Loading...
Note: File does not exist in v6.13.7.
1 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:
12 int (*d_revalidate)(struct dentry *, unsigned int);
13 int (*d_weak_revalidate)(struct dentry *, unsigned int);
14 int (*d_hash)(const struct dentry *, struct qstr *);
15 int (*d_compare)(const struct dentry *, const struct dentry *,
16 unsigned int, const char *, const struct qstr *);
17 int (*d_delete)(struct dentry *);
18 void (*d_release)(struct dentry *);
19 void (*d_iput)(struct dentry *, struct inode *);
20 char *(*d_dname)((struct dentry *dentry, char *buffer, int buflen);
21 struct vfsmount *(*d_automount)(struct path *path);
22 int (*d_manage)(struct dentry *, bool);
23
24locking rules:
25 rename_lock ->d_lock may block rcu-walk
26d_revalidate: no no yes (ref-walk) maybe
27d_weak_revalidate:no no yes no
28d_hash no no no maybe
29d_compare: yes no no maybe
30d_delete: no yes no no
31d_release: no no yes no
32d_prune: no yes no no
33d_iput: no no yes no
34d_dname: no no no no
35d_automount: no no yes no
36d_manage: no no yes (ref-walk) maybe
37
38--------------------------- inode_operations ---------------------------
39prototypes:
40 int (*create) (struct inode *,struct dentry *,umode_t, bool);
41 struct dentry * (*lookup) (struct inode *,struct dentry *, unsigned int);
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 *);
45 int (*mkdir) (struct inode *,struct dentry *,umode_t);
46 int (*rmdir) (struct inode *,struct dentry *);
47 int (*mknod) (struct inode *,struct dentry *,umode_t,dev_t);
48 int (*rename) (struct inode *, struct dentry *,
49 struct inode *, struct dentry *);
50 int (*rename2) (struct inode *, struct dentry *,
51 struct inode *, struct dentry *, unsigned int);
52 int (*readlink) (struct dentry *, char __user *,int);
53 const char *(*get_link) (struct dentry *, struct inode *, void **);
54 void (*truncate) (struct inode *);
55 int (*permission) (struct inode *, int, unsigned int);
56 int (*get_acl)(struct inode *, int);
57 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 *);
63 int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start, u64 len);
64 void (*update_time)(struct inode *, struct timespec *, int);
65 int (*atomic_open)(struct inode *, struct dentry *,
66 struct file *, unsigned open_flag,
67 umode_t create_mode, int *opened);
68 int (*tmpfile) (struct inode *, struct dentry *, umode_t);
69 int (*dentry_open)(struct dentry *, struct file *, const struct cred *);
70
71locking rules:
72 all may block
73 i_mutex(inode)
74lookup: yes
75create: yes
76link: yes (both)
77mknod: yes
78symlink: yes
79mkdir: yes
80unlink: yes (both)
81rmdir: yes (both) (see below)
82rename: yes (all) (see below)
83rename2: yes (all) (see below)
84readlink: no
85get_link: no
86setattr: yes
87permission: no (may not block if called in rcu-walk mode)
88get_acl: no
89getattr: no
90setxattr: yes
91getxattr: no
92listxattr: no
93removexattr: yes
94fiemap: no
95update_time: no
96atomic_open: yes
97tmpfile: no
98dentry_open: no
99
100 Additionally, ->rmdir(), ->unlink() and ->rename() have ->i_mutex on
101victim.
102 cross-directory ->rename() and rename2() has (per-superblock)
103->s_vfs_rename_sem.
104
105See Documentation/filesystems/directory-locking for more detailed discussion
106of the locking scheme for directory operations.
107
108--------------------------- super_operations ---------------------------
109prototypes:
110 struct inode *(*alloc_inode)(struct super_block *sb);
111 void (*destroy_inode)(struct inode *);
112 void (*dirty_inode) (struct inode *, int flags);
113 int (*write_inode) (struct inode *, struct writeback_control *wbc);
114 int (*drop_inode) (struct inode *);
115 void (*evict_inode) (struct inode *);
116 void (*put_super) (struct super_block *);
117 int (*sync_fs)(struct super_block *sb, int wait);
118 int (*freeze_fs) (struct super_block *);
119 int (*unfreeze_fs) (struct super_block *);
120 int (*statfs) (struct dentry *, struct kstatfs *);
121 int (*remount_fs) (struct super_block *, int *, char *);
122 void (*umount_begin) (struct super_block *);
123 int (*show_options)(struct seq_file *, struct dentry *);
124 ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
125 ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
126 int (*bdev_try_to_free_page)(struct super_block*, struct page*, gfp_t);
127
128locking rules:
129 All may block [not true, see below]
130 s_umount
131alloc_inode:
132destroy_inode:
133dirty_inode:
134write_inode:
135drop_inode: !!!inode->i_lock!!!
136evict_inode:
137put_super: write
138sync_fs: read
139freeze_fs: write
140unfreeze_fs: write
141statfs: maybe(read) (see below)
142remount_fs: write
143umount_begin: no
144show_options: no (namespace_sem)
145quota_read: no (see below)
146quota_write: no (see below)
147bdev_try_to_free_page: no (see below)
148
149->statfs() has s_umount (shared) when called by ustat(2) (native or
150compat), but that's an accident of bad API; s_umount is used to pin
151the superblock down when we only have dev_t given us by userland to
152identify the superblock. Everything else (statfs(), fstatfs(), etc.)
153doesn't hold it when calling ->statfs() - superblock is pinned down
154by resolving the pathname passed to syscall.
155->quota_read() and ->quota_write() functions are both guaranteed to
156be the only ones operating on the quota file by the quota code (via
157dqio_sem) (unless an admin really wants to screw up something and
158writes to quota files with quotas on). For other details about locking
159see also dquot_operations section.
160->bdev_try_to_free_page is called from the ->releasepage handler of
161the block device inode. See there for more details.
162
163--------------------------- file_system_type ---------------------------
164prototypes:
165 struct dentry *(*mount) (struct file_system_type *, int,
166 const char *, void *);
167 void (*kill_sb) (struct super_block *);
168locking rules:
169 may block
170mount yes
171kill_sb yes
172
173->mount() returns ERR_PTR or the root dentry; its superblock should be locked
174on return.
175->kill_sb() takes a write-locked superblock, does all shutdown work on it,
176unlocks and drops the reference.
177
178--------------------------- address_space_operations --------------------------
179prototypes:
180 int (*writepage)(struct page *page, struct writeback_control *wbc);
181 int (*readpage)(struct file *, struct page *);
182 int (*sync_page)(struct page *);
183 int (*writepages)(struct address_space *, struct writeback_control *);
184 int (*set_page_dirty)(struct page *page);
185 int (*readpages)(struct file *filp, struct address_space *mapping,
186 struct list_head *pages, unsigned nr_pages);
187 int (*write_begin)(struct file *, struct address_space *mapping,
188 loff_t pos, unsigned len, unsigned flags,
189 struct page **pagep, void **fsdata);
190 int (*write_end)(struct file *, struct address_space *mapping,
191 loff_t pos, unsigned len, unsigned copied,
192 struct page *page, void *fsdata);
193 sector_t (*bmap)(struct address_space *, sector_t);
194 void (*invalidatepage) (struct page *, unsigned int, unsigned int);
195 int (*releasepage) (struct page *, int);
196 void (*freepage)(struct page *);
197 int (*direct_IO)(struct kiocb *, struct iov_iter *iter, loff_t offset);
198 int (*migratepage)(struct address_space *, struct page *, struct page *);
199 int (*launder_page)(struct page *);
200 int (*is_partially_uptodate)(struct page *, unsigned long, unsigned long);
201 int (*error_remove_page)(struct address_space *, struct page *);
202 int (*swap_activate)(struct file *);
203 int (*swap_deactivate)(struct file *);
204
205locking rules:
206 All except set_page_dirty and freepage may block
207
208 PageLocked(page) i_mutex
209writepage: yes, unlocks (see below)
210readpage: yes, unlocks
211sync_page: maybe
212writepages:
213set_page_dirty no
214readpages:
215write_begin: locks the page yes
216write_end: yes, unlocks yes
217bmap:
218invalidatepage: yes
219releasepage: yes
220freepage: yes
221direct_IO:
222migratepage: yes (both)
223launder_page: yes
224is_partially_uptodate: yes
225error_remove_page: yes
226swap_activate: no
227swap_deactivate: no
228
229 ->write_begin(), ->write_end(), ->sync_page() and ->readpage()
230may be called from the request handler (/dev/loop).
231
232 ->readpage() unlocks the page, either synchronously or via I/O
233completion.
234
235 ->readpages() populates the pagecache with the passed pages and starts
236I/O against them. They come unlocked upon I/O completion.
237
238 ->writepage() is used for two purposes: for "memory cleansing" and for
239"sync". These are quite different operations and the behaviour may differ
240depending upon the mode.
241
242If writepage is called for sync (wbc->sync_mode != WBC_SYNC_NONE) then
243it *must* start I/O against the page, even if that would involve
244blocking on in-progress I/O.
245
246If writepage is called for memory cleansing (sync_mode ==
247WBC_SYNC_NONE) then its role is to get as much writeout underway as
248possible. So writepage should try to avoid blocking against
249currently-in-progress I/O.
250
251If the filesystem is not called for "sync" and it determines that it
252would need to block against in-progress I/O to be able to start new I/O
253against the page the filesystem should redirty the page with
254redirty_page_for_writepage(), then unlock the page and return zero.
255This may also be done to avoid internal deadlocks, but rarely.
256
257If the filesystem is called for sync then it must wait on any
258in-progress I/O and then start new I/O.
259
260The filesystem should unlock the page synchronously, before returning to the
261caller, unless ->writepage() returns special WRITEPAGE_ACTIVATE
262value. WRITEPAGE_ACTIVATE means that page cannot really be written out
263currently, and VM should stop calling ->writepage() on this page for some
264time. VM does this by moving page to the head of the active list, hence the
265name.
266
267Unless the filesystem is going to redirty_page_for_writepage(), unlock the page
268and return zero, writepage *must* run set_page_writeback() against the page,
269followed by unlocking it. Once set_page_writeback() has been run against the
270page, write I/O can be submitted and the write I/O completion handler must run
271end_page_writeback() once the I/O is complete. If no I/O is submitted, the
272filesystem must run end_page_writeback() against the page before returning from
273writepage.
274
275That is: after 2.5.12, pages which are under writeout are *not* locked. Note,
276if the filesystem needs the page to be locked during writeout, that is ok, too,
277the page is allowed to be unlocked at any point in time between the calls to
278set_page_writeback() and end_page_writeback().
279
280Note, failure to run either redirty_page_for_writepage() or the combination of
281set_page_writeback()/end_page_writeback() on a page submitted to writepage
282will leave the page itself marked clean but it will be tagged as dirty in the
283radix tree. This incoherency can lead to all sorts of hard-to-debug problems
284in the filesystem like having dirty inodes at umount and losing written data.
285
286 ->sync_page() locking rules are not well-defined - usually it is called
287with lock on page, but that is not guaranteed. Considering the currently
288existing instances of this method ->sync_page() itself doesn't look
289well-defined...
290
291 ->writepages() is used for periodic writeback and for syscall-initiated
292sync operations. The address_space should start I/O against at least
293*nr_to_write pages. *nr_to_write must be decremented for each page which is
294written. The address_space implementation may write more (or less) pages
295than *nr_to_write asks for, but it should try to be reasonably close. If
296nr_to_write is NULL, all dirty pages must be written.
297
298writepages should _only_ write pages which are present on
299mapping->io_pages.
300
301 ->set_page_dirty() is called from various places in the kernel
302when the target page is marked as needing writeback. It may be called
303under spinlock (it cannot block) and is sometimes called with the page
304not locked.
305
306 ->bmap() is currently used by legacy ioctl() (FIBMAP) provided by some
307filesystems and by the swapper. The latter will eventually go away. Please,
308keep it that way and don't breed new callers.
309
310 ->invalidatepage() is called when the filesystem must attempt to drop
311some or all of the buffers from the page when it is being truncated. It
312returns zero on success. If ->invalidatepage is zero, the kernel uses
313block_invalidatepage() instead.
314
315 ->releasepage() is called when the kernel is about to try to drop the
316buffers from the page in preparation for freeing it. It returns zero to
317indicate that the buffers are (or may be) freeable. If ->releasepage is zero,
318the kernel assumes that the fs has no private interest in the buffers.
319
320 ->freepage() is called when the kernel is done dropping the page
321from the page cache.
322
323 ->launder_page() may be called prior to releasing a page if
324it is still found to be dirty. It returns zero if the page was successfully
325cleaned, or an error value if not. Note that in order to prevent the page
326getting mapped back in and redirtied, it needs to be kept locked
327across the entire operation.
328
329 ->swap_activate will be called with a non-zero argument on
330files backing (non block device backed) swapfiles. A return value
331of zero indicates success, in which case this file can be used for
332backing swapspace. The swapspace operations will be proxied to the
333address space operations.
334
335 ->swap_deactivate() will be called in the sys_swapoff()
336path after ->swap_activate() returned success.
337
338----------------------- file_lock_operations ------------------------------
339prototypes:
340 void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
341 void (*fl_release_private)(struct file_lock *);
342
343
344locking rules:
345 inode->i_lock may block
346fl_copy_lock: yes no
347fl_release_private: maybe maybe[1]
348
349[1]: ->fl_release_private for flock or POSIX locks is currently allowed
350to block. Leases however can still be freed while the i_lock is held and
351so fl_release_private called on a lease should not block.
352
353----------------------- lock_manager_operations ---------------------------
354prototypes:
355 int (*lm_compare_owner)(struct file_lock *, struct file_lock *);
356 unsigned long (*lm_owner_key)(struct file_lock *);
357 void (*lm_notify)(struct file_lock *); /* unblock callback */
358 int (*lm_grant)(struct file_lock *, struct file_lock *, int);
359 void (*lm_break)(struct file_lock *); /* break_lease callback */
360 int (*lm_change)(struct file_lock **, int);
361
362locking rules:
363
364 inode->i_lock blocked_lock_lock may block
365lm_compare_owner: yes[1] maybe no
366lm_owner_key yes[1] yes no
367lm_notify: yes yes no
368lm_grant: no no no
369lm_break: yes no no
370lm_change yes no no
371
372[1]: ->lm_compare_owner and ->lm_owner_key are generally called with
373*an* inode->i_lock held. It may not be the i_lock of the inode
374associated with either file_lock argument! This is the case with deadlock
375detection, since the code has to chase down the owners of locks that may
376be entirely unrelated to the one on which the lock is being acquired.
377For deadlock detection however, the blocked_lock_lock is also held. The
378fact that these locks are held ensures that the file_locks do not
379disappear out from under you while doing the comparison or generating an
380owner key.
381
382--------------------------- buffer_head -----------------------------------
383prototypes:
384 void (*b_end_io)(struct buffer_head *bh, int uptodate);
385
386locking rules:
387 called from interrupts. In other words, extreme care is needed here.
388bh is locked, but that's all warranties we have here. Currently only RAID1,
389highmem, fs/buffer.c, and fs/ntfs/aops.c are providing these. Block devices
390call this method upon the IO completion.
391
392--------------------------- block_device_operations -----------------------
393prototypes:
394 int (*open) (struct block_device *, fmode_t);
395 int (*release) (struct gendisk *, fmode_t);
396 int (*ioctl) (struct block_device *, fmode_t, unsigned, unsigned long);
397 int (*compat_ioctl) (struct block_device *, fmode_t, unsigned, unsigned long);
398 int (*direct_access) (struct block_device *, sector_t, void __pmem **,
399 unsigned long *);
400 int (*media_changed) (struct gendisk *);
401 void (*unlock_native_capacity) (struct gendisk *);
402 int (*revalidate_disk) (struct gendisk *);
403 int (*getgeo)(struct block_device *, struct hd_geometry *);
404 void (*swap_slot_free_notify) (struct block_device *, unsigned long);
405
406locking rules:
407 bd_mutex
408open: yes
409release: yes
410ioctl: no
411compat_ioctl: no
412direct_access: no
413media_changed: no
414unlock_native_capacity: no
415revalidate_disk: no
416getgeo: no
417swap_slot_free_notify: no (see below)
418
419media_changed, unlock_native_capacity and revalidate_disk are called only from
420check_disk_change().
421
422swap_slot_free_notify is called with swap_lock and sometimes the page lock
423held.
424
425
426--------------------------- file_operations -------------------------------
427prototypes:
428 loff_t (*llseek) (struct file *, loff_t, int);
429 ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
430 ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
431 ssize_t (*read_iter) (struct kiocb *, struct iov_iter *);
432 ssize_t (*write_iter) (struct kiocb *, struct iov_iter *);
433 int (*iterate) (struct file *, struct dir_context *);
434 unsigned int (*poll) (struct file *, struct poll_table_struct *);
435 long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
436 long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
437 int (*mmap) (struct file *, struct vm_area_struct *);
438 int (*open) (struct inode *, struct file *);
439 int (*flush) (struct file *);
440 int (*release) (struct inode *, struct file *);
441 int (*fsync) (struct file *, loff_t start, loff_t end, int datasync);
442 int (*aio_fsync) (struct kiocb *, int datasync);
443 int (*fasync) (int, struct file *, int);
444 int (*lock) (struct file *, int, struct file_lock *);
445 ssize_t (*readv) (struct file *, const struct iovec *, unsigned long,
446 loff_t *);
447 ssize_t (*writev) (struct file *, const struct iovec *, unsigned long,
448 loff_t *);
449 ssize_t (*sendfile) (struct file *, loff_t *, size_t, read_actor_t,
450 void __user *);
451 ssize_t (*sendpage) (struct file *, struct page *, int, size_t,
452 loff_t *, int);
453 unsigned long (*get_unmapped_area)(struct file *, unsigned long,
454 unsigned long, unsigned long, unsigned long);
455 int (*check_flags)(int);
456 int (*flock) (struct file *, int, struct file_lock *);
457 ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *,
458 size_t, unsigned int);
459 ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *,
460 size_t, unsigned int);
461 int (*setlease)(struct file *, long, struct file_lock **, void **);
462 long (*fallocate)(struct file *, int, loff_t, loff_t);
463};
464
465locking rules:
466 All may block.
467
468->llseek() locking has moved from llseek to the individual llseek
469implementations. If your fs is not using generic_file_llseek, you
470need to acquire and release the appropriate locks in your ->llseek().
471For many filesystems, it is probably safe to acquire the inode
472mutex or just to use i_size_read() instead.
473Note: this does not protect the file->f_pos against concurrent modifications
474since this is something the userspace has to take care about.
475
476->fasync() is responsible for maintaining the FASYNC bit in filp->f_flags.
477Most instances call fasync_helper(), which does that maintenance, so it's
478not normally something one needs to worry about. Return values > 0 will be
479mapped to zero in the VFS layer.
480
481->readdir() and ->ioctl() on directories must be changed. Ideally we would
482move ->readdir() to inode_operations and use a separate method for directory
483->ioctl() or kill the latter completely. One of the problems is that for
484anything that resembles union-mount we won't have a struct file for all
485components. And there are other reasons why the current interface is a mess...
486
487->read on directories probably must go away - we should just enforce -EISDIR
488in sys_read() and friends.
489
490->setlease operations should call generic_setlease() before or after setting
491the lease within the individual filesystem to record the result of the
492operation
493
494--------------------------- dquot_operations -------------------------------
495prototypes:
496 int (*write_dquot) (struct dquot *);
497 int (*acquire_dquot) (struct dquot *);
498 int (*release_dquot) (struct dquot *);
499 int (*mark_dirty) (struct dquot *);
500 int (*write_info) (struct super_block *, int);
501
502These operations are intended to be more or less wrapping functions that ensure
503a proper locking wrt the filesystem and call the generic quota operations.
504
505What filesystem should expect from the generic quota functions:
506
507 FS recursion Held locks when called
508write_dquot: yes dqonoff_sem or dqptr_sem
509acquire_dquot: yes dqonoff_sem or dqptr_sem
510release_dquot: yes dqonoff_sem or dqptr_sem
511mark_dirty: no -
512write_info: yes dqonoff_sem
513
514FS recursion means calling ->quota_read() and ->quota_write() from superblock
515operations.
516
517More details about quota locking can be found in fs/dquot.c.
518
519--------------------------- vm_operations_struct -----------------------------
520prototypes:
521 void (*open)(struct vm_area_struct*);
522 void (*close)(struct vm_area_struct*);
523 int (*fault)(struct vm_area_struct*, struct vm_fault *);
524 int (*page_mkwrite)(struct vm_area_struct *, struct vm_fault *);
525 int (*pfn_mkwrite)(struct vm_area_struct *, struct vm_fault *);
526 int (*access)(struct vm_area_struct *, unsigned long, void*, int, int);
527
528locking rules:
529 mmap_sem PageLocked(page)
530open: yes
531close: yes
532fault: yes can return with page locked
533map_pages: yes
534page_mkwrite: yes can return with page locked
535pfn_mkwrite: yes
536access: yes
537
538 ->fault() is called when a previously not present pte is about
539to be faulted in. The filesystem must find and return the page associated
540with the passed in "pgoff" in the vm_fault structure. If it is possible that
541the page may be truncated and/or invalidated, then the filesystem must lock
542the page, then ensure it is not already truncated (the page lock will block
543subsequent truncate), and then return with VM_FAULT_LOCKED, and the page
544locked. The VM will unlock the page.
545
546 ->map_pages() is called when VM asks to map easy accessible pages.
547Filesystem should find and map pages associated with offsets from "pgoff"
548till "max_pgoff". ->map_pages() is called with page table locked and must
549not block. If it's not possible to reach a page without blocking,
550filesystem should skip it. Filesystem should use do_set_pte() to setup
551page table entry. Pointer to entry associated with offset "pgoff" is
552passed in "pte" field in vm_fault structure. Pointers to entries for other
553offsets should be calculated relative to "pte".
554
555 ->page_mkwrite() is called when a previously read-only pte is
556about to become writeable. The filesystem again must ensure that there are
557no truncate/invalidate races, and then return with the page locked. If
558the page has been truncated, the filesystem should not look up a new page
559like the ->fault() handler, but simply return with VM_FAULT_NOPAGE, which
560will cause the VM to retry the fault.
561
562 ->pfn_mkwrite() is the same as page_mkwrite but when the pte is
563VM_PFNMAP or VM_MIXEDMAP with a page-less entry. Expected return is
564VM_FAULT_NOPAGE. Or one of the VM_FAULT_ERROR types. The default behavior
565after this call is to make the pte read-write, unless pfn_mkwrite returns
566an error.
567
568 ->access() is called when get_user_pages() fails in
569access_process_vm(), typically used to debug a process through
570/proc/pid/mem or ptrace. This function is needed only for
571VM_IO | VM_PFNMAP VMAs.
572
573================================================================================
574 Dubious stuff
575
576(if you break something or notice that it is broken and do not fix it yourself
577- at least put it here)