0001 .. SPDX-License-Identifier: GPL-2.0
0002
0003 ========================
0004 ext4 General Information
0005 ========================
0006
0007 Ext4 is an advanced level of the ext3 filesystem which incorporates
0008 scalability and reliability enhancements for supporting large filesystems
0009 (64 bit) in keeping with increasing disk capacities and state-of-the-art
0010 feature requirements.
0011
0012 Mailing list: linux-ext4@vger.kernel.org
0013 Web site: http://ext4.wiki.kernel.org
0014
0015
0016 Quick usage instructions
0017 ========================
0018
0019 Note: More extensive information for getting started with ext4 can be
0020 found at the ext4 wiki site at the URL:
0021 http://ext4.wiki.kernel.org/index.php/Ext4_Howto
0022
0023 - The latest version of e2fsprogs can be found at:
0024
0025 https://www.kernel.org/pub/linux/kernel/people/tytso/e2fsprogs/
0026
0027 or
0028
0029 http://sourceforge.net/project/showfiles.php?group_id=2406
0030
0031 or grab the latest git repository from:
0032
0033 https://git.kernel.org/pub/scm/fs/ext2/e2fsprogs.git
0034
0035 - Create a new filesystem using the ext4 filesystem type:
0036
0037 # mke2fs -t ext4 /dev/hda1
0038
0039 Or to configure an existing ext3 filesystem to support extents:
0040
0041 # tune2fs -O extents /dev/hda1
0042
0043 If the filesystem was created with 128 byte inodes, it can be
0044 converted to use 256 byte for greater efficiency via:
0045
0046 # tune2fs -I 256 /dev/hda1
0047
0048 - Mounting:
0049
0050 # mount -t ext4 /dev/hda1 /wherever
0051
0052 - When comparing performance with other filesystems, it's always
0053 important to try multiple workloads; very often a subtle change in a
0054 workload parameter can completely change the ranking of which
0055 filesystems do well compared to others. When comparing versus ext3,
0056 note that ext4 enables write barriers by default, while ext3 does
0057 not enable write barriers by default. So it is useful to use
0058 explicitly specify whether barriers are enabled or not when via the
0059 '-o barriers=[0|1]' mount option for both ext3 and ext4 filesystems
0060 for a fair comparison. When tuning ext3 for best benchmark numbers,
0061 it is often worthwhile to try changing the data journaling mode; '-o
0062 data=writeback' can be faster for some workloads. (Note however that
0063 running mounted with data=writeback can potentially leave stale data
0064 exposed in recently written files in case of an unclean shutdown,
0065 which could be a security exposure in some situations.) Configuring
0066 the filesystem with a large journal can also be helpful for
0067 metadata-intensive workloads.
0068
0069 Features
0070 ========
0071
0072 Currently Available
0073 -------------------
0074
0075 * ability to use filesystems > 16TB (e2fsprogs support not available yet)
0076 * extent format reduces metadata overhead (RAM, IO for access, transactions)
0077 * extent format more robust in face of on-disk corruption due to magics,
0078 * internal redundancy in tree
0079 * improved file allocation (multi-block alloc)
0080 * lift 32000 subdirectory limit imposed by i_links_count[1]
0081 * nsec timestamps for mtime, atime, ctime, create time
0082 * inode version field on disk (NFSv4, Lustre)
0083 * reduced e2fsck time via uninit_bg feature
0084 * journal checksumming for robustness, performance
0085 * persistent file preallocation (e.g for streaming media, databases)
0086 * ability to pack bitmaps and inode tables into larger virtual groups via the
0087 flex_bg feature
0088 * large file support
0089 * inode allocation using large virtual block groups via flex_bg
0090 * delayed allocation
0091 * large block (up to pagesize) support
0092 * efficient new ordered mode in JBD2 and ext4 (avoid using buffer head to force
0093 the ordering)
0094 * Case-insensitive file name lookups
0095 * file-based encryption support (fscrypt)
0096 * file-based verity support (fsverity)
0097
0098 [1] Filesystems with a block size of 1k may see a limit imposed by the
0099 directory hash tree having a maximum depth of two.
0100
0101 case-insensitive file name lookups
0102 ======================================================
0103
0104 The case-insensitive file name lookup feature is supported on a
0105 per-directory basis, allowing the user to mix case-insensitive and
0106 case-sensitive directories in the same filesystem. It is enabled by
0107 flipping the +F inode attribute of an empty directory. The
0108 case-insensitive string match operation is only defined when we know how
0109 text in encoded in a byte sequence. For that reason, in order to enable
0110 case-insensitive directories, the filesystem must have the
0111 casefold feature, which stores the filesystem-wide encoding
0112 model used. By default, the charset adopted is the latest version of
0113 Unicode (12.1.0, by the time of this writing), encoded in the UTF-8
0114 form. The comparison algorithm is implemented by normalizing the
0115 strings to the Canonical decomposition form, as defined by Unicode,
0116 followed by a byte per byte comparison.
0117
0118 The case-awareness is name-preserving on the disk, meaning that the file
0119 name provided by userspace is a byte-per-byte match to what is actually
0120 written in the disk. The Unicode normalization format used by the
0121 kernel is thus an internal representation, and not exposed to the
0122 userspace nor to the disk, with the important exception of disk hashes,
0123 used on large case-insensitive directories with DX feature. On DX
0124 directories, the hash must be calculated using the casefolded version of
0125 the filename, meaning that the normalization format used actually has an
0126 impact on where the directory entry is stored.
0127
0128 When we change from viewing filenames as opaque byte sequences to seeing
0129 them as encoded strings we need to address what happens when a program
0130 tries to create a file with an invalid name. The Unicode subsystem
0131 within the kernel leaves the decision of what to do in this case to the
0132 filesystem, which select its preferred behavior by enabling/disabling
0133 the strict mode. When Ext4 encounters one of those strings and the
0134 filesystem did not require strict mode, it falls back to considering the
0135 entire string as an opaque byte sequence, which still allows the user to
0136 operate on that file, but the case-insensitive lookups won't work.
0137
0138 Options
0139 =======
0140
0141 When mounting an ext4 filesystem, the following option are accepted:
0142 (*) == default
0143
0144 ro
0145 Mount filesystem read only. Note that ext4 will replay the journal (and
0146 thus write to the partition) even when mounted "read only". The mount
0147 options "ro,noload" can be used to prevent writes to the filesystem.
0148
0149 journal_checksum
0150 Enable checksumming of the journal transactions. This will allow the
0151 recovery code in e2fsck and the kernel to detect corruption in the
0152 kernel. It is a compatible change and will be ignored by older
0153 kernels.
0154
0155 journal_async_commit
0156 Commit block can be written to disk without waiting for descriptor
0157 blocks. If enabled older kernels cannot mount the device. This will
0158 enable 'journal_checksum' internally.
0159
0160 journal_path=path, journal_dev=devnum
0161 When the external journal device's major/minor numbers have changed,
0162 these options allow the user to specify the new journal location. The
0163 journal device is identified through either its new major/minor numbers
0164 encoded in devnum, or via a path to the device.
0165
0166 norecovery, noload
0167 Don't load the journal on mounting. Note that if the filesystem was
0168 not unmounted cleanly, skipping the journal replay will lead to the
0169 filesystem containing inconsistencies that can lead to any number of
0170 problems.
0171
0172 data=journal
0173 All data are committed into the journal prior to being written into the
0174 main file system. Enabling this mode will disable delayed allocation
0175 and O_DIRECT support.
0176
0177 data=ordered (*)
0178 All data are forced directly out to the main file system prior to its
0179 metadata being committed to the journal.
0180
0181 data=writeback
0182 Data ordering is not preserved, data may be written into the main file
0183 system after its metadata has been committed to the journal.
0184
0185 commit=nrsec (*)
0186 This setting limits the maximum age of the running transaction to
0187 'nrsec' seconds. The default value is 5 seconds. This means that if
0188 you lose your power, you will lose as much as the latest 5 seconds of
0189 metadata changes (your filesystem will not be damaged though, thanks
0190 to the journaling). This default value (or any low value) will hurt
0191 performance, but it's good for data-safety. Setting it to 0 will have
0192 the same effect as leaving it at the default (5 seconds). Setting it
0193 to very large values will improve performance. Note that due to
0194 delayed allocation even older data can be lost on power failure since
0195 writeback of those data begins only after time set in
0196 /proc/sys/vm/dirty_expire_centisecs.
0197
0198 barrier=<0|1(*)>, barrier(*), nobarrier
0199 This enables/disables the use of write barriers in the jbd code.
0200 barrier=0 disables, barrier=1 enables. This also requires an IO stack
0201 which can support barriers, and if jbd gets an error on a barrier
0202 write, it will disable again with a warning. Write barriers enforce
0203 proper on-disk ordering of journal commits, making volatile disk write
0204 caches safe to use, at some performance penalty. If your disks are
0205 battery-backed in one way or another, disabling barriers may safely
0206 improve performance. The mount options "barrier" and "nobarrier" can
0207 also be used to enable or disable barriers, for consistency with other
0208 ext4 mount options.
0209
0210 inode_readahead_blks=n
0211 This tuning parameter controls the maximum number of inode table blocks
0212 that ext4's inode table readahead algorithm will pre-read into the
0213 buffer cache. The default value is 32 blocks.
0214
0215 nouser_xattr
0216 Disables Extended User Attributes. See the attr(5) manual page for
0217 more information about extended attributes.
0218
0219 noacl
0220 This option disables POSIX Access Control List support. If ACL support
0221 is enabled in the kernel configuration (CONFIG_EXT4_FS_POSIX_ACL), ACL
0222 is enabled by default on mount. See the acl(5) manual page for more
0223 information about acl.
0224
0225 bsddf (*)
0226 Make 'df' act like BSD.
0227
0228 minixdf
0229 Make 'df' act like Minix.
0230
0231 debug
0232 Extra debugging information is sent to syslog.
0233
0234 abort
0235 Simulate the effects of calling ext4_abort() for debugging purposes.
0236 This is normally used while remounting a filesystem which is already
0237 mounted.
0238
0239 errors=remount-ro
0240 Remount the filesystem read-only on an error.
0241
0242 errors=continue
0243 Keep going on a filesystem error.
0244
0245 errors=panic
0246 Panic and halt the machine if an error occurs. (These mount options
0247 override the errors behavior specified in the superblock, which can be
0248 configured using tune2fs)
0249
0250 data_err=ignore(*)
0251 Just print an error message if an error occurs in a file data buffer in
0252 ordered mode.
0253 data_err=abort
0254 Abort the journal if an error occurs in a file data buffer in ordered
0255 mode.
0256
0257 grpid | bsdgroups
0258 New objects have the group ID of their parent.
0259
0260 nogrpid (*) | sysvgroups
0261 New objects have the group ID of their creator.
0262
0263 resgid=n
0264 The group ID which may use the reserved blocks.
0265
0266 resuid=n
0267 The user ID which may use the reserved blocks.
0268
0269 sb=
0270 Use alternate superblock at this location.
0271
0272 quota, noquota, grpquota, usrquota
0273 These options are ignored by the filesystem. They are used only by
0274 quota tools to recognize volumes where quota should be turned on. See
0275 documentation in the quota-tools package for more details
0276 (http://sourceforge.net/projects/linuxquota).
0277
0278 jqfmt=<quota type>, usrjquota=<file>, grpjquota=<file>
0279 These options tell filesystem details about quota so that quota
0280 information can be properly updated during journal replay. They replace
0281 the above quota options. See documentation in the quota-tools package
0282 for more details (http://sourceforge.net/projects/linuxquota).
0283
0284 stripe=n
0285 Number of filesystem blocks that mballoc will try to use for allocation
0286 size and alignment. For RAID5/6 systems this should be the number of
0287 data disks * RAID chunk size in file system blocks.
0288
0289 delalloc (*)
0290 Defer block allocation until just before ext4 writes out the block(s)
0291 in question. This allows ext4 to better allocation decisions more
0292 efficiently.
0293
0294 nodelalloc
0295 Disable delayed allocation. Blocks are allocated when the data is
0296 copied from userspace to the page cache, either via the write(2) system
0297 call or when an mmap'ed page which was previously unallocated is
0298 written for the first time.
0299
0300 max_batch_time=usec
0301 Maximum amount of time ext4 should wait for additional filesystem
0302 operations to be batch together with a synchronous write operation.
0303 Since a synchronous write operation is going to force a commit and then
0304 a wait for the I/O complete, it doesn't cost much, and can be a huge
0305 throughput win, we wait for a small amount of time to see if any other
0306 transactions can piggyback on the synchronous write. The algorithm
0307 used is designed to automatically tune for the speed of the disk, by
0308 measuring the amount of time (on average) that it takes to finish
0309 committing a transaction. Call this time the "commit time". If the
0310 time that the transaction has been running is less than the commit
0311 time, ext4 will try sleeping for the commit time to see if other
0312 operations will join the transaction. The commit time is capped by
0313 the max_batch_time, which defaults to 15000us (15ms). This
0314 optimization can be turned off entirely by setting max_batch_time to 0.
0315
0316 min_batch_time=usec
0317 This parameter sets the commit time (as described above) to be at least
0318 min_batch_time. It defaults to zero microseconds. Increasing this
0319 parameter may improve the throughput of multi-threaded, synchronous
0320 workloads on very fast disks, at the cost of increasing latency.
0321
0322 journal_ioprio=prio
0323 The I/O priority (from 0 to 7, where 0 is the highest priority) which
0324 should be used for I/O operations submitted by kjournald2 during a
0325 commit operation. This defaults to 3, which is a slightly higher
0326 priority than the default I/O priority.
0327
0328 auto_da_alloc(*), noauto_da_alloc
0329 Many broken applications don't use fsync() when replacing existing
0330 files via patterns such as fd = open("foo.new")/write(fd,..)/close(fd)/
0331 rename("foo.new", "foo"), or worse yet, fd = open("foo",
0332 O_TRUNC)/write(fd,..)/close(fd). If auto_da_alloc is enabled, ext4
0333 will detect the replace-via-rename and replace-via-truncate patterns
0334 and force that any delayed allocation blocks are allocated such that at
0335 the next journal commit, in the default data=ordered mode, the data
0336 blocks of the new file are forced to disk before the rename() operation
0337 is committed. This provides roughly the same level of guarantees as
0338 ext3, and avoids the "zero-length" problem that can happen when a
0339 system crashes before the delayed allocation blocks are forced to disk.
0340
0341 noinit_itable
0342 Do not initialize any uninitialized inode table blocks in the
0343 background. This feature may be used by installation CD's so that the
0344 install process can complete as quickly as possible; the inode table
0345 initialization process would then be deferred until the next time the
0346 file system is unmounted.
0347
0348 init_itable=n
0349 The lazy itable init code will wait n times the number of milliseconds
0350 it took to zero out the previous block group's inode table. This
0351 minimizes the impact on the system performance while file system's
0352 inode table is being initialized.
0353
0354 discard, nodiscard(*)
0355 Controls whether ext4 should issue discard/TRIM commands to the
0356 underlying block device when blocks are freed. This is useful for SSD
0357 devices and sparse/thinly-provisioned LUNs, but it is off by default
0358 until sufficient testing has been done.
0359
0360 nouid32
0361 Disables 32-bit UIDs and GIDs. This is for interoperability with
0362 older kernels which only store and expect 16-bit values.
0363
0364 block_validity(*), noblock_validity
0365 These options enable or disable the in-kernel facility for tracking
0366 filesystem metadata blocks within internal data structures. This
0367 allows multi- block allocator and other routines to notice bugs or
0368 corrupted allocation bitmaps which cause blocks to be allocated which
0369 overlap with filesystem metadata blocks.
0370
0371 dioread_lock, dioread_nolock
0372 Controls whether or not ext4 should use the DIO read locking. If the
0373 dioread_nolock option is specified ext4 will allocate uninitialized
0374 extent before buffer write and convert the extent to initialized after
0375 IO completes. This approach allows ext4 code to avoid using inode
0376 mutex, which improves scalability on high speed storages. However this
0377 does not work with data journaling and dioread_nolock option will be
0378 ignored with kernel warning. Note that dioread_nolock code path is only
0379 used for extent-based files. Because of the restrictions this options
0380 comprises it is off by default (e.g. dioread_lock).
0381
0382 max_dir_size_kb=n
0383 This limits the size of directories so that any attempt to expand them
0384 beyond the specified limit in kilobytes will cause an ENOSPC error.
0385 This is useful in memory constrained environments, where a very large
0386 directory can cause severe performance problems or even provoke the Out
0387 Of Memory killer. (For example, if there is only 512mb memory
0388 available, a 176mb directory may seriously cramp the system's style.)
0389
0390 i_version
0391 Enable 64-bit inode version support. This option is off by default.
0392
0393 dax
0394 Use direct access (no page cache). See
0395 Documentation/filesystems/dax.rst. Note that this option is
0396 incompatible with data=journal.
0397
0398 inlinecrypt
0399 When possible, encrypt/decrypt the contents of encrypted files using the
0400 blk-crypto framework rather than filesystem-layer encryption. This
0401 allows the use of inline encryption hardware. The on-disk format is
0402 unaffected. For more details, see
0403 Documentation/block/inline-encryption.rst.
0404
0405 Data Mode
0406 =========
0407 There are 3 different data modes:
0408
0409 * writeback mode
0410
0411 In data=writeback mode, ext4 does not journal data at all. This mode provides
0412 a similar level of journaling as that of XFS, JFS, and ReiserFS in its default
0413 mode - metadata journaling. A crash+recovery can cause incorrect data to
0414 appear in files which were written shortly before the crash. This mode will
0415 typically provide the best ext4 performance.
0416
0417 * ordered mode
0418
0419 In data=ordered mode, ext4 only officially journals metadata, but it logically
0420 groups metadata information related to data changes with the data blocks into
0421 a single unit called a transaction. When it's time to write the new metadata
0422 out to disk, the associated data blocks are written first. In general, this
0423 mode performs slightly slower than writeback but significantly faster than
0424 journal mode.
0425
0426 * journal mode
0427
0428 data=journal mode provides full data and metadata journaling. All new data is
0429 written to the journal first, and then to its final location. In the event of
0430 a crash, the journal can be replayed, bringing both data and metadata into a
0431 consistent state. This mode is the slowest except when data needs to be read
0432 from and written to disk at the same time where it outperforms all others
0433 modes. Enabling this mode will disable delayed allocation and O_DIRECT
0434 support.
0435
0436 /proc entries
0437 =============
0438
0439 Information about mounted ext4 file systems can be found in
0440 /proc/fs/ext4. Each mounted filesystem will have a directory in
0441 /proc/fs/ext4 based on its device name (i.e., /proc/fs/ext4/hdc or
0442 /proc/fs/ext4/dm-0). The files in each per-device directory are shown
0443 in table below.
0444
0445 Files in /proc/fs/ext4/<devname>
0446
0447 mb_groups
0448 details of multiblock allocator buddy cache of free blocks
0449
0450 /sys entries
0451 ============
0452
0453 Information about mounted ext4 file systems can be found in
0454 /sys/fs/ext4. Each mounted filesystem will have a directory in
0455 /sys/fs/ext4 based on its device name (i.e., /sys/fs/ext4/hdc or
0456 /sys/fs/ext4/dm-0). The files in each per-device directory are shown
0457 in table below.
0458
0459 Files in /sys/fs/ext4/<devname>:
0460
0461 (see also Documentation/ABI/testing/sysfs-fs-ext4)
0462
0463 delayed_allocation_blocks
0464 This file is read-only and shows the number of blocks that are dirty in
0465 the page cache, but which do not have their location in the filesystem
0466 allocated yet.
0467
0468 inode_goal
0469 Tuning parameter which (if non-zero) controls the goal inode used by
0470 the inode allocator in preference to all other allocation heuristics.
0471 This is intended for debugging use only, and should be 0 on production
0472 systems.
0473
0474 inode_readahead_blks
0475 Tuning parameter which controls the maximum number of inode table
0476 blocks that ext4's inode table readahead algorithm will pre-read into
0477 the buffer cache.
0478
0479 lifetime_write_kbytes
0480 This file is read-only and shows the number of kilobytes of data that
0481 have been written to this filesystem since it was created.
0482
0483 max_writeback_mb_bump
0484 The maximum number of megabytes the writeback code will try to write
0485 out before move on to another inode.
0486
0487 mb_group_prealloc
0488 The multiblock allocator will round up allocation requests to a
0489 multiple of this tuning parameter if the stripe size is not set in the
0490 ext4 superblock
0491
0492 mb_max_inode_prealloc
0493 The maximum length of per-inode ext4_prealloc_space list.
0494
0495 mb_max_to_scan
0496 The maximum number of extents the multiblock allocator will search to
0497 find the best extent.
0498
0499 mb_min_to_scan
0500 The minimum number of extents the multiblock allocator will search to
0501 find the best extent.
0502
0503 mb_order2_req
0504 Tuning parameter which controls the minimum size for requests (as a
0505 power of 2) where the buddy cache is used.
0506
0507 mb_stats
0508 Controls whether the multiblock allocator should collect statistics,
0509 which are shown during the unmount. 1 means to collect statistics, 0
0510 means not to collect statistics.
0511
0512 mb_stream_req
0513 Files which have fewer blocks than this tunable parameter will have
0514 their blocks allocated out of a block group specific preallocation
0515 pool, so that small files are packed closely together. Each large file
0516 will have its blocks allocated out of its own unique preallocation
0517 pool.
0518
0519 session_write_kbytes
0520 This file is read-only and shows the number of kilobytes of data that
0521 have been written to this filesystem since it was mounted.
0522
0523 reserved_clusters
0524 This is RW file and contains number of reserved clusters in the file
0525 system which will be used in the specific situations to avoid costly
0526 zeroout, unexpected ENOSPC, or possible data loss. The default is 2% or
0527 4096 clusters, whichever is smaller and this can be changed however it
0528 can never exceed number of clusters in the file system. If there is not
0529 enough space for the reserved space when mounting the file mount will
0530 _not_ fail.
0531
0532 Ioctls
0533 ======
0534
0535 Ext4 implements various ioctls which can be used by applications to access
0536 ext4-specific functionality. An incomplete list of these ioctls is shown in the
0537 table below. This list includes truly ext4-specific ioctls (``EXT4_IOC_*``) as
0538 well as ioctls that may have been ext4-specific originally but are now supported
0539 by some other filesystem(s) too (``FS_IOC_*``).
0540
0541 Table of Ext4 ioctls
0542
0543 FS_IOC_GETFLAGS
0544 Get additional attributes associated with inode. The ioctl argument is
0545 an integer bitfield, with bit values described in ext4.h.
0546
0547 FS_IOC_SETFLAGS
0548 Set additional attributes associated with inode. The ioctl argument is
0549 an integer bitfield, with bit values described in ext4.h.
0550
0551 EXT4_IOC_GETVERSION, EXT4_IOC_GETVERSION_OLD
0552 Get the inode i_generation number stored for each inode. The
0553 i_generation number is normally changed only when new inode is created
0554 and it is particularly useful for network filesystems. The '_OLD'
0555 version of this ioctl is an alias for FS_IOC_GETVERSION.
0556
0557 EXT4_IOC_SETVERSION, EXT4_IOC_SETVERSION_OLD
0558 Set the inode i_generation number stored for each inode. The '_OLD'
0559 version of this ioctl is an alias for FS_IOC_SETVERSION.
0560
0561 EXT4_IOC_GROUP_EXTEND
0562 This ioctl has the same purpose as the resize mount option. It allows
0563 to resize filesystem to the end of the last existing block group,
0564 further resize has to be done with resize2fs, either online, or
0565 offline. The argument points to the unsigned logn number representing
0566 the filesystem new block count.
0567
0568 EXT4_IOC_MOVE_EXT
0569 Move the block extents from orig_fd (the one this ioctl is pointing to)
0570 to the donor_fd (the one specified in move_extent structure passed as
0571 an argument to this ioctl). Then, exchange inode metadata between
0572 orig_fd and donor_fd. This is especially useful for online
0573 defragmentation, because the allocator has the opportunity to allocate
0574 moved blocks better, ideally into one contiguous extent.
0575
0576 EXT4_IOC_GROUP_ADD
0577 Add a new group descriptor to an existing or new group descriptor
0578 block. The new group descriptor is described by ext4_new_group_input
0579 structure, which is passed as an argument to this ioctl. This is
0580 especially useful in conjunction with EXT4_IOC_GROUP_EXTEND, which
0581 allows online resize of the filesystem to the end of the last existing
0582 block group. Those two ioctls combined is used in userspace online
0583 resize tool (e.g. resize2fs).
0584
0585 EXT4_IOC_MIGRATE
0586 This ioctl operates on the filesystem itself. It converts (migrates)
0587 ext3 indirect block mapped inode to ext4 extent mapped inode by walking
0588 through indirect block mapping of the original inode and converting
0589 contiguous block ranges into ext4 extents of the temporary inode. Then,
0590 inodes are swapped. This ioctl might help, when migrating from ext3 to
0591 ext4 filesystem, however suggestion is to create fresh ext4 filesystem
0592 and copy data from the backup. Note, that filesystem has to support
0593 extents for this ioctl to work.
0594
0595 EXT4_IOC_ALLOC_DA_BLKS
0596 Force all of the delay allocated blocks to be allocated to preserve
0597 application-expected ext3 behaviour. Note that this will also start
0598 triggering a write of the data blocks, but this behaviour may change in
0599 the future as it is not necessary and has been done this way only for
0600 sake of simplicity.
0601
0602 EXT4_IOC_RESIZE_FS
0603 Resize the filesystem to a new size. The number of blocks of resized
0604 filesystem is passed in via 64 bit integer argument. The kernel
0605 allocates bitmaps and inode table, the userspace tool thus just passes
0606 the new number of blocks.
0607
0608 EXT4_IOC_SWAP_BOOT
0609 Swap i_blocks and associated attributes (like i_blocks, i_size,
0610 i_flags, ...) from the specified inode with inode EXT4_BOOT_LOADER_INO
0611 (#5). This is typically used to store a boot loader in a secure part of
0612 the filesystem, where it can't be changed by a normal user by accident.
0613 The data blocks of the previous boot loader will be associated with the
0614 given inode.
0615
0616 References
0617 ==========
0618
0619 kernel source: <file:fs/ext4/>
0620 <file:fs/jbd2/>
0621
0622 programs: http://e2fsprogs.sourceforge.net/
0623
0624 useful links: https://fedoraproject.org/wiki/ext3-devel
0625 http://www.bullopensource.org/ext4/
0626 http://ext4.wiki.kernel.org/index.php/Main_Page
0627 https://fedoraproject.org/wiki/Features/Ext4
