2 Overview of the Linux Virtual File System
4 Original author: Richard Gooch <rgooch@atnf.csiro.au>
6 Last updated on June 24, 2007.
8 Copyright (C) 1999 Richard Gooch
9 Copyright (C) 2005 Pekka Enberg
11 This file is released under the GPLv2.
17 The Virtual File System (also known as the Virtual Filesystem Switch)
18 is the software layer in the kernel that provides the filesystem
19 interface to userspace programs. It also provides an abstraction
20 within the kernel which allows different filesystem implementations to
23 VFS system calls open(2), stat(2), read(2), write(2), chmod(2) and so
24 on are called from a process context. Filesystem locking is described
25 in the document Documentation/filesystems/Locking.
28 Directory Entry Cache (dcache)
29 ------------------------------
31 The VFS implements the open(2), stat(2), chmod(2), and similar system
32 calls. The pathname argument that is passed to them is used by the VFS
33 to search through the directory entry cache (also known as the dentry
34 cache or dcache). This provides a very fast look-up mechanism to
35 translate a pathname (filename) into a specific dentry. Dentries live
36 in RAM and are never saved to disc: they exist only for performance.
38 The dentry cache is meant to be a view into your entire filespace. As
39 most computers cannot fit all dentries in the RAM at the same time,
40 some bits of the cache are missing. In order to resolve your pathname
41 into a dentry, the VFS may have to resort to creating dentries along
42 the way, and then loading the inode. This is done by looking up the
49 An individual dentry usually has a pointer to an inode. Inodes are
50 filesystem objects such as regular files, directories, FIFOs and other
51 beasts. They live either on the disc (for block device filesystems)
52 or in the memory (for pseudo filesystems). Inodes that live on the
53 disc are copied into the memory when required and changes to the inode
54 are written back to disc. A single inode can be pointed to by multiple
55 dentries (hard links, for example, do this).
57 To look up an inode requires that the VFS calls the lookup() method of
58 the parent directory inode. This method is installed by the specific
59 filesystem implementation that the inode lives in. Once the VFS has
60 the required dentry (and hence the inode), we can do all those boring
61 things like open(2) the file, or stat(2) it to peek at the inode
62 data. The stat(2) operation is fairly simple: once the VFS has the
63 dentry, it peeks at the inode data and passes some of it back to
70 Opening a file requires another operation: allocation of a file
71 structure (this is the kernel-side implementation of file
72 descriptors). The freshly allocated file structure is initialized with
73 a pointer to the dentry and a set of file operation member functions.
74 These are taken from the inode data. The open() file method is then
75 called so the specific filesystem implementation can do its work. You
76 can see that this is another switch performed by the VFS. The file
77 structure is placed into the file descriptor table for the process.
79 Reading, writing and closing files (and other assorted VFS operations)
80 is done by using the userspace file descriptor to grab the appropriate
81 file structure, and then calling the required file structure method to
82 do whatever is required. For as long as the file is open, it keeps the
83 dentry in use, which in turn means that the VFS inode is still in use.
86 Registering and Mounting a Filesystem
87 =====================================
89 To register and unregister a filesystem, use the following API
94 extern int register_filesystem(struct file_system_type *);
95 extern int unregister_filesystem(struct file_system_type *);
97 The passed struct file_system_type describes your filesystem. When a
98 request is made to mount a filesystem onto a directory in your namespace,
99 the VFS will call the appropriate mount() method for the specific
100 filesystem. New vfsmount referring to the tree returned by ->mount()
101 will be attached to the mountpoint, so that when pathname resolution
102 reaches the mountpoint it will jump into the root of that vfsmount.
104 You can see all filesystems that are registered to the kernel in the
105 file /proc/filesystems.
108 struct file_system_type
109 -----------------------
111 This describes the filesystem. As of kernel 2.6.39, the following
114 struct file_system_type {
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;
127 name: the name of the filesystem type, such as "ext2", "iso9660",
130 fs_flags: various flags (i.e. FS_REQUIRES_DEV, FS_NO_DCACHE, etc.)
132 mount: the method to call when a new instance of this
133 filesystem should be mounted
135 kill_sb: the method to call when an instance of this filesystem
138 owner: for internal VFS use: you should initialize this to THIS_MODULE in
141 next: for internal VFS use: you should initialize this to NULL
143 s_lock_key, s_umount_key: lockdep-specific
145 The mount() method has the following arguments:
147 struct file_system_type *fs_type: describes the filesystem, partly initialized
148 by the specific filesystem code
150 int flags: mount flags
152 const char *dev_name: the device name we are mounting.
154 void *data: arbitrary mount options, usually comes as an ASCII
155 string (see "Mount Options" section)
157 The mount() method must return the root dentry of the tree requested by
158 caller. An active reference to its superblock must be grabbed and the
159 superblock must be locked. On failure it should return ERR_PTR(error).
161 The arguments match those of mount(2) and their interpretation
162 depends on filesystem type. E.g. for block filesystems, dev_name is
163 interpreted as block device name, that device is opened and if it
164 contains a suitable filesystem image the method creates and initializes
165 struct super_block accordingly, returning its root dentry to caller.
167 ->mount() may choose to return a subtree of existing filesystem - it
168 doesn't have to create a new one. The main result from the caller's
169 point of view is a reference to dentry at the root of (sub)tree to
170 be attached; creation of new superblock is a common side effect.
172 The most interesting member of the superblock structure that the
173 mount() method fills in is the "s_op" field. This is a pointer to
174 a "struct super_operations" which describes the next level of the
175 filesystem implementation.
177 Usually, a filesystem uses one of the generic mount() implementations
178 and provides a fill_super() callback instead. The generic variants are:
180 mount_bdev: mount a filesystem residing on a block device
182 mount_nodev: mount a filesystem that is not backed by a device
184 mount_single: mount a filesystem which shares the instance between
187 A fill_super() callback implementation has the following arguments:
189 struct super_block *sb: the superblock structure. The callback
190 must initialize this properly.
192 void *data: arbitrary mount options, usually comes as an ASCII
193 string (see "Mount Options" section)
195 int silent: whether or not to be silent on error
198 The Superblock Object
199 =====================
201 A superblock object represents a mounted filesystem.
204 struct super_operations
205 -----------------------
207 This describes how the VFS can manipulate the superblock of your
208 filesystem. As of kernel 2.6.22, the following members are defined:
210 struct super_operations {
211 struct inode *(*alloc_inode)(struct super_block *sb);
212 void (*destroy_inode)(struct inode *);
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 int (*sync_fs)(struct super_block *sb, int wait);
220 int (*freeze_fs) (struct super_block *);
221 int (*unfreeze_fs) (struct super_block *);
222 int (*statfs) (struct dentry *, struct kstatfs *);
223 int (*remount_fs) (struct super_block *, int *, char *);
224 void (*clear_inode) (struct inode *);
225 void (*umount_begin) (struct super_block *);
227 int (*show_options)(struct seq_file *, struct dentry *);
229 ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
230 ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
231 int (*nr_cached_objects)(struct super_block *);
232 void (*free_cached_objects)(struct super_block *, int);
235 All methods are called without any locks being held, unless otherwise
236 noted. This means that most methods can block safely. All methods are
237 only called from a process context (i.e. not from an interrupt handler
240 alloc_inode: this method is called by alloc_inode() to allocate memory
241 for struct inode and initialize it. If this function is not
242 defined, a simple 'struct inode' is allocated. Normally
243 alloc_inode will be used to allocate a larger structure which
244 contains a 'struct inode' embedded within it.
246 destroy_inode: this method is called by destroy_inode() to release
247 resources allocated for struct inode. It is only required if
248 ->alloc_inode was defined and simply undoes anything done by
251 dirty_inode: this method is called by the VFS to mark an inode dirty.
253 write_inode: this method is called when the VFS needs to write an
254 inode to disc. The second parameter indicates whether the write
255 should be synchronous or not, not all filesystems check this flag.
257 drop_inode: called when the last access to the inode is dropped,
258 with the inode->i_lock spinlock held.
260 This method should be either NULL (normal UNIX filesystem
261 semantics) or "generic_delete_inode" (for filesystems that do not
262 want to cache inodes - causing "delete_inode" to always be
263 called regardless of the value of i_nlink)
265 The "generic_delete_inode()" behavior is equivalent to the
266 old practice of using "force_delete" in the put_inode() case,
267 but does not have the races that the "force_delete()" approach
270 delete_inode: called when the VFS wants to delete an inode
272 put_super: called when the VFS wishes to free the superblock
273 (i.e. unmount). This is called with the superblock lock held
275 sync_fs: called when VFS is writing out all dirty data associated with
276 a superblock. The second parameter indicates whether the method
277 should wait until the write out has been completed. Optional.
279 freeze_fs: called when VFS is locking a filesystem and
280 forcing it into a consistent state. This method is currently
281 used by the Logical Volume Manager (LVM).
283 unfreeze_fs: called when VFS is unlocking a filesystem and making it writable
286 statfs: called when the VFS needs to get filesystem statistics.
288 remount_fs: called when the filesystem is remounted. This is called
289 with the kernel lock held
291 clear_inode: called then the VFS clears the inode. Optional
293 umount_begin: called when the VFS is unmounting a filesystem.
295 show_options: called by the VFS to show mount options for
296 /proc/<pid>/mounts. (see "Mount Options" section)
298 quota_read: called by the VFS to read from filesystem quota file.
300 quota_write: called by the VFS to write to filesystem quota file.
302 nr_cached_objects: called by the sb cache shrinking function for the
303 filesystem to return the number of freeable cached objects it contains.
306 free_cache_objects: called by the sb cache shrinking function for the
307 filesystem to scan the number of objects indicated to try to free them.
308 Optional, but any filesystem implementing this method needs to also
309 implement ->nr_cached_objects for it to be called correctly.
311 We can't do anything with any errors that the filesystem might
312 encountered, hence the void return type. This will never be called if
313 the VM is trying to reclaim under GFP_NOFS conditions, hence this
314 method does not need to handle that situation itself.
316 Implementations must include conditional reschedule calls inside any
317 scanning loop that is done. This allows the VFS to determine
318 appropriate scan batch sizes without having to worry about whether
319 implementations will cause holdoff problems due to large scan batch
322 Whoever sets up the inode is responsible for filling in the "i_op" field. This
323 is a pointer to a "struct inode_operations" which describes the methods that
324 can be performed on individual inodes.
326 struct xattr_handlers
327 ---------------------
329 On filesystems that support extended attributes (xattrs), the s_xattr
330 superblock field points to a NULL-terminated array of xattr handlers. Extended
331 attributes are name:value pairs.
333 name: Indicates that the handler matches attributes with the specified name
334 (such as "system.posix_acl_access"); the prefix field must be NULL.
336 prefix: Indicates that the handler matches all attributes with the specified
337 name prefix (such as "user."); the name field must be NULL.
339 list: Determine if attributes matching this xattr handler should be listed
340 for a particular dentry. Used by some listxattr implementations like
343 get: Called by the VFS to get the value of a particular extended attribute.
344 This method is called by the getxattr(2) system call.
346 set: Called by the VFS to set the value of a particular extended attribute.
347 When the new value is NULL, called to remove a particular extended
348 attribute. This method is called by the the setxattr(2) and
349 removexattr(2) system calls.
351 When none of the xattr handlers of a filesystem match the specified attribute
352 name or when a filesystem doesn't support extended attributes, the various
353 *xattr(2) system calls return -EOPNOTSUPP.
359 An inode object represents an object within the filesystem.
362 struct inode_operations
363 -----------------------
365 This describes how the VFS can manipulate an inode in your
366 filesystem. As of kernel 2.6.22, the following members are defined:
368 struct inode_operations {
369 int (*create) (struct inode *,struct dentry *, umode_t, bool);
370 struct dentry * (*lookup) (struct inode *,struct dentry *, unsigned int);
371 int (*link) (struct dentry *,struct inode *,struct dentry *);
372 int (*unlink) (struct inode *,struct dentry *);
373 int (*symlink) (struct inode *,struct dentry *,const char *);
374 int (*mkdir) (struct inode *,struct dentry *,umode_t);
375 int (*rmdir) (struct inode *,struct dentry *);
376 int (*mknod) (struct inode *,struct dentry *,umode_t,dev_t);
377 int (*rename) (struct inode *, struct dentry *,
378 struct inode *, struct dentry *, unsigned int);
379 int (*readlink) (struct dentry *, char __user *,int);
380 const char *(*get_link) (struct dentry *, struct inode *,
381 struct delayed_call *);
382 int (*permission) (struct inode *, int);
383 int (*get_acl)(struct inode *, int);
384 int (*setattr) (struct dentry *, struct iattr *);
385 int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
386 ssize_t (*listxattr) (struct dentry *, char *, size_t);
387 void (*update_time)(struct inode *, struct timespec *, int);
388 int (*atomic_open)(struct inode *, struct dentry *, struct file *,
389 unsigned open_flag, umode_t create_mode, int *opened);
390 int (*tmpfile) (struct inode *, struct dentry *, umode_t);
393 Again, all methods are called without any locks being held, unless
396 create: called by the open(2) and creat(2) system calls. Only
397 required if you want to support regular files. The dentry you
398 get should not have an inode (i.e. it should be a negative
399 dentry). Here you will probably call d_instantiate() with the
400 dentry and the newly created inode
402 lookup: called when the VFS needs to look up an inode in a parent
403 directory. The name to look for is found in the dentry. This
404 method must call d_add() to insert the found inode into the
405 dentry. The "i_count" field in the inode structure should be
406 incremented. If the named inode does not exist a NULL inode
407 should be inserted into the dentry (this is called a negative
408 dentry). Returning an error code from this routine must only
409 be done on a real error, otherwise creating inodes with system
410 calls like create(2), mknod(2), mkdir(2) and so on will fail.
411 If you wish to overload the dentry methods then you should
412 initialise the "d_dop" field in the dentry; this is a pointer
413 to a struct "dentry_operations".
414 This method is called with the directory inode semaphore held
416 link: called by the link(2) system call. Only required if you want
417 to support hard links. You will probably need to call
418 d_instantiate() just as you would in the create() method
420 unlink: called by the unlink(2) system call. Only required if you
421 want to support deleting inodes
423 symlink: called by the symlink(2) system call. Only required if you
424 want to support symlinks. You will probably need to call
425 d_instantiate() just as you would in the create() method
427 mkdir: called by the mkdir(2) system call. Only required if you want
428 to support creating subdirectories. You will probably need to
429 call d_instantiate() just as you would in the create() method
431 rmdir: called by the rmdir(2) system call. Only required if you want
432 to support deleting subdirectories
434 mknod: called by the mknod(2) system call to create a device (char,
435 block) inode or a named pipe (FIFO) or socket. Only required
436 if you want to support creating these types of inodes. You
437 will probably need to call d_instantiate() just as you would
438 in the create() method
440 rename: called by the rename(2) system call to rename the object to
441 have the parent and name given by the second inode and dentry.
443 The filesystem must return -EINVAL for any unsupported or
444 unknown flags. Currently the following flags are implemented:
445 (1) RENAME_NOREPLACE: this flag indicates that if the target
446 of the rename exists the rename should fail with -EEXIST
447 instead of replacing the target. The VFS already checks for
448 existence, so for local filesystems the RENAME_NOREPLACE
449 implementation is equivalent to plain rename.
450 (2) RENAME_EXCHANGE: exchange source and target. Both must
451 exist; this is checked by the VFS. Unlike plain rename,
452 source and target may be of different type.
454 readlink: called by the readlink(2) system call. Only required if
455 you want to support reading symbolic links
457 get_link: called by the VFS to follow a symbolic link to the
458 inode it points to. Only required if you want to support
459 symbolic links. This method returns the symlink body
460 to traverse (and possibly resets the current position with
461 nd_jump_link()). If the body won't go away until the inode
462 is gone, nothing else is needed; if it needs to be otherwise
463 pinned, arrange for its release by having get_link(..., ..., done)
464 do set_delayed_call(done, destructor, argument).
465 In that case destructor(argument) will be called once VFS is
466 done with the body you've returned.
467 May be called in RCU mode; that is indicated by NULL dentry
468 argument. If request can't be handled without leaving RCU mode,
469 have it return ERR_PTR(-ECHILD).
471 permission: called by the VFS to check for access rights on a POSIX-like
474 May be called in rcu-walk mode (mask & MAY_NOT_BLOCK). If in rcu-walk
475 mode, the filesystem must check the permission without blocking or
476 storing to the inode.
478 If a situation is encountered that rcu-walk cannot handle, return
479 -ECHILD and it will be called again in ref-walk mode.
481 setattr: called by the VFS to set attributes for a file. This method
482 is called by chmod(2) and related system calls.
484 getattr: called by the VFS to get attributes of a file. This method
485 is called by stat(2) and related system calls.
487 listxattr: called by the VFS to list all extended attributes for a
488 given file. This method is called by the listxattr(2) system call.
490 update_time: called by the VFS to update a specific time or the i_version of
491 an inode. If this is not defined the VFS will update the inode itself
492 and call mark_inode_dirty_sync.
494 atomic_open: called on the last component of an open. Using this optional
495 method the filesystem can look up, possibly create and open the file in
496 one atomic operation. If it cannot perform this (e.g. the file type
497 turned out to be wrong) it may signal this by returning 1 instead of
498 usual 0 or -ve . This method is only called if the last component is
499 negative or needs lookup. Cached positive dentries are still handled by
500 f_op->open(). If the file was created, the FILE_CREATED flag should be
501 set in "opened". In case of O_EXCL the method must only succeed if the
502 file didn't exist and hence FILE_CREATED shall always be set on success.
504 tmpfile: called in the end of O_TMPFILE open(). Optional, equivalent to
505 atomically creating, opening and unlinking a file in given directory.
507 The Address Space Object
508 ========================
510 The address space object is used to group and manage pages in the page
511 cache. It can be used to keep track of the pages in a file (or
512 anything else) and also track the mapping of sections of the file into
513 process address spaces.
515 There are a number of distinct yet related services that an
516 address-space can provide. These include communicating memory
517 pressure, page lookup by address, and keeping track of pages tagged as
520 The first can be used independently to the others. The VM can try to
521 either write dirty pages in order to clean them, or release clean
522 pages in order to reuse them. To do this it can call the ->writepage
523 method on dirty pages, and ->releasepage on clean pages with
524 PagePrivate set. Clean pages without PagePrivate and with no external
525 references will be released without notice being given to the
528 To achieve this functionality, pages need to be placed on an LRU with
529 lru_cache_add and mark_page_active needs to be called whenever the
532 Pages are normally kept in a radix tree index by ->index. This tree
533 maintains information about the PG_Dirty and PG_Writeback status of
534 each page, so that pages with either of these flags can be found
537 The Dirty tag is primarily used by mpage_writepages - the default
538 ->writepages method. It uses the tag to find dirty pages to call
539 ->writepage on. If mpage_writepages is not used (i.e. the address
540 provides its own ->writepages) , the PAGECACHE_TAG_DIRTY tag is
541 almost unused. write_inode_now and sync_inode do use it (through
542 __sync_single_inode) to check if ->writepages has been successful in
543 writing out the whole address_space.
545 The Writeback tag is used by filemap*wait* and sync_page* functions,
546 via filemap_fdatawait_range, to wait for all writeback to complete.
548 An address_space handler may attach extra information to a page,
549 typically using the 'private' field in the 'struct page'. If such
550 information is attached, the PG_Private flag should be set. This will
551 cause various VM routines to make extra calls into the address_space
552 handler to deal with that data.
554 An address space acts as an intermediate between storage and
555 application. Data is read into the address space a whole page at a
556 time, and provided to the application either by copying of the page,
557 or by memory-mapping the page.
558 Data is written into the address space by the application, and then
559 written-back to storage typically in whole pages, however the
560 address_space has finer control of write sizes.
562 The read process essentially only requires 'readpage'. The write
563 process is more complicated and uses write_begin/write_end or
564 set_page_dirty to write data into the address_space, and writepage
565 and writepages to writeback data to storage.
567 Adding and removing pages to/from an address_space is protected by the
570 When data is written to a page, the PG_Dirty flag should be set. It
571 typically remains set until writepage asks for it to be written. This
572 should clear PG_Dirty and set PG_Writeback. It can be actually
573 written at any point after PG_Dirty is clear. Once it is known to be
574 safe, PG_Writeback is cleared.
576 Writeback makes use of a writeback_control structure...
578 struct address_space_operations
579 -------------------------------
581 This describes how the VFS can manipulate mapping of a file to page cache in
582 your filesystem. The following members are defined:
584 struct address_space_operations {
585 int (*writepage)(struct page *page, struct writeback_control *wbc);
586 int (*readpage)(struct file *, struct page *);
587 int (*writepages)(struct address_space *, struct writeback_control *);
588 int (*set_page_dirty)(struct page *page);
589 int (*readpages)(struct file *filp, struct address_space *mapping,
590 struct list_head *pages, unsigned nr_pages);
591 int (*write_begin)(struct file *, struct address_space *mapping,
592 loff_t pos, unsigned len, unsigned flags,
593 struct page **pagep, void **fsdata);
594 int (*write_end)(struct file *, struct address_space *mapping,
595 loff_t pos, unsigned len, unsigned copied,
596 struct page *page, void *fsdata);
597 sector_t (*bmap)(struct address_space *, sector_t);
598 void (*invalidatepage) (struct page *, unsigned int, unsigned int);
599 int (*releasepage) (struct page *, int);
600 void (*freepage)(struct page *);
601 ssize_t (*direct_IO)(struct kiocb *, struct iov_iter *iter);
602 /* isolate a page for migration */
603 bool (*isolate_page) (struct page *, isolate_mode_t);
604 /* migrate the contents of a page to the specified target */
605 int (*migratepage) (struct page *, struct page *);
606 /* put migration-failed page back to right list */
607 void (*putback_page) (struct page *);
608 int (*launder_page) (struct page *);
610 int (*is_partially_uptodate) (struct page *, unsigned long,
612 void (*is_dirty_writeback) (struct page *, bool *, bool *);
613 int (*error_remove_page) (struct mapping *mapping, struct page *page);
614 int (*swap_activate)(struct file *);
615 int (*swap_deactivate)(struct file *);
618 writepage: called by the VM to write a dirty page to backing store.
619 This may happen for data integrity reasons (i.e. 'sync'), or
620 to free up memory (flush). The difference can be seen in
622 The PG_Dirty flag has been cleared and PageLocked is true.
623 writepage should start writeout, should set PG_Writeback,
624 and should make sure the page is unlocked, either synchronously
625 or asynchronously when the write operation completes.
627 If wbc->sync_mode is WB_SYNC_NONE, ->writepage doesn't have to
628 try too hard if there are problems, and may choose to write out
629 other pages from the mapping if that is easier (e.g. due to
630 internal dependencies). If it chooses not to start writeout, it
631 should return AOP_WRITEPAGE_ACTIVATE so that the VM will not keep
632 calling ->writepage on that page.
634 See the file "Locking" for more details.
636 readpage: called by the VM to read a page from backing store.
637 The page will be Locked when readpage is called, and should be
638 unlocked and marked uptodate once the read completes.
639 If ->readpage discovers that it needs to unlock the page for
640 some reason, it can do so, and then return AOP_TRUNCATED_PAGE.
641 In this case, the page will be relocated, relocked and if
642 that all succeeds, ->readpage will be called again.
644 writepages: called by the VM to write out pages associated with the
645 address_space object. If wbc->sync_mode is WBC_SYNC_ALL, then
646 the writeback_control will specify a range of pages that must be
647 written out. If it is WBC_SYNC_NONE, then a nr_to_write is given
648 and that many pages should be written if possible.
649 If no ->writepages is given, then mpage_writepages is used
650 instead. This will choose pages from the address space that are
651 tagged as DIRTY and will pass them to ->writepage.
653 set_page_dirty: called by the VM to set a page dirty.
654 This is particularly needed if an address space attaches
655 private data to a page, and that data needs to be updated when
656 a page is dirtied. This is called, for example, when a memory
657 mapped page gets modified.
658 If defined, it should set the PageDirty flag, and the
659 PAGECACHE_TAG_DIRTY tag in the radix tree.
661 readpages: called by the VM to read pages associated with the address_space
662 object. This is essentially just a vector version of
663 readpage. Instead of just one page, several pages are
665 readpages is only used for read-ahead, so read errors are
666 ignored. If anything goes wrong, feel free to give up.
669 Called by the generic buffered write code to ask the filesystem to
670 prepare to write len bytes at the given offset in the file. The
671 address_space should check that the write will be able to complete,
672 by allocating space if necessary and doing any other internal
673 housekeeping. If the write will update parts of any basic-blocks on
674 storage, then those blocks should be pre-read (if they haven't been
675 read already) so that the updated blocks can be written out properly.
677 The filesystem must return the locked pagecache page for the specified
678 offset, in *pagep, for the caller to write into.
680 It must be able to cope with short writes (where the length passed to
681 write_begin is greater than the number of bytes copied into the page).
683 flags is a field for AOP_FLAG_xxx flags, described in
686 A void * may be returned in fsdata, which then gets passed into
689 Returns 0 on success; < 0 on failure (which is the error code), in
690 which case write_end is not called.
692 write_end: After a successful write_begin, and data copy, write_end must
693 be called. len is the original len passed to write_begin, and copied
694 is the amount that was able to be copied (copied == len is always true
695 if write_begin was called with the AOP_FLAG_UNINTERRUPTIBLE flag).
697 The filesystem must take care of unlocking the page and releasing it
698 refcount, and updating i_size.
700 Returns < 0 on failure, otherwise the number of bytes (<= 'copied')
701 that were able to be copied into pagecache.
703 bmap: called by the VFS to map a logical block offset within object to
704 physical block number. This method is used by the FIBMAP
705 ioctl and for working with swap-files. To be able to swap to
706 a file, the file must have a stable mapping to a block
707 device. The swap system does not go through the filesystem
708 but instead uses bmap to find out where the blocks in the file
709 are and uses those addresses directly.
711 invalidatepage: If a page has PagePrivate set, then invalidatepage
712 will be called when part or all of the page is to be removed
713 from the address space. This generally corresponds to either a
714 truncation, punch hole or a complete invalidation of the address
715 space (in the latter case 'offset' will always be 0 and 'length'
716 will be PAGE_SIZE). Any private data associated with the page
717 should be updated to reflect this truncation. If offset is 0 and
718 length is PAGE_SIZE, then the private data should be released,
719 because the page must be able to be completely discarded. This may
720 be done by calling the ->releasepage function, but in this case the
721 release MUST succeed.
723 releasepage: releasepage is called on PagePrivate pages to indicate
724 that the page should be freed if possible. ->releasepage
725 should remove any private data from the page and clear the
726 PagePrivate flag. If releasepage() fails for some reason, it must
727 indicate failure with a 0 return value.
728 releasepage() is used in two distinct though related cases. The
729 first is when the VM finds a clean page with no active users and
730 wants to make it a free page. If ->releasepage succeeds, the
731 page will be removed from the address_space and become free.
733 The second case is when a request has been made to invalidate
734 some or all pages in an address_space. This can happen
735 through the fadvise(POSIX_FADV_DONTNEED) system call or by the
736 filesystem explicitly requesting it as nfs and 9fs do (when
737 they believe the cache may be out of date with storage) by
738 calling invalidate_inode_pages2().
739 If the filesystem makes such a call, and needs to be certain
740 that all pages are invalidated, then its releasepage will
741 need to ensure this. Possibly it can clear the PageUptodate
742 bit if it cannot free private data yet.
744 freepage: freepage is called once the page is no longer visible in
745 the page cache in order to allow the cleanup of any private
746 data. Since it may be called by the memory reclaimer, it
747 should not assume that the original address_space mapping still
748 exists, and it should not block.
750 direct_IO: called by the generic read/write routines to perform
751 direct_IO - that is IO requests which bypass the page cache
752 and transfer data directly between the storage and the
753 application's address space.
755 isolate_page: Called by the VM when isolating a movable non-lru page.
756 If page is successfully isolated, VM marks the page as PG_isolated
757 via __SetPageIsolated.
759 migrate_page: This is used to compact the physical memory usage.
760 If the VM wants to relocate a page (maybe off a memory card
761 that is signalling imminent failure) it will pass a new page
762 and an old page to this function. migrate_page should
763 transfer any private data across and update any references
764 that it has to the page.
766 putback_page: Called by the VM when isolated page's migration fails.
768 launder_page: Called before freeing a page - it writes back the dirty page. To
769 prevent redirtying the page, it is kept locked during the whole
772 is_partially_uptodate: Called by the VM when reading a file through the
773 pagecache when the underlying blocksize != pagesize. If the required
774 block is up to date then the read can complete without needing the IO
775 to bring the whole page up to date.
777 is_dirty_writeback: Called by the VM when attempting to reclaim a page.
778 The VM uses dirty and writeback information to determine if it needs
779 to stall to allow flushers a chance to complete some IO. Ordinarily
780 it can use PageDirty and PageWriteback but some filesystems have
781 more complex state (unstable pages in NFS prevent reclaim) or
782 do not set those flags due to locking problems. This callback
783 allows a filesystem to indicate to the VM if a page should be
784 treated as dirty or writeback for the purposes of stalling.
786 error_remove_page: normally set to generic_error_remove_page if truncation
787 is ok for this address space. Used for memory failure handling.
788 Setting this implies you deal with pages going away under you,
789 unless you have them locked or reference counts increased.
791 swap_activate: Called when swapon is used on a file to allocate
792 space if necessary and pin the block lookup information in
793 memory. A return value of zero indicates success,
794 in which case this file can be used to back swapspace. The
795 swapspace operations will be proxied to this address space's
796 ->swap_{out,in} methods.
798 swap_deactivate: Called during swapoff on files where swap_activate
805 A file object represents a file opened by a process.
808 struct file_operations
809 ----------------------
811 This describes how the VFS can manipulate an open file. As of kernel
812 4.1, the following members are defined:
814 struct file_operations {
815 struct module *owner;
816 loff_t (*llseek) (struct file *, loff_t, int);
817 ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
818 ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
819 ssize_t (*read_iter) (struct kiocb *, struct iov_iter *);
820 ssize_t (*write_iter) (struct kiocb *, struct iov_iter *);
821 int (*iterate) (struct file *, struct dir_context *);
822 unsigned int (*poll) (struct file *, struct poll_table_struct *);
823 long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
824 long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
825 int (*mmap) (struct file *, struct vm_area_struct *);
826 int (*mremap)(struct file *, struct vm_area_struct *);
827 int (*open) (struct inode *, struct file *);
828 int (*flush) (struct file *, fl_owner_t id);
829 int (*release) (struct inode *, struct file *);
830 int (*fsync) (struct file *, loff_t, loff_t, int datasync);
831 int (*aio_fsync) (struct kiocb *, int datasync);
832 int (*fasync) (int, struct file *, int);
833 int (*lock) (struct file *, int, struct file_lock *);
834 ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
835 unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
836 int (*check_flags)(int);
837 int (*flock) (struct file *, int, struct file_lock *);
838 ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int);
839 ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int);
840 int (*setlease)(struct file *, long, struct file_lock **, void **);
841 long (*fallocate)(struct file *file, int mode, loff_t offset,
843 void (*show_fdinfo)(struct seq_file *m, struct file *f);
845 unsigned (*mmap_capabilities)(struct file *);
849 Again, all methods are called without any locks being held, unless
852 llseek: called when the VFS needs to move the file position index
854 read: called by read(2) and related system calls
856 read_iter: possibly asynchronous read with iov_iter as destination
858 write: called by write(2) and related system calls
860 write_iter: possibly asynchronous write with iov_iter as source
862 iterate: called when the VFS needs to read the directory contents
864 poll: called by the VFS when a process wants to check if there is
865 activity on this file and (optionally) go to sleep until there
866 is activity. Called by the select(2) and poll(2) system calls
868 unlocked_ioctl: called by the ioctl(2) system call.
870 compat_ioctl: called by the ioctl(2) system call when 32 bit system calls
871 are used on 64 bit kernels.
873 mmap: called by the mmap(2) system call
875 open: called by the VFS when an inode should be opened. When the VFS
876 opens a file, it creates a new "struct file". It then calls the
877 open method for the newly allocated file structure. You might
878 think that the open method really belongs in
879 "struct inode_operations", and you may be right. I think it's
880 done the way it is because it makes filesystems simpler to
881 implement. The open() method is a good place to initialize the
882 "private_data" member in the file structure if you want to point
883 to a device structure
885 flush: called by the close(2) system call to flush a file
887 release: called when the last reference to an open file is closed
889 fsync: called by the fsync(2) system call
891 fasync: called by the fcntl(2) system call when asynchronous
892 (non-blocking) mode is enabled for a file
894 lock: called by the fcntl(2) system call for F_GETLK, F_SETLK, and F_SETLKW
897 get_unmapped_area: called by the mmap(2) system call
899 check_flags: called by the fcntl(2) system call for F_SETFL command
901 flock: called by the flock(2) system call
903 splice_write: called by the VFS to splice data from a pipe to a file. This
904 method is used by the splice(2) system call
906 splice_read: called by the VFS to splice data from file to a pipe. This
907 method is used by the splice(2) system call
909 setlease: called by the VFS to set or release a file lock lease. setlease
910 implementations should call generic_setlease to record or remove
911 the lease in the inode after setting it.
913 fallocate: called by the VFS to preallocate blocks or punch a hole.
915 Note that the file operations are implemented by the specific
916 filesystem in which the inode resides. When opening a device node
917 (character or block special) most filesystems will call special
918 support routines in the VFS which will locate the required device
919 driver information. These support routines replace the filesystem file
920 operations with those for the device driver, and then proceed to call
921 the new open() method for the file. This is how opening a device file
922 in the filesystem eventually ends up calling the device driver open()
926 Directory Entry Cache (dcache)
927 ==============================
930 struct dentry_operations
931 ------------------------
933 This describes how a filesystem can overload the standard dentry
934 operations. Dentries and the dcache are the domain of the VFS and the
935 individual filesystem implementations. Device drivers have no business
936 here. These methods may be set to NULL, as they are either optional or
937 the VFS uses a default. As of kernel 2.6.22, the following members are
940 struct dentry_operations {
941 int (*d_revalidate)(struct dentry *, unsigned int);
942 int (*d_weak_revalidate)(struct dentry *, unsigned int);
943 int (*d_hash)(const struct dentry *, struct qstr *);
944 int (*d_compare)(const struct dentry *,
945 unsigned int, const char *, const struct qstr *);
946 int (*d_delete)(const struct dentry *);
947 int (*d_init)(struct dentry *);
948 void (*d_release)(struct dentry *);
949 void (*d_iput)(struct dentry *, struct inode *);
950 char *(*d_dname)(struct dentry *, char *, int);
951 struct vfsmount *(*d_automount)(struct path *);
952 int (*d_manage)(struct dentry *, bool);
953 struct dentry *(*d_real)(struct dentry *, const struct inode *,
957 d_revalidate: called when the VFS needs to revalidate a dentry. This
958 is called whenever a name look-up finds a dentry in the
959 dcache. Most local filesystems leave this as NULL, because all their
960 dentries in the dcache are valid. Network filesystems are different
961 since things can change on the server without the client necessarily
964 This function should return a positive value if the dentry is still
965 valid, and zero or a negative error code if it isn't.
967 d_revalidate may be called in rcu-walk mode (flags & LOOKUP_RCU).
968 If in rcu-walk mode, the filesystem must revalidate the dentry without
969 blocking or storing to the dentry, d_parent and d_inode should not be
970 used without care (because they can change and, in d_inode case, even
971 become NULL under us).
973 If a situation is encountered that rcu-walk cannot handle, return
974 -ECHILD and it will be called again in ref-walk mode.
976 d_weak_revalidate: called when the VFS needs to revalidate a "jumped" dentry.
977 This is called when a path-walk ends at dentry that was not acquired by
978 doing a lookup in the parent directory. This includes "/", "." and "..",
979 as well as procfs-style symlinks and mountpoint traversal.
981 In this case, we are less concerned with whether the dentry is still
982 fully correct, but rather that the inode is still valid. As with
983 d_revalidate, most local filesystems will set this to NULL since their
984 dcache entries are always valid.
986 This function has the same return code semantics as d_revalidate.
988 d_weak_revalidate is only called after leaving rcu-walk mode.
990 d_hash: called when the VFS adds a dentry to the hash table. The first
991 dentry passed to d_hash is the parent directory that the name is
994 Same locking and synchronisation rules as d_compare regarding
995 what is safe to dereference etc.
997 d_compare: called to compare a dentry name with a given name. The first
998 dentry is the parent of the dentry to be compared, the second is
999 the child dentry. len and name string are properties of the dentry
1000 to be compared. qstr is the name to compare it with.
1002 Must be constant and idempotent, and should not take locks if
1003 possible, and should not or store into the dentry.
1004 Should not dereference pointers outside the dentry without
1005 lots of care (eg. d_parent, d_inode, d_name should not be used).
1007 However, our vfsmount is pinned, and RCU held, so the dentries and
1008 inodes won't disappear, neither will our sb or filesystem module.
1011 It is a tricky calling convention because it needs to be called under
1012 "rcu-walk", ie. without any locks or references on things.
1014 d_delete: called when the last reference to a dentry is dropped and the
1015 dcache is deciding whether or not to cache it. Return 1 to delete
1016 immediately, or 0 to cache the dentry. Default is NULL which means to
1017 always cache a reachable dentry. d_delete must be constant and
1020 d_init: called when a dentry is allocated
1022 d_release: called when a dentry is really deallocated
1024 d_iput: called when a dentry loses its inode (just prior to its
1025 being deallocated). The default when this is NULL is that the
1026 VFS calls iput(). If you define this method, you must call
1029 d_dname: called when the pathname of a dentry should be generated.
1030 Useful for some pseudo filesystems (sockfs, pipefs, ...) to delay
1031 pathname generation. (Instead of doing it when dentry is created,
1032 it's done only when the path is needed.). Real filesystems probably
1033 dont want to use it, because their dentries are present in global
1034 dcache hash, so their hash should be an invariant. As no lock is
1035 held, d_dname() should not try to modify the dentry itself, unless
1036 appropriate SMP safety is used. CAUTION : d_path() logic is quite
1037 tricky. The correct way to return for example "Hello" is to put it
1038 at the end of the buffer, and returns a pointer to the first char.
1039 dynamic_dname() helper function is provided to take care of this.
1043 static char *pipefs_dname(struct dentry *dent, char *buffer, int buflen)
1045 return dynamic_dname(dentry, buffer, buflen, "pipe:[%lu]",
1046 dentry->d_inode->i_ino);
1049 d_automount: called when an automount dentry is to be traversed (optional).
1050 This should create a new VFS mount record and return the record to the
1051 caller. The caller is supplied with a path parameter giving the
1052 automount directory to describe the automount target and the parent
1053 VFS mount record to provide inheritable mount parameters. NULL should
1054 be returned if someone else managed to make the automount first. If
1055 the vfsmount creation failed, then an error code should be returned.
1056 If -EISDIR is returned, then the directory will be treated as an
1057 ordinary directory and returned to pathwalk to continue walking.
1059 If a vfsmount is returned, the caller will attempt to mount it on the
1060 mountpoint and will remove the vfsmount from its expiration list in
1061 the case of failure. The vfsmount should be returned with 2 refs on
1062 it to prevent automatic expiration - the caller will clean up the
1065 This function is only used if DCACHE_NEED_AUTOMOUNT is set on the
1066 dentry. This is set by __d_instantiate() if S_AUTOMOUNT is set on the
1069 d_manage: called to allow the filesystem to manage the transition from a
1070 dentry (optional). This allows autofs, for example, to hold up clients
1071 waiting to explore behind a 'mountpoint' whilst letting the daemon go
1072 past and construct the subtree there. 0 should be returned to let the
1073 calling process continue. -EISDIR can be returned to tell pathwalk to
1074 use this directory as an ordinary directory and to ignore anything
1075 mounted on it and not to check the automount flag. Any other error
1076 code will abort pathwalk completely.
1078 If the 'rcu_walk' parameter is true, then the caller is doing a
1079 pathwalk in RCU-walk mode. Sleeping is not permitted in this mode,
1080 and the caller can be asked to leave it and call again by returning
1081 -ECHILD. -EISDIR may also be returned to tell pathwalk to
1082 ignore d_automount or any mounts.
1084 This function is only used if DCACHE_MANAGE_TRANSIT is set on the
1085 dentry being transited from.
1087 d_real: overlay/union type filesystems implement this method to return one of
1088 the underlying dentries hidden by the overlay. It is used in three
1091 Called from open it may need to copy-up the file depending on the
1092 supplied open flags. This mode is selected with a non-zero flags
1093 argument. In this mode the d_real method can return an error.
1095 Called from file_dentry() it returns the real dentry matching the inode
1096 argument. The real dentry may be from a lower layer already copied up,
1097 but still referenced from the file. This mode is selected with a
1098 non-NULL inode argument. This will always succeed.
1100 With NULL inode and zero flags the topmost real underlying dentry is
1101 returned. This will always succeed.
1103 This method is never called with both non-NULL inode and non-zero flags.
1105 Each dentry has a pointer to its parent dentry, as well as a hash list
1106 of child dentries. Child dentries are basically like files in a
1110 Directory Entry Cache API
1111 --------------------------
1113 There are a number of functions defined which permit a filesystem to
1114 manipulate dentries:
1116 dget: open a new handle for an existing dentry (this just increments
1119 dput: close a handle for a dentry (decrements the usage count). If
1120 the usage count drops to 0, and the dentry is still in its
1121 parent's hash, the "d_delete" method is called to check whether
1122 it should be cached. If it should not be cached, or if the dentry
1123 is not hashed, it is deleted. Otherwise cached dentries are put
1124 into an LRU list to be reclaimed on memory shortage.
1126 d_drop: this unhashes a dentry from its parents hash list. A
1127 subsequent call to dput() will deallocate the dentry if its
1128 usage count drops to 0
1130 d_delete: delete a dentry. If there are no other open references to
1131 the dentry then the dentry is turned into a negative dentry
1132 (the d_iput() method is called). If there are other
1133 references, then d_drop() is called instead
1135 d_add: add a dentry to its parents hash list and then calls
1138 d_instantiate: add a dentry to the alias hash list for the inode and
1139 updates the "d_inode" member. The "i_count" member in the
1140 inode structure should be set/incremented. If the inode
1141 pointer is NULL, the dentry is called a "negative
1142 dentry". This function is commonly called when an inode is
1143 created for an existing negative dentry
1145 d_lookup: look up a dentry given its parent and path name component
1146 It looks up the child of that given name from the dcache
1147 hash table. If it is found, the reference count is incremented
1148 and the dentry is returned. The caller must use dput()
1149 to free the dentry when it finishes using it.
1157 On mount and remount the filesystem is passed a string containing a
1158 comma separated list of mount options. The options can have either of
1164 The <linux/parser.h> header defines an API that helps parse these
1165 options. There are plenty of examples on how to use it in existing
1171 If a filesystem accepts mount options, it must define show_options()
1172 to show all the currently active options. The rules are:
1174 - options MUST be shown which are not default or their values differ
1177 - options MAY be shown which are enabled by default or have their
1180 Options used only internally between a mount helper and the kernel
1181 (such as file descriptors), or which only have an effect during the
1182 mounting (such as ones controlling the creation of a journal) are exempt
1183 from the above rules.
1185 The underlying reason for the above rules is to make sure, that a
1186 mount can be accurately replicated (e.g. umounting and mounting again)
1187 based on the information found in /proc/mounts.
1189 A simple method of saving options at mount/remount time and showing
1190 them is provided with the save_mount_options() and
1191 generic_show_options() helper functions. Please note, that using
1192 these may have drawbacks. For more info see header comments for these
1193 functions in fs/namespace.c.
1198 (Note some of these resources are not up-to-date with the latest kernel
1201 Creating Linux virtual filesystems. 2002
1202 <http://lwn.net/Articles/13325/>
1204 The Linux Virtual File-system Layer by Neil Brown. 1999
1205 <http://www.cse.unsw.edu.au/~neilb/oss/linux-commentary/vfs.html>
1207 A tour of the Linux VFS by Michael K. Johnson. 1996
1208 <http://www.tldp.org/LDP/khg/HyperNews/get/fs/vfstour.html>
1210 A small trail through the Linux kernel by Andries Brouwer. 2001
1211 <http://www.win.tue.nl/~aeb/linux/vfs/trail.html>
projects / cascardo / linux.git / blob