Documentation/filesystems/Locking

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