blob: d449e632e6a09fca5ca0fedd6bdc2cf87eb44a96 [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -07001 The text below describes the locking rules for VFS-related methods.
2It is (believed to be) up-to-date. *Please*, if you change anything in
3prototypes or locking protocols - update this file. And update the relevant
4instances in the tree, don't leave that to maintainers of filesystems/devices/
5etc. At the very least, put the list of dubious cases in the end of this file.
6Don't turn it into log - maintainers of out-of-the-tree code are supposed to
7be able to use diff(1).
8 Thing currently missing here: socket operations. Alexey?
9
10--------------------------- dentry_operations --------------------------
11prototypes:
Nick Piggin34286d62011-01-07 17:49:57 +110012 int (*d_revalidate)(struct dentry *, struct nameidata *);
Nick Pigginb1e6a012011-01-07 17:49:28 +110013 int (*d_hash)(const struct dentry *, const struct inode *,
14 struct qstr *);
Nick Piggin621e1552011-01-07 17:49:27 +110015 int (*d_compare)(const struct dentry *, const struct inode *,
16 const struct dentry *, const struct inode *,
17 unsigned int, const char *, const struct qstr *);
Linus Torvalds1da177e2005-04-16 15:20:36 -070018 int (*d_delete)(struct dentry *);
19 void (*d_release)(struct dentry *);
20 void (*d_iput)(struct dentry *, struct inode *);
Eric Dumazetc23fbb62007-05-08 00:26:18 -070021 char *(*d_dname)((struct dentry *dentry, char *buffer, int buflen);
David Howells9875cf82011-01-14 18:45:21 +000022 struct vfsmount *(*d_automount)(struct path *path);
David Howellscc53ce52011-01-14 18:45:26 +000023 int (*d_manage)(struct dentry *, bool);
Linus Torvalds1da177e2005-04-16 15:20:36 -070024
25locking rules:
Nick Piggin34286d62011-01-07 17:49:57 +110026 rename_lock ->d_lock may block rcu-walk
27d_revalidate: no no yes (ref-walk) maybe
28d_hash no no no maybe
29d_compare: yes no no maybe
30d_delete: no yes no no
31d_release: no no yes no
Sage Weilf0023bc2011-10-28 10:02:42 -070032d_prune: no yes no no
Nick Piggin34286d62011-01-07 17:49:57 +110033d_iput: no no yes no
34d_dname: no no no no
David Howells9875cf82011-01-14 18:45:21 +000035d_automount: no no yes no
David Howellsab909112011-01-14 18:46:51 +000036d_manage: no no yes (ref-walk) maybe
Linus Torvalds1da177e2005-04-16 15:20:36 -070037
38--------------------------- inode_operations ---------------------------
39prototypes:
Al Viro4acdaf22011-07-26 01:42:34 -040040 int (*create) (struct inode *,struct dentry *,umode_t, struct nameidata *);
Linus Torvalds1da177e2005-04-16 15:20:36 -070041 struct dentry * (*lookup) (struct inode *,struct dentry *, struct nameid
42ata *);
43 int (*link) (struct dentry *,struct inode *,struct dentry *);
44 int (*unlink) (struct inode *,struct dentry *);
45 int (*symlink) (struct inode *,struct dentry *,const char *);
Al Viro18bb1db2011-07-26 01:41:39 -040046 int (*mkdir) (struct inode *,struct dentry *,umode_t);
Linus Torvalds1da177e2005-04-16 15:20:36 -070047 int (*rmdir) (struct inode *,struct dentry *);
Al Viro1a67aaf2011-07-26 01:52:52 -040048 int (*mknod) (struct inode *,struct dentry *,umode_t,dev_t);
Linus Torvalds1da177e2005-04-16 15:20:36 -070049 int (*rename) (struct inode *, struct dentry *,
50 struct inode *, struct dentry *);
51 int (*readlink) (struct dentry *, char __user *,int);
Christoph Hellwigb83be6f2010-12-16 12:04:54 +010052 void * (*follow_link) (struct dentry *, struct nameidata *);
53 void (*put_link) (struct dentry *, struct nameidata *, void *);
Linus Torvalds1da177e2005-04-16 15:20:36 -070054 void (*truncate) (struct inode *);
Nick Pigginb74c79e2011-01-07 17:49:58 +110055 int (*permission) (struct inode *, int, unsigned int);
Christoph Hellwig4e34e712011-07-23 17:37:31 +020056 int (*get_acl)(struct inode *, int);
Linus Torvalds1da177e2005-04-16 15:20:36 -070057 int (*setattr) (struct dentry *, struct iattr *);
58 int (*getattr) (struct vfsmount *, struct dentry *, struct kstat *);
59 int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
60 ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
61 ssize_t (*listxattr) (struct dentry *, char *, size_t);
62 int (*removexattr) (struct dentry *, const char *);
Christoph Hellwigb83be6f2010-12-16 12:04:54 +010063 int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start, u64 len);
Linus Torvalds1da177e2005-04-16 15:20:36 -070064
65locking rules:
Christoph Hellwigb83be6f2010-12-16 12:04:54 +010066 all may block
Artem Bityutskiya7bc02f2007-05-09 07:53:16 +020067 i_mutex(inode)
Linus Torvalds1da177e2005-04-16 15:20:36 -070068lookup: yes
69create: yes
70link: yes (both)
71mknod: yes
72symlink: yes
73mkdir: yes
74unlink: yes (both)
75rmdir: yes (both) (see below)
76rename: yes (all) (see below)
77readlink: no
78follow_link: no
Christoph Hellwigb83be6f2010-12-16 12:04:54 +010079put_link: no
Linus Torvalds1da177e2005-04-16 15:20:36 -070080truncate: yes (see below)
81setattr: yes
Nick Pigginb74c79e2011-01-07 17:49:58 +110082permission: no (may not block if called in rcu-walk mode)
Christoph Hellwig4e34e712011-07-23 17:37:31 +020083get_acl: no
Linus Torvalds1da177e2005-04-16 15:20:36 -070084getattr: no
85setxattr: yes
86getxattr: no
87listxattr: no
88removexattr: yes
Christoph Hellwigb83be6f2010-12-16 12:04:54 +010089fiemap: no
Artem Bityutskiya7bc02f2007-05-09 07:53:16 +020090 Additionally, ->rmdir(), ->unlink() and ->rename() have ->i_mutex on
Linus Torvalds1da177e2005-04-16 15:20:36 -070091victim.
92 cross-directory ->rename() has (per-superblock) ->s_vfs_rename_sem.
93 ->truncate() is never called directly - it's a callback, not a
Christoph Hellwigb83be6f2010-12-16 12:04:54 +010094method. It's called by vmtruncate() - deprecated library function used by
Linus Torvalds1da177e2005-04-16 15:20:36 -070095->setattr(). Locking information above applies to that call (i.e. is
96inherited from ->setattr() - vmtruncate() is used when ATTR_SIZE had been
97passed).
98
99See Documentation/filesystems/directory-locking for more detailed discussion
100of the locking scheme for directory operations.
101
102--------------------------- super_operations ---------------------------
103prototypes:
104 struct inode *(*alloc_inode)(struct super_block *sb);
105 void (*destroy_inode)(struct inode *);
Christoph Hellwigaa385722011-05-27 06:53:02 -0400106 void (*dirty_inode) (struct inode *, int flags);
Christoph Hellwigb83be6f2010-12-16 12:04:54 +0100107 int (*write_inode) (struct inode *, struct writeback_control *wbc);
Al Viro336fb3b2010-06-08 00:37:12 -0400108 int (*drop_inode) (struct inode *);
109 void (*evict_inode) (struct inode *);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700110 void (*put_super) (struct super_block *);
111 void (*write_super) (struct super_block *);
112 int (*sync_fs)(struct super_block *sb, int wait);
Takashi Satoc4be0c12009-01-09 16:40:58 -0800113 int (*freeze_fs) (struct super_block *);
114 int (*unfreeze_fs) (struct super_block *);
David Howells726c3342006-06-23 02:02:58 -0700115 int (*statfs) (struct dentry *, struct kstatfs *);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700116 int (*remount_fs) (struct super_block *, int *, char *);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700117 void (*umount_begin) (struct super_block *);
Al Viro34c80b12011-12-08 21:32:45 -0500118 int (*show_options)(struct seq_file *, struct dentry *);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700119 ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
120 ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
Christoph Hellwigb83be6f2010-12-16 12:04:54 +0100121 int (*bdev_try_to_free_page)(struct super_block*, struct page*, gfp_t);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700122
123locking rules:
Al Viro336fb3b2010-06-08 00:37:12 -0400124 All may block [not true, see below]
Christoph Hellwig7e325d32009-06-19 20:22:37 +0200125 s_umount
126alloc_inode:
127destroy_inode:
Christoph Hellwigaa385722011-05-27 06:53:02 -0400128dirty_inode:
Christoph Hellwig7e325d32009-06-19 20:22:37 +0200129write_inode:
Dave Chinnerf283c862011-03-22 22:23:39 +1100130drop_inode: !!!inode->i_lock!!!
Al Viro336fb3b2010-06-08 00:37:12 -0400131evict_inode:
Christoph Hellwig7e325d32009-06-19 20:22:37 +0200132put_super: write
133write_super: read
134sync_fs: read
135freeze_fs: read
136unfreeze_fs: read
Al Viro336fb3b2010-06-08 00:37:12 -0400137statfs: maybe(read) (see below)
138remount_fs: write
Christoph Hellwig7e325d32009-06-19 20:22:37 +0200139umount_begin: no
140show_options: no (namespace_sem)
141quota_read: no (see below)
142quota_write: no (see below)
Christoph Hellwigb83be6f2010-12-16 12:04:54 +0100143bdev_try_to_free_page: no (see below)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700144
Al Viro336fb3b2010-06-08 00:37:12 -0400145->statfs() has s_umount (shared) when called by ustat(2) (native or
146compat), but that's an accident of bad API; s_umount is used to pin
147the superblock down when we only have dev_t given us by userland to
148identify the superblock. Everything else (statfs(), fstatfs(), etc.)
149doesn't hold it when calling ->statfs() - superblock is pinned down
150by resolving the pathname passed to syscall.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700151->quota_read() and ->quota_write() functions are both guaranteed to
152be the only ones operating on the quota file by the quota code (via
153dqio_sem) (unless an admin really wants to screw up something and
154writes to quota files with quotas on). For other details about locking
155see also dquot_operations section.
Christoph Hellwigb83be6f2010-12-16 12:04:54 +0100156->bdev_try_to_free_page is called from the ->releasepage handler of
157the block device inode. See there for more details.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700158
159--------------------------- file_system_type ---------------------------
160prototypes:
Jonathan Corbet5d8b2eb2006-07-10 04:44:07 -0700161 int (*get_sb) (struct file_system_type *, int,
162 const char *, void *, struct vfsmount *);
Christoph Hellwigb83be6f2010-12-16 12:04:54 +0100163 struct dentry *(*mount) (struct file_system_type *, int,
164 const char *, void *);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700165 void (*kill_sb) (struct super_block *);
166locking rules:
Christoph Hellwigb83be6f2010-12-16 12:04:54 +0100167 may block
Christoph Hellwigb83be6f2010-12-16 12:04:54 +0100168mount yes
169kill_sb yes
Linus Torvalds1da177e2005-04-16 15:20:36 -0700170
Al Viro1a102ff2011-03-16 09:07:58 -0400171->mount() returns ERR_PTR or the root dentry; its superblock should be locked
172on return.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700173->kill_sb() takes a write-locked superblock, does all shutdown work on it,
174unlocks and drops the reference.
175
176--------------------------- address_space_operations --------------------------
177prototypes:
178 int (*writepage)(struct page *page, struct writeback_control *wbc);
179 int (*readpage)(struct file *, struct page *);
180 int (*sync_page)(struct page *);
181 int (*writepages)(struct address_space *, struct writeback_control *);
182 int (*set_page_dirty)(struct page *page);
183 int (*readpages)(struct file *filp, struct address_space *mapping,
184 struct list_head *pages, unsigned nr_pages);
Nick Piggin4e02ed42008-10-29 14:00:55 -0700185 int (*write_begin)(struct file *, struct address_space *mapping,
186 loff_t pos, unsigned len, unsigned flags,
187 struct page **pagep, void **fsdata);
188 int (*write_end)(struct file *, struct address_space *mapping,
189 loff_t pos, unsigned len, unsigned copied,
190 struct page *page, void *fsdata);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700191 sector_t (*bmap)(struct address_space *, sector_t);
192 int (*invalidatepage) (struct page *, unsigned long);
193 int (*releasepage) (struct page *, int);
Linus Torvalds6072d132010-12-01 13:35:19 -0500194 void (*freepage)(struct page *);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700195 int (*direct_IO)(int, struct kiocb *, const struct iovec *iov,
196 loff_t offset, unsigned long nr_segs);
Christoph Hellwigb83be6f2010-12-16 12:04:54 +0100197 int (*get_xip_mem)(struct address_space *, pgoff_t, int, void **,
198 unsigned long *);
199 int (*migratepage)(struct address_space *, struct page *, struct page *);
200 int (*launder_page)(struct page *);
201 int (*is_partially_uptodate)(struct page *, read_descriptor_t *, unsigned long);
202 int (*error_remove_page)(struct address_space *, struct page *);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700203
204locking rules:
Linus Torvalds6072d132010-12-01 13:35:19 -0500205 All except set_page_dirty and freepage may block
Linus Torvalds1da177e2005-04-16 15:20:36 -0700206
Christoph Hellwigb83be6f2010-12-16 12:04:54 +0100207 PageLocked(page) i_mutex
208writepage: yes, unlocks (see below)
209readpage: yes, unlocks
210sync_page: maybe
211writepages:
212set_page_dirty no
213readpages:
214write_begin: locks the page yes
215write_end: yes, unlocks yes
216bmap:
217invalidatepage: yes
218releasepage: yes
219freepage: yes
220direct_IO:
221get_xip_mem: maybe
222migratepage: yes (both)
223launder_page: yes
224is_partially_uptodate: yes
225error_remove_page: yes
Linus Torvalds1da177e2005-04-16 15:20:36 -0700226
Nick Piggin4e02ed42008-10-29 14:00:55 -0700227 ->write_begin(), ->write_end(), ->sync_page() and ->readpage()
Linus Torvalds1da177e2005-04-16 15:20:36 -0700228may be called from the request handler (/dev/loop).
229
230 ->readpage() unlocks the page, either synchronously or via I/O
231completion.
232
233 ->readpages() populates the pagecache with the passed pages and starts
234I/O against them. They come unlocked upon I/O completion.
235
236 ->writepage() is used for two purposes: for "memory cleansing" and for
237"sync". These are quite different operations and the behaviour may differ
238depending upon the mode.
239
240If writepage is called for sync (wbc->sync_mode != WBC_SYNC_NONE) then
241it *must* start I/O against the page, even if that would involve
242blocking on in-progress I/O.
243
244If writepage is called for memory cleansing (sync_mode ==
245WBC_SYNC_NONE) then its role is to get as much writeout underway as
246possible. So writepage should try to avoid blocking against
247currently-in-progress I/O.
248
249If the filesystem is not called for "sync" and it determines that it
250would need to block against in-progress I/O to be able to start new I/O
251against the page the filesystem should redirty the page with
252redirty_page_for_writepage(), then unlock the page and return zero.
253This may also be done to avoid internal deadlocks, but rarely.
254
Robert P. J. Day3a4fa0a2007-10-19 23:10:43 +0200255If the filesystem is called for sync then it must wait on any
Linus Torvalds1da177e2005-04-16 15:20:36 -0700256in-progress I/O and then start new I/O.
257
Nikita Danilov20546062005-05-01 08:58:37 -0700258The filesystem should unlock the page synchronously, before returning to the
259caller, unless ->writepage() returns special WRITEPAGE_ACTIVATE
260value. WRITEPAGE_ACTIVATE means that page cannot really be written out
261currently, and VM should stop calling ->writepage() on this page for some
262time. VM does this by moving page to the head of the active list, hence the
263name.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700264
265Unless the filesystem is going to redirty_page_for_writepage(), unlock the page
266and return zero, writepage *must* run set_page_writeback() against the page,
267followed by unlocking it. Once set_page_writeback() has been run against the
268page, write I/O can be submitted and the write I/O completion handler must run
269end_page_writeback() once the I/O is complete. If no I/O is submitted, the
270filesystem must run end_page_writeback() against the page before returning from
271writepage.
272
273That is: after 2.5.12, pages which are under writeout are *not* locked. Note,
274if the filesystem needs the page to be locked during writeout, that is ok, too,
275the page is allowed to be unlocked at any point in time between the calls to
276set_page_writeback() and end_page_writeback().
277
278Note, failure to run either redirty_page_for_writepage() or the combination of
279set_page_writeback()/end_page_writeback() on a page submitted to writepage
280will leave the page itself marked clean but it will be tagged as dirty in the
281radix tree. This incoherency can lead to all sorts of hard-to-debug problems
282in the filesystem like having dirty inodes at umount and losing written data.
283
284 ->sync_page() locking rules are not well-defined - usually it is called
285with lock on page, but that is not guaranteed. Considering the currently
286existing instances of this method ->sync_page() itself doesn't look
287well-defined...
288
289 ->writepages() is used for periodic writeback and for syscall-initiated
290sync operations. The address_space should start I/O against at least
291*nr_to_write pages. *nr_to_write must be decremented for each page which is
292written. The address_space implementation may write more (or less) pages
293than *nr_to_write asks for, but it should try to be reasonably close. If
294nr_to_write is NULL, all dirty pages must be written.
295
296writepages should _only_ write pages which are present on
297mapping->io_pages.
298
299 ->set_page_dirty() is called from various places in the kernel
300when the target page is marked as needing writeback. It may be called
301under spinlock (it cannot block) and is sometimes called with the page
302not locked.
303
304 ->bmap() is currently used by legacy ioctl() (FIBMAP) provided by some
Christoph Hellwigb83be6f2010-12-16 12:04:54 +0100305filesystems and by the swapper. The latter will eventually go away. Please,
306keep it that way and don't breed new callers.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700307
308 ->invalidatepage() is called when the filesystem must attempt to drop
309some or all of the buffers from the page when it is being truncated. It
310returns zero on success. If ->invalidatepage is zero, the kernel uses
311block_invalidatepage() instead.
312
313 ->releasepage() is called when the kernel is about to try to drop the
314buffers from the page in preparation for freeing it. It returns zero to
315indicate that the buffers are (or may be) freeable. If ->releasepage is zero,
316the kernel assumes that the fs has no private interest in the buffers.
317
Linus Torvalds6072d132010-12-01 13:35:19 -0500318 ->freepage() is called when the kernel is done dropping the page
319from the page cache.
320
Trond Myklebuste3db7692007-01-10 23:15:39 -0800321 ->launder_page() may be called prior to releasing a page if
322it is still found to be dirty. It returns zero if the page was successfully
323cleaned, or an error value if not. Note that in order to prevent the page
324getting mapped back in and redirtied, it needs to be kept locked
325across the entire operation.
326
Linus Torvalds1da177e2005-04-16 15:20:36 -0700327----------------------- file_lock_operations ------------------------------
328prototypes:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700329 void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
330 void (*fl_release_private)(struct file_lock *);
331
332
333locking rules:
Christoph Hellwigb83be6f2010-12-16 12:04:54 +0100334 file_lock_lock may block
335fl_copy_lock: yes no
336fl_release_private: maybe no
Linus Torvalds1da177e2005-04-16 15:20:36 -0700337
338----------------------- lock_manager_operations ---------------------------
339prototypes:
J. Bruce Fields8fb47a42011-07-20 20:21:59 -0400340 int (*lm_compare_owner)(struct file_lock *, struct file_lock *);
341 void (*lm_notify)(struct file_lock *); /* unblock callback */
342 int (*lm_grant)(struct file_lock *, struct file_lock *, int);
343 void (*lm_release_private)(struct file_lock *);
344 void (*lm_break)(struct file_lock *); /* break_lease callback */
345 int (*lm_change)(struct file_lock **, int);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700346
347locking rules:
Christoph Hellwigb83be6f2010-12-16 12:04:54 +0100348 file_lock_lock may block
J. Bruce Fields8fb47a42011-07-20 20:21:59 -0400349lm_compare_owner: yes no
350lm_notify: yes no
351lm_grant: no no
352lm_release_private: maybe no
353lm_break: yes no
354lm_change yes no
Linus Torvalds1da177e2005-04-16 15:20:36 -0700355
Linus Torvalds1da177e2005-04-16 15:20:36 -0700356--------------------------- buffer_head -----------------------------------
357prototypes:
358 void (*b_end_io)(struct buffer_head *bh, int uptodate);
359
360locking rules:
361 called from interrupts. In other words, extreme care is needed here.
362bh is locked, but that's all warranties we have here. Currently only RAID1,
363highmem, fs/buffer.c, and fs/ntfs/aops.c are providing these. Block devices
364call this method upon the IO completion.
365
366--------------------------- block_device_operations -----------------------
367prototypes:
Christoph Hellwige1455d12010-10-06 10:46:53 +0200368 int (*open) (struct block_device *, fmode_t);
369 int (*release) (struct gendisk *, fmode_t);
370 int (*ioctl) (struct block_device *, fmode_t, unsigned, unsigned long);
371 int (*compat_ioctl) (struct block_device *, fmode_t, unsigned, unsigned long);
372 int (*direct_access) (struct block_device *, sector_t, void **, unsigned long *);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700373 int (*media_changed) (struct gendisk *);
Christoph Hellwige1455d12010-10-06 10:46:53 +0200374 void (*unlock_native_capacity) (struct gendisk *);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700375 int (*revalidate_disk) (struct gendisk *);
Christoph Hellwige1455d12010-10-06 10:46:53 +0200376 int (*getgeo)(struct block_device *, struct hd_geometry *);
377 void (*swap_slot_free_notify) (struct block_device *, unsigned long);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700378
379locking rules:
Christoph Hellwigb83be6f2010-12-16 12:04:54 +0100380 bd_mutex
381open: yes
382release: yes
383ioctl: no
384compat_ioctl: no
385direct_access: no
386media_changed: no
387unlock_native_capacity: no
388revalidate_disk: no
389getgeo: no
390swap_slot_free_notify: no (see below)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700391
Christoph Hellwige1455d12010-10-06 10:46:53 +0200392media_changed, unlock_native_capacity and revalidate_disk are called only from
393check_disk_change().
394
395swap_slot_free_notify is called with swap_lock and sometimes the page lock
396held.
397
Linus Torvalds1da177e2005-04-16 15:20:36 -0700398
399--------------------------- file_operations -------------------------------
400prototypes:
401 loff_t (*llseek) (struct file *, loff_t, int);
402 ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700403 ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
Badari Pulavarty027445c2006-09-30 23:28:46 -0700404 ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
405 ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700406 int (*readdir) (struct file *, void *, filldir_t);
407 unsigned int (*poll) (struct file *, struct poll_table_struct *);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700408 long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
409 long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
410 int (*mmap) (struct file *, struct vm_area_struct *);
411 int (*open) (struct inode *, struct file *);
412 int (*flush) (struct file *);
413 int (*release) (struct inode *, struct file *);
Josef Bacik02c24a82011-07-16 20:44:56 -0400414 int (*fsync) (struct file *, loff_t start, loff_t end, int datasync);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700415 int (*aio_fsync) (struct kiocb *, int datasync);
416 int (*fasync) (int, struct file *, int);
417 int (*lock) (struct file *, int, struct file_lock *);
418 ssize_t (*readv) (struct file *, const struct iovec *, unsigned long,
419 loff_t *);
420 ssize_t (*writev) (struct file *, const struct iovec *, unsigned long,
421 loff_t *);
422 ssize_t (*sendfile) (struct file *, loff_t *, size_t, read_actor_t,
423 void __user *);
424 ssize_t (*sendpage) (struct file *, struct page *, int, size_t,
425 loff_t *, int);
426 unsigned long (*get_unmapped_area)(struct file *, unsigned long,
427 unsigned long, unsigned long, unsigned long);
428 int (*check_flags)(int);
Christoph Hellwigb83be6f2010-12-16 12:04:54 +0100429 int (*flock) (struct file *, int, struct file_lock *);
430 ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *,
431 size_t, unsigned int);
432 ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *,
433 size_t, unsigned int);
434 int (*setlease)(struct file *, long, struct file_lock **);
Christoph Hellwig2fe17c12011-01-14 13:07:43 +0100435 long (*fallocate)(struct file *, int, loff_t, loff_t);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700436};
437
438locking rules:
Christoph Hellwigb83be6f2010-12-16 12:04:54 +0100439 All may block except for ->setlease.
Josef Bacik02c24a82011-07-16 20:44:56 -0400440 No VFS locks held on entry except for ->setlease.
Christoph Hellwigb83be6f2010-12-16 12:04:54 +0100441
442->setlease has the file_list_lock held and must not sleep.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700443
444->llseek() locking has moved from llseek to the individual llseek
445implementations. If your fs is not using generic_file_llseek, you
446need to acquire and release the appropriate locks in your ->llseek().
447For many filesystems, it is probably safe to acquire the inode
Jan Blunck866707f2010-05-26 14:44:54 -0700448mutex or just to use i_size_read() instead.
449Note: this does not protect the file->f_pos against concurrent modifications
450since this is something the userspace has to take care about.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700451
Christoph Hellwigb83be6f2010-12-16 12:04:54 +0100452->fasync() is responsible for maintaining the FASYNC bit in filp->f_flags.
453Most instances call fasync_helper(), which does that maintenance, so it's
454not normally something one needs to worry about. Return values > 0 will be
455mapped to zero in the VFS layer.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700456
457->readdir() and ->ioctl() on directories must be changed. Ideally we would
458move ->readdir() to inode_operations and use a separate method for directory
459->ioctl() or kill the latter completely. One of the problems is that for
460anything that resembles union-mount we won't have a struct file for all
461components. And there are other reasons why the current interface is a mess...
462
Linus Torvalds1da177e2005-04-16 15:20:36 -0700463->read on directories probably must go away - we should just enforce -EISDIR
464in sys_read() and friends.
465
Linus Torvalds1da177e2005-04-16 15:20:36 -0700466--------------------------- dquot_operations -------------------------------
467prototypes:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700468 int (*write_dquot) (struct dquot *);
469 int (*acquire_dquot) (struct dquot *);
470 int (*release_dquot) (struct dquot *);
471 int (*mark_dirty) (struct dquot *);
472 int (*write_info) (struct super_block *, int);
473
474These operations are intended to be more or less wrapping functions that ensure
475a proper locking wrt the filesystem and call the generic quota operations.
476
477What filesystem should expect from the generic quota functions:
478
479 FS recursion Held locks when called
Linus Torvalds1da177e2005-04-16 15:20:36 -0700480write_dquot: yes dqonoff_sem or dqptr_sem
481acquire_dquot: yes dqonoff_sem or dqptr_sem
482release_dquot: yes dqonoff_sem or dqptr_sem
483mark_dirty: no -
484write_info: yes dqonoff_sem
485
486FS recursion means calling ->quota_read() and ->quota_write() from superblock
487operations.
488
Linus Torvalds1da177e2005-04-16 15:20:36 -0700489More details about quota locking can be found in fs/dquot.c.
490
491--------------------------- vm_operations_struct -----------------------------
492prototypes:
493 void (*open)(struct vm_area_struct*);
494 void (*close)(struct vm_area_struct*);
Nick Piggind0217ac2007-07-19 01:47:03 -0700495 int (*fault)(struct vm_area_struct*, struct vm_fault *);
Nick Pigginc2ec1752009-03-31 15:23:21 -0700496 int (*page_mkwrite)(struct vm_area_struct *, struct vm_fault *);
Rik van Riel28b2ee22008-07-23 21:27:05 -0700497 int (*access)(struct vm_area_struct *, unsigned long, void*, int, int);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700498
499locking rules:
Christoph Hellwigb83be6f2010-12-16 12:04:54 +0100500 mmap_sem PageLocked(page)
501open: yes
502close: yes
503fault: yes can return with page locked
504page_mkwrite: yes can return with page locked
505access: yes
Mark Fashehed2f2f92007-07-19 01:47:01 -0700506
Nick Pigginb827e492009-04-30 15:08:16 -0700507 ->fault() is called when a previously not present pte is about
508to be faulted in. The filesystem must find and return the page associated
509with the passed in "pgoff" in the vm_fault structure. If it is possible that
510the page may be truncated and/or invalidated, then the filesystem must lock
511the page, then ensure it is not already truncated (the page lock will block
512subsequent truncate), and then return with VM_FAULT_LOCKED, and the page
513locked. The VM will unlock the page.
514
515 ->page_mkwrite() is called when a previously read-only pte is
516about to become writeable. The filesystem again must ensure that there are
517no truncate/invalidate races, and then return with the page locked. If
518the page has been truncated, the filesystem should not look up a new page
519like the ->fault() handler, but simply return with VM_FAULT_NOPAGE, which
520will cause the VM to retry the fault.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700521
Rik van Riel28b2ee22008-07-23 21:27:05 -0700522 ->access() is called when get_user_pages() fails in
523acces_process_vm(), typically used to debug a process through
524/proc/pid/mem or ptrace. This function is needed only for
525VM_IO | VM_PFNMAP VMAs.
526
Linus Torvalds1da177e2005-04-16 15:20:36 -0700527================================================================================
528 Dubious stuff
529
530(if you break something or notice that it is broken and do not fix it yourself
531- at least put it here)