Loading...
1
2 Overview of the Linux Virtual File System
3
4 Original author: Richard Gooch <rgooch@atnf.csiro.au>
5
6 Last updated on June 24, 2007.
7
8 Copyright (C) 1999 Richard Gooch
9 Copyright (C) 2005 Pekka Enberg
10
11 This file is released under the GPLv2.
12
13
14Introduction
15============
16
17The 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.
26
27
28Directory Entry Cache (dcache)
29------------------------------
30
31The 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.
44
45
46The Inode Object
47----------------
48
49An 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).
56
57To 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.
65
66
67The File Object
68---------------
69
70Opening a file requires another operation: allocation of a file
71structure (this is the kernel-side implementation of file
72descriptors). The freshly allocated file structure is initialized with
73a 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
75called so the specific filesystem implementation can do its work. You
76can see that this is another switch performed by the VFS. The file
77structure is placed into the file descriptor table for the process.
78
79Reading, writing and closing files (and other assorted VFS operations)
80is done by using the userspace file descriptor to grab the appropriate
81file 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.
84
85
86Registering and Mounting a Filesystem
87=====================================
88
89To register and unregister a filesystem, use the following API
90functions:
91
92 #include <linux/fs.h>
93
94 extern int register_filesystem(struct file_system_type *);
95 extern int unregister_filesystem(struct file_system_type *);
96
97The passed struct file_system_type describes your filesystem. When a
98request is made to mount a filesystem onto a directory in your namespace,
99the VFS will call the appropriate mount() method for the specific
100filesystem. New vfsmount referring to the tree returned by ->mount()
101will be attached to the mountpoint, so that when pathname resolution
102reaches the mountpoint it will jump into the root of that vfsmount.
103
104You can see all filesystems that are registered to the kernel in the
105file /proc/filesystems.
106
107
108struct file_system_type
109-----------------------
110
111This describes the filesystem. As of kernel 2.6.39, the following
112members are defined:
113
114struct file_system_type {
115 const char *name;
116 int fs_flags;
117 struct dentry *(*mount) (struct file_system_type *, int,
118 const char *, void *);
119 void (*kill_sb) (struct super_block *);
120 struct module *owner;
121 struct file_system_type * next;
122 struct list_head fs_supers;
123 struct lock_class_key s_lock_key;
124 struct lock_class_key s_umount_key;
125};
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
132 mount: the method to call when a new instance of this
133 filesystem should be mounted
134
135 kill_sb: the method to call when an instance of this filesystem
136 should be shut down
137
138 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
143 s_lock_key, s_umount_key: lockdep-specific
144
145The mount() method has the following arguments:
146
147 struct file_system_type *fs_type: describes the filesystem, partly initialized
148 by the specific filesystem code
149
150 int flags: mount flags
151
152 const char *dev_name: the device name we are mounting.
153
154 void *data: arbitrary mount options, usually comes as an ASCII
155 string (see "Mount Options" section)
156
157The 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).
160
161The 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.
171
172The most interesting member of the superblock structure that the
173mount() method fills in is the "s_op" field. This is a pointer to
174a "struct super_operations" which describes the next level of the
175filesystem implementation.
176
177Usually, a filesystem uses one of the generic mount() implementations
178and provides a fill_super() callback instead. The generic variants are:
179
180 mount_bdev: mount a filesystem residing on a block device
181
182 mount_nodev: mount a filesystem that is not backed by a device
183
184 mount_single: mount a filesystem which shares the instance between
185 all mounts
186
187A fill_super() callback implementation has the following arguments:
188
189 struct super_block *sb: the superblock structure. The callback
190 must initialize this properly.
191
192 void *data: arbitrary mount options, usually comes as an ASCII
193 string (see "Mount Options" section)
194
195 int silent: whether or not to be silent on error
196
197
198The Superblock Object
199=====================
200
201A superblock object represents a mounted filesystem.
202
203
204struct super_operations
205-----------------------
206
207This describes how the VFS can manipulate the superblock of your
208filesystem. As of kernel 2.6.22, the following members are defined:
209
210struct super_operations {
211 struct inode *(*alloc_inode)(struct super_block *sb);
212 void (*destroy_inode)(struct inode *);
213
214 void (*dirty_inode) (struct inode *, int flags);
215 int (*write_inode) (struct inode *, int);
216 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);
221 int (*freeze_fs) (struct super_block *);
222 int (*unfreeze_fs) (struct super_block *);
223 int (*statfs) (struct dentry *, struct kstatfs *);
224 int (*remount_fs) (struct super_block *, int *, char *);
225 void (*clear_inode) (struct inode *);
226 void (*umount_begin) (struct super_block *);
227
228 int (*show_options)(struct seq_file *, struct dentry *);
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);
232 int (*nr_cached_objects)(struct super_block *);
233 void (*free_cached_objects)(struct super_block *, int);
234};
235
236All methods are called without any locks being held, unless otherwise
237noted. This means that most methods can block safely. All methods are
238only called from a process context (i.e. not from an interrupt handler
239or bottom half).
240
241 alloc_inode: this method is called by inode_alloc() to allocate memory
242 for struct inode and initialize it. If this function is not
243 defined, a simple 'struct inode' is allocated. Normally
244 alloc_inode will be used to allocate a larger structure which
245 contains a 'struct inode' embedded within it.
246
247 destroy_inode: this method is called by destroy_inode() to release
248 resources allocated for struct inode. It is only required if
249 ->alloc_inode was defined and simply undoes anything done by
250 ->alloc_inode.
251
252 dirty_inode: this method is called by the VFS to mark an inode dirty.
253
254 write_inode: this method is called when the VFS needs to write an
255 inode to disc. The second parameter indicates whether the write
256 should be synchronous or not, not all filesystems check this flag.
257
258 drop_inode: called when the last access to the inode is dropped,
259 with the inode->i_lock spinlock held.
260
261 This method should be either NULL (normal UNIX filesystem
262 semantics) or "generic_delete_inode" (for filesystems that do not
263 want to cache inodes - causing "delete_inode" to always be
264 called regardless of the value of i_nlink)
265
266 The "generic_delete_inode()" behavior is equivalent to the
267 old practice of using "force_delete" in the put_inode() case,
268 but does not have the races that the "force_delete()" approach
269 had.
270
271 delete_inode: called when the VFS wants to delete an inode
272
273 put_super: called when the VFS wishes to free the superblock
274 (i.e. unmount). This is called with the superblock lock held
275
276 write_super: called when the VFS superblock needs to be written to
277 disc. This method is optional
278
279 sync_fs: called when VFS is writing out all dirty data associated with
280 a superblock. The second parameter indicates whether the method
281 should wait until the write out has been completed. Optional.
282
283 freeze_fs: called when VFS is locking a filesystem and
284 forcing it into a consistent state. This method is currently
285 used by the Logical Volume Manager (LVM).
286
287 unfreeze_fs: called when VFS is unlocking a filesystem and making it writable
288 again.
289
290 statfs: called when the VFS needs to get filesystem statistics.
291
292 remount_fs: called when the filesystem is remounted. This is called
293 with the kernel lock held
294
295 clear_inode: called then the VFS clears the inode. Optional
296
297 umount_begin: called when the VFS is unmounting a filesystem.
298
299 show_options: called by the VFS to show mount options for
300 /proc/<pid>/mounts. (see "Mount Options" section)
301
302 quota_read: called by the VFS to read from filesystem quota file.
303
304 quota_write: called by the VFS to write to filesystem quota file.
305
306 nr_cached_objects: called by the sb cache shrinking function for the
307 filesystem to return the number of freeable cached objects it contains.
308 Optional.
309
310 free_cache_objects: called by the sb cache shrinking function for the
311 filesystem to scan the number of objects indicated to try to free them.
312 Optional, but any filesystem implementing this method needs to also
313 implement ->nr_cached_objects for it to be called correctly.
314
315 We can't do anything with any errors that the filesystem might
316 encountered, hence the void return type. This will never be called if
317 the VM is trying to reclaim under GFP_NOFS conditions, hence this
318 method does not need to handle that situation itself.
319
320 Implementations must include conditional reschedule calls inside any
321 scanning loop that is done. This allows the VFS to determine
322 appropriate scan batch sizes without having to worry about whether
323 implementations will cause holdoff problems due to large scan batch
324 sizes.
325
326Whoever sets up the inode is responsible for filling in the "i_op" field. This
327is a pointer to a "struct inode_operations" which describes the methods that
328can be performed on individual inodes.
329
330
331The Inode Object
332================
333
334An inode object represents an object within the filesystem.
335
336
337struct inode_operations
338-----------------------
339
340This describes how the VFS can manipulate an inode in your
341filesystem. As of kernel 2.6.22, the following members are defined:
342
343struct inode_operations {
344 int (*create) (struct inode *,struct dentry *, umode_t, struct nameidata *);
345 struct dentry * (*lookup) (struct inode *,struct dentry *, struct nameidata *);
346 int (*link) (struct dentry *,struct inode *,struct dentry *);
347 int (*unlink) (struct inode *,struct dentry *);
348 int (*symlink) (struct inode *,struct dentry *,const char *);
349 int (*mkdir) (struct inode *,struct dentry *,umode_t);
350 int (*rmdir) (struct inode *,struct dentry *);
351 int (*mknod) (struct inode *,struct dentry *,umode_t,dev_t);
352 int (*rename) (struct inode *, struct dentry *,
353 struct inode *, struct dentry *);
354 int (*readlink) (struct dentry *, char __user *,int);
355 void * (*follow_link) (struct dentry *, struct nameidata *);
356 void (*put_link) (struct dentry *, struct nameidata *, void *);
357 void (*truncate) (struct inode *);
358 int (*permission) (struct inode *, int);
359 int (*get_acl)(struct inode *, int);
360 int (*setattr) (struct dentry *, struct iattr *);
361 int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
362 int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
363 ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
364 ssize_t (*listxattr) (struct dentry *, char *, size_t);
365 int (*removexattr) (struct dentry *, const char *);
366 void (*update_time)(struct inode *, struct timespec *, int);
367};
368
369Again, all methods are called without any locks being held, unless
370otherwise noted.
371
372 create: called by the open(2) and creat(2) system calls. Only
373 required if you want to support regular files. The dentry you
374 get should not have an inode (i.e. it should be a negative
375 dentry). Here you will probably call d_instantiate() with the
376 dentry and the newly created inode
377
378 lookup: called when the VFS needs to look up an inode in a parent
379 directory. The name to look for is found in the dentry. This
380 method must call d_add() to insert the found inode into the
381 dentry. The "i_count" field in the inode structure should be
382 incremented. If the named inode does not exist a NULL inode
383 should be inserted into the dentry (this is called a negative
384 dentry). Returning an error code from this routine must only
385 be done on a real error, otherwise creating inodes with system
386 calls like create(2), mknod(2), mkdir(2) and so on will fail.
387 If you wish to overload the dentry methods then you should
388 initialise the "d_dop" field in the dentry; this is a pointer
389 to a struct "dentry_operations".
390 This method is called with the directory inode semaphore held
391
392 link: called by the link(2) system call. Only required if you want
393 to support hard links. You will probably need to call
394 d_instantiate() just as you would in the create() method
395
396 unlink: called by the unlink(2) system call. Only required if you
397 want to support deleting inodes
398
399 symlink: called by the symlink(2) system call. Only required if you
400 want to support symlinks. You will probably need to call
401 d_instantiate() just as you would in the create() method
402
403 mkdir: called by the mkdir(2) system call. Only required if you want
404 to support creating subdirectories. You will probably need to
405 call d_instantiate() just as you would in the create() method
406
407 rmdir: called by the rmdir(2) system call. Only required if you want
408 to support deleting subdirectories
409
410 mknod: called by the mknod(2) system call to create a device (char,
411 block) inode or a named pipe (FIFO) or socket. Only required
412 if you want to support creating these types of inodes. You
413 will probably need to call d_instantiate() just as you would
414 in the create() method
415
416 rename: called by the rename(2) system call to rename the object to
417 have the parent and name given by the second inode and dentry.
418
419 readlink: called by the readlink(2) system call. Only required if
420 you want to support reading symbolic links
421
422 follow_link: called by the VFS to follow a symbolic link to the
423 inode it points to. Only required if you want to support
424 symbolic links. This method returns a void pointer cookie
425 that is passed to put_link().
426
427 put_link: called by the VFS to release resources allocated by
428 follow_link(). The cookie returned by follow_link() is passed
429 to this method as the last parameter. It is used by
430 filesystems such as NFS where page cache is not stable
431 (i.e. page that was installed when the symbolic link walk
432 started might not be in the page cache at the end of the
433 walk).
434
435 truncate: Deprecated. This will not be called if ->setsize is defined.
436 Called by the VFS to change the size of a file. The
437 i_size field of the inode is set to the desired size by the
438 VFS before this method is called. This method is called by
439 the truncate(2) system call and related functionality.
440
441 Note: ->truncate and vmtruncate are deprecated. Do not add new
442 instances/calls of these. Filesystems should be converted to do their
443 truncate sequence via ->setattr().
444
445 permission: called by the VFS to check for access rights on a POSIX-like
446 filesystem.
447
448 May be called in rcu-walk mode (mask & MAY_NOT_BLOCK). If in rcu-walk
449 mode, the filesystem must check the permission without blocking or
450 storing to the inode.
451
452 If a situation is encountered that rcu-walk cannot handle, return
453 -ECHILD and it will be called again in ref-walk mode.
454
455 setattr: called by the VFS to set attributes for a file. This method
456 is called by chmod(2) and related system calls.
457
458 getattr: called by the VFS to get attributes of a file. This method
459 is called by stat(2) and related system calls.
460
461 setxattr: called by the VFS to set an extended attribute for a file.
462 Extended attribute is a name:value pair associated with an
463 inode. This method is called by setxattr(2) system call.
464
465 getxattr: called by the VFS to retrieve the value of an extended
466 attribute name. This method is called by getxattr(2) function
467 call.
468
469 listxattr: called by the VFS to list all extended attributes for a
470 given file. This method is called by listxattr(2) system call.
471
472 removexattr: called by the VFS to remove an extended attribute from
473 a file. This method is called by removexattr(2) system call.
474
475 update_time: called by the VFS to update a specific time or the i_version of
476 an inode. If this is not defined the VFS will update the inode itself
477 and call mark_inode_dirty_sync.
478
479The Address Space Object
480========================
481
482The address space object is used to group and manage pages in the page
483cache. It can be used to keep track of the pages in a file (or
484anything else) and also track the mapping of sections of the file into
485process address spaces.
486
487There are a number of distinct yet related services that an
488address-space can provide. These include communicating memory
489pressure, page lookup by address, and keeping track of pages tagged as
490Dirty or Writeback.
491
492The first can be used independently to the others. The VM can try to
493either write dirty pages in order to clean them, or release clean
494pages in order to reuse them. To do this it can call the ->writepage
495method on dirty pages, and ->releasepage on clean pages with
496PagePrivate set. Clean pages without PagePrivate and with no external
497references will be released without notice being given to the
498address_space.
499
500To achieve this functionality, pages need to be placed on an LRU with
501lru_cache_add and mark_page_active needs to be called whenever the
502page is used.
503
504Pages are normally kept in a radix tree index by ->index. This tree
505maintains information about the PG_Dirty and PG_Writeback status of
506each page, so that pages with either of these flags can be found
507quickly.
508
509The Dirty tag is primarily used by mpage_writepages - the default
510->writepages method. It uses the tag to find dirty pages to call
511->writepage on. If mpage_writepages is not used (i.e. the address
512provides its own ->writepages) , the PAGECACHE_TAG_DIRTY tag is
513almost unused. write_inode_now and sync_inode do use it (through
514__sync_single_inode) to check if ->writepages has been successful in
515writing out the whole address_space.
516
517The Writeback tag is used by filemap*wait* and sync_page* functions,
518via filemap_fdatawait_range, to wait for all writeback to
519complete. While waiting ->sync_page (if defined) will be called on
520each page that is found to require writeback.
521
522An address_space handler may attach extra information to a page,
523typically using the 'private' field in the 'struct page'. If such
524information is attached, the PG_Private flag should be set. This will
525cause various VM routines to make extra calls into the address_space
526handler to deal with that data.
527
528An address space acts as an intermediate between storage and
529application. Data is read into the address space a whole page at a
530time, and provided to the application either by copying of the page,
531or by memory-mapping the page.
532Data is written into the address space by the application, and then
533written-back to storage typically in whole pages, however the
534address_space has finer control of write sizes.
535
536The read process essentially only requires 'readpage'. The write
537process is more complicated and uses write_begin/write_end or
538set_page_dirty to write data into the address_space, and writepage,
539sync_page, and writepages to writeback data to storage.
540
541Adding and removing pages to/from an address_space is protected by the
542inode's i_mutex.
543
544When data is written to a page, the PG_Dirty flag should be set. It
545typically remains set until writepage asks for it to be written. This
546should clear PG_Dirty and set PG_Writeback. It can be actually
547written at any point after PG_Dirty is clear. Once it is known to be
548safe, PG_Writeback is cleared.
549
550Writeback makes use of a writeback_control structure...
551
552struct address_space_operations
553-------------------------------
554
555This describes how the VFS can manipulate mapping of a file to page cache in
556your filesystem. As of kernel 2.6.22, the following members are defined:
557
558struct address_space_operations {
559 int (*writepage)(struct page *page, struct writeback_control *wbc);
560 int (*readpage)(struct file *, struct page *);
561 int (*sync_page)(struct page *);
562 int (*writepages)(struct address_space *, struct writeback_control *);
563 int (*set_page_dirty)(struct page *page);
564 int (*readpages)(struct file *filp, struct address_space *mapping,
565 struct list_head *pages, unsigned nr_pages);
566 int (*write_begin)(struct file *, struct address_space *mapping,
567 loff_t pos, unsigned len, unsigned flags,
568 struct page **pagep, void **fsdata);
569 int (*write_end)(struct file *, struct address_space *mapping,
570 loff_t pos, unsigned len, unsigned copied,
571 struct page *page, void *fsdata);
572 sector_t (*bmap)(struct address_space *, sector_t);
573 int (*invalidatepage) (struct page *, unsigned long);
574 int (*releasepage) (struct page *, int);
575 void (*freepage)(struct page *);
576 ssize_t (*direct_IO)(int, struct kiocb *, const struct iovec *iov,
577 loff_t offset, unsigned long nr_segs);
578 struct page* (*get_xip_page)(struct address_space *, sector_t,
579 int);
580 /* migrate the contents of a page to the specified target */
581 int (*migratepage) (struct page *, struct page *);
582 int (*launder_page) (struct page *);
583 int (*error_remove_page) (struct mapping *mapping, struct page *page);
584};
585
586 writepage: called by the VM to write a dirty page to backing store.
587 This may happen for data integrity reasons (i.e. 'sync'), or
588 to free up memory (flush). The difference can be seen in
589 wbc->sync_mode.
590 The PG_Dirty flag has been cleared and PageLocked is true.
591 writepage should start writeout, should set PG_Writeback,
592 and should make sure the page is unlocked, either synchronously
593 or asynchronously when the write operation completes.
594
595 If wbc->sync_mode is WB_SYNC_NONE, ->writepage doesn't have to
596 try too hard if there are problems, and may choose to write out
597 other pages from the mapping if that is easier (e.g. due to
598 internal dependencies). If it chooses not to start writeout, it
599 should return AOP_WRITEPAGE_ACTIVATE so that the VM will not keep
600 calling ->writepage on that page.
601
602 See the file "Locking" for more details.
603
604 readpage: called by the VM to read a page from backing store.
605 The page will be Locked when readpage is called, and should be
606 unlocked and marked uptodate once the read completes.
607 If ->readpage discovers that it needs to unlock the page for
608 some reason, it can do so, and then return AOP_TRUNCATED_PAGE.
609 In this case, the page will be relocated, relocked and if
610 that all succeeds, ->readpage will be called again.
611
612 sync_page: called by the VM to notify the backing store to perform all
613 queued I/O operations for a page. I/O operations for other pages
614 associated with this address_space object may also be performed.
615
616 This function is optional and is called only for pages with
617 PG_Writeback set while waiting for the writeback to complete.
618
619 writepages: called by the VM to write out pages associated with the
620 address_space object. If wbc->sync_mode is WBC_SYNC_ALL, then
621 the writeback_control will specify a range of pages that must be
622 written out. If it is WBC_SYNC_NONE, then a nr_to_write is given
623 and that many pages should be written if possible.
624 If no ->writepages is given, then mpage_writepages is used
625 instead. This will choose pages from the address space that are
626 tagged as DIRTY and will pass them to ->writepage.
627
628 set_page_dirty: called by the VM to set a page dirty.
629 This is particularly needed if an address space attaches
630 private data to a page, and that data needs to be updated when
631 a page is dirtied. This is called, for example, when a memory
632 mapped page gets modified.
633 If defined, it should set the PageDirty flag, and the
634 PAGECACHE_TAG_DIRTY tag in the radix tree.
635
636 readpages: called by the VM to read pages associated with the address_space
637 object. This is essentially just a vector version of
638 readpage. Instead of just one page, several pages are
639 requested.
640 readpages is only used for read-ahead, so read errors are
641 ignored. If anything goes wrong, feel free to give up.
642
643 write_begin:
644 Called by the generic buffered write code to ask the filesystem to
645 prepare to write len bytes at the given offset in the file. The
646 address_space should check that the write will be able to complete,
647 by allocating space if necessary and doing any other internal
648 housekeeping. If the write will update parts of any basic-blocks on
649 storage, then those blocks should be pre-read (if they haven't been
650 read already) so that the updated blocks can be written out properly.
651
652 The filesystem must return the locked pagecache page for the specified
653 offset, in *pagep, for the caller to write into.
654
655 It must be able to cope with short writes (where the length passed to
656 write_begin is greater than the number of bytes copied into the page).
657
658 flags is a field for AOP_FLAG_xxx flags, described in
659 include/linux/fs.h.
660
661 A void * may be returned in fsdata, which then gets passed into
662 write_end.
663
664 Returns 0 on success; < 0 on failure (which is the error code), in
665 which case write_end is not called.
666
667 write_end: After a successful write_begin, and data copy, write_end must
668 be called. len is the original len passed to write_begin, and copied
669 is the amount that was able to be copied (copied == len is always true
670 if write_begin was called with the AOP_FLAG_UNINTERRUPTIBLE flag).
671
672 The filesystem must take care of unlocking the page and releasing it
673 refcount, and updating i_size.
674
675 Returns < 0 on failure, otherwise the number of bytes (<= 'copied')
676 that were able to be copied into pagecache.
677
678 bmap: called by the VFS to map a logical block offset within object to
679 physical block number. This method is used by the FIBMAP
680 ioctl and for working with swap-files. To be able to swap to
681 a file, the file must have a stable mapping to a block
682 device. The swap system does not go through the filesystem
683 but instead uses bmap to find out where the blocks in the file
684 are and uses those addresses directly.
685
686
687 invalidatepage: If a page has PagePrivate set, then invalidatepage
688 will be called when part or all of the page is to be removed
689 from the address space. This generally corresponds to either a
690 truncation or a complete invalidation of the address space
691 (in the latter case 'offset' will always be 0).
692 Any private data associated with the page should be updated
693 to reflect this truncation. If offset is 0, then
694 the private data should be released, because the page
695 must be able to be completely discarded. This may be done by
696 calling the ->releasepage function, but in this case the
697 release MUST succeed.
698
699 releasepage: releasepage is called on PagePrivate pages to indicate
700 that the page should be freed if possible. ->releasepage
701 should remove any private data from the page and clear the
702 PagePrivate flag. If releasepage() fails for some reason, it must
703 indicate failure with a 0 return value.
704 releasepage() is used in two distinct though related cases. The
705 first is when the VM finds a clean page with no active users and
706 wants to make it a free page. If ->releasepage succeeds, the
707 page will be removed from the address_space and become free.
708
709 The second case is when a request has been made to invalidate
710 some or all pages in an address_space. This can happen
711 through the fadvice(POSIX_FADV_DONTNEED) system call or by the
712 filesystem explicitly requesting it as nfs and 9fs do (when
713 they believe the cache may be out of date with storage) by
714 calling invalidate_inode_pages2().
715 If the filesystem makes such a call, and needs to be certain
716 that all pages are invalidated, then its releasepage will
717 need to ensure this. Possibly it can clear the PageUptodate
718 bit if it cannot free private data yet.
719
720 freepage: freepage is called once the page is no longer visible in
721 the page cache in order to allow the cleanup of any private
722 data. Since it may be called by the memory reclaimer, it
723 should not assume that the original address_space mapping still
724 exists, and it should not block.
725
726 direct_IO: called by the generic read/write routines to perform
727 direct_IO - that is IO requests which bypass the page cache
728 and transfer data directly between the storage and the
729 application's address space.
730
731 get_xip_page: called by the VM to translate a block number to a page.
732 The page is valid until the corresponding filesystem is unmounted.
733 Filesystems that want to use execute-in-place (XIP) need to implement
734 it. An example implementation can be found in fs/ext2/xip.c.
735
736 migrate_page: This is used to compact the physical memory usage.
737 If the VM wants to relocate a page (maybe off a memory card
738 that is signalling imminent failure) it will pass a new page
739 and an old page to this function. migrate_page should
740 transfer any private data across and update any references
741 that it has to the page.
742
743 launder_page: Called before freeing a page - it writes back the dirty page. To
744 prevent redirtying the page, it is kept locked during the whole
745 operation.
746
747 error_remove_page: normally set to generic_error_remove_page if truncation
748 is ok for this address space. Used for memory failure handling.
749 Setting this implies you deal with pages going away under you,
750 unless you have them locked or reference counts increased.
751
752
753The File Object
754===============
755
756A file object represents a file opened by a process.
757
758
759struct file_operations
760----------------------
761
762This describes how the VFS can manipulate an open file. As of kernel
7633.5, the following members are defined:
764
765struct file_operations {
766 struct module *owner;
767 loff_t (*llseek) (struct file *, loff_t, int);
768 ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
769 ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
770 ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
771 ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
772 int (*readdir) (struct file *, void *, filldir_t);
773 unsigned int (*poll) (struct file *, struct poll_table_struct *);
774 long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
775 long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
776 int (*mmap) (struct file *, struct vm_area_struct *);
777 int (*open) (struct inode *, struct file *);
778 int (*flush) (struct file *);
779 int (*release) (struct inode *, struct file *);
780 int (*fsync) (struct file *, loff_t, loff_t, int datasync);
781 int (*aio_fsync) (struct kiocb *, int datasync);
782 int (*fasync) (int, struct file *, int);
783 int (*lock) (struct file *, int, struct file_lock *);
784 ssize_t (*readv) (struct file *, const struct iovec *, unsigned long, loff_t *);
785 ssize_t (*writev) (struct file *, const struct iovec *, unsigned long, loff_t *);
786 ssize_t (*sendfile) (struct file *, loff_t *, size_t, read_actor_t, void *);
787 ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
788 unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
789 int (*check_flags)(int);
790 int (*flock) (struct file *, int, struct file_lock *);
791 ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, size_t, unsigned int);
792 ssize_t (*splice_read)(struct file *, struct pipe_inode_info *, size_t, unsigned int);
793 int (*setlease)(struct file *, long arg, struct file_lock **);
794 long (*fallocate)(struct file *, int mode, loff_t offset, loff_t len);
795};
796
797Again, all methods are called without any locks being held, unless
798otherwise noted.
799
800 llseek: called when the VFS needs to move the file position index
801
802 read: called by read(2) and related system calls
803
804 aio_read: called by io_submit(2) and other asynchronous I/O operations
805
806 write: called by write(2) and related system calls
807
808 aio_write: called by io_submit(2) and other asynchronous I/O operations
809
810 readdir: called when the VFS needs to read the directory contents
811
812 poll: called by the VFS when a process wants to check if there is
813 activity on this file and (optionally) go to sleep until there
814 is activity. Called by the select(2) and poll(2) system calls
815
816 unlocked_ioctl: called by the ioctl(2) system call.
817
818 compat_ioctl: called by the ioctl(2) system call when 32 bit system calls
819 are used on 64 bit kernels.
820
821 mmap: called by the mmap(2) system call
822
823 open: called by the VFS when an inode should be opened. When the VFS
824 opens a file, it creates a new "struct file". It then calls the
825 open method for the newly allocated file structure. You might
826 think that the open method really belongs in
827 "struct inode_operations", and you may be right. I think it's
828 done the way it is because it makes filesystems simpler to
829 implement. The open() method is a good place to initialize the
830 "private_data" member in the file structure if you want to point
831 to a device structure
832
833 flush: called by the close(2) system call to flush a file
834
835 release: called when the last reference to an open file is closed
836
837 fsync: called by the fsync(2) system call
838
839 fasync: called by the fcntl(2) system call when asynchronous
840 (non-blocking) mode is enabled for a file
841
842 lock: called by the fcntl(2) system call for F_GETLK, F_SETLK, and F_SETLKW
843 commands
844
845 readv: called by the readv(2) system call
846
847 writev: called by the writev(2) system call
848
849 sendfile: called by the sendfile(2) system call
850
851 get_unmapped_area: called by the mmap(2) system call
852
853 check_flags: called by the fcntl(2) system call for F_SETFL command
854
855 flock: called by the flock(2) system call
856
857 splice_write: called by the VFS to splice data from a pipe to a file. This
858 method is used by the splice(2) system call
859
860 splice_read: called by the VFS to splice data from file to a pipe. This
861 method is used by the splice(2) system call
862
863 setlease: called by the VFS to set or release a file lock lease.
864 setlease has the file_lock_lock held and must not sleep.
865
866 fallocate: called by the VFS to preallocate blocks or punch a hole.
867
868Note that the file operations are implemented by the specific
869filesystem in which the inode resides. When opening a device node
870(character or block special) most filesystems will call special
871support routines in the VFS which will locate the required device
872driver information. These support routines replace the filesystem file
873operations with those for the device driver, and then proceed to call
874the new open() method for the file. This is how opening a device file
875in the filesystem eventually ends up calling the device driver open()
876method.
877
878
879Directory Entry Cache (dcache)
880==============================
881
882
883struct dentry_operations
884------------------------
885
886This describes how a filesystem can overload the standard dentry
887operations. Dentries and the dcache are the domain of the VFS and the
888individual filesystem implementations. Device drivers have no business
889here. These methods may be set to NULL, as they are either optional or
890the VFS uses a default. As of kernel 2.6.22, the following members are
891defined:
892
893struct dentry_operations {
894 int (*d_revalidate)(struct dentry *, struct nameidata *);
895 int (*d_hash)(const struct dentry *, const struct inode *,
896 struct qstr *);
897 int (*d_compare)(const struct dentry *, const struct inode *,
898 const struct dentry *, const struct inode *,
899 unsigned int, const char *, const struct qstr *);
900 int (*d_delete)(const struct dentry *);
901 void (*d_release)(struct dentry *);
902 void (*d_iput)(struct dentry *, struct inode *);
903 char *(*d_dname)(struct dentry *, char *, int);
904 struct vfsmount *(*d_automount)(struct path *);
905 int (*d_manage)(struct dentry *, bool);
906};
907
908 d_revalidate: called when the VFS needs to revalidate a dentry. This
909 is called whenever a name look-up finds a dentry in the
910 dcache. Most filesystems leave this as NULL, because all their
911 dentries in the dcache are valid
912
913 d_revalidate may be called in rcu-walk mode (nd->flags & LOOKUP_RCU).
914 If in rcu-walk mode, the filesystem must revalidate the dentry without
915 blocking or storing to the dentry, d_parent and d_inode should not be
916 used without care (because they can go NULL), instead nd->inode should
917 be used.
918
919 If a situation is encountered that rcu-walk cannot handle, return
920 -ECHILD and it will be called again in ref-walk mode.
921
922 d_hash: called when the VFS adds a dentry to the hash table. The first
923 dentry passed to d_hash is the parent directory that the name is
924 to be hashed into. The inode is the dentry's inode.
925
926 Same locking and synchronisation rules as d_compare regarding
927 what is safe to dereference etc.
928
929 d_compare: called to compare a dentry name with a given name. The first
930 dentry is the parent of the dentry to be compared, the second is
931 the parent's inode, then the dentry and inode (may be NULL) of the
932 child dentry. len and name string are properties of the dentry to be
933 compared. qstr is the name to compare it with.
934
935 Must be constant and idempotent, and should not take locks if
936 possible, and should not or store into the dentry or inodes.
937 Should not dereference pointers outside the dentry or inodes without
938 lots of care (eg. d_parent, d_inode, d_name should not be used).
939
940 However, our vfsmount is pinned, and RCU held, so the dentries and
941 inodes won't disappear, neither will our sb or filesystem module.
942 ->i_sb and ->d_sb may be used.
943
944 It is a tricky calling convention because it needs to be called under
945 "rcu-walk", ie. without any locks or references on things.
946
947 d_delete: called when the last reference to a dentry is dropped and the
948 dcache is deciding whether or not to cache it. Return 1 to delete
949 immediately, or 0 to cache the dentry. Default is NULL which means to
950 always cache a reachable dentry. d_delete must be constant and
951 idempotent.
952
953 d_release: called when a dentry is really deallocated
954
955 d_iput: called when a dentry loses its inode (just prior to its
956 being deallocated). The default when this is NULL is that the
957 VFS calls iput(). If you define this method, you must call
958 iput() yourself
959
960 d_dname: called when the pathname of a dentry should be generated.
961 Useful for some pseudo filesystems (sockfs, pipefs, ...) to delay
962 pathname generation. (Instead of doing it when dentry is created,
963 it's done only when the path is needed.). Real filesystems probably
964 dont want to use it, because their dentries are present in global
965 dcache hash, so their hash should be an invariant. As no lock is
966 held, d_dname() should not try to modify the dentry itself, unless
967 appropriate SMP safety is used. CAUTION : d_path() logic is quite
968 tricky. The correct way to return for example "Hello" is to put it
969 at the end of the buffer, and returns a pointer to the first char.
970 dynamic_dname() helper function is provided to take care of this.
971
972 d_automount: called when an automount dentry is to be traversed (optional).
973 This should create a new VFS mount record and return the record to the
974 caller. The caller is supplied with a path parameter giving the
975 automount directory to describe the automount target and the parent
976 VFS mount record to provide inheritable mount parameters. NULL should
977 be returned if someone else managed to make the automount first. If
978 the vfsmount creation failed, then an error code should be returned.
979 If -EISDIR is returned, then the directory will be treated as an
980 ordinary directory and returned to pathwalk to continue walking.
981
982 If a vfsmount is returned, the caller will attempt to mount it on the
983 mountpoint and will remove the vfsmount from its expiration list in
984 the case of failure. The vfsmount should be returned with 2 refs on
985 it to prevent automatic expiration - the caller will clean up the
986 additional ref.
987
988 This function is only used if DCACHE_NEED_AUTOMOUNT is set on the
989 dentry. This is set by __d_instantiate() if S_AUTOMOUNT is set on the
990 inode being added.
991
992 d_manage: called to allow the filesystem to manage the transition from a
993 dentry (optional). This allows autofs, for example, to hold up clients
994 waiting to explore behind a 'mountpoint' whilst letting the daemon go
995 past and construct the subtree there. 0 should be returned to let the
996 calling process continue. -EISDIR can be returned to tell pathwalk to
997 use this directory as an ordinary directory and to ignore anything
998 mounted on it and not to check the automount flag. Any other error
999 code will abort pathwalk completely.
1000
1001 If the 'rcu_walk' parameter is true, then the caller is doing a
1002 pathwalk in RCU-walk mode. Sleeping is not permitted in this mode,
1003 and the caller can be asked to leave it and call again by returning
1004 -ECHILD.
1005
1006 This function is only used if DCACHE_MANAGE_TRANSIT is set on the
1007 dentry being transited from.
1008
1009Example :
1010
1011static char *pipefs_dname(struct dentry *dent, char *buffer, int buflen)
1012{
1013 return dynamic_dname(dentry, buffer, buflen, "pipe:[%lu]",
1014 dentry->d_inode->i_ino);
1015}
1016
1017Each dentry has a pointer to its parent dentry, as well as a hash list
1018of child dentries. Child dentries are basically like files in a
1019directory.
1020
1021
1022Directory Entry Cache API
1023--------------------------
1024
1025There are a number of functions defined which permit a filesystem to
1026manipulate dentries:
1027
1028 dget: open a new handle for an existing dentry (this just increments
1029 the usage count)
1030
1031 dput: close a handle for a dentry (decrements the usage count). If
1032 the usage count drops to 0, and the dentry is still in its
1033 parent's hash, the "d_delete" method is called to check whether
1034 it should be cached. If it should not be cached, or if the dentry
1035 is not hashed, it is deleted. Otherwise cached dentries are put
1036 into an LRU list to be reclaimed on memory shortage.
1037
1038 d_drop: this unhashes a dentry from its parents hash list. A
1039 subsequent call to dput() will deallocate the dentry if its
1040 usage count drops to 0
1041
1042 d_delete: delete a dentry. If there are no other open references to
1043 the dentry then the dentry is turned into a negative dentry
1044 (the d_iput() method is called). If there are other
1045 references, then d_drop() is called instead
1046
1047 d_add: add a dentry to its parents hash list and then calls
1048 d_instantiate()
1049
1050 d_instantiate: add a dentry to the alias hash list for the inode and
1051 updates the "d_inode" member. The "i_count" member in the
1052 inode structure should be set/incremented. If the inode
1053 pointer is NULL, the dentry is called a "negative
1054 dentry". This function is commonly called when an inode is
1055 created for an existing negative dentry
1056
1057 d_lookup: look up a dentry given its parent and path name component
1058 It looks up the child of that given name from the dcache
1059 hash table. If it is found, the reference count is incremented
1060 and the dentry is returned. The caller must use dput()
1061 to free the dentry when it finishes using it.
1062
1063Mount Options
1064=============
1065
1066Parsing options
1067---------------
1068
1069On mount and remount the filesystem is passed a string containing a
1070comma separated list of mount options. The options can have either of
1071these forms:
1072
1073 option
1074 option=value
1075
1076The <linux/parser.h> header defines an API that helps parse these
1077options. There are plenty of examples on how to use it in existing
1078filesystems.
1079
1080Showing options
1081---------------
1082
1083If a filesystem accepts mount options, it must define show_options()
1084to show all the currently active options. The rules are:
1085
1086 - options MUST be shown which are not default or their values differ
1087 from the default
1088
1089 - options MAY be shown which are enabled by default or have their
1090 default value
1091
1092Options used only internally between a mount helper and the kernel
1093(such as file descriptors), or which only have an effect during the
1094mounting (such as ones controlling the creation of a journal) are exempt
1095from the above rules.
1096
1097The underlying reason for the above rules is to make sure, that a
1098mount can be accurately replicated (e.g. umounting and mounting again)
1099based on the information found in /proc/mounts.
1100
1101A simple method of saving options at mount/remount time and showing
1102them is provided with the save_mount_options() and
1103generic_show_options() helper functions. Please note, that using
1104these may have drawbacks. For more info see header comments for these
1105functions in fs/namespace.c.
1106
1107Resources
1108=========
1109
1110(Note some of these resources are not up-to-date with the latest kernel
1111 version.)
1112
1113Creating Linux virtual filesystems. 2002
1114 <http://lwn.net/Articles/13325/>
1115
1116The Linux Virtual File-system Layer by Neil Brown. 1999
1117 <http://www.cse.unsw.edu.au/~neilb/oss/linux-commentary/vfs.html>
1118
1119A tour of the Linux VFS by Michael K. Johnson. 1996
1120 <http://www.tldp.org/LDP/khg/HyperNews/get/fs/vfstour.html>
1121
1122A small trail through the Linux kernel by Andries Brouwer. 2001
1123 <http://www.win.tue.nl/~aeb/linux/vfs/trail.html>