2006-10-30 06:46:40 +00:00
|
|
|
<?xml version="1.0" encoding="UTF-8"?>
|
|
|
|
<!DOCTYPE book PUBLIC "-//OASIS//DTD DocBook XML V4.1.2//EN"
|
|
|
|
"http://www.oasis-open.org/docbook/xml/4.1.2/docbookx.dtd" []>
|
|
|
|
|
|
|
|
<book id="Linux-filesystems-API">
|
|
|
|
<bookinfo>
|
|
|
|
<title>Linux Filesystems API</title>
|
|
|
|
|
|
|
|
<legalnotice>
|
|
|
|
<para>
|
|
|
|
This documentation is free software; you can redistribute
|
|
|
|
it and/or modify it under the terms of the GNU General Public
|
|
|
|
License as published by the Free Software Foundation; either
|
|
|
|
version 2 of the License, or (at your option) any later
|
|
|
|
version.
|
|
|
|
</para>
|
|
|
|
|
|
|
|
<para>
|
|
|
|
This program is distributed in the hope that it will be
|
|
|
|
useful, but WITHOUT ANY WARRANTY; without even the implied
|
|
|
|
warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
|
|
|
|
See the GNU General Public License for more details.
|
|
|
|
</para>
|
|
|
|
|
|
|
|
<para>
|
|
|
|
You should have received a copy of the GNU General Public
|
|
|
|
License along with this program; if not, write to the Free
|
|
|
|
Software Foundation, Inc., 59 Temple Place, Suite 330, Boston,
|
|
|
|
MA 02111-1307 USA
|
|
|
|
</para>
|
|
|
|
|
|
|
|
<para>
|
|
|
|
For more details see the file COPYING in the source
|
|
|
|
distribution of Linux.
|
|
|
|
</para>
|
|
|
|
</legalnotice>
|
|
|
|
</bookinfo>
|
|
|
|
|
|
|
|
<toc></toc>
|
|
|
|
|
|
|
|
<chapter id="vfs">
|
|
|
|
<title>The Linux VFS</title>
|
2007-10-17 06:31:31 +00:00
|
|
|
<sect1 id="the_filesystem_types"><title>The Filesystem types</title>
|
2006-10-30 06:46:40 +00:00
|
|
|
!Iinclude/linux/fs.h
|
|
|
|
</sect1>
|
2007-10-17 06:31:31 +00:00
|
|
|
<sect1 id="the_directory_cache"><title>The Directory Cache</title>
|
2006-10-30 06:46:40 +00:00
|
|
|
!Efs/dcache.c
|
|
|
|
!Iinclude/linux/dcache.h
|
|
|
|
</sect1>
|
2007-10-17 06:31:31 +00:00
|
|
|
<sect1 id="inode_handling"><title>Inode Handling</title>
|
2006-10-30 06:46:40 +00:00
|
|
|
!Efs/inode.c
|
|
|
|
!Efs/bad_inode.c
|
|
|
|
</sect1>
|
2007-10-17 06:31:31 +00:00
|
|
|
<sect1 id="registration_and_superblocks"><title>Registration and Superblocks</title>
|
2006-10-30 06:46:40 +00:00
|
|
|
!Efs/super.c
|
|
|
|
</sect1>
|
2007-10-17 06:31:31 +00:00
|
|
|
<sect1 id="file_locks"><title>File Locks</title>
|
2006-10-30 06:46:40 +00:00
|
|
|
!Efs/locks.c
|
|
|
|
!Ifs/locks.c
|
|
|
|
</sect1>
|
2007-10-17 06:31:31 +00:00
|
|
|
<sect1 id="other_functions"><title>Other Functions</title>
|
2006-10-30 06:46:40 +00:00
|
|
|
!Efs/mpage.c
|
|
|
|
!Efs/namei.c
|
|
|
|
!Efs/buffer.c
|
2014-05-20 14:17:35 +00:00
|
|
|
!Eblock/bio.c
|
2006-10-30 06:46:40 +00:00
|
|
|
!Efs/seq_file.c
|
|
|
|
!Efs/filesystems.c
|
|
|
|
!Efs/fs-writeback.c
|
|
|
|
!Efs/block_dev.c
|
|
|
|
</sect1>
|
|
|
|
</chapter>
|
|
|
|
|
|
|
|
<chapter id="proc">
|
|
|
|
<title>The proc filesystem</title>
|
|
|
|
|
2007-10-17 06:31:31 +00:00
|
|
|
<sect1 id="sysctl_interface"><title>sysctl interface</title>
|
2006-10-30 06:46:40 +00:00
|
|
|
!Ekernel/sysctl.c
|
|
|
|
</sect1>
|
|
|
|
|
2007-10-17 06:31:31 +00:00
|
|
|
<sect1 id="proc_filesystem_interface"><title>proc filesystem interface</title>
|
2006-10-30 06:46:40 +00:00
|
|
|
!Ifs/proc/base.c
|
|
|
|
</sect1>
|
|
|
|
</chapter>
|
|
|
|
|
2011-02-21 04:08:35 +00:00
|
|
|
<chapter id="fs_events">
|
|
|
|
<title>Events based on file descriptors</title>
|
|
|
|
!Efs/eventfd.c
|
|
|
|
</chapter>
|
|
|
|
|
2006-10-30 06:46:40 +00:00
|
|
|
<chapter id="sysfs">
|
|
|
|
<title>The Filesystem for Exporting Kernel Objects</title>
|
|
|
|
!Efs/sysfs/file.c
|
|
|
|
!Efs/sysfs/symlink.c
|
|
|
|
</chapter>
|
|
|
|
|
|
|
|
<chapter id="debugfs">
|
|
|
|
<title>The debugfs filesystem</title>
|
|
|
|
|
2007-10-17 06:31:31 +00:00
|
|
|
<sect1 id="debugfs_interface"><title>debugfs interface</title>
|
2006-10-30 06:46:40 +00:00
|
|
|
!Efs/debugfs/inode.c
|
|
|
|
!Efs/debugfs/file.c
|
|
|
|
</sect1>
|
|
|
|
</chapter>
|
|
|
|
|
2006-11-03 06:07:02 +00:00
|
|
|
<chapter id="LinuxJDBAPI">
|
|
|
|
<chapterinfo>
|
|
|
|
<title>The Linux Journalling API</title>
|
|
|
|
|
|
|
|
<authorgroup>
|
|
|
|
<author>
|
|
|
|
<firstname>Roger</firstname>
|
|
|
|
<surname>Gammans</surname>
|
|
|
|
<affiliation>
|
|
|
|
<address>
|
|
|
|
<email>rgammans@computer-surgery.co.uk</email>
|
|
|
|
</address>
|
|
|
|
</affiliation>
|
|
|
|
</author>
|
|
|
|
</authorgroup>
|
|
|
|
|
|
|
|
<authorgroup>
|
|
|
|
<author>
|
|
|
|
<firstname>Stephen</firstname>
|
|
|
|
<surname>Tweedie</surname>
|
|
|
|
<affiliation>
|
|
|
|
<address>
|
|
|
|
<email>sct@redhat.com</email>
|
|
|
|
</address>
|
|
|
|
</affiliation>
|
|
|
|
</author>
|
|
|
|
</authorgroup>
|
|
|
|
|
|
|
|
<copyright>
|
|
|
|
<year>2002</year>
|
|
|
|
<holder>Roger Gammans</holder>
|
|
|
|
</copyright>
|
|
|
|
</chapterinfo>
|
|
|
|
|
|
|
|
<title>The Linux Journalling API</title>
|
|
|
|
|
2007-10-17 06:31:31 +00:00
|
|
|
<sect1 id="journaling_overview">
|
2006-11-03 06:07:02 +00:00
|
|
|
<title>Overview</title>
|
2007-10-17 06:31:31 +00:00
|
|
|
<sect2 id="journaling_details">
|
2006-11-03 06:07:02 +00:00
|
|
|
<title>Details</title>
|
|
|
|
<para>
|
|
|
|
The journalling layer is easy to use. You need to
|
|
|
|
first of all create a journal_t data structure. There are
|
|
|
|
two calls to do this dependent on how you decide to allocate the physical
|
2015-07-14 12:55:05 +00:00
|
|
|
media on which the journal resides. The jbd2_journal_init_inode() call
|
|
|
|
is for journals stored in filesystem inodes, or the jbd2_journal_init_dev()
|
|
|
|
call can be used for journal stored on a raw device (in a continuous range
|
2006-11-03 06:07:02 +00:00
|
|
|
of blocks). A journal_t is a typedef for a struct pointer, so when
|
2015-07-14 12:55:05 +00:00
|
|
|
you are finally finished make sure you call jbd2_journal_destroy() on it
|
2006-11-03 06:07:02 +00:00
|
|
|
to free up any used kernel memory.
|
|
|
|
</para>
|
|
|
|
|
|
|
|
<para>
|
|
|
|
Once you have got your journal_t object you need to 'mount' or load the journal
|
2015-07-14 12:55:05 +00:00
|
|
|
file. The journalling layer expects the space for the journal was already
|
|
|
|
allocated and initialized properly by the userspace tools. When loading the
|
|
|
|
journal you must call jbd2_journal_load() to process journal contents. If the
|
|
|
|
client file system detects the journal contents does not need to be processed
|
|
|
|
(or even need not have valid contents), it may call jbd2_journal_wipe() to
|
|
|
|
clear the journal contents before calling jbd2_journal_load().
|
2006-11-03 06:07:02 +00:00
|
|
|
</para>
|
|
|
|
|
|
|
|
<para>
|
2015-07-14 12:55:05 +00:00
|
|
|
Note that jbd2_journal_wipe(..,0) calls jbd2_journal_skip_recovery() for you if
|
|
|
|
it detects any outstanding transactions in the journal and similarly
|
|
|
|
jbd2_journal_load() will call jbd2_journal_recover() if necessary. I would
|
|
|
|
advise reading ext4_load_journal() in fs/ext4/super.c for examples on this
|
|
|
|
stage.
|
2006-11-03 06:07:02 +00:00
|
|
|
</para>
|
|
|
|
|
|
|
|
<para>
|
|
|
|
Now you can go ahead and start modifying the underlying
|
|
|
|
filesystem. Almost.
|
|
|
|
</para>
|
|
|
|
|
|
|
|
<para>
|
|
|
|
|
|
|
|
You still need to actually journal your filesystem changes, this
|
|
|
|
is done by wrapping them into transactions. Additionally you
|
|
|
|
also need to wrap the modification of each of the buffers
|
|
|
|
with calls to the journal layer, so it knows what the modifications
|
2015-07-14 12:55:05 +00:00
|
|
|
you are actually making are. To do this use jbd2_journal_start() which
|
2006-11-03 06:07:02 +00:00
|
|
|
returns a transaction handle.
|
|
|
|
</para>
|
|
|
|
|
|
|
|
<para>
|
2015-07-14 12:55:05 +00:00
|
|
|
jbd2_journal_start()
|
|
|
|
and its counterpart jbd2_journal_stop(), which indicates the end of a
|
|
|
|
transaction are nestable calls, so you can reenter a transaction if necessary,
|
|
|
|
but remember you must call jbd2_journal_stop() the same number of times as
|
|
|
|
jbd2_journal_start() before the transaction is completed (or more accurately
|
|
|
|
leaves the update phase). Ext4/VFS makes use of this feature to simplify
|
|
|
|
handling of inode dirtying, quota support, etc.
|
2006-11-03 06:07:02 +00:00
|
|
|
</para>
|
|
|
|
|
|
|
|
<para>
|
|
|
|
Inside each transaction you need to wrap the modifications to the
|
|
|
|
individual buffers (blocks). Before you start to modify a buffer you
|
2015-07-14 12:55:05 +00:00
|
|
|
need to call jbd2_journal_get_{create,write,undo}_access() as appropriate,
|
2006-11-03 06:07:02 +00:00
|
|
|
this allows the journalling layer to copy the unmodified data if it
|
|
|
|
needs to. After all the buffer may be part of a previously uncommitted
|
|
|
|
transaction.
|
|
|
|
At this point you are at last ready to modify a buffer, and once
|
2015-07-14 12:55:05 +00:00
|
|
|
you are have done so you need to call jbd2_journal_dirty_{meta,}data().
|
2006-11-03 06:07:02 +00:00
|
|
|
Or if you've asked for access to a buffer you now know is now longer
|
2015-07-14 12:55:05 +00:00
|
|
|
required to be pushed back on the device you can call jbd2_journal_forget()
|
2006-11-03 06:07:02 +00:00
|
|
|
in much the same way as you might have used bforget() in the past.
|
|
|
|
</para>
|
|
|
|
|
|
|
|
<para>
|
2015-07-14 12:55:05 +00:00
|
|
|
A jbd2_journal_flush() may be called at any time to commit and checkpoint
|
2006-11-03 06:07:02 +00:00
|
|
|
all your transactions.
|
|
|
|
</para>
|
|
|
|
|
|
|
|
<para>
|
2015-07-14 12:55:05 +00:00
|
|
|
Then at umount time , in your put_super() you can then call jbd2_journal_destroy()
|
2012-07-25 15:12:00 +00:00
|
|
|
to clean up your in-core journal object.
|
2006-11-03 06:07:02 +00:00
|
|
|
</para>
|
|
|
|
|
|
|
|
<para>
|
|
|
|
Unfortunately there a couple of ways the journal layer can cause a deadlock.
|
|
|
|
The first thing to note is that each task can only have
|
|
|
|
a single outstanding transaction at any one time, remember nothing
|
2015-07-14 12:55:05 +00:00
|
|
|
commits until the outermost jbd2_journal_stop(). This means
|
2006-11-03 06:07:02 +00:00
|
|
|
you must complete the transaction at the end of each file/inode/address
|
|
|
|
etc. operation you perform, so that the journalling system isn't re-entered
|
|
|
|
on another journal. Since transactions can't be nested/batched
|
|
|
|
across differing journals, and another filesystem other than
|
2015-07-14 12:55:05 +00:00
|
|
|
yours (say ext4) may be modified in a later syscall.
|
2006-11-03 06:07:02 +00:00
|
|
|
</para>
|
|
|
|
|
|
|
|
<para>
|
2015-07-14 12:55:05 +00:00
|
|
|
The second case to bear in mind is that jbd2_journal_start() can
|
2006-11-03 06:07:02 +00:00
|
|
|
block if there isn't enough space in the journal for your transaction
|
|
|
|
(based on the passed nblocks param) - when it blocks it merely(!) needs to
|
|
|
|
wait for transactions to complete and be committed from other tasks,
|
2015-07-14 12:55:05 +00:00
|
|
|
so essentially we are waiting for jbd2_journal_stop(). So to avoid
|
|
|
|
deadlocks you must treat jbd2_journal_start/stop() as if they
|
2006-11-03 06:07:02 +00:00
|
|
|
were semaphores and include them in your semaphore ordering rules to prevent
|
2015-07-14 12:55:05 +00:00
|
|
|
deadlocks. Note that jbd2_journal_extend() has similar blocking behaviour to
|
|
|
|
jbd2_journal_start() so you can deadlock here just as easily as on
|
|
|
|
jbd2_journal_start().
|
2006-11-03 06:07:02 +00:00
|
|
|
</para>
|
|
|
|
|
|
|
|
<para>
|
|
|
|
Try to reserve the right number of blocks the first time. ;-). This will
|
|
|
|
be the maximum number of blocks you are going to touch in this transaction.
|
2015-07-14 12:55:05 +00:00
|
|
|
I advise having a look at at least ext4_jbd.h to see the basis on which
|
|
|
|
ext4 uses to make these decisions.
|
2006-11-03 06:07:02 +00:00
|
|
|
</para>
|
|
|
|
|
|
|
|
<para>
|
|
|
|
Another wriggle to watch out for is your on-disk block allocation strategy.
|
2015-07-14 12:55:05 +00:00
|
|
|
Why? Because, if you do a delete, you need to ensure you haven't reused any
|
|
|
|
of the freed blocks until the transaction freeing these blocks commits. If you
|
|
|
|
reused these blocks and crash happens, there is no way to restore the contents
|
|
|
|
of the reallocated blocks at the end of the last fully committed transaction.
|
|
|
|
|
|
|
|
One simple way of doing this is to mark blocks as free in internal in-memory
|
|
|
|
block allocation structures only after the transaction freeing them commits.
|
|
|
|
Ext4 uses journal commit callback for this purpose.
|
|
|
|
</para>
|
|
|
|
|
|
|
|
<para>
|
|
|
|
With journal commit callbacks you can ask the journalling layer to call a
|
|
|
|
callback function when the transaction is finally committed to disk, so that
|
|
|
|
you can do some of your own management. You ask the journalling layer for
|
|
|
|
calling the callback by simply setting journal->j_commit_callback function
|
|
|
|
pointer and that function is called after each transaction commit. You can also
|
|
|
|
use transaction->t_private_list for attaching entries to a transaction that
|
|
|
|
need processing when the transaction commits.
|
2006-11-03 06:07:02 +00:00
|
|
|
</para>
|
|
|
|
|
|
|
|
<para>
|
2015-07-14 12:55:05 +00:00
|
|
|
JBD2 also provides a way to block all transaction updates via
|
|
|
|
jbd2_journal_{un,}lock_updates(). Ext4 uses this when it wants a window with a
|
|
|
|
clean and stable fs for a moment. E.g.
|
2006-11-03 06:07:02 +00:00
|
|
|
</para>
|
|
|
|
|
|
|
|
<programlisting>
|
|
|
|
|
2015-07-14 12:55:05 +00:00
|
|
|
jbd2_journal_lock_updates() //stop new stuff happening..
|
|
|
|
jbd2_journal_flush() // checkpoint everything.
|
2006-11-03 06:07:02 +00:00
|
|
|
..do stuff on stable fs
|
2015-07-14 12:55:05 +00:00
|
|
|
jbd2_journal_unlock_updates() // carry on with filesystem use.
|
2006-11-03 06:07:02 +00:00
|
|
|
</programlisting>
|
|
|
|
|
|
|
|
<para>
|
|
|
|
The opportunities for abuse and DOS attacks with this should be obvious,
|
|
|
|
if you allow unprivileged userspace to trigger codepaths containing these
|
|
|
|
calls.
|
|
|
|
</para>
|
|
|
|
|
|
|
|
</sect2>
|
|
|
|
|
2007-10-17 06:31:31 +00:00
|
|
|
<sect2 id="jbd_summary">
|
2006-11-03 06:07:02 +00:00
|
|
|
<title>Summary</title>
|
|
|
|
<para>
|
|
|
|
Using the journal is a matter of wrapping the different context changes,
|
|
|
|
being each mount, each modification (transaction) and each changed buffer
|
|
|
|
to tell the journalling layer about them.
|
|
|
|
</para>
|
|
|
|
|
|
|
|
</sect2>
|
|
|
|
|
|
|
|
</sect1>
|
|
|
|
|
2007-10-17 06:31:31 +00:00
|
|
|
<sect1 id="data_types">
|
2006-11-03 06:07:02 +00:00
|
|
|
<title>Data Types</title>
|
|
|
|
<para>
|
|
|
|
The journalling layer uses typedefs to 'hide' the concrete definitions
|
2015-07-14 12:55:05 +00:00
|
|
|
of the structures used. As a client of the JBD2 layer you can
|
2006-11-03 06:07:02 +00:00
|
|
|
just rely on the using the pointer as a magic cookie of some sort.
|
|
|
|
|
|
|
|
Obviously the hiding is not enforced as this is 'C'.
|
|
|
|
</para>
|
2007-10-17 06:31:31 +00:00
|
|
|
<sect2 id="structures"><title>Structures</title>
|
2015-07-14 12:55:05 +00:00
|
|
|
!Iinclude/linux/jbd2.h
|
2006-11-03 06:07:02 +00:00
|
|
|
</sect2>
|
|
|
|
</sect1>
|
|
|
|
|
2007-10-17 06:31:31 +00:00
|
|
|
<sect1 id="functions">
|
2006-11-03 06:07:02 +00:00
|
|
|
<title>Functions</title>
|
|
|
|
<para>
|
|
|
|
The functions here are split into two groups those that
|
|
|
|
affect a journal as a whole, and those which are used to
|
|
|
|
manage transactions
|
|
|
|
</para>
|
2007-10-17 06:31:31 +00:00
|
|
|
<sect2 id="journal_level"><title>Journal Level</title>
|
2015-07-14 12:55:05 +00:00
|
|
|
!Efs/jbd2/journal.c
|
|
|
|
!Ifs/jbd2/recovery.c
|
2006-11-03 06:07:02 +00:00
|
|
|
</sect2>
|
2007-10-17 06:31:31 +00:00
|
|
|
<sect2 id="transaction_level"><title>Transasction Level</title>
|
2015-07-14 12:55:05 +00:00
|
|
|
!Efs/jbd2/transaction.c
|
2006-11-03 06:07:02 +00:00
|
|
|
</sect2>
|
|
|
|
</sect1>
|
2007-10-17 06:31:31 +00:00
|
|
|
<sect1 id="see_also">
|
2006-11-03 06:07:02 +00:00
|
|
|
<title>See also</title>
|
|
|
|
<para>
|
|
|
|
<citation>
|
2012-02-15 06:11:56 +00:00
|
|
|
<ulink url="http://kernel.org/pub/linux/kernel/people/sct/ext3/journal-design.ps.gz">
|
2006-11-03 06:07:02 +00:00
|
|
|
Journaling the Linux ext2fs Filesystem, LinuxExpo 98, Stephen Tweedie
|
|
|
|
</ulink>
|
|
|
|
</citation>
|
|
|
|
</para>
|
|
|
|
<para>
|
|
|
|
<citation>
|
|
|
|
<ulink url="http://olstrans.sourceforge.net/release/OLS2000-ext3/OLS2000-ext3.html">
|
|
|
|
Ext3 Journalling FileSystem, OLS 2000, Dr. Stephen Tweedie
|
|
|
|
</ulink>
|
|
|
|
</citation>
|
|
|
|
</para>
|
|
|
|
</sect1>
|
|
|
|
|
|
|
|
</chapter>
|
|
|
|
|
2008-02-13 23:03:23 +00:00
|
|
|
<chapter id="splice">
|
|
|
|
<title>splice API</title>
|
|
|
|
<para>
|
|
|
|
splice is a method for moving blocks of data around inside the
|
|
|
|
kernel, without continually transferring them between the kernel
|
|
|
|
and user space.
|
|
|
|
</para>
|
|
|
|
!Ffs/splice.c
|
|
|
|
</chapter>
|
|
|
|
|
|
|
|
<chapter id="pipes">
|
|
|
|
<title>pipes API</title>
|
|
|
|
<para>
|
|
|
|
Pipe interfaces are all for in-kernel (builtin image) use.
|
|
|
|
They are not exported for use by modules.
|
|
|
|
</para>
|
|
|
|
!Iinclude/linux/pipe_fs_i.h
|
|
|
|
!Ffs/pipe.c
|
|
|
|
</chapter>
|
|
|
|
|
2006-10-30 06:46:40 +00:00
|
|
|
</book>
|