blob: 21a7dc467bba197758caceb4d18285f784da6bb2 [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -07001
Pekka J Enberg5ea626a2005-09-09 13:10:19 -07002 Overview of the Linux Virtual File System
Linus Torvalds1da177e2005-04-16 15:20:36 -07003
Pekka J Enberg5ea626a2005-09-09 13:10:19 -07004 Original author: Richard Gooch <rgooch@atnf.csiro.au>
Linus Torvalds1da177e2005-04-16 15:20:36 -07005
Borislav Petkov0746aec2007-07-15 23:41:19 -07006 Last updated on June 24, 2007.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -07007
8 Copyright (C) 1999 Richard Gooch
9 Copyright (C) 2005 Pekka Enberg
10
11 This file is released under the GPLv2.
Linus Torvalds1da177e2005-04-16 15:20:36 -070012
13
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080014Introduction
15============
Linus Torvalds1da177e2005-04-16 15:20:36 -070016
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080017The Virtual File System (also known as the Virtual Filesystem Switch)
18is the software layer in the kernel that provides the filesystem
19interface to userspace programs. It also provides an abstraction
20within the kernel which allows different filesystem implementations to
21coexist.
22
23VFS system calls open(2), stat(2), read(2), write(2), chmod(2) and so
24on are called from a process context. Filesystem locking is described
25in the document Documentation/filesystems/Locking.
Linus Torvalds1da177e2005-04-16 15:20:36 -070026
27
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080028Directory Entry Cache (dcache)
29------------------------------
Linus Torvalds1da177e2005-04-16 15:20:36 -070030
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080031The VFS implements the open(2), stat(2), chmod(2), and similar system
32calls. The pathname argument that is passed to them is used by the VFS
33to search through the directory entry cache (also known as the dentry
34cache or dcache). This provides a very fast look-up mechanism to
35translate a pathname (filename) into a specific dentry. Dentries live
36in RAM and are never saved to disc: they exist only for performance.
37
38The dentry cache is meant to be a view into your entire filespace. As
39most computers cannot fit all dentries in the RAM at the same time,
40some bits of the cache are missing. In order to resolve your pathname
41into a dentry, the VFS may have to resort to creating dentries along
42the way, and then loading the inode. This is done by looking up the
43inode.
Linus Torvalds1da177e2005-04-16 15:20:36 -070044
Pekka J Enberg5ea626a2005-09-09 13:10:19 -070045
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080046The Inode Object
47----------------
Linus Torvalds1da177e2005-04-16 15:20:36 -070048
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080049An individual dentry usually has a pointer to an inode. Inodes are
50filesystem objects such as regular files, directories, FIFOs and other
51beasts. They live either on the disc (for block device filesystems)
52or in the memory (for pseudo filesystems). Inodes that live on the
53disc are copied into the memory when required and changes to the inode
54are written back to disc. A single inode can be pointed to by multiple
55dentries (hard links, for example, do this).
Linus Torvalds1da177e2005-04-16 15:20:36 -070056
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080057To look up an inode requires that the VFS calls the lookup() method of
58the parent directory inode. This method is installed by the specific
59filesystem implementation that the inode lives in. Once the VFS has
60the required dentry (and hence the inode), we can do all those boring
61things like open(2) the file, or stat(2) it to peek at the inode
62data. The stat(2) operation is fairly simple: once the VFS has the
63dentry, it peeks at the inode data and passes some of it back to
64userspace.
Linus Torvalds1da177e2005-04-16 15:20:36 -070065
Linus Torvalds1da177e2005-04-16 15:20:36 -070066
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080067The File Object
68---------------
Linus Torvalds1da177e2005-04-16 15:20:36 -070069
70Opening a file requires another operation: allocation of a file
71structure (this is the kernel-side implementation of file
Pekka J Enberg5ea626a2005-09-09 13:10:19 -070072descriptors). The freshly allocated file structure is initialized with
Linus Torvalds1da177e2005-04-16 15:20:36 -070073a pointer to the dentry and a set of file operation member functions.
74These are taken from the inode data. The open() file method is then
Francis Galieguea33f3222010-04-23 00:08:02 +020075called so the specific filesystem implementation can do its work. You
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080076can see that this is another switch performed by the VFS. The file
77structure is placed into the file descriptor table for the process.
Linus Torvalds1da177e2005-04-16 15:20:36 -070078
79Reading, writing and closing files (and other assorted VFS operations)
80is done by using the userspace file descriptor to grab the appropriate
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080081file structure, and then calling the required file structure method to
82do whatever is required. For as long as the file is open, it keeps the
83dentry in use, which in turn means that the VFS inode is still in use.
Linus Torvalds1da177e2005-04-16 15:20:36 -070084
Pekka J Enberg5ea626a2005-09-09 13:10:19 -070085
86Registering and Mounting a Filesystem
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080087=====================================
Linus Torvalds1da177e2005-04-16 15:20:36 -070088
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080089To register and unregister a filesystem, use the following API
90functions:
Linus Torvalds1da177e2005-04-16 15:20:36 -070091
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080092 #include <linux/fs.h>
Linus Torvalds1da177e2005-04-16 15:20:36 -070093
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080094 extern int register_filesystem(struct file_system_type *);
95 extern int unregister_filesystem(struct file_system_type *);
Linus Torvalds1da177e2005-04-16 15:20:36 -070096
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080097The passed struct file_system_type describes your filesystem. When a
Al Viro1a102ff2011-03-16 09:07:58 -040098request is made to mount a filesystem onto a directory in your namespace,
99the VFS will call the appropriate mount() method for the specific
Lucas De Marchi25985ed2011-03-30 22:57:33 -0300100filesystem. New vfsmount referring to the tree returned by ->mount()
Al Viro1a102ff2011-03-16 09:07:58 -0400101will be attached to the mountpoint, so that when pathname resolution
102reaches the mountpoint it will jump into the root of that vfsmount.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700103
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800104You can see all filesystems that are registered to the kernel in the
105file /proc/filesystems.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700106
107
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700108struct file_system_type
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800109-----------------------
Linus Torvalds1da177e2005-04-16 15:20:36 -0700110
Al Viro1a102ff2011-03-16 09:07:58 -0400111This describes the filesystem. As of kernel 2.6.39, the following
Linus Torvalds1da177e2005-04-16 15:20:36 -0700112members are defined:
113
114struct file_system_type {
115 const char *name;
116 int fs_flags;
Al Viro1a102ff2011-03-16 09:07:58 -0400117 struct dentry (*mount) (struct file_system_type *, int,
118 const char *, void *);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700119 void (*kill_sb) (struct super_block *);
120 struct module *owner;
121 struct file_system_type * next;
122 struct list_head fs_supers;
Borislav Petkov0746aec2007-07-15 23:41:19 -0700123 struct lock_class_key s_lock_key;
124 struct lock_class_key s_umount_key;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700125};
126
127 name: the name of the filesystem type, such as "ext2", "iso9660",
128 "msdos" and so on
129
130 fs_flags: various flags (i.e. FS_REQUIRES_DEV, FS_NO_DCACHE, etc.)
131
Al Viro1a102ff2011-03-16 09:07:58 -0400132 mount: the method to call when a new instance of this
Linus Torvalds1da177e2005-04-16 15:20:36 -0700133 filesystem should be mounted
134
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700135 kill_sb: the method to call when an instance of this filesystem
Al Viro1a102ff2011-03-16 09:07:58 -0400136 should be shut down
Linus Torvalds1da177e2005-04-16 15:20:36 -0700137
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700138 owner: for internal VFS use: you should initialize this to THIS_MODULE in
139 most cases.
140
141 next: for internal VFS use: you should initialize this to NULL
142
Borislav Petkov0746aec2007-07-15 23:41:19 -0700143 s_lock_key, s_umount_key: lockdep-specific
144
Al Viro1a102ff2011-03-16 09:07:58 -0400145The mount() method has the following arguments:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700146
Matt LaPlanted9195882008-07-25 19:45:33 -0700147 struct file_system_type *fs_type: describes the filesystem, partly initialized
Borislav Petkov0746aec2007-07-15 23:41:19 -0700148 by the specific filesystem code
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700149
150 int flags: mount flags
151
152 const char *dev_name: the device name we are mounting.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700153
154 void *data: arbitrary mount options, usually comes as an ASCII
Miklos Szeredif84e3f52008-02-08 04:21:34 -0800155 string (see "Mount Options" section)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700156
Al Viro1a102ff2011-03-16 09:07:58 -0400157The mount() method must return the root dentry of the tree requested by
158caller. An active reference to its superblock must be grabbed and the
159superblock must be locked. On failure it should return ERR_PTR(error).
Linus Torvalds1da177e2005-04-16 15:20:36 -0700160
Al Viro1a102ff2011-03-16 09:07:58 -0400161The arguments match those of mount(2) and their interpretation
162depends on filesystem type. E.g. for block filesystems, dev_name is
163interpreted as block device name, that device is opened and if it
164contains a suitable filesystem image the method creates and initializes
165struct super_block accordingly, returning its root dentry to caller.
166
167->mount() may choose to return a subtree of existing filesystem - it
168doesn't have to create a new one. The main result from the caller's
169point of view is a reference to dentry at the root of (sub)tree to
170be attached; creation of new superblock is a common side effect.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700171
172The most interesting member of the superblock structure that the
Al Viro1a102ff2011-03-16 09:07:58 -0400173mount() method fills in is the "s_op" field. This is a pointer to
Linus Torvalds1da177e2005-04-16 15:20:36 -0700174a "struct super_operations" which describes the next level of the
175filesystem implementation.
176
Al Viro1a102ff2011-03-16 09:07:58 -0400177Usually, a filesystem uses one of the generic mount() implementations
178and provides a fill_super() callback instead. The generic variants are:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700179
Al Viro1a102ff2011-03-16 09:07:58 -0400180 mount_bdev: mount a filesystem residing on a block device
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700181
Al Viro1a102ff2011-03-16 09:07:58 -0400182 mount_nodev: mount a filesystem that is not backed by a device
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700183
Al Viro1a102ff2011-03-16 09:07:58 -0400184 mount_single: mount a filesystem which shares the instance between
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700185 all mounts
186
Al Viro1a102ff2011-03-16 09:07:58 -0400187A fill_super() callback implementation has the following arguments:
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700188
Al Viro1a102ff2011-03-16 09:07:58 -0400189 struct super_block *sb: the superblock structure. The callback
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700190 must initialize this properly.
191
192 void *data: arbitrary mount options, usually comes as an ASCII
Miklos Szeredif84e3f52008-02-08 04:21:34 -0800193 string (see "Mount Options" section)
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700194
195 int silent: whether or not to be silent on error
196
197
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800198The Superblock Object
199=====================
200
201A superblock object represents a mounted filesystem.
202
203
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700204struct super_operations
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800205-----------------------
Linus Torvalds1da177e2005-04-16 15:20:36 -0700206
207This describes how the VFS can manipulate the superblock of your
Borislav Petkov422b14c2007-07-15 23:41:43 -0700208filesystem. As of kernel 2.6.22, the following members are defined:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700209
210struct super_operations {
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700211 struct inode *(*alloc_inode)(struct super_block *sb);
212 void (*destroy_inode)(struct inode *);
213
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700214 void (*dirty_inode) (struct inode *);
215 int (*write_inode) (struct inode *, int);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700216 void (*drop_inode) (struct inode *);
217 void (*delete_inode) (struct inode *);
218 void (*put_super) (struct super_block *);
219 void (*write_super) (struct super_block *);
220 int (*sync_fs)(struct super_block *sb, int wait);
Takashi Satoc4be0c12009-01-09 16:40:58 -0800221 int (*freeze_fs) (struct super_block *);
222 int (*unfreeze_fs) (struct super_block *);
David Howells726c3342006-06-23 02:02:58 -0700223 int (*statfs) (struct dentry *, struct kstatfs *);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700224 int (*remount_fs) (struct super_block *, int *, char *);
225 void (*clear_inode) (struct inode *);
226 void (*umount_begin) (struct super_block *);
227
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700228 int (*show_options)(struct seq_file *, struct vfsmount *);
229
230 ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
231 ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700232};
233
234All methods are called without any locks being held, unless otherwise
235noted. This means that most methods can block safely. All methods are
236only called from a process context (i.e. not from an interrupt handler
237or bottom half).
238
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700239 alloc_inode: this method is called by inode_alloc() to allocate memory
NeilBrown341546f2006-03-25 03:07:56 -0800240 for struct inode and initialize it. If this function is not
241 defined, a simple 'struct inode' is allocated. Normally
242 alloc_inode will be used to allocate a larger structure which
243 contains a 'struct inode' embedded within it.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700244
245 destroy_inode: this method is called by destroy_inode() to release
NeilBrown341546f2006-03-25 03:07:56 -0800246 resources allocated for struct inode. It is only required if
247 ->alloc_inode was defined and simply undoes anything done by
248 ->alloc_inode.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700249
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700250 dirty_inode: this method is called by the VFS to mark an inode dirty.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700251
252 write_inode: this method is called when the VFS needs to write an
253 inode to disc. The second parameter indicates whether the write
254 should be synchronous or not, not all filesystems check this flag.
255
Linus Torvalds1da177e2005-04-16 15:20:36 -0700256 drop_inode: called when the last access to the inode is dropped,
Dave Chinnerf283c862011-03-22 22:23:39 +1100257 with the inode->i_lock spinlock held.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700258
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700259 This method should be either NULL (normal UNIX filesystem
Linus Torvalds1da177e2005-04-16 15:20:36 -0700260 semantics) or "generic_delete_inode" (for filesystems that do not
261 want to cache inodes - causing "delete_inode" to always be
262 called regardless of the value of i_nlink)
263
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700264 The "generic_delete_inode()" behavior is equivalent to the
Linus Torvalds1da177e2005-04-16 15:20:36 -0700265 old practice of using "force_delete" in the put_inode() case,
266 but does not have the races that the "force_delete()" approach
267 had.
268
269 delete_inode: called when the VFS wants to delete an inode
270
Linus Torvalds1da177e2005-04-16 15:20:36 -0700271 put_super: called when the VFS wishes to free the superblock
272 (i.e. unmount). This is called with the superblock lock held
273
274 write_super: called when the VFS superblock needs to be written to
275 disc. This method is optional
276
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700277 sync_fs: called when VFS is writing out all dirty data associated with
278 a superblock. The second parameter indicates whether the method
279 should wait until the write out has been completed. Optional.
280
Takashi Satoc4be0c12009-01-09 16:40:58 -0800281 freeze_fs: called when VFS is locking a filesystem and
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800282 forcing it into a consistent state. This method is currently
283 used by the Logical Volume Manager (LVM).
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700284
Takashi Satoc4be0c12009-01-09 16:40:58 -0800285 unfreeze_fs: called when VFS is unlocking a filesystem and making it writable
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700286 again.
287
Adrian McMenamin66672fe2009-04-20 18:38:28 -0700288 statfs: called when the VFS needs to get filesystem statistics.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700289
290 remount_fs: called when the filesystem is remounted. This is called
291 with the kernel lock held
292
293 clear_inode: called then the VFS clears the inode. Optional
294
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700295 umount_begin: called when the VFS is unmounting a filesystem.
296
Miklos Szeredif84e3f52008-02-08 04:21:34 -0800297 show_options: called by the VFS to show mount options for
298 /proc/<pid>/mounts. (see "Mount Options" section)
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700299
300 quota_read: called by the VFS to read from filesystem quota file.
301
302 quota_write: called by the VFS to write to filesystem quota file.
303
David Howells12debc42008-02-07 00:15:52 -0800304Whoever sets up the inode is responsible for filling in the "i_op" field. This
305is a pointer to a "struct inode_operations" which describes the methods that
306can be performed on individual inodes.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700307
308
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800309The Inode Object
310================
311
312An inode object represents an object within the filesystem.
313
314
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700315struct inode_operations
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800316-----------------------
Linus Torvalds1da177e2005-04-16 15:20:36 -0700317
318This describes how the VFS can manipulate an inode in your
Borislav Petkov422b14c2007-07-15 23:41:43 -0700319filesystem. As of kernel 2.6.22, the following members are defined:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700320
321struct inode_operations {
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700322 int (*create) (struct inode *,struct dentry *,int, struct nameidata *);
323 struct dentry * (*lookup) (struct inode *,struct dentry *, struct nameidata *);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700324 int (*link) (struct dentry *,struct inode *,struct dentry *);
325 int (*unlink) (struct inode *,struct dentry *);
326 int (*symlink) (struct inode *,struct dentry *,const char *);
327 int (*mkdir) (struct inode *,struct dentry *,int);
328 int (*rmdir) (struct inode *,struct dentry *);
329 int (*mknod) (struct inode *,struct dentry *,int,dev_t);
330 int (*rename) (struct inode *, struct dentry *,
331 struct inode *, struct dentry *);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700332 int (*readlink) (struct dentry *, char __user *,int);
333 void * (*follow_link) (struct dentry *, struct nameidata *);
334 void (*put_link) (struct dentry *, struct nameidata *, void *);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700335 void (*truncate) (struct inode *);
Nick Pigginb74c79e2011-01-07 17:49:58 +1100336 int (*permission) (struct inode *, int, unsigned int);
337 int (*check_acl)(struct inode *, int, unsigned int);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700338 int (*setattr) (struct dentry *, struct iattr *);
339 int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
340 int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
341 ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
342 ssize_t (*listxattr) (struct dentry *, char *, size_t);
343 int (*removexattr) (struct dentry *, const char *);
Borislav Petkov422b14c2007-07-15 23:41:43 -0700344 void (*truncate_range)(struct inode *, loff_t, loff_t);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700345};
346
347Again, all methods are called without any locks being held, unless
348otherwise noted.
349
Linus Torvalds1da177e2005-04-16 15:20:36 -0700350 create: called by the open(2) and creat(2) system calls. Only
351 required if you want to support regular files. The dentry you
352 get should not have an inode (i.e. it should be a negative
353 dentry). Here you will probably call d_instantiate() with the
354 dentry and the newly created inode
355
356 lookup: called when the VFS needs to look up an inode in a parent
357 directory. The name to look for is found in the dentry. This
358 method must call d_add() to insert the found inode into the
359 dentry. The "i_count" field in the inode structure should be
360 incremented. If the named inode does not exist a NULL inode
361 should be inserted into the dentry (this is called a negative
362 dentry). Returning an error code from this routine must only
363 be done on a real error, otherwise creating inodes with system
364 calls like create(2), mknod(2), mkdir(2) and so on will fail.
365 If you wish to overload the dentry methods then you should
366 initialise the "d_dop" field in the dentry; this is a pointer
367 to a struct "dentry_operations".
368 This method is called with the directory inode semaphore held
369
370 link: called by the link(2) system call. Only required if you want
371 to support hard links. You will probably need to call
372 d_instantiate() just as you would in the create() method
373
374 unlink: called by the unlink(2) system call. Only required if you
375 want to support deleting inodes
376
377 symlink: called by the symlink(2) system call. Only required if you
378 want to support symlinks. You will probably need to call
379 d_instantiate() just as you would in the create() method
380
381 mkdir: called by the mkdir(2) system call. Only required if you want
382 to support creating subdirectories. You will probably need to
383 call d_instantiate() just as you would in the create() method
384
385 rmdir: called by the rmdir(2) system call. Only required if you want
386 to support deleting subdirectories
387
388 mknod: called by the mknod(2) system call to create a device (char,
389 block) inode or a named pipe (FIFO) or socket. Only required
390 if you want to support creating these types of inodes. You
391 will probably need to call d_instantiate() just as you would
392 in the create() method
393
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800394 rename: called by the rename(2) system call to rename the object to
395 have the parent and name given by the second inode and dentry.
396
Linus Torvalds1da177e2005-04-16 15:20:36 -0700397 readlink: called by the readlink(2) system call. Only required if
398 you want to support reading symbolic links
399
400 follow_link: called by the VFS to follow a symbolic link to the
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700401 inode it points to. Only required if you want to support
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800402 symbolic links. This method returns a void pointer cookie
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700403 that is passed to put_link().
404
405 put_link: called by the VFS to release resources allocated by
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800406 follow_link(). The cookie returned by follow_link() is passed
Paolo Ornati670e9f32006-10-03 22:57:56 +0200407 to this method as the last parameter. It is used by
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800408 filesystems such as NFS where page cache is not stable
409 (i.e. page that was installed when the symbolic link walk
410 started might not be in the page cache at the end of the
411 walk).
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700412
npiggin@suse.de7bb46a62010-05-27 01:05:33 +1000413 truncate: Deprecated. This will not be called if ->setsize is defined.
414 Called by the VFS to change the size of a file. The
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800415 i_size field of the inode is set to the desired size by the
416 VFS before this method is called. This method is called by
417 the truncate(2) system call and related functionality.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700418
npiggin@suse.de7bb46a62010-05-27 01:05:33 +1000419 Note: ->truncate and vmtruncate are deprecated. Do not add new
420 instances/calls of these. Filesystems should be converted to do their
421 truncate sequence via ->setattr().
422
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700423 permission: called by the VFS to check for access rights on a POSIX-like
424 filesystem.
425
Nick Piggina82416d2011-01-14 02:26:53 +0000426 May be called in rcu-walk mode (flags & IPERM_FLAG_RCU). If in rcu-walk
427 mode, the filesystem must check the permission without blocking or
Nick Pigginb74c79e2011-01-07 17:49:58 +1100428 storing to the inode.
429
430 If a situation is encountered that rcu-walk cannot handle, return
431 -ECHILD and it will be called again in ref-walk mode.
432
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800433 setattr: called by the VFS to set attributes for a file. This method
434 is called by chmod(2) and related system calls.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700435
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800436 getattr: called by the VFS to get attributes of a file. This method
437 is called by stat(2) and related system calls.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700438
439 setxattr: called by the VFS to set an extended attribute for a file.
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800440 Extended attribute is a name:value pair associated with an
441 inode. This method is called by setxattr(2) system call.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700442
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800443 getxattr: called by the VFS to retrieve the value of an extended
444 attribute name. This method is called by getxattr(2) function
445 call.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700446
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800447 listxattr: called by the VFS to list all extended attributes for a
448 given file. This method is called by listxattr(2) system call.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700449
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800450 removexattr: called by the VFS to remove an extended attribute from
451 a file. This method is called by removexattr(2) system call.
452
Borislav Petkov422b14c2007-07-15 23:41:43 -0700453 truncate_range: a method provided by the underlying filesystem to truncate a
454 range of blocks , i.e. punch a hole somewhere in a file.
455
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800456
457The Address Space Object
458========================
459
NeilBrown341546f2006-03-25 03:07:56 -0800460The address space object is used to group and manage pages in the page
461cache. It can be used to keep track of the pages in a file (or
462anything else) and also track the mapping of sections of the file into
463process address spaces.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700464
NeilBrown341546f2006-03-25 03:07:56 -0800465There are a number of distinct yet related services that an
466address-space can provide. These include communicating memory
467pressure, page lookup by address, and keeping track of pages tagged as
468Dirty or Writeback.
469
NeilBrowna9e102b2006-03-25 03:08:29 -0800470The first can be used independently to the others. The VM can try to
NeilBrown341546f2006-03-25 03:07:56 -0800471either write dirty pages in order to clean them, or release clean
472pages in order to reuse them. To do this it can call the ->writepage
473method on dirty pages, and ->releasepage on clean pages with
474PagePrivate set. Clean pages without PagePrivate and with no external
475references will be released without notice being given to the
476address_space.
477
NeilBrowna9e102b2006-03-25 03:08:29 -0800478To achieve this functionality, pages need to be placed on an LRU with
NeilBrown341546f2006-03-25 03:07:56 -0800479lru_cache_add and mark_page_active needs to be called whenever the
480page is used.
481
482Pages are normally kept in a radix tree index by ->index. This tree
483maintains information about the PG_Dirty and PG_Writeback status of
484each page, so that pages with either of these flags can be found
485quickly.
486
487The Dirty tag is primarily used by mpage_writepages - the default
488->writepages method. It uses the tag to find dirty pages to call
489->writepage on. If mpage_writepages is not used (i.e. the address
NeilBrowna9e102b2006-03-25 03:08:29 -0800490provides its own ->writepages) , the PAGECACHE_TAG_DIRTY tag is
NeilBrown341546f2006-03-25 03:07:56 -0800491almost unused. write_inode_now and sync_inode do use it (through
492__sync_single_inode) to check if ->writepages has been successful in
493writing out the whole address_space.
494
495The Writeback tag is used by filemap*wait* and sync_page* functions,
Christoph Hellwig94004ed2009-09-30 22:16:33 +0200496via filemap_fdatawait_range, to wait for all writeback to
NeilBrown341546f2006-03-25 03:07:56 -0800497complete. While waiting ->sync_page (if defined) will be called on
NeilBrowna9e102b2006-03-25 03:08:29 -0800498each page that is found to require writeback.
NeilBrown341546f2006-03-25 03:07:56 -0800499
500An address_space handler may attach extra information to a page,
501typically using the 'private' field in the 'struct page'. If such
502information is attached, the PG_Private flag should be set. This will
NeilBrowna9e102b2006-03-25 03:08:29 -0800503cause various VM routines to make extra calls into the address_space
NeilBrown341546f2006-03-25 03:07:56 -0800504handler to deal with that data.
505
506An address space acts as an intermediate between storage and
507application. Data is read into the address space a whole page at a
508time, and provided to the application either by copying of the page,
509or by memory-mapping the page.
510Data is written into the address space by the application, and then
511written-back to storage typically in whole pages, however the
NeilBrowna9e102b2006-03-25 03:08:29 -0800512address_space has finer control of write sizes.
NeilBrown341546f2006-03-25 03:07:56 -0800513
514The read process essentially only requires 'readpage'. The write
Nick Piggin4e02ed42008-10-29 14:00:55 -0700515process is more complicated and uses write_begin/write_end or
NeilBrown341546f2006-03-25 03:07:56 -0800516set_page_dirty to write data into the address_space, and writepage,
517sync_page, and writepages to writeback data to storage.
518
519Adding and removing pages to/from an address_space is protected by the
520inode's i_mutex.
521
522When data is written to a page, the PG_Dirty flag should be set. It
523typically remains set until writepage asks for it to be written. This
524should clear PG_Dirty and set PG_Writeback. It can be actually
525written at any point after PG_Dirty is clear. Once it is known to be
526safe, PG_Writeback is cleared.
527
528Writeback makes use of a writeback_control structure...
Linus Torvalds1da177e2005-04-16 15:20:36 -0700529
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700530struct address_space_operations
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800531-------------------------------
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700532
533This describes how the VFS can manipulate mapping of a file to page cache in
Borislav Petkov422b14c2007-07-15 23:41:43 -0700534your filesystem. As of kernel 2.6.22, the following members are defined:
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700535
536struct address_space_operations {
537 int (*writepage)(struct page *page, struct writeback_control *wbc);
538 int (*readpage)(struct file *, struct page *);
539 int (*sync_page)(struct page *);
540 int (*writepages)(struct address_space *, struct writeback_control *);
541 int (*set_page_dirty)(struct page *page);
542 int (*readpages)(struct file *filp, struct address_space *mapping,
543 struct list_head *pages, unsigned nr_pages);
Nick Pigginafddba42007-10-16 01:25:01 -0700544 int (*write_begin)(struct file *, struct address_space *mapping,
545 loff_t pos, unsigned len, unsigned flags,
546 struct page **pagep, void **fsdata);
547 int (*write_end)(struct file *, struct address_space *mapping,
548 loff_t pos, unsigned len, unsigned copied,
549 struct page *page, void *fsdata);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700550 sector_t (*bmap)(struct address_space *, sector_t);
551 int (*invalidatepage) (struct page *, unsigned long);
552 int (*releasepage) (struct page *, int);
Linus Torvalds6072d132010-12-01 13:35:19 -0500553 void (*freepage)(struct page *);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700554 ssize_t (*direct_IO)(int, struct kiocb *, const struct iovec *iov,
555 loff_t offset, unsigned long nr_segs);
556 struct page* (*get_xip_page)(struct address_space *, sector_t,
557 int);
NeilBrown341546f2006-03-25 03:07:56 -0800558 /* migrate the contents of a page to the specified target */
559 int (*migratepage) (struct page *, struct page *);
Borislav Petkov422b14c2007-07-15 23:41:43 -0700560 int (*launder_page) (struct page *);
Andi Kleen25718732009-09-16 11:50:13 +0200561 int (*error_remove_page) (struct mapping *mapping, struct page *page);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700562};
563
NeilBrown341546f2006-03-25 03:07:56 -0800564 writepage: called by the VM to write a dirty page to backing store.
NeilBrowna9e102b2006-03-25 03:08:29 -0800565 This may happen for data integrity reasons (i.e. 'sync'), or
NeilBrown341546f2006-03-25 03:07:56 -0800566 to free up memory (flush). The difference can be seen in
567 wbc->sync_mode.
568 The PG_Dirty flag has been cleared and PageLocked is true.
569 writepage should start writeout, should set PG_Writeback,
570 and should make sure the page is unlocked, either synchronously
571 or asynchronously when the write operation completes.
572
573 If wbc->sync_mode is WB_SYNC_NONE, ->writepage doesn't have to
NeilBrowna9e102b2006-03-25 03:08:29 -0800574 try too hard if there are problems, and may choose to write out
575 other pages from the mapping if that is easier (e.g. due to
576 internal dependencies). If it chooses not to start writeout, it
577 should return AOP_WRITEPAGE_ACTIVATE so that the VM will not keep
NeilBrown341546f2006-03-25 03:07:56 -0800578 calling ->writepage on that page.
579
580 See the file "Locking" for more details.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700581
582 readpage: called by the VM to read a page from backing store.
NeilBrown341546f2006-03-25 03:07:56 -0800583 The page will be Locked when readpage is called, and should be
584 unlocked and marked uptodate once the read completes.
585 If ->readpage discovers that it needs to unlock the page for
586 some reason, it can do so, and then return AOP_TRUNCATED_PAGE.
NeilBrowna9e102b2006-03-25 03:08:29 -0800587 In this case, the page will be relocated, relocked and if
NeilBrown341546f2006-03-25 03:07:56 -0800588 that all succeeds, ->readpage will be called again.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700589
590 sync_page: called by the VM to notify the backing store to perform all
591 queued I/O operations for a page. I/O operations for other pages
592 associated with this address_space object may also be performed.
593
NeilBrown341546f2006-03-25 03:07:56 -0800594 This function is optional and is called only for pages with
595 PG_Writeback set while waiting for the writeback to complete.
596
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700597 writepages: called by the VM to write out pages associated with the
NeilBrowna9e102b2006-03-25 03:08:29 -0800598 address_space object. If wbc->sync_mode is WBC_SYNC_ALL, then
599 the writeback_control will specify a range of pages that must be
600 written out. If it is WBC_SYNC_NONE, then a nr_to_write is given
NeilBrown341546f2006-03-25 03:07:56 -0800601 and that many pages should be written if possible.
602 If no ->writepages is given, then mpage_writepages is used
NeilBrowna9e102b2006-03-25 03:08:29 -0800603 instead. This will choose pages from the address space that are
NeilBrown341546f2006-03-25 03:07:56 -0800604 tagged as DIRTY and will pass them to ->writepage.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700605
606 set_page_dirty: called by the VM to set a page dirty.
NeilBrown341546f2006-03-25 03:07:56 -0800607 This is particularly needed if an address space attaches
608 private data to a page, and that data needs to be updated when
609 a page is dirtied. This is called, for example, when a memory
610 mapped page gets modified.
611 If defined, it should set the PageDirty flag, and the
612 PAGECACHE_TAG_DIRTY tag in the radix tree.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700613
614 readpages: called by the VM to read pages associated with the address_space
NeilBrown341546f2006-03-25 03:07:56 -0800615 object. This is essentially just a vector version of
616 readpage. Instead of just one page, several pages are
617 requested.
NeilBrowna9e102b2006-03-25 03:08:29 -0800618 readpages is only used for read-ahead, so read errors are
NeilBrown341546f2006-03-25 03:07:56 -0800619 ignored. If anything goes wrong, feel free to give up.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700620
Nick Piggin4e02ed42008-10-29 14:00:55 -0700621 write_begin:
Nick Pigginafddba42007-10-16 01:25:01 -0700622 Called by the generic buffered write code to ask the filesystem to
623 prepare to write len bytes at the given offset in the file. The
624 address_space should check that the write will be able to complete,
625 by allocating space if necessary and doing any other internal
626 housekeeping. If the write will update parts of any basic-blocks on
627 storage, then those blocks should be pre-read (if they haven't been
628 read already) so that the updated blocks can be written out properly.
629
630 The filesystem must return the locked pagecache page for the specified
631 offset, in *pagep, for the caller to write into.
632
Nick Piggin4e02ed42008-10-29 14:00:55 -0700633 It must be able to cope with short writes (where the length passed to
634 write_begin is greater than the number of bytes copied into the page).
635
Nick Pigginafddba42007-10-16 01:25:01 -0700636 flags is a field for AOP_FLAG_xxx flags, described in
637 include/linux/fs.h.
638
639 A void * may be returned in fsdata, which then gets passed into
640 write_end.
641
642 Returns 0 on success; < 0 on failure (which is the error code), in
643 which case write_end is not called.
644
645 write_end: After a successful write_begin, and data copy, write_end must
646 be called. len is the original len passed to write_begin, and copied
647 is the amount that was able to be copied (copied == len is always true
648 if write_begin was called with the AOP_FLAG_UNINTERRUPTIBLE flag).
649
650 The filesystem must take care of unlocking the page and releasing it
651 refcount, and updating i_size.
652
653 Returns < 0 on failure, otherwise the number of bytes (<= 'copied')
654 that were able to be copied into pagecache.
655
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700656 bmap: called by the VFS to map a logical block offset within object to
NeilBrowna9e102b2006-03-25 03:08:29 -0800657 physical block number. This method is used by the FIBMAP
NeilBrown341546f2006-03-25 03:07:56 -0800658 ioctl and for working with swap-files. To be able to swap to
NeilBrowna9e102b2006-03-25 03:08:29 -0800659 a file, the file must have a stable mapping to a block
NeilBrown341546f2006-03-25 03:07:56 -0800660 device. The swap system does not go through the filesystem
661 but instead uses bmap to find out where the blocks in the file
662 are and uses those addresses directly.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700663
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700664
NeilBrown341546f2006-03-25 03:07:56 -0800665 invalidatepage: If a page has PagePrivate set, then invalidatepage
666 will be called when part or all of the page is to be removed
NeilBrowna9e102b2006-03-25 03:08:29 -0800667 from the address space. This generally corresponds to either a
NeilBrown341546f2006-03-25 03:07:56 -0800668 truncation or a complete invalidation of the address space
669 (in the latter case 'offset' will always be 0).
670 Any private data associated with the page should be updated
671 to reflect this truncation. If offset is 0, then
672 the private data should be released, because the page
673 must be able to be completely discarded. This may be done by
674 calling the ->releasepage function, but in this case the
675 release MUST succeed.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700676
NeilBrown341546f2006-03-25 03:07:56 -0800677 releasepage: releasepage is called on PagePrivate pages to indicate
678 that the page should be freed if possible. ->releasepage
679 should remove any private data from the page and clear the
Andrew Morton4fe65ca2010-12-02 14:31:19 -0800680 PagePrivate flag. If releasepage() fails for some reason, it must
681 indicate failure with a 0 return value.
682 releasepage() is used in two distinct though related cases. The
683 first is when the VM finds a clean page with no active users and
NeilBrown341546f2006-03-25 03:07:56 -0800684 wants to make it a free page. If ->releasepage succeeds, the
685 page will be removed from the address_space and become free.
686
Shaun Zinckbc5b1d52007-10-20 02:35:36 +0200687 The second case is when a request has been made to invalidate
NeilBrown341546f2006-03-25 03:07:56 -0800688 some or all pages in an address_space. This can happen
689 through the fadvice(POSIX_FADV_DONTNEED) system call or by the
690 filesystem explicitly requesting it as nfs and 9fs do (when
691 they believe the cache may be out of date with storage) by
692 calling invalidate_inode_pages2().
693 If the filesystem makes such a call, and needs to be certain
NeilBrowna9e102b2006-03-25 03:08:29 -0800694 that all pages are invalidated, then its releasepage will
NeilBrown341546f2006-03-25 03:07:56 -0800695 need to ensure this. Possibly it can clear the PageUptodate
696 bit if it cannot free private data yet.
697
Linus Torvalds6072d132010-12-01 13:35:19 -0500698 freepage: freepage is called once the page is no longer visible in
699 the page cache in order to allow the cleanup of any private
700 data. Since it may be called by the memory reclaimer, it
701 should not assume that the original address_space mapping still
702 exists, and it should not block.
703
NeilBrown341546f2006-03-25 03:07:56 -0800704 direct_IO: called by the generic read/write routines to perform
705 direct_IO - that is IO requests which bypass the page cache
NeilBrowna9e102b2006-03-25 03:08:29 -0800706 and transfer data directly between the storage and the
NeilBrown341546f2006-03-25 03:07:56 -0800707 application's address space.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700708
709 get_xip_page: called by the VM to translate a block number to a page.
710 The page is valid until the corresponding filesystem is unmounted.
711 Filesystems that want to use execute-in-place (XIP) need to implement
712 it. An example implementation can be found in fs/ext2/xip.c.
713
NeilBrown341546f2006-03-25 03:07:56 -0800714 migrate_page: This is used to compact the physical memory usage.
715 If the VM wants to relocate a page (maybe off a memory card
716 that is signalling imminent failure) it will pass a new page
717 and an old page to this function. migrate_page should
718 transfer any private data across and update any references
719 that it has to the page.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700720
Borislav Petkov422b14c2007-07-15 23:41:43 -0700721 launder_page: Called before freeing a page - it writes back the dirty page. To
722 prevent redirtying the page, it is kept locked during the whole
723 operation.
724
Andi Kleen25718732009-09-16 11:50:13 +0200725 error_remove_page: normally set to generic_error_remove_page if truncation
726 is ok for this address space. Used for memory failure handling.
727 Setting this implies you deal with pages going away under you,
728 unless you have them locked or reference counts increased.
729
730
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800731The File Object
732===============
733
734A file object represents a file opened by a process.
735
736
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700737struct file_operations
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800738----------------------
Linus Torvalds1da177e2005-04-16 15:20:36 -0700739
740This describes how the VFS can manipulate an open file. As of kernel
Borislav Petkov422b14c2007-07-15 23:41:43 -07007412.6.22, the following members are defined:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700742
743struct file_operations {
Borislav Petkov422b14c2007-07-15 23:41:43 -0700744 struct module *owner;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700745 loff_t (*llseek) (struct file *, loff_t, int);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700746 ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700747 ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
Badari Pulavarty027445c2006-09-30 23:28:46 -0700748 ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
749 ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700750 int (*readdir) (struct file *, void *, filldir_t);
751 unsigned int (*poll) (struct file *, struct poll_table_struct *);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700752 long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
753 long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700754 int (*mmap) (struct file *, struct vm_area_struct *);
755 int (*open) (struct inode *, struct file *);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700756 int (*flush) (struct file *);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700757 int (*release) (struct inode *, struct file *);
Christoph Hellwig7ea80852010-05-26 17:53:25 +0200758 int (*fsync) (struct file *, int datasync);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700759 int (*aio_fsync) (struct kiocb *, int datasync);
760 int (*fasync) (int, struct file *, int);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700761 int (*lock) (struct file *, int, struct file_lock *);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700762 ssize_t (*readv) (struct file *, const struct iovec *, unsigned long, loff_t *);
763 ssize_t (*writev) (struct file *, const struct iovec *, unsigned long, loff_t *);
764 ssize_t (*sendfile) (struct file *, loff_t *, size_t, read_actor_t, void *);
765 ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
766 unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
767 int (*check_flags)(int);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700768 int (*flock) (struct file *, int, struct file_lock *);
Borislav Petkov422b14c2007-07-15 23:41:43 -0700769 ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, size_t, unsigned int);
770 ssize_t (*splice_read)(struct file *, struct pipe_inode_info *, size_t, unsigned int);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700771};
772
773Again, all methods are called without any locks being held, unless
774otherwise noted.
775
776 llseek: called when the VFS needs to move the file position index
777
778 read: called by read(2) and related system calls
779
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700780 aio_read: called by io_submit(2) and other asynchronous I/O operations
781
Linus Torvalds1da177e2005-04-16 15:20:36 -0700782 write: called by write(2) and related system calls
783
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700784 aio_write: called by io_submit(2) and other asynchronous I/O operations
785
Linus Torvalds1da177e2005-04-16 15:20:36 -0700786 readdir: called when the VFS needs to read the directory contents
787
788 poll: called by the VFS when a process wants to check if there is
789 activity on this file and (optionally) go to sleep until there
790 is activity. Called by the select(2) and poll(2) system calls
791
Arnd Bergmannb19dd422010-07-04 00:15:10 +0200792 unlocked_ioctl: called by the ioctl(2) system call.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700793
794 compat_ioctl: called by the ioctl(2) system call when 32 bit system calls
795 are used on 64 bit kernels.
796
Linus Torvalds1da177e2005-04-16 15:20:36 -0700797 mmap: called by the mmap(2) system call
798
799 open: called by the VFS when an inode should be opened. When the VFS
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700800 opens a file, it creates a new "struct file". It then calls the
801 open method for the newly allocated file structure. You might
802 think that the open method really belongs in
803 "struct inode_operations", and you may be right. I think it's
804 done the way it is because it makes filesystems simpler to
805 implement. The open() method is a good place to initialize the
806 "private_data" member in the file structure if you want to point
807 to a device structure
808
809 flush: called by the close(2) system call to flush a file
Linus Torvalds1da177e2005-04-16 15:20:36 -0700810
811 release: called when the last reference to an open file is closed
812
813 fsync: called by the fsync(2) system call
814
815 fasync: called by the fcntl(2) system call when asynchronous
816 (non-blocking) mode is enabled for a file
817
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700818 lock: called by the fcntl(2) system call for F_GETLK, F_SETLK, and F_SETLKW
819 commands
820
821 readv: called by the readv(2) system call
822
823 writev: called by the writev(2) system call
824
825 sendfile: called by the sendfile(2) system call
826
827 get_unmapped_area: called by the mmap(2) system call
828
829 check_flags: called by the fcntl(2) system call for F_SETFL command
830
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700831 flock: called by the flock(2) system call
832
Pekka J Enbergd1195c52006-04-11 14:21:59 +0200833 splice_write: called by the VFS to splice data from a pipe to a file. This
834 method is used by the splice(2) system call
835
836 splice_read: called by the VFS to splice data from file to a pipe. This
837 method is used by the splice(2) system call
838
Linus Torvalds1da177e2005-04-16 15:20:36 -0700839Note that the file operations are implemented by the specific
840filesystem in which the inode resides. When opening a device node
841(character or block special) most filesystems will call special
842support routines in the VFS which will locate the required device
843driver information. These support routines replace the filesystem file
844operations with those for the device driver, and then proceed to call
845the new open() method for the file. This is how opening a device file
846in the filesystem eventually ends up calling the device driver open()
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700847method.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700848
849
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700850Directory Entry Cache (dcache)
851==============================
852
Linus Torvalds1da177e2005-04-16 15:20:36 -0700853
854struct dentry_operations
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700855------------------------
Linus Torvalds1da177e2005-04-16 15:20:36 -0700856
857This describes how a filesystem can overload the standard dentry
858operations. Dentries and the dcache are the domain of the VFS and the
859individual filesystem implementations. Device drivers have no business
860here. These methods may be set to NULL, as they are either optional or
Eric Dumazetc23fbb62007-05-08 00:26:18 -0700861the VFS uses a default. As of kernel 2.6.22, the following members are
Linus Torvalds1da177e2005-04-16 15:20:36 -0700862defined:
863
864struct dentry_operations {
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700865 int (*d_revalidate)(struct dentry *, struct nameidata *);
Nick Pigginb1e6a012011-01-07 17:49:28 +1100866 int (*d_hash)(const struct dentry *, const struct inode *,
867 struct qstr *);
Nick Piggin621e1552011-01-07 17:49:27 +1100868 int (*d_compare)(const struct dentry *, const struct inode *,
869 const struct dentry *, const struct inode *,
870 unsigned int, const char *, const struct qstr *);
Nick Pigginfe15ce42011-01-07 17:49:23 +1100871 int (*d_delete)(const struct dentry *);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700872 void (*d_release)(struct dentry *);
873 void (*d_iput)(struct dentry *, struct inode *);
Eric Dumazetc23fbb62007-05-08 00:26:18 -0700874 char *(*d_dname)(struct dentry *, char *, int);
David Howells9875cf82011-01-14 18:45:21 +0000875 struct vfsmount *(*d_automount)(struct path *);
Al Viro1aed3e42011-03-18 09:09:02 -0400876 int (*d_manage)(struct dentry *, bool);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700877};
878
879 d_revalidate: called when the VFS needs to revalidate a dentry. This
880 is called whenever a name look-up finds a dentry in the
881 dcache. Most filesystems leave this as NULL, because all their
882 dentries in the dcache are valid
883
Nick Piggin34286d62011-01-07 17:49:57 +1100884 d_revalidate may be called in rcu-walk mode (nd->flags & LOOKUP_RCU).
885 If in rcu-walk mode, the filesystem must revalidate the dentry without
886 blocking or storing to the dentry, d_parent and d_inode should not be
887 used without care (because they can go NULL), instead nd->inode should
888 be used.
889
890 If a situation is encountered that rcu-walk cannot handle, return
891 -ECHILD and it will be called again in ref-walk mode.
892
Nick Piggin621e1552011-01-07 17:49:27 +1100893 d_hash: called when the VFS adds a dentry to the hash table. The first
894 dentry passed to d_hash is the parent directory that the name is
Nick Pigginb1e6a012011-01-07 17:49:28 +1100895 to be hashed into. The inode is the dentry's inode.
896
897 Same locking and synchronisation rules as d_compare regarding
898 what is safe to dereference etc.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700899
Nick Piggin621e1552011-01-07 17:49:27 +1100900 d_compare: called to compare a dentry name with a given name. The first
901 dentry is the parent of the dentry to be compared, the second is
902 the parent's inode, then the dentry and inode (may be NULL) of the
903 child dentry. len and name string are properties of the dentry to be
904 compared. qstr is the name to compare it with.
905
906 Must be constant and idempotent, and should not take locks if
907 possible, and should not or store into the dentry or inodes.
908 Should not dereference pointers outside the dentry or inodes without
909 lots of care (eg. d_parent, d_inode, d_name should not be used).
910
911 However, our vfsmount is pinned, and RCU held, so the dentries and
912 inodes won't disappear, neither will our sb or filesystem module.
913 ->i_sb and ->d_sb may be used.
914
915 It is a tricky calling convention because it needs to be called under
916 "rcu-walk", ie. without any locks or references on things.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700917
Nick Pigginfe15ce42011-01-07 17:49:23 +1100918 d_delete: called when the last reference to a dentry is dropped and the
919 dcache is deciding whether or not to cache it. Return 1 to delete
920 immediately, or 0 to cache the dentry. Default is NULL which means to
921 always cache a reachable dentry. d_delete must be constant and
922 idempotent.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700923
924 d_release: called when a dentry is really deallocated
925
926 d_iput: called when a dentry loses its inode (just prior to its
927 being deallocated). The default when this is NULL is that the
928 VFS calls iput(). If you define this method, you must call
929 iput() yourself
930
Eric Dumazetc23fbb62007-05-08 00:26:18 -0700931 d_dname: called when the pathname of a dentry should be generated.
Matt LaPlanted9195882008-07-25 19:45:33 -0700932 Useful for some pseudo filesystems (sockfs, pipefs, ...) to delay
Eric Dumazetc23fbb62007-05-08 00:26:18 -0700933 pathname generation. (Instead of doing it when dentry is created,
Matt LaPlanted9195882008-07-25 19:45:33 -0700934 it's done only when the path is needed.). Real filesystems probably
Eric Dumazetc23fbb62007-05-08 00:26:18 -0700935 dont want to use it, because their dentries are present in global
936 dcache hash, so their hash should be an invariant. As no lock is
937 held, d_dname() should not try to modify the dentry itself, unless
938 appropriate SMP safety is used. CAUTION : d_path() logic is quite
939 tricky. The correct way to return for example "Hello" is to put it
940 at the end of the buffer, and returns a pointer to the first char.
941 dynamic_dname() helper function is provided to take care of this.
942
David Howells9875cf82011-01-14 18:45:21 +0000943 d_automount: called when an automount dentry is to be traversed (optional).
David Howellsea5b7782011-01-14 19:10:03 +0000944 This should create a new VFS mount record and return the record to the
945 caller. The caller is supplied with a path parameter giving the
946 automount directory to describe the automount target and the parent
947 VFS mount record to provide inheritable mount parameters. NULL should
948 be returned if someone else managed to make the automount first. If
949 the vfsmount creation failed, then an error code should be returned.
950 If -EISDIR is returned, then the directory will be treated as an
951 ordinary directory and returned to pathwalk to continue walking.
952
953 If a vfsmount is returned, the caller will attempt to mount it on the
954 mountpoint and will remove the vfsmount from its expiration list in
955 the case of failure. The vfsmount should be returned with 2 refs on
956 it to prevent automatic expiration - the caller will clean up the
957 additional ref.
David Howells9875cf82011-01-14 18:45:21 +0000958
959 This function is only used if DCACHE_NEED_AUTOMOUNT is set on the
960 dentry. This is set by __d_instantiate() if S_AUTOMOUNT is set on the
961 inode being added.
962
David Howellscc53ce52011-01-14 18:45:26 +0000963 d_manage: called to allow the filesystem to manage the transition from a
964 dentry (optional). This allows autofs, for example, to hold up clients
965 waiting to explore behind a 'mountpoint' whilst letting the daemon go
966 past and construct the subtree there. 0 should be returned to let the
967 calling process continue. -EISDIR can be returned to tell pathwalk to
968 use this directory as an ordinary directory and to ignore anything
969 mounted on it and not to check the automount flag. Any other error
970 code will abort pathwalk completely.
971
David Howellsab909112011-01-14 18:46:51 +0000972 If the 'rcu_walk' parameter is true, then the caller is doing a
973 pathwalk in RCU-walk mode. Sleeping is not permitted in this mode,
974 and the caller can be asked to leave it and call again by returing
975 -ECHILD.
976
David Howellscc53ce52011-01-14 18:45:26 +0000977 This function is only used if DCACHE_MANAGE_TRANSIT is set on the
978 dentry being transited from.
979
Eric Dumazetc23fbb62007-05-08 00:26:18 -0700980Example :
981
982static char *pipefs_dname(struct dentry *dent, char *buffer, int buflen)
983{
984 return dynamic_dname(dentry, buffer, buflen, "pipe:[%lu]",
985 dentry->d_inode->i_ino);
986}
987
Linus Torvalds1da177e2005-04-16 15:20:36 -0700988Each dentry has a pointer to its parent dentry, as well as a hash list
989of child dentries. Child dentries are basically like files in a
990directory.
991
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700992
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800993Directory Entry Cache API
Linus Torvalds1da177e2005-04-16 15:20:36 -0700994--------------------------
995
996There are a number of functions defined which permit a filesystem to
997manipulate dentries:
998
999 dget: open a new handle for an existing dentry (this just increments
1000 the usage count)
1001
1002 dput: close a handle for a dentry (decrements the usage count). If
Nick Pigginfe15ce42011-01-07 17:49:23 +11001003 the usage count drops to 0, and the dentry is still in its
1004 parent's hash, the "d_delete" method is called to check whether
1005 it should be cached. If it should not be cached, or if the dentry
1006 is not hashed, it is deleted. Otherwise cached dentries are put
1007 into an LRU list to be reclaimed on memory shortage.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001008
1009 d_drop: this unhashes a dentry from its parents hash list. A
Pekka J Enberg5ea626a2005-09-09 13:10:19 -07001010 subsequent call to dput() will deallocate the dentry if its
Linus Torvalds1da177e2005-04-16 15:20:36 -07001011 usage count drops to 0
1012
1013 d_delete: delete a dentry. If there are no other open references to
1014 the dentry then the dentry is turned into a negative dentry
1015 (the d_iput() method is called). If there are other
1016 references, then d_drop() is called instead
1017
1018 d_add: add a dentry to its parents hash list and then calls
1019 d_instantiate()
1020
1021 d_instantiate: add a dentry to the alias hash list for the inode and
1022 updates the "d_inode" member. The "i_count" member in the
1023 inode structure should be set/incremented. If the inode
1024 pointer is NULL, the dentry is called a "negative
1025 dentry". This function is commonly called when an inode is
1026 created for an existing negative dentry
1027
1028 d_lookup: look up a dentry given its parent and path name component
1029 It looks up the child of that given name from the dcache
1030 hash table. If it is found, the reference count is incremented
Zhaoleibe42c4c2008-12-01 14:34:58 -08001031 and the dentry is returned. The caller must use dput()
Linus Torvalds1da177e2005-04-16 15:20:36 -07001032 to free the dentry when it finishes using it.
1033
Pekka Enbergcbf8f0f2005-11-07 01:01:09 -08001034For further information on dentry locking, please refer to the document
1035Documentation/filesystems/dentry-locking.txt.
Pekka Enbergcc7d1f82005-11-07 01:01:08 -08001036
Miklos Szeredif84e3f52008-02-08 04:21:34 -08001037Mount Options
1038=============
1039
1040Parsing options
1041---------------
1042
1043On mount and remount the filesystem is passed a string containing a
1044comma separated list of mount options. The options can have either of
1045these forms:
1046
1047 option
1048 option=value
1049
1050The <linux/parser.h> header defines an API that helps parse these
1051options. There are plenty of examples on how to use it in existing
1052filesystems.
1053
1054Showing options
1055---------------
1056
1057If a filesystem accepts mount options, it must define show_options()
1058to show all the currently active options. The rules are:
1059
1060 - options MUST be shown which are not default or their values differ
1061 from the default
1062
1063 - options MAY be shown which are enabled by default or have their
1064 default value
1065
1066Options used only internally between a mount helper and the kernel
1067(such as file descriptors), or which only have an effect during the
1068mounting (such as ones controlling the creation of a journal) are exempt
1069from the above rules.
1070
1071The underlying reason for the above rules is to make sure, that a
1072mount can be accurately replicated (e.g. umounting and mounting again)
1073based on the information found in /proc/mounts.
1074
1075A simple method of saving options at mount/remount time and showing
1076them is provided with the save_mount_options() and
1077generic_show_options() helper functions. Please note, that using
1078these may have drawbacks. For more info see header comments for these
1079functions in fs/namespace.c.
Pekka Enbergcc7d1f82005-11-07 01:01:08 -08001080
1081Resources
1082=========
1083
1084(Note some of these resources are not up-to-date with the latest kernel
1085 version.)
1086
1087Creating Linux virtual filesystems. 2002
1088 <http://lwn.net/Articles/13325/>
1089
1090The Linux Virtual File-system Layer by Neil Brown. 1999
1091 <http://www.cse.unsw.edu.au/~neilb/oss/linux-commentary/vfs.html>
1092
1093A tour of the Linux VFS by Michael K. Johnson. 1996
1094 <http://www.tldp.org/LDP/khg/HyperNews/get/fs/vfstour.html>
1095
1096A small trail through the Linux kernel by Andries Brouwer. 2001
1097 <http://www.win.tue.nl/~aeb/linux/vfs/trail.html>