Commit | Line | Data |
---|---|---|
1da177e4 LT |
1 | The text below describes the locking rules for VFS-related methods. |
2 | It is (believed to be) up-to-date. *Please*, if you change anything in | |
3 | prototypes or locking protocols - update this file. And update the relevant | |
4 | instances in the tree, don't leave that to maintainers of filesystems/devices/ | |
5 | etc. At the very least, put the list of dubious cases in the end of this file. | |
6 | Don't turn it into log - maintainers of out-of-the-tree code are supposed to | |
7 | be able to use diff(1). | |
8 | Thing currently missing here: socket operations. Alexey? | |
9 | ||
10 | --------------------------- dentry_operations -------------------------- | |
11 | prototypes: | |
0b728e19 | 12 | int (*d_revalidate)(struct dentry *, unsigned int); |
b1e6a015 NP |
13 | int (*d_hash)(const struct dentry *, const struct inode *, |
14 | struct qstr *); | |
621e155a NP |
15 | int (*d_compare)(const struct dentry *, const struct inode *, |
16 | const struct dentry *, const struct inode *, | |
17 | unsigned int, const char *, const struct qstr *); | |
1da177e4 LT |
18 | int (*d_delete)(struct dentry *); |
19 | void (*d_release)(struct dentry *); | |
20 | void (*d_iput)(struct dentry *, struct inode *); | |
c23fbb6b | 21 | char *(*d_dname)((struct dentry *dentry, char *buffer, int buflen); |
9875cf80 | 22 | struct vfsmount *(*d_automount)(struct path *path); |
cc53ce53 | 23 | int (*d_manage)(struct dentry *, bool); |
1da177e4 LT |
24 | |
25 | locking rules: | |
34286d66 NP |
26 | rename_lock ->d_lock may block rcu-walk |
27 | d_revalidate: no no yes (ref-walk) maybe | |
28 | d_hash no no no maybe | |
29 | d_compare: yes no no maybe | |
30 | d_delete: no yes no no | |
31 | d_release: no no yes no | |
f0023bc6 | 32 | d_prune: no yes no no |
34286d66 NP |
33 | d_iput: no no yes no |
34 | d_dname: no no no no | |
9875cf80 | 35 | d_automount: no no yes no |
ab90911f | 36 | d_manage: no no yes (ref-walk) maybe |
1da177e4 LT |
37 | |
38 | --------------------------- inode_operations --------------------------- | |
39 | prototypes: | |
ebfc3b49 | 40 | int (*create) (struct inode *,struct dentry *,umode_t, bool); |
00cd8dd3 | 41 | struct dentry * (*lookup) (struct inode *,struct dentry *, unsigned int); |
1da177e4 LT |
42 | int (*link) (struct dentry *,struct inode *,struct dentry *); |
43 | int (*unlink) (struct inode *,struct dentry *); | |
44 | int (*symlink) (struct inode *,struct dentry *,const char *); | |
18bb1db3 | 45 | int (*mkdir) (struct inode *,struct dentry *,umode_t); |
1da177e4 | 46 | int (*rmdir) (struct inode *,struct dentry *); |
1a67aafb | 47 | int (*mknod) (struct inode *,struct dentry *,umode_t,dev_t); |
1da177e4 LT |
48 | int (*rename) (struct inode *, struct dentry *, |
49 | struct inode *, struct dentry *); | |
50 | int (*readlink) (struct dentry *, char __user *,int); | |
b83be6f2 CH |
51 | void * (*follow_link) (struct dentry *, struct nameidata *); |
52 | void (*put_link) (struct dentry *, struct nameidata *, void *); | |
1da177e4 | 53 | void (*truncate) (struct inode *); |
b74c79e9 | 54 | int (*permission) (struct inode *, int, unsigned int); |
4e34e719 | 55 | int (*get_acl)(struct inode *, int); |
1da177e4 LT |
56 | int (*setattr) (struct dentry *, struct iattr *); |
57 | int (*getattr) (struct vfsmount *, struct dentry *, struct kstat *); | |
58 | int (*setxattr) (struct dentry *, const char *,const void *,size_t,int); | |
59 | ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t); | |
60 | ssize_t (*listxattr) (struct dentry *, char *, size_t); | |
61 | int (*removexattr) (struct dentry *, const char *); | |
b83be6f2 | 62 | int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start, u64 len); |
c3b2da31 | 63 | void (*update_time)(struct inode *, struct timespec *, int); |
d9585277 | 64 | int (*atomic_open)(struct inode *, struct dentry *, |
30d90494 | 65 | struct file *, unsigned open_flag, |
47237687 | 66 | umode_t create_mode, int *opened); |
1da177e4 LT |
67 | |
68 | locking rules: | |
b83be6f2 | 69 | all may block |
a7bc02f4 | 70 | i_mutex(inode) |
1da177e4 LT |
71 | lookup: yes |
72 | create: yes | |
73 | link: yes (both) | |
74 | mknod: yes | |
75 | symlink: yes | |
76 | mkdir: yes | |
77 | unlink: yes (both) | |
78 | rmdir: yes (both) (see below) | |
79 | rename: yes (all) (see below) | |
80 | readlink: no | |
81 | follow_link: no | |
b83be6f2 | 82 | put_link: no |
1da177e4 LT |
83 | truncate: yes (see below) |
84 | setattr: yes | |
b74c79e9 | 85 | permission: no (may not block if called in rcu-walk mode) |
4e34e719 | 86 | get_acl: no |
1da177e4 LT |
87 | getattr: no |
88 | setxattr: yes | |
89 | getxattr: no | |
90 | listxattr: no | |
91 | removexattr: yes | |
b83be6f2 | 92 | fiemap: no |
c3b2da31 | 93 | update_time: no |
d18e9008 | 94 | atomic_open: yes |
c3b2da31 | 95 | |
a7bc02f4 | 96 | Additionally, ->rmdir(), ->unlink() and ->rename() have ->i_mutex on |
1da177e4 LT |
97 | victim. |
98 | cross-directory ->rename() has (per-superblock) ->s_vfs_rename_sem. | |
99 | ->truncate() is never called directly - it's a callback, not a | |
b83be6f2 | 100 | method. It's called by vmtruncate() - deprecated library function used by |
1da177e4 LT |
101 | ->setattr(). Locking information above applies to that call (i.e. is |
102 | inherited from ->setattr() - vmtruncate() is used when ATTR_SIZE had been | |
103 | passed). | |
104 | ||
105 | See Documentation/filesystems/directory-locking for more detailed discussion | |
106 | of the locking scheme for directory operations. | |
107 | ||
108 | --------------------------- super_operations --------------------------- | |
109 | prototypes: | |
110 | struct inode *(*alloc_inode)(struct super_block *sb); | |
111 | void (*destroy_inode)(struct inode *); | |
aa385729 | 112 | void (*dirty_inode) (struct inode *, int flags); |
b83be6f2 | 113 | int (*write_inode) (struct inode *, struct writeback_control *wbc); |
336fb3b9 AV |
114 | int (*drop_inode) (struct inode *); |
115 | void (*evict_inode) (struct inode *); | |
1da177e4 LT |
116 | void (*put_super) (struct super_block *); |
117 | void (*write_super) (struct super_block *); | |
118 | int (*sync_fs)(struct super_block *sb, int wait); | |
c4be0c1d TS |
119 | int (*freeze_fs) (struct super_block *); |
120 | int (*unfreeze_fs) (struct super_block *); | |
726c3342 | 121 | int (*statfs) (struct dentry *, struct kstatfs *); |
1da177e4 | 122 | int (*remount_fs) (struct super_block *, int *, char *); |
1da177e4 | 123 | void (*umount_begin) (struct super_block *); |
34c80b1d | 124 | int (*show_options)(struct seq_file *, struct dentry *); |
1da177e4 LT |
125 | ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t); |
126 | ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t); | |
b83be6f2 | 127 | int (*bdev_try_to_free_page)(struct super_block*, struct page*, gfp_t); |
1da177e4 LT |
128 | |
129 | locking rules: | |
336fb3b9 | 130 | All may block [not true, see below] |
7e325d3a CH |
131 | s_umount |
132 | alloc_inode: | |
133 | destroy_inode: | |
aa385729 | 134 | dirty_inode: |
7e325d3a | 135 | write_inode: |
f283c86a | 136 | drop_inode: !!!inode->i_lock!!! |
336fb3b9 | 137 | evict_inode: |
7e325d3a CH |
138 | put_super: write |
139 | write_super: read | |
140 | sync_fs: read | |
141 | freeze_fs: read | |
142 | unfreeze_fs: read | |
336fb3b9 AV |
143 | statfs: maybe(read) (see below) |
144 | remount_fs: write | |
7e325d3a CH |
145 | umount_begin: no |
146 | show_options: no (namespace_sem) | |
147 | quota_read: no (see below) | |
148 | quota_write: no (see below) | |
b83be6f2 | 149 | bdev_try_to_free_page: no (see below) |
1da177e4 | 150 | |
336fb3b9 AV |
151 | ->statfs() has s_umount (shared) when called by ustat(2) (native or |
152 | compat), but that's an accident of bad API; s_umount is used to pin | |
153 | the superblock down when we only have dev_t given us by userland to | |
154 | identify the superblock. Everything else (statfs(), fstatfs(), etc.) | |
155 | doesn't hold it when calling ->statfs() - superblock is pinned down | |
156 | by resolving the pathname passed to syscall. | |
1da177e4 LT |
157 | ->quota_read() and ->quota_write() functions are both guaranteed to |
158 | be the only ones operating on the quota file by the quota code (via | |
159 | dqio_sem) (unless an admin really wants to screw up something and | |
160 | writes to quota files with quotas on). For other details about locking | |
161 | see also dquot_operations section. | |
b83be6f2 CH |
162 | ->bdev_try_to_free_page is called from the ->releasepage handler of |
163 | the block device inode. See there for more details. | |
1da177e4 LT |
164 | |
165 | --------------------------- file_system_type --------------------------- | |
166 | prototypes: | |
5d8b2ebf JC |
167 | int (*get_sb) (struct file_system_type *, int, |
168 | const char *, void *, struct vfsmount *); | |
b83be6f2 CH |
169 | struct dentry *(*mount) (struct file_system_type *, int, |
170 | const char *, void *); | |
1da177e4 LT |
171 | void (*kill_sb) (struct super_block *); |
172 | locking rules: | |
b83be6f2 | 173 | may block |
b83be6f2 CH |
174 | mount yes |
175 | kill_sb yes | |
1da177e4 | 176 | |
1a102ff9 AV |
177 | ->mount() returns ERR_PTR or the root dentry; its superblock should be locked |
178 | on return. | |
1da177e4 LT |
179 | ->kill_sb() takes a write-locked superblock, does all shutdown work on it, |
180 | unlocks and drops the reference. | |
181 | ||
182 | --------------------------- address_space_operations -------------------------- | |
183 | prototypes: | |
184 | int (*writepage)(struct page *page, struct writeback_control *wbc); | |
185 | int (*readpage)(struct file *, struct page *); | |
186 | int (*sync_page)(struct page *); | |
187 | int (*writepages)(struct address_space *, struct writeback_control *); | |
188 | int (*set_page_dirty)(struct page *page); | |
189 | int (*readpages)(struct file *filp, struct address_space *mapping, | |
190 | struct list_head *pages, unsigned nr_pages); | |
4e02ed4b NP |
191 | int (*write_begin)(struct file *, struct address_space *mapping, |
192 | loff_t pos, unsigned len, unsigned flags, | |
193 | struct page **pagep, void **fsdata); | |
194 | int (*write_end)(struct file *, struct address_space *mapping, | |
195 | loff_t pos, unsigned len, unsigned copied, | |
196 | struct page *page, void *fsdata); | |
1da177e4 LT |
197 | sector_t (*bmap)(struct address_space *, sector_t); |
198 | int (*invalidatepage) (struct page *, unsigned long); | |
199 | int (*releasepage) (struct page *, int); | |
6072d13c | 200 | void (*freepage)(struct page *); |
1da177e4 LT |
201 | int (*direct_IO)(int, struct kiocb *, const struct iovec *iov, |
202 | loff_t offset, unsigned long nr_segs); | |
b83be6f2 CH |
203 | int (*get_xip_mem)(struct address_space *, pgoff_t, int, void **, |
204 | unsigned long *); | |
205 | int (*migratepage)(struct address_space *, struct page *, struct page *); | |
206 | int (*launder_page)(struct page *); | |
207 | int (*is_partially_uptodate)(struct page *, read_descriptor_t *, unsigned long); | |
208 | int (*error_remove_page)(struct address_space *, struct page *); | |
62c230bc MG |
209 | int (*swap_activate)(struct file *); |
210 | int (*swap_deactivate)(struct file *); | |
1da177e4 LT |
211 | |
212 | locking rules: | |
6072d13c | 213 | All except set_page_dirty and freepage may block |
1da177e4 | 214 | |
b83be6f2 CH |
215 | PageLocked(page) i_mutex |
216 | writepage: yes, unlocks (see below) | |
217 | readpage: yes, unlocks | |
218 | sync_page: maybe | |
219 | writepages: | |
220 | set_page_dirty no | |
221 | readpages: | |
222 | write_begin: locks the page yes | |
223 | write_end: yes, unlocks yes | |
224 | bmap: | |
225 | invalidatepage: yes | |
226 | releasepage: yes | |
227 | freepage: yes | |
228 | direct_IO: | |
229 | get_xip_mem: maybe | |
230 | migratepage: yes (both) | |
231 | launder_page: yes | |
232 | is_partially_uptodate: yes | |
233 | error_remove_page: yes | |
62c230bc MG |
234 | swap_activate: no |
235 | swap_deactivate: no | |
1da177e4 | 236 | |
4e02ed4b | 237 | ->write_begin(), ->write_end(), ->sync_page() and ->readpage() |
1da177e4 LT |
238 | may be called from the request handler (/dev/loop). |
239 | ||
240 | ->readpage() unlocks the page, either synchronously or via I/O | |
241 | completion. | |
242 | ||
243 | ->readpages() populates the pagecache with the passed pages and starts | |
244 | I/O against them. They come unlocked upon I/O completion. | |
245 | ||
246 | ->writepage() is used for two purposes: for "memory cleansing" and for | |
247 | "sync". These are quite different operations and the behaviour may differ | |
248 | depending upon the mode. | |
249 | ||
250 | If writepage is called for sync (wbc->sync_mode != WBC_SYNC_NONE) then | |
251 | it *must* start I/O against the page, even if that would involve | |
252 | blocking on in-progress I/O. | |
253 | ||
254 | If writepage is called for memory cleansing (sync_mode == | |
255 | WBC_SYNC_NONE) then its role is to get as much writeout underway as | |
256 | possible. So writepage should try to avoid blocking against | |
257 | currently-in-progress I/O. | |
258 | ||
259 | If the filesystem is not called for "sync" and it determines that it | |
260 | would need to block against in-progress I/O to be able to start new I/O | |
261 | against the page the filesystem should redirty the page with | |
262 | redirty_page_for_writepage(), then unlock the page and return zero. | |
263 | This may also be done to avoid internal deadlocks, but rarely. | |
264 | ||
3a4fa0a2 | 265 | If the filesystem is called for sync then it must wait on any |
1da177e4 LT |
266 | in-progress I/O and then start new I/O. |
267 | ||
2054606a ND |
268 | The filesystem should unlock the page synchronously, before returning to the |
269 | caller, unless ->writepage() returns special WRITEPAGE_ACTIVATE | |
270 | value. WRITEPAGE_ACTIVATE means that page cannot really be written out | |
271 | currently, and VM should stop calling ->writepage() on this page for some | |
272 | time. VM does this by moving page to the head of the active list, hence the | |
273 | name. | |
1da177e4 LT |
274 | |
275 | Unless the filesystem is going to redirty_page_for_writepage(), unlock the page | |
276 | and return zero, writepage *must* run set_page_writeback() against the page, | |
277 | followed by unlocking it. Once set_page_writeback() has been run against the | |
278 | page, write I/O can be submitted and the write I/O completion handler must run | |
279 | end_page_writeback() once the I/O is complete. If no I/O is submitted, the | |
280 | filesystem must run end_page_writeback() against the page before returning from | |
281 | writepage. | |
282 | ||
283 | That is: after 2.5.12, pages which are under writeout are *not* locked. Note, | |
284 | if the filesystem needs the page to be locked during writeout, that is ok, too, | |
285 | the page is allowed to be unlocked at any point in time between the calls to | |
286 | set_page_writeback() and end_page_writeback(). | |
287 | ||
288 | Note, failure to run either redirty_page_for_writepage() or the combination of | |
289 | set_page_writeback()/end_page_writeback() on a page submitted to writepage | |
290 | will leave the page itself marked clean but it will be tagged as dirty in the | |
291 | radix tree. This incoherency can lead to all sorts of hard-to-debug problems | |
292 | in the filesystem like having dirty inodes at umount and losing written data. | |
293 | ||
294 | ->sync_page() locking rules are not well-defined - usually it is called | |
295 | with lock on page, but that is not guaranteed. Considering the currently | |
296 | existing instances of this method ->sync_page() itself doesn't look | |
297 | well-defined... | |
298 | ||
299 | ->writepages() is used for periodic writeback and for syscall-initiated | |
300 | sync operations. The address_space should start I/O against at least | |
301 | *nr_to_write pages. *nr_to_write must be decremented for each page which is | |
302 | written. The address_space implementation may write more (or less) pages | |
303 | than *nr_to_write asks for, but it should try to be reasonably close. If | |
304 | nr_to_write is NULL, all dirty pages must be written. | |
305 | ||
306 | writepages should _only_ write pages which are present on | |
307 | mapping->io_pages. | |
308 | ||
309 | ->set_page_dirty() is called from various places in the kernel | |
310 | when the target page is marked as needing writeback. It may be called | |
311 | under spinlock (it cannot block) and is sometimes called with the page | |
312 | not locked. | |
313 | ||
314 | ->bmap() is currently used by legacy ioctl() (FIBMAP) provided by some | |
b83be6f2 CH |
315 | filesystems and by the swapper. The latter will eventually go away. Please, |
316 | keep it that way and don't breed new callers. | |
1da177e4 LT |
317 | |
318 | ->invalidatepage() is called when the filesystem must attempt to drop | |
319 | some or all of the buffers from the page when it is being truncated. It | |
320 | returns zero on success. If ->invalidatepage is zero, the kernel uses | |
321 | block_invalidatepage() instead. | |
322 | ||
323 | ->releasepage() is called when the kernel is about to try to drop the | |
324 | buffers from the page in preparation for freeing it. It returns zero to | |
325 | indicate that the buffers are (or may be) freeable. If ->releasepage is zero, | |
326 | the kernel assumes that the fs has no private interest in the buffers. | |
327 | ||
6072d13c LT |
328 | ->freepage() is called when the kernel is done dropping the page |
329 | from the page cache. | |
330 | ||
e3db7691 TM |
331 | ->launder_page() may be called prior to releasing a page if |
332 | it is still found to be dirty. It returns zero if the page was successfully | |
333 | cleaned, or an error value if not. Note that in order to prevent the page | |
334 | getting mapped back in and redirtied, it needs to be kept locked | |
335 | across the entire operation. | |
336 | ||
62c230bc MG |
337 | ->swap_activate will be called with a non-zero argument on |
338 | files backing (non block device backed) swapfiles. A return value | |
339 | of zero indicates success, in which case this file can be used for | |
340 | backing swapspace. The swapspace operations will be proxied to the | |
341 | address space operations. | |
342 | ||
343 | ->swap_deactivate() will be called in the sys_swapoff() | |
344 | path after ->swap_activate() returned success. | |
345 | ||
1da177e4 LT |
346 | ----------------------- file_lock_operations ------------------------------ |
347 | prototypes: | |
1da177e4 LT |
348 | void (*fl_copy_lock)(struct file_lock *, struct file_lock *); |
349 | void (*fl_release_private)(struct file_lock *); | |
350 | ||
351 | ||
352 | locking rules: | |
b83be6f2 CH |
353 | file_lock_lock may block |
354 | fl_copy_lock: yes no | |
355 | fl_release_private: maybe no | |
1da177e4 LT |
356 | |
357 | ----------------------- lock_manager_operations --------------------------- | |
358 | prototypes: | |
8fb47a4f BF |
359 | int (*lm_compare_owner)(struct file_lock *, struct file_lock *); |
360 | void (*lm_notify)(struct file_lock *); /* unblock callback */ | |
361 | int (*lm_grant)(struct file_lock *, struct file_lock *, int); | |
362 | void (*lm_release_private)(struct file_lock *); | |
363 | void (*lm_break)(struct file_lock *); /* break_lease callback */ | |
364 | int (*lm_change)(struct file_lock **, int); | |
1da177e4 LT |
365 | |
366 | locking rules: | |
b83be6f2 | 367 | file_lock_lock may block |
8fb47a4f BF |
368 | lm_compare_owner: yes no |
369 | lm_notify: yes no | |
370 | lm_grant: no no | |
371 | lm_release_private: maybe no | |
372 | lm_break: yes no | |
373 | lm_change yes no | |
b83be6f2 | 374 | |
1da177e4 LT |
375 | --------------------------- buffer_head ----------------------------------- |
376 | prototypes: | |
377 | void (*b_end_io)(struct buffer_head *bh, int uptodate); | |
378 | ||
379 | locking rules: | |
380 | called from interrupts. In other words, extreme care is needed here. | |
381 | bh is locked, but that's all warranties we have here. Currently only RAID1, | |
382 | highmem, fs/buffer.c, and fs/ntfs/aops.c are providing these. Block devices | |
383 | call this method upon the IO completion. | |
384 | ||
385 | --------------------------- block_device_operations ----------------------- | |
386 | prototypes: | |
e1455d1b CH |
387 | int (*open) (struct block_device *, fmode_t); |
388 | int (*release) (struct gendisk *, fmode_t); | |
389 | int (*ioctl) (struct block_device *, fmode_t, unsigned, unsigned long); | |
390 | int (*compat_ioctl) (struct block_device *, fmode_t, unsigned, unsigned long); | |
391 | int (*direct_access) (struct block_device *, sector_t, void **, unsigned long *); | |
1da177e4 | 392 | int (*media_changed) (struct gendisk *); |
e1455d1b | 393 | void (*unlock_native_capacity) (struct gendisk *); |
1da177e4 | 394 | int (*revalidate_disk) (struct gendisk *); |
e1455d1b CH |
395 | int (*getgeo)(struct block_device *, struct hd_geometry *); |
396 | void (*swap_slot_free_notify) (struct block_device *, unsigned long); | |
1da177e4 LT |
397 | |
398 | locking rules: | |
b83be6f2 CH |
399 | bd_mutex |
400 | open: yes | |
401 | release: yes | |
402 | ioctl: no | |
403 | compat_ioctl: no | |
404 | direct_access: no | |
405 | media_changed: no | |
406 | unlock_native_capacity: no | |
407 | revalidate_disk: no | |
408 | getgeo: no | |
409 | swap_slot_free_notify: no (see below) | |
e1455d1b CH |
410 | |
411 | media_changed, unlock_native_capacity and revalidate_disk are called only from | |
412 | check_disk_change(). | |
413 | ||
414 | swap_slot_free_notify is called with swap_lock and sometimes the page lock | |
415 | held. | |
1da177e4 | 416 | |
1da177e4 LT |
417 | |
418 | --------------------------- file_operations ------------------------------- | |
419 | prototypes: | |
420 | loff_t (*llseek) (struct file *, loff_t, int); | |
421 | ssize_t (*read) (struct file *, char __user *, size_t, loff_t *); | |
1da177e4 | 422 | ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *); |
027445c3 BP |
423 | ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t); |
424 | ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t); | |
1da177e4 LT |
425 | int (*readdir) (struct file *, void *, filldir_t); |
426 | unsigned int (*poll) (struct file *, struct poll_table_struct *); | |
1da177e4 LT |
427 | long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long); |
428 | long (*compat_ioctl) (struct file *, unsigned int, unsigned long); | |
429 | int (*mmap) (struct file *, struct vm_area_struct *); | |
430 | int (*open) (struct inode *, struct file *); | |
431 | int (*flush) (struct file *); | |
432 | int (*release) (struct inode *, struct file *); | |
02c24a82 | 433 | int (*fsync) (struct file *, loff_t start, loff_t end, int datasync); |
1da177e4 LT |
434 | int (*aio_fsync) (struct kiocb *, int datasync); |
435 | int (*fasync) (int, struct file *, int); | |
436 | int (*lock) (struct file *, int, struct file_lock *); | |
437 | ssize_t (*readv) (struct file *, const struct iovec *, unsigned long, | |
438 | loff_t *); | |
439 | ssize_t (*writev) (struct file *, const struct iovec *, unsigned long, | |
440 | loff_t *); | |
441 | ssize_t (*sendfile) (struct file *, loff_t *, size_t, read_actor_t, | |
442 | void __user *); | |
443 | ssize_t (*sendpage) (struct file *, struct page *, int, size_t, | |
444 | loff_t *, int); | |
445 | unsigned long (*get_unmapped_area)(struct file *, unsigned long, | |
446 | unsigned long, unsigned long, unsigned long); | |
447 | int (*check_flags)(int); | |
b83be6f2 CH |
448 | int (*flock) (struct file *, int, struct file_lock *); |
449 | ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, | |
450 | size_t, unsigned int); | |
451 | ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, | |
452 | size_t, unsigned int); | |
453 | int (*setlease)(struct file *, long, struct file_lock **); | |
2fe17c10 | 454 | long (*fallocate)(struct file *, int, loff_t, loff_t); |
1da177e4 LT |
455 | }; |
456 | ||
457 | locking rules: | |
b83be6f2 | 458 | All may block except for ->setlease. |
02c24a82 | 459 | No VFS locks held on entry except for ->setlease. |
b83be6f2 CH |
460 | |
461 | ->setlease has the file_list_lock held and must not sleep. | |
1da177e4 LT |
462 | |
463 | ->llseek() locking has moved from llseek to the individual llseek | |
464 | implementations. If your fs is not using generic_file_llseek, you | |
465 | need to acquire and release the appropriate locks in your ->llseek(). | |
466 | For many filesystems, it is probably safe to acquire the inode | |
866707fc JB |
467 | mutex or just to use i_size_read() instead. |
468 | Note: this does not protect the file->f_pos against concurrent modifications | |
469 | since this is something the userspace has to take care about. | |
1da177e4 | 470 | |
b83be6f2 CH |
471 | ->fasync() is responsible for maintaining the FASYNC bit in filp->f_flags. |
472 | Most instances call fasync_helper(), which does that maintenance, so it's | |
473 | not normally something one needs to worry about. Return values > 0 will be | |
474 | mapped to zero in the VFS layer. | |
1da177e4 LT |
475 | |
476 | ->readdir() and ->ioctl() on directories must be changed. Ideally we would | |
477 | move ->readdir() to inode_operations and use a separate method for directory | |
478 | ->ioctl() or kill the latter completely. One of the problems is that for | |
479 | anything that resembles union-mount we won't have a struct file for all | |
480 | components. And there are other reasons why the current interface is a mess... | |
481 | ||
1da177e4 LT |
482 | ->read on directories probably must go away - we should just enforce -EISDIR |
483 | in sys_read() and friends. | |
484 | ||
1da177e4 LT |
485 | --------------------------- dquot_operations ------------------------------- |
486 | prototypes: | |
1da177e4 LT |
487 | int (*write_dquot) (struct dquot *); |
488 | int (*acquire_dquot) (struct dquot *); | |
489 | int (*release_dquot) (struct dquot *); | |
490 | int (*mark_dirty) (struct dquot *); | |
491 | int (*write_info) (struct super_block *, int); | |
492 | ||
493 | These operations are intended to be more or less wrapping functions that ensure | |
494 | a proper locking wrt the filesystem and call the generic quota operations. | |
495 | ||
496 | What filesystem should expect from the generic quota functions: | |
497 | ||
498 | FS recursion Held locks when called | |
1da177e4 LT |
499 | write_dquot: yes dqonoff_sem or dqptr_sem |
500 | acquire_dquot: yes dqonoff_sem or dqptr_sem | |
501 | release_dquot: yes dqonoff_sem or dqptr_sem | |
502 | mark_dirty: no - | |
503 | write_info: yes dqonoff_sem | |
504 | ||
505 | FS recursion means calling ->quota_read() and ->quota_write() from superblock | |
506 | operations. | |
507 | ||
1da177e4 LT |
508 | More details about quota locking can be found in fs/dquot.c. |
509 | ||
510 | --------------------------- vm_operations_struct ----------------------------- | |
511 | prototypes: | |
512 | void (*open)(struct vm_area_struct*); | |
513 | void (*close)(struct vm_area_struct*); | |
d0217ac0 | 514 | int (*fault)(struct vm_area_struct*, struct vm_fault *); |
c2ec175c | 515 | int (*page_mkwrite)(struct vm_area_struct *, struct vm_fault *); |
28b2ee20 | 516 | int (*access)(struct vm_area_struct *, unsigned long, void*, int, int); |
1da177e4 LT |
517 | |
518 | locking rules: | |
b83be6f2 CH |
519 | mmap_sem PageLocked(page) |
520 | open: yes | |
521 | close: yes | |
522 | fault: yes can return with page locked | |
523 | page_mkwrite: yes can return with page locked | |
524 | access: yes | |
ed2f2f9b | 525 | |
b827e496 NP |
526 | ->fault() is called when a previously not present pte is about |
527 | to be faulted in. The filesystem must find and return the page associated | |
528 | with the passed in "pgoff" in the vm_fault structure. If it is possible that | |
529 | the page may be truncated and/or invalidated, then the filesystem must lock | |
530 | the page, then ensure it is not already truncated (the page lock will block | |
531 | subsequent truncate), and then return with VM_FAULT_LOCKED, and the page | |
532 | locked. The VM will unlock the page. | |
533 | ||
534 | ->page_mkwrite() is called when a previously read-only pte is | |
535 | about to become writeable. The filesystem again must ensure that there are | |
536 | no truncate/invalidate races, and then return with the page locked. If | |
537 | the page has been truncated, the filesystem should not look up a new page | |
538 | like the ->fault() handler, but simply return with VM_FAULT_NOPAGE, which | |
539 | will cause the VM to retry the fault. | |
1da177e4 | 540 | |
28b2ee20 RR |
541 | ->access() is called when get_user_pages() fails in |
542 | acces_process_vm(), typically used to debug a process through | |
543 | /proc/pid/mem or ptrace. This function is needed only for | |
544 | VM_IO | VM_PFNMAP VMAs. | |
545 | ||
1da177e4 LT |
546 | ================================================================================ |
547 | Dubious stuff | |
548 | ||
549 | (if you break something or notice that it is broken and do not fix it yourself | |
550 | - at least put it here) |