Skip to content

L
linux
Commit c0e3e040 authored by Darrick J. Wong's avatar Darrick J. Wong Committed by Theodore Ts'o
Browse files
Options

docs: make ext4 readme tables readable 

The tables in the ext4 readme are not particularly space efficient in
the text or html outputs, and they're totally broken in the pdf output.
Convert them into titled paragraphs so that they render more nicely.
Signed-off-by: default avatarDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: default avatarTheodore Ts'o <tytso@mit.edu>
parent de7abd7b
Hide whitespace changes
Inline Side-by-side
Showing with 391 additions and 430 deletions
Documentation/filesystems/ext4/ext4.rst
View file @ c0e3e040
...@@ -101,269 +101,256 @@ Options ...@@ -101,269 +101,256 @@ Options
When mounting an ext4 filesystem, the following option are accepted: When mounting an ext4 filesystem, the following option are accepted:
(*) == default (*) == default
======================= ======================================================= ro
Mount Option Description Mount filesystem read only. Note that ext4 will replay the journal (and
======================= ======================================================= thus write to the partition) even when mounted "read only". The mount
ro Mount filesystem read only. Note that ext4 will options "ro,noload" can be used to prevent writes to the filesystem.
replay the journal (and thus write to the
partition) even when mounted "read only". The journal_checksum
mount options "ro,noload" can be used to prevent Enable checksumming of the journal transactions. This will allow the
writes to the filesystem. recovery code in e2fsck and the kernel to detect corruption in the
kernel. It is a compatible change and will be ignored by older
journal_checksum Enable checksumming of the journal transactions. kernels.
This will allow the recovery code in e2fsck and the
kernel to detect corruption in the kernel. It is a journal_async_commit
compatible change and will be ignored by older kernels. Commit block can be written to disk without waiting for descriptor
blocks. If enabled older kernels cannot mount the device. This will
journal_async_commit Commit block can be written to disk without waiting enable 'journal_checksum' internally.
for descriptor blocks. If enabled older kernels cannot
mount the device. This will enable 'journal_checksum' journal_path=path, journal_dev=devnum
internally. When the external journal device's major/minor numbers have changed,
these options allow the user to specify the new journal location. The
journal_path=path journal device is identified through either its new major/minor numbers
journal_dev=devnum When the external journal device's major/minor numbers encoded in devnum, or via a path to the device.
have changed, these options allow the user to specify
the new journal location. The journal device is norecovery, noload
identified through either its new major/minor numbers Don't load the journal on mounting. Note that if the filesystem was
encoded in devnum, or via a path to the device. not unmounted cleanly, skipping the journal replay will lead to the
filesystem containing inconsistencies that can lead to any number of
norecovery Don't load the journal on mounting. Note that problems.
noload if the filesystem was not unmounted cleanly,
skipping the journal replay will lead to the data=journal
filesystem containing inconsistencies that can All data are committed into the journal prior to being written into the
lead to any number of problems. main file system. Enabling this mode will disable delayed allocation
and O_DIRECT support.
data=journal All data are committed into the journal prior to being
written into the main file system. Enabling data=ordered (*)
this mode will disable delayed allocation and All data are forced directly out to the main file system prior to its
O_DIRECT support. metadata being committed to the journal.
data=ordered (*) All data are forced directly out to the main file data=writeback
system prior to its metadata being committed to the Data ordering is not preserved, data may be written into the main file
journal. system after its metadata has been committed to the journal.
data=writeback Data ordering is not preserved, data may be written commit=nrsec (*)
into the main file system after its metadata has been Ext4 can be told to sync all its data and metadata every 'nrsec'
committed to the journal. seconds. The default value is 5 seconds. This means that if you lose
your power, you will lose as much as the latest 5 seconds of work (your
commit=nrsec (*) Ext4 can be told to sync all its data and metadata filesystem will not be damaged though, thanks to the journaling). This
every 'nrsec' seconds. The default value is 5 seconds. default value (or any low value) will hurt performance, but it's good
This means that if you lose your power, you will lose for data-safety. Setting it to 0 will have the same effect as leaving
as much as the latest 5 seconds of work (your it at the default (5 seconds). Setting it to very large values will
filesystem will not be damaged though, thanks to the improve performance.
journaling). This default value (or any low value)
will hurt performance, but it's good for data-safety. barrier=<0|1(*)>, barrier(*), nobarrier
Setting it to 0 will have the same effect as leaving This enables/disables the use of write barriers in the jbd code.
it at the default (5 seconds). barrier=0 disables, barrier=1 enables. This also requires an IO stack
Setting it to very large values will improve which can support barriers, and if jbd gets an error on a barrier
performance. write, it will disable again with a warning. Write barriers enforce
proper on-disk ordering of journal commits, making volatile disk write
barrier=<0|1(*)> This enables/disables the use of write barriers in caches safe to use, at some performance penalty. If your disks are
barrier(*) the jbd code. barrier=0 disables, barrier=1 enables. battery-backed in one way or another, disabling barriers may safely
nobarrier This also requires an IO stack which can support improve performance. The mount options "barrier" and "nobarrier" can
barriers, and if jbd gets an error on a barrier also be used to enable or disable barriers, for consistency with other
write, it will disable again with a warning. ext4 mount options.
Write barriers enforce proper on-disk ordering
of journal commits, making volatile disk write caches inode_readahead_blks=n
safe to use, at some performance penalty. If This tuning parameter controls the maximum number of inode table blocks
your disks are battery-backed in one way or another, that ext4's inode table readahead algorithm will pre-read into the
disabling barriers may safely improve performance. buffer cache. The default value is 32 blocks.
The mount options "barrier" and "nobarrier" can
also be used to enable or disable barriers, for nouser_xattr
consistency with other ext4 mount options. Disables Extended User Attributes. See the attr(5) manual page for
more information about extended attributes.
inode_readahead_blks=n This tuning parameter controls the maximum
number of inode table blocks that ext4's inode noacl
table readahead algorithm will pre-read into This option disables POSIX Access Control List support. If ACL support
the buffer cache. The default value is 32 blocks. is enabled in the kernel configuration (CONFIG_EXT4_FS_POSIX_ACL), ACL
is enabled by default on mount. See the acl(5) manual page for more
nouser_xattr Disables Extended User Attributes. See the information about acl.
attr(5) manual page for more information about
extended attributes. bsddf (*)
Make 'df' act like BSD.
noacl This option disables POSIX Access Control List
support. If ACL support is enabled in the kernel minixdf
configuration (CONFIG_EXT4_FS_POSIX_ACL), ACL is Make 'df' act like Minix.
enabled by default on mount. See the acl(5) manual
page for more information about acl. debug
Extra debugging information is sent to syslog.
bsddf (*) Make 'df' act like BSD.
minixdf Make 'df' act like Minix. abort
Simulate the effects of calling ext4_abort() for debugging purposes.
debug Extra debugging information is sent to syslog. This is normally used while remounting a filesystem which is already
mounted.
abort Simulate the effects of calling ext4_abort() for
debugging purposes. This is normally used while errors=remount-ro
remounting a filesystem which is already mounted. Remount the filesystem read-only on an error.
errors=remount-ro Remount the filesystem read-only on an error. errors=continue
errors=continue Keep going on a filesystem error. Keep going on a filesystem error.
errors=panic Panic and halt the machine if an error occurs.
(These mount options override the errors behavior errors=panic
specified in the superblock, which can be configured Panic and halt the machine if an error occurs. (These mount options
using tune2fs) override the errors behavior specified in the superblock, which can be
configured using tune2fs)
data_err=ignore(*) Just print an error message if an error occurs
in a file data buffer in ordered mode. data_err=ignore(*)
data_err=abort Abort the journal if an error occurs in a file Just print an error message if an error occurs in a file data buffer in
data buffer in ordered mode. ordered mode.
data_err=abort
grpid New objects have the group ID of their parent. Abort the journal if an error occurs in a file data buffer in ordered
bsdgroups mode.
nogrpid (*) New objects have the group ID of their creator. grpid | bsdgroups
sysvgroups New objects have the group ID of their parent.
resgid=n The group ID which may use the reserved blocks. nogrpid (*) | sysvgroups
New objects have the group ID of their creator.
resuid=n The user ID which may use the reserved blocks.
resgid=n
sb=n Use alternate superblock at this location. The group ID which may use the reserved blocks.
quota These options are ignored by the filesystem. They resuid=n
noquota are used only by quota tools to recognize volumes The user ID which may use the reserved blocks.
grpquota where quota should be turned on. See documentation
usrquota in the quota-tools package for more details sb=
(http://sourceforge.net/projects/linuxquota). Use alternate superblock at this location.
jqfmt=<quota type> These options tell filesystem details about quota quota, noquota, grpquota, usrquota
usrjquota=<file> so that quota information can be properly updated These options are ignored by the filesystem. They are used only by
grpjquota=<file> during journal replay. They replace the above quota tools to recognize volumes where quota should be turned on. See
quota options. See documentation in the quota-tools documentation in the quota-tools package for more details
package for more details (http://sourceforge.net/projects/linuxquota).
(http://sourceforge.net/projects/linuxquota).
jqfmt=<quota type>, usrjquota=<file>, grpjquota=<file>
stripe=n Number of filesystem blocks that mballoc will try These options tell filesystem details about quota so that quota
to use for allocation size and alignment. For RAID5/6 information can be properly updated during journal replay. They replace
systems this should be the number of data the above quota options. See documentation in the quota-tools package
disks * RAID chunk size in file system blocks. for more details (http://sourceforge.net/projects/linuxquota).
delalloc (*) Defer block allocation until just before ext4 stripe=n
writes out the block(s) in question. This Number of filesystem blocks that mballoc will try to use for allocation
allows ext4 to better allocation decisions size and alignment. For RAID5/6 systems this should be the number of
more efficiently. data disks * RAID chunk size in file system blocks.
nodelalloc Disable delayed allocation. Blocks are allocated
when the data is copied from userspace to the delalloc (*)
page cache, either via the write(2) system call Defer block allocation until just before ext4 writes out the block(s)
or when an mmap'ed page which was previously in question. This allows ext4 to better allocation decisions more
unallocated is written for the first time. efficiently.
max_batch_time=usec Maximum amount of time ext4 should wait for nodelalloc
additional filesystem operations to be batch Disable delayed allocation. Blocks are allocated when the data is
together with a synchronous write operation. copied from userspace to the page cache, either via the write(2) system
Since a synchronous write operation is going to call or when an mmap'ed page which was previously unallocated is
force a commit and then a wait for the I/O written for the first time.
complete, it doesn't cost much, and can be a
huge throughput win, we wait for a small amount max_batch_time=usec
of time to see if any other transactions can Maximum amount of time ext4 should wait for additional filesystem
piggyback on the synchronous write. The operations to be batch together with a synchronous write operation.
algorithm used is designed to automatically tune Since a synchronous write operation is going to force a commit and then
for the speed of the disk, by measuring the a wait for the I/O complete, it doesn't cost much, and can be a huge
amount of time (on average) that it takes to throughput win, we wait for a small amount of time to see if any other
finish committing a transaction. Call this time transactions can piggyback on the synchronous write. The algorithm
the "commit time". If the time that the used is designed to automatically tune for the speed of the disk, by
transaction has been running is less than the measuring the amount of time (on average) that it takes to finish
commit time, ext4 will try sleeping for the committing a transaction. Call this time the "commit time". If the
commit time to see if other operations will join time that the transaction has been running is less than the commit
the transaction. The commit time is capped by time, ext4 will try sleeping for the commit time to see if other
the max_batch_time, which defaults to 15000us operations will join the transaction. The commit time is capped by
(15ms). This optimization can be turned off the max_batch_time, which defaults to 15000us (15ms). This
entirely by setting max_batch_time to 0. optimization can be turned off entirely by setting max_batch_time to 0.
min_batch_time=usec This parameter sets the commit time (as min_batch_time=usec
described above) to be at least min_batch_time. This parameter sets the commit time (as described above) to be at least
It defaults to zero microseconds. Increasing min_batch_time. It defaults to zero microseconds. Increasing this
this parameter may improve the throughput of parameter may improve the throughput of multi-threaded, synchronous
multi-threaded, synchronous workloads on very workloads on very fast disks, at the cost of increasing latency.
fast disks, at the cost of increasing latency.
journal_ioprio=prio
journal_ioprio=prio The I/O priority (from 0 to 7, where 0 is the The I/O priority (from 0 to 7, where 0 is the highest priority) which
highest priority) which should be used for I/O should be used for I/O operations submitted by kjournald2 during a
operations submitted by kjournald2 during a commit operation. This defaults to 3, which is a slightly higher
commit operation. This defaults to 3, which is priority than the default I/O priority.
a slightly higher priority than the default I/O
priority. auto_da_alloc(*), noauto_da_alloc
Many broken applications don't use fsync() when replacing existing
auto_da_alloc(*) Many broken applications don't use fsync() when files via patterns such as fd = open("foo.new")/write(fd,..)/close(fd)/
noauto_da_alloc replacing existing files via patterns such as rename("foo.new", "foo"), or worse yet, fd = open("foo",
fd = open("foo.new")/write(fd,..)/close(fd)/ O_TRUNC)/write(fd,..)/close(fd). If auto_da_alloc is enabled, ext4
rename("foo.new", "foo"), or worse yet, will detect the replace-via-rename and replace-via-truncate patterns
fd = open("foo", O_TRUNC)/write(fd,..)/close(fd). and force that any delayed allocation blocks are allocated such that at
If auto_da_alloc is enabled, ext4 will detect the next journal commit, in the default data=ordered mode, the data
the replace-via-rename and replace-via-truncate blocks of the new file are forced to disk before the rename() operation
patterns and force that any delayed allocation is committed. This provides roughly the same level of guarantees as
blocks are allocated such that at the next ext3, and avoids the "zero-length" problem that can happen when a
journal commit, in the default data=ordered system crashes before the delayed allocation blocks are forced to disk.
mode, the data blocks of the new file are forced
to disk before the rename() operation is noinit_itable
committed. This provides roughly the same level Do not initialize any uninitialized inode table blocks in the
of guarantees as ext3, and avoids the background. This feature may be used by installation CD's so that the
"zero-length" problem that can happen when a install process can complete as quickly as possible; the inode table
system crashes before the delayed allocation initialization process would then be deferred until the next time the
blocks are forced to disk. file system is unmounted.
noinit_itable Do not initialize any uninitialized inode table init_itable=n
blocks in the background. This feature may be The lazy itable init code will wait n times the number of milliseconds
used by installation CD's so that the install it took to zero out the previous block group's inode table. This
process can complete as quickly as possible; the minimizes the impact on the system performance while file system's
inode table initialization process would then be inode table is being initialized.
deferred until the next time the file system
is unmounted. discard, nodiscard(*)
Controls whether ext4 should issue discard/TRIM commands to the
init_itable=n The lazy itable init code will wait n times the underlying block device when blocks are freed. This is useful for SSD
number of milliseconds it took to zero out the devices and sparse/thinly-provisioned LUNs, but it is off by default
previous block group's inode table. This until sufficient testing has been done.
minimizes the impact on the system performance
while file system's inode table is being initialized. nouid32
Disables 32-bit UIDs and GIDs. This is for interoperability with
discard Controls whether ext4 should issue discard/TRIM older kernels which only store and expect 16-bit values.
nodiscard(*) commands to the underlying block device when
blocks are freed. This is useful for SSD devices block_validity(*), noblock_validity
and sparse/thinly-provisioned LUNs, but it is off These options enable or disable the in-kernel facility for tracking
by default until sufficient testing has been done. filesystem metadata blocks within internal data structures. This
allows multi- block allocator and other routines to notice bugs or
nouid32 Disables 32-bit UIDs and GIDs. This is for corrupted allocation bitmaps which cause blocks to be allocated which
interoperability with older kernels which only overlap with filesystem metadata blocks.
store and expect 16-bit values.
dioread_lock, dioread_nolock
block_validity(*) These options enable or disable the in-kernel Controls whether or not ext4 should use the DIO read locking. If the
noblock_validity facility for tracking filesystem metadata blocks dioread_nolock option is specified ext4 will allocate uninitialized
within internal data structures. This allows multi- extent before buffer write and convert the extent to initialized after
block allocator and other routines to notice IO completes. This approach allows ext4 code to avoid using inode
bugs or corrupted allocation bitmaps which cause mutex, which improves scalability on high speed storages. However this
blocks to be allocated which overlap with does not work with data journaling and dioread_nolock option will be
filesystem metadata blocks. ignored with kernel warning. Note that dioread_nolock code path is only
used for extent-based files. Because of the restrictions this options
dioread_lock Controls whether or not ext4 should use the DIO read comprises it is off by default (e.g. dioread_lock).
dioread_nolock locking. If the dioread_nolock option is specified
ext4 will allocate uninitialized extent before buffer max_dir_size_kb=n
write and convert the extent to initialized after IO This limits the size of directories so that any attempt to expand them
completes. This approach allows ext4 code to avoid beyond the specified limit in kilobytes will cause an ENOSPC error.
using inode mutex, which improves scalability on high This is useful in memory constrained environments, where a very large
speed storages. However this does not work with directory can cause severe performance problems or even provoke the Out
data journaling and dioread_nolock option will be Of Memory killer. (For example, if there is only 512mb memory
ignored with kernel warning. Note that dioread_nolock available, a 176mb directory may seriously cramp the system's style.)
code path is only used for extent-based files.
Because of the restrictions this options comprises i_version
it is off by default (e.g. dioread_lock). Enable 64-bit inode version support. This option is off by default.
max_dir_size_kb=n This limits the size of directories so that any dax
attempt to expand them beyond the specified Use direct access (no page cache). See
limit in kilobytes will cause an ENOSPC error. Documentation/filesystems/dax.txt. Note that this option is
This is useful in memory constrained incompatible with data=journal.
environments, where a very large directory can
cause severe performance problems or even
provoke the Out Of Memory killer. (For example,
if there is only 512mb memory available, a 176mb
directory may seriously cramp the system's style.)
i_version Enable 64-bit inode version support. This option is
off by default.
dax Use direct access (no page cache). See
Documentation/filesystems/dax.txt. Note that
this option is incompatible with data=journal.
======================= =======================================================
Data Mode Data Mode
========= =========
...@@ -407,11 +394,8 @@ in table below. ...@@ -407,11 +394,8 @@ in table below.
Files in /proc/fs/ext4/<devname> Files in /proc/fs/ext4/<devname>
================ ======= mb_groups
File Content details of multiblock allocator buddy cache of free blocks
================ =======
mb_groups details of multiblock allocator buddy cache of free blocks
================ =======
/sys entries /sys entries
============ ============
...@@ -426,74 +410,71 @@ Files in /sys/fs/ext4/<devname>: ...@@ -426,74 +410,71 @@ Files in /sys/fs/ext4/<devname>:
(see also Documentation/ABI/testing/sysfs-fs-ext4) (see also Documentation/ABI/testing/sysfs-fs-ext4)
============================= ================================================= delayed_allocation_blocks
File Content This file is read-only and shows the number of blocks that are dirty in
============================= ================================================= the page cache, but which do not have their location in the filesystem
delayed_allocation_blocks This file is read-only and shows the number of allocated yet.
blocks that are dirty in the page cache, but
which do not have their location in the inode_goal
filesystem allocated yet. Tuning parameter which (if non-zero) controls the goal inode used by
the inode allocator in preference to all other allocation heuristics.
inode_goal Tuning parameter which (if non-zero) controls This is intended for debugging use only, and should be 0 on production
the goal inode used by the inode allocator in systems.
preference to all other allocation heuristics.
This is intended for debugging use only, and inode_readahead_blks
should be 0 on production systems. Tuning parameter which controls the maximum number of inode table
blocks that ext4's inode table readahead algorithm will pre-read into
inode_readahead_blks Tuning parameter which controls the maximum the buffer cache.
number of inode table blocks that ext4's inode
table readahead algorithm will pre-read into lifetime_write_kbytes
the buffer cache This file is read-only and shows the number of kilobytes of data that
have been written to this filesystem since it was created.
lifetime_write_kbytes This file is read-only and shows the number of
kilobytes of data that have been written to this max_writeback_mb_bump
filesystem since it was created. The maximum number of megabytes the writeback code will try to write
out before move on to another inode.
max_writeback_mb_bump The maximum number of megabytes the writeback
code will try to write out before move on to mb_group_prealloc
another inode. The multiblock allocator will round up allocation requests to a
multiple of this tuning parameter if the stripe size is not set in the
mb_group_prealloc The multiblock allocator will round up allocation ext4 superblock
requests to a multiple of this tuning parameter if
the stripe size is not set in the ext4 superblock mb_max_to_scan
The maximum number of extents the multiblock allocator will search to
mb_max_to_scan The maximum number of extents the multiblock find the best extent.
allocator will search to find the best extent
mb_min_to_scan
mb_min_to_scan The minimum number of extents the multiblock The minimum number of extents the multiblock allocator will search to
allocator will search to find the best extent find the best extent.
mb_order2_req Tuning parameter which controls the minimum size mb_order2_req
for requests (as a power of 2) where the buddy Tuning parameter which controls the minimum size for requests (as a
cache is used power of 2) where the buddy cache is used.
mb_stats Controls whether the multiblock allocator should mb_stats
collect statistics, which are shown during the Controls whether the multiblock allocator should collect statistics,
unmount. 1 means to collect statistics, 0 means which are shown during the unmount. 1 means to collect statistics, 0
not to collect statistics means not to collect statistics.
mb_stream_req Files which have fewer blocks than this tunable mb_stream_req
parameter will have their blocks allocated out Files which have fewer blocks than this tunable parameter will have
of a block group specific preallocation pool, so their blocks allocated out of a block group specific preallocation
that small files are packed closely together. pool, so that small files are packed closely together. Each large file
Each large file will have its blocks allocated will have its blocks allocated out of its own unique preallocation
out of its own unique preallocation pool. pool.
session_write_kbytes This file is read-only and shows the number of session_write_kbytes
kilobytes of data that have been written to this This file is read-only and shows the number of kilobytes of data that
filesystem since it was mounted. have been written to this filesystem since it was mounted.
reserved_clusters This is RW file and contains number of reserved reserved_clusters
clusters in the file system which will be used This is RW file and contains number of reserved clusters in the file
in the specific situations to avoid costly system which will be used in the specific situations to avoid costly
zeroout, unexpected ENOSPC, or possible data zeroout, unexpected ENOSPC, or possible data loss. The default is 2% or
loss. The default is 2% or 4096 clusters, 4096 clusters, whichever is smaller and this can be changed however it
whichever is smaller and this can be changed can never exceed number of clusters in the file system. If there is not
however it can never exceed number of clusters enough space for the reserved space when mounting the file mount will
in the file system. If there is not enough space _not_ fail.
for the reserved space when mounting the file
mount will _not_ fail.
============================= =================================================
Ioctls Ioctls
====== ======
...@@ -504,100 +485,80 @@ shown in the table below. ...@@ -504,100 +485,80 @@ shown in the table below.
Table of Ext4 specific ioctls Table of Ext4 specific ioctls
============================= ================================================= EXT4_IOC_GETFLAGS
Ioctl Description Get additional attributes associated with inode. The ioctl argument is
============================= ================================================= an integer bitfield, with bit values described in ext4.h. This ioctl is
EXT4_IOC_GETFLAGS Get additional attributes associated with inode. an alias for FS_IOC_GETFLAGS.
The ioctl argument is an integer bitfield, with
bit values described in ext4.h. This ioctl is an EXT4_IOC_SETFLAGS
alias for FS_IOC_GETFLAGS. Set additional attributes associated with inode. The ioctl argument is
an integer bitfield, with bit values described in ext4.h. This ioctl is
EXT4_IOC_SETFLAGS Set additional attributes associated with inode. an alias for FS_IOC_SETFLAGS.
The ioctl argument is an integer bitfield, with
bit values described in ext4.h. This ioctl is an EXT4_IOC_GETVERSION, EXT4_IOC_GETVERSION_OLD
alias for FS_IOC_SETFLAGS. Get the inode i_generation number stored for each inode. The
i_generation number is normally changed only when new inode is created
EXT4_IOC_GETVERSION and it is particularly useful for network filesystems. The '_OLD'
EXT4_IOC_GETVERSION_OLD version of this ioctl is an alias for FS_IOC_GETVERSION.
Get the inode i_generation number stored for
each inode. The i_generation number is normally EXT4_IOC_SETVERSION, EXT4_IOC_SETVERSION_OLD
changed only when new inode is created and it is Set the inode i_generation number stored for each inode. The '_OLD'
particularly useful for network filesystems. The version of this ioctl is an alias for FS_IOC_SETVERSION.
'_OLD' version of this ioctl is an alias for
FS_IOC_GETVERSION. EXT4_IOC_GROUP_EXTEND
This ioctl has the same purpose as the resize mount option. It allows
EXT4_IOC_SETVERSION to resize filesystem to the end of the last existing block group,
EXT4_IOC_SETVERSION_OLD further resize has to be done with resize2fs, either online, or
Set the inode i_generation number stored for offline. The argument points to the unsigned logn number representing
each inode. The '_OLD' version of this ioctl the filesystem new block count.
is an alias for FS_IOC_SETVERSION.
EXT4_IOC_MOVE_EXT
EXT4_IOC_GROUP_EXTEND This ioctl has the same purpose as the resize Move the block extents from orig_fd (the one this ioctl is pointing to)
mount option. It allows to resize filesystem to the donor_fd (the one specified in move_extent structure passed as
to the end of the last existing block group, an argument to this ioctl). Then, exchange inode metadata between
further resize has to be done with resize2fs, orig_fd and donor_fd. This is especially useful for online
either online, or offline. The argument points defragmentation, because the allocator has the opportunity to allocate
to the unsigned logn number representing the moved blocks better, ideally into one contiguous extent.
filesystem new block count.
EXT4_IOC_GROUP_ADD
EXT4_IOC_MOVE_EXT Move the block extents from orig_fd (the one Add a new group descriptor to an existing or new group descriptor
this ioctl is pointing to) to the donor_fd (the block. The new group descriptor is described by ext4_new_group_input
one specified in move_extent structure passed structure, which is passed as an argument to this ioctl. This is
as an argument to this ioctl). Then, exchange especially useful in conjunction with EXT4_IOC_GROUP_EXTEND, which
inode metadata between orig_fd and donor_fd. allows online resize of the filesystem to the end of the last existing
This is especially useful for online block group. Those two ioctls combined is used in userspace online
defragmentation, because the allocator has the resize tool (e.g. resize2fs).
opportunity to allocate moved blocks better,
ideally into one contiguous extent. EXT4_IOC_MIGRATE
This ioctl operates on the filesystem itself. It converts (migrates)
EXT4_IOC_GROUP_ADD Add a new group descriptor to an existing or ext3 indirect block mapped inode to ext4 extent mapped inode by walking
new group descriptor block. The new group through indirect block mapping of the original inode and converting
descriptor is described by ext4_new_group_input contiguous block ranges into ext4 extents of the temporary inode. Then,
structure, which is passed as an argument to inodes are swapped. This ioctl might help, when migrating from ext3 to
this ioctl. This is especially useful in ext4 filesystem, however suggestion is to create fresh ext4 filesystem
conjunction with EXT4_IOC_GROUP_EXTEND, and copy data from the backup. Note, that filesystem has to support
which allows online resize of the filesystem extents for this ioctl to work.
to the end of the last existing block group.
Those two ioctls combined is used in userspace EXT4_IOC_ALLOC_DA_BLKS
online resize tool (e.g. resize2fs). Force all of the delay allocated blocks to be allocated to preserve
application-expected ext3 behaviour. Note that this will also start
EXT4_IOC_MIGRATE This ioctl operates on the filesystem itself. triggering a write of the data blocks, but this behaviour may change in
It converts (migrates) ext3 indirect block mapped the future as it is not necessary and has been done this way only for
inode to ext4 extent mapped inode by walking sake of simplicity.
through indirect block mapping of the original
inode and converting contiguous block ranges EXT4_IOC_RESIZE_FS
into ext4 extents of the temporary inode. Then, Resize the filesystem to a new size. The number of blocks of resized
inodes are swapped. This ioctl might help, when filesystem is passed in via 64 bit integer argument. The kernel
migrating from ext3 to ext4 filesystem, however allocates bitmaps and inode table, the userspace tool thus just passes
suggestion is to create fresh ext4 filesystem the new number of blocks.
and copy data from the backup. Note, that
filesystem has to support extents for this ioctl EXT4_IOC_SWAP_BOOT
to work. Swap i_blocks and associated attributes (like i_blocks, i_size,
i_flags, ...) from the specified inode with inode EXT4_BOOT_LOADER_INO
EXT4_IOC_ALLOC_DA_BLKS Force all of the delay allocated blocks to be (#5). This is typically used to store a boot loader in a secure part of
allocated to preserve application-expected ext3 the filesystem, where it can't be changed by a normal user by accident.
behaviour. Note that this will also start The data blocks of the previous boot loader will be associated with the
triggering a write of the data blocks, but this given inode.
behaviour may change in the future as it is
not necessary and has been done this way only
for sake of simplicity.
EXT4_IOC_RESIZE_FS Resize the filesystem to a new size. The number
of blocks of resized filesystem is passed in via
64 bit integer argument. The kernel allocates
bitmaps and inode table, the userspace tool thus
just passes the new number of blocks.
EXT4_IOC_SWAP_BOOT Swap i_blocks and associated attributes
(like i_blocks, i_size, i_flags, ...) from
the specified inode with inode
EXT4_BOOT_LOADER_INO (#5). This is typically
used to store a boot loader in a secure part of
the filesystem, where it can't be changed by a
normal user by accident.
The data blocks of the previous boot loader
will be associated with the given inode.
============================= =================================================
References References
========== ==========
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7