Blame view
Documentation/filesystems/Locking
24 KB
1da177e4c Linux-2.6.12-rc2 |
1 2 3 4 5 6 7 8 9 10 11 |
The text below describes the locking rules for VFS-related methods. It is (believed to be) up-to-date. *Please*, if you change anything in prototypes or locking protocols - update this file. And update the relevant instances in the tree, don't leave that to maintainers of filesystems/devices/ etc. At the very least, put the list of dubious cases in the end of this file. Don't turn it into log - maintainers of out-of-the-tree code are supposed to be able to use diff(1). Thing currently missing here: socket operations. Alexey? --------------------------- dentry_operations -------------------------- prototypes: |
0b728e191 stop passing name... |
12 |
int (*d_revalidate)(struct dentry *, unsigned int); |
ecf3d1f1a vfs: kill FS_REVA... |
13 |
int (*d_weak_revalidate)(struct dentry *, unsigned int); |
da53be12b Don't pass inode ... |
14 |
int (*d_hash)(const struct dentry *, struct qstr *); |
6fa67e707 get rid of 'paren... |
15 |
int (*d_compare)(const struct dentry *, |
621e155a3 fs: change d_comp... |
16 |
unsigned int, const char *, const struct qstr *); |
1da177e4c Linux-2.6.12-rc2 |
17 |
int (*d_delete)(struct dentry *); |
285b102d3 vfs: new d_init m... |
18 |
int (*d_init)(struct dentry *); |
1da177e4c Linux-2.6.12-rc2 |
19 20 |
void (*d_release)(struct dentry *); void (*d_iput)(struct dentry *, struct inode *); |
c23fbb6bc VFS: delay the de... |
21 |
char *(*d_dname)((struct dentry *dentry, char *buffer, int buflen); |
9875cf806 Add a dentry op t... |
22 |
struct vfsmount *(*d_automount)(struct path *path); |
fb5f51c74 vfs: change d_man... |
23 |
int (*d_manage)(const struct path *, bool); |
fb16043b4 vfs: remove open_... |
24 |
struct dentry *(*d_real)(struct dentry *, const struct inode *); |
1da177e4c Linux-2.6.12-rc2 |
25 26 |
locking rules: |
34286d666 fs: rcu-walk awar... |
27 28 |
rename_lock ->d_lock may block rcu-walk d_revalidate: no no yes (ref-walk) maybe |
ecf3d1f1a vfs: kill FS_REVA... |
29 |
d_weak_revalidate:no no yes no |
34286d666 fs: rcu-walk awar... |
30 31 32 |
d_hash no no no maybe d_compare: yes no no maybe d_delete: no yes no no |
285b102d3 vfs: new d_init m... |
33 |
d_init: no no yes no |
34286d666 fs: rcu-walk awar... |
34 |
d_release: no no yes no |
f0023bc61 vfs: add d_prune ... |
35 |
d_prune: no yes no no |
34286d666 fs: rcu-walk awar... |
36 37 |
d_iput: no no yes no d_dname: no no no no |
9875cf806 Add a dentry op t... |
38 |
d_automount: no no yes no |
ab90911ff Allow d_manage() ... |
39 |
d_manage: no no yes (ref-walk) maybe |
e698b8a43 vfs: document ->d... |
40 |
d_real no no yes no |
1da177e4c Linux-2.6.12-rc2 |
41 42 43 |
--------------------------- inode_operations --------------------------- prototypes: |
ebfc3b49a don't pass nameid... |
44 |
int (*create) (struct inode *,struct dentry *,umode_t, bool); |
00cd8dd3b stop passing name... |
45 |
struct dentry * (*lookup) (struct inode *,struct dentry *, unsigned int); |
1da177e4c Linux-2.6.12-rc2 |
46 47 48 |
int (*link) (struct dentry *,struct inode *,struct dentry *); int (*unlink) (struct inode *,struct dentry *); int (*symlink) (struct inode *,struct dentry *,const char *); |
18bb1db3e switch vfs_mkdir(... |
49 |
int (*mkdir) (struct inode *,struct dentry *,umode_t); |
1da177e4c Linux-2.6.12-rc2 |
50 |
int (*rmdir) (struct inode *,struct dentry *); |
1a67aafb5 switch ->mknod() ... |
51 |
int (*mknod) (struct inode *,struct dentry *,umode_t,dev_t); |
1da177e4c Linux-2.6.12-rc2 |
52 |
int (*rename) (struct inode *, struct dentry *, |
520c8b165 vfs: add renameat... |
53 |
struct inode *, struct dentry *, unsigned int); |
1da177e4c Linux-2.6.12-rc2 |
54 |
int (*readlink) (struct dentry *, char __user *,int); |
6b2553918 replace ->follow_... |
55 |
const char *(*get_link) (struct dentry *, struct inode *, void **); |
1da177e4c Linux-2.6.12-rc2 |
56 |
void (*truncate) (struct inode *); |
b74c79e99 fs: provide rcu-w... |
57 |
int (*permission) (struct inode *, int, unsigned int); |
4e34e719e fs: take the ACL ... |
58 |
int (*get_acl)(struct inode *, int); |
1da177e4c Linux-2.6.12-rc2 |
59 |
int (*setattr) (struct dentry *, struct iattr *); |
75dd7e4bb Documentation/fil... |
60 |
int (*getattr) (const struct path *, struct kstat *, u32, unsigned int); |
1da177e4c Linux-2.6.12-rc2 |
61 |
ssize_t (*listxattr) (struct dentry *, char *, size_t); |
b83be6f20 update Documentat... |
62 |
int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start, u64 len); |
c3b2da314 fs: introduce ino... |
63 |
void (*update_time)(struct inode *, struct timespec *, int); |
d95852777 make ->atomic_ope... |
64 |
int (*atomic_open)(struct inode *, struct dentry *, |
30d904947 kill struct opendata |
65 |
struct file *, unsigned open_flag, |
6c9b1de17 document ->atomic... |
66 |
umode_t create_mode); |
48bde8d36 Document ->tmpfile() |
67 |
int (*tmpfile) (struct inode *, struct dentry *, umode_t); |
1da177e4c Linux-2.6.12-rc2 |
68 69 |
locking rules: |
b83be6f20 update Documentat... |
70 |
all may block |
965de0ec3 Documentation: fi... |
71 72 73 74 75 76 77 78 79 80 |
i_rwsem(inode) lookup: shared create: exclusive link: exclusive (both) mknod: exclusive symlink: exclusive mkdir: exclusive unlink: exclusive (both) rmdir: exclusive (both)(see below) rename: exclusive (all) (see below) |
1da177e4c Linux-2.6.12-rc2 |
81 |
readlink: no |
6b2553918 replace ->follow_... |
82 |
get_link: no |
965de0ec3 Documentation: fi... |
83 |
setattr: exclusive |
b74c79e99 fs: provide rcu-w... |
84 |
permission: no (may not block if called in rcu-walk mode) |
4e34e719e fs: take the ACL ... |
85 |
get_acl: no |
1da177e4c Linux-2.6.12-rc2 |
86 |
getattr: no |
1da177e4c Linux-2.6.12-rc2 |
87 |
listxattr: no |
b83be6f20 update Documentat... |
88 |
fiemap: no |
c3b2da314 fs: introduce ino... |
89 |
update_time: no |
965de0ec3 Documentation: fi... |
90 |
atomic_open: exclusive |
48bde8d36 Document ->tmpfile() |
91 |
tmpfile: no |
c3b2da314 fs: introduce ino... |
92 |
|
6c6ef9f26 xattr: Stop calli... |
93 |
|
965de0ec3 Documentation: fi... |
94 95 |
Additionally, ->rmdir(), ->unlink() and ->rename() have ->i_rwsem exclusive on victim. |
2773bf00a fs: rename "renam... |
96 |
cross-directory ->rename() has (per-superblock) ->s_vfs_rename_sem. |
1da177e4c Linux-2.6.12-rc2 |
97 98 99 |
See Documentation/filesystems/directory-locking for more detailed discussion of the locking scheme for directory operations. |
6c6ef9f26 xattr: Stop calli... |
100 101 102 103 104 105 106 107 108 109 110 111 |
----------------------- xattr_handler operations ----------------------- prototypes: bool (*list)(struct dentry *dentry); int (*get)(const struct xattr_handler *handler, struct dentry *dentry, struct inode *inode, const char *name, void *buffer, size_t size); int (*set)(const struct xattr_handler *handler, struct dentry *dentry, struct inode *inode, const char *name, const void *buffer, size_t size, int flags); locking rules: all may block |
965de0ec3 Documentation: fi... |
112 |
i_rwsem(inode) |
6c6ef9f26 xattr: Stop calli... |
113 114 |
list: no get: no |
965de0ec3 Documentation: fi... |
115 |
set: exclusive |
6c6ef9f26 xattr: Stop calli... |
116 |
|
1da177e4c Linux-2.6.12-rc2 |
117 118 119 120 |
--------------------------- super_operations --------------------------- prototypes: struct inode *(*alloc_inode)(struct super_block *sb); void (*destroy_inode)(struct inode *); |
aa3857295 fs: pass exact ty... |
121 |
void (*dirty_inode) (struct inode *, int flags); |
b83be6f20 update Documentat... |
122 |
int (*write_inode) (struct inode *, struct writeback_control *wbc); |
336fb3b97 update VFS docume... |
123 124 |
int (*drop_inode) (struct inode *); void (*evict_inode) (struct inode *); |
1da177e4c Linux-2.6.12-rc2 |
125 |
void (*put_super) (struct super_block *); |
1da177e4c Linux-2.6.12-rc2 |
126 |
int (*sync_fs)(struct super_block *sb, int wait); |
c4be0c1dc filesystem freeze... |
127 128 |
int (*freeze_fs) (struct super_block *); int (*unfreeze_fs) (struct super_block *); |
726c33422 [PATCH] VFS: Perm... |
129 |
int (*statfs) (struct dentry *, struct kstatfs *); |
1da177e4c Linux-2.6.12-rc2 |
130 |
int (*remount_fs) (struct super_block *, int *, char *); |
1da177e4c Linux-2.6.12-rc2 |
131 |
void (*umount_begin) (struct super_block *); |
34c80b1d9 vfs: switch ->sho... |
132 |
int (*show_options)(struct seq_file *, struct dentry *); |
1da177e4c Linux-2.6.12-rc2 |
133 134 |
ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t); ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t); |
b83be6f20 update Documentat... |
135 |
int (*bdev_try_to_free_page)(struct super_block*, struct page*, gfp_t); |
1da177e4c Linux-2.6.12-rc2 |
136 137 |
locking rules: |
336fb3b97 update VFS docume... |
138 |
All may block [not true, see below] |
7e325d3a6 update Documentat... |
139 140 141 |
s_umount alloc_inode: destroy_inode: |
aa3857295 fs: pass exact ty... |
142 |
dirty_inode: |
7e325d3a6 update Documentat... |
143 |
write_inode: |
f283c86af fs: remove inode_... |
144 |
drop_inode: !!!inode->i_lock!!! |
336fb3b97 update VFS docume... |
145 |
evict_inode: |
7e325d3a6 update Documentat... |
146 |
put_super: write |
7e325d3a6 update Documentat... |
147 |
sync_fs: read |
06fd516c1 Documentation: Co... |
148 149 |
freeze_fs: write unfreeze_fs: write |
336fb3b97 update VFS docume... |
150 151 |
statfs: maybe(read) (see below) remount_fs: write |
7e325d3a6 update Documentat... |
152 153 154 155 |
umount_begin: no show_options: no (namespace_sem) quota_read: no (see below) quota_write: no (see below) |
b83be6f20 update Documentat... |
156 |
bdev_try_to_free_page: no (see below) |
1da177e4c Linux-2.6.12-rc2 |
157 |
|
336fb3b97 update VFS docume... |
158 159 160 161 162 163 |
->statfs() has s_umount (shared) when called by ustat(2) (native or compat), but that's an accident of bad API; s_umount is used to pin the superblock down when we only have dev_t given us by userland to identify the superblock. Everything else (statfs(), fstatfs(), etc.) doesn't hold it when calling ->statfs() - superblock is pinned down by resolving the pathname passed to syscall. |
1da177e4c Linux-2.6.12-rc2 |
164 165 166 167 168 |
->quota_read() and ->quota_write() functions are both guaranteed to be the only ones operating on the quota file by the quota code (via dqio_sem) (unless an admin really wants to screw up something and writes to quota files with quotas on). For other details about locking see also dquot_operations section. |
b83be6f20 update Documentat... |
169 170 |
->bdev_try_to_free_page is called from the ->releasepage handler of the block device inode. See there for more details. |
1da177e4c Linux-2.6.12-rc2 |
171 172 173 |
--------------------------- file_system_type --------------------------- prototypes: |
b83be6f20 update Documentat... |
174 175 |
struct dentry *(*mount) (struct file_system_type *, int, const char *, void *); |
1da177e4c Linux-2.6.12-rc2 |
176 177 |
void (*kill_sb) (struct super_block *); locking rules: |
b83be6f20 update Documentat... |
178 |
may block |
b83be6f20 update Documentat... |
179 180 |
mount yes kill_sb yes |
1da177e4c Linux-2.6.12-rc2 |
181 |
|
1a102ff92 vfs: bury ->get_sb() |
182 183 |
->mount() returns ERR_PTR or the root dentry; its superblock should be locked on return. |
1da177e4c Linux-2.6.12-rc2 |
184 185 186 187 188 189 190 |
->kill_sb() takes a write-locked superblock, does all shutdown work on it, unlocks and drops the reference. --------------------------- address_space_operations -------------------------- prototypes: int (*writepage)(struct page *page, struct writeback_control *wbc); int (*readpage)(struct file *, struct page *); |
1da177e4c Linux-2.6.12-rc2 |
191 192 193 194 |
int (*writepages)(struct address_space *, struct writeback_control *); int (*set_page_dirty)(struct page *page); int (*readpages)(struct file *filp, struct address_space *mapping, struct list_head *pages, unsigned nr_pages); |
4e02ed4b4 fs: remove prepar... |
195 196 197 198 199 200 |
int (*write_begin)(struct file *, struct address_space *mapping, loff_t pos, unsigned len, unsigned flags, struct page **pagep, void **fsdata); int (*write_end)(struct file *, struct address_space *mapping, loff_t pos, unsigned len, unsigned copied, struct page *page, void *fsdata); |
1da177e4c Linux-2.6.12-rc2 |
201 |
sector_t (*bmap)(struct address_space *, sector_t); |
d47992f86 mm: change invali... |
202 |
void (*invalidatepage) (struct page *, unsigned int, unsigned int); |
1da177e4c Linux-2.6.12-rc2 |
203 |
int (*releasepage) (struct page *, int); |
6072d13c4 Call the filesyst... |
204 |
void (*freepage)(struct page *); |
c8b8e32d7 direct-io: elimin... |
205 |
int (*direct_IO)(struct kiocb *, struct iov_iter *iter); |
bda807d44 mm: migrate: supp... |
206 |
bool (*isolate_page) (struct page *, isolate_mode_t); |
b83be6f20 update Documentat... |
207 |
int (*migratepage)(struct address_space *, struct page *, struct page *); |
bda807d44 mm: migrate: supp... |
208 |
void (*putback_page) (struct page *); |
b83be6f20 update Documentat... |
209 |
int (*launder_page)(struct page *); |
c186afb4d switch ->is_parti... |
210 |
int (*is_partially_uptodate)(struct page *, unsigned long, unsigned long); |
b83be6f20 update Documentat... |
211 |
int (*error_remove_page)(struct address_space *, struct page *); |
62c230bc1 mm: add support f... |
212 213 |
int (*swap_activate)(struct file *); int (*swap_deactivate)(struct file *); |
1da177e4c Linux-2.6.12-rc2 |
214 215 |
locking rules: |
6072d13c4 Call the filesyst... |
216 |
All except set_page_dirty and freepage may block |
1da177e4c Linux-2.6.12-rc2 |
217 |
|
965de0ec3 Documentation: fi... |
218 |
PageLocked(page) i_rwsem |
b83be6f20 update Documentat... |
219 220 |
writepage: yes, unlocks (see below) readpage: yes, unlocks |
b83be6f20 update Documentat... |
221 222 223 |
writepages: set_page_dirty no readpages: |
965de0ec3 Documentation: fi... |
224 225 |
write_begin: locks the page exclusive write_end: yes, unlocks exclusive |
b83be6f20 update Documentat... |
226 227 228 229 230 |
bmap: invalidatepage: yes releasepage: yes freepage: yes direct_IO: |
bda807d44 mm: migrate: supp... |
231 |
isolate_page: yes |
b83be6f20 update Documentat... |
232 |
migratepage: yes (both) |
bda807d44 mm: migrate: supp... |
233 |
putback_page: yes |
b83be6f20 update Documentat... |
234 235 236 |
launder_page: yes is_partially_uptodate: yes error_remove_page: yes |
62c230bc1 mm: add support f... |
237 238 |
swap_activate: no swap_deactivate: no |
1da177e4c Linux-2.6.12-rc2 |
239 |
|
f4e6d844b Remove last trace... |
240 241 |
->write_begin(), ->write_end() and ->readpage() may be called from the request handler (/dev/loop). |
1da177e4c Linux-2.6.12-rc2 |
242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 |
->readpage() unlocks the page, either synchronously or via I/O completion. ->readpages() populates the pagecache with the passed pages and starts I/O against them. They come unlocked upon I/O completion. ->writepage() is used for two purposes: for "memory cleansing" and for "sync". These are quite different operations and the behaviour may differ depending upon the mode. If writepage is called for sync (wbc->sync_mode != WBC_SYNC_NONE) then it *must* start I/O against the page, even if that would involve blocking on in-progress I/O. If writepage is called for memory cleansing (sync_mode == WBC_SYNC_NONE) then its role is to get as much writeout underway as possible. So writepage should try to avoid blocking against currently-in-progress I/O. If the filesystem is not called for "sync" and it determines that it would need to block against in-progress I/O to be able to start new I/O against the page the filesystem should redirty the page with redirty_page_for_writepage(), then unlock the page and return zero. This may also be done to avoid internal deadlocks, but rarely. |
3a4fa0a25 Fix misspellings ... |
267 |
If the filesystem is called for sync then it must wait on any |
1da177e4c Linux-2.6.12-rc2 |
268 |
in-progress I/O and then start new I/O. |
2054606ad [PATCH] doc: Lock... |
269 270 271 272 273 274 |
The filesystem should unlock the page synchronously, before returning to the caller, unless ->writepage() returns special WRITEPAGE_ACTIVATE value. WRITEPAGE_ACTIVATE means that page cannot really be written out currently, and VM should stop calling ->writepage() on this page for some time. VM does this by moving page to the head of the active list, hence the name. |
1da177e4c Linux-2.6.12-rc2 |
275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 |
Unless the filesystem is going to redirty_page_for_writepage(), unlock the page and return zero, writepage *must* run set_page_writeback() against the page, followed by unlocking it. Once set_page_writeback() has been run against the page, write I/O can be submitted and the write I/O completion handler must run end_page_writeback() once the I/O is complete. If no I/O is submitted, the filesystem must run end_page_writeback() against the page before returning from writepage. That is: after 2.5.12, pages which are under writeout are *not* locked. Note, if the filesystem needs the page to be locked during writeout, that is ok, too, the page is allowed to be unlocked at any point in time between the calls to set_page_writeback() and end_page_writeback(). Note, failure to run either redirty_page_for_writepage() or the combination of set_page_writeback()/end_page_writeback() on a page submitted to writepage will leave the page itself marked clean but it will be tagged as dirty in the radix tree. This incoherency can lead to all sorts of hard-to-debug problems in the filesystem like having dirty inodes at umount and losing written data. |
1da177e4c Linux-2.6.12-rc2 |
294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 |
->writepages() is used for periodic writeback and for syscall-initiated sync operations. The address_space should start I/O against at least *nr_to_write pages. *nr_to_write must be decremented for each page which is written. The address_space implementation may write more (or less) pages than *nr_to_write asks for, but it should try to be reasonably close. If nr_to_write is NULL, all dirty pages must be written. writepages should _only_ write pages which are present on mapping->io_pages. ->set_page_dirty() is called from various places in the kernel when the target page is marked as needing writeback. It may be called under spinlock (it cannot block) and is sometimes called with the page not locked. ->bmap() is currently used by legacy ioctl() (FIBMAP) provided by some |
b83be6f20 update Documentat... |
310 311 |
filesystems and by the swapper. The latter will eventually go away. Please, keep it that way and don't breed new callers. |
1da177e4c Linux-2.6.12-rc2 |
312 313 |
->invalidatepage() is called when the filesystem must attempt to drop |
d47992f86 mm: change invali... |
314 315 |
some or all of the buffers from the page when it is being truncated. It returns zero on success. If ->invalidatepage is zero, the kernel uses |
1da177e4c Linux-2.6.12-rc2 |
316 317 318 319 320 321 |
block_invalidatepage() instead. ->releasepage() is called when the kernel is about to try to drop the buffers from the page in preparation for freeing it. It returns zero to indicate that the buffers are (or may be) freeable. If ->releasepage is zero, the kernel assumes that the fs has no private interest in the buffers. |
6072d13c4 Call the filesyst... |
322 323 |
->freepage() is called when the kernel is done dropping the page from the page cache. |
e3db7691e [PATCH] NFS: Fix ... |
324 325 326 327 328 |
->launder_page() may be called prior to releasing a page if it is still found to be dirty. It returns zero if the page was successfully cleaned, or an error value if not. Note that in order to prevent the page getting mapped back in and redirtied, it needs to be kept locked across the entire operation. |
62c230bc1 mm: add support f... |
329 330 331 332 333 334 335 336 |
->swap_activate will be called with a non-zero argument on files backing (non block device backed) swapfiles. A return value of zero indicates success, in which case this file can be used for backing swapspace. The swapspace operations will be proxied to the address space operations. ->swap_deactivate() will be called in the sys_swapoff() path after ->swap_activate() returned success. |
1da177e4c Linux-2.6.12-rc2 |
337 338 |
----------------------- file_lock_operations ------------------------------ prototypes: |
1da177e4c Linux-2.6.12-rc2 |
339 340 341 342 343 |
void (*fl_copy_lock)(struct file_lock *, struct file_lock *); void (*fl_release_private)(struct file_lock *); locking rules: |
1c8c601a8 locks: protect mo... |
344 |
inode->i_lock may block |
b83be6f20 update Documentat... |
345 |
fl_copy_lock: yes no |
2ece173e4 locks: update Loc... |
346 347 348 349 350 |
fl_release_private: maybe maybe[1] [1]: ->fl_release_private for flock or POSIX locks is currently allowed to block. Leases however can still be freed while the i_lock is held and so fl_release_private called on a lease should not block. |
1da177e4c Linux-2.6.12-rc2 |
351 352 353 |
----------------------- lock_manager_operations --------------------------- prototypes: |
8fb47a4fb locks: rename loc... |
354 |
int (*lm_compare_owner)(struct file_lock *, struct file_lock *); |
3999e4936 locks: add a new ... |
355 |
unsigned long (*lm_owner_key)(struct file_lock *); |
8fb47a4fb locks: rename loc... |
356 357 |
void (*lm_notify)(struct file_lock *); /* unblock callback */ int (*lm_grant)(struct file_lock *, struct file_lock *, int); |
8fb47a4fb locks: rename loc... |
358 359 |
void (*lm_break)(struct file_lock *); /* break_lease callback */ int (*lm_change)(struct file_lock **, int); |
1da177e4c Linux-2.6.12-rc2 |
360 361 |
locking rules: |
1c8c601a8 locks: protect mo... |
362 |
|
7b2296afb locks: give the b... |
363 364 365 366 367 368 369 |
inode->i_lock blocked_lock_lock may block lm_compare_owner: yes[1] maybe no lm_owner_key yes[1] yes no lm_notify: yes yes no lm_grant: no no no lm_break: yes no no lm_change yes no no |
1c8c601a8 locks: protect mo... |
370 |
|
3999e4936 locks: add a new ... |
371 372 373 374 375 |
[1]: ->lm_compare_owner and ->lm_owner_key are generally called with *an* inode->i_lock held. It may not be the i_lock of the inode associated with either file_lock argument! This is the case with deadlock detection, since the code has to chase down the owners of locks that may be entirely unrelated to the one on which the lock is being acquired. |
7b2296afb locks: give the b... |
376 |
For deadlock detection however, the blocked_lock_lock is also held. The |
3999e4936 locks: add a new ... |
377 378 379 |
fact that these locks are held ensures that the file_locks do not disappear out from under you while doing the comparison or generating an owner key. |
b83be6f20 update Documentat... |
380 |
|
1da177e4c Linux-2.6.12-rc2 |
381 382 383 384 385 386 387 388 389 390 391 392 |
--------------------------- buffer_head ----------------------------------- prototypes: void (*b_end_io)(struct buffer_head *bh, int uptodate); locking rules: called from interrupts. In other words, extreme care is needed here. bh is locked, but that's all warranties we have here. Currently only RAID1, highmem, fs/buffer.c, and fs/ntfs/aops.c are providing these. Block devices call this method upon the IO completion. --------------------------- block_device_operations ----------------------- prototypes: |
e1455d1bd update block_devi... |
393 394 395 396 |
int (*open) (struct block_device *, fmode_t); int (*release) (struct gendisk *, fmode_t); int (*ioctl) (struct block_device *, fmode_t, unsigned, unsigned long); int (*compat_ioctl) (struct block_device *, fmode_t, unsigned, unsigned long); |
7a9eb2066 pmem: kill __pmem... |
397 |
int (*direct_access) (struct block_device *, sector_t, void **, |
e2e05394e pmem, dax: have d... |
398 |
unsigned long *); |
1da177e4c Linux-2.6.12-rc2 |
399 |
int (*media_changed) (struct gendisk *); |
e1455d1bd update block_devi... |
400 |
void (*unlock_native_capacity) (struct gendisk *); |
1da177e4c Linux-2.6.12-rc2 |
401 |
int (*revalidate_disk) (struct gendisk *); |
e1455d1bd update block_devi... |
402 403 |
int (*getgeo)(struct block_device *, struct hd_geometry *); void (*swap_slot_free_notify) (struct block_device *, unsigned long); |
1da177e4c Linux-2.6.12-rc2 |
404 405 |
locking rules: |
b83be6f20 update Documentat... |
406 407 408 409 410 411 412 413 414 415 416 |
bd_mutex open: yes release: yes ioctl: no compat_ioctl: no direct_access: no media_changed: no unlock_native_capacity: no revalidate_disk: no getgeo: no swap_slot_free_notify: no (see below) |
e1455d1bd update block_devi... |
417 418 419 420 421 422 |
media_changed, unlock_native_capacity and revalidate_disk are called only from check_disk_change(). swap_slot_free_notify is called with swap_lock and sometimes the page lock held. |
1da177e4c Linux-2.6.12-rc2 |
423 |
|
1da177e4c Linux-2.6.12-rc2 |
424 425 426 427 428 |
--------------------------- file_operations ------------------------------- prototypes: loff_t (*llseek) (struct file *, loff_t, int); ssize_t (*read) (struct file *, char __user *, size_t, loff_t *); |
1da177e4c Linux-2.6.12-rc2 |
429 |
ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *); |
293bc9822 new methods: ->re... |
430 431 |
ssize_t (*read_iter) (struct kiocb *, struct iov_iter *); ssize_t (*write_iter) (struct kiocb *, struct iov_iter *); |
2233f31aa [readdir] ->readd... |
432 |
int (*iterate) (struct file *, struct dir_context *); |
965de0ec3 Documentation: fi... |
433 |
int (*iterate_shared) (struct file *, struct dir_context *); |
6e8b704df fs: update docume... |
434 |
__poll_t (*poll) (struct file *, struct poll_table_struct *); |
1da177e4c Linux-2.6.12-rc2 |
435 436 437 438 439 440 |
long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long); long (*compat_ioctl) (struct file *, unsigned int, unsigned long); int (*mmap) (struct file *, struct vm_area_struct *); int (*open) (struct inode *, struct file *); int (*flush) (struct file *); int (*release) (struct inode *, struct file *); |
02c24a821 fs: push i_mutex ... |
441 |
int (*fsync) (struct file *, loff_t start, loff_t end, int datasync); |
1da177e4c Linux-2.6.12-rc2 |
442 443 444 445 446 447 448 449 450 451 452 453 454 |
int (*fasync) (int, struct file *, int); int (*lock) (struct file *, int, struct file_lock *); ssize_t (*readv) (struct file *, const struct iovec *, unsigned long, loff_t *); ssize_t (*writev) (struct file *, const struct iovec *, unsigned long, loff_t *); ssize_t (*sendfile) (struct file *, loff_t *, size_t, read_actor_t, void __user *); ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int); unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long); int (*check_flags)(int); |
b83be6f20 update Documentat... |
455 456 457 458 459 |
int (*flock) (struct file *, int, struct file_lock *); ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int); ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int); |
e6f5c7893 locks: plumb a "p... |
460 |
int (*setlease)(struct file *, long, struct file_lock **, void **); |
2fe17c107 fallocate should ... |
461 |
long (*fallocate)(struct file *, int, loff_t, loff_t); |
1da177e4c Linux-2.6.12-rc2 |
462 463 464 |
}; locking rules: |
a11e1d432 Revert changes to... |
465 |
All may block. |
b83be6f20 update Documentat... |
466 |
|
1da177e4c Linux-2.6.12-rc2 |
467 468 469 470 |
->llseek() locking has moved from llseek to the individual llseek implementations. If your fs is not using generic_file_llseek, you need to acquire and release the appropriate locks in your ->llseek(). For many filesystems, it is probably safe to acquire the inode |
866707fc2 Documentation/fil... |
471 472 473 |
mutex or just to use i_size_read() instead. Note: this does not protect the file->f_pos against concurrent modifications since this is something the userspace has to take care about. |
1da177e4c Linux-2.6.12-rc2 |
474 |
|
965de0ec3 Documentation: fi... |
475 476 477 |
->iterate() is called with i_rwsem exclusive. ->iterate_shared() is called with i_rwsem at least shared. |
b83be6f20 update Documentat... |
478 479 480 481 |
->fasync() is responsible for maintaining the FASYNC bit in filp->f_flags. Most instances call fasync_helper(), which does that maintenance, so it's not normally something one needs to worry about. Return values > 0 will be mapped to zero in the VFS layer. |
1da177e4c Linux-2.6.12-rc2 |
482 483 484 485 486 487 |
->readdir() and ->ioctl() on directories must be changed. Ideally we would move ->readdir() to inode_operations and use a separate method for directory ->ioctl() or kill the latter completely. One of the problems is that for anything that resembles union-mount we won't have a struct file for all components. And there are other reasons why the current interface is a mess... |
1da177e4c Linux-2.6.12-rc2 |
488 489 |
->read on directories probably must go away - we should just enforce -EISDIR in sys_read() and friends. |
f82b4b678 locks: move i_loc... |
490 491 492 |
->setlease operations should call generic_setlease() before or after setting the lease within the individual filesystem to record the result of the operation |
1da177e4c Linux-2.6.12-rc2 |
493 494 |
--------------------------- dquot_operations ------------------------------- prototypes: |
1da177e4c Linux-2.6.12-rc2 |
495 496 497 498 499 500 501 502 503 504 505 506 |
int (*write_dquot) (struct dquot *); int (*acquire_dquot) (struct dquot *); int (*release_dquot) (struct dquot *); int (*mark_dirty) (struct dquot *); int (*write_info) (struct super_block *, int); These operations are intended to be more or less wrapping functions that ensure a proper locking wrt the filesystem and call the generic quota operations. What filesystem should expect from the generic quota functions: FS recursion Held locks when called |
1da177e4c Linux-2.6.12-rc2 |
507 508 509 510 511 512 513 514 |
write_dquot: yes dqonoff_sem or dqptr_sem acquire_dquot: yes dqonoff_sem or dqptr_sem release_dquot: yes dqonoff_sem or dqptr_sem mark_dirty: no - write_info: yes dqonoff_sem FS recursion means calling ->quota_read() and ->quota_write() from superblock operations. |
1da177e4c Linux-2.6.12-rc2 |
515 516 517 518 519 520 |
More details about quota locking can be found in fs/dquot.c. --------------------------- vm_operations_struct ----------------------------- prototypes: void (*open)(struct vm_area_struct*); void (*close)(struct vm_area_struct*); |
fe3136f44 mm: Introduce new... |
521 522 523 |
vm_fault_t (*fault)(struct vm_area_struct*, struct vm_fault *); vm_fault_t (*page_mkwrite)(struct vm_area_struct *, struct vm_fault *); vm_fault_t (*pfn_mkwrite)(struct vm_area_struct *, struct vm_fault *); |
28b2ee20c access_process_vm... |
524 |
int (*access)(struct vm_area_struct *, unsigned long, void*, int, int); |
1da177e4c Linux-2.6.12-rc2 |
525 526 |
locking rules: |
b83be6f20 update Documentat... |
527 528 529 530 |
mmap_sem PageLocked(page) open: yes close: yes fault: yes can return with page locked |
8c6e50b02 mm: introduce vm_... |
531 |
map_pages: yes |
b83be6f20 update Documentat... |
532 |
page_mkwrite: yes can return with page locked |
dd9061846 mm: new pfn_mkwri... |
533 |
pfn_mkwrite: yes |
b83be6f20 update Documentat... |
534 |
access: yes |
ed2f2f9b3 Document ->page_m... |
535 |
|
b827e496c mm: close page_mk... |
536 537 538 539 540 541 542 |
->fault() is called when a previously not present pte is about to be faulted in. The filesystem must find and return the page associated with the passed in "pgoff" in the vm_fault structure. If it is possible that the page may be truncated and/or invalidated, then the filesystem must lock the page, then ensure it is not already truncated (the page lock will block subsequent truncate), and then return with VM_FAULT_LOCKED, and the page locked. The VM will unlock the page. |
8c6e50b02 mm: introduce vm_... |
543 |
->map_pages() is called when VM asks to map easy accessible pages. |
bae473a42 mm: introduce fau... |
544 545 |
Filesystem should find and map pages associated with offsets from "start_pgoff" till "end_pgoff". ->map_pages() is called with page table locked and must |
8c6e50b02 mm: introduce vm_... |
546 547 |
not block. If it's not possible to reach a page without blocking, filesystem should skip it. Filesystem should use do_set_pte() to setup |
bae473a42 mm: introduce fau... |
548 |
page table entry. Pointer to entry associated with the page is passed in |
82b0f8c39 mm: join struct f... |
549 |
"pte" field in vm_fault structure. Pointers to entries for other offsets |
bae473a42 mm: introduce fau... |
550 |
should be calculated relative to "pte". |
8c6e50b02 mm: introduce vm_... |
551 |
|
b827e496c mm: close page_mk... |
552 553 554 555 556 557 |
->page_mkwrite() is called when a previously read-only pte is about to become writeable. The filesystem again must ensure that there are no truncate/invalidate races, and then return with the page locked. If the page has been truncated, the filesystem should not look up a new page like the ->fault() handler, but simply return with VM_FAULT_NOPAGE, which will cause the VM to retry the fault. |
1da177e4c Linux-2.6.12-rc2 |
558 |
|
dd9061846 mm: new pfn_mkwri... |
559 560 561 562 563 |
->pfn_mkwrite() is the same as page_mkwrite but when the pte is VM_PFNMAP or VM_MIXEDMAP with a page-less entry. Expected return is VM_FAULT_NOPAGE. Or one of the VM_FAULT_ERROR types. The default behavior after this call is to make the pte read-write, unless pfn_mkwrite returns an error. |
28b2ee20c access_process_vm... |
564 |
->access() is called when get_user_pages() fails in |
507da6a1f doc: Fix typo (ac... |
565 |
access_process_vm(), typically used to debug a process through |
28b2ee20c access_process_vm... |
566 567 |
/proc/pid/mem or ptrace. This function is needed only for VM_IO | VM_PFNMAP VMAs. |
1da177e4c Linux-2.6.12-rc2 |
568 569 570 571 572 |
================================================================================ Dubious stuff (if you break something or notice that it is broken and do not fix it yourself - at least put it here) |