Linux Filesystems API 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. 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. 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 For more details see the file COPYING in the source distribution of Linux. The Linux VFS The Filesystem types !Iinclude/linux/fs.h The Directory Cache !Efs/dcache.c !Iinclude/linux/dcache.h Inode Handling !Efs/inode.c !Efs/bad_inode.c Registration and Superblocks !Efs/super.c File Locks !Efs/locks.c !Ifs/locks.c Other Functions !Efs/mpage.c !Efs/namei.c !Efs/buffer.c !Efs/bio.c !Efs/seq_file.c !Efs/filesystems.c !Efs/fs-writeback.c !Efs/block_dev.c The proc filesystem sysctl interface !Ekernel/sysctl.c proc filesystem interface !Ifs/proc/base.c The Filesystem for Exporting Kernel Objects !Efs/sysfs/file.c !Efs/sysfs/symlink.c !Efs/sysfs/bin.c The debugfs filesystem debugfs interface !Efs/debugfs/inode.c !Efs/debugfs/file.c The Linux Journalling API Roger Gammans
rgammans@computer-surgery.co.uk
Stephen Tweedie
sct@redhat.com
2002 Roger Gammans
The Linux Journalling API Overview Details 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 media on which the journal resides. The journal_init_inode() call is for journals stored in filesystem inodes, or the journal_init_dev() call can be use for journal stored on a raw device (in a continuous range of blocks). A journal_t is a typedef for a struct pointer, so when you are finally finished make sure you call journal_destroy() on it to free up any used kernel memory. Once you have got your journal_t object you need to 'mount' or load the journal file, unless of course you haven't initialised it yet - in which case you need to call journal_create(). Most of the time however your journal file will already have been created, but before you load it you must call journal_wipe() to empty the journal file. Hang on, you say , what if the filesystem wasn't cleanly umount()'d . Well, it is the job of the client file system to detect this and skip the call to journal_wipe(). In either case the next call should be to journal_load() which prepares the journal file for use. Note that journal_wipe(..,0) calls journal_skip_recovery() for you if it detects any outstanding transactions in the journal and similarly journal_load() will call journal_recover() if necessary. I would advise reading fs/ext3/super.c for examples on this stage. [RGG: Why is the journal_wipe() call necessary - doesn't this needlessly complicate the API. Or isn't a good idea for the journal layer to hide dirty mounts from the client fs] Now you can go ahead and start modifying the underlying filesystem. Almost. 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 you are actually making are. To do this use journal_start() which returns a transaction handle. journal_start() and its counterpart 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 journal_stop() the same number of times as journal_start() before the transaction is completed (or more accurately leaves the update phase). Ext3/VFS makes use of this feature to simplify quota support. Inside each transaction you need to wrap the modifications to the individual buffers (blocks). Before you start to modify a buffer you need to call journal_get_{create,write,undo}_access() as appropriate, 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 you are have done so you need to call journal_dirty_{meta,}data(). Or if you've asked for access to a buffer you now know is now longer required to be pushed back on the device you can call journal_forget() in much the same way as you might have used bforget() in the past. A journal_flush() may be called at any time to commit and checkpoint all your transactions. Then at umount time , in your put_super() (2.4) or write_super() (2.5) you can then call journal_destroy() to clean up your in-core journal object. 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 commits until the outermost journal_stop(). This means 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 yours (say ext3) may be modified in a later syscall. The second case to bear in mind is that journal_start() can 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, so essentially we are waiting for journal_stop(). So to avoid deadlocks you must treat journal_start/stop() as if they were semaphores and include them in your semaphore ordering rules to prevent deadlocks. Note that journal_extend() has similar blocking behaviour to journal_start() so you can deadlock here just as easily as on journal_start(). 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. I advise having a look at at least ext3_jbd.h to see the basis on which ext3 uses to make these decisions. Another wriggle to watch out for is your on-disk block allocation strategy. why? Because, if you undo a delete, you need to ensure you haven't reused any of the freed blocks in a later transaction. One simple way of doing this is make sure any blocks you allocate only have checkpointed transactions listed against them. Ext3 does this in ext3_test_allocatable(). Lock is also providing through journal_{un,}lock_updates(), ext3 uses this when it wants a window with a clean and stable fs for a moment. eg. journal_lock_updates() //stop new stuff happening.. journal_flush() // checkpoint everything. ..do stuff on stable fs journal_unlock_updates() // carry on with filesystem use. The opportunities for abuse and DOS attacks with this should be obvious, if you allow unprivileged userspace to trigger codepaths containing these calls. A new feature of jbd since 2.5.25 is commit callbacks with the new journal_callback_set() function you can now ask the journalling layer to call you back when the transaction is finally committed to disk, so that you can do some of your own management. The key to this is the journal_callback struct, this maintains the internal callback information but you can extend it like this:- struct myfs_callback_s { //Data structure element required by jbd.. struct journal_callback for_jbd; // Stuff for myfs allocated together. myfs_inode* i_commited; } this would be useful if you needed to know when data was committed to a particular inode. Summary 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. Here is a some pseudo code to give you an idea of how it works, as an example. journal_t* my_jnrl = journal_create(); journal_init_{dev,inode}(jnrl,...) if (clean) journal_wipe(); journal_load(); foreach(transaction) { /*transactions must be completed before a syscall returns to userspace*/ handle_t * xct=journal_start(my_jnrl); foreach(bh) { journal_get_{create,write,undo}_access(xact,bh); if ( myfs_modify(bh) ) { /* returns true if makes changes */ journal_dirty_{meta,}data(xact,bh); } else { journal_forget(bh); } } journal_stop(xct); } journal_destroy(my_jrnl); Data Types The journalling layer uses typedefs to 'hide' the concrete definitions of the structures used. As a client of the JBD layer you can just rely on the using the pointer as a magic cookie of some sort. Obviously the hiding is not enforced as this is 'C'. Structures !Iinclude/linux/jbd.h Functions The functions here are split into two groups those that affect a journal as a whole, and those which are used to manage transactions Journal Level !Efs/jbd/journal.c !Ifs/jbd/recovery.c Transasction Level !Efs/jbd/transaction.c See also Journaling the Linux ext2fs Filesystem, LinuxExpo 98, Stephen Tweedie Ext3 Journalling FileSystem, OLS 2000, Dr. Stephen Tweedie
splice API splice is a method for moving blocks of data around inside the kernel, without continually transferring them between the kernel and user space. !Ffs/splice.c pipes API Pipe interfaces are all for in-kernel (builtin image) use. They are not exported for use by modules. !Iinclude/linux/pipe_fs_i.h !Ffs/pipe.c