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 it's 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 device onto a directory in your filespace,
99 the VFS will call the appropriate get_sb() method for the specific
100 filesystem. The dentry for the mount point will then be updated to
101 point to the root inode for the new filesystem.
103 You can see all filesystems that are registered to the kernel in the
104 file /proc/filesystems.
107 struct file_system_type
108 -----------------------
110 This describes the filesystem. As of kernel 2.6.22, the following
113 struct file_system_type {
116 int (*get_sb) (struct file_system_type *, int,
117 const char *, void *, struct vfsmount *);
118 void (*kill_sb) (struct super_block *);
119 struct module *owner;
120 struct file_system_type * next;
121 struct list_head fs_supers;
122 struct lock_class_key s_lock_key;
123 struct lock_class_key s_umount_key;
126 name: the name of the filesystem type, such as "ext2", "iso9660",
129 fs_flags: various flags (i.e. FS_REQUIRES_DEV, FS_NO_DCACHE, etc.)
131 get_sb: the method to call when a new instance of this
132 filesystem should be mounted
134 kill_sb: the method to call when an instance of this filesystem
137 owner: for internal VFS use: you should initialize this to THIS_MODULE in
140 next: for internal VFS use: you should initialize this to NULL
142 s_lock_key, s_umount_key: lockdep-specific
144 The get_sb() method has the following arguments:
146 struct file_system_type *fs_type: decribes the filesystem, partly initialized
147 by the specific filesystem code
149 int flags: mount flags
151 const char *dev_name: the device name we are mounting.
153 void *data: arbitrary mount options, usually comes as an ASCII
154 string (see "Mount Options" section)
156 struct vfsmount *mnt: a vfs-internal representation of a mount point
158 The get_sb() method must determine if the block device specified
159 in the dev_name and fs_type contains a filesystem of the type the method
160 supports. If it succeeds in opening the named block device, it initializes a
161 struct super_block descriptor for the filesystem contained by the block device.
162 On failure it returns an error.
164 The most interesting member of the superblock structure that the
165 get_sb() method fills in is the "s_op" field. This is a pointer to
166 a "struct super_operations" which describes the next level of the
167 filesystem implementation.
169 Usually, a filesystem uses one of the generic get_sb() implementations
170 and provides a fill_super() method instead. The generic methods are:
172 get_sb_bdev: mount a filesystem residing on a block device
174 get_sb_nodev: mount a filesystem that is not backed by a device
176 get_sb_single: mount a filesystem which shares the instance between
179 A fill_super() method implementation has the following arguments:
181 struct super_block *sb: the superblock structure. The method fill_super()
182 must initialize this properly.
184 void *data: arbitrary mount options, usually comes as an ASCII
185 string (see "Mount Options" section)
187 int silent: whether or not to be silent on error
190 The Superblock Object
191 =====================
193 A superblock object represents a mounted filesystem.
196 struct super_operations
197 -----------------------
199 This describes how the VFS can manipulate the superblock of your
200 filesystem. As of kernel 2.6.22, the following members are defined:
202 struct super_operations {
203 struct inode *(*alloc_inode)(struct super_block *sb);
204 void (*destroy_inode)(struct inode *);
206 void (*dirty_inode) (struct inode *);
207 int (*write_inode) (struct inode *, int);
208 void (*drop_inode) (struct inode *);
209 void (*delete_inode) (struct inode *);
210 void (*put_super) (struct super_block *);
211 void (*write_super) (struct super_block *);
212 int (*sync_fs)(struct super_block *sb, int wait);
213 void (*write_super_lockfs) (struct super_block *);
214 void (*unlockfs) (struct super_block *);
215 int (*statfs) (struct dentry *, struct kstatfs *);
216 int (*remount_fs) (struct super_block *, int *, char *);
217 void (*clear_inode) (struct inode *);
218 void (*umount_begin) (struct super_block *);
220 int (*show_options)(struct seq_file *, struct vfsmount *);
222 ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
223 ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
226 All methods are called without any locks being held, unless otherwise
227 noted. This means that most methods can block safely. All methods are
228 only called from a process context (i.e. not from an interrupt handler
231 alloc_inode: this method is called by inode_alloc() to allocate memory
232 for struct inode and initialize it. If this function is not
233 defined, a simple 'struct inode' is allocated. Normally
234 alloc_inode will be used to allocate a larger structure which
235 contains a 'struct inode' embedded within it.
237 destroy_inode: this method is called by destroy_inode() to release
238 resources allocated for struct inode. It is only required if
239 ->alloc_inode was defined and simply undoes anything done by
242 dirty_inode: this method is called by the VFS to mark an inode dirty.
244 write_inode: this method is called when the VFS needs to write an
245 inode to disc. The second parameter indicates whether the write
246 should be synchronous or not, not all filesystems check this flag.
248 drop_inode: called when the last access to the inode is dropped,
249 with the inode_lock spinlock held.
251 This method should be either NULL (normal UNIX filesystem
252 semantics) or "generic_delete_inode" (for filesystems that do not
253 want to cache inodes - causing "delete_inode" to always be
254 called regardless of the value of i_nlink)
256 The "generic_delete_inode()" behavior is equivalent to the
257 old practice of using "force_delete" in the put_inode() case,
258 but does not have the races that the "force_delete()" approach
261 delete_inode: called when the VFS wants to delete an inode
263 put_super: called when the VFS wishes to free the superblock
264 (i.e. unmount). This is called with the superblock lock held
266 write_super: called when the VFS superblock needs to be written to
267 disc. This method is optional
269 sync_fs: called when VFS is writing out all dirty data associated with
270 a superblock. The second parameter indicates whether the method
271 should wait until the write out has been completed. Optional.
273 write_super_lockfs: called when VFS is locking a filesystem and
274 forcing it into a consistent state. This method is currently
275 used by the Logical Volume Manager (LVM).
277 unlockfs: called when VFS is unlocking a filesystem and making it writable
280 statfs: called when the VFS needs to get filesystem statistics. This
281 is called with the kernel lock held
283 remount_fs: called when the filesystem is remounted. This is called
284 with the kernel lock held
286 clear_inode: called then the VFS clears the inode. Optional
288 umount_begin: called when the VFS is unmounting a filesystem.
290 show_options: called by the VFS to show mount options for
291 /proc/<pid>/mounts. (see "Mount Options" section)
293 quota_read: called by the VFS to read from filesystem quota file.
295 quota_write: called by the VFS to write to filesystem quota file.
297 Whoever sets up the inode is responsible for filling in the "i_op" field. This
298 is a pointer to a "struct inode_operations" which describes the methods that
299 can be performed on individual inodes.
305 An inode object represents an object within the filesystem.
308 struct inode_operations
309 -----------------------
311 This describes how the VFS can manipulate an inode in your
312 filesystem. As of kernel 2.6.22, the following members are defined:
314 struct inode_operations {
315 int (*create) (struct inode *,struct dentry *,int, struct nameidata *);
316 struct dentry * (*lookup) (struct inode *,struct dentry *, struct nameidata *);
317 int (*link) (struct dentry *,struct inode *,struct dentry *);
318 int (*unlink) (struct inode *,struct dentry *);
319 int (*symlink) (struct inode *,struct dentry *,const char *);
320 int (*mkdir) (struct inode *,struct dentry *,int);
321 int (*rmdir) (struct inode *,struct dentry *);
322 int (*mknod) (struct inode *,struct dentry *,int,dev_t);
323 int (*rename) (struct inode *, struct dentry *,
324 struct inode *, struct dentry *);
325 int (*readlink) (struct dentry *, char __user *,int);
326 void * (*follow_link) (struct dentry *, struct nameidata *);
327 void (*put_link) (struct dentry *, struct nameidata *, void *);
328 void (*truncate) (struct inode *);
329 int (*permission) (struct inode *, int, struct nameidata *);
330 int (*setattr) (struct dentry *, struct iattr *);
331 int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
332 int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
333 ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
334 ssize_t (*listxattr) (struct dentry *, char *, size_t);
335 int (*removexattr) (struct dentry *, const char *);
336 void (*truncate_range)(struct inode *, loff_t, loff_t);
339 Again, all methods are called without any locks being held, unless
342 create: called by the open(2) and creat(2) system calls. Only
343 required if you want to support regular files. The dentry you
344 get should not have an inode (i.e. it should be a negative
345 dentry). Here you will probably call d_instantiate() with the
346 dentry and the newly created inode
348 lookup: called when the VFS needs to look up an inode in a parent
349 directory. The name to look for is found in the dentry. This
350 method must call d_add() to insert the found inode into the
351 dentry. The "i_count" field in the inode structure should be
352 incremented. If the named inode does not exist a NULL inode
353 should be inserted into the dentry (this is called a negative
354 dentry). Returning an error code from this routine must only
355 be done on a real error, otherwise creating inodes with system
356 calls like create(2), mknod(2), mkdir(2) and so on will fail.
357 If you wish to overload the dentry methods then you should
358 initialise the "d_dop" field in the dentry; this is a pointer
359 to a struct "dentry_operations".
360 This method is called with the directory inode semaphore held
362 link: called by the link(2) system call. Only required if you want
363 to support hard links. You will probably need to call
364 d_instantiate() just as you would in the create() method
366 unlink: called by the unlink(2) system call. Only required if you
367 want to support deleting inodes
369 symlink: called by the symlink(2) system call. Only required if you
370 want to support symlinks. You will probably need to call
371 d_instantiate() just as you would in the create() method
373 mkdir: called by the mkdir(2) system call. Only required if you want
374 to support creating subdirectories. You will probably need to
375 call d_instantiate() just as you would in the create() method
377 rmdir: called by the rmdir(2) system call. Only required if you want
378 to support deleting subdirectories
380 mknod: called by the mknod(2) system call to create a device (char,
381 block) inode or a named pipe (FIFO) or socket. Only required
382 if you want to support creating these types of inodes. You
383 will probably need to call d_instantiate() just as you would
384 in the create() method
386 rename: called by the rename(2) system call to rename the object to
387 have the parent and name given by the second inode and dentry.
389 readlink: called by the readlink(2) system call. Only required if
390 you want to support reading symbolic links
392 follow_link: called by the VFS to follow a symbolic link to the
393 inode it points to. Only required if you want to support
394 symbolic links. This method returns a void pointer cookie
395 that is passed to put_link().
397 put_link: called by the VFS to release resources allocated by
398 follow_link(). The cookie returned by follow_link() is passed
399 to this method as the last parameter. It is used by
400 filesystems such as NFS where page cache is not stable
401 (i.e. page that was installed when the symbolic link walk
402 started might not be in the page cache at the end of the
405 truncate: called by the VFS to change the size of a file. The
406 i_size field of the inode is set to the desired size by the
407 VFS before this method is called. This method is called by
408 the truncate(2) system call and related functionality.
410 permission: called by the VFS to check for access rights on a POSIX-like
413 setattr: called by the VFS to set attributes for a file. This method
414 is called by chmod(2) and related system calls.
416 getattr: called by the VFS to get attributes of a file. This method
417 is called by stat(2) and related system calls.
419 setxattr: called by the VFS to set an extended attribute for a file.
420 Extended attribute is a name:value pair associated with an
421 inode. This method is called by setxattr(2) system call.
423 getxattr: called by the VFS to retrieve the value of an extended
424 attribute name. This method is called by getxattr(2) function
427 listxattr: called by the VFS to list all extended attributes for a
428 given file. This method is called by listxattr(2) system call.
430 removexattr: called by the VFS to remove an extended attribute from
431 a file. This method is called by removexattr(2) system call.
433 truncate_range: a method provided by the underlying filesystem to truncate a
434 range of blocks , i.e. punch a hole somewhere in a file.
437 The Address Space Object
438 ========================
440 The address space object is used to group and manage pages in the page
441 cache. It can be used to keep track of the pages in a file (or
442 anything else) and also track the mapping of sections of the file into
443 process address spaces.
445 There are a number of distinct yet related services that an
446 address-space can provide. These include communicating memory
447 pressure, page lookup by address, and keeping track of pages tagged as
450 The first can be used independently to the others. The VM can try to
451 either write dirty pages in order to clean them, or release clean
452 pages in order to reuse them. To do this it can call the ->writepage
453 method on dirty pages, and ->releasepage on clean pages with
454 PagePrivate set. Clean pages without PagePrivate and with no external
455 references will be released without notice being given to the
458 To achieve this functionality, pages need to be placed on an LRU with
459 lru_cache_add and mark_page_active needs to be called whenever the
462 Pages are normally kept in a radix tree index by ->index. This tree
463 maintains information about the PG_Dirty and PG_Writeback status of
464 each page, so that pages with either of these flags can be found
467 The Dirty tag is primarily used by mpage_writepages - the default
468 ->writepages method. It uses the tag to find dirty pages to call
469 ->writepage on. If mpage_writepages is not used (i.e. the address
470 provides its own ->writepages) , the PAGECACHE_TAG_DIRTY tag is
471 almost unused. write_inode_now and sync_inode do use it (through
472 __sync_single_inode) to check if ->writepages has been successful in
473 writing out the whole address_space.
475 The Writeback tag is used by filemap*wait* and sync_page* functions,
476 via wait_on_page_writeback_range, to wait for all writeback to
477 complete. While waiting ->sync_page (if defined) will be called on
478 each page that is found to require writeback.
480 An address_space handler may attach extra information to a page,
481 typically using the 'private' field in the 'struct page'. If such
482 information is attached, the PG_Private flag should be set. This will
483 cause various VM routines to make extra calls into the address_space
484 handler to deal with that data.
486 An address space acts as an intermediate between storage and
487 application. Data is read into the address space a whole page at a
488 time, and provided to the application either by copying of the page,
489 or by memory-mapping the page.
490 Data is written into the address space by the application, and then
491 written-back to storage typically in whole pages, however the
492 address_space has finer control of write sizes.
494 The read process essentially only requires 'readpage'. The write
495 process is more complicated and uses prepare_write/commit_write or
496 set_page_dirty to write data into the address_space, and writepage,
497 sync_page, and writepages to writeback data to storage.
499 Adding and removing pages to/from an address_space is protected by the
502 When data is written to a page, the PG_Dirty flag should be set. It
503 typically remains set until writepage asks for it to be written. This
504 should clear PG_Dirty and set PG_Writeback. It can be actually
505 written at any point after PG_Dirty is clear. Once it is known to be
506 safe, PG_Writeback is cleared.
508 Writeback makes use of a writeback_control structure...
510 struct address_space_operations
511 -------------------------------
513 This describes how the VFS can manipulate mapping of a file to page cache in
514 your filesystem. As of kernel 2.6.22, the following members are defined:
516 struct address_space_operations {
517 int (*writepage)(struct page *page, struct writeback_control *wbc);
518 int (*readpage)(struct file *, struct page *);
519 int (*sync_page)(struct page *);
520 int (*writepages)(struct address_space *, struct writeback_control *);
521 int (*set_page_dirty)(struct page *page);
522 int (*readpages)(struct file *filp, struct address_space *mapping,
523 struct list_head *pages, unsigned nr_pages);
524 int (*prepare_write)(struct file *, struct page *, unsigned, unsigned);
525 int (*commit_write)(struct file *, struct page *, unsigned, unsigned);
526 int (*write_begin)(struct file *, struct address_space *mapping,
527 loff_t pos, unsigned len, unsigned flags,
528 struct page **pagep, void **fsdata);
529 int (*write_end)(struct file *, struct address_space *mapping,
530 loff_t pos, unsigned len, unsigned copied,
531 struct page *page, void *fsdata);
532 sector_t (*bmap)(struct address_space *, sector_t);
533 int (*invalidatepage) (struct page *, unsigned long);
534 int (*releasepage) (struct page *, int);
535 ssize_t (*direct_IO)(int, struct kiocb *, const struct iovec *iov,
536 loff_t offset, unsigned long nr_segs);
537 struct page* (*get_xip_page)(struct address_space *, sector_t,
539 /* migrate the contents of a page to the specified target */
540 int (*migratepage) (struct page *, struct page *);
541 int (*launder_page) (struct page *);
544 writepage: called by the VM to write a dirty page to backing store.
545 This may happen for data integrity reasons (i.e. 'sync'), or
546 to free up memory (flush). The difference can be seen in
548 The PG_Dirty flag has been cleared and PageLocked is true.
549 writepage should start writeout, should set PG_Writeback,
550 and should make sure the page is unlocked, either synchronously
551 or asynchronously when the write operation completes.
553 If wbc->sync_mode is WB_SYNC_NONE, ->writepage doesn't have to
554 try too hard if there are problems, and may choose to write out
555 other pages from the mapping if that is easier (e.g. due to
556 internal dependencies). If it chooses not to start writeout, it
557 should return AOP_WRITEPAGE_ACTIVATE so that the VM will not keep
558 calling ->writepage on that page.
560 See the file "Locking" for more details.
562 readpage: called by the VM to read a page from backing store.
563 The page will be Locked when readpage is called, and should be
564 unlocked and marked uptodate once the read completes.
565 If ->readpage discovers that it needs to unlock the page for
566 some reason, it can do so, and then return AOP_TRUNCATED_PAGE.
567 In this case, the page will be relocated, relocked and if
568 that all succeeds, ->readpage will be called again.
570 sync_page: called by the VM to notify the backing store to perform all
571 queued I/O operations for a page. I/O operations for other pages
572 associated with this address_space object may also be performed.
574 This function is optional and is called only for pages with
575 PG_Writeback set while waiting for the writeback to complete.
577 writepages: called by the VM to write out pages associated with the
578 address_space object. If wbc->sync_mode is WBC_SYNC_ALL, then
579 the writeback_control will specify a range of pages that must be
580 written out. If it is WBC_SYNC_NONE, then a nr_to_write is given
581 and that many pages should be written if possible.
582 If no ->writepages is given, then mpage_writepages is used
583 instead. This will choose pages from the address space that are
584 tagged as DIRTY and will pass them to ->writepage.
586 set_page_dirty: called by the VM to set a page dirty.
587 This is particularly needed if an address space attaches
588 private data to a page, and that data needs to be updated when
589 a page is dirtied. This is called, for example, when a memory
590 mapped page gets modified.
591 If defined, it should set the PageDirty flag, and the
592 PAGECACHE_TAG_DIRTY tag in the radix tree.
594 readpages: called by the VM to read pages associated with the address_space
595 object. This is essentially just a vector version of
596 readpage. Instead of just one page, several pages are
598 readpages is only used for read-ahead, so read errors are
599 ignored. If anything goes wrong, feel free to give up.
601 prepare_write: called by the generic write path in VM to set up a write
602 request for a page. This indicates to the address space that
603 the given range of bytes is about to be written. The
604 address_space should check that the write will be able to
605 complete, by allocating space if necessary and doing any other
606 internal housekeeping. If the write will update parts of
607 any basic-blocks on storage, then those blocks should be
608 pre-read (if they haven't been read already) so that the
609 updated blocks can be written out properly.
610 The page will be locked.
612 Note: the page _must not_ be marked uptodate in this function
613 (or anywhere else) unless it actually is uptodate right now. As
614 soon as a page is marked uptodate, it is possible for a concurrent
615 read(2) to copy it to userspace.
617 commit_write: If prepare_write succeeds, new data will be copied
618 into the page and then commit_write will be called. It will
619 typically update the size of the file (if appropriate) and
620 mark the inode as dirty, and do any other related housekeeping
621 operations. It should avoid returning an error if possible -
622 errors should have been handled by prepare_write.
624 write_begin: This is intended as a replacement for prepare_write. The
625 key differences being that:
626 - it returns a locked page (in *pagep) rather than being
627 given a pre locked page;
628 - it must be able to cope with short writes (where the
629 length passed to write_begin is greater than the number
630 of bytes copied into the page).
632 Called by the generic buffered write code to ask the filesystem to
633 prepare to write len bytes at the given offset in the file. The
634 address_space should check that the write will be able to complete,
635 by allocating space if necessary and doing any other internal
636 housekeeping. If the write will update parts of any basic-blocks on
637 storage, then those blocks should be pre-read (if they haven't been
638 read already) so that the updated blocks can be written out properly.
640 The filesystem must return the locked pagecache page for the specified
641 offset, in *pagep, for the caller to write into.
643 flags is a field for AOP_FLAG_xxx flags, described in
646 A void * may be returned in fsdata, which then gets passed into
649 Returns 0 on success; < 0 on failure (which is the error code), in
650 which case write_end is not called.
652 write_end: After a successful write_begin, and data copy, write_end must
653 be called. len is the original len passed to write_begin, and copied
654 is the amount that was able to be copied (copied == len is always true
655 if write_begin was called with the AOP_FLAG_UNINTERRUPTIBLE flag).
657 The filesystem must take care of unlocking the page and releasing it
658 refcount, and updating i_size.
660 Returns < 0 on failure, otherwise the number of bytes (<= 'copied')
661 that were able to be copied into pagecache.
663 bmap: called by the VFS to map a logical block offset within object to
664 physical block number. This method is used by the FIBMAP
665 ioctl and for working with swap-files. To be able to swap to
666 a file, the file must have a stable mapping to a block
667 device. The swap system does not go through the filesystem
668 but instead uses bmap to find out where the blocks in the file
669 are and uses those addresses directly.
672 invalidatepage: If a page has PagePrivate set, then invalidatepage
673 will be called when part or all of the page is to be removed
674 from the address space. This generally corresponds to either a
675 truncation or a complete invalidation of the address space
676 (in the latter case 'offset' will always be 0).
677 Any private data associated with the page should be updated
678 to reflect this truncation. If offset is 0, then
679 the private data should be released, because the page
680 must be able to be completely discarded. This may be done by
681 calling the ->releasepage function, but in this case the
682 release MUST succeed.
684 releasepage: releasepage is called on PagePrivate pages to indicate
685 that the page should be freed if possible. ->releasepage
686 should remove any private data from the page and clear the
687 PagePrivate flag. It may also remove the page from the
688 address_space. If this fails for some reason, it may indicate
689 failure with a 0 return value.
690 This is used in two distinct though related cases. The first
691 is when the VM finds a clean page with no active users and
692 wants to make it a free page. If ->releasepage succeeds, the
693 page will be removed from the address_space and become free.
695 The second case is when a request has been made to invalidate
696 some or all pages in an address_space. This can happen
697 through the fadvice(POSIX_FADV_DONTNEED) system call or by the
698 filesystem explicitly requesting it as nfs and 9fs do (when
699 they believe the cache may be out of date with storage) by
700 calling invalidate_inode_pages2().
701 If the filesystem makes such a call, and needs to be certain
702 that all pages are invalidated, then its releasepage will
703 need to ensure this. Possibly it can clear the PageUptodate
704 bit if it cannot free private data yet.
706 direct_IO: called by the generic read/write routines to perform
707 direct_IO - that is IO requests which bypass the page cache
708 and transfer data directly between the storage and the
709 application's address space.
711 get_xip_page: called by the VM to translate a block number to a page.
712 The page is valid until the corresponding filesystem is unmounted.
713 Filesystems that want to use execute-in-place (XIP) need to implement
714 it. An example implementation can be found in fs/ext2/xip.c.
716 migrate_page: This is used to compact the physical memory usage.
717 If the VM wants to relocate a page (maybe off a memory card
718 that is signalling imminent failure) it will pass a new page
719 and an old page to this function. migrate_page should
720 transfer any private data across and update any references
721 that it has to the page.
723 launder_page: Called before freeing a page - it writes back the dirty page. To
724 prevent redirtying the page, it is kept locked during the whole
730 A file object represents a file opened by a process.
733 struct file_operations
734 ----------------------
736 This describes how the VFS can manipulate an open file. As of kernel
737 2.6.22, the following members are defined:
739 struct file_operations {
740 struct module *owner;
741 loff_t (*llseek) (struct file *, loff_t, int);
742 ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
743 ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
744 ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
745 ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
746 int (*readdir) (struct file *, void *, filldir_t);
747 unsigned int (*poll) (struct file *, struct poll_table_struct *);
748 int (*ioctl) (struct inode *, struct file *, unsigned int, unsigned long);
749 long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
750 long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
751 int (*mmap) (struct file *, struct vm_area_struct *);
752 int (*open) (struct inode *, struct file *);
753 int (*flush) (struct file *);
754 int (*release) (struct inode *, struct file *);
755 int (*fsync) (struct file *, struct dentry *, int datasync);
756 int (*aio_fsync) (struct kiocb *, int datasync);
757 int (*fasync) (int, struct file *, int);
758 int (*lock) (struct file *, int, struct file_lock *);
759 ssize_t (*readv) (struct file *, const struct iovec *, unsigned long, loff_t *);
760 ssize_t (*writev) (struct file *, const struct iovec *, unsigned long, loff_t *);
761 ssize_t (*sendfile) (struct file *, loff_t *, size_t, read_actor_t, void *);
762 ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
763 unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
764 int (*check_flags)(int);
765 int (*dir_notify)(struct file *filp, unsigned long arg);
766 int (*flock) (struct file *, int, struct file_lock *);
767 ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, size_t, unsigned int);
768 ssize_t (*splice_read)(struct file *, struct pipe_inode_info *, size_t, unsigned int);
771 Again, all methods are called without any locks being held, unless
774 llseek: called when the VFS needs to move the file position index
776 read: called by read(2) and related system calls
778 aio_read: called by io_submit(2) and other asynchronous I/O operations
780 write: called by write(2) and related system calls
782 aio_write: called by io_submit(2) and other asynchronous I/O operations
784 readdir: called when the VFS needs to read the directory contents
786 poll: called by the VFS when a process wants to check if there is
787 activity on this file and (optionally) go to sleep until there
788 is activity. Called by the select(2) and poll(2) system calls
790 ioctl: called by the ioctl(2) system call
792 unlocked_ioctl: called by the ioctl(2) system call. Filesystems that do not
793 require the BKL should use this method instead of the ioctl() above.
795 compat_ioctl: called by the ioctl(2) system call when 32 bit system calls
796 are used on 64 bit kernels.
798 mmap: called by the mmap(2) system call
800 open: called by the VFS when an inode should be opened. When the VFS
801 opens a file, it creates a new "struct file". It then calls the
802 open method for the newly allocated file structure. You might
803 think that the open method really belongs in
804 "struct inode_operations", and you may be right. I think it's
805 done the way it is because it makes filesystems simpler to
806 implement. The open() method is a good place to initialize the
807 "private_data" member in the file structure if you want to point
808 to a device structure
810 flush: called by the close(2) system call to flush a file
812 release: called when the last reference to an open file is closed
814 fsync: called by the fsync(2) system call
816 fasync: called by the fcntl(2) system call when asynchronous
817 (non-blocking) mode is enabled for a file
819 lock: called by the fcntl(2) system call for F_GETLK, F_SETLK, and F_SETLKW
822 readv: called by the readv(2) system call
824 writev: called by the writev(2) system call
826 sendfile: called by the sendfile(2) system call
828 get_unmapped_area: called by the mmap(2) system call
830 check_flags: called by the fcntl(2) system call for F_SETFL command
832 dir_notify: called by the fcntl(2) system call for F_NOTIFY command
834 flock: called by the flock(2) system call
836 splice_write: called by the VFS to splice data from a pipe to a file. This
837 method is used by the splice(2) system call
839 splice_read: called by the VFS to splice data from file to a pipe. This
840 method is used by the splice(2) system call
842 Note that the file operations are implemented by the specific
843 filesystem in which the inode resides. When opening a device node
844 (character or block special) most filesystems will call special
845 support routines in the VFS which will locate the required device
846 driver information. These support routines replace the filesystem file
847 operations with those for the device driver, and then proceed to call
848 the new open() method for the file. This is how opening a device file
849 in the filesystem eventually ends up calling the device driver open()
853 Directory Entry Cache (dcache)
854 ==============================
857 struct dentry_operations
858 ------------------------
860 This describes how a filesystem can overload the standard dentry
861 operations. Dentries and the dcache are the domain of the VFS and the
862 individual filesystem implementations. Device drivers have no business
863 here. These methods may be set to NULL, as they are either optional or
864 the VFS uses a default. As of kernel 2.6.22, the following members are
867 struct dentry_operations {
868 int (*d_revalidate)(struct dentry *, struct nameidata *);
869 int (*d_hash) (struct dentry *, struct qstr *);
870 int (*d_compare) (struct dentry *, struct qstr *, struct qstr *);
871 int (*d_delete)(struct dentry *);
872 void (*d_release)(struct dentry *);
873 void (*d_iput)(struct dentry *, struct inode *);
874 char *(*d_dname)(struct dentry *, char *, int);
877 d_revalidate: called when the VFS needs to revalidate a dentry. This
878 is called whenever a name look-up finds a dentry in the
879 dcache. Most filesystems leave this as NULL, because all their
880 dentries in the dcache are valid
882 d_hash: called when the VFS adds a dentry to the hash table
884 d_compare: called when a dentry should be compared with another
886 d_delete: called when the last reference to a dentry is
887 deleted. This means no-one is using the dentry, however it is
888 still valid and in the dcache
890 d_release: called when a dentry is really deallocated
892 d_iput: called when a dentry loses its inode (just prior to its
893 being deallocated). The default when this is NULL is that the
894 VFS calls iput(). If you define this method, you must call
897 d_dname: called when the pathname of a dentry should be generated.
898 Usefull for some pseudo filesystems (sockfs, pipefs, ...) to delay
899 pathname generation. (Instead of doing it when dentry is created,
900 its done only when the path is needed.). Real filesystems probably
901 dont want to use it, because their dentries are present in global
902 dcache hash, so their hash should be an invariant. As no lock is
903 held, d_dname() should not try to modify the dentry itself, unless
904 appropriate SMP safety is used. CAUTION : d_path() logic is quite
905 tricky. The correct way to return for example "Hello" is to put it
906 at the end of the buffer, and returns a pointer to the first char.
907 dynamic_dname() helper function is provided to take care of this.
911 static char *pipefs_dname(struct dentry *dent, char *buffer, int buflen)
913 return dynamic_dname(dentry, buffer, buflen, "pipe:[%lu]",
914 dentry->d_inode->i_ino);
917 Each dentry has a pointer to its parent dentry, as well as a hash list
918 of child dentries. Child dentries are basically like files in a
922 Directory Entry Cache API
923 --------------------------
925 There are a number of functions defined which permit a filesystem to
928 dget: open a new handle for an existing dentry (this just increments
931 dput: close a handle for a dentry (decrements the usage count). If
932 the usage count drops to 0, the "d_delete" method is called
933 and the dentry is placed on the unused list if the dentry is
934 still in its parents hash list. Putting the dentry on the
935 unused list just means that if the system needs some RAM, it
936 goes through the unused list of dentries and deallocates them.
937 If the dentry has already been unhashed and the usage count
938 drops to 0, in this case the dentry is deallocated after the
939 "d_delete" method is called
941 d_drop: this unhashes a dentry from its parents hash list. A
942 subsequent call to dput() will deallocate the dentry if its
943 usage count drops to 0
945 d_delete: delete a dentry. If there are no other open references to
946 the dentry then the dentry is turned into a negative dentry
947 (the d_iput() method is called). If there are other
948 references, then d_drop() is called instead
950 d_add: add a dentry to its parents hash list and then calls
953 d_instantiate: add a dentry to the alias hash list for the inode and
954 updates the "d_inode" member. The "i_count" member in the
955 inode structure should be set/incremented. If the inode
956 pointer is NULL, the dentry is called a "negative
957 dentry". This function is commonly called when an inode is
958 created for an existing negative dentry
960 d_lookup: look up a dentry given its parent and path name component
961 It looks up the child of that given name from the dcache
962 hash table. If it is found, the reference count is incremented
963 and the dentry is returned. The caller must use d_put()
964 to free the dentry when it finishes using it.
966 For further information on dentry locking, please refer to the document
967 Documentation/filesystems/dentry-locking.txt.
975 On mount and remount the filesystem is passed a string containing a
976 comma separated list of mount options. The options can have either of
982 The <linux/parser.h> header defines an API that helps parse these
983 options. There are plenty of examples on how to use it in existing
989 If a filesystem accepts mount options, it must define show_options()
990 to show all the currently active options. The rules are:
992 - options MUST be shown which are not default or their values differ
995 - options MAY be shown which are enabled by default or have their
998 Options used only internally between a mount helper and the kernel
999 (such as file descriptors), or which only have an effect during the
1000 mounting (such as ones controlling the creation of a journal) are exempt
1001 from the above rules.
1003 The underlying reason for the above rules is to make sure, that a
1004 mount can be accurately replicated (e.g. umounting and mounting again)
1005 based on the information found in /proc/mounts.
1007 A simple method of saving options at mount/remount time and showing
1008 them is provided with the save_mount_options() and
1009 generic_show_options() helper functions. Please note, that using
1010 these may have drawbacks. For more info see header comments for these
1011 functions in fs/namespace.c.
1016 (Note some of these resources are not up-to-date with the latest kernel
1019 Creating Linux virtual filesystems. 2002
1020 <http://lwn.net/Articles/13325/>
1022 The Linux Virtual File-system Layer by Neil Brown. 1999
1023 <http://www.cse.unsw.edu.au/~neilb/oss/linux-commentary/vfs.html>
1025 A tour of the Linux VFS by Michael K. Johnson. 1996
1026 <http://www.tldp.org/LDP/khg/HyperNews/get/fs/vfstour.html>
1028 A small trail through the Linux kernel by Andries Brouwer. 2001
1029 <http://www.win.tue.nl/~aeb/linux/vfs/trail.html>