1 .. SPDX-License-Identifier: GPL-2.0
2
3 Index Nodes
4 -----------
5
6 In a regular UNIX filesystem, the inode stores
7 pertaining to the file (time stamps, block map
8 etc), not the directory entry. To find the inf
9 file, one must traverse the directory files to
10 associated with a file, then load the inode to
11 that file. ext4 appears to cheat (for performa
12 by storing a copy of the file type (normally s
13 directory entry. (Compare all this to FAT, whi
14 information directly in the directory entry, b
15 links and is in general more seek-happy than e
16 block allocator and extensive use of linked li
17
18 The inode table is a linear array of ``struct
19 sized to have enough blocks to store at least
20 ``sb.s_inode_size * sb.s_inodes_per_group`` by
21 block group containing an inode can be calcula
22 ``(inode_number - 1) / sb.s_inodes_per_group``
23 group's table is ``(inode_number - 1) % sb.s_i
24 is no inode 0.
25
26 The inode checksum is calculated against the F
27 and the inode structure itself.
28
29 The inode table entry is laid out in ``struct
30
31 .. list-table::
32 :widths: 8 8 24 40
33 :header-rows: 1
34 :class: longtable
35
36 * - Offset
37 - Size
38 - Name
39 - Description
40 * - 0x0
41 - __le16
42 - i_mode
43 - File mode. See the table i_mode_ below.
44 * - 0x2
45 - __le16
46 - i_uid
47 - Lower 16-bits of Owner UID.
48 * - 0x4
49 - __le32
50 - i_size_lo
51 - Lower 32-bits of size in bytes.
52 * - 0x8
53 - __le32
54 - i_atime
55 - Last access time, in seconds since the
56 inode flag is set, this inode stores an
57 this field contains the checksum of the
58 * - 0xC
59 - __le32
60 - i_ctime
61 - Last inode change time, in seconds sinc
62 EA_INODE inode flag is set, this inode
63 value and this field contains the lower
64 reference count.
65 * - 0x10
66 - __le32
67 - i_mtime
68 - Last data modification time, in seconds
69 EA_INODE inode flag is set, this inode
70 value and this field contains the numbe
71 extended attribute.
72 * - 0x14
73 - __le32
74 - i_dtime
75 - Deletion Time, in seconds since the epo
76 * - 0x18
77 - __le16
78 - i_gid
79 - Lower 16-bits of GID.
80 * - 0x1A
81 - __le16
82 - i_links_count
83 - Hard link count. Normally, ext4 does no
84 than 65,000 hard links. This applies to
85 which means that there cannot be more t
86 directory (each subdirectory's '..' ent
87 the '.' entry in the directory itself).
88 enabled, ext4 supports more than 64,998
89 field to 1 to indicate that the number
90 * - 0x1C
91 - __le32
92 - i_blocks_lo
93 - Lower 32-bits of “block” count. If
94 set on the filesystem, the file consume
95 on disk. If huge_file is set and EXT4_H
96 ``inode.i_flags``, then the file consum
97 << 32)`` 512-byte blocks on disk. If hu
98 EXT4_HUGE_FILE_FL IS set in ``inode.i_f
99 consumes (``i_blocks_lo + i_blocks_hi``
100 disk.
101 * - 0x20
102 - __le32
103 - i_flags
104 - Inode flags. See the table i_flags_ bel
105 * - 0x24
106 - 4 bytes
107 - i_osd1
108 - See the table i_osd1_ for more details.
109 * - 0x28
110 - 60 bytes
111 - i_block[EXT4_N_BLOCKS=15]
112 - Block map or extent tree. See the secti
113 * - 0x64
114 - __le32
115 - i_generation
116 - File version (for NFS).
117 * - 0x68
118 - __le32
119 - i_file_acl_lo
120 - Lower 32-bits of extended attribute blo
121 many possible extended attributes; I th
122 result of the first use of extended att
123 * - 0x6C
124 - __le32
125 - i_size_high / i_dir_acl
126 - Upper 32-bits of file/directory size. I
127 i_dir_acl, though it was usually set to
128 * - 0x70
129 - __le32
130 - i_obso_faddr
131 - (Obsolete) fragment address.
132 * - 0x74
133 - 12 bytes
134 - i_osd2
135 - See the table i_osd2_ for more details.
136 * - 0x80
137 - __le16
138 - i_extra_isize
139 - Size of this inode - 128. Alternately,
140 fields beyond the original ext2 inode,
141 * - 0x82
142 - __le16
143 - i_checksum_hi
144 - Upper 16-bits of the inode checksum.
145 * - 0x84
146 - __le32
147 - i_ctime_extra
148 - Extra change time bits. This provides s
149 Timestamps section.
150 * - 0x88
151 - __le32
152 - i_mtime_extra
153 - Extra modification time bits. This prov
154 * - 0x8C
155 - __le32
156 - i_atime_extra
157 - Extra access time bits. This provides s
158 * - 0x90
159 - __le32
160 - i_crtime
161 - File creation time, in seconds since th
162 * - 0x94
163 - __le32
164 - i_crtime_extra
165 - Extra file creation time bits. This pro
166 * - 0x98
167 - __le32
168 - i_version_hi
169 - Upper 32-bits for version number.
170 * - 0x9C
171 - __le32
172 - i_projid
173 - Project ID.
174
175 .. _i_mode:
176
177 The ``i_mode`` value is a combination of the f
178
179 .. list-table::
180 :widths: 16 64
181 :header-rows: 1
182
183 * - Value
184 - Description
185 * - 0x1
186 - S_IXOTH (Others may execute)
187 * - 0x2
188 - S_IWOTH (Others may write)
189 * - 0x4
190 - S_IROTH (Others may read)
191 * - 0x8
192 - S_IXGRP (Group members may execute)
193 * - 0x10
194 - S_IWGRP (Group members may write)
195 * - 0x20
196 - S_IRGRP (Group members may read)
197 * - 0x40
198 - S_IXUSR (Owner may execute)
199 * - 0x80
200 - S_IWUSR (Owner may write)
201 * - 0x100
202 - S_IRUSR (Owner may read)
203 * - 0x200
204 - S_ISVTX (Sticky bit)
205 * - 0x400
206 - S_ISGID (Set GID)
207 * - 0x800
208 - S_ISUID (Set UID)
209 * -
210 - These are mutually-exclusive file types
211 * - 0x1000
212 - S_IFIFO (FIFO)
213 * - 0x2000
214 - S_IFCHR (Character device)
215 * - 0x4000
216 - S_IFDIR (Directory)
217 * - 0x6000
218 - S_IFBLK (Block device)
219 * - 0x8000
220 - S_IFREG (Regular file)
221 * - 0xA000
222 - S_IFLNK (Symbolic link)
223 * - 0xC000
224 - S_IFSOCK (Socket)
225
226 .. _i_flags:
227
228 The ``i_flags`` field is a combination of thes
229
230 .. list-table::
231 :widths: 16 64
232 :header-rows: 1
233
234 * - Value
235 - Description
236 * - 0x1
237 - This file requires secure deletion (EXT
238 * - 0x2
239 - This file should be preserved, should u
240 (EXT4_UNRM_FL). (not implemented)
241 * - 0x4
242 - File is compressed (EXT4_COMPR_FL). (no
243 * - 0x8
244 - All writes to the file must be synchron
245 * - 0x10
246 - File is immutable (EXT4_IMMUTABLE_FL).
247 * - 0x20
248 - File can only be appended (EXT4_APPEND_
249 * - 0x40
250 - The dump(1) utility should not dump thi
251 * - 0x80
252 - Do not update access time (EXT4_NOATIME
253 * - 0x100
254 - Dirty compressed file (EXT4_DIRTY_FL).
255 * - 0x200
256 - File has one or more compressed cluster
257 * - 0x400
258 - Do not compress file (EXT4_NOCOMPR_FL).
259 * - 0x800
260 - Encrypted inode (EXT4_ENCRYPT_FL). This
261 EXT4_ECOMPR_FL (compression error), whi
262 * - 0x1000
263 - Directory has hashed indexes (EXT4_INDE
264 * - 0x2000
265 - AFS magic directory (EXT4_IMAGIC_FL).
266 * - 0x4000
267 - File data must always be written throug
268 (EXT4_JOURNAL_DATA_FL).
269 * - 0x8000
270 - File tail should not be merged (EXT4_NO
271 * - 0x10000
272 - All directory entry data should be writ
273 ``dirsync``) (EXT4_DIRSYNC_FL).
274 * - 0x20000
275 - Top of directory hierarchy (EXT4_TOPDIR
276 * - 0x40000
277 - This is a huge file (EXT4_HUGE_FILE_FL)
278 * - 0x80000
279 - Inode uses extents (EXT4_EXTENTS_FL).
280 * - 0x100000
281 - Verity protected file (EXT4_VERITY_FL).
282 * - 0x200000
283 - Inode stores a large extended attribute
284 (EXT4_EA_INODE_FL).
285 * - 0x400000
286 - This file has blocks allocated past EOF
287 (deprecated)
288 * - 0x01000000
289 - Inode is a snapshot (``EXT4_SNAPFILE_FL
290 * - 0x04000000
291 - Snapshot is being deleted (``EXT4_SNAPF
292 mainline)
293 * - 0x08000000
294 - Snapshot shrink has completed (``EXT4_S
295 mainline)
296 * - 0x10000000
297 - Inode has inline data (EXT4_INLINE_DATA
298 * - 0x20000000
299 - Create children with the same project I
300 * - 0x80000000
301 - Reserved for ext4 library (EXT4_RESERVE
302 * -
303 - Aggregate flags:
304 * - 0x705BDFFF
305 - User-visible flags.
306 * - 0x604BC0FF
307 - User-modifiable flags. Note that while
308 EXT4_EXTENTS_FL can be set with setattr
309 EXT4_FL_USER_MODIFIABLE mask, since it
310 these flags in a special manner and the
311 flags that are saved directly to i_flag
312
313 .. _i_osd1:
314
315 The ``osd1`` field has multiple meanings depen
316
317 Linux:
318
319 .. list-table::
320 :widths: 8 8 24 40
321 :header-rows: 1
322
323 * - Offset
324 - Size
325 - Name
326 - Description
327 * - 0x0
328 - __le32
329 - l_i_version
330 - Inode version. However, if the EA_INODE
331 stores an extended attribute value and
332 bits of the attribute value's reference
333
334 Hurd:
335
336 .. list-table::
337 :widths: 8 8 24 40
338 :header-rows: 1
339
340 * - Offset
341 - Size
342 - Name
343 - Description
344 * - 0x0
345 - __le32
346 - h_i_translator
347 - ??
348
349 Masix:
350
351 .. list-table::
352 :widths: 8 8 24 40
353 :header-rows: 1
354
355 * - Offset
356 - Size
357 - Name
358 - Description
359 * - 0x0
360 - __le32
361 - m_i_reserved
362 - ??
363
364 .. _i_osd2:
365
366 The ``osd2`` field has multiple meanings depen
367
368 Linux:
369
370 .. list-table::
371 :widths: 8 8 24 40
372 :header-rows: 1
373
374 * - Offset
375 - Size
376 - Name
377 - Description
378 * - 0x0
379 - __le16
380 - l_i_blocks_high
381 - Upper 16-bits of the block count. Pleas
382 i_blocks_lo.
383 * - 0x2
384 - __le16
385 - l_i_file_acl_high
386 - Upper 16-bits of the extended attribute
387 ACL location). See the Extended Attribu
388 * - 0x4
389 - __le16
390 - l_i_uid_high
391 - Upper 16-bits of the Owner UID.
392 * - 0x6
393 - __le16
394 - l_i_gid_high
395 - Upper 16-bits of the GID.
396 * - 0x8
397 - __le16
398 - l_i_checksum_lo
399 - Lower 16-bits of the inode checksum.
400 * - 0xA
401 - __le16
402 - l_i_reserved
403 - Unused.
404
405 Hurd:
406
407 .. list-table::
408 :widths: 8 8 24 40
409 :header-rows: 1
410
411 * - Offset
412 - Size
413 - Name
414 - Description
415 * - 0x0
416 - __le16
417 - h_i_reserved1
418 - ??
419 * - 0x2
420 - __u16
421 - h_i_mode_high
422 - Upper 16-bits of the file mode.
423 * - 0x4
424 - __le16
425 - h_i_uid_high
426 - Upper 16-bits of the Owner UID.
427 * - 0x6
428 - __le16
429 - h_i_gid_high
430 - Upper 16-bits of the GID.
431 * - 0x8
432 - __u32
433 - h_i_author
434 - Author code?
435
436 Masix:
437
438 .. list-table::
439 :widths: 8 8 24 40
440 :header-rows: 1
441
442 * - Offset
443 - Size
444 - Name
445 - Description
446 * - 0x0
447 - __le16
448 - h_i_reserved1
449 - ??
450 * - 0x2
451 - __u16
452 - m_i_file_acl_high
453 - Upper 16-bits of the extended attribute
454 ACL location).
455 * - 0x4
456 - __u32
457 - m_i_reserved2[2]
458 - ??
459
460 Inode Size
461 ~~~~~~~~~~
462
463 In ext2 and ext3, the inode structure size was
464 (``EXT2_GOOD_OLD_INODE_SIZE``) and each inode
465 128 bytes. Starting with ext4, it is possible
466 on-disk inode at format time for all inodes in
467 space beyond the end of the original ext2 inod
468 record size is recorded in the superblock as `
469 number of bytes actually used by struct ext4_i
470 128-byte ext2 inode is recorded in the ``i_ext
471 inode, which allows struct ext4_inode to grow
472 having to upgrade all of the on-disk inodes. A
473 EXT2_GOOD_OLD_INODE_SIZE should be verified to
474 ``i_extra_isize``. By default, ext4 inode reco
475 of August 2019) the inode structure is 160 byt
476 (``i_extra_isize = 32``). The extra space betw
477 structure and the end of the inode record can
478 attributes. Each inode record can be as large
479 size, though this is not terribly efficient.
480
481 Finding an Inode
482 ~~~~~~~~~~~~~~~~
483
484 Each block group contains ``sb->s_inodes_per_g
485 inode 0 is defined not to exist, this formula
486 block group that an inode lives in:
487 ``bg = (inode_num - 1) / sb->s_inodes_per_grou
488 can be found within the block group's inode ta
489 ``index = (inode_num - 1) % sb->s_inodes_per_g
490 address within the inode table, use
491 ``offset = index * sb->s_inode_size``.
492
493 Inode Timestamps
494 ~~~~~~~~~~~~~~~~
495
496 Four timestamps are recorded in the lower 128
497 structure -- inode change time (ctime), access
498 modification time (mtime), and deletion time (
499 are 32-bit signed integers that represent seco
500 (1970-01-01 00:00:00 GMT), which means that th
501 January 2038. If the filesystem does not have
502 that are not linked from any directory but are
503 the dtime field overloaded for use with the or
504 ``s_last_orphan`` points to the first inode in
505 the number of the next orphaned inode, or zero
506
507 If the inode structure size ``sb->s_inode_size
508 bytes and the ``i_inode_extra`` field is large
509 respective ``i_[cma]time_extra`` field, the ct
510 inode fields are widened to 64 bits. Within th
511 the lower two bits are used to extend the 32-b
512 bit wide; the upper 30 bits are used to provid
513 accuracy. Therefore, timestamps should not ove
514 dtime was not widened. There is also a fifth t
515 creation time (crtime); this field is 64-bits
516 same manner as 64-bit [cma]time. Neither crtim
517 through the regular stat() interface, though d
518
519 We use the 32-bit signed time value plus (2^32
520 In other words:
521
522 .. list-table::
523 :widths: 20 20 20 20 20
524 :header-rows: 1
525
526 * - Extra epoch bits
527 - MSB of 32-bit time
528 - Adjustment for signed 32-bit to 64-bit
529 - Decoded 64-bit tv_sec
530 - valid time range
531 * - 0 0
532 - 1
533 - 0
534 - ``-0x80000000 - -0x00000001``
535 - 1901-12-13 to 1969-12-31
536 * - 0 0
537 - 0
538 - 0
539 - ``0x000000000 - 0x07fffffff``
540 - 1970-01-01 to 2038-01-19
541 * - 0 1
542 - 1
543 - 0x100000000
544 - ``0x080000000 - 0x0ffffffff``
545 - 2038-01-19 to 2106-02-07
546 * - 0 1
547 - 0
548 - 0x100000000
549 - ``0x100000000 - 0x17fffffff``
550 - 2106-02-07 to 2174-02-25
551 * - 1 0
552 - 1
553 - 0x200000000
554 - ``0x180000000 - 0x1ffffffff``
555 - 2174-02-25 to 2242-03-16
556 * - 1 0
557 - 0
558 - 0x200000000
559 - ``0x200000000 - 0x27fffffff``
560 - 2242-03-16 to 2310-04-04
561 * - 1 1
562 - 1
563 - 0x300000000
564 - ``0x280000000 - 0x2ffffffff``
565 - 2310-04-04 to 2378-04-22
566 * - 1 1
567 - 0
568 - 0x300000000
569 - ``0x300000000 - 0x37fffffff``
570 - 2378-04-22 to 2446-05-10
571
572 This is a somewhat odd encoding since there ar
573 as many positive values as negative values. Th
574 long-standing bugs decoding and encoding dates
575 seem to be fixed as of kernel 3.12 and e2fspro
576 incorrectly use the extra epoch bits 1,1 for d
577 1970. At some point the kernel will be fixed a
578 situation, assuming that it is run before 2310
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.