1.. SPDX-License-Identifier: GPL-2.0 2 3========================================= 4Overview of the Linux Virtual File System 5========================================= 6 7Original author: Richard Gooch <rgooch@atnf.csiro.au> 8 9- Copyright (C) 1999 Richard Gooch 10- Copyright (C) 2005 Pekka Enberg 11 12 13Introduction 14============ 15 16The Virtual File System (also known as the Virtual Filesystem Switch) is 17the software layer in the kernel that provides the filesystem interface 18to userspace programs. It also provides an abstraction within the 19kernel which allows different filesystem implementations to coexist. 20 21VFS system calls open(2), stat(2), read(2), write(2), chmod(2) and so on 22are called from a process context. Filesystem locking is described in 23the document Documentation/filesystems/locking.rst. 24 25 26Directory Entry Cache (dcache) 27------------------------------ 28 29The VFS implements the open(2), stat(2), chmod(2), and similar system 30calls. The pathname argument that is passed to them is used by the VFS 31to search through the directory entry cache (also known as the dentry 32cache or dcache). This provides a very fast look-up mechanism to 33translate a pathname (filename) into a specific dentry. Dentries live 34in RAM and are never saved to disc: they exist only for performance. 35 36The dentry cache is meant to be a view into your entire filespace. As 37most computers cannot fit all dentries in the RAM at the same time, some 38bits of the cache are missing. In order to resolve your pathname into a 39dentry, the VFS may have to resort to creating dentries along the way, 40and then loading the inode. This is done by looking up the inode. 41 42 43The Inode Object 44---------------- 45 46An individual dentry usually has a pointer to an inode. Inodes are 47filesystem objects such as regular files, directories, FIFOs and other 48beasts. They live either on the disc (for block device filesystems) or 49in the memory (for pseudo filesystems). Inodes that live on the disc 50are copied into the memory when required and changes to the inode are 51written back to disc. A single inode can be pointed to by multiple 52dentries (hard links, for example, do this). 53 54To look up an inode requires that the VFS calls the lookup() method of 55the parent directory inode. This method is installed by the specific 56filesystem implementation that the inode lives in. Once the VFS has the 57required dentry (and hence the inode), we can do all those boring things 58like open(2) the file, or stat(2) it to peek at the inode data. The 59stat(2) operation is fairly simple: once the VFS has the dentry, it 60peeks at the inode data and passes some of it back to userspace. 61 62 63The File Object 64--------------- 65 66Opening a file requires another operation: allocation of a file 67structure (this is the kernel-side implementation of file descriptors). 68The freshly allocated file structure is initialized with a pointer to 69the dentry and a set of file operation member functions. These are 70taken from the inode data. The open() file method is then called so the 71specific filesystem implementation can do its work. You can see that 72this is another switch performed by the VFS. The file structure is 73placed into the file descriptor table for the process. 74 75Reading, writing and closing files (and other assorted VFS operations) 76is done by using the userspace file descriptor to grab the appropriate 77file structure, and then calling the required file structure method to 78do whatever is required. For as long as the file is open, it keeps the 79dentry in use, which in turn means that the VFS inode is still in use. 80 81 82Registering and Mounting a Filesystem 83===================================== 84 85To register and unregister a filesystem, use the following API 86functions: 87 88.. code-block:: c 89 90 #include <linux/fs.h> 91 92 extern int register_filesystem(struct file_system_type *); 93 extern int unregister_filesystem(struct file_system_type *); 94 95The passed struct file_system_type describes your filesystem. When a 96request is made to mount a filesystem onto a directory in your 97namespace, the VFS will call the appropriate mount() method for the 98specific filesystem. New vfsmount referring to the tree returned by 99->mount() will be attached to the mountpoint, so that when pathname 100resolution reaches the mountpoint it will jump into the root of that 101vfsmount. 102 103You can see all filesystems that are registered to the kernel in the 104file /proc/filesystems. 105 106 107struct file_system_type 108----------------------- 109 110This describes the filesystem. The following 111members are defined: 112 113.. code-block:: c 114 115 struct file_system_type { 116 const char *name; 117 int fs_flags; 118 int (*init_fs_context)(struct fs_context *); 119 const struct fs_parameter_spec *parameters; 120 struct dentry *(*mount) (struct file_system_type *, int, 121 const char *, void *); 122 void (*kill_sb) (struct super_block *); 123 struct module *owner; 124 struct file_system_type * next; 125 struct hlist_head fs_supers; 126 127 struct lock_class_key s_lock_key; 128 struct lock_class_key s_umount_key; 129 struct lock_class_key s_vfs_rename_key; 130 struct lock_class_key s_writers_key[SB_FREEZE_LEVELS]; 131 132 struct lock_class_key i_lock_key; 133 struct lock_class_key i_mutex_key; 134 struct lock_class_key invalidate_lock_key; 135 struct lock_class_key i_mutex_dir_key; 136 }; 137 138``name`` 139 the name of the filesystem type, such as "ext2", "iso9660", 140 "msdos" and so on 141 142``fs_flags`` 143 various flags (i.e. FS_REQUIRES_DEV, FS_NO_DCACHE, etc.) 144 145``init_fs_context`` 146 Initializes 'struct fs_context' ->ops and ->fs_private fields with 147 filesystem-specific data. 148 149``parameters`` 150 Pointer to the array of filesystem parameters descriptors 151 'struct fs_parameter_spec'. 152 More info in Documentation/filesystems/mount_api.rst. 153 154``mount`` 155 the method to call when a new instance of this filesystem should 156 be mounted 157 158``kill_sb`` 159 the method to call when an instance of this filesystem should be 160 shut down 161 162 163``owner`` 164 for internal VFS use: you should initialize this to THIS_MODULE 165 in most cases. 166 167``next`` 168 for internal VFS use: you should initialize this to NULL 169 170``fs_supers`` 171 for internal VFS use: hlist of filesystem instances (superblocks) 172 173 s_lock_key, s_umount_key, s_vfs_rename_key, s_writers_key, 174 i_lock_key, i_mutex_key, invalidate_lock_key, i_mutex_dir_key: lockdep-specific 175 176The mount() method has the following arguments: 177 178``struct file_system_type *fs_type`` 179 describes the filesystem, partly initialized by the specific 180 filesystem code 181 182``int flags`` 183 mount flags 184 185``const char *dev_name`` 186 the device name we are mounting. 187 188``void *data`` 189 arbitrary mount options, usually comes as an ASCII string (see 190 "Mount Options" section) 191 192The mount() method must return the root dentry of the tree requested by 193caller. An active reference to its superblock must be grabbed and the 194superblock must be locked. On failure it should return ERR_PTR(error). 195 196The arguments match those of mount(2) and their interpretation depends 197on filesystem type. E.g. for block filesystems, dev_name is interpreted 198as block device name, that device is opened and if it contains a 199suitable filesystem image the method creates and initializes struct 200super_block accordingly, returning its root dentry to caller. 201 202->mount() may choose to return a subtree of existing filesystem - it 203doesn't have to create a new one. The main result from the caller's 204point of view is a reference to dentry at the root of (sub)tree to be 205attached; creation of new superblock is a common side effect. 206 207The most interesting member of the superblock structure that the mount() 208method fills in is the "s_op" field. This is a pointer to a "struct 209super_operations" which describes the next level of the filesystem 210implementation. 211 212Usually, a filesystem uses one of the generic mount() implementations 213and provides a fill_super() callback instead. The generic variants are: 214 215``mount_bdev`` 216 mount a filesystem residing on a block device 217 218``mount_nodev`` 219 mount a filesystem that is not backed by a device 220 221``mount_single`` 222 mount a filesystem which shares the instance between all mounts 223 224A fill_super() callback implementation has the following arguments: 225 226``struct super_block *sb`` 227 the superblock structure. The callback must initialize this 228 properly. 229 230``void *data`` 231 arbitrary mount options, usually comes as an ASCII string (see 232 "Mount Options" section) 233 234``int silent`` 235 whether or not to be silent on error 236 237 238The Superblock Object 239===================== 240 241A superblock object represents a mounted filesystem. 242 243 244struct super_operations 245----------------------- 246 247This describes how the VFS can manipulate the superblock of your 248filesystem. The following members are defined: 249 250.. code-block:: c 251 252 struct super_operations { 253 struct inode *(*alloc_inode)(struct super_block *sb); 254 void (*destroy_inode)(struct inode *); 255 void (*free_inode)(struct inode *); 256 257 void (*dirty_inode) (struct inode *, int flags); 258 int (*write_inode) (struct inode *, struct writeback_control *wbc); 259 int (*drop_inode) (struct inode *); 260 void (*evict_inode) (struct inode *); 261 void (*put_super) (struct super_block *); 262 int (*sync_fs)(struct super_block *sb, int wait); 263 int (*freeze_super) (struct super_block *sb, 264 enum freeze_holder who); 265 int (*freeze_fs) (struct super_block *); 266 int (*thaw_super) (struct super_block *sb, 267 enum freeze_wholder who); 268 int (*unfreeze_fs) (struct super_block *); 269 int (*statfs) (struct dentry *, struct kstatfs *); 270 int (*remount_fs) (struct super_block *, int *, char *); 271 void (*umount_begin) (struct super_block *); 272 273 int (*show_options)(struct seq_file *, struct dentry *); 274 int (*show_devname)(struct seq_file *, struct dentry *); 275 int (*show_path)(struct seq_file *, struct dentry *); 276 int (*show_stats)(struct seq_file *, struct dentry *); 277 278 ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t); 279 ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t); 280 struct dquot **(*get_dquots)(struct inode *); 281 282 long (*nr_cached_objects)(struct super_block *, 283 struct shrink_control *); 284 long (*free_cached_objects)(struct super_block *, 285 struct shrink_control *); 286 }; 287 288All methods are called without any locks being held, unless otherwise 289noted. This means that most methods can block safely. All methods are 290only called from a process context (i.e. not from an interrupt handler 291or bottom half). 292 293``alloc_inode`` 294 this method is called by alloc_inode() to allocate memory for 295 struct inode and initialize it. If this function is not 296 defined, a simple 'struct inode' is allocated. Normally 297 alloc_inode will be used to allocate a larger structure which 298 contains a 'struct inode' embedded within it. 299 300``destroy_inode`` 301 this method is called by destroy_inode() to release resources 302 allocated for struct inode. It is only required if 303 ->alloc_inode was defined and simply undoes anything done by 304 ->alloc_inode. 305 306``free_inode`` 307 this method is called from RCU callback. If you use call_rcu() 308 in ->destroy_inode to free 'struct inode' memory, then it's 309 better to release memory in this method. 310 311``dirty_inode`` 312 this method is called by the VFS when an inode is marked dirty. 313 This is specifically for the inode itself being marked dirty, 314 not its data. If the update needs to be persisted by fdatasync(), 315 then I_DIRTY_DATASYNC will be set in the flags argument. 316 I_DIRTY_TIME will be set in the flags in case lazytime is enabled 317 and struct inode has times updated since the last ->dirty_inode 318 call. 319 320``write_inode`` 321 this method is called when the VFS needs to write an inode to 322 disc. The second parameter indicates whether the write should 323 be synchronous or not, not all filesystems check this flag. 324 325``drop_inode`` 326 called when the last access to the inode is dropped, with the 327 inode->i_lock spinlock held. 328 329 This method should be either NULL (normal UNIX filesystem 330 semantics) or "generic_delete_inode" (for filesystems that do 331 not want to cache inodes - causing "delete_inode" to always be 332 called regardless of the value of i_nlink) 333 334 The "generic_delete_inode()" behavior is equivalent to the old 335 practice of using "force_delete" in the put_inode() case, but 336 does not have the races that the "force_delete()" approach had. 337 338``evict_inode`` 339 called when the VFS wants to evict an inode. Caller does 340 *not* evict the pagecache or inode-associated metadata buffers; 341 the method has to use truncate_inode_pages_final() to get rid 342 of those. Caller makes sure async writeback cannot be running for 343 the inode while (or after) ->evict_inode() is called. Optional. 344 345``put_super`` 346 called when the VFS wishes to free the superblock 347 (i.e. unmount). This is called with the superblock lock held 348 349``sync_fs`` 350 called when VFS is writing out all dirty data associated with a 351 superblock. The second parameter indicates whether the method 352 should wait until the write out has been completed. Optional. 353 354``freeze_super`` 355 Called instead of ->freeze_fs callback if provided. 356 Main difference is that ->freeze_super is called without taking 357 down_write(&sb->s_umount). If filesystem implements it and wants 358 ->freeze_fs to be called too, then it has to call ->freeze_fs 359 explicitly from this callback. Optional. 360 361``freeze_fs`` 362 called when VFS is locking a filesystem and forcing it into a 363 consistent state. This method is currently used by the Logical 364 Volume Manager (LVM) and ioctl(FIFREEZE). Optional. 365 366``thaw_super`` 367 called when VFS is unlocking a filesystem and making it writable 368 again after ->freeze_super. Optional. 369 370``unfreeze_fs`` 371 called when VFS is unlocking a filesystem and making it writable 372 again after ->freeze_fs. Optional. 373 374``statfs`` 375 called when the VFS needs to get filesystem statistics. 376 377``remount_fs`` 378 called when the filesystem is remounted. This is called with 379 the kernel lock held 380 381``umount_begin`` 382 called when the VFS is unmounting a filesystem. 383 384``show_options`` 385 called by the VFS to show mount options for /proc/<pid>/mounts 386 and /proc/<pid>/mountinfo. 387 (see "Mount Options" section) 388 389``show_devname`` 390 Optional. Called by the VFS to show device name for 391 /proc/<pid>/{mounts,mountinfo,mountstats}. If not provided then 392 '(struct mount).mnt_devname' will be used. 393 394``show_path`` 395 Optional. Called by the VFS (for /proc/<pid>/mountinfo) to show 396 the mount root dentry path relative to the filesystem root. 397 398``show_stats`` 399 Optional. Called by the VFS (for /proc/<pid>/mountstats) to show 400 filesystem-specific mount statistics. 401 402``quota_read`` 403 called by the VFS to read from filesystem quota file. 404 405``quota_write`` 406 called by the VFS to write to filesystem quota file. 407 408``get_dquots`` 409 called by quota to get 'struct dquot' array for a particular inode. 410 Optional. 411 412``nr_cached_objects`` 413 called by the sb cache shrinking function for the filesystem to 414 return the number of freeable cached objects it contains. 415 Optional. 416 417``free_cache_objects`` 418 called by the sb cache shrinking function for the filesystem to 419 scan the number of objects indicated to try to free them. 420 Optional, but any filesystem implementing this method needs to 421 also implement ->nr_cached_objects for it to be called 422 correctly. 423 424 We can't do anything with any errors that the filesystem might 425 encountered, hence the void return type. This will never be 426 called if the VM is trying to reclaim under GFP_NOFS conditions, 427 hence this method does not need to handle that situation itself. 428 429 Implementations must include conditional reschedule calls inside 430 any scanning loop that is done. This allows the VFS to 431 determine appropriate scan batch sizes without having to worry 432 about whether implementations will cause holdoff problems due to 433 large scan batch sizes. 434 435Whoever sets up the inode is responsible for filling in the "i_op" 436field. This is a pointer to a "struct inode_operations" which describes 437the methods that can be performed on individual inodes. 438 439 440struct xattr_handler 441--------------------- 442 443On filesystems that support extended attributes (xattrs), the s_xattr 444superblock field points to a NULL-terminated array of xattr handlers. 445Extended attributes are name:value pairs. 446 447``name`` 448 Indicates that the handler matches attributes with the specified 449 name (such as "system.posix_acl_access"); the prefix field must 450 be NULL. 451 452``prefix`` 453 Indicates that the handler matches all attributes with the 454 specified name prefix (such as "user."); the name field must be 455 NULL. 456 457``list`` 458 Determine if attributes matching this xattr handler should be 459 listed for a particular dentry. Used by some listxattr 460 implementations like generic_listxattr. 461 462``get`` 463 Called by the VFS to get the value of a particular extended 464 attribute. This method is called by the getxattr(2) system 465 call. 466 467``set`` 468 Called by the VFS to set the value of a particular extended 469 attribute. When the new value is NULL, called to remove a 470 particular extended attribute. This method is called by the 471 setxattr(2) and removexattr(2) system calls. 472 473When none of the xattr handlers of a filesystem match the specified 474attribute name or when a filesystem doesn't support extended attributes, 475the various ``*xattr(2)`` system calls return -EOPNOTSUPP. 476 477 478The Inode Object 479================ 480 481An inode object represents an object within the filesystem. 482 483 484struct inode_operations 485----------------------- 486 487This describes how the VFS can manipulate an inode in your filesystem. 488As of kernel 2.6.22, the following members are defined: 489 490.. code-block:: c 491 492 struct inode_operations { 493 int (*create) (struct mnt_idmap *, struct inode *,struct dentry *, umode_t, bool); 494 struct dentry * (*lookup) (struct inode *,struct dentry *, unsigned int); 495 int (*link) (struct dentry *,struct inode *,struct dentry *); 496 int (*unlink) (struct inode *,struct dentry *); 497 int (*symlink) (struct mnt_idmap *, struct inode *,struct dentry *,const char *); 498 struct dentry *(*mkdir) (struct mnt_idmap *, struct inode *,struct dentry *,umode_t); 499 int (*rmdir) (struct inode *,struct dentry *); 500 int (*mknod) (struct mnt_idmap *, struct inode *,struct dentry *,umode_t,dev_t); 501 int (*rename) (struct mnt_idmap *, struct inode *, struct dentry *, 502 struct inode *, struct dentry *, unsigned int); 503 int (*readlink) (struct dentry *, char __user *,int); 504 const char *(*get_link) (struct dentry *, struct inode *, 505 struct delayed_call *); 506 int (*permission) (struct mnt_idmap *, struct inode *, int); 507 struct posix_acl * (*get_inode_acl)(struct inode *, int, bool); 508 int (*setattr) (struct mnt_idmap *, struct dentry *, struct iattr *); 509 int (*getattr) (struct mnt_idmap *, const struct path *, struct kstat *, u32, unsigned int); 510 ssize_t (*listxattr) (struct dentry *, char *, size_t); 511 void (*update_time)(struct inode *, struct timespec *, int); 512 int (*atomic_open)(struct inode *, struct dentry *, struct file *, 513 unsigned open_flag, umode_t create_mode); 514 int (*tmpfile) (struct mnt_idmap *, struct inode *, struct file *, umode_t); 515 struct posix_acl * (*get_acl)(struct mnt_idmap *, struct dentry *, int); 516 int (*set_acl)(struct mnt_idmap *, struct dentry *, struct posix_acl *, int); 517 int (*fileattr_set)(struct mnt_idmap *idmap, 518 struct dentry *dentry, struct file_kattr *fa); 519 int (*fileattr_get)(struct dentry *dentry, struct file_kattr *fa); 520 struct offset_ctx *(*get_offset_ctx)(struct inode *inode); 521 }; 522 523Again, all methods are called without any locks being held, unless 524otherwise noted. 525 526``create`` 527 called by the open(2) and creat(2) system calls. Only required 528 if you want to support regular files. The dentry you get should 529 not have an inode (i.e. it should be a negative dentry). Here 530 you will probably call d_instantiate() with the dentry and the 531 newly created inode 532 533``lookup`` 534 called when the VFS needs to look up an inode in a parent 535 directory. The name to look for is found in the dentry. This 536 method must call d_add() to insert the found inode into the 537 dentry. The "i_count" field in the inode structure should be 538 incremented. If the named inode does not exist a NULL inode 539 should be inserted into the dentry (this is called a negative 540 dentry). Returning an error code from this routine must only be 541 done on a real error, otherwise creating inodes with system 542 calls like create(2), mknod(2), mkdir(2) and so on will fail. 543 If you wish to overload the dentry methods then you should 544 initialise the "d_dop" field in the dentry; this is a pointer to 545 a struct "dentry_operations". This method is called with the 546 directory inode semaphore held 547 548``link`` 549 called by the link(2) system call. Only required if you want to 550 support hard links. You will probably need to call 551 d_instantiate() just as you would in the create() method 552 553``unlink`` 554 called by the unlink(2) system call. Only required if you want 555 to support deleting inodes 556 557``symlink`` 558 called by the symlink(2) system call. Only required if you want 559 to support symlinks. You will probably need to call 560 d_instantiate() just as you would in the create() method 561 562``mkdir`` 563 called by the mkdir(2) system call. Only required if you want 564 to support creating subdirectories. You will probably need to 565 call d_instantiate_new() just as you would in the create() method. 566 567 If d_instantiate_new() is not used and if the fh_to_dentry() 568 export operation is provided, or if the storage might be 569 accessible by another path (e.g. with a network filesystem) 570 then more care may be needed. Importantly d_instantate() 571 should not be used with an inode that is no longer I_NEW if there 572 any chance that the inode could already be attached to a dentry. 573 This is because of a hard rule in the VFS that a directory must 574 only ever have one dentry. 575 576 For example, if an NFS filesystem is mounted twice the new directory 577 could be visible on the other mount before it is on the original 578 mount, and a pair of name_to_handle_at(), open_by_handle_at() 579 calls could instantiate the directory inode with an IS_ROOT() 580 dentry before the first mkdir returns. 581 582 If there is any chance this could happen, then the new inode 583 should be d_drop()ed and attached with d_splice_alias(). The 584 returned dentry (if any) should be returned by ->mkdir(). 585 586``rmdir`` 587 called by the rmdir(2) system call. Only required if you want 588 to support deleting subdirectories 589 590``mknod`` 591 called by the mknod(2) system call to create a device (char, 592 block) inode or a named pipe (FIFO) or socket. Only required if 593 you want to support creating these types of inodes. You will 594 probably need to call d_instantiate() just as you would in the 595 create() method 596 597``rename`` 598 called by the rename(2) system call to rename the object to have 599 the parent and name given by the second inode and dentry. 600 601 The filesystem must return -EINVAL for any unsupported or 602 unknown flags. Currently the following flags are implemented: 603 (1) RENAME_NOREPLACE: this flag indicates that if the target of 604 the rename exists the rename should fail with -EEXIST instead of 605 replacing the target. The VFS already checks for existence, so 606 for local filesystems the RENAME_NOREPLACE implementation is 607 equivalent to plain rename. 608 (2) RENAME_EXCHANGE: exchange source and target. Both must 609 exist; this is checked by the VFS. Unlike plain rename, source 610 and target may be of different type. 611 612``get_link`` 613 called by the VFS to follow a symbolic link to the inode it 614 points to. Only required if you want to support symbolic links. 615 This method returns the symlink body to traverse (and possibly 616 resets the current position with nd_jump_link()). If the body 617 won't go away until the inode is gone, nothing else is needed; 618 if it needs to be otherwise pinned, arrange for its release by 619 having get_link(..., ..., done) do set_delayed_call(done, 620 destructor, argument). In that case destructor(argument) will 621 be called once VFS is done with the body you've returned. May 622 be called in RCU mode; that is indicated by NULL dentry 623 argument. If request can't be handled without leaving RCU mode, 624 have it return ERR_PTR(-ECHILD). 625 626 If the filesystem stores the symlink target in ->i_link, the 627 VFS may use it directly without calling ->get_link(); however, 628 ->get_link() must still be provided. ->i_link must not be 629 freed until after an RCU grace period. Writing to ->i_link 630 post-iget() time requires a 'release' memory barrier. 631 632``readlink`` 633 this is now just an override for use by readlink(2) for the 634 cases when ->get_link uses nd_jump_link() or object is not in 635 fact a symlink. Normally filesystems should only implement 636 ->get_link for symlinks and readlink(2) will automatically use 637 that. 638 639``permission`` 640 called by the VFS to check for access rights on a POSIX-like 641 filesystem. 642 643 May be called in rcu-walk mode (mask & MAY_NOT_BLOCK). If in 644 rcu-walk mode, the filesystem must check the permission without 645 blocking or storing to the inode. 646 647 If a situation is encountered that rcu-walk cannot handle, 648 return 649 -ECHILD and it will be called again in ref-walk mode. 650 651``setattr`` 652 called by the VFS to set attributes for a file. This method is 653 called by chmod(2) and related system calls. 654 655``getattr`` 656 called by the VFS to get attributes of a file. This method is 657 called by stat(2) and related system calls. 658 659``listxattr`` 660 called by the VFS to list all extended attributes for a given 661 file. This method is called by the listxattr(2) system call. 662 663``update_time`` 664 called by the VFS to update a specific time or the i_version of 665 an inode. If this is not defined the VFS will update the inode 666 itself and call mark_inode_dirty_sync. 667 668``atomic_open`` 669 called on the last component of an open. Using this optional 670 method the filesystem can look up, possibly create and open the 671 file in one atomic operation. If it wants to leave actual 672 opening to the caller (e.g. if the file turned out to be a 673 symlink, device, or just something filesystem won't do atomic 674 open for), it may signal this by returning finish_no_open(file, 675 dentry). This method is only called if the last component is 676 negative or needs lookup. Cached positive dentries are still 677 handled by f_op->open(). If the file was created, FMODE_CREATED 678 flag should be set in file->f_mode. In case of O_EXCL the 679 method must only succeed if the file didn't exist and hence 680 FMODE_CREATED shall always be set on success. 681 682``tmpfile`` 683 called in the end of O_TMPFILE open(). Optional, equivalent to 684 atomically creating, opening and unlinking a file in given 685 directory. On success needs to return with the file already 686 open; this can be done by calling finish_open_simple() right at 687 the end. 688 689``fileattr_get`` 690 called on ioctl(FS_IOC_GETFLAGS) and ioctl(FS_IOC_FSGETXATTR) to 691 retrieve miscellaneous file flags and attributes. Also called 692 before the relevant SET operation to check what is being changed 693 (in this case with i_rwsem locked exclusive). If unset, then 694 fall back to f_op->ioctl(). 695 696``fileattr_set`` 697 called on ioctl(FS_IOC_SETFLAGS) and ioctl(FS_IOC_FSSETXATTR) to 698 change miscellaneous file flags and attributes. Callers hold 699 i_rwsem exclusive. If unset, then fall back to f_op->ioctl(). 700``get_offset_ctx`` 701 called to get the offset context for a directory inode. A 702 filesystem must define this operation to use 703 simple_offset_dir_operations. 704 705The Address Space Object 706======================== 707 708The address space object is used to group and manage pages in the page 709cache. It can be used to keep track of the pages in a file (or anything 710else) and also track the mapping of sections of the file into process 711address spaces. 712 713There are a number of distinct yet related services that an 714address-space can provide. These include communicating memory pressure, 715page lookup by address, and keeping track of pages tagged as Dirty or 716Writeback. 717 718The first can be used independently to the others. The VM can try to 719release clean pages in order to reuse them. To do this it can call 720->release_folio on clean folios with the private 721flag set. Clean pages without PagePrivate and with no external references 722will be released without notice being given to the address_space. 723 724To achieve this functionality, pages need to be placed on an LRU with 725lru_cache_add and mark_page_active needs to be called whenever the page 726is used. 727 728Pages are normally kept in a radix tree index by ->index. This tree 729maintains information about the PG_Dirty and PG_Writeback status of each 730page, so that pages with either of these flags can be found quickly. 731 732The Dirty tag is primarily used by mpage_writepages - the default 733->writepages method. It uses the tag to find dirty pages to 734write back. If mpage_writepages is not used (i.e. the address 735provides its own ->writepages) , the PAGECACHE_TAG_DIRTY tag is almost 736unused. write_inode_now and sync_inode do use it (through 737__sync_single_inode) to check if ->writepages has been successful in 738writing out the whole address_space. 739 740The Writeback tag is used by filemap*wait* and sync_page* functions, via 741filemap_fdatawait_range, to wait for all writeback to complete. 742 743An address_space handler may attach extra information to a page, 744typically using the 'private' field in the 'struct page'. If such 745information is attached, the PG_Private flag should be set. This will 746cause various VM routines to make extra calls into the address_space 747handler to deal with that data. 748 749An address space acts as an intermediate between storage and 750application. Data is read into the address space a whole page at a 751time, and provided to the application either by copying of the page, or 752by memory-mapping the page. Data is written into the address space by 753the application, and then written-back to storage typically in whole 754pages, however the address_space has finer control of write sizes. 755 756The read process essentially only requires 'read_folio'. The write 757process is more complicated and uses write_begin/write_end or 758dirty_folio to write data into the address_space, and 759writepages to writeback data to storage. 760 761Removing pages from an address_space requires holding the inode's i_rwsem 762exclusively, while adding pages to the address_space requires holding the 763inode's i_mapping->invalidate_lock exclusively. 764 765When data is written to a page, the PG_Dirty flag should be set. It 766typically remains set until writepages asks for it to be written. This 767should clear PG_Dirty and set PG_Writeback. It can be actually written 768at any point after PG_Dirty is clear. Once it is known to be safe, 769PG_Writeback is cleared. 770 771Writeback makes use of a writeback_control structure to direct the 772operations. This gives the writepages operation some 773information about the nature of and reason for the writeback request, 774and the constraints under which it is being done. It is also used to 775return information back to the caller about the result of a 776writepages request. 777 778 779Handling errors during writeback 780-------------------------------- 781 782Most applications that do buffered I/O will periodically call a file 783synchronization call (fsync, fdatasync, msync or sync_file_range) to 784ensure that data written has made it to the backing store. When there 785is an error during writeback, they expect that error to be reported when 786a file sync request is made. After an error has been reported on one 787request, subsequent requests on the same file descriptor should return 7880, unless further writeback errors have occurred since the previous file 789synchronization. 790 791Ideally, the kernel would report errors only on file descriptions on 792which writes were done that subsequently failed to be written back. The 793generic pagecache infrastructure does not track the file descriptions 794that have dirtied each individual page however, so determining which 795file descriptors should get back an error is not possible. 796 797Instead, the generic writeback error tracking infrastructure in the 798kernel settles for reporting errors to fsync on all file descriptions 799that were open at the time that the error occurred. In a situation with 800multiple writers, all of them will get back an error on a subsequent 801fsync, even if all of the writes done through that particular file 802descriptor succeeded (or even if there were no writes on that file 803descriptor at all). 804 805Filesystems that wish to use this infrastructure should call 806mapping_set_error to record the error in the address_space when it 807occurs. Then, after writing back data from the pagecache in their 808file->fsync operation, they should call file_check_and_advance_wb_err to 809ensure that the struct file's error cursor has advanced to the correct 810point in the stream of errors emitted by the backing device(s). 811 812 813struct address_space_operations 814------------------------------- 815 816This describes how the VFS can manipulate mapping of a file to page 817cache in your filesystem. The following members are defined: 818 819.. code-block:: c 820 821 struct address_space_operations { 822 int (*read_folio)(struct file *, struct folio *); 823 int (*writepages)(struct address_space *, struct writeback_control *); 824 bool (*dirty_folio)(struct address_space *, struct folio *); 825 void (*readahead)(struct readahead_control *); 826 int (*write_begin)(const struct kiocb *, struct address_space *mapping, 827 loff_t pos, unsigned len, 828 struct page **pagep, void **fsdata); 829 int (*write_end)(const struct kiocb *, struct address_space *mapping, 830 loff_t pos, unsigned len, unsigned copied, 831 struct folio *folio, void *fsdata); 832 sector_t (*bmap)(struct address_space *, sector_t); 833 void (*invalidate_folio) (struct folio *, size_t start, size_t len); 834 bool (*release_folio)(struct folio *, gfp_t); 835 void (*free_folio)(struct folio *); 836 ssize_t (*direct_IO)(struct kiocb *, struct iov_iter *iter); 837 int (*migrate_folio)(struct mapping *, struct folio *dst, 838 struct folio *src, enum migrate_mode); 839 int (*launder_folio) (struct folio *); 840 841 bool (*is_partially_uptodate) (struct folio *, size_t from, 842 size_t count); 843 void (*is_dirty_writeback)(struct folio *, bool *, bool *); 844 int (*error_remove_folio)(struct mapping *mapping, struct folio *); 845 int (*swap_activate)(struct swap_info_struct *sis, struct file *f, sector_t *span) 846 int (*swap_deactivate)(struct file *); 847 int (*swap_rw)(struct kiocb *iocb, struct iov_iter *iter); 848 }; 849 850``read_folio`` 851 Called by the page cache to read a folio from the backing store. 852 The 'file' argument supplies authentication information to network 853 filesystems, and is generally not used by block based filesystems. 854 It may be NULL if the caller does not have an open file (eg if 855 the kernel is performing a read for itself rather than on behalf 856 of a userspace process with an open file). 857 858 If the mapping does not support large folios, the folio will 859 contain a single page. The folio will be locked when read_folio 860 is called. If the read completes successfully, the folio should 861 be marked uptodate. The filesystem should unlock the folio 862 once the read has completed, whether it was successful or not. 863 The filesystem does not need to modify the refcount on the folio; 864 the page cache holds a reference count and that will not be 865 released until the folio is unlocked. 866 867 Filesystems may implement ->read_folio() synchronously. 868 In normal operation, folios are read through the ->readahead() 869 method. Only if this fails, or if the caller needs to wait for 870 the read to complete will the page cache call ->read_folio(). 871 Filesystems should not attempt to perform their own readahead 872 in the ->read_folio() operation. 873 874 If the filesystem cannot perform the read at this time, it can 875 unlock the folio, do whatever action it needs to ensure that the 876 read will succeed in the future and return AOP_TRUNCATED_PAGE. 877 In this case, the caller should look up the folio, lock it, 878 and call ->read_folio again. 879 880 Callers may invoke the ->read_folio() method directly, but using 881 read_mapping_folio() will take care of locking, waiting for the 882 read to complete and handle cases such as AOP_TRUNCATED_PAGE. 883 884``writepages`` 885 called by the VM to write out pages associated with the 886 address_space object. If wbc->sync_mode is WB_SYNC_ALL, then 887 the writeback_control will specify a range of pages that must be 888 written out. If it is WB_SYNC_NONE, then a nr_to_write is 889 given and that many pages should be written if possible. If no 890 ->writepages is given, then mpage_writepages is used instead. 891 This will choose pages from the address space that are tagged as 892 DIRTY and will write them back. 893 894``dirty_folio`` 895 called by the VM to mark a folio as dirty. This is particularly 896 needed if an address space attaches private data to a folio, and 897 that data needs to be updated when a folio is dirtied. This is 898 called, for example, when a memory mapped page gets modified. 899 If defined, it should set the folio dirty flag, and the 900 PAGECACHE_TAG_DIRTY search mark in i_pages. 901 902``readahead`` 903 Called by the VM to read pages associated with the address_space 904 object. The pages are consecutive in the page cache and are 905 locked. The implementation should decrement the page refcount 906 after starting I/O on each page. Usually the page will be 907 unlocked by the I/O completion handler. The set of pages are 908 divided into some sync pages followed by some async pages, 909 rac->ra->async_size gives the number of async pages. The 910 filesystem should attempt to read all sync pages but may decide 911 to stop once it reaches the async pages. If it does decide to 912 stop attempting I/O, it can simply return. The caller will 913 remove the remaining pages from the address space, unlock them 914 and decrement the page refcount. Set PageUptodate if the I/O 915 completes successfully. 916 917``write_begin`` 918 Called by the generic buffered write code to ask the filesystem 919 to prepare to write len bytes at the given offset in the file. 920 The address_space should check that the write will be able to 921 complete, by allocating space if necessary and doing any other 922 internal housekeeping. If the write will update parts of any 923 basic-blocks on storage, then those blocks should be pre-read 924 (if they haven't been read already) so that the updated blocks 925 can be written out properly. 926 927 The filesystem must return the locked pagecache folio for the 928 specified offset, in ``*foliop``, for the caller to write into. 929 930 It must be able to cope with short writes (where the length 931 passed to write_begin is greater than the number of bytes copied 932 into the folio). 933 934 A void * may be returned in fsdata, which then gets passed into 935 write_end. 936 937 Returns 0 on success; < 0 on failure (which is the error code), 938 in which case write_end is not called. 939 940``write_end`` 941 After a successful write_begin, and data copy, write_end must be 942 called. len is the original len passed to write_begin, and 943 copied is the amount that was able to be copied. 944 945 The filesystem must take care of unlocking the folio, 946 decrementing its refcount, and updating i_size. 947 948 Returns < 0 on failure, otherwise the number of bytes (<= 949 'copied') that were able to be copied into pagecache. 950 951``bmap`` 952 called by the VFS to map a logical block offset within object to 953 physical block number. This method is used by the FIBMAP ioctl 954 and for working with swap-files. To be able to swap to a file, 955 the file must have a stable mapping to a block device. The swap 956 system does not go through the filesystem but instead uses bmap 957 to find out where the blocks in the file are and uses those 958 addresses directly. 959 960``invalidate_folio`` 961 If a folio has private data, then invalidate_folio will be 962 called when part or all of the folio is to be removed from the 963 address space. This generally corresponds to either a 964 truncation, punch hole or a complete invalidation of the address 965 space (in the latter case 'offset' will always be 0 and 'length' 966 will be folio_size()). Any private data associated with the folio 967 should be updated to reflect this truncation. If offset is 0 968 and length is folio_size(), then the private data should be 969 released, because the folio must be able to be completely 970 discarded. This may be done by calling the ->release_folio 971 function, but in this case the release MUST succeed. 972 973``release_folio`` 974 release_folio is called on folios with private data to tell the 975 filesystem that the folio is about to be freed. ->release_folio 976 should remove any private data from the folio and clear the 977 private flag. If release_folio() fails, it should return false. 978 release_folio() is used in two distinct though related cases. 979 The first is when the VM wants to free a clean folio with no 980 active users. If ->release_folio succeeds, the folio will be 981 removed from the address_space and be freed. 982 983 The second case is when a request has been made to invalidate 984 some or all folios in an address_space. This can happen 985 through the fadvise(POSIX_FADV_DONTNEED) system call or by the 986 filesystem explicitly requesting it as nfs and 9p do (when they 987 believe the cache may be out of date with storage) by calling 988 invalidate_inode_pages2(). If the filesystem makes such a call, 989 and needs to be certain that all folios are invalidated, then 990 its release_folio will need to ensure this. Possibly it can 991 clear the uptodate flag if it cannot free private data yet. 992 993``free_folio`` 994 free_folio is called once the folio is no longer visible in the 995 page cache in order to allow the cleanup of any private data. 996 Since it may be called by the memory reclaimer, it should not 997 assume that the original address_space mapping still exists, and 998 it should not block. 999 1000``direct_IO`` 1001 called by the generic read/write routines to perform direct_IO - 1002 that is IO requests which bypass the page cache and transfer 1003 data directly between the storage and the application's address 1004 space. 1005 1006``migrate_folio`` 1007 This is used to compact the physical memory usage. If the VM 1008 wants to relocate a folio (maybe from a memory device that is 1009 signalling imminent failure) it will pass a new folio and an old 1010 folio to this function. migrate_folio should transfer any private 1011 data across and update any references that it has to the folio. 1012 1013``launder_folio`` 1014 Called before freeing a folio - it writes back the dirty folio. 1015 To prevent redirtying the folio, it is kept locked during the 1016 whole operation. 1017 1018``is_partially_uptodate`` 1019 Called by the VM when reading a file through the pagecache when 1020 the underlying blocksize is smaller than the size of the folio. 1021 If the required block is up to date then the read can complete 1022 without needing I/O to bring the whole page up to date. 1023 1024``is_dirty_writeback`` 1025 Called by the VM when attempting to reclaim a folio. The VM uses 1026 dirty and writeback information to determine if it needs to 1027 stall to allow flushers a chance to complete some IO. 1028 Ordinarily it can use folio_test_dirty and folio_test_writeback but 1029 some filesystems have more complex state (unstable folios in NFS 1030 prevent reclaim) or do not set those flags due to locking 1031 problems. This callback allows a filesystem to indicate to the 1032 VM if a folio should be treated as dirty or writeback for the 1033 purposes of stalling. 1034 1035``error_remove_folio`` 1036 normally set to generic_error_remove_folio if truncation is ok 1037 for this address space. Used for memory failure handling. 1038 Setting this implies you deal with pages going away under you, 1039 unless you have them locked or reference counts increased. 1040 1041``swap_activate`` 1042 1043 Called to prepare the given file for swap. It should perform 1044 any validation and preparation necessary to ensure that writes 1045 can be performed with minimal memory allocation. It should call 1046 add_swap_extent(), or the helper iomap_swapfile_activate(), and 1047 return the number of extents added. If IO should be submitted 1048 through ->swap_rw(), it should set SWP_FS_OPS, otherwise IO will 1049 be submitted directly to the block device ``sis->bdev``. 1050 1051``swap_deactivate`` 1052 Called during swapoff on files where swap_activate was 1053 successful. 1054 1055``swap_rw`` 1056 Called to read or write swap pages when SWP_FS_OPS is set. 1057 1058The File Object 1059=============== 1060 1061A file object represents a file opened by a process. This is also known 1062as an "open file description" in POSIX parlance. 1063 1064 1065struct file_operations 1066---------------------- 1067 1068This describes how the VFS can manipulate an open file. As of kernel 10694.18, the following members are defined: 1070 1071.. code-block:: c 1072 1073 struct file_operations { 1074 struct module *owner; 1075 fop_flags_t fop_flags; 1076 loff_t (*llseek) (struct file *, loff_t, int); 1077 ssize_t (*read) (struct file *, char __user *, size_t, loff_t *); 1078 ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *); 1079 ssize_t (*read_iter) (struct kiocb *, struct iov_iter *); 1080 ssize_t (*write_iter) (struct kiocb *, struct iov_iter *); 1081 int (*iopoll)(struct kiocb *kiocb, struct io_comp_batch *, 1082 unsigned int flags); 1083 int (*iterate_shared) (struct file *, struct dir_context *); 1084 __poll_t (*poll) (struct file *, struct poll_table_struct *); 1085 long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long); 1086 long (*compat_ioctl) (struct file *, unsigned int, unsigned long); 1087 int (*mmap) (struct file *, struct vm_area_struct *); 1088 int (*open) (struct inode *, struct file *); 1089 int (*flush) (struct file *, fl_owner_t id); 1090 int (*release) (struct inode *, struct file *); 1091 int (*fsync) (struct file *, loff_t, loff_t, int datasync); 1092 int (*fasync) (int, struct file *, int); 1093 int (*lock) (struct file *, int, struct file_lock *); 1094 unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long); 1095 int (*check_flags)(int); 1096 int (*flock) (struct file *, int, struct file_lock *); 1097 ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int); 1098 ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int); 1099 void (*splice_eof)(struct file *file); 1100 int (*setlease)(struct file *, int, struct file_lease **, void **); 1101 long (*fallocate)(struct file *file, int mode, loff_t offset, 1102 loff_t len); 1103 void (*show_fdinfo)(struct seq_file *m, struct file *f); 1104 #ifndef CONFIG_MMU 1105 unsigned (*mmap_capabilities)(struct file *); 1106 #endif 1107 ssize_t (*copy_file_range)(struct file *, loff_t, struct file *, 1108 loff_t, size_t, unsigned int); 1109 loff_t (*remap_file_range)(struct file *file_in, loff_t pos_in, 1110 struct file *file_out, loff_t pos_out, 1111 loff_t len, unsigned int remap_flags); 1112 int (*fadvise)(struct file *, loff_t, loff_t, int); 1113 int (*uring_cmd)(struct io_uring_cmd *ioucmd, unsigned int issue_flags); 1114 int (*uring_cmd_iopoll)(struct io_uring_cmd *, struct io_comp_batch *, 1115 unsigned int poll_flags); 1116 int (*mmap_prepare)(struct vm_area_desc *); 1117 }; 1118 1119Again, all methods are called without any locks being held, unless 1120otherwise noted. 1121 1122``llseek`` 1123 called when the VFS needs to move the file position index 1124 1125``read`` 1126 called by read(2) and related system calls 1127 1128``read_iter`` 1129 possibly asynchronous read with iov_iter as destination 1130 1131``write`` 1132 called by write(2) and related system calls 1133 1134``write_iter`` 1135 possibly asynchronous write with iov_iter as source 1136 1137``iopoll`` 1138 called when aio wants to poll for completions on HIPRI iocbs 1139 1140``iterate_shared`` 1141 called when the VFS needs to read the directory contents 1142 1143``poll`` 1144 called by the VFS when a process wants to check if there is 1145 activity on this file and (optionally) go to sleep until there 1146 is activity. Called by the select(2) and poll(2) system calls 1147 1148``unlocked_ioctl`` 1149 called by the ioctl(2) system call. 1150 1151``compat_ioctl`` 1152 called by the ioctl(2) system call when 32 bit system calls are 1153 used on 64 bit kernels. 1154 1155``mmap`` 1156 called by the mmap(2) system call. Deprecated in favour of 1157 ``mmap_prepare``. 1158 1159``open`` 1160 called by the VFS when an inode should be opened. When the VFS 1161 opens a file, it creates a new "struct file". It then calls the 1162 open method for the newly allocated file structure. You might 1163 think that the open method really belongs in "struct 1164 inode_operations", and you may be right. I think it's done the 1165 way it is because it makes filesystems simpler to implement. 1166 The open() method is a good place to initialize the 1167 "private_data" member in the file structure if you want to point 1168 to a device structure 1169 1170``flush`` 1171 called by the close(2) system call to flush a file 1172 1173``release`` 1174 called when the last reference to an open file is closed 1175 1176``fsync`` 1177 called by the fsync(2) system call. Also see the section above 1178 entitled "Handling errors during writeback". 1179 1180``fasync`` 1181 called by the fcntl(2) system call when asynchronous 1182 (non-blocking) mode is enabled for a file 1183 1184``lock`` 1185 called by the fcntl(2) system call for F_GETLK, F_SETLK, and 1186 F_SETLKW commands 1187 1188``get_unmapped_area`` 1189 called by the mmap(2) system call 1190 1191``check_flags`` 1192 called by the fcntl(2) system call for F_SETFL command 1193 1194``flock`` 1195 called by the flock(2) system call 1196 1197``splice_write`` 1198 called by the VFS to splice data from a pipe to a file. This 1199 method is used by the splice(2) system call 1200 1201``splice_read`` 1202 called by the VFS to splice data from file to a pipe. This 1203 method is used by the splice(2) system call 1204 1205``setlease`` 1206 called by the VFS to set or release a file lock lease. setlease 1207 implementations should call generic_setlease to record or remove 1208 the lease in the inode after setting it. 1209 1210``fallocate`` 1211 called by the VFS to preallocate blocks or punch a hole. 1212 1213``copy_file_range`` 1214 called by the copy_file_range(2) system call. 1215 1216``remap_file_range`` 1217 called by the ioctl(2) system call for FICLONERANGE and FICLONE 1218 and FIDEDUPERANGE commands to remap file ranges. An 1219 implementation should remap len bytes at pos_in of the source 1220 file into the dest file at pos_out. Implementations must handle 1221 callers passing in len == 0; this means "remap to the end of the 1222 source file". The return value should the number of bytes 1223 remapped, or the usual negative error code if errors occurred 1224 before any bytes were remapped. The remap_flags parameter 1225 accepts REMAP_FILE_* flags. If REMAP_FILE_DEDUP is set then the 1226 implementation must only remap if the requested file ranges have 1227 identical contents. If REMAP_FILE_CAN_SHORTEN is set, the caller is 1228 ok with the implementation shortening the request length to 1229 satisfy alignment or EOF requirements (or any other reason). 1230 1231``fadvise`` 1232 possibly called by the fadvise64() system call. 1233 1234``mmap_prepare`` 1235 Called by the mmap(2) system call. Allows a VFS to set up a 1236 file-backed memory mapping, most notably establishing relevant 1237 private state and VMA callbacks. 1238 1239Note that the file operations are implemented by the specific 1240filesystem in which the inode resides. When opening a device node 1241(character or block special) most filesystems will call special 1242support routines in the VFS which will locate the required device 1243driver information. These support routines replace the filesystem file 1244operations with those for the device driver, and then proceed to call 1245the new open() method for the file. This is how opening a device file 1246in the filesystem eventually ends up calling the device driver open() 1247method. 1248 1249 1250Directory Entry Cache (dcache) 1251============================== 1252 1253 1254struct dentry_operations 1255------------------------ 1256 1257This describes how a filesystem can overload the standard dentry 1258operations. Dentries and the dcache are the domain of the VFS and the 1259individual filesystem implementations. Device drivers have no business 1260here. These methods may be set to NULL, as they are either optional or 1261the VFS uses a default. As of kernel 2.6.22, the following members are 1262defined: 1263 1264.. code-block:: c 1265 1266 struct dentry_operations { 1267 int (*d_revalidate)(struct inode *, const struct qstr *, 1268 struct dentry *, unsigned int); 1269 int (*d_weak_revalidate)(struct dentry *, unsigned int); 1270 int (*d_hash)(const struct dentry *, struct qstr *); 1271 int (*d_compare)(const struct dentry *, 1272 unsigned int, const char *, const struct qstr *); 1273 int (*d_delete)(const struct dentry *); 1274 int (*d_init)(struct dentry *); 1275 void (*d_release)(struct dentry *); 1276 void (*d_iput)(struct dentry *, struct inode *); 1277 char *(*d_dname)(struct dentry *, char *, int); 1278 struct vfsmount *(*d_automount)(struct path *); 1279 int (*d_manage)(const struct path *, bool); 1280 struct dentry *(*d_real)(struct dentry *, enum d_real_type type); 1281 bool (*d_unalias_trylock)(const struct dentry *); 1282 void (*d_unalias_unlock)(const struct dentry *); 1283 }; 1284 1285``d_revalidate`` 1286 called when the VFS needs to revalidate a dentry. This is 1287 called whenever a name look-up finds a dentry in the dcache. 1288 Most local filesystems leave this as NULL, because all their 1289 dentries in the dcache are valid. Network filesystems are 1290 different since things can change on the server without the 1291 client necessarily being aware of it. 1292 1293 This function should return a positive value if the dentry is 1294 still valid, and zero or a negative error code if it isn't. 1295 1296 d_revalidate may be called in rcu-walk mode (flags & 1297 LOOKUP_RCU). If in rcu-walk mode, the filesystem must 1298 revalidate the dentry without blocking or storing to the dentry, 1299 d_parent and d_inode should not be used without care (because 1300 they can change and, in d_inode case, even become NULL under 1301 us). 1302 1303 If a situation is encountered that rcu-walk cannot handle, 1304 return 1305 -ECHILD and it will be called again in ref-walk mode. 1306 1307``d_weak_revalidate`` 1308 called when the VFS needs to revalidate a "jumped" dentry. This 1309 is called when a path-walk ends at dentry that was not acquired 1310 by doing a lookup in the parent directory. This includes "/", 1311 "." and "..", as well as procfs-style symlinks and mountpoint 1312 traversal. 1313 1314 In this case, we are less concerned with whether the dentry is 1315 still fully correct, but rather that the inode is still valid. 1316 As with d_revalidate, most local filesystems will set this to 1317 NULL since their dcache entries are always valid. 1318 1319 This function has the same return code semantics as 1320 d_revalidate. 1321 1322 d_weak_revalidate is only called after leaving rcu-walk mode. 1323 1324``d_hash`` 1325 called when the VFS adds a dentry to the hash table. The first 1326 dentry passed to d_hash is the parent directory that the name is 1327 to be hashed into. 1328 1329 Same locking and synchronisation rules as d_compare regarding 1330 what is safe to dereference etc. 1331 1332``d_compare`` 1333 called to compare a dentry name with a given name. The first 1334 dentry is the parent of the dentry to be compared, the second is 1335 the child dentry. len and name string are properties of the 1336 dentry to be compared. qstr is the name to compare it with. 1337 1338 Must be constant and idempotent, and should not take locks if 1339 possible, and should not or store into the dentry. Should not 1340 dereference pointers outside the dentry without lots of care 1341 (eg. d_parent, d_inode, d_name should not be used). 1342 1343 However, our vfsmount is pinned, and RCU held, so the dentries 1344 and inodes won't disappear, neither will our sb or filesystem 1345 module. ->d_sb may be used. 1346 1347 It is a tricky calling convention because it needs to be called 1348 under "rcu-walk", ie. without any locks or references on things. 1349 1350``d_delete`` 1351 called when the last reference to a dentry is dropped and the 1352 dcache is deciding whether or not to cache it. Return 1 to 1353 delete immediately, or 0 to cache the dentry. Default is NULL 1354 which means to always cache a reachable dentry. d_delete must 1355 be constant and idempotent. 1356 1357``d_init`` 1358 called when a dentry is allocated 1359 1360``d_release`` 1361 called when a dentry is really deallocated 1362 1363``d_iput`` 1364 called when a dentry loses its inode (just prior to its being 1365 deallocated). The default when this is NULL is that the VFS 1366 calls iput(). If you define this method, you must call iput() 1367 yourself 1368 1369``d_dname`` 1370 called when the pathname of a dentry should be generated. 1371 Useful for some pseudo filesystems (sockfs, pipefs, ...) to 1372 delay pathname generation. (Instead of doing it when dentry is 1373 created, it's done only when the path is needed.). Real 1374 filesystems probably dont want to use it, because their dentries 1375 are present in global dcache hash, so their hash should be an 1376 invariant. As no lock is held, d_dname() should not try to 1377 modify the dentry itself, unless appropriate SMP safety is used. 1378 CAUTION : d_path() logic is quite tricky. The correct way to 1379 return for example "Hello" is to put it at the end of the 1380 buffer, and returns a pointer to the first char. 1381 dynamic_dname() helper function is provided to take care of 1382 this. 1383 1384 Example : 1385 1386.. code-block:: c 1387 1388 static char *pipefs_dname(struct dentry *dent, char *buffer, int buflen) 1389 { 1390 return dynamic_dname(dentry, buffer, buflen, "pipe:[%lu]", 1391 dentry->d_inode->i_ino); 1392 } 1393 1394``d_automount`` 1395 called when an automount dentry is to be traversed (optional). 1396 This should create a new VFS mount record and return the record 1397 to the caller. The caller is supplied with a path parameter 1398 giving the automount directory to describe the automount target 1399 and the parent VFS mount record to provide inheritable mount 1400 parameters. NULL should be returned if someone else managed to 1401 make the automount first. If the vfsmount creation failed, then 1402 an error code should be returned. If -EISDIR is returned, then 1403 the directory will be treated as an ordinary directory and 1404 returned to pathwalk to continue walking. 1405 1406 If a vfsmount is returned, the caller will attempt to mount it 1407 on the mountpoint and will remove the vfsmount from its 1408 expiration list in the case of failure. 1409 1410 This function is only used if DCACHE_NEED_AUTOMOUNT is set on 1411 the dentry. This is set by __d_instantiate() if S_AUTOMOUNT is 1412 set on the inode being added. 1413 1414``d_manage`` 1415 called to allow the filesystem to manage the transition from a 1416 dentry (optional). This allows autofs, for example, to hold up 1417 clients waiting to explore behind a 'mountpoint' while letting 1418 the daemon go past and construct the subtree there. 0 should be 1419 returned to let the calling process continue. -EISDIR can be 1420 returned to tell pathwalk to use this directory as an ordinary 1421 directory and to ignore anything mounted on it and not to check 1422 the automount flag. Any other error code will abort pathwalk 1423 completely. 1424 1425 If the 'rcu_walk' parameter is true, then the caller is doing a 1426 pathwalk in RCU-walk mode. Sleeping is not permitted in this 1427 mode, and the caller can be asked to leave it and call again by 1428 returning -ECHILD. -EISDIR may also be returned to tell 1429 pathwalk to ignore d_automount or any mounts. 1430 1431 This function is only used if DCACHE_MANAGE_TRANSIT is set on 1432 the dentry being transited from. 1433 1434``d_real`` 1435 overlay/union type filesystems implement this method to return one 1436 of the underlying dentries of a regular file hidden by the overlay. 1437 1438 The 'type' argument takes the values D_REAL_DATA or D_REAL_METADATA 1439 for returning the real underlying dentry that refers to the inode 1440 hosting the file's data or metadata respectively. 1441 1442 For non-regular files, the 'dentry' argument is returned. 1443 1444``d_unalias_trylock`` 1445 if present, will be called by d_splice_alias() before moving a 1446 preexisting attached alias. Returning false prevents __d_move(), 1447 making d_splice_alias() fail with -ESTALE. 1448 1449 Rationale: setting FS_RENAME_DOES_D_MOVE will prevent d_move() 1450 and d_exchange() calls from the outside of filesystem methods; 1451 however, it does not guarantee that attached dentries won't 1452 be renamed or moved by d_splice_alias() finding a preexisting 1453 alias for a directory inode. Normally we would not care; 1454 however, something that wants to stabilize the entire path to 1455 root over a blocking operation might need that. See 9p for one 1456 (and hopefully only) example. 1457 1458``d_unalias_unlock`` 1459 should be paired with ``d_unalias_trylock``; that one is called after 1460 __d_move() call in __d_unalias(). 1461 1462 1463Each dentry has a pointer to its parent dentry, as well as a hash list 1464of child dentries. Child dentries are basically like files in a 1465directory. 1466 1467 1468Directory Entry Cache API 1469-------------------------- 1470 1471There are a number of functions defined which permit a filesystem to 1472manipulate dentries: 1473 1474``dget`` 1475 open a new handle for an existing dentry (this just increments 1476 the usage count) 1477 1478``dput`` 1479 close a handle for a dentry (decrements the usage count). If 1480 the usage count drops to 0, and the dentry is still in its 1481 parent's hash, the "d_delete" method is called to check whether 1482 it should be cached. If it should not be cached, or if the 1483 dentry is not hashed, it is deleted. Otherwise cached dentries 1484 are put into an LRU list to be reclaimed on memory shortage. 1485 1486``d_drop`` 1487 this unhashes a dentry from its parents hash list. A subsequent 1488 call to dput() will deallocate the dentry if its usage count 1489 drops to 0 1490 1491``d_delete`` 1492 delete a dentry. If there are no other open references to the 1493 dentry then the dentry is turned into a negative dentry (the 1494 d_iput() method is called). If there are other references, then 1495 d_drop() is called instead 1496 1497``d_add`` 1498 add a dentry to its parents hash list and then calls 1499 d_instantiate() 1500 1501``d_instantiate`` 1502 add a dentry to the alias hash list for the inode and updates 1503 the "d_inode" member. The "i_count" member in the inode 1504 structure should be set/incremented. If the inode pointer is 1505 NULL, the dentry is called a "negative dentry". This function 1506 is commonly called when an inode is created for an existing 1507 negative dentry 1508 1509``d_lookup`` 1510 look up a dentry given its parent and path name component It 1511 looks up the child of that given name from the dcache hash 1512 table. If it is found, the reference count is incremented and 1513 the dentry is returned. The caller must use dput() to free the 1514 dentry when it finishes using it. 1515 1516 1517Mount Options 1518============= 1519 1520 1521Parsing options 1522--------------- 1523 1524On mount and remount the filesystem is passed a string containing a 1525comma separated list of mount options. The options can have either of 1526these forms: 1527 1528 option 1529 option=value 1530 1531The <linux/parser.h> header defines an API that helps parse these 1532options. There are plenty of examples on how to use it in existing 1533filesystems. 1534 1535 1536Showing options 1537--------------- 1538 1539If a filesystem accepts mount options, it must define show_options() to 1540show all the currently active options. The rules are: 1541 1542 - options MUST be shown which are not default or their values differ 1543 from the default 1544 1545 - options MAY be shown which are enabled by default or have their 1546 default value 1547 1548Options used only internally between a mount helper and the kernel (such 1549as file descriptors), or which only have an effect during the mounting 1550(such as ones controlling the creation of a journal) are exempt from the 1551above rules. 1552 1553The underlying reason for the above rules is to make sure, that a mount 1554can be accurately replicated (e.g. umounting and mounting again) based 1555on the information found in /proc/mounts. 1556 1557 1558Resources 1559========= 1560 1561(Note some of these resources are not up-to-date with the latest kernel 1562 version.) 1563 1564Creating Linux virtual filesystems. 2002 1565 <https://lwn.net/Articles/13325/> 1566 1567The Linux Virtual File-system Layer by Neil Brown. 1999 1568 <http://www.cse.unsw.edu.au/~neilb/oss/linux-commentary/vfs.html> 1569 1570A tour of the Linux VFS by Michael K. Johnson. 1996 1571 <https://www.tldp.org/LDP/khg/HyperNews/get/fs/vfstour.html> 1572 1573A small trail through the Linux kernel by Andries Brouwer. 2001 1574 <https://www.win.tue.nl/~aeb/linux/vfs/trail.html> 1575