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_hash)(const struct dentry *, const struct inode *,
  14                         struct qstr *);
  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 *);
  18         int (*d_delete)(struct dentry *);
  19         void (*d_release)(struct dentry *);
  20         void (*d_iput)(struct dentry *, struct inode *);
  21         char *(*d_dname)((struct dentry *dentry, char *buffer, int buflen);
  22         struct vfsmount *(*d_automount)(struct path *path);
  23         int (*d_manage)(struct dentry *, bool);
  24
  25 locking rules:
  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
  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 (*readlink) (struct dentry *, char __user *,int);
  51         void * (*follow_link) (struct dentry *, struct nameidata *);
  52         void (*put_link) (struct dentry *, struct nameidata *, void *);
  53         void (*truncate) (struct inode *);
  54         int (*permission) (struct inode *, int, unsigned int);
  55         int (*get_acl)(struct inode *, int);
  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 *);
  62         int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start, u64 len);
  63         void (*update_time)(struct inode *, struct timespec *, int);
  64         int (*atomic_open)(struct inode *, struct dentry *,
  65                                 struct file *, unsigned open_flag,
  66                                 umode_t create_mode, int *opened);
  67
  68 locking rules:
  69         all may block
  70                 i_mutex(inode)
  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
  82 put_link:       no
  83 truncate:       yes             (see below)
  84 setattr:        yes
  85 permission:     no (may not block if called in rcu-walk mode)
  86 get_acl:        no
  87 getattr:        no
  88 setxattr:       yes
  89 getxattr:       no
  90 listxattr:      no
  91 removexattr:    yes
  92 fiemap:         no
  93 update_time:    no
  94 atomic_open:    yes
  95
  96         Additionally, ->rmdir(), ->unlink() and ->rename() have ->i_mutex on
  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
 100 method. It's called by vmtruncate() - deprecated library function used by
 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 *);
 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         int (*invalidatepage) (struct page *, unsigned long);
 197         int (*releasepage) (struct page *, int);
 198         void (*freepage)(struct page *);
 199         int (*direct_IO)(int, struct kiocb *, const struct iovec *iov,
 200                         loff_t offset, unsigned long nr_segs);
 201         int (*get_xip_mem)(struct address_space *, pgoff_t, int, void **,
 202                                 unsigned long *);
 203         int (*migratepage)(struct address_space *, struct page *, struct page *);
 204         int (*launder_page)(struct page *);
 205         int (*is_partially_uptodate)(struct page *, read_descriptor_t *, unsigned long);
 206         int (*error_remove_page)(struct address_space *, struct page *);
 207         int (*swap_activate)(struct file *);
 208         int (*swap_deactivate)(struct file *);
 209
 210 locking rules:
 211         All except set_page_dirty and freepage may block
 212
 213                         PageLocked(page)        i_mutex
 214 writepage:              yes, unlocks (see below)
 215 readpage:               yes, unlocks
 216 sync_page:              maybe
 217 writepages:
 218 set_page_dirty          no
 219 readpages:
 220 write_begin:            locks the page          yes
 221 write_end:              yes, unlocks            yes
 222 bmap:
 223 invalidatepage:         yes
 224 releasepage:            yes
 225 freepage:               yes
 226 direct_IO:
 227 get_xip_mem:                                    maybe
 228 migratepage:            yes (both)
 229 launder_page:           yes
 230 is_partially_uptodate:  yes
 231 error_remove_page:      yes
 232 swap_activate:          no
 233 swap_deactivate:        no
 234
 235         ->write_begin(), ->write_end(), ->sync_page() and ->readpage()
 236 may be called from the request handler (/dev/loop).
 237
 238         ->readpage() unlocks the page, either synchronously or via I/O
 239 completion.
 240
 241         ->readpages() populates the pagecache with the passed pages and starts
 242 I/O against them.  They come unlocked upon I/O completion.
 243
 244         ->writepage() is used for two purposes: for "memory cleansing" and for
 245 "sync".  These are quite different operations and the behaviour may differ
 246 depending upon the mode.
 247
 248 If writepage is called for sync (wbc->sync_mode != WBC_SYNC_NONE) then
 249 it *must* start I/O against the page, even if that would involve
 250 blocking on in-progress I/O.
 251
 252 If writepage is called for memory cleansing (sync_mode ==
 253 WBC_SYNC_NONE) then its role is to get as much writeout underway as
 254 possible.  So writepage should try to avoid blocking against
 255 currently-in-progress I/O.
 256
 257 If the filesystem is not called for "sync" and it determines that it
 258 would need to block against in-progress I/O to be able to start new I/O
 259 against the page the filesystem should redirty the page with
 260 redirty_page_for_writepage(), then unlock the page and return zero.
 261 This may also be done to avoid internal deadlocks, but rarely.
 262
 263 If the filesystem is called for sync then it must wait on any
 264 in-progress I/O and then start new I/O.
 265
 266 The filesystem should unlock the page synchronously, before returning to the
 267 caller, unless ->writepage() returns special WRITEPAGE_ACTIVATE
 268 value. WRITEPAGE_ACTIVATE means that page cannot really be written out
 269 currently, and VM should stop calling ->writepage() on this page for some
 270 time. VM does this by moving page to the head of the active list, hence the
 271 name.
 272
 273 Unless the filesystem is going to redirty_page_for_writepage(), unlock the page
 274 and return zero, writepage *must* run set_page_writeback() against the page,
 275 followed by unlocking it.  Once set_page_writeback() has been run against the
 276 page, write I/O can be submitted and the write I/O completion handler must run
 277 end_page_writeback() once the I/O is complete.  If no I/O is submitted, the
 278 filesystem must run end_page_writeback() against the page before returning from
 279 writepage.
 280
 281 That is: after 2.5.12, pages which are under writeout are *not* locked.  Note,
 282 if the filesystem needs the page to be locked during writeout, that is ok, too,
 283 the page is allowed to be unlocked at any point in time between the calls to
 284 set_page_writeback() and end_page_writeback().
 285
 286 Note, failure to run either redirty_page_for_writepage() or the combination of
 287 set_page_writeback()/end_page_writeback() on a page submitted to writepage
 288 will leave the page itself marked clean but it will be tagged as dirty in the
 289 radix tree.  This incoherency can lead to all sorts of hard-to-debug problems
 290 in the filesystem like having dirty inodes at umount and losing written data.
 291
 292         ->sync_page() locking rules are not well-defined - usually it is called
 293 with lock on page, but that is not guaranteed. Considering the currently
 294 existing instances of this method ->sync_page() itself doesn't look
 295 well-defined...
 296
 297         ->writepages() is used for periodic writeback and for syscall-initiated
 298 sync operations.  The address_space should start I/O against at least
 299 *nr_to_write pages.  *nr_to_write must be decremented for each page which is
 300 written.  The address_space implementation may write more (or less) pages
 301 than *nr_to_write asks for, but it should try to be reasonably close.  If
 302 nr_to_write is NULL, all dirty pages must be written.
 303
 304 writepages should _only_ write pages which are present on
 305 mapping->io_pages.
 306
 307         ->set_page_dirty() is called from various places in the kernel
 308 when the target page is marked as needing writeback.  It may be called
 309 under spinlock (it cannot block) and is sometimes called with the page
 310 not locked.
 311
 312         ->bmap() is currently used by legacy ioctl() (FIBMAP) provided by some
 313 filesystems and by the swapper. The latter will eventually go away.  Please,
 314 keep it that way and don't breed new callers.
 315
 316         ->invalidatepage() is called when the filesystem must attempt to drop
 317 some or all of the buffers from the page when it is being truncated.  It
 318 returns zero on success.  If ->invalidatepage is zero, the kernel uses
 319 block_invalidatepage() instead.
 320
 321         ->releasepage() is called when the kernel is about to try to drop the
 322 buffers from the page in preparation for freeing it.  It returns zero to
 323 indicate that the buffers are (or may be) freeable.  If ->releasepage is zero,
 324 the kernel assumes that the fs has no private interest in the buffers.
 325
 326         ->freepage() is called when the kernel is done dropping the page
 327 from the page cache.
 328
 329         ->launder_page() may be called prior to releasing a page if
 330 it is still found to be dirty. It returns zero if the page was successfully
 331 cleaned, or an error value if not. Note that in order to prevent the page
 332 getting mapped back in and redirtied, it needs to be kept locked
 333 across the entire operation.
 334
 335         ->swap_activate will be called with a non-zero argument on
 336 files backing (non block device backed) swapfiles. A return value
 337 of zero indicates success, in which case this file can be used for
 338 backing swapspace. The swapspace operations will be proxied to the
 339 address space operations.
 340
 341         ->swap_deactivate() will be called in the sys_swapoff()
 342 path after ->swap_activate() returned success.
 343
 344 ----------------------- file_lock_operations ------------------------------
 345 prototypes:
 346         void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
 347         void (*fl_release_private)(struct file_lock *);
 348
 349
 350 locking rules:
 351                         file_lock_lock  may block
 352 fl_copy_lock:           yes             no
 353 fl_release_private:     maybe           no
 354
 355 ----------------------- lock_manager_operations ---------------------------
 356 prototypes:
 357         int (*lm_compare_owner)(struct file_lock *, struct file_lock *);
 358         void (*lm_notify)(struct file_lock *);  /* unblock callback */
 359         int (*lm_grant)(struct file_lock *, struct file_lock *, int);
 360         void (*lm_break)(struct file_lock *); /* break_lease callback */
 361         int (*lm_change)(struct file_lock **, int);
 362
 363 locking rules:
 364                         file_lock_lock  may block
 365 lm_compare_owner:       yes             no
 366 lm_notify:              yes             no
 367 lm_grant:               no              no
 368 lm_break:               yes             no
 369 lm_change               yes             no
 370
 371 --------------------------- buffer_head -----------------------------------
 372 prototypes:
 373         void (*b_end_io)(struct buffer_head *bh, int uptodate);
 374
 375 locking rules:
 376         called from interrupts. In other words, extreme care is needed here.
 377 bh is locked, but that's all warranties we have here. Currently only RAID1,
 378 highmem, fs/buffer.c, and fs/ntfs/aops.c are providing these. Block devices
 379 call this method upon the IO completion.
 380
 381 --------------------------- block_device_operations -----------------------
 382 prototypes:
 383         int (*open) (struct block_device *, fmode_t);
 384         int (*release) (struct gendisk *, fmode_t);
 385         int (*ioctl) (struct block_device *, fmode_t, unsigned, unsigned long);
 386         int (*compat_ioctl) (struct block_device *, fmode_t, unsigned, unsigned long);
 387         int (*direct_access) (struct block_device *, sector_t, void **, unsigned long *);
 388         int (*media_changed) (struct gendisk *);
 389         void (*unlock_native_capacity) (struct gendisk *);
 390         int (*revalidate_disk) (struct gendisk *);
 391         int (*getgeo)(struct block_device *, struct hd_geometry *);
 392         void (*swap_slot_free_notify) (struct block_device *, unsigned long);
 393
 394 locking rules:
 395                         bd_mutex
 396 open:                   yes
 397 release:                yes
 398 ioctl:                  no
 399 compat_ioctl:           no
 400 direct_access:          no
 401 media_changed:          no
 402 unlock_native_capacity: no
 403 revalidate_disk:        no
 404 getgeo:                 no
 405 swap_slot_free_notify:  no      (see below)
 406
 407 media_changed, unlock_native_capacity and revalidate_disk are called only from
 408 check_disk_change().
 409
 410 swap_slot_free_notify is called with swap_lock and sometimes the page lock
 411 held.
 412
 413
 414 --------------------------- file_operations -------------------------------
 415 prototypes:
 416         loff_t (*llseek) (struct file *, loff_t, int);
 417         ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
 418         ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
 419         ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
 420         ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
 421         int (*readdir) (struct file *, void *, filldir_t);
 422         unsigned int (*poll) (struct file *, struct poll_table_struct *);
 423         long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
 424         long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
 425         int (*mmap) (struct file *, struct vm_area_struct *);
 426         int (*open) (struct inode *, struct file *);
 427         int (*flush) (struct file *);
 428         int (*release) (struct inode *, struct file *);
 429         int (*fsync) (struct file *, loff_t start, loff_t end, int datasync);
 430         int (*aio_fsync) (struct kiocb *, int datasync);
 431         int (*fasync) (int, struct file *, int);
 432         int (*lock) (struct file *, int, struct file_lock *);
 433         ssize_t (*readv) (struct file *, const struct iovec *, unsigned long,
 434                         loff_t *);
 435         ssize_t (*writev) (struct file *, const struct iovec *, unsigned long,
 436                         loff_t *);
 437         ssize_t (*sendfile) (struct file *, loff_t *, size_t, read_actor_t,
 438                         void __user *);
 439         ssize_t (*sendpage) (struct file *, struct page *, int, size_t,
 440                         loff_t *, int);
 441         unsigned long (*get_unmapped_area)(struct file *, unsigned long,
 442                         unsigned long, unsigned long, unsigned long);
 443         int (*check_flags)(int);
 444         int (*flock) (struct file *, int, struct file_lock *);
 445         ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *,
 446                         size_t, unsigned int);
 447         ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *,
 448                         size_t, unsigned int);
 449         int (*setlease)(struct file *, long, struct file_lock **);
 450         long (*fallocate)(struct file *, int, loff_t, loff_t);
 451 };
 452
 453 locking rules:
 454         All may block except for ->setlease.
 455         No VFS locks held on entry except for ->setlease.
 456
 457 ->setlease has the file_list_lock held and must not sleep.
 458
 459 ->llseek() locking has moved from llseek to the individual llseek
 460 implementations.  If your fs is not using generic_file_llseek, you
 461 need to acquire and release the appropriate locks in your ->llseek().
 462 For many filesystems, it is probably safe to acquire the inode
 463 mutex or just to use i_size_read() instead.
 464 Note: this does not protect the file->f_pos against concurrent modifications
 465 since this is something the userspace has to take care about.
 466
 467 ->fasync() is responsible for maintaining the FASYNC bit in filp->f_flags.
 468 Most instances call fasync_helper(), which does that maintenance, so it's
 469 not normally something one needs to worry about.  Return values > 0 will be
 470 mapped to zero in the VFS layer.
 471
 472 ->readdir() and ->ioctl() on directories must be changed. Ideally we would
 473 move ->readdir() to inode_operations and use a separate method for directory
 474 ->ioctl() or kill the latter completely. One of the problems is that for
 475 anything that resembles union-mount we won't have a struct file for all
 476 components. And there are other reasons why the current interface is a mess...
 477
 478 ->read on directories probably must go away - we should just enforce -EISDIR
 479 in sys_read() and friends.
 480
 481 --------------------------- dquot_operations -------------------------------
 482 prototypes:
 483         int (*write_dquot) (struct dquot *);
 484         int (*acquire_dquot) (struct dquot *);
 485         int (*release_dquot) (struct dquot *);
 486         int (*mark_dirty) (struct dquot *);
 487         int (*write_info) (struct super_block *, int);
 488
 489 These operations are intended to be more or less wrapping functions that ensure
 490 a proper locking wrt the filesystem and call the generic quota operations.
 491
 492 What filesystem should expect from the generic quota functions:
 493
 494                 FS recursion    Held locks when called
 495 write_dquot:    yes             dqonoff_sem or dqptr_sem
 496 acquire_dquot:  yes             dqonoff_sem or dqptr_sem
 497 release_dquot:  yes             dqonoff_sem or dqptr_sem
 498 mark_dirty:     no              -
 499 write_info:     yes             dqonoff_sem
 500
 501 FS recursion means calling ->quota_read() and ->quota_write() from superblock
 502 operations.
 503
 504 More details about quota locking can be found in fs/dquot.c.
 505
 506 --------------------------- vm_operations_struct -----------------------------
 507 prototypes:
 508         void (*open)(struct vm_area_struct*);
 509         void (*close)(struct vm_area_struct*);
 510         int (*fault)(struct vm_area_struct*, struct vm_fault *);
 511         int (*page_mkwrite)(struct vm_area_struct *, struct vm_fault *);
 512         int (*access)(struct vm_area_struct *, unsigned long, void*, int, int);
 513
 514 locking rules:
 515                 mmap_sem        PageLocked(page)
 516 open:           yes
 517 close:          yes
 518 fault:          yes             can return with page locked
 519 page_mkwrite:   yes             can return with page locked
 520 access:         yes
 521
 522         ->fault() is called when a previously not present pte is about
 523 to be faulted in. The filesystem must find and return the page associated
 524 with the passed in "pgoff" in the vm_fault structure. If it is possible that
 525 the page may be truncated and/or invalidated, then the filesystem must lock
 526 the page, then ensure it is not already truncated (the page lock will block
 527 subsequent truncate), and then return with VM_FAULT_LOCKED, and the page
 528 locked. The VM will unlock the page.
 529
 530         ->page_mkwrite() is called when a previously read-only pte is
 531 about to become writeable. The filesystem again must ensure that there are
 532 no truncate/invalidate races, and then return with the page locked. If
 533 the page has been truncated, the filesystem should not look up a new page
 534 like the ->fault() handler, but simply return with VM_FAULT_NOPAGE, which
 535 will cause the VM to retry the fault.
 536
 537         ->access() is called when get_user_pages() fails in
 538 acces_process_vm(), typically used to debug a process through
 539 /proc/pid/mem or ptrace.  This function is needed only for
 540 VM_IO | VM_PFNMAP VMAs.
 541
 542 ================================================================================
 543                         Dubious stuff
 544
 545 (if you break something or notice that it is broken and do not fix it yourself
 546 - at least put it here)