 * Layout of f2fs page.private:
1380 * Layout A: lowest bit should be 1
1381 * | bit0 = 1 | bit1 | bit2 | ... | bit MAX | private data .... |
1382 * bit 0 PAGE_PRIVATE_NOT_POINTER
1383 * bit 1 PAGE_PRIVATE_DUMMY_WRITE
1384 * bit 2 PAGE_PRIVATE_ONGOING_MIGRATION
1385 * bit 3 PAGE_PRIVATE_INLINE_INODE
1386 * bit 4 PAGE_PRIVATE_REF_RESOURCE
1387 * bit 5- f2fs private data
1389 * Layout B: lowest bit should be 0
1390 * page.private is a wrapped pointer.
1393 PAGE_PRIVATE_NOT_POINTER, /* private contains non-pointer data */
1394 PAGE_PRIVATE_DUMMY_WRITE, /* data page for padding aligned IO */
1395 PAGE_PRIVATE_ONGOING_MIGRATION, /* data page which is on-going migrating */
1396 PAGE_PRIVATE_INLINE_INODE, /* inode page contains inline data */
1397 PAGE_PRIVATE_REF_RESOURCE, /* dirty page has referenced resources */
1401 /* For compression */
1402 enum compress_algorithm_type {
1410 enum compress_flag {
1415 #define COMPRESS_WATERMARK 20
1416 #define COMPRESS_PERCENT 20
1418 #define COMPRESS_DATA_RESERVED_SIZE 4
1419 struct compress_data {
1420 __le32 clen; /* compressed data size */
1421 __le32 chksum; /* compressed data chksum */
1422 __le32 reserved[COMPRESS_DATA_RESERVED_SIZE]; /* reserved */
1423 u8 cdata[]; /* compressed data */
1426 #define COMPRESS_HEADER_SIZE (sizeof(struct compress_data))
1428 #define F2FS_COMPRESSED_PAGE_MAGIC 0xF5F2C000
1430 #define COMPRESS_LEVEL_OFFSET 8
1432 /* compress context */
1433 struct compress_ctx {
1434 struct inode *inode; /* inode the context belong to */
1435 pgoff_t cluster_idx; /* cluster index number */
1436 unsigned int cluster_size; /* page count in cluster */
1437 unsigned int log_cluster_size; /* log of cluster size */
1438 struct page **rpages; /* pages store raw data in cluster */
1439 unsigned int nr_rpages; /* total page number in rpages */
1440 struct page **cpages; /* pages store compressed data in cluster */
1441 unsigned int nr_cpages; /* total page number in cpages */
1442 unsigned int valid_nr_cpages; /* valid page number in cpages */
1443 void *rbuf; /* virtual mapped address on rpages */
1444 struct compress_data *cbuf; /* virtual mapped address on cpages */
1445 size_t rlen; /* valid data length in rbuf */
1446 size_t clen; /* valid data length in cbuf */
1447 void *private; /* payload buffer for specified compression algorithm */
1448 void *private2; /* extra payload buffer */
1451 /* compress context for write IO path */
1452 struct compress_io_ctx {
1453 u32 magic; /* magic number to indicate page is compressed */
1454 struct inode *inode; /* inode the context belong to */
1455 struct page **rpages; /* pages store raw data in cluster */
1456 unsigned int nr_rpages; /* total page number in rpages */
1457 atomic_t pending_pages; /* in-flight compressed page count */
1460 /* Context for decompressing one cluster on the read IO path */
1461 struct decompress_io_ctx {
1462 u32 magic; /* magic number to indicate page is compressed */
1463 struct inode *inode; /* inode the context belong to */
1464 pgoff_t cluster_idx; /* cluster index number */
1465 unsigned int cluster_size; /* page count in cluster */
1466 unsigned int log_cluster_size; /* log of cluster size */
1467 struct page **rpages; /* pages store raw data in cluster */
1468 unsigned int nr_rpages; /* total page number in rpages */
1469 struct page **cpages; /* pages store compressed data in cluster */
1470 unsigned int nr_cpages; /* total page number in cpages */
1471 struct page **tpages; /* temp pages to pad holes in cluster */
1472 void *rbuf; /* virtual mapped address on rpages */
1473 struct compress_data *cbuf; /* virtual mapped address on cpages */
1474 size_t rlen; /* valid data length in rbuf */
1475 size_t clen; /* valid data length in cbuf */
1478 * The number of compressed pages remaining to be read in this cluster.
1479 * This is initially nr_cpages. It is decremented by 1 each time a page
1480 * has been read (or failed to be read). When it reaches 0, the cluster
1481 * is decompressed (or an error is reported).
1483 * If an error occurs before all the pages have been submitted for I/O,
1484 * then this will never reach 0. In this case the I/O submitter is
1485 * responsible for calling f2fs_decompress_end_io() instead.
1487 atomic_t remaining_pages;
1490 * Number of references to this decompress_io_ctx.
1492 * One reference is held for I/O completion. This reference is dropped
1493 * after the pagecache pages are updated and unlocked -- either after
1494 * decompression (and verity if enabled), or after an error.
1496 * In addition, each compressed page holds a reference while it is in a
1497 * bio. These references are necessary prevent compressed pages from
1498 * being freed while they are still in a bio.
1502 bool failed; /* IO error occurred before decompression? */
1503 bool need_verity; /* need fs-verity verification after decompression? */
1504 void *private; /* payload buffer for specified decompression algorithm */
1505 void *private2; /* extra payload buffer */
1506 struct work_struct verity_work; /* work to verify the decompressed pages */
1507 struct work_struct free_work; /* work for late free this structure itself */
1510 #define NULL_CLUSTER ((unsigned int)(~0))
1511 #define MIN_COMPRESS_LOG_SIZE 2
1512 #define MAX_COMPRESS_LOG_SIZE 8
1513 #define MAX_COMPRESS_WINDOW_SIZE(log_size) ((PAGE_SIZE) << (log_size))
1515 struct f2fs_sb_info {
1516 struct super_block *sb; /* pointer to VFS super block */
1517 struct proc_dir_entry *s_proc; /* proc entry */
1518 struct f2fs_super_block *raw_super; /* raw super block pointer */
1519 struct f2fs_rwsem sb_lock; /* lock for raw super block */
1520 int valid_super_block; /* valid super block no */
1521 unsigned long s_flag; /* flags for sbi */
1522 struct mutex writepages; /* mutex for writepages() */
1524 #ifdef CONFIG_BLK_DEV_ZONED
1525 unsigned int blocks_per_blkz; /* F2FS blocks per zone */
1528 /* for node-related operations */
1529 struct f2fs_nm_info *nm_info; /* node manager */
1530 struct inode *node_inode; /* cache node blocks */
1532 /* for segment-related operations */
1533 struct f2fs_sm_info *sm_info; /* segment manager */
1535 /* for bio operations */
1536 struct f2fs_bio_info *write_io[NR_PAGE_TYPE]; /* for write bios */
1537 /* keep migration IO order for LFS mode */
1538 struct f2fs_rwsem io_order_lock;
1539 mempool_t *write_io_dummy; /* Dummy pages */
1540 pgoff_t page_eio_ofs[NR_PAGE_TYPE]; /* EIO page offset */
1541 int page_eio_cnt[NR_PAGE_TYPE]; /* EIO count */
1543 /* for checkpoint */
1544 struct f2fs_checkpoint *ckpt; /* raw checkpoint pointer */
1545 int cur_cp_pack; /* remain current cp pack */
1546 spinlock_t cp_lock; /* for flag in ckpt */
1547 struct inode *meta_inode; /* cache meta blocks */
1548 struct f2fs_rwsem cp_global_sem; /* checkpoint procedure lock */
1549 struct f2fs_rwsem cp_rwsem; /* blocking FS operations */
1550 struct f2fs_rwsem node_write; /* locking node writes */
1551 struct f2fs_rwsem node_change; /* locking node change */
1552 wait_queue_head_t cp_wait;
1553 unsigned long last_time[MAX_TIME]; /* to store time in jiffies */
1554 long interval_time[MAX_TIME]; /* to store thresholds */
1555 struct ckpt_req_control cprc_info; /* for checkpoint request control */
1557 struct inode_management im[MAX_INO_ENTRY]; /* manage inode cache */
1559 spinlock_t fsync_node_lock; /* for node entry lock */
1560 struct list_head fsync_node_list; /* node list head */
1561 unsigned int fsync_seg_id; /* sequence id */
1562 unsigned int fsync_node_num; /* number of node entries */
1564 /* for orphan inode, use 0'th array */
1565 unsigned int max_orphans; /* max orphan inodes */
1567 /* for inode management */
1568 struct list_head inode_list[NR_INODE_TYPE]; /* dirty inode list */
1569 spinlock_t inode_lock[NR_INODE_TYPE]; /* for dirty inode list lock */
1570 struct mutex flush_lock; /* for flush exclusion */
1572 /* for extent tree cache */
1573 struct extent_tree_info extent_tree[NR_EXTENT_CACHES];
1574 atomic64_t allocated_data_blocks; /* for block age extent_cache */
1576 /* The threshold used for hot and warm data seperation*/
1577 unsigned int hot_data_age_threshold;
1578 unsigned int warm_data_age_threshold;
1579 unsigned int last_age_weight;
1581 /* basic filesystem units */
1582 unsigned int log_sectors_per_block; /* log2 sectors per block */
1583 unsigned int log_blocksize; /* log2 block size */
1584 unsigned int blocksize; /* block size */
1585 unsigned int root_ino_num; /* root inode number*/
1586 unsigned int node_ino_num; /* node inode number*/
1587 unsigned int meta_ino_num; /* meta inode number*/
1588 unsigned int log_blocks_per_seg; /* log2 blocks per segment */
1589 unsigned int blocks_per_seg; /* blocks per segment */
1590 unsigned int unusable_blocks_per_sec; /* unusable blocks per section */
1591 unsigned int segs_per_sec; /* segments per section */
1592 unsigned int secs_per_zone; /* sections per zone */
1593 unsigned int total_sections; /* total section count */
1594 unsigned int total_node_count; /* total node block count */
1595 unsigned int total_valid_node_count; /* valid node block count */
1596 int dir_level; /* directory level */
1597 bool readdir_ra; /* readahead inode in readdir */
1598 u64 max_io_bytes; /* max io bytes to merge IOs */
1600 block_t user_block_count; /* # of user blocks */
1601 block_t total_valid_block_count; /* # of valid blocks */
1602 block_t discard_blks; /* discard command candidats */
1603 block_t last_valid_block_count; /* for recovery */
1604 block_t reserved_blocks; /* configurable reserved blocks */
1605 block_t current_reserved_blocks; /* current reserved blocks */
1607 /* Additional tracking for no checkpoint mode */
1608 block_t unusable_block_count; /* # of blocks saved by last cp */
1610 unsigned int nquota_files; /* # of quota sysfile */
1611 struct f2fs_rwsem quota_sem; /* blocking cp for flags */
1613 /* # of pages, see count_type */
1614 atomic_t nr_pages[NR_COUNT_TYPE];
1615 /* # of allocated blocks */
1616 struct percpu_counter alloc_valid_block_count;
1617 /* # of node block writes as roll forward recovery */
1618 struct percpu_counter rf_node_block_count;
1620 /* writeback control */
1621 atomic_t wb_sync_req[META]; /* count # of WB_SYNC threads */
1623 /* valid inode count */
1624 struct percpu_counter total_valid_inode_count;
1626 struct f2fs_mount_info mount_opt; /* mount options */
1628 /* for cleaning operations */
1629 struct f2fs_rwsem gc_lock; /*
1630 * semaphore for GC, avoid
1631 * race between GC and GC or CP
1633 struct f2fs_gc_kthread *gc_thread; /* GC thread */
1634 struct atgc_management am; /* atgc management */
1635 unsigned int cur_victim_sec; /* current victim section num */
1636 unsigned int gc_mode; /* current GC state */
1637 unsigned int next_victim_seg[2]; /* next segment in victim section */
1638 spinlock_t gc_remaining_trials_lock;
1639 /* remaining trial count for GC_URGENT_* and GC_IDLE_* */
1640 unsigned int gc_remaining_trials;
1642 /* for skip statistic */
1643 unsigned long long skipped_gc_rwsem; /* FG_GC only */
1645 /* threshold for gc trials on pinned files */
1646 u64 gc_pin_file_threshold;
1647 struct f2fs_rwsem pin_sem;
1649 /* maximum # of trials to find a victim segment for SSR and GC */
1650 unsigned int max_victim_search;
1651 /* migration granularity of garbage collection, unit: segment */
1652 unsigned int migration_granularity;
1655 * for stat information.
1656 * one is for the LFS mode, and the other is for the SSR mode.
1658 #ifdef CONFIG_F2FS_STAT_FS
1659 struct f2fs_stat_info *stat_info; /* FS status information */
1660 atomic_t meta_count[META_MAX]; /* # of meta blocks */
1661 unsigned int segment_count[2]; /* # of allocated segments */
1662 unsigned int block_count[2]; /* # of allocated blocks */
1663 atomic_t inplace_count; /* # of inplace update */
1664 /* # of lookup extent cache */
1665 atomic64_t total_hit_ext[NR_EXTENT_CACHES];
1666 /* # of hit rbtree extent node */
1667 atomic64_t read_hit_rbtree[NR_EXTENT_CACHES];
1668 /* # of hit cached extent node */
1669 atomic64_t read_hit_cached[NR_EXTENT_CACHES];
1670 /* # of hit largest extent node in read extent cache */
1671 atomic64_t read_hit_largest;
1672 atomic_t inline_xattr; /* # of inline_xattr inodes */
1673 atomic_t inline_inode; /* # of inline_data inodes */
1674 atomic_t inline_dir; /* # of inline_dentry inodes */
1675 atomic_t compr_inode; /* # of compressed inodes */
1676 atomic64_t compr_blocks; /* # of compressed blocks */
1677 atomic_t swapfile_inode; /* # of swapfile inodes */
1678 atomic_t atomic_files; /* # of opened atomic file */
1679 atomic_t max_aw_cnt; /* max # of atomic writes */
1680 unsigned int io_skip_bggc; /* skip background gc for in-flight IO */
1681 unsigned int other_skip_bggc; /* skip background gc for other reasons */
1682 unsigned int ndirty_inode[NR_INODE_TYPE]; /* # of dirty inodes */
1684 spinlock_t stat_lock; /* lock for stat operations */
1686 /* to attach REQ_META|REQ_FUA flags */
1687 unsigned int data_io_flag;
1688 unsigned int node_io_flag;
1690 /* For sysfs support */
1691 struct kobject s_kobj; /* /sys/fs/f2fs/<devname> */
1692 struct completion s_kobj_unregister;
1694 struct kobject s_stat_kobj; /* /sys/fs/f2fs/<devname>/stat */
1695 struct completion s_stat_kobj_unregister;
1697 struct kobject s_feature_list_kobj; /* /sys/fs/f2fs/<devname>/feature_list */
1698 struct completion s_feature_list_kobj_unregister;
1700 /* For shrinker support */
1701 struct list_head s_list;
1702 struct mutex umount_mutex;
1703 unsigned int shrinker_run_no;
1705 /* For multi devices */
1706 int s_ndevs; /* number of devices */
1707 struct f2fs_dev_info *devs; /* for device list */
1708 unsigned int dirty_device; /* for checkpoint data flush */
1709 spinlock_t dev_lock; /* protect dirty_device */
1710 bool aligned_blksize; /* all devices has the same logical blksize */
1712 /* For write statistics */
1713 u64 sectors_written_start;
1716 /* Reference to checksum algorithm driver via cryptoapi */
1717 struct crypto_shash *s_chksum_driver;
1719 /* Precomputed FS UUID checksum for seeding other checksums */
1720 __u32 s_chksum_seed;
1722 struct workqueue_struct *post_read_wq; /* post read workqueue */
1724 unsigned char errors[MAX_F2FS_ERRORS]; /* error flags */
1725 spinlock_t error_lock; /* protect errors array */
1726 bool error_dirty; /* errors of sb is dirty */
1728 struct kmem_cache *inline_xattr_slab; /* inline xattr entry */
1729 unsigned int inline_xattr_slab_size; /* default inline xattr slab size */
1731 /* For reclaimed segs statistics per each GC mode */
1732 unsigned int gc_segment_mode; /* GC state for reclaimed segments */
1733 unsigned int gc_reclaimed_segs[MAX_GC_MODE]; /* Reclaimed segs for each mode */
1735 unsigned long seq_file_ra_mul; /* multiplier for ra_pages of seq. files in fadvise */
1737 int max_fragment_chunk; /* max chunk size for block fragmentation mode */
1738 int max_fragment_hole; /* max hole size for block fragmentation mode */
1740 /* For atomic write statistics */
1741 atomic64_t current_atomic_write;
1742 s64 peak_atomic_write;
1743 u64 committed_atomic_block;
1744 u64 revoked_atomic_block;
1746 #ifdef CONFIG_F2FS_FS_COMPRESSION
1747 struct kmem_cache *page_array_slab; /* page array entry */
1748 unsigned int page_array_slab_size; /* default page array slab size */
1750 /* For runtime compression statistics */
1751 u64 compr_written_block;
1752 u64 compr_saved_block;
1753 u32 compr_new_inode;
1755 /* For compressed block cache */
1756 struct inode *compress_inode; /* cache compressed blocks */
1757 unsigned int compress_percent; /* cache page percentage */
1758 unsigned int compress_watermark; /* cache page watermark */
1759 atomic_t compress_page_hit; /* cache hit count */
1762 #ifdef CONFIG_F2FS_IOSTAT
1763 /* For app/fs IO statistics */
1764 spinlock_t iostat_lock;
1765 unsigned long long iostat_count[NR_IO_TYPE];
1766 unsigned long long iostat_bytes[NR_IO_TYPE];
1767 unsigned long long prev_iostat_bytes[NR_IO_TYPE];
1769 unsigned long iostat_next_period;
1770 unsigned int iostat_period_ms;
1772 /* For io latency related statistics info in one iostat period */
1773 spinlock_t iostat_lat_lock;
1774 struct iostat_lat_info *iostat_io_lat;
1778 #ifdef CONFIG_F2FS_FAULT_INJECTION
1779 #define time_to_inject(sbi, type) __time_to_inject(sbi, type, __func__, \
1780 __builtin_return_address(0))
1781 static inline bool __time_to_inject(struct f2fs_sb_info *sbi, int type,
1782 const char *func, const char *parent_func)
1784 struct f2fs_fault_info *ffi = &F2FS_OPTION(sbi).fault_info;
1786 if (!ffi->inject_rate)
1789 if (!IS_FAULT_SET(ffi, type))
1792 atomic_inc(&ffi->inject_ops);
1793 if (atomic_read(&ffi->inject_ops) >= ffi->inject_rate) {
1794 atomic_set(&ffi->inject_ops, 0);
1795 printk_ratelimited("%sF2FS-fs (%s) : inject %s in %s of %pS\n",
1796 KERN_INFO, sbi->sb->s_id, f2fs_fault_name[type],
1803 static inline bool time_to_inject(struct f2fs_sb_info *sbi, int type)
1810 * Test if the mounted volume is a multi-device volume.
1811 * - For a single regular disk volume, sbi->s_ndevs is 0.
1812 * - For a single zoned disk volume, sbi->s_ndevs is 1.
1813 * - For a multi-device volume, sbi->s_ndevs is always 2 or more.
1815 static inline bool f2fs_is_multi_device(struct f2fs_sb_info *sbi)
1817 return sbi->s_ndevs > 1;
1820 static inline void f2fs_update_time(struct f2fs_sb_info *sbi, int type)
1822 unsigned long now = jiffies;
1824 sbi->last_time[type] = now;
1826 /* DISCARD_TIME and GC_TIME are based on REQ_TIME */
1827 if (type == REQ_TIME) {
1828 sbi->last_time[DISCARD_TIME] = now;
1829 sbi->last_time[GC_TIME] = now;
1833 static inline bool f2fs_time_over(struct f2fs_sb_info *sbi, int type)
1835 unsigned long interval = sbi->interval_time[type] * HZ;
1837 return time_after(jiffies, sbi->last_time[type] + interval);
1840 static inline unsigned int f2fs_time_to_wait(struct f2fs_sb_info *sbi,
1843 unsigned long interval = sbi->interval_time[type] * HZ;
1844 unsigned int wait_ms = 0;
1847 delta = (sbi->last_time[type] + interval) - jiffies;
1849 wait_ms = jiffies_to_msecs(delta);
1857 static inline u32 __f2fs_crc32(struct f2fs_sb_info *sbi, u32 crc,
1858 const void *address, unsigned int length)
1861 struct shash_desc shash;
1866 BUG_ON(crypto_shash_descsize(sbi->s_chksum_driver) != sizeof(desc.ctx));
1868 desc.shash.tfm = sbi->s_chksum_driver;
1869 *(u32 *)desc.ctx = crc;
1871 err = crypto_shash_update(&desc.shash, address, length);
1874 return *(u32 *)desc.ctx;
1877 static inline u32 f2fs_crc32(struct f2fs_sb_info *sbi, const void *address,
1878 unsigned int length)
1880 return __f2fs_crc32(sbi, F2FS_SUPER_MAGIC, address, length);
1883 static inline bool f2fs_crc_valid(struct f2fs_sb_info *sbi, __u32 blk_crc,
1884 void *buf, size_t buf_size)
1886 return f2fs_crc32(sbi, buf, buf_size) == blk_crc;
1889 static inline u32 f2fs_chksum(struct f2fs_sb_info *sbi, u32 crc,
1890 const void *address, unsigned int length)
1892 return __f2fs_crc32(sbi, crc, address, length);
1895 static inline struct f2fs_inode_info *F2FS_I(struct inode *inode)
1897 return container_of(inode, struct f2fs_inode_info, vfs_inode);
1900 static inline struct f2fs_sb_info *F2FS_SB(struct super_block *sb)
1902 return sb->s_fs_info;
1905 static inline struct f2fs_sb_info *F2FS_I_SB(struct inode *inode)
1907 return F2FS_SB(inode->i_sb);
1910 static inline struct f2fs_sb_info *F2FS_M_SB(struct address_space *mapping)
1912 return F2FS_I_SB(mapping->host);
1915 static inline struct f2fs_sb_info *F2FS_P_SB(struct page *page)
1917 return F2FS_M_SB(page_file_mapping(page));
1920 static inline struct f2fs_super_block *F2FS_RAW_SUPER(struct f2fs_sb_info *sbi)
1922 return (struct f2fs_super_block *)(sbi->raw_super);
1925 static inline struct f2fs_checkpoint *F2FS_CKPT(struct f2fs_sb_info *sbi)
1927 return (struct f2fs_checkpoint *)(sbi->ckpt);
1930 static inline struct f2fs_node *F2FS_NODE(struct page *page)
1932 return (struct f2fs_node *)page_address(page);
1935 static inline struct f2fs_inode *F2FS_INODE(struct page *page)
1937 return &((struct f2fs_node *)page_address(page))->i;
1940 static inline struct f2fs_nm_info *NM_I(struct f2fs_sb_info *sbi)
1942 return (struct f2fs_nm_info *)(sbi->nm_info);
1945 static inline struct f2fs_sm_info *SM_I(struct f2fs_sb_info *sbi)
1947 return (struct f2fs_sm_info *)(sbi->sm_info);
1950 static inline struct sit_info *SIT_I(struct f2fs_sb_info *sbi)
1952 return (struct sit_info *)(SM_I(sbi)->sit_info);
1955 static inline struct free_segmap_info *FREE_I(struct f2fs_sb_info *sbi)
1957 return (struct free_segmap_info *)(SM_I(sbi)->free_info);
1960 static inline struct dirty_seglist_info *DIRTY_I(struct f2fs_sb_info *sbi)
1962 return (struct dirty_seglist_info *)(SM_I(sbi)->dirty_info);
1965 static inline struct address_space *META_MAPPING(struct f2fs_sb_info *sbi)
1967 return sbi->meta_inode->i_mapping;
1970 static inline struct address_space *NODE_MAPPING(struct f2fs_sb_info *sbi)
1972 return sbi->node_inode->i_mapping;
1975 static inline bool is_sbi_flag_set(struct f2fs_sb_info *sbi, unsigned int type)
1977 return test_bit(type, &sbi->s_flag);
1980 static inline void set_sbi_flag(struct f2fs_sb_info *sbi, unsigned int type)
1982 set_bit(type, &sbi->s_flag);
1985 static inline void clear_sbi_flag(struct f2fs_sb_info *sbi, unsigned int type)
1987 clear_bit(type, &sbi->s_flag);
1990 static inline unsigned long long cur_cp_version(struct f2fs_checkpoint *cp)
1992 return le64_to_cpu(cp->checkpoint_ver);
1995 static inline unsigned long f2fs_qf_ino(struct super_block *sb, int type)
1997 if (type < F2FS_MAX_QUOTAS)
1998 return le32_to_cpu(F2FS_SB(sb)->raw_super->qf_ino[type]);
2002 static inline __u64 cur_cp_crc(struct f2fs_checkpoint *cp)
2004 size_t crc_offset = le32_to_cpu(cp->checksum_offset);
2005 return le32_to_cpu(*((__le32 *)((unsigned char *)cp + crc_offset)));
2008 static inline bool __is_set_ckpt_flags(struct f2fs_checkpoint *cp, unsigned int f)
2010 unsigned int ckpt_flags = le32_to_cpu(cp->ckpt_flags);
2012 return ckpt_flags & f;
2015 static inline bool is_set_ckpt_flags(struct f2fs_sb_info *sbi, unsigned int f)
2017 return __is_set_ckpt_flags(F2FS_CKPT(sbi), f);
2020 static inline void __set_ckpt_flags(struct f2fs_checkpoint *cp, unsigned int f)
2022 unsigned int ckpt_flags;
2024 ckpt_flags = le32_to_cpu(cp->ckpt_flags);
2026 cp->ckpt_flags = cpu_to_le32(ckpt_flags);
2029 static inline void set_ckpt_flags(struct f2fs_sb_info *sbi, unsigned int f)
2031 unsigned long flags;
2033 spin_lock_irqsave(&sbi->cp_lock, flags);
2034 __set_ckpt_flags(F2FS_CKPT(sbi), f);
2035 spin_unlock_irqrestore(&sbi->cp_lock, flags);
2038 static inline void __clear_ckpt_flags(struct f2fs_checkpoint *cp, unsigned int f)
2040 unsigned int ckpt_flags;
2042 ckpt_flags = le32_to_cpu(cp->ckpt_flags);
2044 cp->ckpt_flags = cpu_to_le32(ckpt_flags);
2047 static inline void clear_ckpt_flags(struct f2fs_sb_info *sbi, unsigned int f)
2049 unsigned long flags;
2051 spin_lock_irqsave(&sbi->cp_lock, flags);
2052 __clear_ckpt_flags(F2FS_CKPT(sbi), f);
2053 spin_unlock_irqrestore(&sbi->cp_lock, flags);
2056 #define init_f2fs_rwsem(sem) \
2058 static struct lock_class_key __key; \
2060 __init_f2fs_rwsem((sem), #sem, &__key); \
2063 static inline void __init_f2fs_rwsem(struct f2fs_rwsem *sem,
2064 const char *sem_name, struct lock_class_key *key)
2066 __init_rwsem(&sem->internal_rwsem, sem_name, key);
2067 #ifdef CONFIG_F2FS_UNFAIR_RWSEM
2068 init_waitqueue_head(&sem->read_waiters);
2072 static inline int f2fs_rwsem_is_locked(struct f2fs_rwsem *sem)
2074 return rwsem_is_locked(&sem->internal_rwsem);
2077 static inline int f2fs_rwsem_is_contended(struct f2fs_rwsem *sem)
2079 return rwsem_is_contended(&sem->internal_rwsem);
2082 static inline void f2fs_down_read(struct f2fs_rwsem *sem)
2084 #ifdef CONFIG_F2FS_UNFAIR_RWSEM
2085 wait_event(sem->read_waiters, down_read_trylock(&sem->internal_rwsem));
2087 down_read(&sem->internal_rwsem);
2091 static inline int f2fs_down_read_trylock(struct f2fs_rwsem *sem)
2093 return down_read_trylock(&sem->internal_rwsem);
2096 #ifdef CONFIG_DEBUG_LOCK_ALLOC
2097 static inline void f2fs_down_read_nested(struct f2fs_rwsem *sem, int subclass)
2099 down_read_nested(&sem->internal_rwsem, subclass);
2102 #define f2fs_down_read_nested(sem, subclass) f2fs_down_read(sem)
2105 static inline void f2fs_up_read(struct f2fs_rwsem *sem)
2107 up_read(&sem->internal_rwsem);
2110 static inline void f2fs_down_write(struct f2fs_rwsem *sem)
2112 down_write(&sem->internal_rwsem);
2115 static inline int f2fs_down_write_trylock(struct f2fs_rwsem *sem)
2117 return down_write_trylock(&sem->internal_rwsem);
2120 static inline void f2fs_up_write(struct f2fs_rwsem *sem)
2122 up_write(&sem->internal_rwsem);
2123 #ifdef CONFIG_F2FS_UNFAIR_RWSEM
2124 wake_up_all(&sem->read_waiters);
2128 static inline void f2fs_lock_op(struct f2fs_sb_info *sbi)
2130 f2fs_down_read(&sbi->cp_rwsem);
2133 static inline int f2fs_trylock_op(struct f2fs_sb_info *sbi)
2135 if (time_to_inject(sbi, FAULT_LOCK_OP))
2137 return f2fs_down_read_trylock(&sbi->cp_rwsem);
2140 static inline void f2fs_unlock_op(struct f2fs_sb_info *sbi)
2142 f2fs_up_read(&sbi->cp_rwsem);
2145 static inline void f2fs_lock_all(struct f2fs_sb_info *sbi)
2147 f2fs_down_write(&sbi->cp_rwsem);
2150 static inline void f2fs_unlock_all(struct f2fs_sb_info *sbi)
2152 f2fs_up_write(&sbi->cp_rwsem);
2155 static inline int __get_cp_reason(struct f2fs_sb_info *sbi)
2157 int reason = CP_SYNC;
2159 if (test_opt(sbi, FASTBOOT))
2160 reason = CP_FASTBOOT;
2161 if (is_sbi_flag_set(sbi, SBI_IS_CLOSE))
2166 static inline bool __remain_node_summaries(int reason)
2168 return (reason & (CP_UMOUNT | CP_FASTBOOT));
2171 static inline bool __exist_node_summaries(struct f2fs_sb_info *sbi)
2173 return (is_set_ckpt_flags(sbi, CP_UMOUNT_FLAG) ||
2174 is_set_ckpt_flags(sbi, CP_FASTBOOT_FLAG));
2178 * Check whether the inode has blocks or not
2180 static inline int F2FS_HAS_BLOCKS(struct inode *inode)
2182 block_t xattr_block = F2FS_I(inode)->i_xattr_nid ? 1 : 0;
2184 return (inode->i_blocks >> F2FS_LOG_SECTORS_PER_BLOCK) > xattr_block;
2187 static inline bool f2fs_has_xattr_block(unsigned int ofs)
2189 return ofs == XATTR_NODE_OFFSET;
2192 static inline bool __allow_reserved_blocks(struct f2fs_sb_info *sbi,
2193 struct inode *inode, bool cap)
2197 if (!test_opt(sbi, RESERVE_ROOT))
2199 if (IS_NOQUOTA(inode))
2201 if (uid_eq(F2FS_OPTION(sbi).s_resuid, current_fsuid()))
2203 if (!gid_eq(F2FS_OPTION(sbi).s_resgid, GLOBAL_ROOT_GID) &&
2204 in_group_p(F2FS_OPTION(sbi).s_resgid))
2206 if (cap && capable(CAP_SYS_RESOURCE))
2211 static inline void f2fs_i_blocks_write(struct inode *, block_t, bool, bool);
2212 static inline int inc_valid_block_count(struct f2fs_sb_info *sbi,
2213 struct inode *inode, blkcnt_t *count)
2215 blkcnt_t diff = 0, release = 0;
2216 block_t avail_user_block_count;
2219 ret = dquot_reserve_block(inode, *count);
2223 if (time_to_inject(sbi, FAULT_BLOCK)) {
2229 * let's increase this in prior to actual block count change in order
2230 * for f2fs_sync_file to avoid data races when deciding checkpoint.
2232 percpu_counter_add(&sbi->alloc_valid_block_count, (*count));
2234 spin_lock(&sbi->stat_lock);
2235 sbi->total_valid_block_count += (block_t)(*count);
2236 avail_user_block_count = sbi->user_block_count -
2237 sbi->current_reserved_blocks;
2239 if (!__allow_reserved_blocks(sbi, inode, true))
2240 avail_user_block_count -= F2FS_OPTION(sbi).root_reserved_blocks;
2242 if (F2FS_IO_ALIGNED(sbi))
2243 avail_user_block_count -= sbi->blocks_per_seg *
2244 SM_I(sbi)->additional_reserved_segments;
2246 if (unlikely(is_sbi_flag_set(sbi, SBI_CP_DISABLED))) {
2247 if (avail_user_block_count > sbi->unusable_block_count)
2248 avail_user_block_count -= sbi->unusable_block_count;
2250 avail_user_block_count = 0;
2252 if (unlikely(sbi->total_valid_block_count > avail_user_block_count)) {
2253 diff = sbi->total_valid_block_count - avail_user_block_count;
2258 sbi->total_valid_block_count -= diff;
2260 spin_unlock(&sbi->stat_lock);
2264 spin_unlock(&sbi->stat_lock);
2266 if (unlikely(release)) {
2267 percpu_counter_sub(&sbi->alloc_valid_block_count, release);
2268 dquot_release_reservation_block(inode, release);
2270 f2fs_i_blocks_write(inode, *count, true, true);
2274 percpu_counter_sub(&sbi->alloc_valid_block_count, release);
2276 dquot_release_reservation_block(inode, release);
2281 void f2fs_printk(struct f2fs_sb_info *sbi, const char *fmt, ...);
2283 #define f2fs_err(sbi, fmt, ...) \
2284 f2fs_printk(sbi, KERN_ERR fmt, ##__VA_ARGS__)
2285 #define f2fs_warn(sbi, fmt, ...) \
2286 f2fs_printk(sbi, KERN_WARNING fmt, ##__VA_ARGS__)
2287 #define f2fs_notice(sbi, fmt, ...) \
2288 f2fs_printk(sbi, KERN_NOTICE fmt, ##__VA_ARGS__)
2289 #define f2fs_info(sbi, fmt, ...) \
2290 f2fs_printk(sbi, KERN_INFO fmt, ##__VA_ARGS__)
2291 #define f2fs_debug(sbi, fmt, ...) \
2292 f2fs_printk(sbi, KERN_DEBUG fmt, ##__VA_ARGS__)
2294 #define PAGE_PRIVATE_GET_FUNC(name, flagname) \
2295 static inline bool page_private_##name(struct page *page) \
2297 return PagePrivate(page) && \
2298 test_bit(PAGE_PRIVATE_NOT_POINTER, &page_private(page)) && \
2299 test_bit(PAGE_PRIVATE_##flagname, &page_private(page)); \
2302 #define PAGE_PRIVATE_SET_FUNC(name, flagname) \
2303 static inline void set_page_private_##name(struct page *page) \
2305 if (!PagePrivate(page)) \
2306 attach_page_private(page, (void *)0); \
2307 set_bit(PAGE_PRIVATE_NOT_POINTER, &page_private(page)); \
2308 set_bit(PAGE_PRIVATE_##flagname, &page_private(page)); \
2311 #define PAGE_PRIVATE_CLEAR_FUNC(name, flagname) \
2312 static inline void clear_page_private_##name(struct page *page) \
2314 clear_bit(PAGE_PRIVATE_##flagname, &page_private(page)); \
2315 if (page_private(page) == BIT(PAGE_PRIVATE_NOT_POINTER)) \
2316 detach_page_private(page); \
2319 PAGE_PRIVATE_GET_FUNC(nonpointer, NOT_POINTER);
2320 PAGE_PRIVATE_GET_FUNC(inline, INLINE_INODE);
2321 PAGE_PRIVATE_GET_FUNC(gcing, ONGOING_MIGRATION);
2322 PAGE_PRIVATE_GET_FUNC(dummy, DUMMY_WRITE);
2324 PAGE_PRIVATE_SET_FUNC(reference, REF_RESOURCE);
2325 PAGE_PRIVATE_SET_FUNC(inline, INLINE_INODE);
2326 PAGE_PRIVATE_SET_FUNC(gcing, ONGOING_MIGRATION);
2327 PAGE_PRIVATE_SET_FUNC(dummy, DUMMY_WRITE);
2329 PAGE_PRIVATE_CLEAR_FUNC(reference, REF_RESOURCE);
2330 PAGE_PRIVATE_CLEAR_FUNC(inline, INLINE_INODE);
2331 PAGE_PRIVATE_CLEAR_FUNC(gcing, ONGOING_MIGRATION);
2332 PAGE_PRIVATE_CLEAR_FUNC(dummy, DUMMY_WRITE);
2334 static inline unsigned long get_page_private_data(struct page *page)
2336 unsigned long data = page_private(page);
2338 if (!test_bit(PAGE_PRIVATE_NOT_POINTER, &data))
2340 return data >> PAGE_PRIVATE_MAX;
2343 static inline void set_page_private_data(struct page *page, unsigned long data)
2345 if (!PagePrivate(page))
2346 attach_page_private(page, (void *)0);
2347 set_bit(PAGE_PRIVATE_NOT_POINTER, &page_private(page));
2348 page_private(page) |= data << PAGE_PRIVATE_MAX;
2351 static inline void clear_page_private_data(struct page *page)
2353 page_private(page) &= GENMASK(PAGE_PRIVATE_MAX - 1, 0);
2354 if (page_private(page) == BIT(PAGE_PRIVATE_NOT_POINTER))
2355 detach_page_private(page);
2358 static inline void clear_page_private_all(struct page *page)
2360 clear_page_private_data(page);
2361 clear_page_private_reference(page);
2362 clear_page_private_gcing(page);
2363 clear_page_private_inline(page);
2365 f2fs_bug_on(F2FS_P_SB(page), page_private(page));
2368 static inline void dec_valid_block_count(struct f2fs_sb_info *sbi,
2369 struct inode *inode,
2372 blkcnt_t sectors = count << F2FS_LOG_SECTORS_PER_BLOCK;
2374 spin_lock(&sbi->stat_lock);
2375 f2fs_bug_on(sbi, sbi->total_valid_block_count < (block_t) count);
2376 sbi->total_valid_block_count -= (block_t)count;
2377 if (sbi->reserved_blocks &&
2378 sbi->current_reserved_blocks < sbi->reserved_blocks)
2379 sbi->current_reserved_blocks = min(sbi->reserved_blocks,
2380 sbi->current_reserved_blocks + count);
2381 spin_unlock(&sbi->stat_lock);
2382 if (unlikely(inode->i_blocks < sectors)) {
2383 f2fs_warn(sbi, "Inconsistent i_blocks, ino:%lu, iblocks:%llu, sectors:%llu",
2385 (unsigned long long)inode->i_blocks,
2386 (unsigned long long)sectors);
2387 set_sbi_flag(sbi, SBI_NEED_FSCK);
2390 f2fs_i_blocks_write(inode, count, false, true);
2393 static inline void inc_page_count(struct f2fs_sb_info *sbi, int count_type)
2395 atomic_inc(&sbi->nr_pages[count_type]);
2397 if (count_type == F2FS_DIRTY_DENTS ||
2398 count_type == F2FS_DIRTY_NODES ||
2399 count_type == F2FS_DIRTY_META ||
2400 count_type == F2FS_DIRTY_QDATA ||
2401 count_type == F2FS_DIRTY_IMETA)
2402 set_sbi_flag(sbi, SBI_IS_DIRTY);
2405 static inline void inode_inc_dirty_pages(struct inode *inode)
2407 atomic_inc(&F2FS_I(inode)->dirty_pages);
2408 inc_page_count(F2FS_I_SB(inode), S_ISDIR(inode->i_mode) ?
2409 F2FS_DIRTY_DENTS : F2FS_DIRTY_DATA);
2410 if (IS_NOQUOTA(inode))
2411 inc_page_count(F2FS_I_SB(inode), F2FS_DIRTY_QDATA);
2414 static inline void dec_page_count(struct f2fs_sb_info *sbi, int count_type)
2416 atomic_dec(&sbi->nr_pages[count_type]);
2419 static inline void inode_dec_dirty_pages(struct inode *inode)
2421 if (!S_ISDIR(inode->i_mode) && !S_ISREG(inode->i_mode) &&
2422 !S_ISLNK(inode->i_mode))
2425 atomic_dec(&F2FS_I(inode)->dirty_pages);
2426 dec_page_count(F2FS_I_SB(inode), S_ISDIR(inode->i_mode) ?
2427 F2FS_DIRTY_DENTS : F2FS_DIRTY_DATA);
2428 if (IS_NOQUOTA(inode))
2429 dec_page_count(F2FS_I_SB(inode), F2FS_DIRTY_QDATA);
2432 static inline void inc_atomic_write_cnt(struct inode *inode)
2434 struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
2435 struct f2fs_inode_info *fi = F2FS_I(inode);
2438 fi->atomic_write_cnt++;
2439 atomic64_inc(&sbi->current_atomic_write);
2440 current_write = atomic64_read(&sbi->current_atomic_write);
2441 if (current_write > sbi->peak_atomic_write)
2442 sbi->peak_atomic_write = current_write;
2445 static inline void release_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);
2450 atomic64_sub(fi->atomic_write_cnt, &sbi->current_atomic_write);
2451 fi->atomic_write_cnt = 0;
2454 static inline s64 get_pages(struct f2fs_sb_info *sbi, int count_type)
2456 return atomic_read(&sbi->nr_pages[count_type]);
2459 static inline int get_dirty_pages(struct inode *inode)
2461 return atomic_read(&F2FS_I(inode)->dirty_pages);
2464 static inline int get_blocktype_secs(struct f2fs_sb_info *sbi, int block_type)
2466 unsigned int pages_per_sec = sbi->segs_per_sec * sbi->blocks_per_seg;
2467 unsigned int segs = (get_pages(sbi, block_type) + pages_per_sec - 1) >>
2468 sbi->log_blocks_per_seg;
2470 return segs / sbi->segs_per_sec;
2473 static inline block_t valid_user_blocks(struct f2fs_sb_info *sbi)
2475 return sbi->total_valid_block_count;
2478 static inline block_t discard_blocks(struct f2fs_sb_info *sbi)
2480 return sbi->discard_blks;
2483 static inline unsigned long __bitmap_size(struct f2fs_sb_info *sbi, int flag)
2485 struct f2fs_checkpoint *ckpt = F2FS_CKPT(sbi);
2487 /* return NAT or SIT bitmap */
2488 if (flag == NAT_BITMAP)
2489 return le32_to_cpu(ckpt->nat_ver_bitmap_bytesize);
2490 else if (flag == SIT_BITMAP)
2491 return le32_to_cpu(ckpt->sit_ver_bitmap_bytesize);
2496 static inline block_t __cp_payload(struct f2fs_sb_info *sbi)
2498 return le32_to_cpu(F2FS_RAW_SUPER(sbi)->cp_payload);
2501 static inline void *__bitmap_ptr(struct f2fs_sb_info *sbi, int flag)
2503 struct f2fs_checkpoint *ckpt = F2FS_CKPT(sbi);
2504 void *tmp_ptr = &ckpt->sit_nat_version_bitmap;
2507 if (is_set_ckpt_flags(sbi, CP_LARGE_NAT_BITMAP_FLAG)) {
2508 offset = (flag == SIT_BITMAP) ?
2509 le32_to_cpu(ckpt->nat_ver_bitmap_bytesize) : 0;
2511 * if large_nat_bitmap feature is enabled, leave checksum
2512 * protection for all nat/sit bitmaps.
2514 return tmp_ptr + offset + sizeof(__le32);
2517 if (__cp_payload(sbi) > 0) {
2518 if (flag == NAT_BITMAP)
2521 return (unsigned char *)ckpt + F2FS_BLKSIZE;
2523 offset = (flag == NAT_BITMAP) ?
2524 le32_to_cpu(ckpt->sit_ver_bitmap_bytesize) : 0;
2525 return tmp_ptr + offset;
2529 static inline block_t __start_cp_addr(struct f2fs_sb_info *sbi)
2531 block_t start_addr = le32_to_cpu(F2FS_RAW_SUPER(sbi)->cp_blkaddr);
2533 if (sbi->cur_cp_pack == 2)
2534 start_addr += sbi->blocks_per_seg;
2538 static inline block_t __start_cp_next_addr(struct f2fs_sb_info *sbi)
2540 block_t start_addr = le32_to_cpu(F2FS_RAW_SUPER(sbi)->cp_blkaddr);
2542 if (sbi->cur_cp_pack == 1)
2543 start_addr += sbi->blocks_per_seg;
2547 static inline void __set_cp_next_pack(struct f2fs_sb_info *sbi)
2549 sbi->cur_cp_pack = (sbi->cur_cp_pack == 1) ? 2 : 1;
2552 static inline block_t __start_sum_addr(struct f2fs_sb_info *sbi)
2554 return le32_to_cpu(F2FS_CKPT(sbi)->cp_pack_start_sum);
2557 extern void f2fs_mark_inode_dirty_sync(struct inode *inode, bool sync);
2558 static inline int inc_valid_node_count(struct f2fs_sb_info *sbi,
2559 struct inode *inode, bool is_inode)
2561 block_t valid_block_count;
2562 unsigned int valid_node_count, user_block_count;
2567 err = dquot_alloc_inode(inode);
2572 err = dquot_reserve_block(inode, 1);
2577 if (time_to_inject(sbi, FAULT_BLOCK))
2580 spin_lock(&sbi->stat_lock);
2582 valid_block_count = sbi->total_valid_block_count +
2583 sbi->current_reserved_blocks + 1;
2585 if (!__allow_reserved_blocks(sbi, inode, false))
2586 valid_block_count += F2FS_OPTION(sbi).root_reserved_blocks;
2588 if (F2FS_IO_ALIGNED(sbi))
2589 valid_block_count += sbi->blocks_per_seg *
2590 SM_I(sbi)->additional_reserved_segments;
2592 user_block_count = sbi->user_block_count;
2593 if (unlikely(is_sbi_flag_set(sbi, SBI_CP_DISABLED)))
2594 user_block_count -= sbi->unusable_block_count;
2596 if (unlikely(valid_block_count > user_block_count)) {
2597 spin_unlock(&sbi->stat_lock);
2601 valid_node_count = sbi->total_valid_node_count + 1;
2602 if (unlikely(valid_node_count > sbi->total_node_count)) {
2603 spin_unlock(&sbi->stat_lock);
2607 sbi->total_valid_node_count++;
2608 sbi->total_valid_block_count++;
2609 spin_unlock(&sbi->stat_lock);
2613 f2fs_mark_inode_dirty_sync(inode, true);
2615 f2fs_i_blocks_write(inode, 1, true, true);
2618 percpu_counter_inc(&sbi->alloc_valid_block_count);
2624 dquot_free_inode(inode);
2626 dquot_release_reservation_block(inode, 1);
2631 static inline void dec_valid_node_count(struct f2fs_sb_info *sbi,
2632 struct inode *inode, bool is_inode)
2634 spin_lock(&sbi->stat_lock);
2636 if (unlikely(!sbi->total_valid_block_count ||
2637 !sbi->total_valid_node_count)) {
2638 f2fs_warn(sbi, "dec_valid_node_count: inconsistent block counts, total_valid_block:%u, total_valid_node:%u",
2639 sbi->total_valid_block_count,
2640 sbi->total_valid_node_count);
2641 set_sbi_flag(sbi, SBI_NEED_FSCK);
2643 sbi->total_valid_block_count--;
2644 sbi->total_valid_node_count--;
2647 if (sbi->reserved_blocks &&
2648 sbi->current_reserved_blocks < sbi->reserved_blocks)
2649 sbi->current_reserved_blocks++;
2651 spin_unlock(&sbi->stat_lock);
2654 dquot_free_inode(inode);
2656 if (unlikely(inode->i_blocks == 0)) {
2657 f2fs_warn(sbi, "dec_valid_node_count: inconsistent i_blocks, ino:%lu, iblocks:%llu",
2659 (unsigned long long)inode->i_blocks);
2660 set_sbi_flag(sbi, SBI_NEED_FSCK);
2663 f2fs_i_blocks_write(inode, 1, false, true);
2667 static inline unsigned int valid_node_count(struct f2fs_sb_info *sbi)
2669 return sbi->total_valid_node_count;
2672 static inline void inc_valid_inode_count(struct f2fs_sb_info *sbi)
2674 percpu_counter_inc(&sbi->total_valid_inode_count);
2677 static inline void dec_valid_inode_count(struct f2fs_sb_info *sbi)
2679 percpu_counter_dec(&sbi->total_valid_inode_count);
2682 static inline s64 valid_inode_count(struct f2fs_sb_info *sbi)
2684 return percpu_counter_sum_positive(&sbi->total_valid_inode_count);
2687 static inline struct page *f2fs_grab_cache_page(struct address_space *mapping,
2688 pgoff_t index, bool for_write)
2693 if (IS_ENABLED(CONFIG_F2FS_FAULT_INJECTION)) {
2695 page = find_get_page_flags(mapping, index,
2696 FGP_LOCK | FGP_ACCESSED);
2698 page = find_lock_page(mapping, index);
2702 if (time_to_inject(F2FS_M_SB(mapping), FAULT_PAGE_ALLOC))
2707 return grab_cache_page(mapping, index);
2709 flags = memalloc_nofs_save();
2710 page = grab_cache_page_write_begin(mapping, index);
2711 memalloc_nofs_restore(flags);
2716 static inline struct page *f2fs_pagecache_get_page(
2717 struct address_space *mapping, pgoff_t index,
2718 int fgp_flags, gfp_t gfp_mask)
2720 if (time_to_inject(F2FS_M_SB(mapping), FAULT_PAGE_GET))
2723 return pagecache_get_page(mapping, index, fgp_flags, gfp_mask);
2726 static inline void f2fs_put_page(struct page *page, int unlock)
2732 f2fs_bug_on(F2FS_P_SB(page), !PageLocked(page));
2738 static inline void f2fs_put_dnode(struct dnode_of_data *dn)
2741 f2fs_put_page(dn->node_page, 1);
2742 if (dn->inode_page && dn->node_page != dn->inode_page)
2743 f2fs_put_page(dn->inode_page, 0);
2744 dn->node_page = NULL;
2745 dn->inode_page = NULL;
2748 static inline struct kmem_cache *f2fs_kmem_cache_create(const char *name,
2751 return kmem_cache_create(name, size, 0, SLAB_RECLAIM_ACCOUNT, NULL);
2754 static inline void *f2fs_kmem_cache_alloc_nofail(struct kmem_cache *cachep,
2759 entry = kmem_cache_alloc(cachep, flags);
2761 entry = kmem_cache_alloc(cachep, flags | __GFP_NOFAIL);
2765 static inline void *f2fs_kmem_cache_alloc(struct kmem_cache *cachep,
2766 gfp_t flags, bool nofail, struct f2fs_sb_info *sbi)
2769 return f2fs_kmem_cache_alloc_nofail(cachep, flags);
2771 if (time_to_inject(sbi, FAULT_SLAB_ALLOC))
2774 return kmem_cache_alloc(cachep, flags);
2777 static inline bool is_inflight_io(struct f2fs_sb_info *sbi, int type)
2779 if (get_pages(sbi, F2FS_RD_DATA) || get_pages(sbi, F2FS_RD_NODE) ||
2780 get_pages(sbi, F2FS_RD_META) || get_pages(sbi, F2FS_WB_DATA) ||
2781 get_pages(sbi, F2FS_WB_CP_DATA) ||
2782 get_pages(sbi, F2FS_DIO_READ) ||
2783 get_pages(sbi, F2FS_DIO_WRITE))
2786 if (type != DISCARD_TIME && SM_I(sbi) && SM_I(sbi)->dcc_info &&
2787 atomic_read(&SM_I(sbi)->dcc_info->queued_discard))
2790 if (SM_I(sbi) && SM_I(sbi)->fcc_info &&
2791 atomic_read(&SM_I(sbi)->fcc_info->queued_flush))
2796 static inline bool is_idle(struct f2fs_sb_info *sbi, int type)
2798 if (sbi->gc_mode == GC_URGENT_HIGH)
2801 if (is_inflight_io(sbi, type))
2804 if (sbi->gc_mode == GC_URGENT_MID)
2807 if (sbi->gc_mode == GC_URGENT_LOW &&
2808 (type == DISCARD_TIME || type == GC_TIME))
2811 return f2fs_time_over(sbi, type);
2814 static inline void f2fs_radix_tree_insert(struct radix_tree_root *root,
2815 unsigned long index, void *item)
2817 while (radix_tree_insert(root, index, item))
2821 #define RAW_IS_INODE(p) ((p)->footer.nid == (p)->footer.ino)
2823 static inline bool IS_INODE(struct page *page)
2825 struct f2fs_node *p = F2FS_NODE(page);
2827 return RAW_IS_INODE(p);
2830 static inline int offset_in_addr(struct f2fs_inode *i)
2832 return (i->i_inline & F2FS_EXTRA_ATTR) ?
2833 (le16_to_cpu(i->i_extra_isize) / sizeof(__le32)) : 0;
2836 static inline __le32 *blkaddr_in_node(struct f2fs_node *node)
2838 return RAW_IS_INODE(node) ? node->i.i_addr : node->dn.addr;
2841 static inline int f2fs_has_extra_attr(struct inode *inode);
2842 static inline block_t data_blkaddr(struct inode *inode,
2843 struct page *node_page, unsigned int offset)
2845 struct f2fs_node *raw_node;
2848 bool is_inode = IS_INODE(node_page);
2850 raw_node = F2FS_NODE(node_page);
2854 /* from GC path only */
2855 base = offset_in_addr(&raw_node->i);
2856 else if (f2fs_has_extra_attr(inode))
2857 base = get_extra_isize(inode);
2860 addr_array = blkaddr_in_node(raw_node);
2861 return le32_to_cpu(addr_array[base + offset]);
2864 static inline block_t f2fs_data_blkaddr(struct dnode_of_data *dn)
2866 return data_blkaddr(dn->inode, dn->node_page, dn->ofs_in_node);
2869 static inline int f2fs_test_bit(unsigned int nr, char *addr)
2874 mask = BIT(7 - (nr & 0x07));
2875 return mask & *addr;
2878 static inline void f2fs_set_bit(unsigned int nr, char *addr)
2883 mask = BIT(7 - (nr & 0x07));
2887 static inline void f2fs_clear_bit(unsigned int nr, char *addr)
2892 mask = BIT(7 - (nr & 0x07));
2896 static inline int f2fs_test_and_set_bit(unsigned int nr, char *addr)
2902 mask = BIT(7 - (nr & 0x07));
2908 static inline int f2fs_test_and_clear_bit(unsigned int nr, char *addr)
2914 mask = BIT(7 - (nr & 0x07));
2920 static inline void f2fs_change_bit(unsigned int nr, char *addr)
2925 mask = BIT(7 - (nr & 0x07));
2930 * On-disk inode flags (f2fs_inode::i_flags)
2932 #define F2FS_COMPR_FL 0x00000004 /* Compress file */
2933 #define F2FS_SYNC_FL 0x00000008 /* Synchronous updates */
2934 #define F2FS_IMMUTABLE_FL 0x00000010 /* Immutable file */
2935 #define F2FS_APPEND_FL 0x00000020 /* writes to file may only append */
2936 #define F2FS_NODUMP_FL 0x00000040 /* do not dump file */
2937 #define F2FS_NOATIME_FL 0x00000080 /* do not update atime */
2938 #define F2FS_NOCOMP_FL 0x00000400 /* Don't compress */
2939 #define F2FS_INDEX_FL 0x00001000 /* hash-indexed directory */
2940 #define F2FS_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */
2941 #define F2FS_PROJINHERIT_FL 0x20000000 /* Create with parents projid */
2942 #define F2FS_CASEFOLD_FL 0x40000000 /* Casefolded file */
2944 /* Flags that should be inherited by new inodes from their parent. */
2945 #define F2FS_FL_INHERITED (F2FS_SYNC_FL | F2FS_NODUMP_FL | F2FS_NOATIME_FL | \
2946 F2FS_DIRSYNC_FL | F2FS_PROJINHERIT_FL | \
2949 /* Flags that are appropriate for regular files (all but dir-specific ones). */
2950 #define F2FS_REG_FLMASK (~(F2FS_DIRSYNC_FL | F2FS_PROJINHERIT_FL | \
2953 /* Flags that are appropriate for non-directories/regular files. */
2954 #define F2FS_OTHER_FLMASK (F2FS_NODUMP_FL | F2FS_NOATIME_FL)
2956 static inline __u32 f2fs_mask_flags(umode_t mode, __u32 flags)
2960 else if (S_ISREG(mode))
2961 return flags & F2FS_REG_FLMASK;
2963 return flags & F2FS_OTHER_FLMASK;
2966 static inline void __mark_inode_dirty_flag(struct inode *inode,
2970 case FI_INLINE_XATTR:
2971 case FI_INLINE_DATA:
2972 case FI_INLINE_DENTRY:
2978 case FI_INLINE_DOTS:
2980 case FI_COMPRESS_RELEASED:
2981 f2fs_mark_inode_dirty_sync(inode, true);
2985 static inline void set_inode_flag(struct inode *inode, int flag)
2987 set_bit(flag, F2FS_I(inode)->flags);
2988 __mark_inode_dirty_flag(inode, flag, true);
2991 static inline int is_inode_flag_set(struct inode *inode, int flag)
2993 return test_bit(flag, F2FS_I(inode)->flags);
2996 static inline void clear_inode_flag(struct inode *inode, int flag)
2998 clear_bit(flag, F2FS_I(inode)->flags);
2999 __mark_inode_dirty_flag(inode, flag, false);
3002 static inline bool f2fs_verity_in_progress(struct inode *inode)
3004 return IS_ENABLED(CONFIG_FS_VERITY) &&
3005 is_inode_flag_set(inode, FI_VERITY_IN_PROGRESS);
3008 static inline void set_acl_inode(struct inode *inode, umode_t mode)
3010 F2FS_I(inode)->i_acl_mode = mode;
3011 set_inode_flag(inode, FI_ACL_MODE);
3012 f2fs_mark_inode_dirty_sync(inode, false);
3015 static inline void f2fs_i_links_write(struct inode *inode, bool inc)
3021 f2fs_mark_inode_dirty_sync(inode, true);
3024 static inline void f2fs_i_blocks_write(struct inode *inode,
3025 block_t diff, bool add, bool claim)
3027 bool clean = !is_inode_flag_set(inode, FI_DIRTY_INODE);
3028 bool recover = is_inode_flag_set(inode, FI_AUTO_RECOVER);
3030 /* add = 1, claim = 1 should be dquot_reserve_block in pair */
3033 dquot_claim_block(inode, diff);
3035 dquot_alloc_block_nofail(inode, diff);
3037 dquot_free_block(inode, diff);
3040 f2fs_mark_inode_dirty_sync(inode, true);
3041 if (clean || recover)
3042 set_inode_flag(inode, FI_AUTO_RECOVER);
3045 static inline bool f2fs_is_atomic_file(struct inode *inode);
3047 static inline void f2fs_i_size_write(struct inode *inode, loff_t i_size)
3049 bool clean = !is_inode_flag_set(inode, FI_DIRTY_INODE);
3050 bool recover = is_inode_flag_set(inode, FI_AUTO_RECOVER);
3052 if (i_size_read(inode) == i_size)
3055 i_size_write(inode, i_size);
3057 if (f2fs_is_atomic_file(inode))
3060 f2fs_mark_inode_dirty_sync(inode, true);
3061 if (clean || recover)
3062 set_inode_flag(inode, FI_AUTO_RECOVER);
3065 static inline void f2fs_i_depth_write(struct inode *inode, unsigned int depth)
3067 F2FS_I(inode)->i_current_depth = depth;
3068 f2fs_mark_inode_dirty_sync(inode, true);
3071 static inline void f2fs_i_gc_failures_write(struct inode *inode,
3074 F2FS_I(inode)->i_gc_failures[GC_FAILURE_PIN] = count;
3075 f2fs_mark_inode_dirty_sync(inode, true);
3078 static inline void f2fs_i_xnid_write(struct inode *inode, nid_t xnid)
3080 F2FS_I(inode)->i_xattr_nid = xnid;
3081 f2fs_mark_inode_dirty_sync(inode, true);
3084 static inline void f2fs_i_pino_write(struct inode *inode, nid_t pino)
3086 F2FS_I(inode)->i_pino = pino;
3087 f2fs_mark_inode_dirty_sync(inode, true);
3090 static inline void get_inline_info(struct inode *inode, struct f2fs_inode *ri)
3092 struct f2fs_inode_info *fi = F2FS_I(inode);
3094 if (ri->i_inline & F2FS_INLINE_XATTR)
3095 set_bit(FI_INLINE_XATTR, fi->flags);
3096 if (ri->i_inline & F2FS_INLINE_DATA)
3097 set_bit(FI_INLINE_DATA, fi->flags);
3098 if (ri->i_inline & F2FS_INLINE_DENTRY)
3099 set_bit(FI_INLINE_DENTRY, fi->flags);
3100 if (ri->i_inline & F2FS_DATA_EXIST)
3101 set_bit(FI_DATA_EXIST, fi->flags);
3102 if (ri->i_inline & F2FS_INLINE_DOTS)
3103 set_bit(FI_INLINE_DOTS, fi->flags);
3104 if (ri->i_inline & F2FS_EXTRA_ATTR)
3105 set_bit(FI_EXTRA_ATTR, fi->flags);
3106 if (ri->i_inline & F2FS_PIN_FILE)
3107 set_bit(FI_PIN_FILE, fi->flags);
3108 if (ri->i_inline & F2FS_COMPRESS_RELEASED)
3109 set_bit(FI_COMPRESS_RELEASED, fi->flags);
3112 static inline void set_raw_inline(struct inode *inode, struct f2fs_inode *ri)
3116 if (is_inode_flag_set(inode, FI_INLINE_XATTR))
3117 ri->i_inline |= F2FS_INLINE_XATTR;
3118 if (is_inode_flag_set(inode, FI_INLINE_DATA))
3119 ri->i_inline |= F2FS_INLINE_DATA;
3120 if (is_inode_flag_set(inode, FI_INLINE_DENTRY))
3121 ri->i_inline |= F2FS_INLINE_DENTRY;
3122 if (is_inode_flag_set(inode, FI_DATA_EXIST))
3123 ri->i_inline |= F2FS_DATA_EXIST;
3124 if (is_inode_flag_set(inode, FI_INLINE_DOTS))
3125 ri->i_inline |= F2FS_INLINE_DOTS;
3126 if (is_inode_flag_set(inode, FI_EXTRA_ATTR))
3127 ri->i_inline |= F2FS_EXTRA_ATTR;
3128 if (is_inode_flag_set(inode, FI_PIN_FILE))
3129 ri->i_inline |= F2FS_PIN_FILE;
3130 if (is_inode_flag_set(inode, FI_COMPRESS_RELEASED))
3131 ri->i_inline |= F2FS_COMPRESS_RELEASED;
3134 static inline int f2fs_has_extra_attr(struct inode *inode)
3136 return is_inode_flag_set(inode, FI_EXTRA_ATTR);
3139 static inline int f2fs_has_inline_xattr(struct inode *inode)
3141 return is_inode_flag_set(inode, FI_INLINE_XATTR);
3144 static inline int f2fs_compressed_file(struct inode *inode)
3146 return S_ISREG(inode->i_mode) &&
3147 is_inode_flag_set(inode, FI_COMPRESSED_FILE);
3150 static inline bool f2fs_need_compress_data(struct inode *inode)
3152 int compress_mode = F2FS_OPTION(F2FS_I_SB(inode)).compress_mode;
3154 if (!f2fs_compressed_file(inode))
3157 if (compress_mode == COMPR_MODE_FS)
3159 else if (compress_mode == COMPR_MODE_USER &&
3160 is_inode_flag_set(inode, FI_ENABLE_COMPRESS))
3166 static inline unsigned int addrs_per_inode(struct inode *inode)
3168 unsigned int addrs = CUR_ADDRS_PER_INODE(inode) -
3169 get_inline_xattr_addrs(inode);
3171 if (!f2fs_compressed_file(inode))
3173 return ALIGN_DOWN(addrs, F2FS_I(inode)->i_cluster_size);
3176 static inline unsigned int addrs_per_block(struct inode *inode)
3178 if (!f2fs_compressed_file(inode))
3179 return DEF_ADDRS_PER_BLOCK;
3180 return ALIGN_DOWN(DEF_ADDRS_PER_BLOCK, F2FS_I(inode)->i_cluster_size);
3183 static inline void *inline_xattr_addr(struct inode *inode, struct page *page)
3185 struct f2fs_inode *ri = F2FS_INODE(page);
3187 return (void *)&(ri->i_addr[DEF_ADDRS_PER_INODE -
3188 get_inline_xattr_addrs(inode)]);
3191 static inline int inline_xattr_size(struct inode *inode)
3193 if (f2fs_has_inline_xattr(inode))
3194 return get_inline_xattr_addrs(inode) * sizeof(__le32);
3199 * Notice: check inline_data flag without inode page lock is unsafe.
3200 * It could change at any time by f2fs_convert_inline_page().
3202 static inline int f2fs_has_inline_data(struct inode *inode)
3204 return is_inode_flag_set(inode, FI_INLINE_DATA);
3207 static inline int f2fs_exist_data(struct inode *inode)
3209 return is_inode_flag_set(inode, FI_DATA_EXIST);
3212 static inline int f2fs_has_inline_dots(struct inode *inode)
3214 return is_inode_flag_set(inode, FI_INLINE_DOTS);
3217 static inline int f2fs_is_mmap_file(struct inode *inode)
3219 return is_inode_flag_set(inode, FI_MMAP_FILE);
3222 static inline bool f2fs_is_pinned_file(struct inode *inode)
3224 return is_inode_flag_set(inode, FI_PIN_FILE);
3227 static inline bool f2fs_is_atomic_file(struct inode *inode)
3229 return is_inode_flag_set(inode, FI_ATOMIC_FILE);
3232 static inline bool f2fs_is_cow_file(struct inode *inode)
3234 return is_inode_flag_set(inode, FI_COW_FILE);
3237 static inline bool f2fs_is_first_block_written(struct inode *inode)
3239 return is_inode_flag_set(inode, FI_FIRST_BLOCK_WRITTEN);
3242 static inline bool f2fs_is_drop_cache(struct inode *inode)
3244 return is_inode_flag_set(inode, FI_DROP_CACHE);
3247 static inline void *inline_data_addr(struct inode *inode, struct page *page)
3249 struct f2fs_inode *ri = F2FS_INODE(page);
3250 int extra_size = get_extra_isize(inode);
3252 return (void *)&(ri->i_addr[extra_size + DEF_INLINE_RESERVED_SIZE]);
3255 static inline int f2fs_has_inline_dentry(struct inode *inode)
3257 return is_inode_flag_set(inode, FI_INLINE_DENTRY);
3260 static inline int is_file(struct inode *inode, int type)
3262 return F2FS_I(inode)->i_advise & type;
3265 static inline void set_file(struct inode *inode, int type)
3267 if (is_file(inode, type))
3269 F2FS_I(inode)->i_advise |= type;
3270 f2fs_mark_inode_dirty_sync(inode, true);
3273 static inline void clear_file(struct inode *inode, int type)
3275 if (!is_file(inode, type))
3277 F2FS_I(inode)->i_advise &= ~type;
3278 f2fs_mark_inode_dirty_sync(inode, true);
3281 static inline bool f2fs_is_time_consistent(struct inode *inode)
3283 if (!timespec64_equal(F2FS_I(inode)->i_disk_time, &inode->i_atime))
3285 if (!timespec64_equal(F2FS_I(inode)->i_disk_time + 1, &inode->i_ctime))
3287 if (!timespec64_equal(F2FS_I(inode)->i_disk_time + 2, &inode->i_mtime))
3292 static inline bool f2fs_skip_inode_update(struct inode *inode, int dsync)
3297 struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
3299 spin_lock(&sbi->inode_lock[DIRTY_META]);
3300 ret = list_empty(&F2FS_I(inode)->gdirty_list);
3301 spin_unlock(&sbi->inode_lock[DIRTY_META]);
3304 if (!is_inode_flag_set(inode, FI_AUTO_RECOVER) ||
3305 file_keep_isize(inode) ||
3306 i_size_read(inode) & ~PAGE_MASK)
3309 if (!f2fs_is_time_consistent(inode))
3312 spin_lock(&F2FS_I(inode)->i_size_lock);
3313 ret = F2FS_I(inode)->last_disk_size == i_size_read(inode);
3314 spin_unlock(&F2FS_I(inode)->i_size_lock);
3319 static inline bool f2fs_readonly(struct super_block *sb)
3321 return sb_rdonly(sb);
3324 static inline bool f2fs_cp_error(struct f2fs_sb_info *sbi)
3326 return is_set_ckpt_flags(sbi, CP_ERROR_FLAG);
3329 static inline bool is_dot_dotdot(const u8 *name, size_t len)
3331 if (len == 1 && name[0] == '.')
3334 if (len == 2 && name[0] == '.' && name[1] == '.')
3340 static inline void *f2fs_kmalloc(struct f2fs_sb_info *sbi,
3341 size_t size, gfp_t flags)
3343 if (time_to_inject(sbi, FAULT_KMALLOC))
3346 return kmalloc(size, flags);
3349 static inline void *f2fs_getname(struct f2fs_sb_info *sbi)
3351 if (time_to_inject(sbi, FAULT_KMALLOC))
3357 static inline void f2fs_putname(char *buf)
3362 static inline void *f2fs_kzalloc(struct f2fs_sb_info *sbi,
3363 size_t size, gfp_t flags)
3365 return f2fs_kmalloc(sbi, size, flags | __GFP_ZERO);
3368 static inline void *f2fs_kvmalloc(struct f2fs_sb_info *sbi,
3369 size_t size, gfp_t flags)
3371 if (time_to_inject(sbi, FAULT_KVMALLOC))
3374 return kvmalloc(size, flags);
3377 static inline void *f2fs_kvzalloc(struct f2fs_sb_info *sbi,
3378 size_t size, gfp_t flags)
3380 return f2fs_kvmalloc(sbi, size, flags | __GFP_ZERO);
3383 static inline int get_extra_isize(struct inode *inode)
3385 return F2FS_I(inode)->i_extra_isize / sizeof(__le32);
3388 static inline int get_inline_xattr_addrs(struct inode *inode)
3390 return F2FS_I(inode)->i_inline_xattr_size;
3393 #define f2fs_get_inode_mode(i) \
3394 ((is_inode_flag_set(i, FI_ACL_MODE)) ? \
3395 (F2FS_I(i)->i_acl_mode) : ((i)->i_mode))
3397 #define F2FS_TOTAL_EXTRA_ATTR_SIZE \
3398 (offsetof(struct f2fs_inode, i_extra_end) - \
3399 offsetof(struct f2fs_inode, i_extra_isize)) \
3401 #define F2FS_OLD_ATTRIBUTE_SIZE (offsetof(struct f2fs_inode, i_addr))
3402 #define F2FS_FITS_IN_INODE(f2fs_inode, extra_isize, field) \
3403 ((offsetof(typeof(*(f2fs_inode)), field) + \
3404 sizeof((f2fs_inode)->field)) \
3405 <= (F2FS_OLD_ATTRIBUTE_SIZE + (extra_isize))) \
3407 #define __is_large_section(sbi) ((sbi)->segs_per_sec > 1)
3409 #define __is_meta_io(fio) (PAGE_TYPE_OF_BIO((fio)->type) == META)
3411 bool f2fs_is_valid_blkaddr(struct f2fs_sb_info *sbi,
3412 block_t blkaddr, int type);
3413 static inline void verify_blkaddr(struct f2fs_sb_info *sbi,
3414 block_t blkaddr, int type)
3416 if (!f2fs_is_valid_blkaddr(sbi, blkaddr, type)) {
3417 f2fs_err(sbi, "invalid blkaddr: %u, type: %d, run fsck to fix.",
3419 f2fs_bug_on(sbi, 1);
3423 static inline bool __is_valid_data_blkaddr(block_t blkaddr)
3425 if (blkaddr == NEW_ADDR || blkaddr == NULL_ADDR ||
3426 blkaddr == COMPRESS_ADDR)
3434 int f2fs_sync_file(struct file *file, loff_t start, loff_t end, int datasync);
3435 void f2fs_truncate_data_blocks(struct dnode_of_data *dn);
3436 int f2fs_do_truncate_blocks(struct inode *inode, u64 from, bool lock);
3437 int f2fs_truncate_blocks(struct inode *inode, u64 from, bool lock);
3438 int f2fs_truncate(struct inode *inode);
3439 int f2fs_getattr(struct mnt_idmap *idmap, const struct path *path,
3440 struct kstat *stat, u32 request_mask, unsigned int flags);
3441 int f2fs_setattr(struct mnt_idmap *idmap, struct dentry *dentry,
3442 struct iattr *attr);
3443 int f2fs_truncate_hole(struct inode *inode, pgoff_t pg_start, pgoff_t pg_end);
3444 void f2fs_truncate_data_blocks_range(struct dnode_of_data *dn, int count);
3445 int f2fs_precache_extents(struct inode *inode);
3446 int f2fs_fileattr_get(struct dentry *dentry, struct fileattr *fa);
3447 int f2fs_fileattr_set(struct mnt_idmap *idmap,
3448 struct dentry *dentry, struct fileattr *fa);
3449 long f2fs_ioctl(struct file *filp, unsigned int cmd, unsigned long arg);
3450 long f2fs_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
3451 int f2fs_transfer_project_quota(struct inode *inode, kprojid_t kprojid);
3452 int f2fs_pin_file_control(struct inode *inode, bool inc);
3457 void f2fs_set_inode_flags(struct inode *inode);
3458 bool f2fs_inode_chksum_verify(struct f2fs_sb_info *sbi, struct page *page);
3459 void f2fs_inode_chksum_set(struct f2fs_sb_info *sbi, struct page *page);
3460 struct inode *f2fs_iget(struct super_block *sb, unsigned long ino);
3461 struct inode *f2fs_iget_retry(struct super_block *sb, unsigned long ino);
3462 int f2fs_try_to_free_nats(struct f2fs_sb_info *sbi, int nr_shrink);
3463 void f2fs_update_inode(struct inode *inode, struct page *node_page);
3464 void f2fs_update_inode_page(struct inode *inode);
3465 int f2fs_write_inode(struct inode *inode, struct writeback_control *wbc);
3466 void f2fs_evict_inode(struct inode *inode);
3467 void f2fs_handle_failed_inode(struct inode *inode);
3472 int f2fs_update_extension_list(struct f2fs_sb_info *sbi, const char *name,
3473 bool hot, bool set);
3474 struct dentry *f2fs_get_parent(struct dentry *child);
3475 int f2fs_get_tmpfile(struct mnt_idmap *idmap, struct inode *dir,
3476 struct inode **new_inode);
3481 int f2fs_init_casefolded_name(const struct inode *dir,
3482 struct f2fs_filename *fname);
3483 int f2fs_setup_filename(struct inode *dir, const struct qstr *iname,
3484 int lookup, struct f2fs_filename *fname);
3485 int f2fs_prepare_lookup(struct inode *dir, struct dentry *dentry,
3486 struct f2fs_filename *fname);
3487 void f2fs_free_filename(struct f2fs_filename *fname);
3488 struct f2fs_dir_entry *f2fs_find_target_dentry(const struct f2fs_dentry_ptr *d,
3489 const struct f2fs_filename *fname, int *max_slots);
3490 int f2fs_fill_dentries(struct dir_context *ctx, struct f2fs_dentry_ptr *d,
3491 unsigned int start_pos, struct fscrypt_str *fstr);
3492 void f2fs_do_make_empty_dir(struct inode *inode, struct inode *parent,
3493 struct f2fs_dentry_ptr *d);
3494 struct page *f2fs_init_inode_metadata(struct inode *inode, struct inode *dir,
3495 const struct f2fs_filename *fname, struct page *dpage);
3496 void f2fs_update_parent_metadata(struct inode *dir, struct inode *inode,
3497 unsigned int current_depth);
3498 int f2fs_room_for_filename(const void *bitmap, int slots, int max_slots);
3499 void f2fs_drop_nlink(struct inode *dir, struct inode *inode);
3500 struct f2fs_dir_entry *__f2fs_find_entry(struct inode *dir,
3501 const struct f2fs_filename *fname,
3502 struct page **res_page);
3503 struct f2fs_dir_entry *f2fs_find_entry(struct inode *dir,
3504 const struct qstr *child, struct page **res_page);
3505 struct f2fs_dir_entry *f2fs_parent_dir(struct inode *dir, struct page **p);
3506 ino_t f2fs_inode_by_name(struct inode *dir, const struct qstr *qstr,
3507 struct page **page);
3508 void f2fs_set_link(struct inode *dir, struct f2fs_dir_entry *de,
3509 struct page *page, struct inode *inode);
3510 bool f2fs_has_enough_room(struct inode *dir, struct page *ipage,
3511 const struct f2fs_filename *fname);
3512 void f2fs_update_dentry(nid_t ino, umode_t mode, struct f2fs_dentry_ptr *d,
3513 const struct fscrypt_str *name, f2fs_hash_t name_hash,
3514 unsigned int bit_pos);
3515 int f2fs_add_regular_entry(struct inode *dir, const struct f2fs_filename *fname,
3516 struct inode *inode, nid_t ino, umode_t mode);
3517 int f2fs_add_dentry(struct inode *dir, const struct f2fs_filename *fname,
3518 struct inode *inode, nid_t ino, umode_t mode);
3519 int f2fs_do_add_link(struct inode *dir, const struct qstr *name,
3520 struct inode *inode, nid_t ino, umode_t mode);
3521 void f2fs_delete_entry(struct f2fs_dir_entry *dentry, struct page *page,
3522 struct inode *dir, struct inode *inode);
3523 int f2fs_do_tmpfile(struct inode *inode, struct inode *dir);
3524 bool f2fs_empty_dir(struct inode *dir);
3526 static inline int f2fs_add_link(struct dentry *dentry, struct inode *inode)
3528 if (fscrypt_is_nokey_name(dentry))
3530 return f2fs_do_add_link(d_inode(dentry->d_parent), &dentry->d_name,
3531 inode, inode->i_ino, inode->i_mode);
3537 int f2fs_inode_dirtied(struct inode *inode, bool sync);
3538 void f2fs_inode_synced(struct inode *inode);
3539 int f2fs_dquot_initialize(struct inode *inode);
3540 int f2fs_enable_quota_files(struct f2fs_sb_info *sbi, bool rdonly);
3541 int f2fs_quota_sync(struct super_block *sb, int type);
3542 loff_t max_file_blocks(struct inode *inode);
3543 void f2fs_quota_off_umount(struct super_block *sb);
3544 void f2fs_handle_stop(struct f2fs_sb_info *sbi, unsigned char reason);
3545 void f2fs_save_errors(struct f2fs_sb_info *sbi, unsigned char flag);
3546 void f2fs_handle_error(struct f2fs_sb_info *sbi, unsigned char error);
3547 int f2fs_commit_super(struct f2fs_sb_info *sbi, bool recover);
3548 int f2fs_sync_fs(struct super_block *sb, int sync);
3549 int f2fs_sanity_check_ckpt(struct f2fs_sb_info *sbi);
3554 void f2fs_hash_filename(const struct inode *dir, struct f2fs_filename *fname);
3561 int f2fs_check_nid_range(struct f2fs_sb_info *sbi, nid_t nid);
3562 bool f2fs_available_free_memory(struct f2fs_sb_info *sbi, int type);
3563 bool f2fs_in_warm_node_list(struct f2fs_sb_info *sbi, struct page *page);
3564 void f2fs_init_fsync_node_info(struct f2fs_sb_info *sbi);
3565 void f2fs_del_fsync_node_entry(struct f2fs_sb_info *sbi, struct page *page);
3566 void f2fs_reset_fsync_node_info(struct f2fs_sb_info *sbi);
3567 int f2fs_need_dentry_mark(struct f2fs_sb_info *sbi, nid_t nid);
3568 bool f2fs_is_checkpointed_node(struct f2fs_sb_info *sbi, nid_t nid);
3569 bool f2fs_need_inode_block_update(struct f2fs_sb_info *sbi, nid_t ino);
3570 int f2fs_get_node_info(struct f2fs_sb_info *sbi, nid_t nid,
3571 struct node_info *ni, bool checkpoint_context);
3572 pgoff_t f2fs_get_next_page_offset(struct dnode_of_data *dn, pgoff_t pgofs);
3573 int f2fs_get_dnode_of_data(struct dnode_of_data *dn, pgoff_t index, int mode);
3574 int f2fs_truncate_inode_blocks(struct inode *inode, pgoff_t from);
3575 int f2fs_truncate_xattr_node(struct inode *inode);
3576 int f2fs_wait_on_node_pages_writeback(struct f2fs_sb_info *sbi,
3577 unsigned int seq_id);
3578 bool f2fs_nat_bitmap_enabled(struct f2fs_sb_info *sbi);
3579 int f2fs_remove_inode_page(struct inode *inode);
3580 struct page *f2fs_new_inode_page(struct inode *inode);
3581 struct page *f2fs_new_node_page(struct dnode_of_data *dn, unsigned int ofs);
3582 void f2fs_ra_node_page(struct f2fs_sb_info *sbi, nid_t nid);
3583 struct page *f2fs_get_node_page(struct f2fs_sb_info *sbi, pgoff_t nid);
3584 struct page *f2fs_get_node_page_ra(struct page *parent, int start);
3585 int f2fs_move_node_page(struct page *node_page, int gc_type);
3586 void f2fs_flush_inline_data(struct f2fs_sb_info *sbi);
3587 int f2fs_fsync_node_pages(struct f2fs_sb_info *sbi, struct inode *inode,
3588 struct writeback_control *wbc, bool atomic,
3589 unsigned int *seq_id);
3590 int f2fs_sync_node_pages(struct f2fs_sb_info *sbi,
3591 struct writeback_control *wbc,
3592 bool do_balance, enum iostat_type io_type);
3593 int f2fs_build_free_nids(struct f2fs_sb_info *sbi, bool sync, bool mount);
3594 bool f2fs_alloc_nid(struct f2fs_sb_info *sbi, nid_t *nid);
3595 void f2fs_alloc_nid_done(struct f2fs_sb_info *sbi, nid_t nid);
3596 void f2fs_alloc_nid_failed(struct f2fs_sb_info *sbi, nid_t nid);
3597 int f2fs_try_to_free_nids(struct f2fs_sb_info *sbi, int nr_shrink);
3598 int f2fs_recover_inline_xattr(struct inode *inode, struct page *page);
3599 int f2fs_recover_xattr_data(struct inode *inode, struct page *page);
3600 int f2fs_recover_inode_page(struct f2fs_sb_info *sbi, struct page *page);
3601 int f2fs_restore_node_summary(struct f2fs_sb_info *sbi,
3602 unsigned int segno, struct f2fs_summary_block *sum);
3603 void f2fs_enable_nat_bits(struct f2fs_sb_info *sbi);
3604 int f2fs_flush_nat_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc);
3605 int f2fs_build_node_manager(struct f2fs_sb_info *sbi);
3606 void f2fs_destroy_node_manager(struct f2fs_sb_info *sbi);
3607 int __init f2fs_create_node_manager_caches(void);
3608 void f2fs_destroy_node_manager_caches(void);
3613 bool f2fs_need_SSR(struct f2fs_sb_info *sbi);
3614 int f2fs_commit_atomic_write(struct inode *inode);
3615 void f2fs_abort_atomic_write(struct inode *inode, bool clean);
3616 void f2fs_balance_fs(struct f2fs_sb_info *sbi, bool need);
3617 void f2fs_balance_fs_bg(struct f2fs_sb_info *sbi, bool from_bg);
3618 int f2fs_issue_flush(struct f2fs_sb_info *sbi, nid_t ino);
3619 int f2fs_create_flush_cmd_control(struct f2fs_sb_info *sbi);
3620 int f2fs_flush_device_cache(struct f2fs_sb_info *sbi);
3621 void f2fs_destroy_flush_cmd_control(struct f2fs_sb_info *sbi, bool free);
3622 void f2fs_invalidate_blocks(struct f2fs_sb_info *sbi, block_t addr);
3623 bool f2fs_is_checkpointed_data(struct f2fs_sb_info *sbi, block_t blkaddr);
3624 int f2fs_start_discard_thread(struct f2fs_sb_info *sbi);
3625 void f2fs_drop_discard_cmd(struct f2fs_sb_info *sbi);
3626 void f2fs_stop_discard_thread(struct f2fs_sb_info *sbi);
3627 bool f2fs_issue_discard_timeout(struct f2fs_sb_info *sbi);
3628 void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi,
3629 struct cp_control *cpc);
3630 void f2fs_dirty_to_prefree(struct f2fs_sb_info *sbi);
3631 block_t f2fs_get_unusable_blocks(struct f2fs_sb_info *sbi);
3632 int f2fs_disable_cp_again(struct f2fs_sb_info *sbi, block_t unusable);
3633 void f2fs_release_discard_addrs(struct f2fs_sb_info *sbi);
3634 int f2fs_npages_for_summary_flush(struct f2fs_sb_info *sbi, bool for_ra);
3635 bool f2fs_segment_has_free_slot(struct f2fs_sb_info *sbi, int segno);
3636 void f2fs_init_inmem_curseg(struct f2fs_sb_info *sbi);
3637 void f2fs_save_inmem_curseg(struct f2fs_sb_info *sbi);
3638 void f2fs_restore_inmem_curseg(struct f2fs_sb_info *sbi);
3639 void f2fs_get_new_segment(struct f2fs_sb_info *sbi,
3640 unsigned int *newseg, bool new_sec, int dir);
3641 void f2fs_allocate_segment_for_resize(struct f2fs_sb_info *sbi, int type,
3642 unsigned int start, unsigned int end);
3643 void f2fs_allocate_new_section(struct f2fs_sb_info *sbi, int type, bool force);
3644 void f2fs_allocate_new_segments(struct f2fs_sb_info *sbi);
3645 int f2fs_trim_fs(struct f2fs_sb_info *sbi, struct fstrim_range *range);
3646 bool f2fs_exist_trim_candidates(struct f2fs_sb_info *sbi,
3647 struct cp_control *cpc);
3648 struct page *f2fs_get_sum_page(struct f2fs_sb_info *sbi, unsigned int segno);
3649 void f2fs_update_meta_page(struct f2fs_sb_info *sbi, void *src,
3651 void f2fs_do_write_meta_page(struct f2fs_sb_info *sbi, struct page *page,
3652 enum iostat_type io_type);
3653 void f2fs_do_write_node_page(unsigned int nid, struct f2fs_io_info *fio);
3654 void f2fs_outplace_write_data(struct dnode_of_data *dn,
3655 struct f2fs_io_info *fio);
3656 int f2fs_inplace_write_data(struct f2fs_io_info *fio);
3657 void f2fs_do_replace_block(struct f2fs_sb_info *sbi, struct f2fs_summary *sum,
3658 block_t old_blkaddr, block_t new_blkaddr,
3659 bool recover_curseg, bool recover_newaddr,
3661 void f2fs_replace_block(struct f2fs_sb_info *sbi, struct dnode_of_data *dn,
3662 block_t old_addr, block_t new_addr,
3663 unsigned char version, bool recover_curseg,
3664 bool recover_newaddr);
3665 void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page,
3666 block_t old_blkaddr, block_t *new_blkaddr,
3667 struct f2fs_summary *sum, int type,
3668 struct f2fs_io_info *fio);
3669 void f2fs_update_device_state(struct f2fs_sb_info *sbi, nid_t ino,
3670 block_t blkaddr, unsigned int blkcnt);
3671 void f2fs_wait_on_page_writeback(struct page *page,
3672 enum page_type type, bool ordered, bool locked);
3673 void f2fs_wait_on_block_writeback(struct inode *inode, block_t blkaddr);
3674 void f2fs_wait_on_block_writeback_range(struct inode *inode, block_t blkaddr,
3676 void f2fs_write_data_summaries(struct f2fs_sb_info *sbi, block_t start_blk);
3677 void f2fs_write_node_summaries(struct f2fs_sb_info *sbi, block_t start_blk);
3678 int f2fs_lookup_journal_in_cursum(struct f2fs_journal *journal, int type,
3679 unsigned int val, int alloc);
3680 void f2fs_flush_sit_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc);
3681 int f2fs_fix_curseg_write_pointer(struct f2fs_sb_info *sbi);
3682 int f2fs_check_write_pointer(struct f2fs_sb_info *sbi);
3683 int f2fs_build_segment_manager(struct f2fs_sb_info *sbi);
3684 void f2fs_destroy_segment_manager(struct f2fs_sb_info *sbi);
3685 int __init f2fs_create_segment_manager_caches(void);
3686 void f2fs_destroy_segment_manager_caches(void);
3687 int f2fs_rw_hint_to_seg_type(enum rw_hint hint);
3688 unsigned int f2fs_usable_segs_in_sec(struct f2fs_sb_info *sbi,
3689 unsigned int segno);
3690 unsigned int f2fs_usable_blks_in_seg(struct f2fs_sb_info *sbi,
3691 unsigned int segno);
3693 #define DEF_FRAGMENT_SIZE 4
3694 #define MIN_FRAGMENT_SIZE 1
3695 #define MAX_FRAGMENT_SIZE 512
3697 static inline bool f2fs_need_rand_seg(struct f2fs_sb_info *sbi)
3699 return F2FS_OPTION(sbi).fs_mode == FS_MODE_FRAGMENT_SEG ||
3700 F2FS_OPTION(sbi).fs_mode == FS_MODE_FRAGMENT_BLK;
3706 void f2fs_stop_checkpoint(struct f2fs_sb_info *sbi, bool end_io,
3707 unsigned char reason);
3708 void f2fs_flush_ckpt_thread(struct f2fs_sb_info *sbi);
3709 struct page *f2fs_grab_meta_page(struct f2fs_sb_info *sbi, pgoff_t index);
3710 struct page *f2fs_get_meta_page(struct f2fs_sb_info *sbi, pgoff_t index);
3711 struct page *f2fs_get_meta_page_retry(struct f2fs_sb_info *sbi, pgoff_t index);
3712 struct page *f2fs_get_tmp_page(struct f2fs_sb_info *sbi, pgoff_t index);
3713 bool f2fs_is_valid_blkaddr(struct f2fs_sb_info *sbi,
3714 block_t blkaddr, int type);
3715 int f2fs_ra_meta_pages(struct f2fs_sb_info *sbi, block_t start, int nrpages,
3716 int type, bool sync);
3717 void f2fs_ra_meta_pages_cond(struct f2fs_sb_info *sbi, pgoff_t index,
3718 unsigned int ra_blocks);
3719 long f2fs_sync_meta_pages(struct f2fs_sb_info *sbi, enum page_type type,
3720 long nr_to_write, enum iostat_type io_type);
3721 void f2fs_add_ino_entry(struct f2fs_sb_info *sbi, nid_t ino, int type);
3722 void f2fs_remove_ino_entry(struct f2fs_sb_info *sbi, nid_t ino, int type);
3723 void f2fs_release_ino_entry(struct f2fs_sb_info *sbi, bool all);
3724 bool f2fs_exist_written_data(struct f2fs_sb_info *sbi, nid_t ino, int mode);
3725 void f2fs_set_dirty_device(struct f2fs_sb_info *sbi, nid_t ino,
3726 unsigned int devidx, int type);
3727 bool f2fs_is_dirty_device(struct f2fs_sb_info *sbi, nid_t ino,
3728 unsigned int devidx, int type);
3729 int f2fs_acquire_orphan_inode(struct f2fs_sb_info *sbi);
3730 void f2fs_release_orphan_inode(struct f2fs_sb_info *sbi);
3731 void f2fs_add_orphan_inode(struct inode *inode);
3732 void f2fs_remove_orphan_inode(struct f2fs_sb_info *sbi, nid_t ino);
3733 int f2fs_recover_orphan_inodes(struct f2fs_sb_info *sbi);
3734 int f2fs_get_valid_checkpoint(struct f2fs_sb_info *sbi);
3735 void f2fs_update_dirty_folio(struct inode *inode, struct folio *folio);
3736 void f2fs_remove_dirty_inode(struct inode *inode);
3737 int f2fs_sync_dirty_inodes(struct f2fs_sb_info *sbi, enum inode_type type,
3739 void f2fs_wait_on_all_pages(struct f2fs_sb_info *sbi, int type);
3740 u64 f2fs_get_sectors_written(struct f2fs_sb_info *sbi);
3741 int f2fs_write_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc);
3742 void f2fs_init_ino_entry_info(struct f2fs_sb_info *sbi);
3743 int __init f2fs_create_checkpoint_caches(void);
3744 void f2fs_destroy_checkpoint_caches(void);
3745 int f2fs_issue_checkpoint(struct f2fs_sb_info *sbi);
3746 int f2fs_start_ckpt_thread(struct f2fs_sb_info *sbi);
3747 void f2fs_stop_ckpt_thread(struct f2fs_sb_info *sbi);
3748 void f2fs_init_ckpt_req_control(struct f2fs_sb_info *sbi);
3753 int __init f2fs_init_bioset(void);
3754 void f2fs_destroy_bioset(void);
3755 int f2fs_init_bio_entry_cache(void);
3756 void f2fs_destroy_bio_entry_cache(void);
3757 void f2fs_submit_read_bio(struct f2fs_sb_info *sbi, struct bio *bio,
3758 enum page_type type);
3759 int f2fs_init_write_merge_io(struct f2fs_sb_info *sbi);
3760 void f2fs_submit_merged_write(struct f2fs_sb_info *sbi, enum page_type type);
3761 void f2fs_submit_merged_write_cond(struct f2fs_sb_info *sbi,
3762 struct inode *inode, struct page *page,
3763 nid_t ino, enum page_type type);
3764 void f2fs_submit_merged_ipu_write(struct f2fs_sb_info *sbi,
3765 struct bio **bio, struct page *page);
3766 void f2fs_flush_merged_writes(struct f2fs_sb_info *sbi);
3767 int f2fs_submit_page_bio(struct f2fs_io_info *fio);
3768 int f2fs_merge_page_bio(struct f2fs_io_info *fio);
3769 void f2fs_submit_page_write(struct f2fs_io_info *fio);
3770 struct block_device *f2fs_target_device(struct f2fs_sb_info *sbi,
3771 block_t blk_addr, sector_t *sector);
3772 int f2fs_target_device_index(struct f2fs_sb_info *sbi, block_t blkaddr);
3773 void f2fs_set_data_blkaddr(struct dnode_of_data *dn);
3774 void f2fs_update_data_blkaddr(struct dnode_of_data *dn, block_t blkaddr);
3775 int f2fs_reserve_new_blocks(struct dnode_of_data *dn, blkcnt_t count);
3776 int f2fs_reserve_new_block(struct dnode_of_data *dn);
3777 int f2fs_get_block_locked(struct dnode_of_data *dn, pgoff_t index);
3778 int f2fs_reserve_block(struct dnode_of_data *dn, pgoff_t index);
3779 struct page *f2fs_get_read_data_page(struct inode *inode, pgoff_t index,
3780 blk_opf_t op_flags, bool for_write, pgoff_t *next_pgofs);
3781 struct page *f2fs_find_data_page(struct inode *inode, pgoff_t index,
3782 pgoff_t *next_pgofs);
3783 struct page *f2fs_get_lock_data_page(struct inode *inode, pgoff_t index,
3785 struct page *f2fs_get_new_data_page(struct inode *inode,
3786 struct page *ipage, pgoff_t index, bool new_i_size);
3787 int f2fs_do_write_data_page(struct f2fs_io_info *fio);
3788 int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map, int flag);
3789 int f2fs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
3790 u64 start, u64 len);
3791 int f2fs_encrypt_one_page(struct f2fs_io_info *fio);
3792 bool f2fs_should_update_inplace(struct inode *inode, struct f2fs_io_info *fio);
3793 bool f2fs_should_update_outplace(struct inode *inode, struct f2fs_io_info *fio);
3794 int f2fs_write_single_data_page(struct page *page, int *submitted,
3795 struct bio **bio, sector_t *last_block,
3796 struct writeback_control *wbc,
3797 enum iostat_type io_type,
3798 int compr_blocks, bool allow_balance);
3799 void f2fs_write_failed(struct inode *inode, loff_t to);
3800 void f2fs_invalidate_folio(struct folio *folio, size_t offset, size_t length);
3801 bool f2fs_release_folio(struct folio *folio, gfp_t wait);
3802 bool f2fs_overwrite_io(struct inode *inode, loff_t pos, size_t len);
3803 void f2fs_clear_page_cache_dirty_tag(struct page *page);
3804 int f2fs_init_post_read_processing(void);
3805 void f2fs_destroy_post_read_processing(void);
3806 int f2fs_init_post_read_wq(struct f2fs_sb_info *sbi);
3807 void f2fs_destroy_post_read_wq(struct f2fs_sb_info *sbi);
3808 extern const struct iomap_ops f2fs_iomap_ops;
3813 int f2fs_start_gc_thread(struct f2fs_sb_info *sbi);
3814 void f2fs_stop_gc_thread(struct f2fs_sb_info *sbi);
3815 block_t f2fs_start_bidx_of_node(unsigned int node_ofs, struct inode *inode);
3816 int f2fs_gc(struct f2fs_sb_info *sbi, struct f2fs_gc_control *gc_control);
3817 void f2fs_build_gc_manager(struct f2fs_sb_info *sbi);
3818 int f2fs_resize_fs(struct f2fs_sb_info *sbi, __u64 block_count);
3819 int __init f2fs_create_garbage_collection_cache(void);
3820 void f2fs_destroy_garbage_collection_cache(void);
3821 /* victim selection function for cleaning and SSR */
3822 int f2fs_get_victim(struct f2fs_sb_info *sbi, unsigned int *result,
3823 int gc_type, int type, char alloc_mode,
3824 unsigned long long age);
3829 int f2fs_recover_fsync_data(struct f2fs_sb_info *sbi, bool check_only);
3830 bool f2fs_space_for_roll_forward(struct f2fs_sb_info *sbi);
3831 int __init f2fs_create_recovery_cache(void);
3832 void f2fs_destroy_recovery_cache(void);
3837 #ifdef CONFIG_F2FS_STAT_FS
3838 struct f2fs_stat_info {
3839 struct list_head stat_list;
3840 struct f2fs_sb_info *sbi;
3841 int all_area_segs, sit_area_segs, nat_area_segs, ssa_area_segs;
3842 int main_area_segs, main_area_sections, main_area_zones;
3843 unsigned long long hit_cached[NR_EXTENT_CACHES];
3844 unsigned long long hit_rbtree[NR_EXTENT_CACHES];
3845 unsigned long long total_ext[NR_EXTENT_CACHES];
3846 unsigned long long hit_total[NR_EXTENT_CACHES];
3847 int ext_tree[NR_EXTENT_CACHES];
3848 int zombie_tree[NR_EXTENT_CACHES];
3849 int ext_node[NR_EXTENT_CACHES];
3850 /* to count memory footprint */
3851 unsigned long long ext_mem[NR_EXTENT_CACHES];
3852 /* for read extent cache */
3853 unsigned long long hit_largest;
3854 /* for block age extent cache */
3855 unsigned long long allocated_data_blocks;
3856 int ndirty_node, ndirty_dent, ndirty_meta, ndirty_imeta;
3857 int ndirty_data, ndirty_qdata;
3858 unsigned int ndirty_dirs, ndirty_files, nquota_files, ndirty_all;
3859 int nats, dirty_nats, sits, dirty_sits;
3860 int free_nids, avail_nids, alloc_nids;
3861 int total_count, utilization;
3862 int bg_gc, nr_wb_cp_data, nr_wb_data;
3863 int nr_rd_data, nr_rd_node, nr_rd_meta;
3864 int nr_dio_read, nr_dio_write;
3865 unsigned int io_skip_bggc, other_skip_bggc;
3866 int nr_flushing, nr_flushed, flush_list_empty;
3867 int nr_discarding, nr_discarded;
3869 unsigned int undiscard_blks;
3870 int nr_issued_ckpt, nr_total_ckpt, nr_queued_ckpt;
3871 unsigned int cur_ckpt_time, peak_ckpt_time;
3872 int inline_xattr, inline_inode, inline_dir, append, update, orphans;
3873 int compr_inode, swapfile_inode;
3874 unsigned long long compr_blocks;
3875 int aw_cnt, max_aw_cnt;
3876 unsigned int valid_count, valid_node_count, valid_inode_count, discard_blks;
3877 unsigned int bimodal, avg_vblocks;
3878 int util_free, util_valid, util_invalid;
3879 int rsvd_segs, overp_segs;
3880 int dirty_count, node_pages, meta_pages, compress_pages;
3881 int compress_page_hit;
3882 int prefree_count, call_count, cp_count, bg_cp_count;
3883 int tot_segs, node_segs, data_segs, free_segs, free_secs;
3884 int bg_node_segs, bg_data_segs;
3885 int tot_blks, data_blks, node_blks;
3886 int bg_data_blks, bg_node_blks;
3887 int curseg[NR_CURSEG_TYPE];
3888 int cursec[NR_CURSEG_TYPE];
3889 int curzone[NR_CURSEG_TYPE];
3890 unsigned int dirty_seg[NR_CURSEG_TYPE];
3891 unsigned int full_seg[NR_CURSEG_TYPE];
3892 unsigned int valid_blks[NR_CURSEG_TYPE];
3894 unsigned int meta_count[META_MAX];
3895 unsigned int segment_count[2];
3896 unsigned int block_count[2];
3897 unsigned int inplace_count;
3898 unsigned long long base_mem, cache_mem, page_mem;
3901 static inline struct f2fs_stat_info *F2FS_STAT(struct f2fs_sb_info *sbi)
3903 return (struct f2fs_stat_info *)sbi->stat_info;
3906 #define stat_inc_cp_count(si) ((si)->cp_count++)
3907 #define stat_inc_bg_cp_count(si) ((si)->bg_cp_count++)
3908 #define stat_inc_call_count(si) ((si)->call_count++)
3909 #define stat_inc_bggc_count(si) ((si)->bg_gc++)
3910 #define stat_io_skip_bggc_count(sbi) ((sbi)->io_skip_bggc++)
3911 #define stat_other_skip_bggc_count(sbi) ((sbi)->other_skip_bggc++)
3912 #define stat_inc_dirty_inode(sbi, type) ((sbi)->ndirty_inode[type]++)
3913 #define stat_dec_dirty_inode(sbi, type) ((sbi)->ndirty_inode[type]--)
3914 #define stat_inc_total_hit(sbi, type) (atomic64_inc(&(sbi)->total_hit_ext[type]))
3915 #define stat_inc_rbtree_node_hit(sbi, type) (atomic64_inc(&(sbi)->read_hit_rbtree[type]))
3916 #define stat_inc_largest_node_hit(sbi) (atomic64_inc(&(sbi)->read_hit_largest))
3917 #define stat_inc_cached_node_hit(sbi, type) (atomic64_inc(&(sbi)->read_hit_cached[type]))
3918 #define stat_inc_inline_xattr(inode) \
3920 if (f2fs_has_inline_xattr(inode)) \
3921 (atomic_inc(&F2FS_I_SB(inode)->inline_xattr)); \
3923 #define stat_dec_inline_xattr(inode) \
3925 if (f2fs_has_inline_xattr(inode)) \
3926 (atomic_dec(&F2FS_I_SB(inode)->inline_xattr)); \
3928 #define stat_inc_inline_inode(inode) \
3930 if (f2fs_has_inline_data(inode)) \
3931 (atomic_inc(&F2FS_I_SB(inode)->inline_inode)); \
3933 #define stat_dec_inline_inode(inode) \
3935 if (f2fs_has_inline_data(inode)) \
3936 (atomic_dec(&F2FS_I_SB(inode)->inline_inode)); \
3938 #define stat_inc_inline_dir(inode) \
3940 if (f2fs_has_inline_dentry(inode)) \
3941 (atomic_inc(&F2FS_I_SB(inode)->inline_dir)); \
3943 #define stat_dec_inline_dir(inode) \
3945 if (f2fs_has_inline_dentry(inode)) \
3946 (atomic_dec(&F2FS_I_SB(inode)->inline_dir)); \
3948 #define stat_inc_compr_inode(inode) \
3950 if (f2fs_compressed_file(inode)) \
3951 (atomic_inc(&F2FS_I_SB(inode)->compr_inode)); \
3953 #define stat_dec_compr_inode(inode) \
3955 if (f2fs_compressed_file(inode)) \
3956 (atomic_dec(&F2FS_I_SB(inode)->compr_inode)); \
3958 #define stat_add_compr_blocks(inode, blocks) \
3959 (atomic64_add(blocks, &F2FS_I_SB(inode)->compr_blocks))
3960 #define stat_sub_compr_blocks(inode, blocks) \
3961 (atomic64_sub(blocks, &F2FS_I_SB(inode)->compr_blocks))
3962 #define stat_inc_swapfile_inode(inode) \
3963 (atomic_inc(&F2FS_I_SB(inode)->swapfile_inode))
3964 #define stat_dec_swapfile_inode(inode) \
3965 (atomic_dec(&F2FS_I_SB(inode)->swapfile_inode))
3966 #define stat_inc_atomic_inode(inode) \
3967 (atomic_inc(&F2FS_I_SB(inode)->atomic_files))
3968 #define stat_dec_atomic_inode(inode) \
3969 (atomic_dec(&F2FS_I_SB(inode)->atomic_files))
3970 #define stat_inc_meta_count(sbi, blkaddr) \
3972 if (blkaddr < SIT_I(sbi)->sit_base_addr) \
3973 atomic_inc(&(sbi)->meta_count[META_CP]); \
3974 else if (blkaddr < NM_I(sbi)->nat_blkaddr) \
3975 atomic_inc(&(sbi)->meta_count[META_SIT]); \
3976 else if (blkaddr < SM_I(sbi)->ssa_blkaddr) \
3977 atomic_inc(&(sbi)->meta_count[META_NAT]); \
3978 else if (blkaddr < SM_I(sbi)->main_blkaddr) \
3979 atomic_inc(&(sbi)->meta_count[META_SSA]); \
3981 #define stat_inc_seg_type(sbi, curseg) \
3982 ((sbi)->segment_count[(curseg)->alloc_type]++)
3983 #define stat_inc_block_count(sbi, curseg) \
3984 ((sbi)->block_count[(curseg)->alloc_type]++)
3985 #define stat_inc_inplace_blocks(sbi) \
3986 (atomic_inc(&(sbi)->inplace_count))
3987 #define stat_update_max_atomic_write(inode) \
3989 int cur = atomic_read(&F2FS_I_SB(inode)->atomic_files); \
3990 int max = atomic_read(&F2FS_I_SB(inode)->max_aw_cnt); \
3992 atomic_set(&F2FS_I_SB(inode)->max_aw_cnt, cur); \
3994 #define stat_inc_seg_count(sbi, type, gc_type) \
3996 struct f2fs_stat_info *si = F2FS_STAT(sbi); \
3998 if ((type) == SUM_TYPE_DATA) { \
4000 si->bg_data_segs += (gc_type == BG_GC) ? 1 : 0; \
4003 si->bg_node_segs += (gc_type == BG_GC) ? 1 : 0; \
4007 #define stat_inc_tot_blk_count(si, blks) \
4008 ((si)->tot_blks += (blks))
4010 #define stat_inc_data_blk_count(sbi, blks, gc_type) \
4012 struct f2fs_stat_info *si = F2FS_STAT(sbi); \
4013 stat_inc_tot_blk_count(si, blks); \
4014 si->data_blks += (blks); \
4015 si->bg_data_blks += ((gc_type) == BG_GC) ? (blks) : 0; \
4018 #define stat_inc_node_blk_count(sbi, blks, gc_type) \
4020 struct f2fs_stat_info *si = F2FS_STAT(sbi); \
4021 stat_inc_tot_blk_count(si, blks); \
4022 si->node_blks += (blks); \
4023 si->bg_node_blks += ((gc_type) == BG_GC) ? (blks) : 0; \
4026 int f2fs_build_stats(struct f2fs_sb_info *sbi);
4027 void f2fs_destroy_stats(struct f2fs_sb_info *sbi);
4028 void __init f2fs_create_root_stats(void);
4029 void f2fs_destroy_root_stats(void);
4030 void f2fs_update_sit_info(struct f2fs_sb_info *sbi);
4032 #define stat_inc_cp_count(si) do { } while (0)
4033 #define stat_inc_bg_cp_count(si) do { } while (0)
4034 #define stat_inc_call_count(si) do { } while (0)
4035 #define stat_inc_bggc_count(si) do { } while (0)
4036 #define stat_io_skip_bggc_count(sbi) do { } while (0)
4037 #define stat_other_skip_bggc_count(sbi) do { } while (0)
4038 #define stat_inc_dirty_inode(sbi, type) do { } while (0)
4039 #define stat_dec_dirty_inode(sbi, type) do { } while (0)
4040 #define stat_inc_total_hit(sbi, type) do { } while (0)
4041 #define stat_inc_rbtree_node_hit(sbi, type) do { } while (0)
4042 #define stat_inc_largest_node_hit(sbi) do { } while (0)
4043 #define stat_inc_cached_node_hit(sbi, type) do { } while (0)
4044 #define stat_inc_inline_xattr(inode) do { } while (0)
4045 #define stat_dec_inline_xattr(inode) do { } while (0)
4046 #define stat_inc_inline_inode(inode) do { } while (0)
4047 #define stat_dec_inline_inode(inode) do { } while (0)
4048 #define stat_inc_inline_dir(inode) do { } while (0)
4049 #define stat_dec_inline_dir(inode) do { } while (0)
4050 #define stat_inc_compr_inode(inode) do { } while (0)
4051 #define stat_dec_compr_inode(inode) do { } while (0)
4052 #define stat_add_compr_blocks(inode, blocks) do { } while (0)
4053 #define stat_sub_compr_blocks(inode, blocks) do { } while (0)
4054 #define stat_inc_swapfile_inode(inode) do { } while (0)
4055 #define stat_dec_swapfile_inode(inode) do { } while (0)
4056 #define stat_inc_atomic_inode(inode) do { } while (0)
4057 #define stat_dec_atomic_inode(inode) do { } while (0)
4058 #define stat_update_max_atomic_write(inode) do { } while (0)
4059 #define stat_inc_meta_count(sbi, blkaddr) do { } while (0)
4060 #define stat_inc_seg_type(sbi, curseg) do { } while (0)
4061 #define stat_inc_block_count(sbi, curseg) do { } while (0)
4062 #define stat_inc_inplace_blocks(sbi) do { } while (0)
4063 #define stat_inc_seg_count(sbi, type, gc_type) do { } while (0)
4064 #define stat_inc_tot_blk_count(si, blks) do { } while (0)
4065 #define stat_inc_data_blk_count(sbi, blks, gc_type) do { } while (0)
4066 #define stat_inc_node_blk_count(sbi, blks, gc_type) do { } while (0)
4068 static inline int f2fs_build_stats(struct f2fs_sb_info *sbi) { return 0; }
4069 static inline void f2fs_destroy_stats(struct f2fs_sb_info *sbi) { }
4070 static inline void __init f2fs_create_root_stats(void) { }
4071 static inline void f2fs_destroy_root_stats(void) { }
4072 static inline void f2fs_update_sit_info(struct f2fs_sb_info *sbi) {}
4075 extern const struct file_operations f2fs_dir_operations;
4076 extern const struct file_operations f2fs_file_operations;
4077 extern const struct inode_operations f2fs_file_inode_operations;
4078 extern const struct address_space_operations f2fs_dblock_aops;
4079 extern const struct address_space_operations f2fs_node_aops;
4080 extern const struct address_space_operations f2fs_meta_aops;
4081 extern const struct inode_operations f2fs_dir_inode_operations;
4082 extern const struct inode_operations f2fs_symlink_inode_operations;
4083 extern const struct inode_operations f2fs_encrypted_symlink_inode_operations;
4084 extern const struct inode_operations f2fs_special_inode_operations;
4085 extern struct kmem_cache *f2fs_inode_entry_slab;
4090 bool f2fs_may_inline_data(struct inode *inode);
4091 bool f2fs_sanity_check_inline_data(struct inode *inode);
4092 bool f2fs_may_inline_dentry(struct inode *inode);
4093 void f2fs_do_read_inline_data(struct page *page, struct page *ipage);
4094 void f2fs_truncate_inline_inode(struct inode *inode,
4095 struct page *ipage, u64 from);
4096 int f2fs_read_inline_data(struct inode *inode, struct page *page);
4097 int f2fs_convert_inline_page(struct dnode_of_data *dn, struct page *page);
4098 int f2fs_convert_inline_inode(struct inode *inode);
4099 int f2fs_try_convert_inline_dir(struct inode *dir, struct dentry *dentry);
4100 int f2fs_write_inline_data(struct inode *inode, struct page *page);
4101 int f2fs_recover_inline_data(struct inode *inode, struct page *npage);
4102 struct f2fs_dir_entry *f2fs_find_in_inline_dir(struct inode *dir,
4103 const struct f2fs_filename *fname,
4104 struct page **res_page);
4105 int f2fs_make_empty_inline_dir(struct inode *inode, struct inode *parent,
4106 struct page *ipage);
4107 int f2fs_add_inline_entry(struct inode *dir, const struct f2fs_filename *fname,
4108 struct inode *inode, nid_t ino, umode_t mode);
4109 void f2fs_delete_inline_entry(struct f2fs_dir_entry *dentry,
4110 struct page *page, struct inode *dir,
4111 struct inode *inode);
4112 bool f2fs_empty_inline_dir(struct inode *dir);
4113 int f2fs_read_inline_dir(struct file *file, struct dir_context *ctx,
4114 struct fscrypt_str *fstr);
4115 int f2fs_inline_data_fiemap(struct inode *inode,
4116 struct fiemap_extent_info *fieinfo,
4117 __u64 start, __u64 len);
4122 unsigned long f2fs_shrink_count(struct shrinker *shrink,
4123 struct shrink_control *sc);
4124 unsigned long f2fs_shrink_scan(struct shrinker *shrink,
4125 struct shrink_control *sc);
4126 void f2fs_join_shrinker(struct f2fs_sb_info *sbi);
4127 void f2fs_leave_shrinker(struct f2fs_sb_info *sbi);
4132 bool sanity_check_extent_cache(struct inode *inode);
4133 void f2fs_init_extent_tree(struct inode *inode);
4134 void f2fs_drop_extent_tree(struct inode *inode);
4135 void f2fs_destroy_extent_node(struct inode *inode);
4136 void f2fs_destroy_extent_tree(struct inode *inode);
4137 void f2fs_init_extent_cache_info(struct f2fs_sb_info *sbi);
4138 int __init f2fs_create_extent_cache(void);
4139 void f2fs_destroy_extent_cache(void);
4141 /* read extent cache ops */
4142 void f2fs_init_read_extent_tree(struct inode *inode, struct page *ipage);
4143 bool f2fs_lookup_read_extent_cache(struct inode *inode, pgoff_t pgofs,
4144 struct extent_info *ei);
4145 bool f2fs_lookup_read_extent_cache_block(struct inode *inode, pgoff_t index,
4147 void f2fs_update_read_extent_cache(struct dnode_of_data *dn);
4148 void f2fs_update_read_extent_cache_range(struct dnode_of_data *dn,
4149 pgoff_t fofs, block_t blkaddr, unsigned int len);
4150 unsigned int f2fs_shrink_read_extent_tree(struct f2fs_sb_info *sbi,
4153 /* block age extent cache ops */
4154 void f2fs_init_age_extent_tree(struct inode *inode);
4155 bool f2fs_lookup_age_extent_cache(struct inode *inode, pgoff_t pgofs,
4156 struct extent_info *ei);
4157 void f2fs_update_age_extent_cache(struct dnode_of_data *dn);
4158 void f2fs_update_age_extent_cache_range(struct dnode_of_data *dn,
4159 pgoff_t fofs, unsigned int len);
4160 unsigned int f2fs_shrink_age_extent_tree(struct f2fs_sb_info *sbi,
4166 #define MIN_RA_MUL 2
4167 #define MAX_RA_MUL 256
4169 int __init f2fs_init_sysfs(void);
4170 void f2fs_exit_sysfs(void);
4171 int f2fs_register_sysfs(struct f2fs_sb_info *sbi);
4172 void f2fs_unregister_sysfs(struct f2fs_sb_info *sbi);
4175 extern const struct fsverity_operations f2fs_verityops;
4180 static inline bool f2fs_encrypted_file(struct inode *inode)
4182 return IS_ENCRYPTED(inode) && S_ISREG(inode->i_mode);
4185 static inline void f2fs_set_encrypted_inode(struct inode *inode)
4187 #ifdef CONFIG_FS_ENCRYPTION
4188 file_set_encrypt(inode);
4189 f2fs_set_inode_flags(inode);
4194 * Returns true if the reads of the inode's data need to undergo some
4195 * postprocessing step, like decryption or authenticity verification.
4197 static inline bool f2fs_post_read_required(struct inode *inode)
4199 return f2fs_encrypted_file(inode) || fsverity_active(inode) ||
4200 f2fs_compressed_file(inode);
4206 #ifdef CONFIG_F2FS_FS_COMPRESSION
4207 bool f2fs_is_compressed_page(struct page *page);
4208 struct page *f2fs_compress_control_page(struct page *page);
4209 int f2fs_prepare_compress_overwrite(struct inode *inode,
4210 struct page **pagep, pgoff_t index, void **fsdata);
4211 bool f2fs_compress_write_end(struct inode *inode, void *fsdata,
4212 pgoff_t index, unsigned copied);
4213 int f2fs_truncate_partial_cluster(struct inode *inode, u64 from, bool lock);
4214 void f2fs_compress_write_end_io(struct bio *bio, struct page *page);
4215 bool f2fs_is_compress_backend_ready(struct inode *inode);
4216 int __init f2fs_init_compress_mempool(void);
4217 void f2fs_destroy_compress_mempool(void);
4218 void f2fs_decompress_cluster(struct decompress_io_ctx *dic, bool in_task);
4219 void f2fs_end_read_compressed_page(struct page *page, bool failed,
4220 block_t blkaddr, bool in_task);
4221 bool f2fs_cluster_is_empty(struct compress_ctx *cc);
4222 bool f2fs_cluster_can_merge_page(struct compress_ctx *cc, pgoff_t index);
4223 bool f2fs_all_cluster_page_ready(struct compress_ctx *cc, struct page **pages,
4224 int index, int nr_pages, bool uptodate);
4225 bool f2fs_sanity_check_cluster(struct dnode_of_data *dn);
4226 void f2fs_compress_ctx_add_page(struct compress_ctx *cc, struct page *page);
4227 int f2fs_write_multi_pages(struct compress_ctx *cc,
4229 struct writeback_control *wbc,
4230 enum iostat_type io_type);
4231 int f2fs_is_compressed_cluster(struct inode *inode, pgoff_t index);
4232 void f2fs_update_read_extent_tree_range_compressed(struct inode *inode,
4233 pgoff_t fofs, block_t blkaddr,
4234 unsigned int llen, unsigned int c_len);
4235 int f2fs_read_multi_pages(struct compress_ctx *cc, struct bio **bio_ret,
4236 unsigned nr_pages, sector_t *last_block_in_bio,
4237 bool is_readahead, bool for_write);
4238 struct decompress_io_ctx *f2fs_alloc_dic(struct compress_ctx *cc);
4239 void f2fs_decompress_end_io(struct decompress_io_ctx *dic, bool failed,
4241 void f2fs_put_page_dic(struct page *page, bool in_task);
4242 unsigned int f2fs_cluster_blocks_are_contiguous(struct dnode_of_data *dn);
4243 int f2fs_init_compress_ctx(struct compress_ctx *cc);
4244 void f2fs_destroy_compress_ctx(struct compress_ctx *cc, bool reuse);
4245 void f2fs_init_compress_info(struct f2fs_sb_info *sbi);
4246 int f2fs_init_compress_inode(struct f2fs_sb_info *sbi);
4247 void f2fs_destroy_compress_inode(struct f2fs_sb_info *sbi);
4248 int f2fs_init_page_array_cache(struct f2fs_sb_info *sbi);
4249 void f2fs_destroy_page_array_cache(struct f2fs_sb_info *sbi);
4250 int __init f2fs_init_compress_cache(void);
4251 void f2fs_destroy_compress_cache(void);
4252 struct address_space *COMPRESS_MAPPING(struct f2fs_sb_info *sbi);
4253 void f2fs_invalidate_compress_page(struct f2fs_sb_info *sbi, block_t blkaddr);
4254 void f2fs_cache_compressed_page(struct f2fs_sb_info *sbi, struct page *page,
4255 nid_t ino, block_t blkaddr);
4256 bool f2fs_load_compressed_page(struct f2fs_sb_info *sbi, struct page *page,
4258 void f2fs_invalidate_compress_pages(struct f2fs_sb_info *sbi, nid_t ino);
4259 #define inc_compr_inode_stat(inode) \
4261 struct f2fs_sb_info *sbi = F2FS_I_SB(inode); \
4262 sbi->compr_new_inode++; \
4264 #define add_compr_block_stat(inode, blocks) \
4266 struct f2fs_sb_info *sbi = F2FS_I_SB(inode); \
4267 int diff = F2FS_I(inode)->i_cluster_size - blocks; \
4268 sbi->compr_written_block += blocks; \
4269 sbi->compr_saved_block += diff; \
4272 static inline bool f2fs_is_compressed_page(struct page *page) { return false; }
4273 static inline bool f2fs_is_compress_backend_ready(struct inode *inode)
4275 if (!f2fs_compressed_file(inode))
4277 /* not support compression */
4280 static inline struct page *f2fs_compress_control_page(struct page *page)
4283 return ERR_PTR(-EINVAL);
4285 static inline int __init f2fs_init_compress_mempool(void) { return 0; }
4286 static inline void f2fs_destroy_compress_mempool(void) { }
4287 static inline void f2fs_decompress_cluster(struct decompress_io_ctx *dic,
4289 static inline void f2fs_end_read_compressed_page(struct page *page,
4290 bool failed, block_t blkaddr, bool in_task)
4294 static inline void f2fs_put_page_dic(struct page *page, bool in_task)
4298 static inline unsigned int f2fs_cluster_blocks_are_contiguous(struct dnode_of_data *dn) { return 0; }
4299 static inline bool f2fs_sanity_check_cluster(struct dnode_of_data *dn) { return false; }
4300 static inline int f2fs_init_compress_inode(struct f2fs_sb_info *sbi) { return 0; }
4301 static inline void f2fs_destroy_compress_inode(struct f2fs_sb_info *sbi) { }
4302 static inline int f2fs_init_page_array_cache(struct f2fs_sb_info *sbi) { return 0; }
4303 static inline void f2fs_destroy_page_array_cache(struct f2fs_sb_info *sbi) { }
4304 static inline int __init f2fs_init_compress_cache(void) { return 0; }
4305 static inline void f2fs_destroy_compress_cache(void) { }
4306 static inline void f2fs_invalidate_compress_page(struct f2fs_sb_info *sbi,
4307 block_t blkaddr) { }
4308 static inline void f2fs_cache_compressed_page(struct f2fs_sb_info *sbi,
4309 struct page *page, nid_t ino, block_t blkaddr) { }
4310 static inline bool f2fs_load_compressed_page(struct f2fs_sb_info *sbi,
4311 struct page *page, block_t blkaddr) { return false; }
4312 static inline void f2fs_invalidate_compress_pages(struct f2fs_sb_info *sbi,
4314 #define inc_compr_inode_stat(inode) do { } while (0)
4315 static inline void f2fs_update_read_extent_tree_range_compressed(
4316 struct inode *inode,
4317 pgoff_t fofs, block_t blkaddr,
4318 unsigned int llen, unsigned int c_len) { }
4321 static inline int set_compress_context(struct inode *inode)
4323 #ifdef CONFIG_F2FS_FS_COMPRESSION
4324 struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
4326 F2FS_I(inode)->i_compress_algorithm =
4327 F2FS_OPTION(sbi).compress_algorithm;
4328 F2FS_I(inode)->i_log_cluster_size =
4329 F2FS_OPTION(sbi).compress_log_size;
4330 F2FS_I(inode)->i_compress_flag =
4331 F2FS_OPTION(sbi).compress_chksum ?
4332 BIT(COMPRESS_CHKSUM) : 0;
4333 F2FS_I(inode)->i_cluster_size =
4334 BIT(F2FS_I(inode)->i_log_cluster_size);
4335 if ((F2FS_I(inode)->i_compress_algorithm == COMPRESS_LZ4 ||
4336 F2FS_I(inode)->i_compress_algorithm == COMPRESS_ZSTD) &&
4337 F2FS_OPTION(sbi).compress_level)
4338 F2FS_I(inode)->i_compress_level =
4339 F2FS_OPTION(sbi).compress_level;
4340 F2FS_I(inode)->i_flags |= F2FS_COMPR_FL;
4341 set_inode_flag(inode, FI_COMPRESSED_FILE);
4342 stat_inc_compr_inode(inode);
4343 inc_compr_inode_stat(inode);
4344 f2fs_mark_inode_dirty_sync(inode, true);
4351 static inline bool f2fs_disable_compressed_file(struct inode *inode)
4353 struct f2fs_inode_info *fi = F2FS_I(inode);
4355 if (!f2fs_compressed_file(inode))
4357 if (S_ISREG(inode->i_mode) && F2FS_HAS_BLOCKS(inode))
4360 fi->i_flags &= ~F2FS_COMPR_FL;
4361 stat_dec_compr_inode(inode);
4362 clear_inode_flag(inode, FI_COMPRESSED_FILE);
4363 f2fs_mark_inode_dirty_sync(inode, true);
4367 #define F2FS_FEATURE_FUNCS(name, flagname) \
4368 static inline bool f2fs_sb_has_##name(struct f2fs_sb_info *sbi) \
4370 return F2FS_HAS_FEATURE(sbi, F2FS_FEATURE_##flagname); \
4373 F2FS_FEATURE_FUNCS(encrypt, ENCRYPT);
4374 F2FS_FEATURE_FUNCS(blkzoned, BLKZONED);
4375 F2FS_FEATURE_FUNCS(extra_attr, EXTRA_ATTR);
4376 F2FS_FEATURE_FUNCS(project_quota, PRJQUOTA);
4377 F2FS_FEATURE_FUNCS(inode_chksum, INODE_CHKSUM);
4378 F2FS_FEATURE_FUNCS(flexible_inline_xattr, FLEXIBLE_INLINE_XATTR);
4379 F2FS_FEATURE_FUNCS(quota_ino, QUOTA_INO);
4380 F2FS_FEATURE_FUNCS(inode_crtime, INODE_CRTIME);
4381 F2FS_FEATURE_FUNCS(lost_found, LOST_FOUND);
4382 F2FS_FEATURE_FUNCS(verity, VERITY);
4383 F2FS_FEATURE_FUNCS(sb_chksum, SB_CHKSUM);
4384 F2FS_FEATURE_FUNCS(casefold, CASEFOLD);
4385 F2FS_FEATURE_FUNCS(compression, COMPRESSION);
4386 F2FS_FEATURE_FUNCS(readonly, RO);
4388 #ifdef CONFIG_BLK_DEV_ZONED
4389 static inline bool f2fs_blkz_is_seq(struct f2fs_sb_info *sbi, int devi,
4392 unsigned int zno = blkaddr / sbi->blocks_per_blkz;
4394 return test_bit(zno, FDEV(devi).blkz_seq);
4398 static inline bool f2fs_hw_should_discard(struct f2fs_sb_info *sbi)
4400 return f2fs_sb_has_blkzoned(sbi);
4403 static inline bool f2fs_bdev_support_discard(struct block_device *bdev)
4405 return bdev_max_discard_sectors(bdev) || bdev_is_zoned(bdev);
4408 static inline bool f2fs_hw_support_discard(struct f2fs_sb_info *sbi)
4412 if (!f2fs_is_multi_device(sbi))
4413 return f2fs_bdev_support_discard(sbi->sb->s_bdev);
4415 for (i = 0; i < sbi->s_ndevs; i++)
4416 if (f2fs_bdev_support_discard(FDEV(i).bdev))
4421 static inline bool f2fs_realtime_discard_enable(struct f2fs_sb_info *sbi)
4423 return (test_opt(sbi, DISCARD) && f2fs_hw_support_discard(sbi)) ||
4424 f2fs_hw_should_discard(sbi);
4427 static inline bool f2fs_hw_is_readonly(struct f2fs_sb_info *sbi)
4431 if (!f2fs_is_multi_device(sbi))
4432 return bdev_read_only(sbi->sb->s_bdev);
4434 for (i = 0; i < sbi->s_ndevs; i++)
4435 if (bdev_read_only(FDEV(i).bdev))
4440 static inline bool f2fs_dev_is_readonly(struct f2fs_sb_info *sbi)
4442 return f2fs_sb_has_readonly(sbi) || f2fs_hw_is_readonly(sbi);
4445 static inline bool f2fs_lfs_mode(struct f2fs_sb_info *sbi)
4447 return F2FS_OPTION(sbi).fs_mode == FS_MODE_LFS;
4450 static inline bool f2fs_low_mem_mode(struct f2fs_sb_info *sbi)
4452 return F2FS_OPTION(sbi).memory_mode == MEMORY_MODE_LOW;
4455 static inline bool f2fs_may_compress(struct inode *inode)
4457 if (IS_SWAPFILE(inode) || f2fs_is_pinned_file(inode) ||
4458 f2fs_is_atomic_file(inode) || f2fs_has_inline_data(inode))
4460 return S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode);
4463 static inline void f2fs_i_compr_blocks_update(struct inode *inode,
4464 u64 blocks, bool add)
4466 struct f2fs_inode_info *fi = F2FS_I(inode);
4467 int diff = fi->i_cluster_size - blocks;
4469 /* don't update i_compr_blocks if saved blocks were released */
4470 if (!add && !atomic_read(&fi->i_compr_blocks))
4474 atomic_add(diff, &fi->i_compr_blocks);
4475 stat_add_compr_blocks(inode, diff);
4477 atomic_sub(diff, &fi->i_compr_blocks);
4478 stat_sub_compr_blocks(inode, diff);
4480 f2fs_mark_inode_dirty_sync(inode, true);
4483 static inline bool f2fs_allow_multi_device_dio(struct f2fs_sb_info *sbi,
4486 if (!f2fs_is_multi_device(sbi))
4488 if (flag != F2FS_GET_BLOCK_DIO)
4490 return sbi->aligned_blksize;
4493 static inline bool f2fs_need_verity(const struct inode *inode, pgoff_t idx)
4495 return fsverity_active(inode) &&
4496 idx < DIV_ROUND_UP(inode->i_size, PAGE_SIZE);
4499 #ifdef CONFIG_F2FS_FAULT_INJECTION
4500 extern void f2fs_build_fault_attr(struct f2fs_sb_info *sbi, unsigned int rate,
4503 #define f2fs_build_fault_attr(sbi, rate, type) do { } while (0)
4506 static inline bool is_journalled_quota(struct f2fs_sb_info *sbi)
4509 if (f2fs_sb_has_quota_ino(sbi))
4511 if (F2FS_OPTION(sbi).s_qf_names[USRQUOTA] ||
4512 F2FS_OPTION(sbi).s_qf_names[GRPQUOTA] ||
4513 F2FS_OPTION(sbi).s_qf_names[PRJQUOTA])
4519 static inline bool f2fs_block_unit_discard(struct f2fs_sb_info *sbi)
4521 return F2FS_OPTION(sbi).discard_unit == DISCARD_UNIT_BLOCK;
4524 static inline void f2fs_io_schedule_timeout(long timeout)
4526 set_current_state(TASK_UNINTERRUPTIBLE);
4527 io_schedule_timeout(timeout);
4530 static inline void f2fs_handle_page_eio(struct f2fs_sb_info *sbi, pgoff_t ofs,
4531 enum page_type type)
4533 if (unlikely(f2fs_cp_error(sbi)))
4536 if (ofs == sbi->page_eio_ofs[type]) {
4537 if (sbi->page_eio_cnt[type]++ == MAX_RETRY_PAGE_EIO)
4538 set_ckpt_flags(sbi, CP_ERROR_FLAG);
4540 sbi->page_eio_ofs[type] = ofs;
4541 sbi->page_eio_cnt[type] = 0;
4545 static inline bool f2fs_is_readonly(struct f2fs_sb_info *sbi)
4547 return f2fs_sb_has_readonly(sbi) || f2fs_readonly(sbi->sb);
4550 #define EFSBADCRC EBADMSG /* Bad CRC detected */
4551 #define EFSCORRUPTED EUCLEAN /* Filesystem is corrupted */
4553 #endif /* _LINUX_F2FS_H */