1EXT4(5) File Formats Manual EXT4(5)
2
6 ext2 - the second extended file system
7 ext3 - the third extended file system
8 ext4 - the fourth extended file system
9
11 The second, third, and fourth extended file systems, or ext2, ext3, and
12 ext4 as they are commonly known, are Linux file systems that have his‐
13 torically been the default file system for many Linux distributions.
14 They are general purpose file systems that have been designed for ex‐
15 tensibility and backwards compatibility. In particular, file systems
16 previously intended for use with the ext2 and ext3 file systems can be
17 mounted using the ext4 file system driver, and indeed in many modern
18 Linux distributions, the ext4 file system driver has been configured to
19 handle mount requests for ext2 and ext3 file systems.
20
22 A file system formatted for ext2, ext3, or ext4 can have some collec‐
23 tion of the following file system feature flags enabled. Some of these
24 features are not supported by all implementations of the ext2, ext3,
25 and ext4 file system drivers, depending on Linux kernel version in use.
26 On other operating systems, such as the GNU/HURD or FreeBSD, only a
27 very restrictive set of file system features may be supported in their
28 implementations of ext2.
29 64bit
31 Enables the file system to be larger than 2^32 blocks. This
32 feature is set automatically, as needed, but it can be useful to
33 specify this feature explicitly if the file system might need to
34 be resized larger than 2^32 blocks, even if it was smaller than
35 that threshold when it was originally created. Note that some
36 older kernels and older versions of e2fsprogs will not support
37 file systems with this ext4 feature enabled.
38 bigalloc
40 This ext4 feature enables clustered block allocation, so that
41 the unit of allocation is a power of two number of blocks. That
42 is, each bit in the what had traditionally been known as the
43 block allocation bitmap now indicates whether a cluster is in
44 use or not, where a cluster is by default composed of 16 blocks.
45 This feature can decrease the time spent on doing block alloca‐
46 tion and brings smaller fragmentation, especially for large
47 files. The size can be specified using the mke2fs -C option.
48 Warning: The bigalloc feature is still under development, and
50 may not be fully supported with your kernel or may have various
51 bugs. Please see the web page http://ext4.wiki.kernel.org/in‐
52 dex.php/Bigalloc for details. May clash with delayed allocation
53 (see nodelalloc mount option).
54 This feature requires that the extent feature be enabled.
56 casefold
58 This ext4 feature provides file system level character encoding
59 support for directories with the casefold (+F) flag enabled.
60 This feature is name-preserving on the disk, but it allows ap‐
61 plications to lookup for a file in the file system using an en‐
62 coding equivalent version of the file name.
63 dir_index
65 Use hashed b-trees to speed up name lookups in large directo‐
66 ries. This feature is supported by ext3 and ext4 file systems,
67 and is ignored by ext2 file systems.
68 dir_nlink
70 Normally, ext4 allows an inode to have no more than 65,000 hard
71 links. This applies to regular files as well as directories,
72 which means that there can be no more than 64,998 subdirectories
73 in a directory (because each of the '.' and '..' entries, as
74 well as the directory entry for the directory in its parent di‐
75 rectory counts as a hard link). This feature lifts this limit
76 by causing ext4 to use a link count of 1 to indicate that the
77 number of hard links to a directory is not known when the link
78 count might exceed the maximum count limit.
79 ea_inode
81 Normally, a file's extended attributes and associated metadata
82 must fit within the inode or the inode's associated extended at‐
83 tribute block. This feature allows the value of each extended
84 attribute to be placed in the data blocks of a separate inode if
85 necessary, increasing the limit on the size and number of ex‐
86 tended attributes per file.
87 encrypt
89 Enables support for file-system level encryption of data blocks
90 and file names. The inode metadata (timestamps, file size,
91 user/group ownership, etc.) is not encrypted.
92 This feature is most useful on file systems with multiple users,
94 or where not all files should be encrypted. In many use cases,
95 especially on single-user systems, encryption at the block de‐
96 vice layer using dm-crypt may provide much better security.
97 ext_attr
99 This feature enables the use of extended attributes. This fea‐
100 ture is supported by ext2, ext3, and ext4.
101 extent
103 This ext4 feature allows the mapping of logical block numbers
104 for a particular inode to physical blocks on the storage device
105 to be stored using an extent tree, which is a more efficient
106 data structure than the traditional indirect block scheme used
107 by the ext2 and ext3 file systems. The use of the extent tree
108 decreases metadata block overhead, improves file system perfor‐
109 mance, and decreases the needed to run e2fsck(8) on the file
110 system. (Note: both extent and extents are accepted as valid
111 names for this feature for historical/backwards compatibility
112 reasons.)
113 extra_isize
115 This ext4 feature reserves a specific amount of space in each
116 inode for extended metadata such as nanosecond timestamps and
117 file creation time, even if the current kernel does not cur‐
118 rently need to reserve this much space. Without this feature,
119 the kernel will reserve the amount of space for features it cur‐
120 rently needs, and the rest may be consumed by extended at‐
121 tributes.
122 For this feature to be useful the inode size must be 256 bytes
124 in size or larger.
125 filetype
127 This feature enables the storage of file type information in di‐
128 rectory entries. This feature is supported by ext2, ext3, and
129 ext4.
130 flex_bg
132 This ext4 feature allows the per-block group metadata (alloca‐
133 tion bitmaps and inode tables) to be placed anywhere on the
134 storage media. In addition, mke2fs will place the per-block
135 group metadata together starting at the first block group of
136 each "flex_bg group". The size of the flex_bg group can be
137 specified using the -G option.
138 has_journal
140 Create a journal to ensure filesystem consistency even across
141 unclean shutdowns. Setting the filesystem feature is equivalent
142 to using the -j option with mke2fs or tune2fs. This feature is
143 supported by ext3 and ext4, and ignored by the ext2 file system
144 driver.
145 huge_file
147 This ext4 feature allows files to be larger than 2 terabytes in
148 size.
149 inline_data
151 Allow data to be stored in the inode and extended attribute
152 area.
153 journal_dev
155 This feature is enabled on the superblock found on an external
156 journal device. The block size for the external journal must be
157 the same as the file system which uses it.
158 The external journal device can be used by a file system by
160 specifying the -J device=<external-device> option to mke2fs(8)
161 or tune2fs(8).
162 large_dir
164 This feature increases the limit on the number of files per di‐
165 rectory by raising the maximum size of directories and, for
166 hashed b-tree directories (see dir_index), the maximum height of
167 the hashed b-tree used to store the directory entries.
168 large_file
170 This feature flag is set automatically by modern kernels when a
171 file larger than 2 gigabytes is created. Very old kernels could
172 not handle large files, so this feature flag was used to pro‐
173 hibit those kernels from mounting file systems that they could
174 not understand.
175 metadata_csum
177 This ext4 feature enables metadata checksumming. This feature
178 stores checksums for all of the filesystem metadata (superblock,
179 group descriptor blocks, inode and block bitmaps, directories,
180 and extent tree blocks). The checksum algorithm used for the
181 metadata blocks is different than the one used for group de‐
182 scriptors with the uninit_bg feature. These two features are
183 incompatible and metadata_csum will be used preferentially in‐
184 stead of uninit_bg.
185 metadata_csum_seed
187 This feature allows the filesystem to store the metadata check‐
188 sum seed in the superblock, which allows the administrator to
189 change the UUID of a filesystem using the metadata_csum feature
190 while it is mounted.
191 meta_bg
193 This ext4 feature allows file systems to be resized on-line
194 without explicitly needing to reserve space for growth in the
195 size of the block group descriptors. This scheme is also used
196 to resize file systems which are larger than 2^32 blocks. It is
197 not recommended that this feature be set when a file system is
198 created, since this alternate method of storing the block group
199 descriptors will slow down the time needed to mount the file
200 system, and newer kernels can automatically set this feature as
201 necessary when doing an online resize and no more reserved space
202 is available in the resize inode.
203 mmp
205 This ext4 feature provides multiple mount protection (MMP). MMP
206 helps to protect the filesystem from being multiply mounted and
207 is useful in shared storage environments.
208 project
210 This ext4 feature provides project quota support. With this fea‐
211 ture, the project ID of inode will be managed when the filesys‐
212 tem is mounted.
213 quota
215 Create quota inodes (inode #3 for userquota and inode #4 for
216 group quota) and set them in the superblock. With this feature,
217 the quotas will be enabled automatically when the filesystem is
218 mounted.
219 Causes the quota files (i.e., user.quota and group.quota which
221 existed in the older quota design) to be hidden inodes.
222 resize_inode
224 This file system feature indicates that space has been reserved
225 so that the block group descriptor table can be extended while
226 resizing a mounted file system. The online resize operation is
227 carried out by the kernel, triggered by resize2fs(8). By de‐
228 fault mke2fs will attempt to reserve enough space so that the
229 filesystem may grow to 1024 times its initial size. This can be
230 changed using the resize extended option.
231 This feature requires that the sparse_super or sparse_super2
233 feature be enabled.
234 sparse_super
236 This file system feature is set on all modern ext2, ext3, and
237 ext4 file systems. It indicates that backup copies of the su‐
238 perblock and block group descriptors are present only in a few
239 block groups, not all of them.
240 sparse_super2
242 This feature indicates that there will only be at most two
243 backup superblocks and block group descriptors. The block
244 groups used to store the backup superblock(s) and blockgroup de‐
245 scriptor(s) are stored in the superblock, but typically, one
246 will be located at the beginning of block group #1, and one in
247 the last block group in the file system. This feature is essen‐
248 tially a more extreme version of sparse_super and is designed to
249 allow a much larger percentage of the disk to have contiguous
250 blocks available for data files.
251 stable_inodes
253 Marks the filesystem's inode numbers and UUID as stable. re‐
254 size2fs(8) will not allow shrinking a filesystem with this fea‐
255 ture, nor will tune2fs(8) allow changing its UUID. This feature
256 allows the use of specialized encryption settings that make use
257 of the inode numbers and UUID. Note that the encrypt feature
258 still needs to be enabled separately. stable_inodes is a "com‐
259 pat" feature, so old kernels will allow it.
260 uninit_bg
262 This ext4 file system feature indicates that the block group de‐
263 scriptors will be protected using checksums, making it safe for
264 mke2fs(8) to create a file system without initializing all of
265 the block groups. The kernel will keep a high watermark of un‐
266 used inodes, and initialize inode tables and blocks lazily.
267 This feature speeds up the time to check the file system using
268 e2fsck(8), and it also speeds up the time required for mke2fs(8)
269 to create the file system.
270 verity
272 Enables support for verity protected files. Verity files are
273 readonly, and their data is transparently verified against a
274 Merkle tree hidden past the end of the file. Using the Merkle
275 tree's root hash, a verity file can be efficiently authenti‐
276 cated, independent of the file's size.
277 This feature is most useful for authenticating important read-
279 only files on read-write file systems. If the file system it‐
280 self is read-only, then using dm-verity to authenticate the en‐
281 tire block device may provide much better security.
282
284 This section describes mount options which are specific to ext2, ext3,
285 and ext4. Other generic mount options may be used as well; see
286 mount(8) for details.
287
289 The `ext2' filesystem is the standard Linux filesystem. Since Linux
290 2.5.46, for most mount options the default is determined by the
291 filesystem superblock. Set them with tune2fs(8).
292 acl|noacl
294 Support POSIX Access Control Lists (or not). See the acl(5)
295 manual page.
296 bsddf|minixdf
298 Set the behavior for the statfs system call. The minixdf behav‐
299 ior is to return in the f_blocks field the total number of
300 blocks of the filesystem, while the bsddf behavior (which is the
301 default) is to subtract the overhead blocks used by the ext2
302 filesystem and not available for file storage. Thus
303 % mount /k -o minixdf; df /k; umount /k
305 Filesystem 1024-blocks Used Available Capacity Mounted on
307 /dev/sda6 2630655 86954 2412169 3% /k
308 % mount /k -o bsddf; df /k; umount /k
310 Filesystem 1024-blocks Used Available Capacity Mounted on
312 /dev/sda6 2543714 13 2412169 0% /k
313 (Note that this example shows that one can add command line op‐
315 tions to the options given in /etc/fstab.)
316 check=none or nocheck
318 No checking is done at mount time. This is the default. This is
319 fast. It is wise to invoke e2fsck(8) every now and then, e.g.
320 at boot time. The non-default behavior is unsupported
321 (check=normal and check=strict options have been removed). Note
322 that these mount options don't have to be supported if ext4 ker‐
323 nel driver is used for ext2 and ext3 filesystems.
324 debug Print debugging info upon each (re)mount.
326 errors={continue|remount-ro|panic}
328 Define the behavior when an error is encountered. (Either ig‐
329 nore errors and just mark the filesystem erroneous and continue,
330 or remount the filesystem read-only, or panic and halt the sys‐
331 tem.) The default is set in the filesystem superblock, and can
332 be changed using tune2fs(8).
333 grpid|bsdgroups and nogrpid|sysvgroups
335 These options define what group id a newly created file gets.
336 When grpid is set, it takes the group id of the directory in
337 which it is created; otherwise (the default) it takes the fsgid
338 of the current process, unless the directory has the setgid bit
339 set, in which case it takes the gid from the parent directory,
340 and also gets the setgid bit set if it is a directory itself.
341 grpquota|noquota|quota|usrquota
343 The usrquota (same as quota) mount option enables user quota
344 support on the filesystem. grpquota enables group quotas sup‐
345 port. You need the quota utilities to actually enable and manage
346 the quota system.
347 nouid32
349 Disables 32-bit UIDs and GIDs. This is for interoperability
350 with older kernels which only store and expect 16-bit values.
351 oldalloc or orlov
353 Use old allocator or Orlov allocator for new inodes. Orlov is
354 default.
355 resgid=n and resuid=n
357 The ext2 filesystem reserves a certain percentage of the avail‐
358 able space (by default 5%, see mke2fs(8) and tune2fs(8)). These
359 options determine who can use the reserved blocks. (Roughly:
360 whoever has the specified uid, or belongs to the specified
361 group.)
362 sb=n Instead of using the normal superblock, use an alternative su‐
364 perblock specified by n. This option is normally used when the
365 primary superblock has been corrupted. The location of backup
366 superblocks is dependent on the filesystem's blocksize, the num‐
367 ber of blocks per group, and features such as sparse_super.
368 Additional backup superblocks can be determined by using the
370 mke2fs program using the -n option to print out where the su‐
371 perblocks exist, supposing mke2fs is supplied with arguments
372 that are consistent with the filesystem's layout (e.g. block‐
373 size, blocks per group, sparse_super, etc.).
374 The block number here uses 1 k units. Thus, if you want to use
376 logical block 32768 on a filesystem with 4 k blocks, use
377 "sb=131072".
378 user_xattr|nouser_xattr
380 Support "user." extended attributes (or not).
381
385 The ext3 filesystem is a version of the ext2 filesystem which has been
386 enhanced with journaling. It supports the same options as ext2 as well
387 as the following additions:
388 journal_dev=devnum/journal_path=path
390 When the external journal device's major/minor numbers have
391 changed, these options allow the user to specify the new journal
392 location. The journal device is identified either through its
393 new major/minor numbers encoded in devnum, or via a path to the
394 device.
395 norecovery/noload
397 Don't load the journal on mounting. Note that if the filesystem
398 was not unmounted cleanly, skipping the journal replay will lead
399 to the filesystem containing inconsistencies that can lead to
400 any number of problems.
401 data={journal|ordered|writeback}
403 Specifies the journaling mode for file data. Metadata is always
404 journaled. To use modes other than ordered on the root filesys‐
405 tem, pass the mode to the kernel as boot parameter, e.g. root‐
406 flags=data=journal.
407 journal
409 All data is committed into the journal prior to being
410 written into the main filesystem.
411 ordered
413 This is the default mode. All data is forced directly
414 out to the main file system prior to its metadata being
415 committed to the journal.
416 writeback
418 Data ordering is not preserved – data may be written into
419 the main filesystem after its metadata has been committed
420 to the journal. This is rumoured to be the highest-
421 throughput option. It guarantees internal filesystem in‐
422 tegrity, however it can allow old data to appear in files
423 after a crash and journal recovery.
424 data_err=ignore
426 Just print an error message if an error occurs in a file data
427 buffer in ordered mode.
428 data_err=abort
430 Abort the journal if an error occurs in a file data buffer in
431 ordered mode.
432 barrier=0 / barrier=1
434 This disables / enables the use of write barriers in the jbd
435 code. barrier=0 disables, barrier=1 enables (default). This
436 also requires an IO stack which can support barriers, and if jbd
437 gets an error on a barrier write, it will disable barriers again
438 with a warning. Write barriers enforce proper on-disk ordering
439 of journal commits, making volatile disk write caches safe to
440 use, at some performance penalty. If your disks are battery-
441 backed in one way or another, disabling barriers may safely im‐
442 prove performance.
443 commit=nrsec
445 Start a journal commit every nrsec seconds. The default value
446 is 5 seconds. Zero means default.
447 user_xattr
449 Enable Extended User Attributes. See the attr(5) manual page.
450 jqfmt={vfsold|vfsv0|vfsv1}
452 Apart from the old quota system (as in ext2, jqfmt=vfsold aka
453 version 1 quota) ext3 also supports journaled quotas (version 2
454 quota). jqfmt=vfsv0 or jqfmt=vfsv1 enables journaled quotas.
455 Journaled quotas have the advantage that even after a crash no
456 quota check is required. When the quota filesystem feature is
457 enabled, journaled quotas are used automatically, and this mount
458 option is ignored.
459 usrjquota=aquota.user|grpjquota=aquota.group
461 For journaled quotas (jqfmt=vfsv0 or jqfmt=vfsv1), the mount op‐
462 tions usrjquota=aquota.user and grpjquota=aquota.group are re‐
463 quired to tell the quota system which quota database files to
464 use. When the quota filesystem feature is enabled, journaled
465 quotas are used automatically, and this mount option is ignored.
466
469 The ext4 filesystem is an advanced level of the ext3 filesystem which
470 incorporates scalability and reliability enhancements for supporting
471 large filesystem.
472 The options journal_dev, journal_path, norecovery, noload, data, com‐
474 mit, orlov, oldalloc, [no]user_xattr, [no]acl, bsddf, minixdf, debug,
475 errors, data_err, grpid, bsdgroups, nogrpid, sysvgroups, resgid, re‐
476 suid, sb, quota, noquota, nouid32, grpquota, usrquota, usrjquota, gr‐
477 pjquota, and jqfmt are backwardly compatible with ext3 or ext2.
478 journal_checksum | nojournal_checksum
480 The journal_checksum option enables checksumming of the journal
481 transactions. This will allow the recovery code in e2fsck and
482 the kernel to detect corruption in the kernel. It is a compati‐
483 ble change and will be ignored by older kernels.
484 journal_async_commit
486 Commit block can be written to disk without waiting for descrip‐
487 tor blocks. If enabled older kernels cannot mount the device.
488 This will enable 'journal_checksum' internally.
489 barrier=0 / barrier=1 / barrier / nobarrier
491 These mount options have the same effect as in ext3. The mount
492 options "barrier" and "nobarrier" are added for consistency with
493 other ext4 mount options.
494 The ext4 filesystem enables write barriers by default.
496 inode_readahead_blks=n
498 This tuning parameter controls the maximum number of inode table
499 blocks that ext4's inode table readahead algorithm will pre-read
500 into the buffer cache. The value must be a power of 2. The de‐
501 fault value is 32 blocks.
502 stripe=n
504 Number of filesystem blocks that mballoc will try to use for al‐
505 location size and alignment. For RAID5/6 systems this should be
506 the number of data disks * RAID chunk size in filesystem blocks.
507 delalloc
509 Deferring block allocation until write-out time.
510 nodelalloc
512 Disable delayed allocation. Blocks are allocated when data is
513 copied from user to page cache.
514 max_batch_time=usec
516 Maximum amount of time ext4 should wait for additional filesys‐
517 tem operations to be batch together with a synchronous write op‐
518 eration. Since a synchronous write operation is going to force a
519 commit and then a wait for the I/O complete, it doesn't cost
520 much, and can be a huge throughput win, we wait for a small
521 amount of time to see if any other transactions can piggyback on
522 the synchronous write. The algorithm used is designed to auto‐
523 matically tune for the speed of the disk, by measuring the
524 amount of time (on average) that it takes to finish committing a
525 transaction. Call this time the "commit time". If the time that
526 the transaction has been running is less than the commit time,
527 ext4 will try sleeping for the commit time to see if other oper‐
528 ations will join the transaction. The commit time is capped by
529 the max_batch_time, which defaults to 15000 µs (15 ms). This op‐
530 timization can be turned off entirely by setting max_batch_time
531 to 0.
532 min_batch_time=usec
534 This parameter sets the commit time (as described above) to be
535 at least min_batch_time. It defaults to zero microseconds. In‐
536 creasing this parameter may improve the throughput of multi-
537 threaded, synchronous workloads on very fast disks, at the cost
538 of increasing latency.
539 journal_ioprio=prio
541 The I/O priority (from 0 to 7, where 0 is the highest priority)
542 which should be used for I/O operations submitted by kjournald2
543 during a commit operation. This defaults to 3, which is a
544 slightly higher priority than the default I/O priority.
545 abort Simulate the effects of calling ext4_abort() for debugging pur‐
547 poses. This is normally used while remounting a filesystem
548 which is already mounted.
549 auto_da_alloc|noauto_da_alloc
551 Many broken applications don't use fsync() when replacing exist‐
552 ing files via patterns such as
553 fd = open("foo.new")/write(fd,...)/close(fd)/ rename("foo.new",
555 "foo")
556 or worse yet
558 fd = open("foo", O_TRUNC)/write(fd,...)/close(fd).
560 If auto_da_alloc is enabled, ext4 will detect the replace-via-
562 rename and replace-via-truncate patterns and force that any de‐
563 layed allocation blocks are allocated such that at the next
564 journal commit, in the default data=ordered mode, the data
565 blocks of the new file are forced to disk before the rename()
566 operation is committed. This provides roughly the same level of
567 guarantees as ext3, and avoids the "zero-length" problem that
568 can happen when a system crashes before the delayed allocation
569 blocks are forced to disk.
570 noinit_itable
572 Do not initialize any uninitialized inode table blocks in the
573 background. This feature may be used by installation CD's so
574 that the install process can complete as quickly as possible;
575 the inode table initialization process would then be deferred
576 until the next time the filesystem is mounted.
577 init_itable=n
579 The lazy itable init code will wait n times the number of mil‐
580 liseconds it took to zero out the previous block group's inode
581 table. This minimizes the impact on system performance while the
582 filesystem's inode table is being initialized.
583 discard/nodiscard
585 Controls whether ext4 should issue discard/TRIM commands to the
586 underlying block device when blocks are freed. This is useful
587 for SSD devices and sparse/thinly-provisioned LUNs, but it is
588 off by default until sufficient testing has been done.
589 block_validity/noblock_validity
591 This option enables/disables the in-kernel facility for tracking
592 filesystem metadata blocks within internal data structures. This
593 allows multi-block allocator and other routines to quickly lo‐
594 cate extents which might overlap with filesystem metadata
595 blocks. This option is intended for debugging purposes and since
596 it negatively affects the performance, it is off by default.
597 dioread_lock/dioread_nolock
599 Controls whether or not ext4 should use the DIO read locking. If
600 the dioread_nolock option is specified ext4 will allocate unini‐
601 tialized extent before buffer write and convert the extent to
602 initialized after IO completes. This approach allows ext4 code
603 to avoid using inode mutex, which improves scalability on high
604 speed storages. However this does not work with data journaling
605 and dioread_nolock option will be ignored with kernel warning.
606 Note that dioread_nolock code path is only used for extent-based
607 files. Because of the restrictions this options comprises it is
608 off by default (e.g. dioread_lock).
609 max_dir_size_kb=n
611 This limits the size of the directories so that any attempt to
612 expand them beyond the specified limit in kilobytes will cause
613 an ENOSPC error. This is useful in memory-constrained environ‐
614 ments, where a very large directory can cause severe performance
615 problems or even provoke the Out Of Memory killer. (For example,
616 if there is only 512 MB memory available, a 176 MB directory may
617 seriously cramp the system's style.)
618 i_version
620 Enable 64-bit inode version support. This option is off by de‐
621 fault.
622 nombcache
624 This option disables use of mbcache for extended attribute dedu‐
625 plication. On systems where extended attributes are rarely or
626 never shared between files, use of mbcache for deduplication
627 adds unnecessary computational overhead.
628 prjquota
630 The prjquota mount option enables project quota support on the
631 filesystem. You need the quota utilities to actually enable and
632 manage the quota system. This mount option requires the project
633 filesystem feature.
634
637 The ext2, ext3, and ext4 filesystems support setting the following file
638 attributes on Linux systems using the chattr(1) utility:
639 a - append only
641 A - no atime updates
643 d - no dump
645 D - synchronous directory updates
647 i - immutable
649 S - synchronous updates
651 u - undeletable
653 In addition, the ext3 and ext4 filesystems support the following flag:
655 j - data journaling
657 Finally, the ext4 filesystem also supports the following flag:
659 e - extents format
661 For descriptions of these attribute flags, please refer to the
663 chattr(1) man page.
664
666 This section lists the file system driver (e.g., ext2, ext3, ext4) and
667 upstream kernel version where a particular file system feature was sup‐
668 ported. Note that in some cases the feature was present in earlier
669 kernel versions, but there were known, serious bugs. In other cases
670 the feature may still be considered in an experimental state. Finally,
671 note that some distributions may have backported features into older
672 kernels; in particular the kernel versions in certain "enterprise dis‐
673 tributions" can be extremely misleading.
674 filetype ext2, 2.2.0
676 sparse_super ext2, 2.2.0
678 large_file ext2, 2.2.0
680 has_journal ext3, 2.4.15
682 ext_attr ext2/ext3, 2.6.0
684 dir_index ext3, 2.6.0
686 resize_inode ext3, 2.6.10 (online resizing)
688 64bit ext4, 2.6.28
690 dir_nlink ext4, 2.6.28
692 extent ext4, 2.6.28
694 extra_isize ext4, 2.6.28
696 flex_bg ext4, 2.6.28
698 huge_file ext4, 2.6.28
700 meta_bg ext4, 2.6.28
702 uninit_bg ext4, 2.6.28
704 mmp ext4, 3.0
706 bigalloc ext4, 3.2
708 quota ext4, 3.6
710 inline_data ext4, 3.8
712 sparse_super2 ext4, 3.16
714 metadata_csum ext4, 3.18
716 encrypt ext4, 4.1
718 metadata_csum_seed ext4, 4.4
720 project ext4, 4.5
722 ea_inode ext4, 4.13
724 large_dir ext4, 4.13
726 casefold ext4, 5.2
728 verity ext4, 5.4
730 stable_inodes ext4, 5.5
732
734 mke2fs(8), mke2fs.conf(5), e2fsck(8), dumpe2fs(8), tune2fs(8), de‐
735 bugfs(8), mount(8), chattr(1)
736E2fsprogs version 1.46.3 July 2021 EXT4(5)