ext4: rename s_error_work to s_sb_upd_work
[linux-block.git] / fs / ext4 / ext4.h
CommitLineData
f5166768 1// SPDX-License-Identifier: GPL-2.0
ac27a0ec 2/*
3dcf5451 3 * ext4.h
ac27a0ec
DK
4 *
5 * Copyright (C) 1992, 1993, 1994, 1995
6 * Remy Card (card@masi.ibp.fr)
7 * Laboratoire MASI - Institut Blaise Pascal
8 * Universite Pierre et Marie Curie (Paris VI)
9 *
10 * from
11 *
12 * linux/include/linux/minix_fs.h
13 *
14 * Copyright (C) 1991, 1992 Linus Torvalds
15 */
16
3dcf5451
CH
17#ifndef _EXT4_H
18#define _EXT4_H
ac27a0ec 19
31d21d21 20#include <linux/refcount.h>
ac27a0ec 21#include <linux/types.h>
3a5b2ecd 22#include <linux/blkdev.h>
ac27a0ec 23#include <linux/magic.h>
2ccb5fb9 24#include <linux/jbd2.h>
60e58e0f 25#include <linux/quota.h>
d444c3c3
TT
26#include <linux/rwsem.h>
27#include <linux/rbtree.h>
28#include <linux/seqlock.h>
29#include <linux/mutex.h>
ca0faba0
TT
30#include <linux/timer.h>
31#include <linux/wait.h>
174cd4b1 32#include <linux/sched/signal.h>
ca0faba0
TT
33#include <linux/blockgroup_lock.h>
34#include <linux/percpu_counter.h>
efbed4dc 35#include <linux/ratelimit.h>
0441984a 36#include <crypto/hash.h>
a633f5a3 37#include <linux/falloc.h>
c8585c6f 38#include <linux/percpu-rwsem.h>
10c5db28 39#include <linux/fiemap.h>
4d92dc0f
BH
40#ifdef __KERNEL__
41#include <linux/compat.h>
42#endif
519fe1ba 43#include <uapi/linux/ext4.h>
c9de560d 44
734f0d24 45#include <linux/fscrypt.h>
c93d8f88 46#include <linux/fsverity.h>
734f0d24 47
072ebb3b
TT
48#include <linux/compiler.h>
49
ac27a0ec 50/*
8a35694e 51 * The fourth extended filesystem constants/structures
ac27a0ec
DK
52 */
53
a2821e34
AZ
54/*
55 * with AGGRESSIVE_CHECK allocator runs consistency checks over
56 * structures. these checks slow things down a lot
57 */
58#define AGGRESSIVE_CHECK__
59
60/*
61 * with DOUBLE_CHECK defined mballoc creates persistent in-core
62 * bitmaps, maintains and uses them to check for double allocations
63 */
64#define DOUBLE_CHECK__
65
ac27a0ec 66/*
617ba13b 67 * Define EXT4FS_DEBUG to produce debug messages
ac27a0ec 68 */
617ba13b 69#undef EXT4FS_DEBUG
ac27a0ec 70
ac27a0ec
DK
71/*
72 * Debug code
73 */
617ba13b
MC
74#ifdef EXT4FS_DEBUG
75#define ext4_debug(f, a...) \
ac27a0ec 76 do { \
af5bc92d 77 printk(KERN_DEBUG "EXT4-fs DEBUG (%s, %d): %s:", \
4db9c54a 78 __FILE__, __LINE__, __func__); \
af5bc92d 79 printk(KERN_DEBUG f, ## a); \
ac27a0ec
DK
80 } while (0)
81#else
ace36ad4 82#define ext4_debug(fmt, ...) no_printk(fmt, ##__VA_ARGS__)
ac27a0ec
DK
83#endif
84
70aa1554
RH
85 /*
86 * Turn on EXT_DEBUG to enable ext4_ext_show_path/leaf/move in extents.c
87 */
88#define EXT_DEBUG__
89
4a092d73 90/*
70aa1554 91 * Dynamic printk for controlled extents debugging.
4a092d73 92 */
70aa1554
RH
93#ifdef CONFIG_EXT4_DEBUG
94#define ext_debug(ino, fmt, ...) \
95 pr_debug("[%s/%d] EXT4-fs (%s): ino %lu: (%s, %d): %s:" fmt, \
96 current->comm, task_pid_nr(current), \
97 ino->i_sb->s_id, ino->i_ino, __FILE__, __LINE__, \
98 __func__, ##__VA_ARGS__)
4a092d73 99#else
70aa1554 100#define ext_debug(ino, fmt, ...) no_printk(fmt, ##__VA_ARGS__)
4a092d73
TT
101#endif
102
837c23fb
CX
103#define ASSERT(assert) \
104do { \
105 if (unlikely(!(assert))) { \
106 printk(KERN_EMERG \
107 "Assertion failure in %s() at %s:%d: '%s'\n", \
108 __func__, __FILE__, __LINE__, #assert); \
109 BUG(); \
110 } \
111} while (0)
112
d444c3c3
TT
113/* data type for block offset of block group */
114typedef int ext4_grpblk_t;
115
116/* data type for filesystem-wide blocks number */
117typedef unsigned long long ext4_fsblk_t;
118
119/* data type for file logical block number */
120typedef __u32 ext4_lblk_t;
121
122/* data type for block group number */
123typedef unsigned int ext4_group_t;
124
331573fe
NJ
125enum SHIFT_DIRECTION {
126 SHIFT_LEFT = 0,
127 SHIFT_RIGHT,
128};
129
4eb7a4a1 130/*
4c0cfebd
TT
131 * For each criteria, mballoc has slightly different way of finding
132 * the required blocks nad usually, higher the criteria the slower the
133 * allocation. We start at lower criterias and keep falling back to
134 * higher ones if we are not able to find any blocks. Lower (earlier)
135 * criteria are faster.
4eb7a4a1
OM
136 */
137enum criteria {
f52f3d2b 138 /*
4c0cfebd
TT
139 * Used when number of blocks needed is a power of 2. This
140 * doesn't trigger any disk IO except prefetch and is the
141 * fastest criteria.
f52f3d2b
OM
142 */
143 CR_POWER2_ALIGNED,
144
145 /*
4c0cfebd
TT
146 * Tries to lookup in-memory data structures to find the most
147 * suitable group that satisfies goal request. No disk IO
148 * except block prefetch.
f52f3d2b
OM
149 */
150 CR_GOAL_LEN_FAST,
151
152 /*
4c0cfebd
TT
153 * Same as CR_GOAL_LEN_FAST but is allowed to reduce the goal
154 * length to the best available length for faster allocation.
f52f3d2b
OM
155 */
156 CR_BEST_AVAIL_LEN,
157
158 /*
4c0cfebd
TT
159 * Reads each block group sequentially, performing disk IO if
160 * necessary, to find find_suitable block group. Tries to
161 * allocate goal length but might trim the request if nothing
162 * is found after enough tries.
f52f3d2b
OM
163 */
164 CR_GOAL_LEN_SLOW,
165
166 /*
4c0cfebd
TT
167 * Finds the first free set of blocks and allocates
168 * those. This is only used in rare cases when
169 * CR_GOAL_LEN_SLOW also fails to allocate anything.
f52f3d2b
OM
170 */
171 CR_ANY_FREE,
4c0cfebd
TT
172
173 /*
174 * Number of criterias defined.
175 */
176 EXT4_MB_NUM_CRS
4eb7a4a1
OM
177};
178
296c355c 179/*
60e6679e 180 * Flags used in mballoc's allocation_context flags field.
296c355c
TT
181 *
182 * Also used to show what's going on for debugging purposes when the
183 * flag field is exported via the traceport interface
184 */
d444c3c3 185
c9de560d 186/* prefer goal again. length */
0ef90db9 187#define EXT4_MB_HINT_MERGE 0x0001
c9de560d 188/* blocks already reserved */
0ef90db9 189#define EXT4_MB_HINT_RESERVED 0x0002
c9de560d 190/* metadata is being allocated */
0ef90db9 191#define EXT4_MB_HINT_METADATA 0x0004
c9de560d 192/* first blocks in the file */
0ef90db9 193#define EXT4_MB_HINT_FIRST 0x0008
c9de560d 194/* search for the best chunk */
0ef90db9 195#define EXT4_MB_HINT_BEST 0x0010
c9de560d 196/* data is being allocated */
0ef90db9 197#define EXT4_MB_HINT_DATA 0x0020
c9de560d 198/* don't preallocate (for tails) */
0ef90db9 199#define EXT4_MB_HINT_NOPREALLOC 0x0040
c9de560d 200/* allocate for locality group */
0ef90db9 201#define EXT4_MB_HINT_GROUP_ALLOC 0x0080
c9de560d 202/* allocate goal blocks or none */
0ef90db9 203#define EXT4_MB_HINT_GOAL_ONLY 0x0100
c9de560d 204/* goal is meaningful */
0ef90db9 205#define EXT4_MB_HINT_TRY_GOAL 0x0200
d2a17637 206/* blocks already pre-reserved by delayed allocation */
0ef90db9 207#define EXT4_MB_DELALLOC_RESERVED 0x0400
4ba74d00
TT
208/* We are doing stream allocation */
209#define EXT4_MB_STREAM_ALLOC 0x0800
55f020db
AH
210/* Use reserved root blocks if needed */
211#define EXT4_MB_USE_ROOT_BLOCKS 0x1000
27dd4385
LC
212/* Use blocks from reserved pool */
213#define EXT4_MB_USE_RESERVED 0x2000
99377830
RH
214/* Do strict check for free blocks while retrying block allocation */
215#define EXT4_MB_STRICT_CHECK 0x4000
196e402a 216/* Large fragment size list lookup succeeded at least once for cr = 0 */
f52f3d2b 217#define EXT4_MB_CR_POWER2_ALIGNED_OPTIMIZED 0x8000
196e402a 218/* Avg fragment size rb tree lookup succeeded at least once for cr = 1 */
f52f3d2b 219#define EXT4_MB_CR_GOAL_LEN_FAST_OPTIMIZED 0x00010000
7e170922 220/* Avg fragment size rb tree lookup succeeded at least once for cr = 1.5 */
f52f3d2b 221#define EXT4_MB_CR_BEST_AVAIL_LEN_OPTIMIZED 0x00020000
7e170922 222
c9de560d
AT
223struct ext4_allocation_request {
224 /* target inode for block we're allocating */
225 struct inode *inode;
726447d8
ES
226 /* how many blocks we want to allocate */
227 unsigned int len;
c9de560d
AT
228 /* logical block in target inode */
229 ext4_lblk_t logical;
c9de560d
AT
230 /* the closest logical allocated block to the left */
231 ext4_lblk_t lleft;
c9de560d
AT
232 /* the closest logical allocated block to the right */
233 ext4_lblk_t lright;
726447d8
ES
234 /* phys. target (a hint) */
235 ext4_fsblk_t goal;
236 /* phys. block for the closest logical allocated block to the left */
237 ext4_fsblk_t pleft;
238 /* phys. block for the closest logical allocated block to the right */
c9de560d 239 ext4_fsblk_t pright;
c9de560d 240 /* flags. see above EXT4_MB_HINT_* */
498e5f24 241 unsigned int flags;
c9de560d
AT
242};
243
e35fd660
TT
244/*
245 * Logical to physical block mapping, used by ext4_map_blocks()
246 *
247 * This structure is used to pass requests into ext4_map_blocks() as
248 * well as to store the information returned by ext4_map_blocks(). It
249 * takes less room on the stack than a struct buffer_head.
250 */
6db07461
RH
251#define EXT4_MAP_NEW BIT(BH_New)
252#define EXT4_MAP_MAPPED BIT(BH_Mapped)
253#define EXT4_MAP_UNWRITTEN BIT(BH_Unwritten)
254#define EXT4_MAP_BOUNDARY BIT(BH_Boundary)
e35fd660 255#define EXT4_MAP_FLAGS (EXT4_MAP_NEW | EXT4_MAP_MAPPED |\
cbd7584e 256 EXT4_MAP_UNWRITTEN | EXT4_MAP_BOUNDARY)
e35fd660
TT
257
258struct ext4_map_blocks {
259 ext4_fsblk_t m_pblk;
260 ext4_lblk_t m_lblk;
261 unsigned int m_len;
262 unsigned int m_flags;
263};
264
7727ae52 265/*
266 * Block validity checking, system zone rbtree.
267 */
268struct ext4_system_blocks {
269 struct rb_root root;
270 struct rcu_head rcu;
271};
272
bd2d0210
TT
273/*
274 * Flags for ext4_io_end->flags
275 */
276#define EXT4_IO_END_UNWRITTEN 0x0001
bd2d0210 277
c8cc8816
RH
278struct ext4_io_end_vec {
279 struct list_head list; /* list of io_end_vec */
280 loff_t offset; /* offset in the file */
281 ssize_t size; /* size of the extent */
282};
283
4188188b 284/*
556615dc 285 * For converting unwritten extents on a work queue. 'handle' is used for
6b523df4 286 * buffered writeback.
4188188b 287 */
0031462b 288typedef struct ext4_io_end {
5b3ff237 289 struct list_head list; /* per-file finished IO list */
6b523df4
JK
290 handle_t *handle; /* handle reserved for extent
291 * conversion */
0031462b 292 struct inode *inode; /* file being written to */
b0857d30
JK
293 struct bio *bio; /* Linked list of completed
294 * bios covering the extent */
8d5d02e6 295 unsigned int flag; /* unwritten or not */
31d21d21 296 refcount_t count; /* reference counter */
c8cc8816 297 struct list_head list_vec; /* list of ext4_io_end_vec */
0031462b
MC
298} ext4_io_end_t;
299
bd2d0210 300struct ext4_io_submit {
5a33911f 301 struct writeback_control *io_wbc;
bd2d0210
TT
302 struct bio *io_bio;
303 ext4_io_end_t *io_end;
bd2d0210
TT
304 sector_t io_next_block;
305};
306
ac27a0ec
DK
307/*
308 * Special inodes numbers
309 */
617ba13b
MC
310#define EXT4_BAD_INO 1 /* Bad blocks inode */
311#define EXT4_ROOT_INO 2 /* Root inode */
ae812306
AK
312#define EXT4_USR_QUOTA_INO 3 /* User quota inode */
313#define EXT4_GRP_QUOTA_INO 4 /* Group quota inode */
617ba13b
MC
314#define EXT4_BOOT_LOADER_INO 5 /* Boot loader inode */
315#define EXT4_UNDEL_DIR_INO 6 /* Undelete directory inode */
316#define EXT4_RESIZE_INO 7 /* Reserved group descriptors inode */
317#define EXT4_JOURNAL_INO 8 /* Journal inode */
ac27a0ec 318
617ba13b
MC
319/* First non-reserved inode for old ext4 filesystems */
320#define EXT4_GOOD_OLD_FIRST_INO 11
ac27a0ec
DK
321
322/*
323 * Maximal count of links to a file
324 */
f8628a14 325#define EXT4_LINK_MAX 65000
ac27a0ec
DK
326
327/*
328 * Macro-instructions used to manage several block sizes
329 */
617ba13b 330#define EXT4_MIN_BLOCK_SIZE 1024
afc7cbca
TS
331#define EXT4_MAX_BLOCK_SIZE 65536
332#define EXT4_MIN_BLOCK_LOG_SIZE 10
fb1813f4 333#define EXT4_MAX_BLOCK_LOG_SIZE 16
9e47a4c9 334#define EXT4_MAX_CLUSTER_LOG_SIZE 30
ac27a0ec 335#ifdef __KERNEL__
617ba13b 336# define EXT4_BLOCK_SIZE(s) ((s)->s_blocksize)
ac27a0ec 337#else
617ba13b 338# define EXT4_BLOCK_SIZE(s) (EXT4_MIN_BLOCK_SIZE << (s)->s_log_block_size)
ac27a0ec 339#endif
af5bc92d 340#define EXT4_ADDR_PER_BLOCK(s) (EXT4_BLOCK_SIZE(s) / sizeof(__u32))
281b5995
TT
341#define EXT4_CLUSTER_SIZE(s) (EXT4_BLOCK_SIZE(s) << \
342 EXT4_SB(s)->s_cluster_bits)
ac27a0ec 343#ifdef __KERNEL__
617ba13b 344# define EXT4_BLOCK_SIZE_BITS(s) ((s)->s_blocksize_bits)
281b5995 345# define EXT4_CLUSTER_BITS(s) (EXT4_SB(s)->s_cluster_bits)
ac27a0ec 346#else
617ba13b 347# define EXT4_BLOCK_SIZE_BITS(s) ((s)->s_log_block_size + 10)
ac27a0ec
DK
348#endif
349#ifdef __KERNEL__
617ba13b
MC
350#define EXT4_ADDR_PER_BLOCK_BITS(s) (EXT4_SB(s)->s_addr_per_block_bits)
351#define EXT4_INODE_SIZE(s) (EXT4_SB(s)->s_inode_size)
352#define EXT4_FIRST_INO(s) (EXT4_SB(s)->s_first_ino)
ac27a0ec 353#else
617ba13b
MC
354#define EXT4_INODE_SIZE(s) (((s)->s_rev_level == EXT4_GOOD_OLD_REV) ? \
355 EXT4_GOOD_OLD_INODE_SIZE : \
ac27a0ec 356 (s)->s_inode_size)
617ba13b
MC
357#define EXT4_FIRST_INO(s) (((s)->s_rev_level == EXT4_GOOD_OLD_REV) ? \
358 EXT4_GOOD_OLD_FIRST_INO : \
ac27a0ec
DK
359 (s)->s_first_ino)
360#endif
a2df2a63 361#define EXT4_BLOCK_ALIGN(size, blkbits) ALIGN((size), (1 << (blkbits)))
518eaa63
FF
362#define EXT4_MAX_BLOCKS(size, offset, blkbits) \
363 ((EXT4_BLOCK_ALIGN(size + offset, blkbits) >> blkbits) - (offset >> \
364 blkbits))
ac27a0ec 365
d5b8f310
TT
366/* Translate a block number to a cluster number */
367#define EXT4_B2C(sbi, blk) ((blk) >> (sbi)->s_cluster_bits)
368/* Translate a cluster number to a block number */
369#define EXT4_C2B(sbi, cluster) ((cluster) << (sbi)->s_cluster_bits)
370/* Translate # of blks to # of clusters */
371#define EXT4_NUM_B2C(sbi, blks) (((blks) + (sbi)->s_cluster_ratio - 1) >> \
372 (sbi)->s_cluster_bits)
f5a44db5
TT
373/* Mask out the low bits to get the starting block of the cluster */
374#define EXT4_PBLK_CMASK(s, pblk) ((pblk) & \
375 ~((ext4_fsblk_t) (s)->s_cluster_ratio - 1))
376#define EXT4_LBLK_CMASK(s, lblk) ((lblk) & \
377 ~((ext4_lblk_t) (s)->s_cluster_ratio - 1))
8fcc3a58
EW
378/* Fill in the low bits to get the last block of the cluster */
379#define EXT4_LBLK_CFILL(sbi, lblk) ((lblk) | \
380 ((ext4_lblk_t) (sbi)->s_cluster_ratio - 1))
f5a44db5
TT
381/* Get the cluster offset */
382#define EXT4_PBLK_COFF(s, pblk) ((pblk) & \
383 ((ext4_fsblk_t) (s)->s_cluster_ratio - 1))
384#define EXT4_LBLK_COFF(s, lblk) ((lblk) & \
385 ((ext4_lblk_t) (s)->s_cluster_ratio - 1))
d5b8f310 386
ac27a0ec
DK
387/*
388 * Structure of a blocks group descriptor
389 */
617ba13b 390struct ext4_group_desc
ac27a0ec 391{
3a14589c 392 __le32 bg_block_bitmap_lo; /* Blocks bitmap block */
5272f837
AK
393 __le32 bg_inode_bitmap_lo; /* Inodes bitmap block */
394 __le32 bg_inode_table_lo; /* Inodes table block */
560671a0
AK
395 __le16 bg_free_blocks_count_lo;/* Free blocks count */
396 __le16 bg_free_inodes_count_lo;/* Free inodes count */
397 __le16 bg_used_dirs_count_lo; /* Directories count */
717d50e4 398 __le16 bg_flags; /* EXT4_BG_flags (INODE_UNINIT, etc) */
e6153918
DW
399 __le32 bg_exclude_bitmap_lo; /* Exclude bitmap for snapshots */
400 __le16 bg_block_bitmap_csum_lo;/* crc32c(s_uuid+grp_num+bbitmap) LE */
401 __le16 bg_inode_bitmap_csum_lo;/* crc32c(s_uuid+grp_num+ibitmap) LE */
560671a0 402 __le16 bg_itable_unused_lo; /* Unused inodes count */
717d50e4 403 __le16 bg_checksum; /* crc16(sb_uuid+group+desc) */
8fadc143
AR
404 __le32 bg_block_bitmap_hi; /* Blocks bitmap block MSB */
405 __le32 bg_inode_bitmap_hi; /* Inodes bitmap block MSB */
406 __le32 bg_inode_table_hi; /* Inodes table block MSB */
91b51a01
CL
407 __le16 bg_free_blocks_count_hi;/* Free blocks count MSB */
408 __le16 bg_free_inodes_count_hi;/* Free inodes count MSB */
409 __le16 bg_used_dirs_count_hi; /* Directories count MSB */
560671a0 410 __le16 bg_itable_unused_hi; /* Unused inodes count MSB */
e6153918
DW
411 __le32 bg_exclude_bitmap_hi; /* Exclude bitmap block MSB */
412 __le16 bg_block_bitmap_csum_hi;/* crc32c(s_uuid+grp_num+bbitmap) BE */
413 __le16 bg_inode_bitmap_csum_hi;/* crc32c(s_uuid+grp_num+ibitmap) BE */
414 __u32 bg_reserved;
ac27a0ec
DK
415};
416
41a246d1
DW
417#define EXT4_BG_INODE_BITMAP_CSUM_HI_END \
418 (offsetof(struct ext4_group_desc, bg_inode_bitmap_csum_hi) + \
419 sizeof(__le16))
420#define EXT4_BG_BLOCK_BITMAP_CSUM_HI_END \
421 (offsetof(struct ext4_group_desc, bg_block_bitmap_csum_hi) + \
422 sizeof(__le16))
423
772cb7c8
JS
424/*
425 * Structure of a flex block group info
426 */
427
428struct flex_groups {
90ba983f
TT
429 atomic64_t free_clusters;
430 atomic_t free_inodes;
431 atomic_t used_dirs;
772cb7c8
JS
432};
433
717d50e4
AD
434#define EXT4_BG_INODE_UNINIT 0x0001 /* Inode table/bitmap not in use */
435#define EXT4_BG_BLOCK_UNINIT 0x0002 /* Block bitmap not in use */
436#define EXT4_BG_INODE_ZEROED 0x0004 /* On-disk itable initialized to zero */
437
ac27a0ec
DK
438/*
439 * Macro-instructions used to manage group descriptors
440 */
0d1ee42f 441#define EXT4_MIN_DESC_SIZE 32
8fadc143 442#define EXT4_MIN_DESC_SIZE_64BIT 64
0d1ee42f
AR
443#define EXT4_MAX_DESC_SIZE EXT4_MIN_BLOCK_SIZE
444#define EXT4_DESC_SIZE(s) (EXT4_SB(s)->s_desc_size)
ac27a0ec 445#ifdef __KERNEL__
617ba13b 446# define EXT4_BLOCKS_PER_GROUP(s) (EXT4_SB(s)->s_blocks_per_group)
281b5995 447# define EXT4_CLUSTERS_PER_GROUP(s) (EXT4_SB(s)->s_clusters_per_group)
617ba13b
MC
448# define EXT4_DESC_PER_BLOCK(s) (EXT4_SB(s)->s_desc_per_block)
449# define EXT4_INODES_PER_GROUP(s) (EXT4_SB(s)->s_inodes_per_group)
450# define EXT4_DESC_PER_BLOCK_BITS(s) (EXT4_SB(s)->s_desc_per_block_bits)
ac27a0ec 451#else
617ba13b 452# define EXT4_BLOCKS_PER_GROUP(s) ((s)->s_blocks_per_group)
0d1ee42f 453# define EXT4_DESC_PER_BLOCK(s) (EXT4_BLOCK_SIZE(s) / EXT4_DESC_SIZE(s))
617ba13b 454# define EXT4_INODES_PER_GROUP(s) ((s)->s_inodes_per_group)
ac27a0ec
DK
455#endif
456
457/*
458 * Constants relative to the data blocks
459 */
617ba13b
MC
460#define EXT4_NDIR_BLOCKS 12
461#define EXT4_IND_BLOCK EXT4_NDIR_BLOCKS
462#define EXT4_DIND_BLOCK (EXT4_IND_BLOCK + 1)
463#define EXT4_TIND_BLOCK (EXT4_DIND_BLOCK + 1)
464#define EXT4_N_BLOCKS (EXT4_TIND_BLOCK + 1)
ac27a0ec
DK
465
466/*
467 * Inode flags
468 */
617ba13b
MC
469#define EXT4_SECRM_FL 0x00000001 /* Secure deletion */
470#define EXT4_UNRM_FL 0x00000002 /* Undelete */
471#define EXT4_COMPR_FL 0x00000004 /* Compress file */
472#define EXT4_SYNC_FL 0x00000008 /* Synchronous updates */
473#define EXT4_IMMUTABLE_FL 0x00000010 /* Immutable file */
474#define EXT4_APPEND_FL 0x00000020 /* writes to file may only append */
475#define EXT4_NODUMP_FL 0x00000040 /* do not dump file */
476#define EXT4_NOATIME_FL 0x00000080 /* do not update atime */
ac27a0ec 477/* Reserved for compression usage... */
617ba13b
MC
478#define EXT4_DIRTY_FL 0x00000100
479#define EXT4_COMPRBLK_FL 0x00000200 /* One or more compressed clusters */
480#define EXT4_NOCOMPR_FL 0x00000400 /* Don't compress */
3edc18d8
TT
481 /* nb: was previously EXT2_ECOMPR_FL */
482#define EXT4_ENCRYPT_FL 0x00000800 /* encrypted file */
ac27a0ec 483/* End compression flags --- maybe not all used */
617ba13b
MC
484#define EXT4_INDEX_FL 0x00001000 /* hash-indexed directory */
485#define EXT4_IMAGIC_FL 0x00002000 /* AFS directory */
486#define EXT4_JOURNAL_DATA_FL 0x00004000 /* file data should be journaled */
487#define EXT4_NOTAIL_FL 0x00008000 /* file tail should not be merged */
488#define EXT4_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */
489#define EXT4_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/
8180a562 490#define EXT4_HUGE_FILE_FL 0x00040000 /* Set to each huge file */
a86c6181 491#define EXT4_EXTENTS_FL 0x00080000 /* Inode uses extents */
c93d8f88 492#define EXT4_VERITY_FL 0x00100000 /* Verity protected inode */
f710b4b9 493#define EXT4_EA_INODE_FL 0x00200000 /* Inode used for large EA */
4337ecd1 494/* 0x00400000 was formerly EXT4_EOFBLOCKS_FL */
b383a73f
IW
495
496#define EXT4_DAX_FL 0x02000000 /* Inode is DAX */
497
67cf5b09 498#define EXT4_INLINE_DATA_FL 0x10000000 /* Inode has inline data. */
8b4953e1 499#define EXT4_PROJINHERIT_FL 0x20000000 /* Create with parents projid */
de8ff14c 500#define EXT4_CASEFOLD_FL 0x40000000 /* Casefolded directory */
8180a562 501#define EXT4_RESERVED_FL 0x80000000 /* reserved for ext4 lib */
ac27a0ec 502
2a12e147
EB
503/* User modifiable flags */
504#define EXT4_FL_USER_MODIFIABLE (EXT4_SECRM_FL | \
505 EXT4_UNRM_FL | \
506 EXT4_COMPR_FL | \
507 EXT4_SYNC_FL | \
508 EXT4_IMMUTABLE_FL | \
509 EXT4_APPEND_FL | \
510 EXT4_NODUMP_FL | \
511 EXT4_NOATIME_FL | \
512 EXT4_JOURNAL_DATA_FL | \
513 EXT4_NOTAIL_FL | \
514 EXT4_DIRSYNC_FL | \
515 EXT4_TOPDIR_FL | \
516 EXT4_EXTENTS_FL | \
517 0x00400000 /* EXT4_EOFBLOCKS_FL */ | \
518 EXT4_DAX_FL | \
519 EXT4_PROJINHERIT_FL | \
520 EXT4_CASEFOLD_FL)
521
522/* User visible flags */
523#define EXT4_FL_USER_VISIBLE (EXT4_FL_USER_MODIFIABLE | \
524 EXT4_DIRTY_FL | \
525 EXT4_COMPRBLK_FL | \
526 EXT4_NOCOMPR_FL | \
527 EXT4_ENCRYPT_FL | \
528 EXT4_INDEX_FL | \
529 EXT4_VERITY_FL | \
530 EXT4_INLINE_DATA_FL)
ac27a0ec 531
8fa43a81
DG
532/* Flags that should be inherited by new inodes from their parent. */
533#define EXT4_FL_INHERITED (EXT4_SECRM_FL | EXT4_UNRM_FL | EXT4_COMPR_FL |\
1cd9f097 534 EXT4_SYNC_FL | EXT4_NODUMP_FL | EXT4_NOATIME_FL |\
8fa43a81 535 EXT4_NOCOMPR_FL | EXT4_JOURNAL_DATA_FL |\
040cb378 536 EXT4_NOTAIL_FL | EXT4_DIRSYNC_FL |\
b383a73f
IW
537 EXT4_PROJINHERIT_FL | EXT4_CASEFOLD_FL |\
538 EXT4_DAX_FL)
8fa43a81 539
2dc6b0d4 540/* Flags that are appropriate for regular files (all but dir-specific ones). */
7ddf79a1
WS
541#define EXT4_REG_FLMASK (~(EXT4_DIRSYNC_FL | EXT4_TOPDIR_FL | EXT4_CASEFOLD_FL |\
542 EXT4_PROJINHERIT_FL))
2dc6b0d4
DG
543
544/* Flags that are appropriate for non-directories/regular files. */
545#define EXT4_OTHER_FLMASK (EXT4_NODUMP_FL | EXT4_NOATIME_FL)
546
abdc644e 547/* The only flags that should be swapped */
548#define EXT4_FL_SHOULD_SWAP (EXT4_HUGE_FILE_FL | EXT4_EXTENTS_FL)
549
b383a73f
IW
550/* Flags which are mutually exclusive to DAX */
551#define EXT4_DAX_MUT_EXCL (EXT4_VERITY_FL | EXT4_ENCRYPT_FL |\
aa2f7792 552 EXT4_JOURNAL_DATA_FL | EXT4_INLINE_DATA_FL)
b383a73f 553
2dc6b0d4
DG
554/* Mask out flags that are inappropriate for the given type of inode. */
555static inline __u32 ext4_mask_flags(umode_t mode, __u32 flags)
556{
557 if (S_ISDIR(mode))
558 return flags;
559 else if (S_ISREG(mode))
560 return flags & EXT4_REG_FLMASK;
561 else
562 return flags & EXT4_OTHER_FLMASK;
563}
564
12e9b892
DM
565/*
566 * Inode flags used for atomic set/get
567 */
568enum {
569 EXT4_INODE_SECRM = 0, /* Secure deletion */
570 EXT4_INODE_UNRM = 1, /* Undelete */
571 EXT4_INODE_COMPR = 2, /* Compress file */
572 EXT4_INODE_SYNC = 3, /* Synchronous updates */
573 EXT4_INODE_IMMUTABLE = 4, /* Immutable file */
574 EXT4_INODE_APPEND = 5, /* writes to file may only append */
575 EXT4_INODE_NODUMP = 6, /* do not dump file */
576 EXT4_INODE_NOATIME = 7, /* do not update atime */
577/* Reserved for compression usage... */
578 EXT4_INODE_DIRTY = 8,
579 EXT4_INODE_COMPRBLK = 9, /* One or more compressed clusters */
580 EXT4_INODE_NOCOMPR = 10, /* Don't compress */
f542fbe8 581 EXT4_INODE_ENCRYPT = 11, /* Encrypted file */
12e9b892
DM
582/* End compression flags --- maybe not all used */
583 EXT4_INODE_INDEX = 12, /* hash-indexed directory */
584 EXT4_INODE_IMAGIC = 13, /* AFS directory */
585 EXT4_INODE_JOURNAL_DATA = 14, /* file data should be journaled */
586 EXT4_INODE_NOTAIL = 15, /* file tail should not be merged */
587 EXT4_INODE_DIRSYNC = 16, /* dirsync behaviour (directories only) */
588 EXT4_INODE_TOPDIR = 17, /* Top of directory hierarchies*/
589 EXT4_INODE_HUGE_FILE = 18, /* Set to each huge file */
590 EXT4_INODE_EXTENTS = 19, /* Inode uses extents */
c93d8f88 591 EXT4_INODE_VERITY = 20, /* Verity protected inode */
12e9b892 592 EXT4_INODE_EA_INODE = 21, /* Inode used for large EA */
4337ecd1 593/* 22 was formerly EXT4_INODE_EOFBLOCKS */
b383a73f 594 EXT4_INODE_DAX = 25, /* Inode is DAX */
67cf5b09 595 EXT4_INODE_INLINE_DATA = 28, /* Data in inode. */
8b4953e1 596 EXT4_INODE_PROJINHERIT = 29, /* Create with parents projid */
de8ff14c 597 EXT4_INODE_CASEFOLD = 30, /* Casefolded directory */
12e9b892
DM
598 EXT4_INODE_RESERVED = 31, /* reserved for ext4 lib */
599};
600
12e9b892 601/*
9a4c8019
CM
602 * Since it's pretty easy to mix up bit numbers and hex values, we use a
603 * build-time check to make sure that EXT4_XXX_FL is consistent with respect to
604 * EXT4_INODE_XXX. If all is well, the macros will be dropped, so, it won't cost
605 * any extra space in the compiled kernel image, otherwise, the build will fail.
606 * It's important that these values are the same, since we are using
607 * EXT4_INODE_XXX to test for flag values, but EXT4_XXX_FL must be consistent
608 * with the values of FS_XXX_FL defined in include/linux/fs.h and the on-disk
609 * values found in ext2, ext3 and ext4 filesystems, and of course the values
610 * defined in e2fsprogs.
12e9b892
DM
611 *
612 * It's not paranoia if the Murphy's Law really *is* out to get you. :-)
613 */
3bf678a0 614#define TEST_FLAG_VALUE(FLAG) (EXT4_##FLAG##_FL == (1U << EXT4_INODE_##FLAG))
9a4c8019
CM
615#define CHECK_FLAG_VALUE(FLAG) BUILD_BUG_ON(!TEST_FLAG_VALUE(FLAG))
616
12e9b892
DM
617static inline void ext4_check_flag_values(void)
618{
619 CHECK_FLAG_VALUE(SECRM);
620 CHECK_FLAG_VALUE(UNRM);
621 CHECK_FLAG_VALUE(COMPR);
622 CHECK_FLAG_VALUE(SYNC);
623 CHECK_FLAG_VALUE(IMMUTABLE);
624 CHECK_FLAG_VALUE(APPEND);
625 CHECK_FLAG_VALUE(NODUMP);
626 CHECK_FLAG_VALUE(NOATIME);
627 CHECK_FLAG_VALUE(DIRTY);
628 CHECK_FLAG_VALUE(COMPRBLK);
629 CHECK_FLAG_VALUE(NOCOMPR);
3edc18d8 630 CHECK_FLAG_VALUE(ENCRYPT);
12e9b892
DM
631 CHECK_FLAG_VALUE(INDEX);
632 CHECK_FLAG_VALUE(IMAGIC);
633 CHECK_FLAG_VALUE(JOURNAL_DATA);
634 CHECK_FLAG_VALUE(NOTAIL);
635 CHECK_FLAG_VALUE(DIRSYNC);
636 CHECK_FLAG_VALUE(TOPDIR);
637 CHECK_FLAG_VALUE(HUGE_FILE);
638 CHECK_FLAG_VALUE(EXTENTS);
c93d8f88 639 CHECK_FLAG_VALUE(VERITY);
12e9b892 640 CHECK_FLAG_VALUE(EA_INODE);
67cf5b09 641 CHECK_FLAG_VALUE(INLINE_DATA);
8b4953e1 642 CHECK_FLAG_VALUE(PROJINHERIT);
de8ff14c 643 CHECK_FLAG_VALUE(CASEFOLD);
12e9b892
DM
644 CHECK_FLAG_VALUE(RESERVED);
645}
646
4d92dc0f
BH
647#if defined(__KERNEL__) && defined(CONFIG_COMPAT)
648struct compat_ext4_new_group_input {
649 u32 group;
650 compat_u64 block_bitmap;
651 compat_u64 inode_bitmap;
652 compat_u64 inode_table;
653 u32 blocks_count;
654 u16 reserved_blocks;
655 u16 unused;
656};
657#endif
658
617ba13b
MC
659/* The struct ext4_new_group_input in kernel space, with free_blocks_count */
660struct ext4_new_group_data {
ac27a0ec 661 __u32 group;
bd81d8ee
LV
662 __u64 block_bitmap;
663 __u64 inode_bitmap;
664 __u64 inode_table;
ac27a0ec
DK
665 __u32 blocks_count;
666 __u16 reserved_blocks;
d77147ff 667 __u16 mdata_blocks;
668 __u32 free_clusters_count;
ac27a0ec
DK
669};
670
33afdcc5
YY
671/* Indexes used to index group tables in ext4_new_group_data */
672enum {
673 BLOCK_BITMAP = 0, /* block bitmap */
674 INODE_BITMAP, /* inode bitmap */
675 INODE_TABLE, /* inode tables */
676 GROUP_TABLE_COUNT,
677};
678
a2df2a63 679/*
79e83036 680 * Flags used by ext4_map_blocks()
a2df2a63 681 */
556615dc 682 /* Allocate any needed blocks and/or convert an unwritten
c2177057 683 extent to be an initialized ext4 */
2ac3b6e0 684#define EXT4_GET_BLOCKS_CREATE 0x0001
556615dc
LC
685 /* Request the creation of an unwritten extent */
686#define EXT4_GET_BLOCKS_UNWRIT_EXT 0x0002
687#define EXT4_GET_BLOCKS_CREATE_UNWRIT_EXT (EXT4_GET_BLOCKS_UNWRIT_EXT|\
c2177057 688 EXT4_GET_BLOCKS_CREATE)
27dd4385
LC
689 /* Caller is from the delayed allocation writeout path
690 * finally doing the actual allocation of delayed blocks */
03f5d8bc 691#define EXT4_GET_BLOCKS_DELALLOC_RESERVE 0x0004
0031462b 692 /* caller is from the direct IO path, request to creation of an
556615dc 693 unwritten extents if not allocated, split the unwritten
0031462b 694 extent if blocks has been preallocated already*/
c7064ef1 695#define EXT4_GET_BLOCKS_PRE_IO 0x0008
1296cc85 696#define EXT4_GET_BLOCKS_CONVERT 0x0010
c7064ef1 697#define EXT4_GET_BLOCKS_IO_CREATE_EXT (EXT4_GET_BLOCKS_PRE_IO|\
556615dc 698 EXT4_GET_BLOCKS_CREATE_UNWRIT_EXT)
c7064ef1
JZ
699 /* Convert extent to initialized after IO complete */
700#define EXT4_GET_BLOCKS_IO_CONVERT_EXT (EXT4_GET_BLOCKS_CONVERT|\
556615dc 701 EXT4_GET_BLOCKS_CREATE_UNWRIT_EXT)
27dd4385
LC
702 /* Eventual metadata allocation (due to growing extent tree)
703 * should not fail, so try to use reserved blocks for that.*/
704#define EXT4_GET_BLOCKS_METADATA_NOFAIL 0x0020
556b27ab
VH
705 /* Don't normalize allocation size (used for fallocate) */
706#define EXT4_GET_BLOCKS_NO_NORMALIZE 0x0040
b8a86845 707 /* Convert written extents to unwritten */
2dcba478 708#define EXT4_GET_BLOCKS_CONVERT_UNWRITTEN 0x0100
c86d8db3
JK
709 /* Write zeros to newly created written extents */
710#define EXT4_GET_BLOCKS_ZERO 0x0200
711#define EXT4_GET_BLOCKS_CREATE_ZERO (EXT4_GET_BLOCKS_CREATE |\
712 EXT4_GET_BLOCKS_ZERO)
ee0876bc
JK
713 /* Caller will submit data before dropping transaction handle. This
714 * allows jbd2 to avoid submitting data before commit. */
715#define EXT4_GET_BLOCKS_IO_SUBMIT 0x0400
9558cf14
ZY
716 /* Caller is in the atomic contex, find extent if it has been cached */
717#define EXT4_GET_BLOCKS_CACHED_NOWAIT 0x0800
ac27a0ec 718
107a7bd3 719/*
7869a4a6 720 * The bit position of these flags must not overlap with any of the
ed8a1a76 721 * EXT4_GET_BLOCKS_*. They are used by ext4_find_extent(),
107a7bd3 722 * read_extent_tree_block(), ext4_split_extent_at(),
7869a4a6
TT
723 * ext4_ext_insert_extent(), and ext4_ext_create_new_leaf().
724 * EXT4_EX_NOCACHE is used to indicate that the we shouldn't be
725 * caching the extents when reading from the extent tree while a
726 * truncate or punch hole operation is in progress.
107a7bd3 727 */
d26e2c4d
TT
728#define EXT4_EX_NOCACHE 0x40000000
729#define EXT4_EX_FORCE_CACHE 0x20000000
73c384c0 730#define EXT4_EX_NOFAIL 0x10000000
107a7bd3 731
e6362609
TT
732/*
733 * Flags used by ext4_free_blocks
734 */
e7093f0d
PK
735#define EXT4_FREE_BLOCKS_METADATA 0x0001
736#define EXT4_FREE_BLOCKS_FORGET 0x0002
737#define EXT4_FREE_BLOCKS_VALIDATED 0x0004
738#define EXT4_FREE_BLOCKS_NO_QUOT_UPDATE 0x0008
84130193
TT
739#define EXT4_FREE_BLOCKS_NOFREE_FIRST_CLUSTER 0x0010
740#define EXT4_FREE_BLOCKS_NOFREE_LAST_CLUSTER 0x0020
9fe67149 741#define EXT4_FREE_BLOCKS_RERESERVE_CLUSTER 0x0040
e6362609 742
899ad0ce 743#if defined(__KERNEL__) && defined(CONFIG_COMPAT)
ac27a0ec
DK
744/*
745 * ioctl commands in 32 bit emulation
746 */
617ba13b
MC
747#define EXT4_IOC32_GETVERSION _IOR('f', 3, int)
748#define EXT4_IOC32_SETVERSION _IOW('f', 4, int)
749#define EXT4_IOC32_GETRSVSZ _IOR('f', 5, int)
750#define EXT4_IOC32_SETRSVSZ _IOW('f', 6, int)
751#define EXT4_IOC32_GROUP_EXTEND _IOW('f', 7, unsigned int)
4d92dc0f 752#define EXT4_IOC32_GROUP_ADD _IOW('f', 8, struct compat_ext4_new_group_input)
617ba13b
MC
753#define EXT4_IOC32_GETVERSION_OLD FS_IOC32_GETVERSION
754#define EXT4_IOC32_SETVERSION_OLD FS_IOC32_SETVERSION
899ad0ce 755#endif
ac27a0ec 756
b595076a 757/* Max physical block we can address w/o extents */
fb0a387d
ES
758#define EXT4_MAX_BLOCK_FILE_PHYS 0xFFFFFFFF
759
bcd8e91f 760/* Max logical block we can support */
9f44eda1 761#define EXT4_MAX_LOGICAL_BLOCK 0xFFFFFFFE
bcd8e91f 762
ac27a0ec
DK
763/*
764 * Structure of an inode on the disk
765 */
617ba13b 766struct ext4_inode {
ac27a0ec
DK
767 __le16 i_mode; /* File mode */
768 __le16 i_uid; /* Low 16 bits of Owner Uid */
a48380f7 769 __le32 i_size_lo; /* Size in bytes */
ac27a0ec 770 __le32 i_atime; /* Access time */
ef7f3835 771 __le32 i_ctime; /* Inode Change time */
ac27a0ec
DK
772 __le32 i_mtime; /* Modification time */
773 __le32 i_dtime; /* Deletion Time */
774 __le16 i_gid; /* Low 16 bits of Group Id */
775 __le16 i_links_count; /* Links count */
0fc1b451 776 __le32 i_blocks_lo; /* Blocks count */
ac27a0ec
DK
777 __le32 i_flags; /* File flags */
778 union {
779 struct {
25ec56b5 780 __le32 l_i_version;
ac27a0ec
DK
781 } linux1;
782 struct {
783 __u32 h_i_translator;
784 } hurd1;
785 struct {
786 __u32 m_i_reserved1;
787 } masix1;
788 } osd1; /* OS dependent 1 */
617ba13b 789 __le32 i_block[EXT4_N_BLOCKS];/* Pointers to blocks */
ac27a0ec 790 __le32 i_generation; /* File version (for NFS) */
7973c0c1 791 __le32 i_file_acl_lo; /* File ACL */
a48380f7 792 __le32 i_size_high;
f077d0d7 793 __le32 i_obso_faddr; /* Obsoleted fragment address */
ac27a0ec
DK
794 union {
795 struct {
0fc1b451 796 __le16 l_i_blocks_high; /* were l_i_reserved1 */
a1ddeb7e 797 __le16 l_i_file_acl_high;
8c55e204 798 __le16 l_i_uid_high; /* these 2 fields */
ac27a0ec 799 __le16 l_i_gid_high; /* were reserved2[0] */
e6153918
DW
800 __le16 l_i_checksum_lo;/* crc32c(uuid+inum+inode) LE */
801 __le16 l_i_reserved;
ac27a0ec
DK
802 } linux2;
803 struct {
f077d0d7 804 __le16 h_i_reserved1; /* Obsoleted fragment number/size which are removed in ext4 */
ac27a0ec
DK
805 __u16 h_i_mode_high;
806 __u16 h_i_uid_high;
807 __u16 h_i_gid_high;
808 __u32 h_i_author;
809 } hurd2;
810 struct {
f077d0d7 811 __le16 h_i_reserved1; /* Obsoleted fragment number/size which are removed in ext4 */
a1ddeb7e 812 __le16 m_i_file_acl_high;
ac27a0ec
DK
813 __u32 m_i_reserved2[2];
814 } masix2;
815 } osd2; /* OS dependent 2 */
816 __le16 i_extra_isize;
e6153918 817 __le16 i_checksum_hi; /* crc32c(uuid+inum+inode) BE */
ef7f3835
KS
818 __le32 i_ctime_extra; /* extra Change time (nsec << 2 | epoch) */
819 __le32 i_mtime_extra; /* extra Modification time(nsec << 2 | epoch) */
820 __le32 i_atime_extra; /* extra Access time (nsec << 2 | epoch) */
821 __le32 i_crtime; /* File Creation time */
822 __le32 i_crtime_extra; /* extra FileCreationtime (nsec << 2 | epoch) */
25ec56b5 823 __le32 i_version_hi; /* high 32 bits for 64-bit version */
8b4953e1 824 __le32 i_projid; /* Project ID */
ac27a0ec
DK
825};
826
ef7f3835
KS
827#define EXT4_EPOCH_BITS 2
828#define EXT4_EPOCH_MASK ((1 << EXT4_EPOCH_BITS) - 1)
829#define EXT4_NSEC_MASK (~0UL << EXT4_EPOCH_BITS)
830
831/*
832 * Extended fields will fit into an inode if the filesystem was formatted
833 * with large inodes (-I 256 or larger) and there are not currently any EAs
834 * consuming all of the available space. For new inodes we always reserve
835 * enough space for the kernel's known extended fields, but for inodes
836 * created with an old kernel this might not have been the case. None of
837 * the extended inode fields is critical for correct filesystem operation.
838 * This macro checks if a certain field fits in the inode. Note that
839 * inode-size = GOOD_OLD_INODE_SIZE + i_extra_isize
840 */
841#define EXT4_FITS_IN_INODE(ext4_inode, einode, field) \
842 ((offsetof(typeof(*ext4_inode), field) + \
843 sizeof((ext4_inode)->field)) \
844 <= (EXT4_GOOD_OLD_INODE_SIZE + \
845 (einode)->i_extra_isize)) \
846
a4dad1ae
DT
847/*
848 * We use an encoding that preserves the times for extra epoch "00":
849 *
850 * extra msb of adjust for signed
851 * epoch 32-bit 32-bit tv_sec to
852 * bits time decoded 64-bit tv_sec 64-bit tv_sec valid time range
853 * 0 0 1 -0x80000000..-0x00000001 0x000000000 1901-12-13..1969-12-31
854 * 0 0 0 0x000000000..0x07fffffff 0x000000000 1970-01-01..2038-01-19
855 * 0 1 1 0x080000000..0x0ffffffff 0x100000000 2038-01-19..2106-02-07
856 * 0 1 0 0x100000000..0x17fffffff 0x100000000 2106-02-07..2174-02-25
857 * 1 0 1 0x180000000..0x1ffffffff 0x200000000 2174-02-25..2242-03-16
858 * 1 0 0 0x200000000..0x27fffffff 0x200000000 2242-03-16..2310-04-04
859 * 1 1 1 0x280000000..0x2ffffffff 0x300000000 2310-04-04..2378-04-22
860 * 1 1 0 0x300000000..0x37fffffff 0x300000000 2378-04-22..2446-05-10
861 *
862 * Note that previous versions of the kernel on 64-bit systems would
863 * incorrectly use extra epoch bits 1,1 for dates between 1901 and
864 * 1970. e2fsck will correct this, assuming that it is run on the
865 * affected filesystem before 2242.
866 */
867
7b62b293 868static inline __le32 ext4_encode_extra_time(struct timespec64 *time)
ef7f3835 869{
7b62b293 870 u32 extra =((time->tv_sec - (s32)time->tv_sec) >> 32) & EXT4_EPOCH_MASK;
a4dad1ae 871 return cpu_to_le32(extra | (time->tv_nsec << EXT4_EPOCH_BITS));
ef7f3835
KS
872}
873
7b62b293
AB
874static inline void ext4_decode_extra_time(struct timespec64 *time,
875 __le32 extra)
ef7f3835 876{
cd2d9922 877 if (unlikely(extra & cpu_to_le32(EXT4_EPOCH_MASK)))
a4dad1ae 878 time->tv_sec += (u64)(le32_to_cpu(extra) & EXT4_EPOCH_MASK) << 32;
a4dad1ae 879 time->tv_nsec = (le32_to_cpu(extra) & EXT4_NSEC_MASK) >> EXT4_EPOCH_BITS;
ef7f3835
KS
880}
881
95582b00
DD
882#define EXT4_INODE_SET_XTIME(xtime, inode, raw_inode) \
883do { \
95582b00 884 if (EXT4_FITS_IN_INODE(raw_inode, EXT4_I(inode), xtime ## _extra)) {\
4881c497 885 (raw_inode)->xtime = cpu_to_le32((inode)->xtime.tv_sec); \
95582b00 886 (raw_inode)->xtime ## _extra = \
7b62b293 887 ext4_encode_extra_time(&(inode)->xtime); \
95582b00 888 } \
cba465b4 889 else \
4881c497 890 (raw_inode)->xtime = cpu_to_le32(clamp_t(int32_t, (inode)->xtime.tv_sec, S32_MIN, S32_MAX)); \
ef7f3835
KS
891} while (0)
892
893#define EXT4_EINODE_SET_XTIME(xtime, einode, raw_inode) \
894do { \
895 if (EXT4_FITS_IN_INODE(raw_inode, einode, xtime)) \
896 (raw_inode)->xtime = cpu_to_le32((einode)->xtime.tv_sec); \
897 if (EXT4_FITS_IN_INODE(raw_inode, einode, xtime ## _extra)) \
898 (raw_inode)->xtime ## _extra = \
899 ext4_encode_extra_time(&(einode)->xtime); \
900} while (0)
901
95582b00
DD
902#define EXT4_INODE_GET_XTIME(xtime, inode, raw_inode) \
903do { \
904 (inode)->xtime.tv_sec = (signed)le32_to_cpu((raw_inode)->xtime); \
905 if (EXT4_FITS_IN_INODE(raw_inode, EXT4_I(inode), xtime ## _extra)) { \
7b62b293 906 ext4_decode_extra_time(&(inode)->xtime, \
95582b00 907 raw_inode->xtime ## _extra); \
95582b00
DD
908 } \
909 else \
910 (inode)->xtime.tv_nsec = 0; \
ef7f3835
KS
911} while (0)
912
95582b00 913
ef7f3835
KS
914#define EXT4_EINODE_GET_XTIME(xtime, einode, raw_inode) \
915do { \
916 if (EXT4_FITS_IN_INODE(raw_inode, einode, xtime)) \
917 (einode)->xtime.tv_sec = \
918 (signed)le32_to_cpu((raw_inode)->xtime); \
19ea8060
TT
919 else \
920 (einode)->xtime.tv_sec = 0; \
ef7f3835
KS
921 if (EXT4_FITS_IN_INODE(raw_inode, einode, xtime ## _extra)) \
922 ext4_decode_extra_time(&(einode)->xtime, \
923 raw_inode->xtime ## _extra); \
af0b44a1
ES
924 else \
925 (einode)->xtime.tv_nsec = 0; \
ef7f3835
KS
926} while (0)
927
25ec56b5
JNC
928#define i_disk_version osd1.linux1.l_i_version
929
ac27a0ec
DK
930#if defined(__KERNEL__) || defined(__linux__)
931#define i_reserved1 osd1.linux1.l_i_reserved1
a1ddeb7e 932#define i_file_acl_high osd2.linux2.l_i_file_acl_high
0fc1b451 933#define i_blocks_high osd2.linux2.l_i_blocks_high
ac27a0ec
DK
934#define i_uid_low i_uid
935#define i_gid_low i_gid
936#define i_uid_high osd2.linux2.l_i_uid_high
937#define i_gid_high osd2.linux2.l_i_gid_high
e6153918 938#define i_checksum_lo osd2.linux2.l_i_checksum_lo
ac27a0ec
DK
939
940#elif defined(__GNU__)
941
942#define i_translator osd1.hurd1.h_i_translator
ac27a0ec
DK
943#define i_uid_high osd2.hurd2.h_i_uid_high
944#define i_gid_high osd2.hurd2.h_i_gid_high
945#define i_author osd2.hurd2.h_i_author
946
947#elif defined(__masix__)
948
949#define i_reserved1 osd1.masix1.m_i_reserved1
a1ddeb7e 950#define i_file_acl_high osd2.masix2.m_i_file_acl_high
ac27a0ec
DK
951#define i_reserved2 osd2.masix2.m_i_reserved2
952
953#endif /* defined(__KERNEL__) || defined(__linux__) */
954
c0677e6d 955#include "extents_status.h"
6866d7b3 956#include "fast_commit.h"
c0677e6d 957
daf647d2
TT
958/*
959 * Lock subclasses for i_data_sem in the ext4_inode_info structure.
960 *
961 * These are needed to avoid lockdep false positives when we need to
962 * allocate blocks to the quota inode during ext4_map_blocks(), while
963 * holding i_data_sem for a normal (non-quota) inode. Since we don't
964 * do quota tracking for the quota inode, this avoids deadlock (as
965 * well as infinite recursion, since it isn't turtles all the way
966 * down...)
967 *
968 * I_DATA_SEM_NORMAL - Used for most inodes
969 * I_DATA_SEM_OTHER - Used by move_inode.c for the second normal inode
970 * where the second inode has larger inode number
971 * than the first
972 * I_DATA_SEM_QUOTA - Used for quota inodes only
aff3bea9 973 * I_DATA_SEM_EA - Used for ea_inodes only
daf647d2
TT
974 */
975enum {
976 I_DATA_SEM_NORMAL = 0,
977 I_DATA_SEM_OTHER,
978 I_DATA_SEM_QUOTA,
aff3bea9 979 I_DATA_SEM_EA
daf647d2
TT
980};
981
982
d444c3c3
TT
983/*
984 * fourth extended file system inode data in memory
985 */
986struct ext4_inode_info {
987 __le32 i_data[15]; /* unconverted */
d444c3c3 988 __u32 i_dtime;
12e9b892 989 ext4_fsblk_t i_file_acl;
d444c3c3
TT
990
991 /*
992 * i_block_group is the number of the block group which contains
993 * this file's inode. Constant across the lifetime of the inode,
a627b0a7 994 * it is used for making block allocation decisions - we try to
d444c3c3
TT
995 * place a file's data blocks near its inode block, and new inodes
996 * near to their parent directory's inode.
997 */
998 ext4_group_t i_block_group;
8a2005d3 999 ext4_lblk_t i_dir_start_lookup;
353eb83c 1000#if (BITS_PER_LONG < 64)
19f5fb7a 1001 unsigned long i_state_flags; /* Dynamic state flags */
353eb83c 1002#endif
12e9b892 1003 unsigned long i_flags;
d444c3c3 1004
d444c3c3
TT
1005 /*
1006 * Extended attributes can be read independently of the main file
f340b3d9 1007 * data. Taking i_rwsem even when reading would cause contention
d444c3c3
TT
1008 * between readers of EAs and writers of regular file data, so
1009 * instead we synchronize on xattr_sem when reading or changing
1010 * EAs.
1011 */
1012 struct rw_semaphore xattr_sem;
d444c3c3 1013
02f310fc
JK
1014 /*
1015 * Inodes with EXT4_STATE_ORPHAN_FILE use i_orphan_idx. Otherwise
1016 * i_orphan is used.
1017 */
1018 union {
1019 struct list_head i_orphan; /* unlinked but open inodes */
1020 unsigned int i_orphan_idx; /* Index in orphan file */
1021 };
d444c3c3 1022
aa75f4d3
HS
1023 /* Fast commit related info */
1024
b3998b3b
RH
1025 /* For tracking dentry create updates */
1026 struct list_head i_fc_dilist;
aa75f4d3
HS
1027 struct list_head i_fc_list; /*
1028 * inodes that need fast commit
1029 * protected by sbi->s_fc_lock.
1030 */
1031
aa75f4d3
HS
1032 /* Start of lblk range that needs to be committed in this fast commit */
1033 ext4_lblk_t i_fc_lblk_start;
1034
1035 /* End of lblk range that needs to be committed in this fast commit */
1036 ext4_lblk_t i_fc_lblk_len;
1037
1038 /* Number of ongoing updates on this inode */
1039 atomic_t i_fc_updates;
1040
1041 /* Fast commit wait queue for this inode */
1042 wait_queue_head_t i_fc_wait;
1043
1044 /* Protect concurrent accesses on i_fc_lblk_start, i_fc_lblk_len */
1045 struct mutex i_fc_lock;
1046
d444c3c3
TT
1047 /*
1048 * i_disksize keeps track of what the inode size is ON DISK, not
1049 * in memory. During truncate, i_size is set to the new size by
1050 * the VFS prior to calling ext4_truncate(), but the filesystem won't
1051 * set i_disksize to 0 until the truncate is actually under way.
1052 *
1053 * The intent is that i_disksize always represents the blocks which
1054 * are used by this file. This allows recovery to restart truncate
1055 * on orphans if we crash during truncate. We actually write i_disksize
1056 * into the on-disk inode when writing inodes out, instead of i_size.
1057 *
1058 * The only time when i_disksize and i_size may be different is when
1059 * a truncate is in progress. The only things which change i_disksize
1060 * are ext4_get_block (growth) and ext4_truncate (shrinkth).
1061 */
1062 loff_t i_disksize;
1063
1064 /*
1065 * i_data_sem is for serialising ext4_truncate() against
1066 * ext4_getblock(). In the 2.4 ext2 design, great chunks of inode's
1067 * data tree are chopped off during truncate. We can't do that in
1068 * ext4 because whenever we perform intermediate commits during
1069 * truncate, the inode and all the metadata blocks *must* be in a
1070 * consistent state which allows truncation of the orphans to restart
1071 * during recovery. Hence we must fix the get_block-vs-truncate race
1072 * by other means, so we have i_data_sem.
1073 */
1074 struct rw_semaphore i_data_sem;
1075 struct inode vfs_inode;
8aefcd55 1076 struct jbd2_inode *jinode;
d444c3c3 1077
202ee5df
TT
1078 spinlock_t i_raw_lock; /* protects updates to the raw inode */
1079
d444c3c3
TT
1080 /*
1081 * File creation time. Its function is same as that of
7b62b293 1082 * struct timespec64 i_{a,c,m}time in the generic inode.
d444c3c3 1083 */
7b62b293 1084 struct timespec64 i_crtime;
d444c3c3
TT
1085
1086 /* mballoc */
27bc446e 1087 atomic_t i_prealloc_active;
38727786
OM
1088 struct rb_root i_prealloc_node;
1089 rwlock_t i_prealloc_lock;
d444c3c3 1090
c0677e6d
ZL
1091 /* extents status tree */
1092 struct ext4_es_tree i_es_tree;
1093 rwlock_t i_es_lock;
edaa53ca 1094 struct list_head i_es_list;
eb68d0e2 1095 unsigned int i_es_all_nr; /* protected by i_es_lock */
edaa53ca 1096 unsigned int i_es_shk_nr; /* protected by i_es_lock */
dd475925
JK
1097 ext4_lblk_t i_es_shrink_lblk; /* Offset where we start searching for
1098 extents to shrink. Protected by
1099 i_es_lock */
c0677e6d 1100
d444c3c3
TT
1101 /* ialloc */
1102 ext4_group_t i_last_alloc_group;
1103
1104 /* allocation reservation info for delalloc */
a627b0a7 1105 /* In case of bigalloc, this refer to clusters rather than blocks */
d444c3c3 1106 unsigned int i_reserved_data_blocks;
d444c3c3 1107
1dc0aa46
EW
1108 /* pending cluster reservations for bigalloc file systems */
1109 struct ext4_pending_tree i_pending_tree;
1110
d444c3c3
TT
1111 /* on-disk additional length */
1112 __u16 i_extra_isize;
1113
67cf5b09
TM
1114 /* Indicate the inline data space. */
1115 u16 i_inline_off;
1116 u16 i_inline_size;
1117
a9e7f447
DM
1118#ifdef CONFIG_QUOTA
1119 /* quota space reservation, managed internally by quota code */
1120 qsize_t i_reserved_quota;
1121#endif
8d5d02e6 1122
2e8fa54e 1123 /* Lock protecting lists below */
744692dc 1124 spinlock_t i_completed_io_lock;
2e8fa54e
JK
1125 /*
1126 * Completed IOs that need unwritten extents handling and have
1127 * transaction reserved
1128 */
1129 struct list_head i_rsv_conversion_list;
2e8fa54e 1130 struct work_struct i_rsv_conversion_work;
600be30a 1131 atomic_t i_unwritten; /* Nr. of inflight conversions pending */
8a2005d3
TT
1132
1133 spinlock_t i_block_reservation_lock;
b436b9be
JK
1134
1135 /*
1136 * Transactions that contain inode's metadata needed to complete
1137 * fsync and fdatasync, respectively.
1138 */
1139 tid_t i_sync_tid;
1140 tid_t i_datasync_tid;
814525f4 1141
96c7e0d9
JK
1142#ifdef CONFIG_QUOTA
1143 struct dquot *i_dquot[MAXQUOTAS];
1144#endif
1145
814525f4
DW
1146 /* Precomputed uuid+inum+igen checksum for seeding inode checksums */
1147 __u32 i_csum_seed;
b30ab0e0 1148
040cb378 1149 kprojid_t i_projid;
d444c3c3
TT
1150};
1151
ac27a0ec
DK
1152/*
1153 * File system states
1154 */
617ba13b
MC
1155#define EXT4_VALID_FS 0x0001 /* Unmounted cleanly */
1156#define EXT4_ERROR_FS 0x0002 /* Errors detected */
1157#define EXT4_ORPHAN_FS 0x0004 /* Orphans being recovered */
8016e29f 1158#define EXT4_FC_REPLAY 0x0020 /* Fast commit replay ongoing */
ac27a0ec 1159
469108ff
TT
1160/*
1161 * Misc. filesystem flags
1162 */
1163#define EXT2_FLAGS_SIGNED_HASH 0x0001 /* Signed dirhash in use */
1164#define EXT2_FLAGS_UNSIGNED_HASH 0x0002 /* Unsigned dirhash in use */
1165#define EXT2_FLAGS_TEST_FILESYS 0x0004 /* to test development code */
1166
ac27a0ec 1167/*
68911009 1168 * Mount flags set via mount options or defaults
ac27a0ec 1169 */
cdb7ee4c 1170#define EXT4_MOUNT_NO_MBCACHE 0x00001 /* Do not use mbcache */
617ba13b
MC
1171#define EXT4_MOUNT_GRPID 0x00004 /* Create files with directory's group */
1172#define EXT4_MOUNT_DEBUG 0x00008 /* Some debugging messages */
1173#define EXT4_MOUNT_ERRORS_CONT 0x00010 /* Continue on errors */
1174#define EXT4_MOUNT_ERRORS_RO 0x00020 /* Remount fs ro on errors */
1175#define EXT4_MOUNT_ERRORS_PANIC 0x00040 /* Panic on errors */
39ef17f1 1176#define EXT4_MOUNT_ERRORS_MASK 0x00070
617ba13b
MC
1177#define EXT4_MOUNT_MINIX_DF 0x00080 /* Mimics the Minix statfs */
1178#define EXT4_MOUNT_NOLOAD 0x00100 /* Don't use existing journal*/
923ae0ff 1179#ifdef CONFIG_FS_DAX
fc626fe3 1180#define EXT4_MOUNT_DAX_ALWAYS 0x00200 /* Direct Access */
923ae0ff 1181#else
fc626fe3 1182#define EXT4_MOUNT_DAX_ALWAYS 0
923ae0ff 1183#endif
617ba13b
MC
1184#define EXT4_MOUNT_DATA_FLAGS 0x00C00 /* Mode for data writes: */
1185#define EXT4_MOUNT_JOURNAL_DATA 0x00400 /* Write data to journal */
1186#define EXT4_MOUNT_ORDERED_DATA 0x00800 /* Flush data before commit */
1187#define EXT4_MOUNT_WRITEBACK_DATA 0x00C00 /* No data ordering */
1188#define EXT4_MOUNT_UPDATE_JOURNAL 0x01000 /* Update the journal format */
1189#define EXT4_MOUNT_NO_UID32 0x02000 /* Disable 32-bit UIDs */
1190#define EXT4_MOUNT_XATTR_USER 0x04000 /* Extended user attributes */
1191#define EXT4_MOUNT_POSIX_ACL 0x08000 /* POSIX Access Control Lists */
afd4672d 1192#define EXT4_MOUNT_NO_AUTO_DA_ALLOC 0x10000 /* No auto delalloc mapping */
617ba13b 1193#define EXT4_MOUNT_BARRIER 0x20000 /* Use block barriers */
49da9392
JK
1194#define EXT4_MOUNT_QUOTA 0x40000 /* Some quota option set */
1195#define EXT4_MOUNT_USRQUOTA 0x80000 /* "old" user quota,
1196 * enable enforcement for hidden
1197 * quota files */
1198#define EXT4_MOUNT_GRPQUOTA 0x100000 /* "old" group quota, enable
1199 * enforcement for hidden quota
1200 * files */
1201#define EXT4_MOUNT_PRJQUOTA 0x200000 /* Enable project quota
1202 * enforcement */
744692dc 1203#define EXT4_MOUNT_DIOREAD_NOLOCK 0x400000 /* Enable support for dio read nolocking */
d4da6c9c 1204#define EXT4_MOUNT_JOURNAL_CHECKSUM 0x800000 /* Journal checksums */
818d276c 1205#define EXT4_MOUNT_JOURNAL_ASYNC_COMMIT 0x1000000 /* Journal Async Commit */
327eaf73 1206#define EXT4_MOUNT_WARN_ON_ERROR 0x2000000 /* Trigger WARN_ON on error */
21175ca4 1207#define EXT4_MOUNT_NO_PREFETCH_BLOCK_BITMAPS 0x4000000
64769240 1208#define EXT4_MOUNT_DELALLOC 0x8000000 /* Delalloc support */
5bf5683a 1209#define EXT4_MOUNT_DATA_ERR_ABORT 0x10000000 /* Abort on file data write */
6fd058f7 1210#define EXT4_MOUNT_BLOCK_VALIDITY 0x20000000 /* Block validity checking */
5328e635 1211#define EXT4_MOUNT_DISCARD 0x40000000 /* Issue DISCARD requests */
bfff6873 1212#define EXT4_MOUNT_INIT_INODE_TABLE 0x80000000 /* Initialize uninitialized itables */
5bf5683a 1213
68911009
LC
1214/*
1215 * Mount flags set either automatically (could not be set by mount option)
1216 * based on per file system feature or property or in special cases such as
1217 * distinguishing between explicit mount option definition and default.
1218 */
56889787
TT
1219#define EXT4_MOUNT2_EXPLICIT_DELALLOC 0x00000001 /* User explicitly
1220 specified delalloc */
68911009
LC
1221#define EXT4_MOUNT2_STD_GROUP_SIZE 0x00000002 /* We have standard group
1222 size of blocksize * 8
1223 blocks */
ed3654eb
TT
1224#define EXT4_MOUNT2_HURD_COMPAT 0x00000004 /* Support HURD-castrated
1225 file systems */
1e381f60
DM
1226#define EXT4_MOUNT2_EXPLICIT_JOURNAL_CHECKSUM 0x00000008 /* User explicitly
1227 specified journal checksum */
1228
995a3ed6 1229#define EXT4_MOUNT2_JOURNAL_FAST_COMMIT 0x00000010 /* Journal fast commit */
a72b38ee
HS
1230#define EXT4_MOUNT2_DAX_NEVER 0x00000020 /* Do not allow Direct Access */
1231#define EXT4_MOUNT2_DAX_INODE 0x00000040 /* For printing options only */
196e402a
HS
1232#define EXT4_MOUNT2_MB_OPTIMIZE_SCAN 0x00000080 /* Optimize group
1233 * scanning in mballoc
1234 */
22b8d707 1235#define EXT4_MOUNT2_ABORT 0x00000100 /* Abort filesystem */
995a3ed6 1236
fd8c37ec
TT
1237#define clear_opt(sb, opt) EXT4_SB(sb)->s_mount_opt &= \
1238 ~EXT4_MOUNT_##opt
1239#define set_opt(sb, opt) EXT4_SB(sb)->s_mount_opt |= \
1240 EXT4_MOUNT_##opt
617ba13b
MC
1241#define test_opt(sb, opt) (EXT4_SB(sb)->s_mount_opt & \
1242 EXT4_MOUNT_##opt)
ac27a0ec 1243
a2595b8a
TT
1244#define clear_opt2(sb, opt) EXT4_SB(sb)->s_mount_opt2 &= \
1245 ~EXT4_MOUNT2_##opt
1246#define set_opt2(sb, opt) EXT4_SB(sb)->s_mount_opt2 |= \
1247 EXT4_MOUNT2_##opt
1248#define test_opt2(sb, opt) (EXT4_SB(sb)->s_mount_opt2 & \
1249 EXT4_MOUNT2_##opt)
1250
597d508c
AM
1251#define ext4_test_and_set_bit __test_and_set_bit_le
1252#define ext4_set_bit __set_bit_le
597d508c
AM
1253#define ext4_test_and_clear_bit __test_and_clear_bit_le
1254#define ext4_clear_bit __clear_bit_le
50e0168c 1255#define ext4_test_bit test_bit_le
50e0168c
AM
1256#define ext4_find_next_zero_bit find_next_zero_bit_le
1257#define ext4_find_next_bit find_next_bit_le
ac27a0ec 1258
123e3016 1259extern void mb_set_bits(void *bm, int cur, int len);
c3e94d1d 1260
ac27a0ec
DK
1261/*
1262 * Maximal mount counts between two filesystem checks
1263 */
617ba13b
MC
1264#define EXT4_DFL_MAX_MNT_COUNT 20 /* Allow 20 mounts */
1265#define EXT4_DFL_CHECKINTERVAL 0 /* Don't use interval check */
ac27a0ec
DK
1266
1267/*
1268 * Behaviour when detecting errors
1269 */
617ba13b
MC
1270#define EXT4_ERRORS_CONTINUE 1 /* Continue execution */
1271#define EXT4_ERRORS_RO 2 /* Remount fs read-only */
1272#define EXT4_ERRORS_PANIC 3 /* Panic */
1273#define EXT4_ERRORS_DEFAULT EXT4_ERRORS_CONTINUE
ac27a0ec 1274
e6153918
DW
1275/* Metadata checksum algorithm codes */
1276#define EXT4_CRC32C_CHKSUM 1
1277
bbc605cd
LC
1278#define EXT4_LABEL_MAX 16
1279
ac27a0ec
DK
1280/*
1281 * Structure of the super block
1282 */
617ba13b 1283struct ext4_super_block {
ac27a0ec 1284/*00*/ __le32 s_inodes_count; /* Inodes count */
6bc9feff 1285 __le32 s_blocks_count_lo; /* Blocks count */
308ba3ec
AK
1286 __le32 s_r_blocks_count_lo; /* Reserved blocks count */
1287 __le32 s_free_blocks_count_lo; /* Free blocks count */
ac27a0ec
DK
1288/*10*/ __le32 s_free_inodes_count; /* Free inodes count */
1289 __le32 s_first_data_block; /* First Data Block */
1290 __le32 s_log_block_size; /* Block size */
281b5995 1291 __le32 s_log_cluster_size; /* Allocation cluster size */
ac27a0ec 1292/*20*/ __le32 s_blocks_per_group; /* # Blocks per group */
281b5995 1293 __le32 s_clusters_per_group; /* # Clusters per group */
ac27a0ec
DK
1294 __le32 s_inodes_per_group; /* # Inodes per group */
1295 __le32 s_mtime; /* Mount time */
1296/*30*/ __le32 s_wtime; /* Write time */
1297 __le16 s_mnt_count; /* Mount count */
1298 __le16 s_max_mnt_count; /* Maximal mount count */
1299 __le16 s_magic; /* Magic signature */
1300 __le16 s_state; /* File system state */
1301 __le16 s_errors; /* Behaviour when detecting errors */
1302 __le16 s_minor_rev_level; /* minor revision level */
1303/*40*/ __le32 s_lastcheck; /* time of last check */
1304 __le32 s_checkinterval; /* max. time between checks */
1305 __le32 s_creator_os; /* OS */
1306 __le32 s_rev_level; /* Revision level */
1307/*50*/ __le16 s_def_resuid; /* Default uid for reserved blocks */
1308 __le16 s_def_resgid; /* Default gid for reserved blocks */
1309 /*
617ba13b 1310 * These fields are for EXT4_DYNAMIC_REV superblocks only.
ac27a0ec
DK
1311 *
1312 * Note: the difference between the compatible feature set and
1313 * the incompatible feature set is that if there is a bit set
1314 * in the incompatible feature set that the kernel doesn't
1315 * know about, it should refuse to mount the filesystem.
1316 *
1317 * e2fsck's requirements are more strict; if it doesn't know
1318 * about a feature in either the compatible or incompatible
1319 * feature set, it must abort and not try to meddle with
1320 * things it doesn't understand...
1321 */
1322 __le32 s_first_ino; /* First non-reserved inode */
0d1ee42f 1323 __le16 s_inode_size; /* size of inode structure */
ac27a0ec
DK
1324 __le16 s_block_group_nr; /* block group # of this superblock */
1325 __le32 s_feature_compat; /* compatible feature set */
1326/*60*/ __le32 s_feature_incompat; /* incompatible feature set */
1327 __le32 s_feature_ro_compat; /* readonly-compatible feature set */
1328/*68*/ __u8 s_uuid[16]; /* 128-bit uuid for volume */
bbc605cd 1329/*78*/ char s_volume_name[EXT4_LABEL_MAX]; /* volume name */
072ebb3b 1330/*88*/ char s_last_mounted[64] __nonstring; /* directory where last mounted */
ac27a0ec
DK
1331/*C8*/ __le32 s_algorithm_usage_bitmap; /* For compression */
1332 /*
1333 * Performance hints. Directory preallocation should only
617ba13b 1334 * happen if the EXT4_FEATURE_COMPAT_DIR_PREALLOC flag is on.
ac27a0ec
DK
1335 */
1336 __u8 s_prealloc_blocks; /* Nr of blocks to try to preallocate*/
1337 __u8 s_prealloc_dir_blocks; /* Nr to preallocate for dirs */
1338 __le16 s_reserved_gdt_blocks; /* Per group desc for online growth */
1339 /*
617ba13b 1340 * Journaling support valid if EXT4_FEATURE_COMPAT_HAS_JOURNAL set.
ac27a0ec
DK
1341 */
1342/*D0*/ __u8 s_journal_uuid[16]; /* uuid of journal superblock */
1343/*E0*/ __le32 s_journal_inum; /* inode number of journal file */
1344 __le32 s_journal_dev; /* device number of journal file */
1345 __le32 s_last_orphan; /* start of list of inodes to delete */
1346 __le32 s_hash_seed[4]; /* HTREE hash seed */
1347 __u8 s_def_hash_version; /* Default hash version to use */
89eeddf0 1348 __u8 s_jnl_backup_type;
0d1ee42f 1349 __le16 s_desc_size; /* size of group descriptor */
bd81d8ee 1350/*100*/ __le32 s_default_mount_opts;
ac27a0ec 1351 __le32 s_first_meta_bg; /* First metablock block group */
bd81d8ee
LV
1352 __le32 s_mkfs_time; /* When the filesystem was created */
1353 __le32 s_jnl_blocks[17]; /* Backup of the journal inode */
7fc1f5c2 1354 /* 64bit support valid if EXT4_FEATURE_INCOMPAT_64BIT */
bd81d8ee
LV
1355/*150*/ __le32 s_blocks_count_hi; /* Blocks count */
1356 __le32 s_r_blocks_count_hi; /* Reserved blocks count */
1357 __le32 s_free_blocks_count_hi; /* Free blocks count */
c1bddad9
AK
1358 __le16 s_min_extra_isize; /* All inodes have at least # bytes */
1359 __le16 s_want_extra_isize; /* New inodes should reserve # bytes */
1360 __le32 s_flags; /* Miscellaneous flags */
1361 __le16 s_raid_stride; /* RAID stride */
c5e06d10 1362 __le16 s_mmp_update_interval; /* # seconds to wait in MMP checking */
c1bddad9
AK
1363 __le64 s_mmp_block; /* Block for multi-mount protection */
1364 __le32 s_raid_stripe_width; /* blocks on all data disks (N*stride)*/
772cb7c8 1365 __u8 s_log_groups_per_flex; /* FLEX_BG group size */
e6153918 1366 __u8 s_checksum_type; /* metadata checksum algorithm used */
f542fbe8
TT
1367 __u8 s_encryption_level; /* versioning level for encryption */
1368 __u8 s_reserved_pad; /* Padding to next 32bits */
afc32f7e 1369 __le64 s_kbytes_written; /* nr of lifetime kilobytes written */
4a9cdec7
TT
1370 __le32 s_snapshot_inum; /* Inode number of active snapshot */
1371 __le32 s_snapshot_id; /* sequential ID of active snapshot */
1372 __le64 s_snapshot_r_blocks_count; /* reserved blocks for active
1373 snapshot's future use */
1374 __le32 s_snapshot_list; /* inode number of the head of the
1375 on-disk snapshot list */
1c13d5c0
TT
1376#define EXT4_S_ERR_START offsetof(struct ext4_super_block, s_error_count)
1377 __le32 s_error_count; /* number of fs errors */
1378 __le32 s_first_error_time; /* first time an error happened */
1379 __le32 s_first_error_ino; /* inode involved in first error */
1380 __le64 s_first_error_block; /* block involved of first error */
072ebb3b 1381 __u8 s_first_error_func[32] __nonstring; /* function where the error happened */
1c13d5c0
TT
1382 __le32 s_first_error_line; /* line number where error happened */
1383 __le32 s_last_error_time; /* most recent time of an error */
1384 __le32 s_last_error_ino; /* inode involved in last error */
1385 __le32 s_last_error_line; /* line number where error happened */
1386 __le64 s_last_error_block; /* block involved of last error */
072ebb3b 1387 __u8 s_last_error_func[32] __nonstring; /* function where the error happened */
8b67f04a
TT
1388#define EXT4_S_ERR_END offsetof(struct ext4_super_block, s_mount_opts)
1389 __u8 s_mount_opts[64];
281b5995
TT
1390 __le32 s_usr_quota_inum; /* inode for tracking user quota */
1391 __le32 s_grp_quota_inum; /* inode for tracking group quota */
1392 __le32 s_overhead_clusters; /* overhead blocks/clusters in fs */
1beeef1b 1393 __le32 s_backup_bgs[2]; /* groups with sparse_super2 SBs */
3edc18d8 1394 __u8 s_encrypt_algos[4]; /* Encryption algorithms in use */
f542fbe8
TT
1395 __u8 s_encrypt_pw_salt[16]; /* Salt used for string2key algorithm */
1396 __le32 s_lpf_ino; /* Location of the lost+found inode */
8b4953e1 1397 __le32 s_prj_quota_inum; /* inode for tracking project quota */
8c81bd8f 1398 __le32 s_checksum_seed; /* crc32c(uuid) if csum_seed set */
6a0678a7
AB
1399 __u8 s_wtime_hi;
1400 __u8 s_mtime_hi;
1401 __u8 s_mkfs_time_hi;
1402 __u8 s_lastcheck_hi;
1403 __u8 s_first_error_time_hi;
1404 __u8 s_last_error_time_hi;
878520ac
TT
1405 __u8 s_first_error_errcode;
1406 __u8 s_last_error_errcode;
c83ad55e
GKB
1407 __le16 s_encoding; /* Filename charset encoding */
1408 __le16 s_encoding_flags; /* Filename charset encoding flags */
02f310fc
JK
1409 __le32 s_orphan_file_inum; /* Inode for tracking orphan inodes */
1410 __le32 s_reserved[94]; /* Padding to the end of the block */
e6153918 1411 __le32 s_checksum; /* crc32c(superblock) */
ac27a0ec
DK
1412};
1413
1c13d5c0
TT
1414#define EXT4_S_ERR_LEN (EXT4_S_ERR_END - EXT4_S_ERR_START)
1415
ac27a0ec 1416#ifdef __KERNEL__
bc0b0d6d 1417
a2d4a646 1418/* Number of quota types we support */
689c958c 1419#define EXT4_MAXQUOTAS 3
a2d4a646 1420
c83ad55e
GKB
1421#define EXT4_ENC_UTF8_12_1 1
1422
188c299e
JK
1423/* Types of ext4 journal triggers */
1424enum ext4_journal_trigger_type {
02f310fc 1425 EXT4_JTR_ORPHAN_FILE,
188c299e
JK
1426 EXT4_JTR_NONE /* This must be the last entry for indexing to work! */
1427};
1428
1429#define EXT4_JOURNAL_TRIGGER_COUNT EXT4_JTR_NONE
1430
1431struct ext4_journal_trigger {
1432 struct jbd2_buffer_trigger_type tr_triggers;
1433 struct super_block *sb;
1434};
1435
1436static inline struct ext4_journal_trigger *EXT4_TRIGGER(
1437 struct jbd2_buffer_trigger_type *trigger)
1438{
1439 return container_of(trigger, struct ext4_journal_trigger, tr_triggers);
1440}
1441
02f310fc
JK
1442#define EXT4_ORPHAN_BLOCK_MAGIC 0x0b10ca04
1443
1444/* Structure at the tail of orphan block */
1445struct ext4_orphan_block_tail {
1446 __le32 ob_magic;
1447 __le32 ob_checksum;
1448};
1449
1450static inline int ext4_inodes_per_orphan_block(struct super_block *sb)
1451{
1452 return (sb->s_blocksize - sizeof(struct ext4_orphan_block_tail)) /
1453 sizeof(u32);
1454}
1455
1456struct ext4_orphan_block {
4a79a98c 1457 atomic_t ob_free_entries; /* Number of free orphan entries in block */
02f310fc
JK
1458 struct buffer_head *ob_bh; /* Buffer for orphan block */
1459};
1460
1461/*
1462 * Info about orphan file.
1463 */
1464struct ext4_orphan_info {
02f310fc
JK
1465 int of_blocks; /* Number of orphan blocks in a file */
1466 __u32 of_csum_seed; /* Checksum seed for orphan file */
1467 struct ext4_orphan_block *of_binfo; /* Array with info about orphan
1468 * file blocks */
1469};
1470
ca0faba0
TT
1471/*
1472 * fourth extended-fs super-block data in memory
1473 */
1474struct ext4_sb_info {
1475 unsigned long s_desc_size; /* Size of a group descriptor in bytes */
1476 unsigned long s_inodes_per_block;/* Number of inodes per block */
1477 unsigned long s_blocks_per_group;/* Number of blocks in a group */
281b5995 1478 unsigned long s_clusters_per_group; /* Number of clusters in a group */
ca0faba0
TT
1479 unsigned long s_inodes_per_group;/* Number of inodes in a group */
1480 unsigned long s_itb_per_group; /* Number of inode table blocks per group */
1481 unsigned long s_gdb_count; /* Number of group descriptor blocks */
1482 unsigned long s_desc_per_block; /* Number of group descriptors per block */
1483 ext4_group_t s_groups_count; /* Number of groups in the fs */
fb0a387d 1484 ext4_group_t s_blockfile_groups;/* Groups acceptable for non-extent files */
952fc18e 1485 unsigned long s_overhead; /* # of fs overhead clusters */
281b5995
TT
1486 unsigned int s_cluster_ratio; /* Number of blocks per cluster */
1487 unsigned int s_cluster_bits; /* log2 of s_cluster_ratio */
ca0faba0
TT
1488 loff_t s_bitmap_maxbytes; /* max bytes for bitmap files */
1489 struct buffer_head * s_sbh; /* Buffer containing the super block */
1490 struct ext4_super_block *s_es; /* Pointer to the super block in the buffer */
1d0c3924 1491 struct buffer_head * __rcu *s_group_desc;
7f4520cc 1492 unsigned int s_mount_opt;
a2595b8a 1493 unsigned int s_mount_opt2;
9b5f6c9b 1494 unsigned long s_mount_flags;
5a916be1 1495 unsigned int s_def_mount_opt;
e3645d72 1496 unsigned int s_def_mount_opt2;
ca0faba0 1497 ext4_fsblk_t s_sb_block;
27dd4385 1498 atomic64_t s_resv_clusters;
08cefc7a
EB
1499 kuid_t s_resuid;
1500 kgid_t s_resgid;
ca0faba0
TT
1501 unsigned short s_mount_state;
1502 unsigned short s_pad;
1503 int s_addr_per_block_bits;
1504 int s_desc_per_block_bits;
1505 int s_inode_size;
1506 int s_first_ino;
1507 unsigned int s_inode_readahead_blks;
11013911 1508 unsigned int s_inode_goal;
ca0faba0
TT
1509 u32 s_hash_seed[4];
1510 int s_def_hash_version;
6d2424a8 1511 int s_hash_unsigned; /* 3 if hash should be unsigned, 0 if not */
57042651 1512 struct percpu_counter s_freeclusters_counter;
ca0faba0
TT
1513 struct percpu_counter s_freeinodes_counter;
1514 struct percpu_counter s_dirs_counter;
57042651 1515 struct percpu_counter s_dirtyclusters_counter;
efc61345 1516 struct percpu_counter s_sra_exceeded_retry_limit;
ca0faba0
TT
1517 struct blockgroup_lock *s_blockgroup_lock;
1518 struct proc_dir_entry *s_proc;
1519 struct kobject s_kobj;
1520 struct completion s_kobj_unregister;
2c0544b2 1521 struct super_block *s_sb;
618f0031 1522 struct buffer_head *s_mmp_bh;
ca0faba0
TT
1523
1524 /* Journaling */
ca0faba0 1525 struct journal_s *s_journal;
9549a168 1526 unsigned long s_ext4_flags; /* Ext4 superblock flags */
02f310fc
JK
1527 struct mutex s_orphan_lock; /* Protects on disk list changes */
1528 struct list_head s_orphan; /* List of orphaned inodes in on disk
1529 list */
1530 struct ext4_orphan_info s_orphan_info;
ca0faba0
TT
1531 unsigned long s_commit_interval;
1532 u32 s_max_batch_time;
1533 u32 s_min_batch_time;
ee7ed3aa 1534 struct block_device *s_journal_bdev;
ca0faba0 1535#ifdef CONFIG_QUOTA
33458eab
TT
1536 /* Names of quota files with journalled quota */
1537 char __rcu *s_qf_names[EXT4_MAXQUOTAS];
ca0faba0
TT
1538 int s_jquota_fmt; /* Format of quota to use */
1539#endif
1540 unsigned int s_want_extra_isize; /* New inodes should reserve # bytes */
dd0db94f 1541 struct ext4_system_blocks __rcu *s_system_blks;
ca0faba0
TT
1542
1543#ifdef EXTENTS_STATS
1544 /* ext4 extents stats */
1545 unsigned long s_ext_min;
1546 unsigned long s_ext_max;
1547 unsigned long s_depth_max;
1548 spinlock_t s_ext_stats_lock;
1549 unsigned long s_ext_blocks;
1550 unsigned long s_ext_extents;
1551#endif
1552
1553 /* for buddy allocator */
df3da4ea 1554 struct ext4_group_info ** __rcu *s_group_info;
ca0faba0 1555 struct inode *s_buddy_cache;
ca0faba0 1556 spinlock_t s_md_lock;
ca0faba0
TT
1557 unsigned short *s_mb_offsets;
1558 unsigned int *s_mb_maxs;
28623c2f 1559 unsigned int s_group_info_size;
d08854f5 1560 unsigned int s_mb_free_pending;
a0154344
DJ
1561 struct list_head s_freed_data_list; /* List of blocks to be freed
1562 after commit completed */
55cdd0af
WJ
1563 struct list_head s_discard_list;
1564 struct work_struct s_discard_work;
5036ab8d 1565 atomic_t s_retry_alloc_pending;
83e80a6e
JK
1566 struct list_head *s_mb_avg_fragment_size;
1567 rwlock_t *s_mb_avg_fragment_size_locks;
196e402a
HS
1568 struct list_head *s_mb_largest_free_orders;
1569 rwlock_t *s_mb_largest_free_orders_locks;
ca0faba0
TT
1570
1571 /* tunables */
1572 unsigned long s_stripe;
196e402a 1573 unsigned int s_mb_max_linear_groups;
ca0faba0
TT
1574 unsigned int s_mb_stream_request;
1575 unsigned int s_mb_max_to_scan;
1576 unsigned int s_mb_min_to_scan;
1577 unsigned int s_mb_stats;
1578 unsigned int s_mb_order2_reqs;
1579 unsigned int s_mb_group_prealloc;
df981d03 1580 unsigned int s_max_dir_size_kb;
ca0faba0
TT
1581 /* where last allocation was done - for stream allocation */
1582 unsigned long s_mb_last_group;
1583 unsigned long s_mb_last_start;
cfd73237
AZ
1584 unsigned int s_mb_prefetch;
1585 unsigned int s_mb_prefetch_limit;
f52f3d2b 1586 unsigned int s_mb_best_avail_max_trim_order;
ca0faba0 1587
ca0faba0 1588 /* stats for buddy allocator */
ca0faba0
TT
1589 atomic_t s_bal_reqs; /* number of reqs with len > 1 */
1590 atomic_t s_bal_success; /* we found long enough chunks */
1591 atomic_t s_bal_allocated; /* in blocks */
1592 atomic_t s_bal_ex_scanned; /* total extents scanned */
fdd9a009 1593 atomic_t s_bal_cX_ex_scanned[EXT4_MB_NUM_CRS]; /* total extents scanned */
a6c75eaf 1594 atomic_t s_bal_groups_scanned; /* number of groups scanned */
ca0faba0 1595 atomic_t s_bal_goals; /* goal hits */
3ef5d263 1596 atomic_t s_bal_len_goals; /* len goal hits */
ca0faba0
TT
1597 atomic_t s_bal_breaks; /* too long searches */
1598 atomic_t s_bal_2orders; /* 2^order hits */
f52f3d2b
OM
1599 atomic_t s_bal_p2_aligned_bad_suggestions;
1600 atomic_t s_bal_goal_fast_bad_suggestions;
1601 atomic_t s_bal_best_avail_bad_suggestions;
4eb7a4a1
OM
1602 atomic64_t s_bal_cX_groups_considered[EXT4_MB_NUM_CRS];
1603 atomic64_t s_bal_cX_hits[EXT4_MB_NUM_CRS];
1604 atomic64_t s_bal_cX_failed[EXT4_MB_NUM_CRS]; /* cX loop didn't find blocks */
67d25186
HS
1605 atomic_t s_mb_buddies_generated; /* number of buddies generated */
1606 atomic64_t s_mb_generation_time;
ca0faba0
TT
1607 atomic_t s_mb_lost_chunks;
1608 atomic_t s_mb_preallocated;
1609 atomic_t s_mb_discarded;
50797481 1610 atomic_t s_lock_busy;
ca0faba0
TT
1611
1612 /* locality groups */
003cb608 1613 struct ext4_locality_group __percpu *s_locality_groups;
ca0faba0
TT
1614
1615 /* for write statistics */
1616 unsigned long s_sectors_written_start;
1617 u64 s_kbytes_written;
1618
67a5da56
ZL
1619 /* the size of zero-out chunk */
1620 unsigned int s_extent_max_zeroout_kb;
1621
ca0faba0 1622 unsigned int s_log_groups_per_flex;
7c990728 1623 struct flex_groups * __rcu *s_flex_groups;
117fff10 1624 ext4_group_t s_flex_groups_allocated;
4c0425ff 1625
2e8fa54e
JK
1626 /* workqueue for reserved extent conversions (buffered io) */
1627 struct workqueue_struct *rsv_conversion_wq;
66e61a9e
TT
1628
1629 /* timer for periodic error stats printing */
1630 struct timer_list s_err_report;
bfff6873
LC
1631
1632 /* Lazy inode table initialization info */
1633 struct ext4_li_request *s_li_request;
1634 /* Wait multiplier for lazy initialization thread */
1635 unsigned int s_li_wait_mult;
c5e06d10
JL
1636
1637 /* Kernel thread for multiple mount protection */
1638 struct task_struct *s_mmp_tsk;
3d56b8d2
TM
1639
1640 /* record the last minlen when FITRIM is called. */
2327fb2e 1641 unsigned long s_last_trim_minblks;
0441984a
DW
1642
1643 /* Reference to checksum algorithm driver via cryptoapi */
1644 struct crypto_shash *s_chksum_driver;
a9c47317
DW
1645
1646 /* Precomputed FS UUID checksum for seeding other checksums */
1647 __u32 s_csum_seed;
74cd15cd
ZL
1648
1649 /* Reclaim extents from extent status tree */
1650 struct shrinker s_es_shrinker;
dd475925 1651 struct list_head s_es_list; /* List of inodes with reclaimable extents */
edaa53ca 1652 long s_es_nr_inode;
eb68d0e2 1653 struct ext4_es_stats s_es_stats;
47387409 1654 struct mb_cache *s_ea_block_cache;
dec214d0 1655 struct mb_cache *s_ea_inode_cache;
edaa53ca 1656 spinlock_t s_es_lock ____cacheline_aligned_in_smp;
efbed4dc 1657
188c299e
JK
1658 /* Journal triggers for checksum computation */
1659 struct ext4_journal_trigger s_journal_triggers[EXT4_JOURNAL_TRIGGER_COUNT];
1660
efbed4dc
TT
1661 /* Ratelimit ext4 messages. */
1662 struct ratelimit_state s_err_ratelimit_state;
1663 struct ratelimit_state s_warning_ratelimit_state;
1664 struct ratelimit_state s_msg_ratelimit_state;
1cf006ed
DM
1665 atomic_t s_warning_count;
1666 atomic_t s_msg_count;
c8585c6f 1667
ac4acb1f
EB
1668 /* Encryption policy for '-o test_dummy_encryption' */
1669 struct fscrypt_dummy_policy s_dummy_enc_policy;
ed318a6c 1670
cb85f4d2
EB
1671 /*
1672 * Barrier between writepages ops and changing any inode's JOURNAL_DATA
1673 * or EXTENTS flag.
1674 */
bbd55937 1675 struct percpu_rw_semaphore s_writepages_rwsem;
5e405595 1676 struct dax_device *s_daxdev;
cd913c76 1677 u64 s_dax_part_off;
46f870d6
TT
1678#ifdef CONFIG_EXT4_DEBUG
1679 unsigned long s_simulate_fail;
1680#endif
bc71726c 1681 /* Record the errseq of the backing block device */
1682 errseq_t s_bdev_wb_err;
1683 spinlock_t s_bdev_wb_lock;
aa75f4d3 1684
c92dc856
JK
1685 /* Information about errors that happened during this mount */
1686 spinlock_t s_error_lock;
1687 int s_add_error_count;
1688 int s_first_error_code;
1689 __u32 s_first_error_line;
1690 __u32 s_first_error_ino;
1691 __u64 s_first_error_block;
1692 const char *s_first_error_func;
1693 time64_t s_first_error_time;
1694 int s_last_error_code;
1695 __u32 s_last_error_line;
1696 __u32 s_last_error_ino;
1697 __u64 s_last_error_block;
1698 const char *s_last_error_func;
1699 time64_t s_last_error_time;
1700 /*
bb15cea2
TT
1701 * If we are in a context where we cannot update the on-disk
1702 * superblock, we queue the work here. This is used to update
1703 * the error information in the superblock, and for periodic
1704 * updates of the superblock called from the commit callback
1705 * function.
c92dc856 1706 */
bb15cea2 1707 struct work_struct s_sb_upd_work;
c92dc856 1708
7bbbe241 1709 /* Ext4 fast commit sub transaction ID */
aa75f4d3 1710 atomic_t s_fc_subtid;
7bbbe241 1711
aa75f4d3
HS
1712 /*
1713 * After commit starts, the main queue gets locked, and the further
1714 * updates get added in the staging queue.
1715 */
1716#define FC_Q_MAIN 0
1717#define FC_Q_STAGING 1
1718 struct list_head s_fc_q[2]; /* Inodes staged for fast commit
1719 * that have data changes in them.
1720 */
1721 struct list_head s_fc_dentry_q[2]; /* directory entry updates */
1722 unsigned int s_fc_bytes;
1723 /*
1724 * Main fast commit lock. This lock protects accesses to the
1725 * following fields:
1726 * ei->i_fc_list, s_fc_dentry_q, s_fc_q, s_fc_bytes, s_fc_bh.
1727 */
1728 spinlock_t s_fc_lock;
1729 struct buffer_head *s_fc_bh;
1730 struct ext4_fc_stats s_fc_stats;
e85c81ba 1731 tid_t s_fc_ineligible_tid;
8016e29f
HS
1732#ifdef CONFIG_EXT4_DEBUG
1733 int s_fc_debug_max_replay;
1734#endif
1735 struct ext4_fc_replay_state s_fc_replay_state;
ca0faba0
TT
1736};
1737
af5bc92d 1738static inline struct ext4_sb_info *EXT4_SB(struct super_block *sb)
ac27a0ec
DK
1739{
1740 return sb->s_fs_info;
1741}
617ba13b 1742static inline struct ext4_inode_info *EXT4_I(struct inode *inode)
ac27a0ec 1743{
617ba13b 1744 return container_of(inode, struct ext4_inode_info, vfs_inode);
ac27a0ec
DK
1745}
1746
00d873c1
JK
1747static inline int ext4_writepages_down_read(struct super_block *sb)
1748{
1749 percpu_down_read(&EXT4_SB(sb)->s_writepages_rwsem);
1750 return memalloc_nofs_save();
1751}
1752
1753static inline void ext4_writepages_up_read(struct super_block *sb, int ctx)
1754{
1755 memalloc_nofs_restore(ctx);
1756 percpu_up_read(&EXT4_SB(sb)->s_writepages_rwsem);
1757}
1758
1759static inline int ext4_writepages_down_write(struct super_block *sb)
1760{
1761 percpu_down_write(&EXT4_SB(sb)->s_writepages_rwsem);
1762 return memalloc_nofs_save();
1763}
1764
1765static inline void ext4_writepages_up_write(struct super_block *sb, int ctx)
1766{
1767 memalloc_nofs_restore(ctx);
1768 percpu_up_write(&EXT4_SB(sb)->s_writepages_rwsem);
1769}
1770
617ba13b 1771static inline int ext4_valid_inum(struct super_block *sb, unsigned long ino)
ac27a0ec 1772{
617ba13b 1773 return ino == EXT4_ROOT_INO ||
617ba13b
MC
1774 (ino >= EXT4_FIRST_INO(sb) &&
1775 ino <= le32_to_cpu(EXT4_SB(sb)->s_es->s_inodes_count));
ac27a0ec 1776}
19f5fb7a 1777
1d0c3924
TT
1778/*
1779 * Returns: sbi->field[index]
1780 * Used to access an array element from the following sbi fields which require
1781 * rcu protection to avoid dereferencing an invalid pointer due to reassignment
1782 * - s_group_desc
1783 * - s_group_info
1784 * - s_flex_group
1785 */
1786#define sbi_array_rcu_deref(sbi, field, index) \
1787({ \
1788 typeof(*((sbi)->field)) _v; \
1789 rcu_read_lock(); \
1790 _v = ((typeof(_v)*)rcu_dereference((sbi)->field))[index]; \
1791 rcu_read_unlock(); \
1792 _v; \
1793})
1794
9b5f6c9b
HS
1795/*
1796 * run-time mount flags
1797 */
1798enum {
1799 EXT4_MF_MNTDIR_SAMPLED,
bdc8a53a 1800 EXT4_MF_FC_INELIGIBLE /* Fast commit ineligible */
9b5f6c9b
HS
1801};
1802
1803static inline void ext4_set_mount_flag(struct super_block *sb, int bit)
1804{
1805 set_bit(bit, &EXT4_SB(sb)->s_mount_flags);
1806}
1807
1808static inline void ext4_clear_mount_flag(struct super_block *sb, int bit)
1809{
1810 clear_bit(bit, &EXT4_SB(sb)->s_mount_flags);
1811}
1812
1813static inline int ext4_test_mount_flag(struct super_block *sb, int bit)
1814{
1815 return test_bit(bit, &EXT4_SB(sb)->s_mount_flags);
1816}
1817
1818
46f870d6
TT
1819/*
1820 * Simulate_fail codes
1821 */
1822#define EXT4_SIM_BBITMAP_EIO 1
1823#define EXT4_SIM_BBITMAP_CRC 2
1824#define EXT4_SIM_IBITMAP_EIO 3
1825#define EXT4_SIM_IBITMAP_CRC 4
1826#define EXT4_SIM_INODE_EIO 5
1827#define EXT4_SIM_INODE_CRC 6
1828#define EXT4_SIM_DIRBLOCK_EIO 7
1829#define EXT4_SIM_DIRBLOCK_CRC 8
1830
1831static inline bool ext4_simulate_fail(struct super_block *sb,
1832 unsigned long code)
1833{
1834#ifdef CONFIG_EXT4_DEBUG
1835 struct ext4_sb_info *sbi = EXT4_SB(sb);
1836
1837 if (unlikely(sbi->s_simulate_fail == code)) {
1838 sbi->s_simulate_fail = 0;
1839 return true;
1840 }
1841#endif
1842 return false;
1843}
1844
1845static inline void ext4_simulate_fail_bh(struct super_block *sb,
1846 struct buffer_head *bh,
1847 unsigned long code)
1848{
1849 if (!IS_ERR(bh) && ext4_simulate_fail(sb, code))
1850 clear_buffer_uptodate(bh);
1851}
1852
878520ac
TT
1853/*
1854 * Error number codes for s_{first,last}_error_errno
1855 *
1856 * Linux errno numbers are architecture specific, so we need to translate
1857 * them into something which is architecture independent. We don't define
1858 * codes for all errno's; just the ones which are most likely to be the cause
1859 * of an ext4_error() call.
1860 */
1861#define EXT4_ERR_UNKNOWN 1
1862#define EXT4_ERR_EIO 2
1863#define EXT4_ERR_ENOMEM 3
1864#define EXT4_ERR_EFSBADCRC 4
1865#define EXT4_ERR_EFSCORRUPTED 5
1866#define EXT4_ERR_ENOSPC 6
1867#define EXT4_ERR_ENOKEY 7
1868#define EXT4_ERR_EROFS 8
1869#define EXT4_ERR_EFBIG 9
1870#define EXT4_ERR_EEXIST 10
1871#define EXT4_ERR_ERANGE 11
1872#define EXT4_ERR_EOVERFLOW 12
1873#define EXT4_ERR_EBUSY 13
1874#define EXT4_ERR_ENOTDIR 14
1875#define EXT4_ERR_ENOTEMPTY 15
1876#define EXT4_ERR_ESHUTDOWN 16
1877#define EXT4_ERR_EFAULT 17
1878
19f5fb7a
TT
1879/*
1880 * Inode dynamic state flags
1881 */
1882enum {
19f5fb7a
TT
1883 EXT4_STATE_NEW, /* inode is newly created */
1884 EXT4_STATE_XATTR, /* has in-inode xattrs */
1885 EXT4_STATE_NO_EXPAND, /* No space for expansion */
1886 EXT4_STATE_DA_ALLOC_CLOSE, /* Alloc DA blks on close */
1887 EXT4_STATE_EXT_MIGRATE, /* Inode is migrating */
14ece102 1888 EXT4_STATE_NEWENTRY, /* File just added to dir */
67cf5b09 1889 EXT4_STATE_MAY_INLINE_DATA, /* may have in-inode data */
7869a4a6 1890 EXT4_STATE_EXT_PRECACHED, /* extents have been precached */
a6d05676 1891 EXT4_STATE_LUSTRE_EA_INODE, /* Lustre-style ea_inode */
c93d8f88 1892 EXT4_STATE_VERITY_IN_PROGRESS, /* building fs-verity Merkle tree */
aa75f4d3 1893 EXT4_STATE_FC_COMMITTING, /* Fast commit ongoing */
02f310fc 1894 EXT4_STATE_ORPHAN_FILE, /* Inode orphaned in orphan file */
19f5fb7a
TT
1895};
1896
353eb83c 1897#define EXT4_INODE_BIT_FNS(name, field, offset) \
12e9b892
DM
1898static inline int ext4_test_inode_##name(struct inode *inode, int bit) \
1899{ \
353eb83c 1900 return test_bit(bit + (offset), &EXT4_I(inode)->i_##field); \
12e9b892
DM
1901} \
1902static inline void ext4_set_inode_##name(struct inode *inode, int bit) \
1903{ \
353eb83c 1904 set_bit(bit + (offset), &EXT4_I(inode)->i_##field); \
12e9b892
DM
1905} \
1906static inline void ext4_clear_inode_##name(struct inode *inode, int bit) \
1907{ \
353eb83c 1908 clear_bit(bit + (offset), &EXT4_I(inode)->i_##field); \
19f5fb7a
TT
1909}
1910
3f61c0cc
AD
1911/* Add these declarations here only so that these functions can be
1912 * found by name. Otherwise, they are very hard to locate. */
1913static inline int ext4_test_inode_flag(struct inode *inode, int bit);
1914static inline void ext4_set_inode_flag(struct inode *inode, int bit);
1915static inline void ext4_clear_inode_flag(struct inode *inode, int bit);
353eb83c 1916EXT4_INODE_BIT_FNS(flag, flags, 0)
3f61c0cc
AD
1917
1918/* Add these declarations here only so that these functions can be
1919 * found by name. Otherwise, they are very hard to locate. */
1920static inline int ext4_test_inode_state(struct inode *inode, int bit);
1921static inline void ext4_set_inode_state(struct inode *inode, int bit);
1922static inline void ext4_clear_inode_state(struct inode *inode, int bit);
353eb83c
TT
1923#if (BITS_PER_LONG < 64)
1924EXT4_INODE_BIT_FNS(state, state_flags, 0)
1925
1926static inline void ext4_clear_state_flags(struct ext4_inode_info *ei)
1927{
1928 (ei)->i_state_flags = 0;
1929}
1930#else
1931EXT4_INODE_BIT_FNS(state, flags, 32)
1932
1933static inline void ext4_clear_state_flags(struct ext4_inode_info *ei)
1934{
1935 /* We depend on the fact that callers will set i_flags */
1936}
1937#endif
ac27a0ec 1938#else
617ba13b 1939/* Assume that user mode programs are passing in an ext4fs superblock, not
ac27a0ec
DK
1940 * a kernel struct super_block. This will allow us to call the feature-test
1941 * macros from user land. */
617ba13b 1942#define EXT4_SB(sb) (sb)
ac27a0ec
DK
1943#endif
1944
c93d8f88
EB
1945static inline bool ext4_verity_in_progress(struct inode *inode)
1946{
1947 return IS_ENABLED(CONFIG_FS_VERITY) &&
1948 ext4_test_inode_state(inode, EXT4_STATE_VERITY_IN_PROGRESS);
1949}
1950
617ba13b 1951#define NEXT_ORPHAN(inode) EXT4_I(inode)->i_dtime
ac27a0ec
DK
1952
1953/*
1954 * Codes for operating systems
1955 */
617ba13b
MC
1956#define EXT4_OS_LINUX 0
1957#define EXT4_OS_HURD 1
1958#define EXT4_OS_MASIX 2
1959#define EXT4_OS_FREEBSD 3
1960#define EXT4_OS_LITES 4
ac27a0ec
DK
1961
1962/*
1963 * Revision levels
1964 */
617ba13b
MC
1965#define EXT4_GOOD_OLD_REV 0 /* The good old (original) format */
1966#define EXT4_DYNAMIC_REV 1 /* V2 format w/ dynamic inode sizes */
ac27a0ec 1967
617ba13b 1968#define EXT4_MAX_SUPP_REV EXT4_DYNAMIC_REV
ac27a0ec 1969
617ba13b 1970#define EXT4_GOOD_OLD_INODE_SIZE 128
ac27a0ec 1971
4881c497
DD
1972#define EXT4_EXTRA_TIMESTAMP_MAX (((s64)1 << 34) - 1 + S32_MIN)
1973#define EXT4_NON_EXTRA_TIMESTAMP_MAX S32_MAX
1974#define EXT4_TIMESTAMP_MIN S32_MIN
1975
ac27a0ec
DK
1976/*
1977 * Feature set definitions
1978 */
1979
617ba13b
MC
1980#define EXT4_FEATURE_COMPAT_DIR_PREALLOC 0x0001
1981#define EXT4_FEATURE_COMPAT_IMAGIC_INODES 0x0002
1982#define EXT4_FEATURE_COMPAT_HAS_JOURNAL 0x0004
1983#define EXT4_FEATURE_COMPAT_EXT_ATTR 0x0008
1984#define EXT4_FEATURE_COMPAT_RESIZE_INODE 0x0010
1985#define EXT4_FEATURE_COMPAT_DIR_INDEX 0x0020
1beeef1b 1986#define EXT4_FEATURE_COMPAT_SPARSE_SUPER2 0x0200
a44ad683
HS
1987/*
1988 * The reason why "FAST_COMMIT" is a compat feature is that, FS becomes
1989 * incompatible only if fast commit blocks are present in the FS. Since we
1990 * clear the journal (and thus the fast commit blocks), we don't mark FS as
1991 * incompatible. We also have a JBD2 incompat feature, which gets set when
1992 * there are fast commit blocks present in the journal.
1993 */
995a3ed6 1994#define EXT4_FEATURE_COMPAT_FAST_COMMIT 0x0400
b925acb8 1995#define EXT4_FEATURE_COMPAT_STABLE_INODES 0x0800
02f310fc 1996#define EXT4_FEATURE_COMPAT_ORPHAN_FILE 0x1000 /* Orphan file exists */
617ba13b
MC
1997
1998#define EXT4_FEATURE_RO_COMPAT_SPARSE_SUPER 0x0001
1999#define EXT4_FEATURE_RO_COMPAT_LARGE_FILE 0x0002
2000#define EXT4_FEATURE_RO_COMPAT_BTREE_DIR 0x0004
0fc1b451 2001#define EXT4_FEATURE_RO_COMPAT_HUGE_FILE 0x0008
717d50e4 2002#define EXT4_FEATURE_RO_COMPAT_GDT_CSUM 0x0010
f8628a14 2003#define EXT4_FEATURE_RO_COMPAT_DIR_NLINK 0x0020
ef7f3835 2004#define EXT4_FEATURE_RO_COMPAT_EXTRA_ISIZE 0x0040
ae812306 2005#define EXT4_FEATURE_RO_COMPAT_QUOTA 0x0100
281b5995 2006#define EXT4_FEATURE_RO_COMPAT_BIGALLOC 0x0200
e6153918
DW
2007/*
2008 * METADATA_CSUM also enables group descriptor checksums (GDT_CSUM). When
2009 * METADATA_CSUM is set, group descriptor checksums use the same algorithm as
2010 * all other data structures' checksums. However, the METADATA_CSUM and
2011 * GDT_CSUM bits are mutually exclusive.
2012 */
9b90e5e0 2013#define EXT4_FEATURE_RO_COMPAT_METADATA_CSUM 0x0400
2cb5cc8b 2014#define EXT4_FEATURE_RO_COMPAT_READONLY 0x1000
8b4953e1 2015#define EXT4_FEATURE_RO_COMPAT_PROJECT 0x2000
c93d8f88 2016#define EXT4_FEATURE_RO_COMPAT_VERITY 0x8000
02f310fc
JK
2017#define EXT4_FEATURE_RO_COMPAT_ORPHAN_PRESENT 0x10000 /* Orphan file may be
2018 non-empty */
617ba13b
MC
2019
2020#define EXT4_FEATURE_INCOMPAT_COMPRESSION 0x0001
2021#define EXT4_FEATURE_INCOMPAT_FILETYPE 0x0002
2022#define EXT4_FEATURE_INCOMPAT_RECOVER 0x0004 /* Needs recovery */
2023#define EXT4_FEATURE_INCOMPAT_JOURNAL_DEV 0x0008 /* Journal device */
2024#define EXT4_FEATURE_INCOMPAT_META_BG 0x0010
a86c6181 2025#define EXT4_FEATURE_INCOMPAT_EXTENTS 0x0040 /* extents support */
bd81d8ee 2026#define EXT4_FEATURE_INCOMPAT_64BIT 0x0080
0fc1b451 2027#define EXT4_FEATURE_INCOMPAT_MMP 0x0100
ce421581 2028#define EXT4_FEATURE_INCOMPAT_FLEX_BG 0x0200
f710b4b9
TT
2029#define EXT4_FEATURE_INCOMPAT_EA_INODE 0x0400 /* EA in inode */
2030#define EXT4_FEATURE_INCOMPAT_DIRDATA 0x1000 /* data in dirent */
8c81bd8f 2031#define EXT4_FEATURE_INCOMPAT_CSUM_SEED 0x2000
9b90e5e0 2032#define EXT4_FEATURE_INCOMPAT_LARGEDIR 0x4000 /* >2GB or 3-lvl htree */
67cf5b09 2033#define EXT4_FEATURE_INCOMPAT_INLINE_DATA 0x8000 /* data in inode */
3edc18d8 2034#define EXT4_FEATURE_INCOMPAT_ENCRYPT 0x10000
c83ad55e 2035#define EXT4_FEATURE_INCOMPAT_CASEFOLD 0x20000
617ba13b 2036
c9e716eb
AD
2037extern void ext4_update_dynamic_rev(struct super_block *sb);
2038
e2b911c5
DW
2039#define EXT4_FEATURE_COMPAT_FUNCS(name, flagname) \
2040static inline bool ext4_has_feature_##name(struct super_block *sb) \
2041{ \
2042 return ((EXT4_SB(sb)->s_es->s_feature_compat & \
2043 cpu_to_le32(EXT4_FEATURE_COMPAT_##flagname)) != 0); \
2044} \
2045static inline void ext4_set_feature_##name(struct super_block *sb) \
2046{ \
c9e716eb 2047 ext4_update_dynamic_rev(sb); \
e2b911c5
DW
2048 EXT4_SB(sb)->s_es->s_feature_compat |= \
2049 cpu_to_le32(EXT4_FEATURE_COMPAT_##flagname); \
2050} \
2051static inline void ext4_clear_feature_##name(struct super_block *sb) \
2052{ \
2053 EXT4_SB(sb)->s_es->s_feature_compat &= \
2054 ~cpu_to_le32(EXT4_FEATURE_COMPAT_##flagname); \
2055}
2056
2057#define EXT4_FEATURE_RO_COMPAT_FUNCS(name, flagname) \
2058static inline bool ext4_has_feature_##name(struct super_block *sb) \
2059{ \
2060 return ((EXT4_SB(sb)->s_es->s_feature_ro_compat & \
2061 cpu_to_le32(EXT4_FEATURE_RO_COMPAT_##flagname)) != 0); \
2062} \
2063static inline void ext4_set_feature_##name(struct super_block *sb) \
2064{ \
c9e716eb 2065 ext4_update_dynamic_rev(sb); \
e2b911c5
DW
2066 EXT4_SB(sb)->s_es->s_feature_ro_compat |= \
2067 cpu_to_le32(EXT4_FEATURE_RO_COMPAT_##flagname); \
2068} \
2069static inline void ext4_clear_feature_##name(struct super_block *sb) \
2070{ \
2071 EXT4_SB(sb)->s_es->s_feature_ro_compat &= \
2072 ~cpu_to_le32(EXT4_FEATURE_RO_COMPAT_##flagname); \
2073}
2074
2075#define EXT4_FEATURE_INCOMPAT_FUNCS(name, flagname) \
2076static inline bool ext4_has_feature_##name(struct super_block *sb) \
2077{ \
2078 return ((EXT4_SB(sb)->s_es->s_feature_incompat & \
2079 cpu_to_le32(EXT4_FEATURE_INCOMPAT_##flagname)) != 0); \
2080} \
2081static inline void ext4_set_feature_##name(struct super_block *sb) \
2082{ \
c9e716eb 2083 ext4_update_dynamic_rev(sb); \
e2b911c5
DW
2084 EXT4_SB(sb)->s_es->s_feature_incompat |= \
2085 cpu_to_le32(EXT4_FEATURE_INCOMPAT_##flagname); \
2086} \
2087static inline void ext4_clear_feature_##name(struct super_block *sb) \
2088{ \
2089 EXT4_SB(sb)->s_es->s_feature_incompat &= \
2090 ~cpu_to_le32(EXT4_FEATURE_INCOMPAT_##flagname); \
2091}
2092
2093EXT4_FEATURE_COMPAT_FUNCS(dir_prealloc, DIR_PREALLOC)
2094EXT4_FEATURE_COMPAT_FUNCS(imagic_inodes, IMAGIC_INODES)
2095EXT4_FEATURE_COMPAT_FUNCS(journal, HAS_JOURNAL)
2096EXT4_FEATURE_COMPAT_FUNCS(xattr, EXT_ATTR)
2097EXT4_FEATURE_COMPAT_FUNCS(resize_inode, RESIZE_INODE)
2098EXT4_FEATURE_COMPAT_FUNCS(dir_index, DIR_INDEX)
2099EXT4_FEATURE_COMPAT_FUNCS(sparse_super2, SPARSE_SUPER2)
995a3ed6 2100EXT4_FEATURE_COMPAT_FUNCS(fast_commit, FAST_COMMIT)
b925acb8 2101EXT4_FEATURE_COMPAT_FUNCS(stable_inodes, STABLE_INODES)
02f310fc 2102EXT4_FEATURE_COMPAT_FUNCS(orphan_file, ORPHAN_FILE)
e2b911c5
DW
2103
2104EXT4_FEATURE_RO_COMPAT_FUNCS(sparse_super, SPARSE_SUPER)
2105EXT4_FEATURE_RO_COMPAT_FUNCS(large_file, LARGE_FILE)
2106EXT4_FEATURE_RO_COMPAT_FUNCS(btree_dir, BTREE_DIR)
2107EXT4_FEATURE_RO_COMPAT_FUNCS(huge_file, HUGE_FILE)
2108EXT4_FEATURE_RO_COMPAT_FUNCS(gdt_csum, GDT_CSUM)
2109EXT4_FEATURE_RO_COMPAT_FUNCS(dir_nlink, DIR_NLINK)
2110EXT4_FEATURE_RO_COMPAT_FUNCS(extra_isize, EXTRA_ISIZE)
2111EXT4_FEATURE_RO_COMPAT_FUNCS(quota, QUOTA)
2112EXT4_FEATURE_RO_COMPAT_FUNCS(bigalloc, BIGALLOC)
2113EXT4_FEATURE_RO_COMPAT_FUNCS(metadata_csum, METADATA_CSUM)
2114EXT4_FEATURE_RO_COMPAT_FUNCS(readonly, READONLY)
2115EXT4_FEATURE_RO_COMPAT_FUNCS(project, PROJECT)
c93d8f88 2116EXT4_FEATURE_RO_COMPAT_FUNCS(verity, VERITY)
02f310fc 2117EXT4_FEATURE_RO_COMPAT_FUNCS(orphan_present, ORPHAN_PRESENT)
e2b911c5
DW
2118
2119EXT4_FEATURE_INCOMPAT_FUNCS(compression, COMPRESSION)
2120EXT4_FEATURE_INCOMPAT_FUNCS(filetype, FILETYPE)
2121EXT4_FEATURE_INCOMPAT_FUNCS(journal_needs_recovery, RECOVER)
2122EXT4_FEATURE_INCOMPAT_FUNCS(journal_dev, JOURNAL_DEV)
2123EXT4_FEATURE_INCOMPAT_FUNCS(meta_bg, META_BG)
2124EXT4_FEATURE_INCOMPAT_FUNCS(extents, EXTENTS)
2125EXT4_FEATURE_INCOMPAT_FUNCS(64bit, 64BIT)
2126EXT4_FEATURE_INCOMPAT_FUNCS(mmp, MMP)
2127EXT4_FEATURE_INCOMPAT_FUNCS(flex_bg, FLEX_BG)
2128EXT4_FEATURE_INCOMPAT_FUNCS(ea_inode, EA_INODE)
2129EXT4_FEATURE_INCOMPAT_FUNCS(dirdata, DIRDATA)
2130EXT4_FEATURE_INCOMPAT_FUNCS(csum_seed, CSUM_SEED)
2131EXT4_FEATURE_INCOMPAT_FUNCS(largedir, LARGEDIR)
2132EXT4_FEATURE_INCOMPAT_FUNCS(inline_data, INLINE_DATA)
2133EXT4_FEATURE_INCOMPAT_FUNCS(encrypt, ENCRYPT)
c83ad55e 2134EXT4_FEATURE_INCOMPAT_FUNCS(casefold, CASEFOLD)
e2b911c5 2135
2035e776
TT
2136#define EXT2_FEATURE_COMPAT_SUPP EXT4_FEATURE_COMPAT_EXT_ATTR
2137#define EXT2_FEATURE_INCOMPAT_SUPP (EXT4_FEATURE_INCOMPAT_FILETYPE| \
2138 EXT4_FEATURE_INCOMPAT_META_BG)
2139#define EXT2_FEATURE_RO_COMPAT_SUPP (EXT4_FEATURE_RO_COMPAT_SPARSE_SUPER| \
2140 EXT4_FEATURE_RO_COMPAT_LARGE_FILE| \
2141 EXT4_FEATURE_RO_COMPAT_BTREE_DIR)
2142
2143#define EXT3_FEATURE_COMPAT_SUPP EXT4_FEATURE_COMPAT_EXT_ATTR
2144#define EXT3_FEATURE_INCOMPAT_SUPP (EXT4_FEATURE_INCOMPAT_FILETYPE| \
2145 EXT4_FEATURE_INCOMPAT_RECOVER| \
2146 EXT4_FEATURE_INCOMPAT_META_BG)
2147#define EXT3_FEATURE_RO_COMPAT_SUPP (EXT4_FEATURE_RO_COMPAT_SPARSE_SUPER| \
2148 EXT4_FEATURE_RO_COMPAT_LARGE_FILE| \
2149 EXT4_FEATURE_RO_COMPAT_BTREE_DIR)
2150
02f310fc
JK
2151#define EXT4_FEATURE_COMPAT_SUPP (EXT4_FEATURE_COMPAT_EXT_ATTR| \
2152 EXT4_FEATURE_COMPAT_ORPHAN_FILE)
617ba13b
MC
2153#define EXT4_FEATURE_INCOMPAT_SUPP (EXT4_FEATURE_INCOMPAT_FILETYPE| \
2154 EXT4_FEATURE_INCOMPAT_RECOVER| \
a86c6181 2155 EXT4_FEATURE_INCOMPAT_META_BG| \
bd81d8ee 2156 EXT4_FEATURE_INCOMPAT_EXTENTS| \
ce421581 2157 EXT4_FEATURE_INCOMPAT_64BIT| \
c5e06d10 2158 EXT4_FEATURE_INCOMPAT_FLEX_BG| \
e50e5129 2159 EXT4_FEATURE_INCOMPAT_EA_INODE| \
6ddb2447
TT
2160 EXT4_FEATURE_INCOMPAT_MMP | \
2161 EXT4_FEATURE_INCOMPAT_INLINE_DATA | \
8c81bd8f 2162 EXT4_FEATURE_INCOMPAT_ENCRYPT | \
c83ad55e 2163 EXT4_FEATURE_INCOMPAT_CASEFOLD | \
e08ac99f
AB
2164 EXT4_FEATURE_INCOMPAT_CSUM_SEED | \
2165 EXT4_FEATURE_INCOMPAT_LARGEDIR)
617ba13b
MC
2166#define EXT4_FEATURE_RO_COMPAT_SUPP (EXT4_FEATURE_RO_COMPAT_SPARSE_SUPER| \
2167 EXT4_FEATURE_RO_COMPAT_LARGE_FILE| \
717d50e4 2168 EXT4_FEATURE_RO_COMPAT_GDT_CSUM| \
f8628a14 2169 EXT4_FEATURE_RO_COMPAT_DIR_NLINK | \
ef7f3835 2170 EXT4_FEATURE_RO_COMPAT_EXTRA_ISIZE | \
0fc1b451 2171 EXT4_FEATURE_RO_COMPAT_BTREE_DIR |\
6f16b606 2172 EXT4_FEATURE_RO_COMPAT_HUGE_FILE |\
e93376c2 2173 EXT4_FEATURE_RO_COMPAT_BIGALLOC |\
7c319d32 2174 EXT4_FEATURE_RO_COMPAT_METADATA_CSUM|\
040cb378 2175 EXT4_FEATURE_RO_COMPAT_QUOTA |\
c93d8f88 2176 EXT4_FEATURE_RO_COMPAT_PROJECT |\
02f310fc
JK
2177 EXT4_FEATURE_RO_COMPAT_VERITY |\
2178 EXT4_FEATURE_RO_COMPAT_ORPHAN_PRESENT)
ac27a0ec 2179
e2b911c5
DW
2180#define EXTN_FEATURE_FUNCS(ver) \
2181static inline bool ext4_has_unknown_ext##ver##_compat_features(struct super_block *sb) \
2182{ \
2183 return ((EXT4_SB(sb)->s_es->s_feature_compat & \
2184 cpu_to_le32(~EXT##ver##_FEATURE_COMPAT_SUPP)) != 0); \
2185} \
2186static inline bool ext4_has_unknown_ext##ver##_ro_compat_features(struct super_block *sb) \
2187{ \
2188 return ((EXT4_SB(sb)->s_es->s_feature_ro_compat & \
2189 cpu_to_le32(~EXT##ver##_FEATURE_RO_COMPAT_SUPP)) != 0); \
2190} \
2191static inline bool ext4_has_unknown_ext##ver##_incompat_features(struct super_block *sb) \
2192{ \
2193 return ((EXT4_SB(sb)->s_es->s_feature_incompat & \
2194 cpu_to_le32(~EXT##ver##_FEATURE_INCOMPAT_SUPP)) != 0); \
2195}
2196
2197EXTN_FEATURE_FUNCS(2)
2198EXTN_FEATURE_FUNCS(3)
2199EXTN_FEATURE_FUNCS(4)
2200
2201static inline bool ext4_has_compat_features(struct super_block *sb)
2202{
2203 return (EXT4_SB(sb)->s_es->s_feature_compat != 0);
2204}
2205static inline bool ext4_has_ro_compat_features(struct super_block *sb)
2206{
2207 return (EXT4_SB(sb)->s_es->s_feature_ro_compat != 0);
2208}
2209static inline bool ext4_has_incompat_features(struct super_block *sb)
2210{
2211 return (EXT4_SB(sb)->s_es->s_feature_incompat != 0);
2212}
2213
25c6d98f
JK
2214extern int ext4_feature_set_ok(struct super_block *sb, int readonly);
2215
9549a168
TT
2216/*
2217 * Superblock flags
2218 */
2219#define EXT4_FLAGS_RESIZING 0
0db1ff22 2220#define EXT4_FLAGS_SHUTDOWN 1
a8ab6d38 2221#define EXT4_FLAGS_BDEV_IS_DAX 2
0db1ff22 2222
eb8ab444 2223static inline int ext4_forced_shutdown(struct super_block *sb)
0db1ff22 2224{
eb8ab444 2225 return test_bit(EXT4_FLAGS_SHUTDOWN, &EXT4_SB(sb)->s_ext4_flags);
0db1ff22 2226}
9549a168 2227
ac27a0ec
DK
2228/*
2229 * Default values for user and/or group using reserved blocks
2230 */
617ba13b
MC
2231#define EXT4_DEF_RESUID 0
2232#define EXT4_DEF_RESGID 0
ac27a0ec 2233
040cb378
LX
2234/*
2235 * Default project ID
2236 */
2237#define EXT4_DEF_PROJID 0
2238
240799cd
TT
2239#define EXT4_DEF_INODE_READAHEAD_BLKS 32
2240
ac27a0ec
DK
2241/*
2242 * Default mount options
2243 */
617ba13b
MC
2244#define EXT4_DEFM_DEBUG 0x0001
2245#define EXT4_DEFM_BSDGROUPS 0x0002
2246#define EXT4_DEFM_XATTR_USER 0x0004
2247#define EXT4_DEFM_ACL 0x0008
2248#define EXT4_DEFM_UID16 0x0010
2249#define EXT4_DEFM_JMODE 0x0060
2250#define EXT4_DEFM_JMODE_DATA 0x0020
2251#define EXT4_DEFM_JMODE_ORDERED 0x0040
2252#define EXT4_DEFM_JMODE_WBACK 0x0060
8b67f04a
TT
2253#define EXT4_DEFM_NOBARRIER 0x0100
2254#define EXT4_DEFM_BLOCK_VALIDITY 0x0200
2255#define EXT4_DEFM_DISCARD 0x0400
2256#define EXT4_DEFM_NODELALLOC 0x0800
ac27a0ec 2257
30773840
TT
2258/*
2259 * Default journal batch times
2260 */
2261#define EXT4_DEF_MIN_BATCH_TIME 0
2262#define EXT4_DEF_MAX_BATCH_TIME 15000 /* 15ms */
2263
a4912123
TT
2264/*
2265 * Minimum number of groups in a flexgroup before we separate out
2266 * directories into the first block group of a flexgroup
2267 */
2268#define EXT4_FLEX_SIZE_DIR_ALLOC_SCHEME 4
2269
ac27a0ec
DK
2270/*
2271 * Structure of a directory entry
2272 */
617ba13b 2273#define EXT4_NAME_LEN 255
c186f088
YB
2274/*
2275 * Base length of the ext4 directory entry excluding the name length
2276 */
2277#define EXT4_BASE_DIR_LEN (sizeof(struct ext4_dir_entry_2) - EXT4_NAME_LEN)
ac27a0ec 2278
617ba13b 2279struct ext4_dir_entry {
ac27a0ec
DK
2280 __le32 inode; /* Inode number */
2281 __le16 rec_len; /* Directory entry length */
2282 __le16 name_len; /* Name length */
617ba13b 2283 char name[EXT4_NAME_LEN]; /* File name */
ac27a0ec
DK
2284};
2285
471fbbea
DR
2286
2287/*
2288 * Encrypted Casefolded entries require saving the hash on disk. This structure
2289 * followed ext4_dir_entry_2's name[name_len] at the next 4 byte aligned
2290 * boundary.
2291 */
2292struct ext4_dir_entry_hash {
2293 __le32 hash;
2294 __le32 minor_hash;
2295};
2296
ac27a0ec 2297/*
617ba13b 2298 * The new version of the directory entry. Since EXT4 structures are
ac27a0ec
DK
2299 * stored in intel byte order, and the name_len field could never be
2300 * bigger than 255 chars, it's safe to reclaim the extra byte for the
2301 * file_type field.
2302 */
617ba13b 2303struct ext4_dir_entry_2 {
ac27a0ec
DK
2304 __le32 inode; /* Inode number */
2305 __le16 rec_len; /* Directory entry length */
2306 __u8 name_len; /* Name length */
9f364e1d 2307 __u8 file_type; /* See file type macros EXT4_FT_* below */
617ba13b 2308 char name[EXT4_NAME_LEN]; /* File name */
ac27a0ec
DK
2309};
2310
471fbbea
DR
2311/*
2312 * Access the hashes at the end of ext4_dir_entry_2
2313 */
2314#define EXT4_DIRENT_HASHES(entry) \
2315 ((struct ext4_dir_entry_hash *) \
2316 (((void *)(entry)) + \
2317 ((8 + (entry)->name_len + EXT4_DIR_ROUND) & ~EXT4_DIR_ROUND)))
2318#define EXT4_DIRENT_HASH(entry) le32_to_cpu(EXT4_DIRENT_HASHES(de)->hash)
2319#define EXT4_DIRENT_MINOR_HASH(entry) \
2320 le32_to_cpu(EXT4_DIRENT_HASHES(de)->minor_hash)
2321
2322static inline bool ext4_hash_in_dirent(const struct inode *inode)
2323{
2324 return IS_CASEFOLDED(inode) && IS_ENCRYPTED(inode);
2325}
2326
e6153918
DW
2327/*
2328 * This is a bogus directory entry at the end of each leaf block that
2329 * records checksums.
2330 */
2331struct ext4_dir_entry_tail {
2332 __le32 det_reserved_zero1; /* Pretend to be unused */
2333 __le16 det_rec_len; /* 12 */
2334 __u8 det_reserved_zero2; /* Zero name length */
2335 __u8 det_reserved_ft; /* 0xDE, fake file type */
2336 __le32 det_checksum; /* crc32c(uuid+inum+dirblock) */
2337};
2338
3c47d541
TM
2339#define EXT4_DIRENT_TAIL(block, blocksize) \
2340 ((struct ext4_dir_entry_tail *)(((void *)(block)) + \
2341 ((blocksize) - \
2342 sizeof(struct ext4_dir_entry_tail))))
2343
ac27a0ec 2344/*
617ba13b 2345 * Ext4 directory file types. Only the low 3 bits are used. The
ac27a0ec
DK
2346 * other bits are reserved for now.
2347 */
617ba13b
MC
2348#define EXT4_FT_UNKNOWN 0
2349#define EXT4_FT_REG_FILE 1
2350#define EXT4_FT_DIR 2
2351#define EXT4_FT_CHRDEV 3
2352#define EXT4_FT_BLKDEV 4
2353#define EXT4_FT_FIFO 5
2354#define EXT4_FT_SOCK 6
2355#define EXT4_FT_SYMLINK 7
ac27a0ec 2356
617ba13b 2357#define EXT4_FT_MAX 8
ac27a0ec 2358
e6153918
DW
2359#define EXT4_FT_DIR_CSUM 0xDE
2360
ac27a0ec 2361/*
617ba13b 2362 * EXT4_DIR_PAD defines the directory entries boundaries
ac27a0ec
DK
2363 *
2364 * NOTE: It must be a multiple of 4
2365 */
617ba13b
MC
2366#define EXT4_DIR_PAD 4
2367#define EXT4_DIR_ROUND (EXT4_DIR_PAD - 1)
a72d7f83
JK
2368#define EXT4_MAX_REC_LEN ((1<<16)-1)
2369
471fbbea
DR
2370/*
2371 * The rec_len is dependent on the type of directory. Directories that are
2372 * casefolded and encrypted need to store the hash as well, so we add room for
2373 * ext4_extended_dir_entry_2. For all entries related to '.' or '..' you should
2374 * pass NULL for dir, as those entries do not use the extra fields.
2375 */
2376static inline unsigned int ext4_dir_rec_len(__u8 name_len,
2377 const struct inode *dir)
2378{
2379 int rec_len = (name_len + 8 + EXT4_DIR_ROUND);
2380
2381 if (dir && ext4_hash_in_dirent(dir))
2382 rec_len += sizeof(struct ext4_dir_entry_hash);
2383 return (rec_len & ~EXT4_DIR_ROUND);
2384}
2385
0cfc9255
ES
2386/*
2387 * If we ever get support for fs block sizes > page_size, we'll need
2388 * to remove the #if statements in the next two functions...
2389 */
2390static inline unsigned int
2391ext4_rec_len_from_disk(__le16 dlen, unsigned blocksize)
2392{
2393 unsigned len = le16_to_cpu(dlen);
2394
ea1754a0 2395#if (PAGE_SIZE >= 65536)
0cfc9255
ES
2396 if (len == EXT4_MAX_REC_LEN || len == 0)
2397 return blocksize;
2398 return (len & 65532) | ((len & 3) << 16);
2399#else
2400 return len;
2401#endif
2402}
2403
2404static inline __le16 ext4_rec_len_to_disk(unsigned len, unsigned blocksize)
2405{
a2e3965d 2406 BUG_ON((len > blocksize) || (blocksize > (1 << 18)) || (len & 3));
ea1754a0 2407#if (PAGE_SIZE >= 65536)
0cfc9255
ES
2408 if (len < 65536)
2409 return cpu_to_le16(len);
2410 if (len == blocksize) {
2411 if (blocksize == 65536)
2412 return cpu_to_le16(EXT4_MAX_REC_LEN);
2413 else
2414 return cpu_to_le16(0);
2415 }
2416 return cpu_to_le16((len & 65532) | ((len >> 16) & 3));
2417#else
2418 return cpu_to_le16(len);
2419#endif
2420}
2421
ac27a0ec
DK
2422/*
2423 * Hash Tree Directory indexing
2424 * (c) Daniel Phillips, 2001
2425 */
2426
e2b911c5 2427#define is_dx(dir) (ext4_has_feature_dir_index((dir)->i_sb) && \
12e9b892 2428 ext4_test_inode_flag((dir), EXT4_INODE_INDEX))
c7414892
AD
2429#define EXT4_DIR_LINK_MAX(dir) unlikely((dir)->i_nlink >= EXT4_LINK_MAX && \
2430 !(ext4_has_feature_dir_nlink((dir)->i_sb) && is_dx(dir)))
617ba13b 2431#define EXT4_DIR_LINK_EMPTY(dir) ((dir)->i_nlink == 2 || (dir)->i_nlink == 1)
ac27a0ec
DK
2432
2433/* Legal values for the dx_root hash_version field: */
2434
e7093f0d
PK
2435#define DX_HASH_LEGACY 0
2436#define DX_HASH_HALF_MD4 1
2437#define DX_HASH_TEA 2
2438#define DX_HASH_LEGACY_UNSIGNED 3
f99b2589
TT
2439#define DX_HASH_HALF_MD4_UNSIGNED 4
2440#define DX_HASH_TEA_UNSIGNED 5
471fbbea 2441#define DX_HASH_SIPHASH 6
ac27a0ec 2442
0441984a
DW
2443static inline u32 ext4_chksum(struct ext4_sb_info *sbi, u32 crc,
2444 const void *address, unsigned int length)
2445{
2446 struct {
2447 struct shash_desc shash;
3108b54b 2448 char ctx[4];
0441984a 2449 } desc;
0441984a 2450
3108b54b
TT
2451 BUG_ON(crypto_shash_descsize(sbi->s_chksum_driver)!=sizeof(desc.ctx));
2452
0441984a 2453 desc.shash.tfm = sbi->s_chksum_driver;
0441984a
DW
2454 *(u32 *)desc.ctx = crc;
2455
e7093f0d 2456 BUG_ON(crypto_shash_update(&desc.shash, address, length));
0441984a
DW
2457
2458 return *(u32 *)desc.ctx;
2459}
2460
ac27a0ec
DK
2461#ifdef __KERNEL__
2462
2463/* hash info structure used by the directory hash */
2464struct dx_hash_info
2465{
2466 u32 hash;
2467 u32 minor_hash;
2468 int hash_version;
2469 u32 *seed;
2470};
2471
d1f5273e
FY
2472
2473/* 32 and 64 bit signed EOF for dx directories */
2474#define EXT4_HTREE_EOF_32BIT ((1UL << (32 - 1)) - 1)
2475#define EXT4_HTREE_EOF_64BIT ((1ULL << (64 - 1)) - 1)
2476
ac27a0ec
DK
2477
2478/*
617ba13b 2479 * Control parameters used by ext4_htree_next_block
ac27a0ec
DK
2480 */
2481#define HASH_NB_ALWAYS 1
2482
5b643f9c
TT
2483struct ext4_filename {
2484 const struct qstr *usr_fname;
a7550b30 2485 struct fscrypt_str disk_name;
5b643f9c 2486 struct dx_hash_info hinfo;
643fa961 2487#ifdef CONFIG_FS_ENCRYPTION
a7550b30 2488 struct fscrypt_str crypto_buf;
5b643f9c 2489#endif
5298d4bf 2490#if IS_ENABLED(CONFIG_UNICODE)
3ae72562
GKB
2491 struct fscrypt_str cf_name;
2492#endif
5b643f9c
TT
2493};
2494
2495#define fname_name(p) ((p)->disk_name.name)
471fbbea 2496#define fname_usr_name(p) ((p)->usr_fname->name)
5b643f9c 2497#define fname_len(p) ((p)->disk_name.len)
ac27a0ec
DK
2498
2499/*
2500 * Describe an inode's exact location on disk and in memory
2501 */
617ba13b 2502struct ext4_iloc
ac27a0ec
DK
2503{
2504 struct buffer_head *bh;
2505 unsigned long offset;
fd2d4291 2506 ext4_group_t block_group;
ac27a0ec
DK
2507};
2508
617ba13b 2509static inline struct ext4_inode *ext4_raw_inode(struct ext4_iloc *iloc)
ac27a0ec 2510{
617ba13b 2511 return (struct ext4_inode *) (iloc->bh->b_data + iloc->offset);
ac27a0ec
DK
2512}
2513
dec214d0
TE
2514static inline bool ext4_is_quota_file(struct inode *inode)
2515{
2516 return IS_NOQUOTA(inode) &&
2517 !(EXT4_I(inode)->i_flags & EXT4_EA_INODE_FL);
2518}
02749a4c 2519
ac27a0ec
DK
2520/*
2521 * This structure is stuffed into the struct file's private_data field
2522 * for directories. It is where we put information so that we can do
2523 * readdir operations in hash tree order.
2524 */
2525struct dir_private_info {
2526 struct rb_root root;
2527 struct rb_node *curr_node;
2528 struct fname *extra_fname;
2529 loff_t last_pos;
2530 __u32 curr_hash;
2531 __u32 curr_minor_hash;
2532 __u32 next_hash;
2533};
2534
2535/* calculate the first block number of the group */
617ba13b 2536static inline ext4_fsblk_t
fd2d4291 2537ext4_group_first_block_no(struct super_block *sb, ext4_group_t group_no)
ac27a0ec 2538{
617ba13b
MC
2539 return group_no * (ext4_fsblk_t)EXT4_BLOCKS_PER_GROUP(sb) +
2540 le32_to_cpu(EXT4_SB(sb)->s_es->s_first_data_block);
ac27a0ec
DK
2541}
2542
2543/*
2544 * Special error return code only used by dx_probe() and its callers.
2545 */
36de9286 2546#define ERR_BAD_DX_DIR (-(MAX_ERRNO - 1))
ac27a0ec 2547
e08ac99f
AB
2548/* htree levels for ext4 */
2549#define EXT4_HTREE_LEVEL_COMPAT 2
2550#define EXT4_HTREE_LEVEL 3
2551
2552static inline int ext4_dir_htree_level(struct super_block *sb)
2553{
2554 return ext4_has_feature_largedir(sb) ?
2555 EXT4_HTREE_LEVEL : EXT4_HTREE_LEVEL_COMPAT;
2556}
2557
bfff6873
LC
2558/*
2559 * Timeout and state flag for lazy initialization inode thread.
2560 */
2561#define EXT4_DEF_LI_WAIT_MULT 10
2562#define EXT4_DEF_LI_MAX_START_DELAY 5
2563#define EXT4_LAZYINIT_QUIT 0x0001
2564#define EXT4_LAZYINIT_RUNNING 0x0002
2565
2566/*
2567 * Lazy inode table initialization info
2568 */
2569struct ext4_lazy_init {
2570 unsigned long li_state;
bfff6873
LC
2571 struct list_head li_request_list;
2572 struct mutex li_list_mtx;
2573};
2574
3d392b26
TT
2575enum ext4_li_mode {
2576 EXT4_LI_MODE_PREFETCH_BBITMAP,
2577 EXT4_LI_MODE_ITABLE,
2578};
2579
bfff6873
LC
2580struct ext4_li_request {
2581 struct super_block *lr_super;
3d392b26
TT
2582 enum ext4_li_mode lr_mode;
2583 ext4_group_t lr_first_not_zeroed;
bfff6873
LC
2584 ext4_group_t lr_next_group;
2585 struct list_head lr_request;
2586 unsigned long lr_next_sched;
2587 unsigned long lr_timeout;
2588};
2589
857ac889
LC
2590struct ext4_features {
2591 struct kobject f_kobj;
2592 struct completion f_kobj_unregister;
2593};
2594
c5e06d10
JL
2595/*
2596 * This structure will be used for multiple mount protection. It will be
2597 * written into the block number saved in the s_mmp_block field in the
2598 * superblock. Programs that check MMP should assume that if
2599 * SEQ_FSCK (or any unknown code above SEQ_MAX) is present then it is NOT safe
2600 * to use the filesystem, regardless of how old the timestamp is.
2601 */
2602#define EXT4_MMP_MAGIC 0x004D4D50U /* ASCII for MMP */
2603#define EXT4_MMP_SEQ_CLEAN 0xFF4D4D50U /* mmp_seq value for clean unmount */
2604#define EXT4_MMP_SEQ_FSCK 0xE24D4D50U /* mmp_seq value when being fscked */
2605#define EXT4_MMP_SEQ_MAX 0xE24D4D4FU /* maximum valid mmp_seq value */
2606
2607struct mmp_struct {
2608 __le32 mmp_magic; /* Magic number for MMP */
2609 __le32 mmp_seq; /* Sequence no. updated periodically */
2610
2611 /*
2612 * mmp_time, mmp_nodename & mmp_bdevname are only used for information
2613 * purposes and do not affect the correctness of the algorithm
2614 */
2615 __le64 mmp_time; /* Time last updated */
2616 char mmp_nodename[64]; /* Node which last updated MMP block */
2617 char mmp_bdevname[32]; /* Bdev which last updated MMP block */
2618
2619 /*
2620 * mmp_check_interval is used to verify if the MMP block has been
2621 * updated on the block device. The value is updated based on the
2622 * maximum time to write the MMP block during an update cycle.
2623 */
2624 __le16 mmp_check_interval;
2625
2626 __le16 mmp_pad1;
e6153918
DW
2627 __le32 mmp_pad2[226];
2628 __le32 mmp_checksum; /* crc32c(uuid+mmp_block) */
c5e06d10
JL
2629};
2630
2631/* arguments passed to the mmp thread */
2632struct mmpd_data {
2633 struct buffer_head *bh; /* bh from initial read_mmp_block() */
2634 struct super_block *sb; /* super block of the fs */
2635};
2636
2637/*
2638 * Check interval multiplier
2639 * The MMP block is written every update interval and initially checked every
2640 * update interval x the multiplier (the value is then adapted based on the
2641 * write latency). The reason is that writes can be delayed under load and we
2642 * don't want readers to incorrectly assume that the filesystem is no longer
2643 * in use.
2644 */
2645#define EXT4_MMP_CHECK_MULT 2UL
2646
2647/*
2648 * Minimum interval for MMP checking in seconds.
2649 */
2650#define EXT4_MMP_MIN_CHECK_INTERVAL 5UL
2651
2652/*
2653 * Maximum interval for MMP checking in seconds.
2654 */
2655#define EXT4_MMP_MAX_CHECK_INTERVAL 300UL
2656
ac27a0ec
DK
2657/*
2658 * Function prototypes
2659 */
2660
2661/*
2662 * Ok, these declarations are also in <linux/kernel.h> but none of the
617ba13b 2663 * ext4 source programs needs to include it so they are duplicated here.
ac27a0ec 2664 */
8c55e204
DK
2665# define NORET_TYPE /**/
2666# define ATTRIB_NORET __attribute__((noreturn))
2667# define NORET_AND noreturn,
ac27a0ec 2668
498e5f24 2669/* bitmap.c */
f6fb99ca 2670extern unsigned int ext4_count_free(char *bitmap, unsigned numchars);
4fd873c8 2671void ext4_inode_bitmap_csum_set(struct super_block *sb,
41a246d1
DW
2672 struct ext4_group_desc *gdp,
2673 struct buffer_head *bh, int sz);
b83acc77 2674int ext4_inode_bitmap_csum_verify(struct super_block *sb,
41a246d1
DW
2675 struct ext4_group_desc *gdp,
2676 struct buffer_head *bh, int sz);
1df9bde4 2677void ext4_block_bitmap_csum_set(struct super_block *sb,
fa77dcfa 2678 struct ext4_group_desc *gdp,
79f1ba49 2679 struct buffer_head *bh);
82483dfe 2680int ext4_block_bitmap_csum_verify(struct super_block *sb,
fa77dcfa 2681 struct ext4_group_desc *gdp,
79f1ba49 2682 struct buffer_head *bh);
498e5f24 2683
ac27a0ec 2684/* balloc.c */
bd86298e
LC
2685extern void ext4_get_group_no_and_offset(struct super_block *sb,
2686 ext4_fsblk_t blocknr,
2687 ext4_group_t *blockgrpp,
2688 ext4_grpblk_t *offsetp);
2689extern ext4_group_t ext4_get_group_number(struct super_block *sb,
2690 ext4_fsblk_t block);
2691
fd2d4291
AM
2692extern int ext4_bg_has_super(struct super_block *sb, ext4_group_t group);
2693extern unsigned long ext4_bg_num_gdb(struct super_block *sb,
2694 ext4_group_t group);
7061eba7 2695extern ext4_fsblk_t ext4_new_meta_blocks(handle_t *handle, struct inode *inode,
55f020db
AH
2696 ext4_fsblk_t goal,
2697 unsigned int flags,
2698 unsigned long *count,
2699 int *errp);
e7d5f315
TT
2700extern int ext4_claim_free_clusters(struct ext4_sb_info *sbi,
2701 s64 nclusters, unsigned int flags);
5dee5437 2702extern ext4_fsblk_t ext4_count_free_clusters(struct super_block *);
617ba13b 2703extern struct ext4_group_desc * ext4_get_group_desc(struct super_block * sb,
fd2d4291 2704 ext4_group_t block_group,
ac27a0ec 2705 struct buffer_head ** bh);
5354b2af
TT
2706extern struct ext4_group_info *ext4_get_group_info(struct super_block *sb,
2707 ext4_group_t group);
617ba13b 2708extern int ext4_should_retry_alloc(struct super_block *sb, int *retries);
813e5727
TT
2709
2710extern struct buffer_head *ext4_read_block_bitmap_nowait(struct super_block *sb,
cfd73237
AZ
2711 ext4_group_t block_group,
2712 bool ignore_locked);
813e5727
TT
2713extern int ext4_wait_block_bitmap(struct super_block *sb,
2714 ext4_group_t block_group,
2715 struct buffer_head *bh);
2716extern struct buffer_head *ext4_read_block_bitmap(struct super_block *sb,
2717 ext4_group_t block_group);
cff1dfd7
TT
2718extern unsigned ext4_free_clusters_after_init(struct super_block *sb,
2719 ext4_group_t block_group,
2720 struct ext4_group_desc *gdp);
f86186b4 2721ext4_fsblk_t ext4_inode_to_goal_block(struct inode *);
ac27a0ec 2722
5298d4bf 2723#if IS_ENABLED(CONFIG_UNICODE)
1ae98e29 2724extern int ext4_fname_setup_ci_filename(struct inode *dir,
3ae72562 2725 const struct qstr *iname,
1ae98e29 2726 struct ext4_filename *fname);
3ae72562
GKB
2727#endif
2728
b1241c8e 2729/* ext4 encryption related stuff goes here crypto.c */
643fa961 2730#ifdef CONFIG_FS_ENCRYPTION
b1241c8e
RH
2731extern const struct fscrypt_operations ext4_cryptops;
2732
3030b59c
RH
2733int ext4_fname_setup_filename(struct inode *dir, const struct qstr *iname,
2734 int lookup, struct ext4_filename *fname);
d5d0e8c7 2735
3030b59c
RH
2736int ext4_fname_prepare_lookup(struct inode *dir, struct dentry *dentry,
2737 struct ext4_filename *fname);
b7236e21 2738
3030b59c 2739void ext4_fname_free_filename(struct ext4_filename *fname);
3ae72562 2740
72f63f4a
RH
2741int ext4_ioctl_get_encryption_pwsalt(struct file *filp, void __user *arg);
2742
b01531db 2743#else /* !CONFIG_FS_ENCRYPTION */
a7550b30 2744static inline int ext4_fname_setup_filename(struct inode *dir,
b01531db
EB
2745 const struct qstr *iname,
2746 int lookup,
2747 struct ext4_filename *fname)
b7236e21 2748{
1ae98e29 2749 int err = 0;
a7550b30
JK
2750 fname->usr_fname = iname;
2751 fname->disk_name.name = (unsigned char *) iname->name;
2752 fname->disk_name.len = iname->len;
3ae72562 2753
5298d4bf 2754#if IS_ENABLED(CONFIG_UNICODE)
1ae98e29 2755 err = ext4_fname_setup_ci_filename(dir, iname, fname);
3ae72562
GKB
2756#endif
2757
1ae98e29 2758 return err;
b7236e21 2759}
88bd6ccd 2760
b01531db
EB
2761static inline int ext4_fname_prepare_lookup(struct inode *dir,
2762 struct dentry *dentry,
2763 struct ext4_filename *fname)
2764{
2765 return ext4_fname_setup_filename(dir, &dentry->d_name, 1, fname);
2766}
2767
3ae72562
GKB
2768static inline void ext4_fname_free_filename(struct ext4_filename *fname)
2769{
5298d4bf 2770#if IS_ENABLED(CONFIG_UNICODE)
3ae72562
GKB
2771 kfree(fname->cf_name.name);
2772 fname->cf_name.name = NULL;
2773#endif
2774}
72f63f4a
RH
2775
2776static inline int ext4_ioctl_get_encryption_pwsalt(struct file *filp,
2777 void __user *arg)
2778{
2779 return -EOPNOTSUPP;
2780}
b01531db 2781#endif /* !CONFIG_FS_ENCRYPTION */
88bd6ccd 2782
ac27a0ec 2783/* dir.c */
60fd4da3 2784extern int __ext4_check_dir_entry(const char *, unsigned int, struct inode *,
f7c21177 2785 struct file *,
60fd4da3 2786 struct ext4_dir_entry_2 *,
226ba972
TM
2787 struct buffer_head *, char *, int,
2788 unsigned int);
471fbbea 2789#define ext4_check_dir_entry(dir, filp, de, bh, buf, size, offset) \
f7c21177 2790 unlikely(__ext4_check_dir_entry(__func__, __LINE__, (dir), (filp), \
471fbbea 2791 (de), (bh), (buf), (size), (offset)))
617ba13b 2792extern int ext4_htree_store_dirent(struct file *dir_file, __u32 hash,
2f61830a
TT
2793 __u32 minor_hash,
2794 struct ext4_dir_entry_2 *dirent,
a7550b30 2795 struct fscrypt_str *ent_name);
617ba13b 2796extern void ext4_htree_free_dir_info(struct dir_private_info *p);
978fef91
TM
2797extern int ext4_find_dest_de(struct inode *dir, struct inode *inode,
2798 struct buffer_head *bh,
2799 void *buf, int buf_size,
5b643f9c 2800 struct ext4_filename *fname,
978fef91 2801 struct ext4_dir_entry_2 **dest_de);
471fbbea 2802void ext4_insert_dentry(struct inode *dir, struct inode *inode,
1bc0af60
EB
2803 struct ext4_dir_entry_2 *de,
2804 int buf_size,
2805 struct ext4_filename *fname);
978fef91
TM
2806static inline void ext4_update_dx_flag(struct inode *inode)
2807{
f902b216
JK
2808 if (!ext4_has_feature_dir_index(inode->i_sb) &&
2809 ext4_test_inode_flag(inode, EXT4_INODE_INDEX)) {
48a34311
JK
2810 /* ext4_iget() should have caught this... */
2811 WARN_ON_ONCE(ext4_has_feature_metadata_csum(inode->i_sb));
978fef91 2812 ext4_clear_inode_flag(inode, EXT4_INODE_INDEX);
48a34311 2813 }
978fef91 2814}
d6006186 2815static const unsigned char ext4_filetype_table[] = {
65d165d9
TM
2816 DT_UNKNOWN, DT_REG, DT_DIR, DT_CHR, DT_BLK, DT_FIFO, DT_SOCK, DT_LNK
2817};
2818
2819static inline unsigned char get_dtype(struct super_block *sb, int filetype)
2820{
e2b911c5 2821 if (!ext4_has_feature_filetype(sb) || filetype >= EXT4_FT_MAX)
65d165d9
TM
2822 return DT_UNKNOWN;
2823
2824 return ext4_filetype_table[filetype];
2825}
40b163f1
DW
2826extern int ext4_check_all_de(struct inode *dir, struct buffer_head *bh,
2827 void *buf, int buf_size);
ac27a0ec
DK
2828
2829/* fsync.c */
02c24a82 2830extern int ext4_sync_file(struct file *, loff_t, loff_t, int);
ac27a0ec
DK
2831
2832/* hash.c */
b886ee3e
GKB
2833extern int ext4fs_dirhash(const struct inode *dir, const char *name, int len,
2834 struct dx_hash_info *hinfo);
ac27a0ec
DK
2835
2836/* ialloc.c */
8016e29f 2837extern int ext4_mark_inode_used(struct super_block *sb, int ino);
f2d40141 2838extern struct inode *__ext4_new_inode(struct mnt_idmap *, handle_t *,
14f3db55 2839 struct inode *, umode_t,
1139575a 2840 const struct qstr *qstr, __u32 goal,
1b917ed8
TE
2841 uid_t *owner, __u32 i_flags,
2842 int handle_type, unsigned int line_no,
2843 int nblocks);
1139575a 2844
14f3db55 2845#define ext4_new_inode(handle, dir, mode, qstr, goal, owner, i_flags) \
f2d40141 2846 __ext4_new_inode(&nop_mnt_idmap, (handle), (dir), (mode), (qstr), \
14f3db55 2847 (goal), (owner), i_flags, 0, 0, 0)
f2d40141 2848#define ext4_new_inode_start_handle(idmap, dir, mode, qstr, goal, owner, \
1139575a 2849 type, nblocks) \
f2d40141 2850 __ext4_new_inode((idmap), NULL, (dir), (mode), (qstr), (goal), (owner), \
1b917ed8 2851 0, (type), __LINE__, (nblocks))
1139575a
TT
2852
2853
af5bc92d
TT
2854extern void ext4_free_inode(handle_t *, struct inode *);
2855extern struct inode * ext4_orphan_get(struct super_block *, unsigned long);
2856extern unsigned long ext4_count_free_inodes(struct super_block *);
2857extern unsigned long ext4_count_dirs(struct super_block *);
61d08673 2858extern void ext4_mark_bitmap_end(int start_bit, int end_bit, char *bitmap);
bfff6873
LC
2859extern int ext4_init_inode_table(struct super_block *sb,
2860 ext4_group_t group, int barrier);
813e5727 2861extern void ext4_end_bitmap_read(struct buffer_head *bh, int uptodate);
ac27a0ec 2862
6866d7b3 2863/* fast_commit.c */
ce8c59d1 2864int ext4_fc_info_show(struct seq_file *seq, void *v);
6866d7b3 2865void ext4_fc_init(struct super_block *sb, journal_t *journal);
aa75f4d3 2866void ext4_fc_init_inode(struct inode *inode);
a80f7fcf 2867void ext4_fc_track_range(handle_t *handle, struct inode *inode, ext4_lblk_t start,
aa75f4d3 2868 ext4_lblk_t end);
a80f7fcf
HS
2869void __ext4_fc_track_unlink(handle_t *handle, struct inode *inode,
2870 struct dentry *dentry);
2871void __ext4_fc_track_link(handle_t *handle, struct inode *inode,
2872 struct dentry *dentry);
2873void ext4_fc_track_unlink(handle_t *handle, struct dentry *dentry);
2874void ext4_fc_track_link(handle_t *handle, struct dentry *dentry);
8210bb29
HS
2875void __ext4_fc_track_create(handle_t *handle, struct inode *inode,
2876 struct dentry *dentry);
a80f7fcf
HS
2877void ext4_fc_track_create(handle_t *handle, struct dentry *dentry);
2878void ext4_fc_track_inode(handle_t *handle, struct inode *inode);
e85c81ba 2879void ext4_fc_mark_ineligible(struct super_block *sb, int reason, handle_t *handle);
aa75f4d3
HS
2880void ext4_fc_start_update(struct inode *inode);
2881void ext4_fc_stop_update(struct inode *inode);
2882void ext4_fc_del(struct inode *inode);
8016e29f
HS
2883bool ext4_fc_replay_check_excluded(struct super_block *sb, ext4_fsblk_t block);
2884void ext4_fc_replay_cleanup(struct super_block *sb);
aa75f4d3
HS
2885int ext4_fc_commit(journal_t *journal, tid_t commit_tid);
2886int __init ext4_fc_init_dentry_cache(void);
ab047d51 2887void ext4_fc_destroy_dentry_cache(void);
599ea31d
XY
2888int ext4_fc_record_regions(struct super_block *sb, int ino,
2889 ext4_lblk_t lblk, ext4_fsblk_t pblk,
2890 int len, int replay);
aa75f4d3 2891
c9de560d 2892/* mballoc.c */
247dbed8 2893extern const struct seq_operations ext4_mb_seq_groups_ops;
f68f4063 2894extern const struct seq_operations ext4_mb_seq_structs_summary_ops;
a6c75eaf 2895extern int ext4_seq_mb_stats_show(struct seq_file *seq, void *offset);
9d99012f 2896extern int ext4_mb_init(struct super_block *);
c9de560d
AT
2897extern int ext4_mb_release(struct super_block *);
2898extern ext4_fsblk_t ext4_mb_new_blocks(handle_t *,
2899 struct ext4_allocation_request *, int *);
27bc446e 2900extern void ext4_discard_preallocations(struct inode *, unsigned int);
5dabfc78
TT
2901extern int __init ext4_init_mballoc(void);
2902extern void ext4_exit_mballoc(void);
3d392b26
TT
2903extern ext4_group_t ext4_mb_prefetch(struct super_block *sb,
2904 ext4_group_t group,
2905 unsigned int nr, int *cnt);
2906extern void ext4_mb_prefetch_fini(struct super_block *sb, ext4_group_t group,
2907 unsigned int nr);
2908
44338711 2909extern void ext4_free_blocks(handle_t *handle, struct inode *inode,
e6362609
TT
2910 struct buffer_head *bh, ext4_fsblk_t block,
2911 unsigned long count, int flags);
28623c2f
TT
2912extern int ext4_mb_alloc_groupinfo(struct super_block *sb,
2913 ext4_group_t ngroups);
920313a7 2914extern int ext4_mb_add_groupinfo(struct super_block *sb,
5f21b0e6 2915 ext4_group_t i, struct ext4_group_desc *desc);
cc7365df 2916extern int ext4_group_add_blocks(handle_t *handle, struct super_block *sb,
2846e820 2917 ext4_fsblk_t block, unsigned long count);
7360d173 2918extern int ext4_trim_fs(struct super_block *, struct fstrim_range *);
a0154344 2919extern void ext4_process_freed_data(struct super_block *sb, tid_t commit_tid);
8016e29f
HS
2920extern void ext4_mb_mark_bb(struct super_block *sb, ext4_fsblk_t block,
2921 int len, int state);
304749c0
OM
2922static inline bool ext4_mb_cr_expensive(enum criteria cr)
2923{
2924 return cr >= CR_GOAL_LEN_SLOW;
2925}
7360d173 2926
ac27a0ec 2927/* inode.c */
8016e29f
HS
2928void ext4_inode_csum_set(struct inode *inode, struct ext4_inode *raw,
2929 struct ext4_inode_info *ei);
f348c252 2930int ext4_inode_is_fast_symlink(struct inode *inode);
10560082 2931struct buffer_head *ext4_getblk(handle_t *, struct inode *, ext4_lblk_t, int);
1c215028 2932struct buffer_head *ext4_bread(handle_t *, struct inode *, ext4_lblk_t, int);
9699d4f9
TE
2933int ext4_bread_batch(struct inode *inode, ext4_lblk_t block, int bh_count,
2934 bool wait, struct buffer_head **bhs);
705965bd
JK
2935int ext4_get_block_unwritten(struct inode *inode, sector_t iblock,
2936 struct buffer_head *bh_result, int create);
6873fa0d 2937int ext4_get_block(struct inode *inode, sector_t iblock,
705965bd 2938 struct buffer_head *bh_result, int create);
9c3569b5
TM
2939int ext4_da_get_block_prep(struct inode *inode, sector_t iblock,
2940 struct buffer_head *bh, int create);
f19d5870 2941int ext4_walk_page_buffers(handle_t *handle,
188c299e 2942 struct inode *inode,
f19d5870
TM
2943 struct buffer_head *head,
2944 unsigned from,
2945 unsigned to,
2946 int *partial,
188c299e 2947 int (*fn)(handle_t *handle, struct inode *inode,
f19d5870 2948 struct buffer_head *bh));
188c299e 2949int do_journal_get_write_access(handle_t *handle, struct inode *inode,
f19d5870 2950 struct buffer_head *bh);
9c3569b5
TM
2951#define FALL_BACK_TO_NONDELALLOC 1
2952#define CONVERT_INLINE_DATA 2
ac27a0ec 2953
8a363970
TT
2954typedef enum {
2955 EXT4_IGET_NORMAL = 0,
2956 EXT4_IGET_SPECIAL = 0x0001, /* OK to iget a system inode */
63b1e9bc 2957 EXT4_IGET_HANDLE = 0x0002, /* Inode # is from a handle */
b3e6bcb9
TT
2958 EXT4_IGET_BAD = 0x0004, /* Allow to iget a bad inode */
2959 EXT4_IGET_EA_INODE = 0x0008 /* Inode should contain an EA value */
8a363970
TT
2960} ext4_iget_flags;
2961
2962extern struct inode *__ext4_iget(struct super_block *sb, unsigned long ino,
2963 ext4_iget_flags flags, const char *function,
2964 unsigned int line);
2965
2966#define ext4_iget(sb, ino, flags) \
2967 __ext4_iget((sb), (ino), (flags), __func__, __LINE__)
2968
a9185b41 2969extern int ext4_write_inode(struct inode *, struct writeback_control *);
c1632a0f 2970extern int ext4_setattr(struct mnt_idmap *, struct dentry *,
549c7297 2971 struct iattr *);
8434ef1d 2972extern u32 ext4_dio_alignment(struct inode *inode);
b74d24f7 2973extern int ext4_getattr(struct mnt_idmap *, const struct path *,
549c7297 2974 struct kstat *, u32, unsigned int);
0930fcc1
AV
2975extern void ext4_evict_inode(struct inode *);
2976extern void ext4_clear_inode(struct inode *);
b74d24f7 2977extern int ext4_file_getattr(struct mnt_idmap *, const struct path *,
549c7297 2978 struct kstat *, u32, unsigned int);
aa385729 2979extern void ext4_dirty_inode(struct inode *, int);
617ba13b
MC
2980extern int ext4_change_inode_journal_flag(struct inode *, int);
2981extern int ext4_get_inode_loc(struct inode *, struct ext4_iloc *);
8016e29f
HS
2982extern int ext4_get_fc_inode_loc(struct super_block *sb, unsigned long ino,
2983 struct ext4_iloc *iloc);
a361293f 2984extern int ext4_inode_attach_jinode(struct inode *inode);
91ef4caf 2985extern int ext4_can_truncate(struct inode *inode);
2c98eb5e 2986extern int ext4_truncate(struct inode *);
430657b6 2987extern int ext4_break_layouts(struct inode *);
ad5cd4f4 2988extern int ext4_punch_hole(struct file *file, loff_t offset, loff_t length);
043546e4 2989extern void ext4_set_inode_flags(struct inode *, bool init);
ccd2506b 2990extern int ext4_alloc_da_blocks(struct inode *inode);
617ba13b 2991extern void ext4_set_aops(struct inode *inode);
a86c6181 2992extern int ext4_writepage_trans_blocks(struct inode *);
59205c8d 2993extern int ext4_normal_submit_inode_data_buffers(struct jbd2_inode *jinode);
f3bd1f3f 2994extern int ext4_chunk_trans_blocks(struct inode *, int nrblocks);
a87dd18c
LC
2995extern int ext4_zero_partial_blocks(handle_t *handle, struct inode *inode,
2996 loff_t lstart, loff_t lend);
401b25aa 2997extern vm_fault_t ext4_page_mkwrite(struct vm_fault *vmf);
a9e7f447 2998extern qsize_t *ext4_get_reserved_space(struct inode *inode);
040cb378 2999extern int ext4_get_projid(struct inode *inode, kprojid_t *projid);
f456767d 3000extern void ext4_da_release_space(struct inode *inode, int to_free);
5f634d06
AK
3001extern void ext4_da_update_reserve_space(struct inode *inode,
3002 int used, int quota_claim);
53085fac
JK
3003extern int ext4_issue_zeroout(struct inode *inode, ext4_lblk_t lblk,
3004 ext4_fsblk_t pblk, ext4_lblk_t len);
dae1e52c
AG
3005
3006/* indirect.c */
3007extern int ext4_ind_map_blocks(handle_t *handle, struct inode *inode,
3008 struct ext4_map_blocks *map, int flags);
fa55a0ed 3009extern int ext4_ind_trans_blocks(struct inode *inode, int nrblocks);
819c4920 3010extern void ext4_ind_truncate(handle_t *, struct inode *inode);
4f579ae7
LC
3011extern int ext4_ind_remove_space(handle_t *handle, struct inode *inode,
3012 ext4_lblk_t start, ext4_lblk_t end);
ff9893dc 3013
ac27a0ec 3014/* ioctl.c */
5cdd7b2d 3015extern long ext4_ioctl(struct file *, unsigned int, unsigned long);
af5bc92d 3016extern long ext4_compat_ioctl(struct file *, unsigned int, unsigned long);
8782a9ae 3017int ext4_fileattr_set(struct mnt_idmap *idmap,
4db5c2e6
MS
3018 struct dentry *dentry, struct fileattr *fa);
3019int ext4_fileattr_get(struct dentry *dentry, struct fileattr *fa);
8016e29f 3020extern void ext4_reset_inode_seed(struct inode *inode);
827891a3 3021int ext4_update_overhead(struct super_block *sb, bool force);
97524b45 3022int ext4_force_shutdown(struct super_block *sb, u32 flags);
ac27a0ec 3023
c14c6fd5 3024/* migrate.c */
2a43a878 3025extern int ext4_ext_migrate(struct inode *);
0d14b098 3026extern int ext4_ind_migrate(struct inode *inode);
3d0518f4 3027
ac27a0ec 3028/* namei.c */
8016e29f
HS
3029extern int ext4_init_new_dir(handle_t *handle, struct inode *dir,
3030 struct inode *inode);
f036adb3
TT
3031extern int ext4_dirblock_csum_verify(struct inode *inode,
3032 struct buffer_head *bh);
617ba13b 3033extern int ext4_htree_fill_tree(struct file *dir_file, __u32 start_hash,
ac27a0ec 3034 __u32 start_minor_hash, __u32 *next_hash);
5b643f9c
TT
3035extern int ext4_search_dir(struct buffer_head *bh,
3036 char *search_buf,
3037 int buf_size,
3038 struct inode *dir,
3039 struct ext4_filename *fname,
5b643f9c
TT
3040 unsigned int offset,
3041 struct ext4_dir_entry_2 **res_dir);
2fe34d29 3042extern int ext4_generic_delete_entry(struct inode *dir,
05019a9e
TM
3043 struct ext4_dir_entry_2 *de_del,
3044 struct buffer_head *bh,
3045 void *entry_buf,
3046 int buf_size,
3047 int csum_size);
a7550b30 3048extern bool ext4_empty_dir(struct inode *inode);
ac27a0ec
DK
3049
3050/* resize.c */
1d0c3924 3051extern void ext4_kvfree_array_rcu(void *to_free);
617ba13b
MC
3052extern int ext4_group_add(struct super_block *sb,
3053 struct ext4_new_group_data *input);
3054extern int ext4_group_extend(struct super_block *sb,
3055 struct ext4_super_block *es,
3056 ext4_fsblk_t n_blocks_count);
19c5246d 3057extern int ext4_resize_fs(struct super_block *sb, ext4_fsblk_t n_blocks_count);
bbc605cd
LC
3058extern unsigned int ext4_list_backups(struct super_block *sb,
3059 unsigned int *three, unsigned int *five,
3060 unsigned int *seven);
ac27a0ec
DK
3061
3062/* super.c */
fb265c9c 3063extern struct buffer_head *ext4_sb_bread(struct super_block *sb,
67c0f556 3064 sector_t block, blk_opf_t op_flags);
8394a6ab 3065extern struct buffer_head *ext4_sb_bread_unmovable(struct super_block *sb,
3066 sector_t block);
67c0f556 3067extern void ext4_read_bh_nowait(struct buffer_head *bh, blk_opf_t op_flags,
fa491b14 3068 bh_end_io_t *end_io);
67c0f556 3069extern int ext4_read_bh(struct buffer_head *bh, blk_opf_t op_flags,
fa491b14 3070 bh_end_io_t *end_io);
67c0f556 3071extern int ext4_read_bh_lock(struct buffer_head *bh, blk_opf_t op_flags, bool wait);
5df1d412 3072extern void ext4_sb_breadahead_unmovable(struct super_block *sb, sector_t block);
ebd173be 3073extern int ext4_seq_options_show(struct seq_file *seq, void *offset);
952fc18e 3074extern int ext4_calculate_overhead(struct super_block *sb);
bbc605cd
LC
3075extern __le32 ext4_superblock_csum(struct super_block *sb,
3076 struct ext4_super_block *es);
06db49e6 3077extern void ext4_superblock_csum_set(struct super_block *sb);
117fff10
TT
3078extern int ext4_alloc_flex_bg_array(struct super_block *sb,
3079 ext4_group_t ngroup);
722887dd
TT
3080extern const char *ext4_decode_error(struct super_block *sb, int errno,
3081 char nbuf[16]);
db79e6d1
WS
3082extern void ext4_mark_group_bitmap_corrupted(struct super_block *sb,
3083 ext4_group_t block_group,
3084 unsigned int flags);
68228da5
WJ
3085extern unsigned int ext4_num_base_meta_blocks(struct super_block *sb,
3086 ext4_group_t block_group);
e7c96e8e 3087
014c9caa
JK
3088extern __printf(7, 8)
3089void __ext4_error(struct super_block *, const char *, unsigned int, bool,
3090 int, __u64, const char *, ...);
54d3adbc
TT
3091extern __printf(6, 7)
3092void __ext4_error_inode(struct inode *, const char *, unsigned int,
3093 ext4_fsblk_t, int, const char *, ...);
b9075fa9 3094extern __printf(5, 6)
e7c96e8e 3095void __ext4_error_file(struct file *, const char *, unsigned int, ext4_fsblk_t,
b9075fa9 3096 const char *, ...);
c398eda0
TT
3097extern void __ext4_std_error(struct super_block *, const char *,
3098 unsigned int, int);
b9075fa9
JP
3099extern __printf(4, 5)
3100void __ext4_warning(struct super_block *, const char *, unsigned int,
3101 const char *, ...);
b03a2f7e
AD
3102extern __printf(4, 5)
3103void __ext4_warning_inode(const struct inode *inode, const char *function,
3104 unsigned int line, const char *fmt, ...);
b9075fa9 3105extern __printf(3, 4)
e7c96e8e 3106void __ext4_msg(struct super_block *, const char *, const char *, ...);
c5e06d10
JL
3107extern void __dump_mmp_msg(struct super_block *, struct mmp_struct *mmp,
3108 const char *, unsigned int, const char *);
b9075fa9
JP
3109extern __printf(7, 8)
3110void __ext4_grp_locked_error(const char *, unsigned int,
3111 struct super_block *, ext4_group_t,
3112 unsigned long, ext4_fsblk_t,
3113 const char *, ...);
e7c96e8e 3114
b03a2f7e
AD
3115#define EXT4_ERROR_INODE(inode, fmt, a...) \
3116 ext4_error_inode((inode), __func__, __LINE__, 0, (fmt), ## a)
3117
54d3adbc
TT
3118#define EXT4_ERROR_INODE_ERR(inode, err, fmt, a...) \
3119 __ext4_error_inode((inode), __func__, __LINE__, 0, (err), (fmt), ## a)
3120
3121#define ext4_error_inode_block(inode, block, err, fmt, a...) \
3122 __ext4_error_inode((inode), __func__, __LINE__, (block), (err), \
3123 (fmt), ## a)
b03a2f7e
AD
3124
3125#define EXT4_ERROR_FILE(file, block, fmt, a...) \
3126 ext4_error_file((file), __func__, __LINE__, (block), (fmt), ## a)
3127
014c9caa
JK
3128#define ext4_abort(sb, err, fmt, a...) \
3129 __ext4_error((sb), __func__, __LINE__, true, (err), 0, (fmt), ## a)
3130
e7c96e8e
JP
3131#ifdef CONFIG_PRINTK
3132
3133#define ext4_error_inode(inode, func, line, block, fmt, ...) \
54d3adbc
TT
3134 __ext4_error_inode(inode, func, line, block, 0, fmt, ##__VA_ARGS__)
3135#define ext4_error_inode_err(inode, func, line, block, err, fmt, ...) \
3136 __ext4_error_inode((inode), (func), (line), (block), \
3137 (err), (fmt), ##__VA_ARGS__)
e7c96e8e
JP
3138#define ext4_error_file(file, func, line, block, fmt, ...) \
3139 __ext4_error_file(file, func, line, block, fmt, ##__VA_ARGS__)
3140#define ext4_error(sb, fmt, ...) \
014c9caa
JK
3141 __ext4_error((sb), __func__, __LINE__, false, 0, 0, (fmt), \
3142 ##__VA_ARGS__)
54d3adbc 3143#define ext4_error_err(sb, err, fmt, ...) \
014c9caa
JK
3144 __ext4_error((sb), __func__, __LINE__, false, (err), 0, (fmt), \
3145 ##__VA_ARGS__)
e7c96e8e
JP
3146#define ext4_warning(sb, fmt, ...) \
3147 __ext4_warning(sb, __func__, __LINE__, fmt, ##__VA_ARGS__)
b03a2f7e
AD
3148#define ext4_warning_inode(inode, fmt, ...) \
3149 __ext4_warning_inode(inode, __func__, __LINE__, fmt, ##__VA_ARGS__)
e7c96e8e
JP
3150#define ext4_msg(sb, level, fmt, ...) \
3151 __ext4_msg(sb, level, fmt, ##__VA_ARGS__)
3152#define dump_mmp_msg(sb, mmp, msg) \
3153 __dump_mmp_msg(sb, mmp, __func__, __LINE__, msg)
3154#define ext4_grp_locked_error(sb, grp, ino, block, fmt, ...) \
3155 __ext4_grp_locked_error(__func__, __LINE__, sb, grp, ino, block, \
3156 fmt, ##__VA_ARGS__)
3157
3158#else
3159
3160#define ext4_error_inode(inode, func, line, block, fmt, ...) \
3161do { \
3162 no_printk(fmt, ##__VA_ARGS__); \
54d3adbc
TT
3163 __ext4_error_inode(inode, "", 0, block, 0, " "); \
3164} while (0)
3165#define ext4_error_inode_err(inode, func, line, block, err, fmt, ...) \
3166do { \
3167 no_printk(fmt, ##__VA_ARGS__); \
3168 __ext4_error_inode(inode, "", 0, block, err, " "); \
e7c96e8e
JP
3169} while (0)
3170#define ext4_error_file(file, func, line, block, fmt, ...) \
3171do { \
3172 no_printk(fmt, ##__VA_ARGS__); \
3173 __ext4_error_file(file, "", 0, block, " "); \
3174} while (0)
3175#define ext4_error(sb, fmt, ...) \
3176do { \
3177 no_printk(fmt, ##__VA_ARGS__); \
014c9caa 3178 __ext4_error(sb, "", 0, false, 0, 0, " "); \
54d3adbc
TT
3179} while (0)
3180#define ext4_error_err(sb, err, fmt, ...) \
3181do { \
3182 no_printk(fmt, ##__VA_ARGS__); \
014c9caa 3183 __ext4_error(sb, "", 0, false, err, 0, " "); \
e7c96e8e
JP
3184} while (0)
3185#define ext4_warning(sb, fmt, ...) \
3186do { \
3187 no_printk(fmt, ##__VA_ARGS__); \
3188 __ext4_warning(sb, "", 0, " "); \
3189} while (0)
b03a2f7e
AD
3190#define ext4_warning_inode(inode, fmt, ...) \
3191do { \
3192 no_printk(fmt, ##__VA_ARGS__); \
3193 __ext4_warning_inode(inode, "", 0, " "); \
3194} while (0)
e7c96e8e
JP
3195#define ext4_msg(sb, level, fmt, ...) \
3196do { \
3197 no_printk(fmt, ##__VA_ARGS__); \
3198 __ext4_msg(sb, "", " "); \
3199} while (0)
3200#define dump_mmp_msg(sb, mmp, msg) \
3201 __dump_mmp_msg(sb, mmp, "", 0, "")
3202#define ext4_grp_locked_error(sb, grp, ino, block, fmt, ...) \
3203do { \
3204 no_printk(fmt, ##__VA_ARGS__); \
3205 __ext4_grp_locked_error("", 0, sb, grp, ino, block, " "); \
3206} while (0)
3207
3208#endif
3209
8fadc143
AR
3210extern ext4_fsblk_t ext4_block_bitmap(struct super_block *sb,
3211 struct ext4_group_desc *bg);
3212extern ext4_fsblk_t ext4_inode_bitmap(struct super_block *sb,
3213 struct ext4_group_desc *bg);
3214extern ext4_fsblk_t ext4_inode_table(struct super_block *sb,
3215 struct ext4_group_desc *bg);
021b65bb
TT
3216extern __u32 ext4_free_group_clusters(struct super_block *sb,
3217 struct ext4_group_desc *bg);
560671a0
AK
3218extern __u32 ext4_free_inodes_count(struct super_block *sb,
3219 struct ext4_group_desc *bg);
3220extern __u32 ext4_used_dirs_count(struct super_block *sb,
3221 struct ext4_group_desc *bg);
3222extern __u32 ext4_itable_unused_count(struct super_block *sb,
3223 struct ext4_group_desc *bg);
8fadc143
AR
3224extern void ext4_block_bitmap_set(struct super_block *sb,
3225 struct ext4_group_desc *bg, ext4_fsblk_t blk);
3226extern void ext4_inode_bitmap_set(struct super_block *sb,
3227 struct ext4_group_desc *bg, ext4_fsblk_t blk);
3228extern void ext4_inode_table_set(struct super_block *sb,
3229 struct ext4_group_desc *bg, ext4_fsblk_t blk);
021b65bb
TT
3230extern void ext4_free_group_clusters_set(struct super_block *sb,
3231 struct ext4_group_desc *bg,
3232 __u32 count);
560671a0
AK
3233extern void ext4_free_inodes_set(struct super_block *sb,
3234 struct ext4_group_desc *bg, __u32 count);
3235extern void ext4_used_dirs_set(struct super_block *sb,
3236 struct ext4_group_desc *bg, __u32 count);
3237extern void ext4_itable_unused_set(struct super_block *sb,
3238 struct ext4_group_desc *bg, __u32 count);
feb0ab32 3239extern int ext4_group_desc_csum_verify(struct super_block *sb, __u32 group,
bb23c20a 3240 struct ext4_group_desc *gdp);
feb0ab32
DW
3241extern void ext4_group_desc_csum_set(struct super_block *sb, __u32 group,
3242 struct ext4_group_desc *gdp);
7f511862
TT
3243extern int ext4_register_li_request(struct super_block *sb,
3244 ext4_group_t first_not_zeroed);
bd81d8ee 3245
9aa5d32b
DM
3246static inline int ext4_has_metadata_csum(struct super_block *sb)
3247{
e2b911c5 3248 WARN_ON_ONCE(ext4_has_feature_metadata_csum(sb) &&
9aa5d32b
DM
3249 !EXT4_SB(sb)->s_chksum_driver);
3250
dec214d0
TE
3251 return ext4_has_feature_metadata_csum(sb) &&
3252 (EXT4_SB(sb)->s_chksum_driver != NULL);
9aa5d32b 3253}
dec214d0
TE
3254
3255static inline int ext4_has_group_desc_csum(struct super_block *sb)
3256{
3257 return ext4_has_feature_gdt_csum(sb) || ext4_has_metadata_csum(sb);
3258}
3259
81e8c3c5
PM
3260#define ext4_read_incompat_64bit_val(es, name) \
3261 (((es)->s_feature_incompat & cpu_to_le32(EXT4_FEATURE_INCOMPAT_64BIT) \
3262 ? (ext4_fsblk_t)le32_to_cpu(es->name##_hi) << 32 : 0) | \
3263 le32_to_cpu(es->name##_lo))
3264
bd81d8ee
LV
3265static inline ext4_fsblk_t ext4_blocks_count(struct ext4_super_block *es)
3266{
81e8c3c5 3267 return ext4_read_incompat_64bit_val(es, s_blocks_count);
bd81d8ee
LV
3268}
3269
3270static inline ext4_fsblk_t ext4_r_blocks_count(struct ext4_super_block *es)
3271{
81e8c3c5 3272 return ext4_read_incompat_64bit_val(es, s_r_blocks_count);
bd81d8ee
LV
3273}
3274
3275static inline ext4_fsblk_t ext4_free_blocks_count(struct ext4_super_block *es)
3276{
81e8c3c5 3277 return ext4_read_incompat_64bit_val(es, s_free_blocks_count);
bd81d8ee
LV
3278}
3279
3280static inline void ext4_blocks_count_set(struct ext4_super_block *es,
3281 ext4_fsblk_t blk)
3282{
6bc9feff 3283 es->s_blocks_count_lo = cpu_to_le32((u32)blk);
bd81d8ee
LV
3284 es->s_blocks_count_hi = cpu_to_le32(blk >> 32);
3285}
3286
3287static inline void ext4_free_blocks_count_set(struct ext4_super_block *es,
3288 ext4_fsblk_t blk)
3289{
308ba3ec 3290 es->s_free_blocks_count_lo = cpu_to_le32((u32)blk);
bd81d8ee
LV
3291 es->s_free_blocks_count_hi = cpu_to_le32(blk >> 32);
3292}
3293
3294static inline void ext4_r_blocks_count_set(struct ext4_super_block *es,
3295 ext4_fsblk_t blk)
3296{
308ba3ec 3297 es->s_r_blocks_count_lo = cpu_to_le32((u32)blk);
bd81d8ee
LV
3298 es->s_r_blocks_count_hi = cpu_to_le32(blk >> 32);
3299}
3300
e08ac99f
AB
3301static inline loff_t ext4_isize(struct super_block *sb,
3302 struct ext4_inode *raw_inode)
a48380f7 3303{
e08ac99f
AB
3304 if (ext4_has_feature_largedir(sb) ||
3305 S_ISREG(le16_to_cpu(raw_inode->i_mode)))
06a279d6
TT
3306 return ((loff_t)le32_to_cpu(raw_inode->i_size_high) << 32) |
3307 le32_to_cpu(raw_inode->i_size_lo);
e08ac99f
AB
3308
3309 return (loff_t) le32_to_cpu(raw_inode->i_size_lo);
a48380f7 3310}
bd81d8ee 3311
a48380f7
AK
3312static inline void ext4_isize_set(struct ext4_inode *raw_inode, loff_t i_size)
3313{
3314 raw_inode->i_size_lo = cpu_to_le32(i_size);
3315 raw_inode->i_size_high = cpu_to_le32(i_size >> 32);
3316}
ac27a0ec 3317
8df9675f
TT
3318/*
3319 * Reading s_groups_count requires using smp_rmb() afterwards. See
3320 * the locking protocol documented in the comments of ext4_group_add()
3321 * in resize.c
3322 */
3323static inline ext4_group_t ext4_get_groups_count(struct super_block *sb)
3324{
3325 ext4_group_t ngroups = EXT4_SB(sb)->s_groups_count;
3326
3327 smp_rmb();
3328 return ngroups;
3329}
c9de560d 3330
772cb7c8
JS
3331static inline ext4_group_t ext4_flex_group(struct ext4_sb_info *sbi,
3332 ext4_group_t block_group)
3333{
3334 return block_group >> sbi->s_log_groups_per_flex;
3335}
3336
3337static inline unsigned int ext4_flex_bg_size(struct ext4_sb_info *sbi)
3338{
3339 return 1 << sbi->s_log_groups_per_flex;
3340}
3341
617ba13b 3342#define ext4_std_error(sb, errno) \
ac27a0ec
DK
3343do { \
3344 if ((errno)) \
c398eda0 3345 __ext4_std_error((sb), __func__, __LINE__, (errno)); \
ac27a0ec
DK
3346} while (0)
3347
a30d542a 3348#ifdef CONFIG_SMP
df55c99d
TT
3349/* Each CPU can accumulate percpu_counter_batch clusters in their local
3350 * counters. So we need to make sure we have free clusters more
179f7ebf 3351 * than percpu_counter_batch * nr_cpu_ids. Also add a window of 4 times.
a30d542a 3352 */
df55c99d 3353#define EXT4_FREECLUSTERS_WATERMARK (4 * (percpu_counter_batch * nr_cpu_ids))
a30d542a 3354#else
df55c99d 3355#define EXT4_FREECLUSTERS_WATERMARK 0
a30d542a
AK
3356#endif
3357
f340b3d9 3358/* Update i_disksize. Requires i_rwsem to avoid races with truncate */
cf17fea6
AK
3359static inline void ext4_update_i_disksize(struct inode *inode, loff_t newsize)
3360{
90e775b7 3361 WARN_ON_ONCE(S_ISREG(inode->i_mode) &&
5955102c 3362 !inode_is_locked(inode));
90e775b7
JK
3363 down_write(&EXT4_I(inode)->i_data_sem);
3364 if (newsize > EXT4_I(inode)->i_disksize)
35df4299 3365 WRITE_ONCE(EXT4_I(inode)->i_disksize, newsize);
90e775b7
JK
3366 up_write(&EXT4_I(inode)->i_data_sem);
3367}
3368
f340b3d9 3369/* Update i_size, i_disksize. Requires i_rwsem to avoid races with truncate */
4631dbf6
DM
3370static inline int ext4_update_inode_size(struct inode *inode, loff_t newsize)
3371{
3372 int changed = 0;
3373
3374 if (newsize > inode->i_size) {
3375 i_size_write(inode, newsize);
3376 changed = 1;
3377 }
3378 if (newsize > EXT4_I(inode)->i_disksize) {
3379 ext4_update_i_disksize(inode, newsize);
3380 changed |= 2;
3381 }
3382 return changed;
3383}
3384
01127848
JK
3385int ext4_update_disksize_before_punch(struct inode *inode, loff_t offset,
3386 loff_t len);
3387
5d1b1b3f
AK
3388struct ext4_group_info {
3389 unsigned long bb_state;
addd752c
CX
3390#ifdef AGGRESSIVE_CHECK
3391 unsigned long bb_check_counter;
3392#endif
5d1b1b3f 3393 struct rb_root bb_free_root;
a36b4498
ES
3394 ext4_grpblk_t bb_first_free; /* first free block */
3395 ext4_grpblk_t bb_free; /* total free blocks */
3396 ext4_grpblk_t bb_fragments; /* nr of freespace fragments */
83e80a6e
JK
3397 int bb_avg_fragment_size_order; /* order of average
3398 fragment in BG */
8a57d9d6 3399 ext4_grpblk_t bb_largest_free_order;/* order of largest frag in BG */
196e402a 3400 ext4_group_t bb_group; /* Group number */
5d1b1b3f
AK
3401 struct list_head bb_prealloc_list;
3402#ifdef DOUBLE_CHECK
3403 void *bb_bitmap;
3404#endif
3405 struct rw_semaphore alloc_sem;
83e80a6e 3406 struct list_head bb_avg_fragment_size_node;
196e402a 3407 struct list_head bb_largest_free_order_node;
a36b4498
ES
3408 ext4_grpblk_t bb_counters[]; /* Nr of free power-of-two-block
3409 * regions, index is order.
3410 * bb_counters[3] = 5 means
3411 * 5 free 8-block regions. */
5d1b1b3f
AK
3412};
3413
3d56b8d2
TM
3414#define EXT4_GROUP_INFO_NEED_INIT_BIT 0
3415#define EXT4_GROUP_INFO_WAS_TRIMMED_BIT 1
163a203d 3416#define EXT4_GROUP_INFO_BBITMAP_CORRUPT_BIT 2
87a39389 3417#define EXT4_GROUP_INFO_IBITMAP_CORRUPT_BIT 3
db79e6d1
WS
3418#define EXT4_GROUP_INFO_BBITMAP_CORRUPT \
3419 (1 << EXT4_GROUP_INFO_BBITMAP_CORRUPT_BIT)
3420#define EXT4_GROUP_INFO_IBITMAP_CORRUPT \
3421 (1 << EXT4_GROUP_INFO_IBITMAP_CORRUPT_BIT)
cfd73237 3422#define EXT4_GROUP_INFO_BBITMAP_READ_BIT 4
5d1b1b3f
AK
3423
3424#define EXT4_MB_GRP_NEED_INIT(grp) \
3425 (test_bit(EXT4_GROUP_INFO_NEED_INIT_BIT, &((grp)->bb_state)))
163a203d
DW
3426#define EXT4_MB_GRP_BBITMAP_CORRUPT(grp) \
3427 (test_bit(EXT4_GROUP_INFO_BBITMAP_CORRUPT_BIT, &((grp)->bb_state)))
87a39389
DW
3428#define EXT4_MB_GRP_IBITMAP_CORRUPT(grp) \
3429 (test_bit(EXT4_GROUP_INFO_IBITMAP_CORRUPT_BIT, &((grp)->bb_state)))
5d1b1b3f 3430
3d56b8d2
TM
3431#define EXT4_MB_GRP_WAS_TRIMMED(grp) \
3432 (test_bit(EXT4_GROUP_INFO_WAS_TRIMMED_BIT, &((grp)->bb_state)))
3433#define EXT4_MB_GRP_SET_TRIMMED(grp) \
3434 (set_bit(EXT4_GROUP_INFO_WAS_TRIMMED_BIT, &((grp)->bb_state)))
3435#define EXT4_MB_GRP_CLEAR_TRIMMED(grp) \
3436 (clear_bit(EXT4_GROUP_INFO_WAS_TRIMMED_BIT, &((grp)->bb_state)))
cfd73237
AZ
3437#define EXT4_MB_GRP_TEST_AND_SET_READ(grp) \
3438 (test_and_set_bit(EXT4_GROUP_INFO_BBITMAP_READ_BIT, &((grp)->bb_state)))
3d56b8d2 3439
50797481
TT
3440#define EXT4_MAX_CONTENTION 8
3441#define EXT4_CONTENTION_THRESHOLD 2
3442
955ce5f5
AK
3443static inline spinlock_t *ext4_group_lock_ptr(struct super_block *sb,
3444 ext4_group_t group)
5d1b1b3f 3445{
955ce5f5
AK
3446 return bgl_lock_ptr(EXT4_SB(sb)->s_blockgroup_lock, group);
3447}
5d1b1b3f 3448
50797481
TT
3449/*
3450 * Returns true if the filesystem is busy enough that attempts to
3451 * access the block group locks has run into contention.
3452 */
3453static inline int ext4_fs_is_busy(struct ext4_sb_info *sbi)
3454{
3455 return (atomic_read(&sbi->s_lock_busy) > EXT4_CONTENTION_THRESHOLD);
3456}
3457
955ce5f5
AK
3458static inline void ext4_lock_group(struct super_block *sb, ext4_group_t group)
3459{
50797481
TT
3460 spinlock_t *lock = ext4_group_lock_ptr(sb, group);
3461 if (spin_trylock(lock))
3462 /*
3463 * We're able to grab the lock right away, so drop the
3464 * lock contention counter.
3465 */
3466 atomic_add_unless(&EXT4_SB(sb)->s_lock_busy, -1, 0);
3467 else {
3468 /*
3469 * The lock is busy, so bump the contention counter,
3470 * and then wait on the spin lock.
3471 */
3472 atomic_add_unless(&EXT4_SB(sb)->s_lock_busy, 1,
3473 EXT4_MAX_CONTENTION);
3474 spin_lock(lock);
3475 }
5d1b1b3f
AK
3476}
3477
3478static inline void ext4_unlock_group(struct super_block *sb,
3479 ext4_group_t group)
3480{
955ce5f5 3481 spin_unlock(ext4_group_lock_ptr(sb, group));
5d1b1b3f
AK
3482}
3483
f177ee08
RA
3484#ifdef CONFIG_QUOTA
3485static inline bool ext4_quota_capable(struct super_block *sb)
3486{
3487 return (test_opt(sb, QUOTA) || ext4_has_feature_quota(sb));
3488}
3489
3490static inline bool ext4_is_quota_journalled(struct super_block *sb)
3491{
3492 struct ext4_sb_info *sbi = EXT4_SB(sb);
3493
3494 return (ext4_has_feature_quota(sb) ||
3495 sbi->s_qf_names[USRQUOTA] || sbi->s_qf_names[GRPQUOTA]);
3496}
25c6d98f 3497int ext4_enable_quotas(struct super_block *sb);
f177ee08
RA
3498#endif
3499
1f7d1e77
TT
3500/*
3501 * Block validity checking
3502 */
3503#define ext4_check_indirect_blockref(inode, bh) \
3504 ext4_check_blockref(__func__, __LINE__, inode, \
3505 (__le32 *)(bh)->b_data, \
3506 EXT4_ADDR_PER_BLOCK((inode)->i_sb))
3507
3508#define ext4_ind_check_inode(inode) \
3509 ext4_check_blockref(__func__, __LINE__, inode, \
3510 EXT4_I(inode)->i_data, \
3511 EXT4_NDIR_BLOCKS)
3512
ac27a0ec
DK
3513/*
3514 * Inodes and files operations
3515 */
3516
3517/* dir.c */
617ba13b 3518extern const struct file_operations ext4_dir_operations;
ac27a0ec
DK
3519
3520/* file.c */
c5ef1c42 3521extern const struct inode_operations ext4_file_inode_operations;
617ba13b 3522extern const struct file_operations ext4_file_operations;
e0d10bfa 3523extern loff_t ext4_llseek(struct file *file, loff_t offset, int origin);
ac27a0ec 3524
95eaefbd 3525/* inline.c */
95eaefbd
TT
3526extern int ext4_get_max_inline_size(struct inode *inode);
3527extern int ext4_find_inline_data_nolock(struct inode *inode);
95eaefbd
TT
3528extern int ext4_destroy_inline_data(handle_t *handle, struct inode *inode);
3529
3edde93e 3530int ext4_readpage_inline(struct inode *inode, struct folio *folio);
95eaefbd
TT
3531extern int ext4_try_to_write_inline_data(struct address_space *mapping,
3532 struct inode *inode,
3533 loff_t pos, unsigned len,
95eaefbd 3534 struct page **pagep);
d19500da
RH
3535int ext4_write_inline_data_end(struct inode *inode, loff_t pos, unsigned len,
3536 unsigned copied, struct folio *folio);
95eaefbd
TT
3537extern int ext4_da_write_inline_data_begin(struct address_space *mapping,
3538 struct inode *inode,
3539 loff_t pos, unsigned len,
95eaefbd
TT
3540 struct page **pagep,
3541 void **fsdata);
5b643f9c
TT
3542extern int ext4_try_add_inline_entry(handle_t *handle,
3543 struct ext4_filename *fname,
56a04915 3544 struct inode *dir, struct inode *inode);
95eaefbd
TT
3545extern int ext4_try_create_inline_dir(handle_t *handle,
3546 struct inode *parent,
3547 struct inode *inode);
3548extern int ext4_read_inline_dir(struct file *filp,
725bebb2 3549 struct dir_context *ctx,
95eaefbd 3550 int *has_inline_data);
7633b08b
TT
3551extern int ext4_inlinedir_to_tree(struct file *dir_file,
3552 struct inode *dir, ext4_lblk_t block,
3553 struct dx_hash_info *hinfo,
3554 __u32 start_hash, __u32 start_minor_hash,
3555 int *has_inline_data);
95eaefbd 3556extern struct buffer_head *ext4_find_inline_entry(struct inode *dir,
5b643f9c 3557 struct ext4_filename *fname,
95eaefbd
TT
3558 struct ext4_dir_entry_2 **res_dir,
3559 int *has_inline_data);
3560extern int ext4_delete_inline_entry(handle_t *handle,
3561 struct inode *dir,
3562 struct ext4_dir_entry_2 *de_del,
3563 struct buffer_head *bh,
3564 int *has_inline_data);
a7550b30 3565extern bool empty_inline_dir(struct inode *dir, int *has_inline_data);
95eaefbd
TT
3566extern struct buffer_head *ext4_get_first_inline_block(struct inode *inode,
3567 struct ext4_dir_entry_2 **parent_de,
3568 int *retval);
5a57bca9 3569extern void *ext4_read_inline_link(struct inode *inode);
7046ae35
AG
3570
3571struct iomap;
3572extern int ext4_inline_data_iomap(struct inode *inode, struct iomap *iomap);
3573
01daf945 3574extern int ext4_inline_data_truncate(struct inode *inode, int *has_inline);
95eaefbd
TT
3575
3576extern int ext4_convert_inline_data(struct inode *inode);
3577
83447ccb
ZL
3578static inline int ext4_has_inline_data(struct inode *inode)
3579{
3580 return ext4_test_inode_flag(inode, EXT4_INODE_INLINE_DATA) &&
3581 EXT4_I(inode)->i_inline_off;
3582}
3583
ac27a0ec 3584/* namei.c */
c5ef1c42
AV
3585extern const struct inode_operations ext4_dir_inode_operations;
3586extern const struct inode_operations ext4_special_inode_operations;
596397b7 3587extern struct dentry *ext4_get_parent(struct dentry *child);
a774f9c2
TM
3588extern struct ext4_dir_entry_2 *ext4_init_dot_dotdot(struct inode *inode,
3589 struct ext4_dir_entry_2 *de,
3590 int blocksize, int csum_size,
3591 unsigned int parent_ino, int dotdot_real_len);
ddce3b94
TT
3592extern void ext4_initialize_dirent_tail(struct buffer_head *bh,
3593 unsigned int blocksize);
f036adb3
TT
3594extern int ext4_handle_dirty_dirblock(handle_t *handle, struct inode *inode,
3595 struct buffer_head *bh);
4c0d5778
EB
3596extern int __ext4_unlink(struct inode *dir, const struct qstr *d_name,
3597 struct inode *inode, struct dentry *dentry);
8016e29f
HS
3598extern int __ext4_link(struct inode *dir, struct inode *inode,
3599 struct dentry *dentry);
b886ee3e 3600
8af0f082 3601#define S_SHIFT 12
381cebfe 3602static const unsigned char ext4_type_by_mode[(S_IFMT >> S_SHIFT) + 1] = {
8af0f082
TM
3603 [S_IFREG >> S_SHIFT] = EXT4_FT_REG_FILE,
3604 [S_IFDIR >> S_SHIFT] = EXT4_FT_DIR,
3605 [S_IFCHR >> S_SHIFT] = EXT4_FT_CHRDEV,
3606 [S_IFBLK >> S_SHIFT] = EXT4_FT_BLKDEV,
3607 [S_IFIFO >> S_SHIFT] = EXT4_FT_FIFO,
3608 [S_IFSOCK >> S_SHIFT] = EXT4_FT_SOCK,
3609 [S_IFLNK >> S_SHIFT] = EXT4_FT_SYMLINK,
3610};
3611
3612static inline void ext4_set_de_type(struct super_block *sb,
3613 struct ext4_dir_entry_2 *de,
3614 umode_t mode) {
e2b911c5 3615 if (ext4_has_feature_filetype(sb))
8af0f082
TM
3616 de->file_type = ext4_type_by_mode[(mode & S_IFMT)>>S_SHIFT];
3617}
3618
f64e02fe 3619/* readpages.c */
a07f624b 3620extern int ext4_mpage_readpages(struct inode *inode,
c0be8e6f 3621 struct readahead_control *rac, struct folio *folio);
22cfe4b4
EB
3622extern int __init ext4_init_post_read_processing(void);
3623extern void ext4_exit_post_read_processing(void);
ac27a0ec
DK
3624
3625/* symlink.c */
a7a67e8a 3626extern const struct inode_operations ext4_encrypted_symlink_inode_operations;
c5ef1c42
AV
3627extern const struct inode_operations ext4_symlink_inode_operations;
3628extern const struct inode_operations ext4_fast_symlink_inode_operations;
ac27a0ec 3629
b5799018 3630/* sysfs.c */
d578b994 3631extern void ext4_notify_error_sysfs(struct ext4_sb_info *sbi);
b5799018 3632extern int ext4_register_sysfs(struct super_block *sb);
ebd173be 3633extern void ext4_unregister_sysfs(struct super_block *sb);
b5799018
TT
3634extern int __init ext4_init_sysfs(void);
3635extern void ext4_exit_sysfs(void);
3636
6fd058f7
TT
3637/* block_validity */
3638extern void ext4_release_system_zone(struct super_block *sb);
3639extern int ext4_setup_system_zone(struct super_block *sb);
5dabfc78
TT
3640extern int __init ext4_init_system_zone(void);
3641extern void ext4_exit_system_zone(void);
ce9f24cc
JK
3642extern int ext4_inode_block_valid(struct inode *inode,
3643 ext4_fsblk_t start_blk,
3644 unsigned int count);
1f7d1e77
TT
3645extern int ext4_check_blockref(const char *, unsigned int,
3646 struct inode *, __le32 *, unsigned int);
6bc6c2bd
RH
3647extern int ext4_sb_block_valid(struct super_block *sb, struct inode *inode,
3648 ext4_fsblk_t start_blk, unsigned int count);
3649
6fd058f7 3650
a86c6181 3651/* extents.c */
4a092d73
TT
3652struct ext4_ext_path;
3653struct ext4_extent;
3654
d7b2a00c
ZL
3655/*
3656 * Maximum number of logical blocks in a file; ext4_extent's ee_block is
3657 * __le32.
3658 */
3659#define EXT_MAX_BLOCKS 0xffffffff
3660
4209ae12 3661extern void ext4_ext_tree_init(handle_t *handle, struct inode *inode);
fffb2739 3662extern int ext4_ext_index_trans_blocks(struct inode *inode, int extents);
e35fd660
TT
3663extern int ext4_ext_map_blocks(handle_t *handle, struct inode *inode,
3664 struct ext4_map_blocks *map, int flags);
d0abb36d 3665extern int ext4_ext_truncate(handle_t *, struct inode *);
26a4c0c6
TT
3666extern int ext4_ext_remove_space(struct inode *inode, ext4_lblk_t start,
3667 ext4_lblk_t end);
a86c6181
AT
3668extern void ext4_ext_init(struct super_block *);
3669extern void ext4_ext_release(struct super_block *);
2fe17c10 3670extern long ext4_fallocate(struct file *file, int mode, loff_t offset,
a2df2a63 3671 loff_t len);
6b523df4
JK
3672extern int ext4_convert_unwritten_extents(handle_t *handle, struct inode *inode,
3673 loff_t offset, ssize_t len);
a00713ea
RH
3674extern int ext4_convert_unwritten_io_end_vec(handle_t *handle,
3675 ext4_io_end_t *io_end);
e35fd660
TT
3676extern int ext4_map_blocks(handle_t *handle, struct inode *inode,
3677 struct ext4_map_blocks *map, int flags);
4a092d73
TT
3678extern int ext4_ext_calc_credits_for_single_extent(struct inode *inode,
3679 int num,
3680 struct ext4_ext_path *path);
4a092d73 3681extern int ext4_ext_insert_extent(handle_t *, struct inode *,
dfe50809 3682 struct ext4_ext_path **,
4a092d73 3683 struct ext4_extent *, int);
ed8a1a76
TT
3684extern struct ext4_ext_path *ext4_find_extent(struct inode *, ext4_lblk_t,
3685 struct ext4_ext_path **,
3686 int flags);
7ff5fdda 3687extern void ext4_free_ext_path(struct ext4_ext_path *);
4a092d73 3688extern int ext4_ext_check_inode(struct inode *inode);
fcf6b1b7 3689extern ext4_lblk_t ext4_ext_next_allocated_block(struct ext4_ext_path *path);
3a06d778
AK
3690extern int ext4_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
3691 __u64 start, __u64 len);
bb5835ed
TT
3692extern int ext4_get_es_cache(struct inode *inode,
3693 struct fiemap_extent_info *fieinfo,
3694 __u64 start, __u64 len);
7869a4a6 3695extern int ext4_ext_precache(struct inode *inode);
fcf6b1b7
DM
3696extern int ext4_swap_extents(handle_t *handle, struct inode *inode1,
3697 struct inode *inode2, ext4_lblk_t lblk1,
3698 ext4_lblk_t lblk2, ext4_lblk_t count,
3699 int mark_unwritten,int *err);
0b02f4c0 3700extern int ext4_clu_mapped(struct inode *inode, ext4_lblk_t lclu);
a4130367 3701extern int ext4_datasem_ensure_credits(handle_t *handle, struct inode *inode,
83448bdf
JK
3702 int check_cred, int restart_cred,
3703 int revoke_cred);
8016e29f
HS
3704extern void ext4_ext_replay_shrink_inode(struct inode *inode, ext4_lblk_t end);
3705extern int ext4_ext_replay_set_iblocks(struct inode *inode);
3706extern int ext4_ext_replay_update_ex(struct inode *inode, ext4_lblk_t start,
3707 int len, int unwritten, ext4_fsblk_t pblk);
3708extern int ext4_ext_clear_bb(struct inode *inode);
a4130367 3709
4a092d73 3710
748de673 3711/* move_extent.c */
393d1d1d
DTB
3712extern void ext4_double_down_write_data_sem(struct inode *first,
3713 struct inode *second);
3714extern void ext4_double_up_write_data_sem(struct inode *orig_inode,
3715 struct inode *donor_inode);
748de673
AF
3716extern int ext4_move_extents(struct file *o_filp, struct file *d_filp,
3717 __u64 start_orig, __u64 start_donor,
3718 __u64 len, __u64 *moved_len);
3719
bd2d0210 3720/* page-io.c */
5dabfc78
TT
3721extern int __init ext4_init_pageio(void);
3722extern void ext4_exit_pageio(void);
bd2d0210 3723extern ext4_io_end_t *ext4_init_io_end(struct inode *inode, gfp_t flags);
97a851ed
JK
3724extern ext4_io_end_t *ext4_get_io_end(ext4_io_end_t *io_end);
3725extern int ext4_put_io_end(ext4_io_end_t *io_end);
3726extern void ext4_put_io_end_defer(ext4_io_end_t *io_end);
3727extern void ext4_io_submit_init(struct ext4_io_submit *io,
3728 struct writeback_control *wbc);
2e8fa54e 3729extern void ext4_end_io_rsv_work(struct work_struct *work);
bd2d0210 3730extern void ext4_io_submit(struct ext4_io_submit *io);
e8d6062c
MW
3731int ext4_bio_write_folio(struct ext4_io_submit *io, struct folio *page,
3732 size_t len);
c8cc8816
RH
3733extern struct ext4_io_end_vec *ext4_alloc_io_end_vec(ext4_io_end_t *io_end);
3734extern struct ext4_io_end_vec *ext4_last_io_end_vec(ext4_io_end_t *io_end);
3a06d778 3735
c5e06d10
JL
3736/* mmp.c */
3737extern int ext4_multi_mount_protect(struct super_block *, ext4_fsblk_t);
3738
618f0031
PS
3739/* mmp.c */
3740extern void ext4_stop_mmpd(struct ext4_sb_info *sbi);
3741
c93d8f88
EB
3742/* verity.c */
3743extern const struct fsverity_operations ext4_verityops;
3744
25c6d98f
JK
3745/* orphan.c */
3746extern int ext4_orphan_add(handle_t *, struct inode *);
3747extern int ext4_orphan_del(handle_t *, struct inode *);
3748extern void ext4_orphan_cleanup(struct super_block *sb,
3749 struct ext4_super_block *es);
02f310fc
JK
3750extern void ext4_release_orphan_info(struct super_block *sb);
3751extern int ext4_init_orphan_info(struct super_block *sb);
3752extern int ext4_orphan_file_empty(struct super_block *sb);
3753extern void ext4_orphan_file_block_trigger(
3754 struct jbd2_buffer_trigger_type *triggers,
3755 struct buffer_head *bh,
3756 void *data, size_t size);
25c6d98f 3757
2ccb5fb9 3758/*
48fc7f7e 3759 * Add new method to test whether block and inode bitmaps are properly
2ccb5fb9
AK
3760 * initialized. With uninit_bg reading the block from disk is not enough
3761 * to mark the bitmap uptodate. We need to also zero-out the bitmap
3762 */
3763#define BH_BITMAP_UPTODATE BH_JBDPrivateStart
3764
3765static inline int bitmap_uptodate(struct buffer_head *bh)
3766{
3767 return (buffer_uptodate(bh) &&
3768 test_bit(BH_BITMAP_UPTODATE, &(bh)->b_state));
3769}
3770static inline void set_bitmap_uptodate(struct buffer_head *bh)
3771{
3772 set_bit(BH_BITMAP_UPTODATE, &(bh)->b_state);
3773}
3774
731eb1a0
AM
3775#define in_range(b, first, len) ((b) >= (first) && (b) <= (first) + (len) - 1)
3776
e9e3bcec
ES
3777/* For ioend & aio unwritten conversion wait queues */
3778#define EXT4_WQ_HASH_SZ 37
3779#define ext4_ioend_wq(v) (&ext4__ioend_wq[((unsigned long)(v)) %\
3780 EXT4_WQ_HASH_SZ])
e9e3bcec 3781extern wait_queue_head_t ext4__ioend_wq[EXT4_WQ_HASH_SZ];
e9e3bcec 3782
8f82f840 3783extern int ext4_resize_begin(struct super_block *sb);
827891a3 3784extern int ext4_resize_end(struct super_block *sb, bool update_backups);
8f82f840 3785
74c66bcb
JK
3786static inline void ext4_set_io_unwritten_flag(struct inode *inode,
3787 struct ext4_io_end *io_end)
3788{
3789 if (!(io_end->flag & EXT4_IO_END_UNWRITTEN)) {
3790 io_end->flag |= EXT4_IO_END_UNWRITTEN;
3791 atomic_inc(&EXT4_I(inode)->i_unwritten);
3792 }
3793}
3794
3795static inline void ext4_clear_io_unwritten_flag(ext4_io_end_t *io_end)
3796{
3797 struct inode *inode = io_end->inode;
3798
3799 if (io_end->flag & EXT4_IO_END_UNWRITTEN) {
3800 io_end->flag &= ~EXT4_IO_END_UNWRITTEN;
3801 /* Wake up anyone waiting on unwritten extent conversion */
3802 if (atomic_dec_and_test(&EXT4_I(inode)->i_unwritten))
3803 wake_up_all(ext4_ioend_wq(inode));
3804 }
3805}
3806
8ff6daa1 3807extern const struct iomap_ops ext4_iomap_ops;
8cd115bd 3808extern const struct iomap_ops ext4_iomap_overwrite_ops;
09edf4d3 3809extern const struct iomap_ops ext4_iomap_report_ops;
364443cb 3810
7963e5ac
Z
3811static inline int ext4_buffer_uptodate(struct buffer_head *bh)
3812{
3813 /*
3814 * If the buffer has the write error flag, we have failed
3815 * to write out data in the block. In this case, we don't
3816 * have to read the block because we may read the old data
3817 * successfully.
3818 */
5c680150 3819 if (buffer_write_io_error(bh))
7963e5ac
Z
3820 set_buffer_uptodate(bh);
3821 return buffer_uptodate(bh);
3822}
3823
ac27a0ec
DK
3824#endif /* __KERNEL__ */
3825
6a797d27
DW
3826#define EFSBADCRC EBADMSG /* Bad CRC detected */
3827#define EFSCORRUPTED EUCLEAN /* Filesystem is corrupted */
3828
3dcf5451 3829#endif /* _EXT4_H */