5 * This file has definitions for some important file table
9 #include <linux/limits.h>
10 #include <linux/ioctl.h>
11 #include <linux/blk_types.h>
14 * It's silly to have NR_OPEN bigger than NR_FILE, but you can change
15 * the file limit at runtime and only root can increase the per-process
16 * nr_file rlimit, so it's safe to set up a ridiculously high absolute
17 * upper limit on files-per-process.
19 * Some programs (notably those using select()) may have to be
20 * recompiled to take full advantage of the new limits..
23 /* Fixed constants first: */
25 #define INR_OPEN 1024 /* Initial setting for nfile rlimits */
27 #define BLOCK_SIZE_BITS 10
28 #define BLOCK_SIZE (1<<BLOCK_SIZE_BITS)
30 #define SEEK_SET 0 /* seek relative to beginning of file */
31 #define SEEK_CUR 1 /* seek relative to current file position */
32 #define SEEK_END 2 /* seek relative to end of file */
33 #define SEEK_MAX SEEK_END
41 /* And dynamically-tunable limits and defaults: */
42 struct files_stat_struct {
43 unsigned long nr_files; /* read only */
44 unsigned long nr_free_files; /* read only */
45 unsigned long max_files; /* tunable */
48 struct inodes_stat_t {
51 int dummy[5]; /* padding for sysctl ABI compatibility */
55 #define NR_FILE 8192 /* this can well be larger on a larger system */
66 * flags in file.f_mode. Note that FMODE_READ and FMODE_WRITE must correspond
67 * to O_WRONLY and O_RDWR via the strange trick in __dentry_open()
70 /* file is open for reading */
71 #define FMODE_READ ((__force fmode_t)0x1)
72 /* file is open for writing */
73 #define FMODE_WRITE ((__force fmode_t)0x2)
74 /* file is seekable */
75 #define FMODE_LSEEK ((__force fmode_t)0x4)
76 /* file can be accessed using pread */
77 #define FMODE_PREAD ((__force fmode_t)0x8)
78 /* file can be accessed using pwrite */
79 #define FMODE_PWRITE ((__force fmode_t)0x10)
80 /* File is opened for execution with sys_execve / sys_uselib */
81 #define FMODE_EXEC ((__force fmode_t)0x20)
82 /* File is opened with O_NDELAY (only set for block devices) */
83 #define FMODE_NDELAY ((__force fmode_t)0x40)
84 /* File is opened with O_EXCL (only set for block devices) */
85 #define FMODE_EXCL ((__force fmode_t)0x80)
86 /* File is opened using open(.., 3, ..) and is writeable only for ioctls
87 (specialy hack for floppy.c) */
88 #define FMODE_WRITE_IOCTL ((__force fmode_t)0x100)
91 * Don't update ctime and mtime.
93 * Currently a special hack for the XFS open_by_handle ioctl, but we'll
94 * hopefully graduate it to a proper O_CMTIME flag supported by open(2) soon.
96 #define FMODE_NOCMTIME ((__force fmode_t)0x800)
98 /* Expect random access pattern */
99 #define FMODE_RANDOM ((__force fmode_t)0x1000)
101 /* File is huge (eg. /dev/kmem): treat loff_t as unsigned */
102 #define FMODE_UNSIGNED_OFFSET ((__force fmode_t)0x2000)
104 /* File was opened by fanotify and shouldn't generate fanotify events */
105 #define FMODE_NONOTIFY ((__force fmode_t)0x1000000)
108 * The below are the various read and write types that we support. Some of
109 * them include behavioral modifiers that send information down to the
110 * block layer and IO scheduler. Terminology:
112 * The block layer uses device plugging to defer IO a little bit, in
113 * the hope that we will see more IO very shortly. This increases
114 * coalescing of adjacent IO and thus reduces the number of IOs we
115 * have to send to the device. It also allows for better queuing,
116 * if the IO isn't mergeable. If the caller is going to be waiting
117 * for the IO, then he must ensure that the device is unplugged so
118 * that the IO is dispatched to the driver.
120 * All IO is handled async in Linux. This is fine for background
121 * writes, but for reads or writes that someone waits for completion
122 * on, we want to notify the block layer and IO scheduler so that they
123 * know about it. That allows them to make better scheduling
124 * decisions. So when the below references 'sync' and 'async', it
125 * is referencing this priority hint.
127 * With that in mind, the available types are:
129 * READ A normal read operation. Device will be plugged.
130 * READ_SYNC A synchronous read. Device is not plugged, caller can
131 * immediately wait on this read without caring about
133 * READA Used for read-ahead operations. Lower priority, and the
134 * block layer could (in theory) choose to ignore this
135 * request if it runs into resource problems.
136 * WRITE A normal async write. Device will be plugged.
137 * WRITE_SYNC_PLUG Synchronous write. Identical to WRITE, but passes down
138 * the hint that someone will be waiting on this IO
139 * shortly. The device must still be unplugged explicitly,
140 * WRITE_SYNC_PLUG does not do this as we could be
141 * submitting more writes before we actually wait on any
143 * WRITE_SYNC Like WRITE_SYNC_PLUG, but also unplugs the device
144 * immediately after submission. The write equivalent
146 * WRITE_ODIRECT_PLUG Special case write for O_DIRECT only.
147 * WRITE_FLUSH Like WRITE_SYNC but with preceding cache flush.
148 * WRITE_FUA Like WRITE_SYNC but data is guaranteed to be on
149 * non-volatile media on completion.
150 * WRITE_FLUSH_FUA Combination of WRITE_FLUSH and FUA. The IO is preceded
151 * by a cache flush and data is guaranteed to be on
152 * non-volatile media on completion.
155 #define RW_MASK REQ_WRITE
156 #define RWA_MASK REQ_RAHEAD
159 #define WRITE RW_MASK
160 #define READA RWA_MASK
162 #define READ_SYNC (READ | REQ_SYNC | REQ_UNPLUG)
163 #define READ_META (READ | REQ_META)
164 #define WRITE_SYNC_PLUG (WRITE | REQ_SYNC | REQ_NOIDLE)
165 #define WRITE_SYNC (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_UNPLUG)
166 #define WRITE_ODIRECT_PLUG (WRITE | REQ_SYNC)
167 #define WRITE_META (WRITE | REQ_META)
168 #define WRITE_FLUSH (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_UNPLUG | \
170 #define WRITE_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_UNPLUG | \
172 #define WRITE_FLUSH_FUA (WRITE | REQ_SYNC | REQ_NOIDLE | REQ_UNPLUG | \
179 /* public flags for file_system_type */
180 #define FS_REQUIRES_DEV 1
181 #define FS_BINARY_MOUNTDATA 2
182 #define FS_HAS_SUBTYPE 4
183 #define FS_REVAL_DOT 16384 /* Check the paths ".", ".." for staleness */
184 #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move()
185 * during rename() internally.
189 * These are the fs-independent mount-flags: up to 32 flags are supported
191 #define MS_RDONLY 1 /* Mount read-only */
192 #define MS_NOSUID 2 /* Ignore suid and sgid bits */
193 #define MS_NODEV 4 /* Disallow access to device special files */
194 #define MS_NOEXEC 8 /* Disallow program execution */
195 #define MS_SYNCHRONOUS 16 /* Writes are synced at once */
196 #define MS_REMOUNT 32 /* Alter flags of a mounted FS */
197 #define MS_MANDLOCK 64 /* Allow mandatory locks on an FS */
198 #define MS_DIRSYNC 128 /* Directory modifications are synchronous */
199 #define MS_NOATIME 1024 /* Do not update access times. */
200 #define MS_NODIRATIME 2048 /* Do not update directory access times */
204 #define MS_VERBOSE 32768 /* War is peace. Verbosity is silence.
205 MS_VERBOSE is deprecated. */
206 #define MS_SILENT 32768
207 #define MS_POSIXACL (1<<16) /* VFS does not apply the umask */
208 #define MS_UNBINDABLE (1<<17) /* change to unbindable */
209 #define MS_PRIVATE (1<<18) /* change to private */
210 #define MS_SLAVE (1<<19) /* change to slave */
211 #define MS_SHARED (1<<20) /* change to shared */
212 #define MS_RELATIME (1<<21) /* Update atime relative to mtime/ctime. */
213 #define MS_KERNMOUNT (1<<22) /* this is a kern_mount call */
214 #define MS_I_VERSION (1<<23) /* Update inode I_version field */
215 #define MS_STRICTATIME (1<<24) /* Always perform atime updates */
216 #define MS_BORN (1<<29)
217 #define MS_ACTIVE (1<<30)
218 #define MS_NOUSER (1<<31)
221 * Superblock flags that can be altered by MS_REMOUNT
223 #define MS_RMT_MASK (MS_RDONLY|MS_SYNCHRONOUS|MS_MANDLOCK|MS_I_VERSION)
226 * Old magic mount flag and mask
228 #define MS_MGC_VAL 0xC0ED0000
229 #define MS_MGC_MSK 0xffff0000
231 /* Inode flags - they have nothing to superblock flags now */
233 #define S_SYNC 1 /* Writes are synced at once */
234 #define S_NOATIME 2 /* Do not update access times */
235 #define S_APPEND 4 /* Append-only file */
236 #define S_IMMUTABLE 8 /* Immutable file */
237 #define S_DEAD 16 /* removed, but still open directory */
238 #define S_NOQUOTA 32 /* Inode is not counted to quota */
239 #define S_DIRSYNC 64 /* Directory modifications are synchronous */
240 #define S_NOCMTIME 128 /* Do not update file c/mtime */
241 #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
242 #define S_PRIVATE 512 /* Inode is fs-internal */
243 #define S_IMA 1024 /* Inode has an associated IMA struct */
246 * Note that nosuid etc flags are inode-specific: setting some file-system
247 * flags just means all the inodes inherit those flags by default. It might be
248 * possible to override it selectively if you really wanted to with some
249 * ioctl() that is not currently implemented.
251 * Exception: MS_RDONLY is always applied to the entire file system.
253 * Unfortunately, it is possible to change a filesystems flags with it mounted
254 * with files in use. This means that all of the inodes will not have their
255 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
256 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
258 #define __IS_FLG(inode,flg) ((inode)->i_sb->s_flags & (flg))
260 #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
261 #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
262 ((inode)->i_flags & S_SYNC))
263 #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
264 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
265 #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
266 #define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
267 #define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
269 #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
270 #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
271 #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
272 #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
274 #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
275 #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
276 #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
277 #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
278 #define IS_IMA(inode) ((inode)->i_flags & S_IMA)
280 /* the read-only stuff doesn't really belong here, but any other place is
281 probably as bad and I don't want to create yet another include file. */
283 #define BLKROSET _IO(0x12,93) /* set device read-only (0 = read-write) */
284 #define BLKROGET _IO(0x12,94) /* get read-only status (0 = read_write) */
285 #define BLKRRPART _IO(0x12,95) /* re-read partition table */
286 #define BLKGETSIZE _IO(0x12,96) /* return device size /512 (long *arg) */
287 #define BLKFLSBUF _IO(0x12,97) /* flush buffer cache */
288 #define BLKRASET _IO(0x12,98) /* set read ahead for block device */
289 #define BLKRAGET _IO(0x12,99) /* get current read ahead setting */
290 #define BLKFRASET _IO(0x12,100)/* set filesystem (mm/filemap.c) read-ahead */
291 #define BLKFRAGET _IO(0x12,101)/* get filesystem (mm/filemap.c) read-ahead */
292 #define BLKSECTSET _IO(0x12,102)/* set max sectors per request (ll_rw_blk.c) */
293 #define BLKSECTGET _IO(0x12,103)/* get max sectors per request (ll_rw_blk.c) */
294 #define BLKSSZGET _IO(0x12,104)/* get block device sector size */
296 #define BLKPG _IO(0x12,105)/* See blkpg.h */
298 /* Some people are morons. Do not use sizeof! */
300 #define BLKELVGET _IOR(0x12,106,size_t)/* elevator get */
301 #define BLKELVSET _IOW(0x12,107,size_t)/* elevator set */
302 /* This was here just to show that the number is taken -
303 probably all these _IO(0x12,*) ioctls should be moved to blkpg.h. */
305 /* A jump here: 108-111 have been used for various private purposes. */
306 #define BLKBSZGET _IOR(0x12,112,size_t)
307 #define BLKBSZSET _IOW(0x12,113,size_t)
308 #define BLKGETSIZE64 _IOR(0x12,114,size_t) /* return device size in bytes (u64 *arg) */
309 #define BLKTRACESETUP _IOWR(0x12,115,struct blk_user_trace_setup)
310 #define BLKTRACESTART _IO(0x12,116)
311 #define BLKTRACESTOP _IO(0x12,117)
312 #define BLKTRACETEARDOWN _IO(0x12,118)
313 #define BLKDISCARD _IO(0x12,119)
314 #define BLKIOMIN _IO(0x12,120)
315 #define BLKIOOPT _IO(0x12,121)
316 #define BLKALIGNOFF _IO(0x12,122)
317 #define BLKPBSZGET _IO(0x12,123)
318 #define BLKDISCARDZEROES _IO(0x12,124)
319 #define BLKSECDISCARD _IO(0x12,125)
321 #define BMAP_IOCTL 1 /* obsolete - kept for compatibility */
322 #define FIBMAP _IO(0x00,1) /* bmap access */
323 #define FIGETBSZ _IO(0x00,2) /* get the block size used for bmap */
324 #define FIFREEZE _IOWR('X', 119, int) /* Freeze */
325 #define FITHAW _IOWR('X', 120, int) /* Thaw */
326 #define FITRIM _IOWR('X', 121, struct fstrim_range) /* Trim */
328 #define FS_IOC_GETFLAGS _IOR('f', 1, long)
329 #define FS_IOC_SETFLAGS _IOW('f', 2, long)
330 #define FS_IOC_GETVERSION _IOR('v', 1, long)
331 #define FS_IOC_SETVERSION _IOW('v', 2, long)
332 #define FS_IOC_FIEMAP _IOWR('f', 11, struct fiemap)
333 #define FS_IOC32_GETFLAGS _IOR('f', 1, int)
334 #define FS_IOC32_SETFLAGS _IOW('f', 2, int)
335 #define FS_IOC32_GETVERSION _IOR('v', 1, int)
336 #define FS_IOC32_SETVERSION _IOW('v', 2, int)
339 * Inode flags (FS_IOC_GETFLAGS / FS_IOC_SETFLAGS)
341 #define FS_SECRM_FL 0x00000001 /* Secure deletion */
342 #define FS_UNRM_FL 0x00000002 /* Undelete */
343 #define FS_COMPR_FL 0x00000004 /* Compress file */
344 #define FS_SYNC_FL 0x00000008 /* Synchronous updates */
345 #define FS_IMMUTABLE_FL 0x00000010 /* Immutable file */
346 #define FS_APPEND_FL 0x00000020 /* writes to file may only append */
347 #define FS_NODUMP_FL 0x00000040 /* do not dump file */
348 #define FS_NOATIME_FL 0x00000080 /* do not update atime */
349 /* Reserved for compression usage... */
350 #define FS_DIRTY_FL 0x00000100
351 #define FS_COMPRBLK_FL 0x00000200 /* One or more compressed clusters */
352 #define FS_NOCOMP_FL 0x00000400 /* Don't compress */
353 #define FS_ECOMPR_FL 0x00000800 /* Compression error */
354 /* End compression flags --- maybe not all used */
355 #define FS_BTREE_FL 0x00001000 /* btree format dir */
356 #define FS_INDEX_FL 0x00001000 /* hash-indexed directory */
357 #define FS_IMAGIC_FL 0x00002000 /* AFS directory */
358 #define FS_JOURNAL_DATA_FL 0x00004000 /* Reserved for ext3 */
359 #define FS_NOTAIL_FL 0x00008000 /* file tail should not be merged */
360 #define FS_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */
361 #define FS_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/
362 #define FS_EXTENT_FL 0x00080000 /* Extents */
363 #define FS_DIRECTIO_FL 0x00100000 /* Use direct i/o */
364 #define FS_RESERVED_FL 0x80000000 /* reserved for ext2 lib */
366 #define FS_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */
367 #define FS_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */
370 #define SYNC_FILE_RANGE_WAIT_BEFORE 1
371 #define SYNC_FILE_RANGE_WRITE 2
372 #define SYNC_FILE_RANGE_WAIT_AFTER 4
376 #include <linux/linkage.h>
377 #include <linux/wait.h>
378 #include <linux/types.h>
379 #include <linux/kdev_t.h>
380 #include <linux/dcache.h>
381 #include <linux/path.h>
382 #include <linux/stat.h>
383 #include <linux/cache.h>
384 #include <linux/kobject.h>
385 #include <linux/list.h>
386 #include <linux/radix-tree.h>
387 #include <linux/prio_tree.h>
388 #include <linux/init.h>
389 #include <linux/pid.h>
390 #include <linux/mutex.h>
391 #include <linux/capability.h>
392 #include <linux/semaphore.h>
393 #include <linux/fiemap.h>
395 #include <asm/atomic.h>
396 #include <asm/byteorder.h>
398 struct export_operations;
403 struct pipe_inode_info;
404 struct poll_table_struct;
406 struct vm_area_struct;
410 extern void __init inode_init(void);
411 extern void __init inode_init_early(void);
412 extern void __init files_init(unsigned long);
414 extern struct files_stat_struct files_stat;
415 extern unsigned long get_max_files(void);
416 extern int sysctl_nr_open;
417 extern struct inodes_stat_t inodes_stat;
418 extern int leases_enable, lease_break_time;
421 typedef int (get_block_t)(struct inode *inode, sector_t iblock,
422 struct buffer_head *bh_result, int create);
423 typedef void (dio_iodone_t)(struct kiocb *iocb, loff_t offset,
424 ssize_t bytes, void *private, int ret,
428 * Attribute flags. These should be or-ed together to figure out what
431 #define ATTR_MODE (1 << 0)
432 #define ATTR_UID (1 << 1)
433 #define ATTR_GID (1 << 2)
434 #define ATTR_SIZE (1 << 3)
435 #define ATTR_ATIME (1 << 4)
436 #define ATTR_MTIME (1 << 5)
437 #define ATTR_CTIME (1 << 6)
438 #define ATTR_ATIME_SET (1 << 7)
439 #define ATTR_MTIME_SET (1 << 8)
440 #define ATTR_FORCE (1 << 9) /* Not a change, but a change it */
441 #define ATTR_ATTR_FLAG (1 << 10)
442 #define ATTR_KILL_SUID (1 << 11)
443 #define ATTR_KILL_SGID (1 << 12)
444 #define ATTR_FILE (1 << 13)
445 #define ATTR_KILL_PRIV (1 << 14)
446 #define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */
447 #define ATTR_TIMES_SET (1 << 16)
450 * This is the Inode Attributes structure, used for notify_change(). It
451 * uses the above definitions as flags, to know which values have changed.
452 * Also, in this manner, a Filesystem can look at only the values it cares
453 * about. Basically, these are the attributes that the VFS layer can
454 * request to change from the FS layer.
456 * Derek Atkins <warlord@MIT.EDU> 94-10-20
459 unsigned int ia_valid;
464 struct timespec ia_atime;
465 struct timespec ia_mtime;
466 struct timespec ia_ctime;
469 * Not an attribute, but an auxilary info for filesystems wanting to
470 * implement an ftruncate() like method. NOTE: filesystem should
471 * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
473 struct file *ia_file;
477 * Includes for diskquotas.
479 #include <linux/quota.h>
482 * enum positive_aop_returns - aop return codes with specific semantics
484 * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
485 * completed, that the page is still locked, and
486 * should be considered active. The VM uses this hint
487 * to return the page to the active list -- it won't
488 * be a candidate for writeback again in the near
489 * future. Other callers must be careful to unlock
490 * the page if they get this return. Returned by
493 * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
494 * unlocked it and the page might have been truncated.
495 * The caller should back up to acquiring a new page and
496 * trying again. The aop will be taking reasonable
497 * precautions not to livelock. If the caller held a page
498 * reference, it should drop it before retrying. Returned
501 * address_space_operation functions return these large constants to indicate
502 * special semantics to the caller. These are much larger than the bytes in a
503 * page to allow for functions that return the number of bytes operated on in a
507 enum positive_aop_returns {
508 AOP_WRITEPAGE_ACTIVATE = 0x80000,
509 AOP_TRUNCATED_PAGE = 0x80001,
512 #define AOP_FLAG_UNINTERRUPTIBLE 0x0001 /* will not do a short write */
513 #define AOP_FLAG_CONT_EXPAND 0x0002 /* called from cont_expand */
514 #define AOP_FLAG_NOFS 0x0004 /* used by filesystem to direct
515 * helper code (eg buffer layer)
516 * to clear GFP_FS from alloc */
519 * oh the beauties of C type declarations.
522 struct address_space;
523 struct writeback_control;
526 const struct iovec *iov;
527 unsigned long nr_segs;
532 size_t iov_iter_copy_from_user_atomic(struct page *page,
533 struct iov_iter *i, unsigned long offset, size_t bytes);
534 size_t iov_iter_copy_from_user(struct page *page,
535 struct iov_iter *i, unsigned long offset, size_t bytes);
536 void iov_iter_advance(struct iov_iter *i, size_t bytes);
537 int iov_iter_fault_in_readable(struct iov_iter *i, size_t bytes);
538 size_t iov_iter_single_seg_count(struct iov_iter *i);
540 static inline void iov_iter_init(struct iov_iter *i,
541 const struct iovec *iov, unsigned long nr_segs,
542 size_t count, size_t written)
545 i->nr_segs = nr_segs;
547 i->count = count + written;
549 iov_iter_advance(i, written);
552 static inline size_t iov_iter_count(struct iov_iter *i)
558 * "descriptor" for what we're up to with a read.
559 * This allows us to use the same read code yet
560 * have multiple different users of the data that
561 * we read from a file.
563 * The simplest case just copies the data to user
576 typedef int (*read_actor_t)(read_descriptor_t *, struct page *,
577 unsigned long, unsigned long);
579 struct address_space_operations {
580 int (*writepage)(struct page *page, struct writeback_control *wbc);
581 int (*readpage)(struct file *, struct page *);
582 void (*sync_page)(struct page *);
584 /* Write back some dirty pages from this mapping. */
585 int (*writepages)(struct address_space *, struct writeback_control *);
587 /* Set a page dirty. Return true if this dirtied it */
588 int (*set_page_dirty)(struct page *page);
590 int (*readpages)(struct file *filp, struct address_space *mapping,
591 struct list_head *pages, unsigned nr_pages);
593 int (*write_begin)(struct file *, struct address_space *mapping,
594 loff_t pos, unsigned len, unsigned flags,
595 struct page **pagep, void **fsdata);
596 int (*write_end)(struct file *, struct address_space *mapping,
597 loff_t pos, unsigned len, unsigned copied,
598 struct page *page, void *fsdata);
600 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
601 sector_t (*bmap)(struct address_space *, sector_t);
602 void (*invalidatepage) (struct page *, unsigned long);
603 int (*releasepage) (struct page *, gfp_t);
604 ssize_t (*direct_IO)(int, struct kiocb *, const struct iovec *iov,
605 loff_t offset, unsigned long nr_segs);
606 int (*get_xip_mem)(struct address_space *, pgoff_t, int,
607 void **, unsigned long *);
608 /* migrate the contents of a page to the specified target */
609 int (*migratepage) (struct address_space *,
610 struct page *, struct page *);
611 int (*launder_page) (struct page *);
612 int (*is_partially_uptodate) (struct page *, read_descriptor_t *,
614 int (*error_remove_page)(struct address_space *, struct page *);
618 * pagecache_write_begin/pagecache_write_end must be used by general code
619 * to write into the pagecache.
621 int pagecache_write_begin(struct file *, struct address_space *mapping,
622 loff_t pos, unsigned len, unsigned flags,
623 struct page **pagep, void **fsdata);
625 int pagecache_write_end(struct file *, struct address_space *mapping,
626 loff_t pos, unsigned len, unsigned copied,
627 struct page *page, void *fsdata);
629 struct backing_dev_info;
630 struct address_space {
631 struct inode *host; /* owner: inode, block_device */
632 struct radix_tree_root page_tree; /* radix tree of all pages */
633 spinlock_t tree_lock; /* and lock protecting it */
634 unsigned int i_mmap_writable;/* count VM_SHARED mappings */
635 struct prio_tree_root i_mmap; /* tree of private and shared mappings */
636 struct list_head i_mmap_nonlinear;/*list VM_NONLINEAR mappings */
637 spinlock_t i_mmap_lock; /* protect tree, count, list */
638 unsigned int truncate_count; /* Cover race condition with truncate */
639 unsigned long nrpages; /* number of total pages */
640 pgoff_t writeback_index;/* writeback starts here */
641 const struct address_space_operations *a_ops; /* methods */
642 unsigned long flags; /* error bits/gfp mask */
643 struct backing_dev_info *backing_dev_info; /* device readahead, etc */
644 spinlock_t private_lock; /* for use by the address_space */
645 struct list_head private_list; /* ditto */
646 struct address_space *assoc_mapping; /* ditto */
647 } __attribute__((aligned(sizeof(long))));
649 * On most architectures that alignment is already the case; but
650 * must be enforced here for CRIS, to let the least signficant bit
651 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
654 struct block_device {
655 dev_t bd_dev; /* not a kdev_t - it's a search key */
656 struct inode * bd_inode; /* will die */
657 struct super_block * bd_super;
659 struct mutex bd_mutex; /* open/close mutex */
660 struct list_head bd_inodes;
665 struct list_head bd_holder_list;
667 struct block_device * bd_contains;
668 unsigned bd_block_size;
669 struct hd_struct * bd_part;
670 /* number of times partitions within this device have been opened. */
671 unsigned bd_part_count;
673 struct gendisk * bd_disk;
674 struct list_head bd_list;
676 * Private data. You must have bd_claim'ed the block_device
677 * to use this. NOTE: bd_claim allows an owner to claim
678 * the same device multiple times, the owner must take special
679 * care to not mess up bd_private for that case.
681 unsigned long bd_private;
683 /* The counter of freeze processes */
684 int bd_fsfreeze_count;
685 /* Mutex for freeze */
686 struct mutex bd_fsfreeze_mutex;
690 * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
693 #define PAGECACHE_TAG_DIRTY 0
694 #define PAGECACHE_TAG_WRITEBACK 1
695 #define PAGECACHE_TAG_TOWRITE 2
697 int mapping_tagged(struct address_space *mapping, int tag);
700 * Might pages of this file be mapped into userspace?
702 static inline int mapping_mapped(struct address_space *mapping)
704 return !prio_tree_empty(&mapping->i_mmap) ||
705 !list_empty(&mapping->i_mmap_nonlinear);
709 * Might pages of this file have been modified in userspace?
710 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
711 * marks vma as VM_SHARED if it is shared, and the file was opened for
712 * writing i.e. vma may be mprotected writable even if now readonly.
714 static inline int mapping_writably_mapped(struct address_space *mapping)
716 return mapping->i_mmap_writable != 0;
720 * Use sequence counter to get consistent i_size on 32-bit processors.
722 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
723 #include <linux/seqlock.h>
724 #define __NEED_I_SIZE_ORDERED
725 #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
727 #define i_size_ordered_init(inode) do { } while (0)
731 #define ACL_NOT_CACHED ((void *)(-1))
734 struct hlist_node i_hash;
735 struct list_head i_wb_list; /* backing dev IO list */
736 struct list_head i_lru; /* inode LRU list */
737 struct list_head i_sb_list;
738 struct list_head i_dentry;
741 unsigned int i_nlink;
745 unsigned int i_blkbits;
748 #ifdef __NEED_I_SIZE_ORDERED
749 seqcount_t i_size_seqcount;
751 struct timespec i_atime;
752 struct timespec i_mtime;
753 struct timespec i_ctime;
755 unsigned short i_bytes;
757 spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */
758 struct mutex i_mutex;
759 struct rw_semaphore i_alloc_sem;
760 const struct inode_operations *i_op;
761 const struct file_operations *i_fop; /* former ->i_op->default_file_ops */
762 struct super_block *i_sb;
763 struct file_lock *i_flock;
764 struct address_space *i_mapping;
765 struct address_space i_data;
767 struct dquot *i_dquot[MAXQUOTAS];
769 struct list_head i_devices;
771 struct pipe_inode_info *i_pipe;
772 struct block_device *i_bdev;
778 #ifdef CONFIG_FSNOTIFY
779 __u32 i_fsnotify_mask; /* all events this inode cares about */
780 struct hlist_head i_fsnotify_marks;
783 unsigned long i_state;
784 unsigned long dirtied_when; /* jiffies of first dirtying */
786 unsigned int i_flags;
789 /* protected by i_lock */
790 unsigned int i_readcount; /* struct files open RO */
792 atomic_t i_writecount;
793 #ifdef CONFIG_SECURITY
796 #ifdef CONFIG_FS_POSIX_ACL
797 struct posix_acl *i_acl;
798 struct posix_acl *i_default_acl;
800 void *i_private; /* fs or device private pointer */
803 static inline int inode_unhashed(struct inode *inode)
805 return hlist_unhashed(&inode->i_hash);
809 * inode->i_mutex nesting subclasses for the lock validator:
811 * 0: the object of the current VFS operation
816 * The locking order between these classes is
817 * parent -> child -> normal -> xattr -> quota
819 enum inode_i_mutex_lock_class
829 * NOTE: in a 32bit arch with a preemptable kernel and
830 * an UP compile the i_size_read/write must be atomic
831 * with respect to the local cpu (unlike with preempt disabled),
832 * but they don't need to be atomic with respect to other cpus like in
833 * true SMP (so they need either to either locally disable irq around
834 * the read or for example on x86 they can be still implemented as a
835 * cmpxchg8b without the need of the lock prefix). For SMP compiles
836 * and 64bit archs it makes no difference if preempt is enabled or not.
838 static inline loff_t i_size_read(const struct inode *inode)
840 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
845 seq = read_seqcount_begin(&inode->i_size_seqcount);
846 i_size = inode->i_size;
847 } while (read_seqcount_retry(&inode->i_size_seqcount, seq));
849 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
853 i_size = inode->i_size;
857 return inode->i_size;
862 * NOTE: unlike i_size_read(), i_size_write() does need locking around it
863 * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
864 * can be lost, resulting in subsequent i_size_read() calls spinning forever.
866 static inline void i_size_write(struct inode *inode, loff_t i_size)
868 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
869 write_seqcount_begin(&inode->i_size_seqcount);
870 inode->i_size = i_size;
871 write_seqcount_end(&inode->i_size_seqcount);
872 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
874 inode->i_size = i_size;
877 inode->i_size = i_size;
881 static inline unsigned iminor(const struct inode *inode)
883 return MINOR(inode->i_rdev);
886 static inline unsigned imajor(const struct inode *inode)
888 return MAJOR(inode->i_rdev);
891 extern struct block_device *I_BDEV(struct inode *inode);
894 rwlock_t lock; /* protects pid, uid, euid fields */
895 struct pid *pid; /* pid or -pgrp where SIGIO should be sent */
896 enum pid_type pid_type; /* Kind of process group SIGIO should be sent to */
897 uid_t uid, euid; /* uid/euid of process setting the owner */
898 int signum; /* posix.1b rt signal to be delivered on IO */
902 * Track a single file's readahead state
904 struct file_ra_state {
905 pgoff_t start; /* where readahead started */
906 unsigned int size; /* # of readahead pages */
907 unsigned int async_size; /* do asynchronous readahead when
908 there are only # of pages ahead */
910 unsigned int ra_pages; /* Maximum readahead window */
911 unsigned int mmap_miss; /* Cache miss stat for mmap accesses */
912 loff_t prev_pos; /* Cache last read() position */
916 * Check if @index falls in the readahead windows.
918 static inline int ra_has_index(struct file_ra_state *ra, pgoff_t index)
920 return (index >= ra->start &&
921 index < ra->start + ra->size);
924 #define FILE_MNT_WRITE_TAKEN 1
925 #define FILE_MNT_WRITE_RELEASED 2
929 * fu_list becomes invalid after file_free is called and queued via
930 * fu_rcuhead for RCU freeing
933 struct list_head fu_list;
934 struct rcu_head fu_rcuhead;
937 #define f_dentry f_path.dentry
938 #define f_vfsmnt f_path.mnt
939 const struct file_operations *f_op;
940 spinlock_t f_lock; /* f_ep_links, f_flags, no IRQ */
944 atomic_long_t f_count;
945 unsigned int f_flags;
948 struct fown_struct f_owner;
949 const struct cred *f_cred;
950 struct file_ra_state f_ra;
953 #ifdef CONFIG_SECURITY
956 /* needed for tty driver, and maybe others */
960 /* Used by fs/eventpoll.c to link all the hooks to this file */
961 struct list_head f_ep_links;
962 #endif /* #ifdef CONFIG_EPOLL */
963 struct address_space *f_mapping;
964 #ifdef CONFIG_DEBUG_WRITECOUNT
965 unsigned long f_mnt_write_state;
969 #define get_file(x) atomic_long_inc(&(x)->f_count)
970 #define fput_atomic(x) atomic_long_add_unless(&(x)->f_count, -1, 1)
971 #define file_count(x) atomic_long_read(&(x)->f_count)
973 #ifdef CONFIG_DEBUG_WRITECOUNT
974 static inline void file_take_write(struct file *f)
976 WARN_ON(f->f_mnt_write_state != 0);
977 f->f_mnt_write_state = FILE_MNT_WRITE_TAKEN;
979 static inline void file_release_write(struct file *f)
981 f->f_mnt_write_state |= FILE_MNT_WRITE_RELEASED;
983 static inline void file_reset_write(struct file *f)
985 f->f_mnt_write_state = 0;
987 static inline void file_check_state(struct file *f)
990 * At this point, either both or neither of these bits
993 WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN);
994 WARN_ON(f->f_mnt_write_state == FILE_MNT_WRITE_RELEASED);
996 static inline int file_check_writeable(struct file *f)
998 if (f->f_mnt_write_state == FILE_MNT_WRITE_TAKEN)
1000 printk(KERN_WARNING "writeable file with no "
1001 "mnt_want_write()\n");
1005 #else /* !CONFIG_DEBUG_WRITECOUNT */
1006 static inline void file_take_write(struct file *filp) {}
1007 static inline void file_release_write(struct file *filp) {}
1008 static inline void file_reset_write(struct file *filp) {}
1009 static inline void file_check_state(struct file *filp) {}
1010 static inline int file_check_writeable(struct file *filp)
1014 #endif /* CONFIG_DEBUG_WRITECOUNT */
1016 #define MAX_NON_LFS ((1UL<<31) - 1)
1018 /* Page cache limit. The filesystems should put that into their s_maxbytes
1019 limits, otherwise bad things can happen in VM. */
1020 #if BITS_PER_LONG==32
1021 #define MAX_LFS_FILESIZE (((u64)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
1022 #elif BITS_PER_LONG==64
1023 #define MAX_LFS_FILESIZE 0x7fffffffffffffffUL
1028 #define FL_ACCESS 8 /* not trying to lock, just looking */
1029 #define FL_EXISTS 16 /* when unlocking, test for existence */
1030 #define FL_LEASE 32 /* lease held on this file */
1031 #define FL_CLOSE 64 /* unlock on close */
1032 #define FL_SLEEP 128 /* A blocking lock */
1035 * Special return value from posix_lock_file() and vfs_lock_file() for
1036 * asynchronous locking.
1038 #define FILE_LOCK_DEFERRED 1
1041 * The POSIX file lock owner is determined by
1042 * the "struct files_struct" in the thread group
1043 * (or NULL for no owner - BSD locks).
1045 * Lockd stuffs a "host" pointer into this.
1047 typedef struct files_struct *fl_owner_t;
1049 struct file_lock_operations {
1050 void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
1051 void (*fl_release_private)(struct file_lock *);
1054 struct lock_manager_operations {
1055 int (*fl_compare_owner)(struct file_lock *, struct file_lock *);
1056 void (*fl_notify)(struct file_lock *); /* unblock callback */
1057 int (*fl_grant)(struct file_lock *, struct file_lock *, int);
1058 void (*fl_copy_lock)(struct file_lock *, struct file_lock *);
1059 void (*fl_release_private)(struct file_lock *);
1060 void (*fl_break)(struct file_lock *);
1061 int (*fl_mylease)(struct file_lock *, struct file_lock *);
1062 int (*fl_change)(struct file_lock **, int);
1065 struct lock_manager {
1066 struct list_head list;
1069 void locks_start_grace(struct lock_manager *);
1070 void locks_end_grace(struct lock_manager *);
1071 int locks_in_grace(void);
1073 /* that will die - we need it for nfs_lock_info */
1074 #include <linux/nfs_fs_i.h>
1077 struct file_lock *fl_next; /* singly linked list for this inode */
1078 struct list_head fl_link; /* doubly linked list of all locks */
1079 struct list_head fl_block; /* circular list of blocked processes */
1080 fl_owner_t fl_owner;
1081 unsigned char fl_flags;
1082 unsigned char fl_type;
1083 unsigned int fl_pid;
1084 struct pid *fl_nspid;
1085 wait_queue_head_t fl_wait;
1086 struct file *fl_file;
1090 struct fasync_struct * fl_fasync; /* for lease break notifications */
1091 unsigned long fl_break_time; /* for nonblocking lease breaks */
1093 const struct file_lock_operations *fl_ops; /* Callbacks for filesystems */
1094 const struct lock_manager_operations *fl_lmops; /* Callbacks for lockmanagers */
1096 struct nfs_lock_info nfs_fl;
1097 struct nfs4_lock_info nfs4_fl;
1099 struct list_head link; /* link in AFS vnode's pending_locks list */
1100 int state; /* state of grant or error if -ve */
1105 /* The following constant reflects the upper bound of the file/locking space */
1107 #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
1108 #define OFFSET_MAX INT_LIMIT(loff_t)
1109 #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
1112 #include <linux/fcntl.h>
1114 extern void send_sigio(struct fown_struct *fown, int fd, int band);
1116 #ifdef CONFIG_FILE_LOCKING
1117 extern int fcntl_getlk(struct file *, struct flock __user *);
1118 extern int fcntl_setlk(unsigned int, struct file *, unsigned int,
1119 struct flock __user *);
1121 #if BITS_PER_LONG == 32
1122 extern int fcntl_getlk64(struct file *, struct flock64 __user *);
1123 extern int fcntl_setlk64(unsigned int, struct file *, unsigned int,
1124 struct flock64 __user *);
1127 extern int fcntl_setlease(unsigned int fd, struct file *filp, long arg);
1128 extern int fcntl_getlease(struct file *filp);
1131 extern void locks_init_lock(struct file_lock *);
1132 extern struct file_lock * locks_alloc_lock(void);
1133 extern void locks_copy_lock(struct file_lock *, struct file_lock *);
1134 extern void __locks_copy_lock(struct file_lock *, const struct file_lock *);
1135 extern void locks_remove_posix(struct file *, fl_owner_t);
1136 extern void locks_remove_flock(struct file *);
1137 extern void locks_release_private(struct file_lock *);
1138 extern void posix_test_lock(struct file *, struct file_lock *);
1139 extern int posix_lock_file(struct file *, struct file_lock *, struct file_lock *);
1140 extern int posix_lock_file_wait(struct file *, struct file_lock *);
1141 extern int posix_unblock_lock(struct file *, struct file_lock *);
1142 extern int vfs_test_lock(struct file *, struct file_lock *);
1143 extern int vfs_lock_file(struct file *, unsigned int, struct file_lock *, struct file_lock *);
1144 extern int vfs_cancel_lock(struct file *filp, struct file_lock *fl);
1145 extern int flock_lock_file_wait(struct file *filp, struct file_lock *fl);
1146 extern int __break_lease(struct inode *inode, unsigned int flags);
1147 extern void lease_get_mtime(struct inode *, struct timespec *time);
1148 extern int generic_setlease(struct file *, long, struct file_lock **);
1149 extern int vfs_setlease(struct file *, long, struct file_lock **);
1150 extern int lease_modify(struct file_lock **, int);
1151 extern int lock_may_read(struct inode *, loff_t start, unsigned long count);
1152 extern int lock_may_write(struct inode *, loff_t start, unsigned long count);
1153 extern void lock_flocks(void);
1154 extern void unlock_flocks(void);
1155 #else /* !CONFIG_FILE_LOCKING */
1156 static inline int fcntl_getlk(struct file *file, struct flock __user *user)
1161 static inline int fcntl_setlk(unsigned int fd, struct file *file,
1162 unsigned int cmd, struct flock __user *user)
1167 #if BITS_PER_LONG == 32
1168 static inline int fcntl_getlk64(struct file *file, struct flock64 __user *user)
1173 static inline int fcntl_setlk64(unsigned int fd, struct file *file,
1174 unsigned int cmd, struct flock64 __user *user)
1179 static inline int fcntl_setlease(unsigned int fd, struct file *filp, long arg)
1184 static inline int fcntl_getlease(struct file *filp)
1189 static inline void locks_init_lock(struct file_lock *fl)
1194 static inline void __locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1199 static inline void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
1204 static inline void locks_remove_posix(struct file *filp, fl_owner_t owner)
1209 static inline void locks_remove_flock(struct file *filp)
1214 static inline void posix_test_lock(struct file *filp, struct file_lock *fl)
1219 static inline int posix_lock_file(struct file *filp, struct file_lock *fl,
1220 struct file_lock *conflock)
1225 static inline int posix_lock_file_wait(struct file *filp, struct file_lock *fl)
1230 static inline int posix_unblock_lock(struct file *filp,
1231 struct file_lock *waiter)
1236 static inline int vfs_test_lock(struct file *filp, struct file_lock *fl)
1241 static inline int vfs_lock_file(struct file *filp, unsigned int cmd,
1242 struct file_lock *fl, struct file_lock *conf)
1247 static inline int vfs_cancel_lock(struct file *filp, struct file_lock *fl)
1252 static inline int flock_lock_file_wait(struct file *filp,
1253 struct file_lock *request)
1258 static inline int __break_lease(struct inode *inode, unsigned int mode)
1263 static inline void lease_get_mtime(struct inode *inode, struct timespec *time)
1268 static inline int generic_setlease(struct file *filp, long arg,
1269 struct file_lock **flp)
1274 static inline int vfs_setlease(struct file *filp, long arg,
1275 struct file_lock **lease)
1280 static inline int lease_modify(struct file_lock **before, int arg)
1285 static inline int lock_may_read(struct inode *inode, loff_t start,
1291 static inline int lock_may_write(struct inode *inode, loff_t start,
1297 static inline void lock_flocks(void)
1301 static inline void unlock_flocks(void)
1305 #endif /* !CONFIG_FILE_LOCKING */
1308 struct fasync_struct {
1312 struct fasync_struct *fa_next; /* singly linked list */
1313 struct file *fa_file;
1314 struct rcu_head fa_rcu;
1317 #define FASYNC_MAGIC 0x4601
1319 /* SMP safe fasync helpers: */
1320 extern int fasync_helper(int, struct file *, int, struct fasync_struct **);
1321 extern struct fasync_struct *fasync_insert_entry(int, struct file *, struct fasync_struct **, struct fasync_struct *);
1322 extern int fasync_remove_entry(struct file *, struct fasync_struct **);
1323 extern struct fasync_struct *fasync_alloc(void);
1324 extern void fasync_free(struct fasync_struct *);
1326 /* can be called from interrupts */
1327 extern void kill_fasync(struct fasync_struct **, int, int);
1329 extern int __f_setown(struct file *filp, struct pid *, enum pid_type, int force);
1330 extern int f_setown(struct file *filp, unsigned long arg, int force);
1331 extern void f_delown(struct file *filp);
1332 extern pid_t f_getown(struct file *filp);
1333 extern int send_sigurg(struct fown_struct *fown);
1339 #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
1340 #define MNT_DETACH 0x00000002 /* Just detach from the tree */
1341 #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
1342 #define UMOUNT_NOFOLLOW 0x00000008 /* Don't follow symlink on umount */
1343 #define UMOUNT_UNUSED 0x80000000 /* Flag guaranteed to be unused */
1345 extern struct list_head super_blocks;
1346 extern spinlock_t sb_lock;
1348 struct super_block {
1349 struct list_head s_list; /* Keep this first */
1350 dev_t s_dev; /* search index; _not_ kdev_t */
1351 unsigned char s_dirt;
1352 unsigned char s_blocksize_bits;
1353 unsigned long s_blocksize;
1354 loff_t s_maxbytes; /* Max file size */
1355 struct file_system_type *s_type;
1356 const struct super_operations *s_op;
1357 const struct dquot_operations *dq_op;
1358 const struct quotactl_ops *s_qcop;
1359 const struct export_operations *s_export_op;
1360 unsigned long s_flags;
1361 unsigned long s_magic;
1362 struct dentry *s_root;
1363 struct rw_semaphore s_umount;
1364 struct mutex s_lock;
1367 #ifdef CONFIG_SECURITY
1370 const struct xattr_handler **s_xattr;
1372 struct list_head s_inodes; /* all inodes */
1373 struct hlist_head s_anon; /* anonymous dentries for (nfs) exporting */
1375 struct list_head __percpu *s_files;
1377 struct list_head s_files;
1379 /* s_dentry_lru and s_nr_dentry_unused are protected by dcache_lock */
1380 struct list_head s_dentry_lru; /* unused dentry lru */
1381 int s_nr_dentry_unused; /* # of dentry on lru */
1383 struct block_device *s_bdev;
1384 struct backing_dev_info *s_bdi;
1385 struct mtd_info *s_mtd;
1386 struct list_head s_instances;
1387 struct quota_info s_dquot; /* Diskquota specific options */
1390 wait_queue_head_t s_wait_unfrozen;
1392 char s_id[32]; /* Informational name */
1394 void *s_fs_info; /* Filesystem private info */
1397 /* Granularity of c/m/atime in ns.
1398 Cannot be worse than a second */
1402 * The next field is for VFS *only*. No filesystems have any business
1403 * even looking at it. You had been warned.
1405 struct mutex s_vfs_rename_mutex; /* Kludge */
1408 * Filesystem subtype. If non-empty the filesystem type field
1409 * in /proc/mounts will be "type.subtype"
1414 * Saved mount options for lazy filesystems using
1415 * generic_show_options()
1417 char __rcu *s_options;
1420 extern struct timespec current_fs_time(struct super_block *sb);
1423 * Snapshotting support.
1427 SB_FREEZE_WRITE = 1,
1428 SB_FREEZE_TRANS = 2,
1431 #define vfs_check_frozen(sb, level) \
1432 wait_event((sb)->s_wait_unfrozen, ((sb)->s_frozen < (level)))
1434 #define get_fs_excl() atomic_inc(¤t->fs_excl)
1435 #define put_fs_excl() atomic_dec(¤t->fs_excl)
1436 #define has_fs_excl() atomic_read(¤t->fs_excl)
1438 #define is_owner_or_cap(inode) \
1439 ((current_fsuid() == (inode)->i_uid) || capable(CAP_FOWNER))
1441 /* not quite ready to be deprecated, but... */
1442 extern void lock_super(struct super_block *);
1443 extern void unlock_super(struct super_block *);
1446 * VFS helper functions..
1448 extern int vfs_create(struct inode *, struct dentry *, int, struct nameidata *);
1449 extern int vfs_mkdir(struct inode *, struct dentry *, int);
1450 extern int vfs_mknod(struct inode *, struct dentry *, int, dev_t);
1451 extern int vfs_symlink(struct inode *, struct dentry *, const char *);
1452 extern int vfs_link(struct dentry *, struct inode *, struct dentry *);
1453 extern int vfs_rmdir(struct inode *, struct dentry *);
1454 extern int vfs_unlink(struct inode *, struct dentry *);
1455 extern int vfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
1458 * VFS dentry helper functions.
1460 extern void dentry_unhash(struct dentry *dentry);
1463 * VFS file helper functions.
1465 extern int file_permission(struct file *, int);
1466 extern void inode_init_owner(struct inode *inode, const struct inode *dir,
1469 * VFS FS_IOC_FIEMAP helper definitions.
1471 struct fiemap_extent_info {
1472 unsigned int fi_flags; /* Flags as passed from user */
1473 unsigned int fi_extents_mapped; /* Number of mapped extents */
1474 unsigned int fi_extents_max; /* Size of fiemap_extent array */
1475 struct fiemap_extent *fi_extents_start; /* Start of fiemap_extent
1478 int fiemap_fill_next_extent(struct fiemap_extent_info *info, u64 logical,
1479 u64 phys, u64 len, u32 flags);
1480 int fiemap_check_flags(struct fiemap_extent_info *fieinfo, u32 fs_flags);
1485 * NOTE! These match bits 12..15 of stat.st_mode
1486 * (ie "(i_mode >> 12) & 15").
1488 #define DT_UNKNOWN 0
1499 * This is the "filldir" function type, used by readdir() to let
1500 * the kernel specify what kind of dirent layout it wants to have.
1501 * This allows the kernel to read directories into kernel space or
1502 * to have different dirent layouts depending on the binary type.
1504 typedef int (*filldir_t)(void *, const char *, int, loff_t, u64, unsigned);
1505 struct block_device_operations;
1507 /* These macros are for out of kernel modules to test that
1508 * the kernel supports the unlocked_ioctl and compat_ioctl
1509 * fields in struct file_operations. */
1510 #define HAVE_COMPAT_IOCTL 1
1511 #define HAVE_UNLOCKED_IOCTL 1
1515 * all file operations except setlease can be called without
1516 * the big kernel lock held in all filesystems.
1518 struct file_operations {
1519 struct module *owner;
1520 loff_t (*llseek) (struct file *, loff_t, int);
1521 ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
1522 ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
1523 ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1524 ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1525 int (*readdir) (struct file *, void *, filldir_t);
1526 unsigned int (*poll) (struct file *, struct poll_table_struct *);
1527 long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
1528 long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
1529 int (*mmap) (struct file *, struct vm_area_struct *);
1530 int (*open) (struct inode *, struct file *);
1531 int (*flush) (struct file *, fl_owner_t id);
1532 int (*release) (struct inode *, struct file *);
1533 int (*fsync) (struct file *, int datasync);
1534 int (*aio_fsync) (struct kiocb *, int datasync);
1535 int (*fasync) (int, struct file *, int);
1536 int (*lock) (struct file *, int, struct file_lock *);
1537 ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
1538 unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
1539 int (*check_flags)(int);
1540 int (*flock) (struct file *, int, struct file_lock *);
1541 ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int);
1542 ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int);
1543 int (*setlease)(struct file *, long, struct file_lock **);
1546 struct inode_operations {
1547 int (*create) (struct inode *,struct dentry *,int, struct nameidata *);
1548 struct dentry * (*lookup) (struct inode *,struct dentry *, struct nameidata *);
1549 int (*link) (struct dentry *,struct inode *,struct dentry *);
1550 int (*unlink) (struct inode *,struct dentry *);
1551 int (*symlink) (struct inode *,struct dentry *,const char *);
1552 int (*mkdir) (struct inode *,struct dentry *,int);
1553 int (*rmdir) (struct inode *,struct dentry *);
1554 int (*mknod) (struct inode *,struct dentry *,int,dev_t);
1555 int (*rename) (struct inode *, struct dentry *,
1556 struct inode *, struct dentry *);
1557 int (*readlink) (struct dentry *, char __user *,int);
1558 void * (*follow_link) (struct dentry *, struct nameidata *);
1559 void (*put_link) (struct dentry *, struct nameidata *, void *);
1560 void (*truncate) (struct inode *);
1561 int (*permission) (struct inode *, int);
1562 int (*check_acl)(struct inode *, int);
1563 int (*setattr) (struct dentry *, struct iattr *);
1564 int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
1565 int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
1566 ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
1567 ssize_t (*listxattr) (struct dentry *, char *, size_t);
1568 int (*removexattr) (struct dentry *, const char *);
1569 void (*truncate_range)(struct inode *, loff_t, loff_t);
1570 long (*fallocate)(struct inode *inode, int mode, loff_t offset,
1572 int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start,
1578 ssize_t rw_copy_check_uvector(int type, const struct iovec __user * uvector,
1579 unsigned long nr_segs, unsigned long fast_segs,
1580 struct iovec *fast_pointer,
1581 struct iovec **ret_pointer);
1583 extern ssize_t vfs_read(struct file *, char __user *, size_t, loff_t *);
1584 extern ssize_t vfs_write(struct file *, const char __user *, size_t, loff_t *);
1585 extern ssize_t vfs_readv(struct file *, const struct iovec __user *,
1586 unsigned long, loff_t *);
1587 extern ssize_t vfs_writev(struct file *, const struct iovec __user *,
1588 unsigned long, loff_t *);
1590 struct super_operations {
1591 struct inode *(*alloc_inode)(struct super_block *sb);
1592 void (*destroy_inode)(struct inode *);
1594 void (*dirty_inode) (struct inode *);
1595 int (*write_inode) (struct inode *, struct writeback_control *wbc);
1596 int (*drop_inode) (struct inode *);
1597 void (*evict_inode) (struct inode *);
1598 void (*put_super) (struct super_block *);
1599 void (*write_super) (struct super_block *);
1600 int (*sync_fs)(struct super_block *sb, int wait);
1601 int (*freeze_fs) (struct super_block *);
1602 int (*unfreeze_fs) (struct super_block *);
1603 int (*statfs) (struct dentry *, struct kstatfs *);
1604 int (*remount_fs) (struct super_block *, int *, char *);
1605 void (*umount_begin) (struct super_block *);
1607 int (*show_options)(struct seq_file *, struct vfsmount *);
1608 int (*show_stats)(struct seq_file *, struct vfsmount *);
1610 ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
1611 ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
1613 int (*bdev_try_to_free_page)(struct super_block*, struct page*, gfp_t);
1614 int (*trim_fs) (struct super_block *, struct fstrim_range *);
1618 * Inode state bits. Protected by inode_lock.
1620 * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
1621 * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
1623 * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
1624 * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
1625 * various stages of removing an inode.
1627 * Two bits are used for locking and completion notification, I_NEW and I_SYNC.
1629 * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
1630 * fdatasync(). i_atime is the usual cause.
1631 * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
1632 * these changes separately from I_DIRTY_SYNC so that we
1633 * don't have to write inode on fdatasync() when only
1634 * mtime has changed in it.
1635 * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
1636 * I_NEW Serves as both a mutex and completion notification.
1637 * New inodes set I_NEW. If two processes both create
1638 * the same inode, one of them will release its inode and
1639 * wait for I_NEW to be released before returning.
1640 * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
1641 * also cause waiting on I_NEW, without I_NEW actually
1642 * being set. find_inode() uses this to prevent returning
1643 * nearly-dead inodes.
1644 * I_WILL_FREE Must be set when calling write_inode_now() if i_count
1645 * is zero. I_FREEING must be set when I_WILL_FREE is
1647 * I_FREEING Set when inode is about to be freed but still has dirty
1648 * pages or buffers attached or the inode itself is still
1650 * I_CLEAR Added by end_writeback(). In this state the inode is clean
1651 * and can be destroyed. Inode keeps I_FREEING.
1653 * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
1654 * prohibited for many purposes. iget() must wait for
1655 * the inode to be completely released, then create it
1656 * anew. Other functions will just ignore such inodes,
1657 * if appropriate. I_NEW is used for waiting.
1659 * I_SYNC Synchonized write of dirty inode data. The bits is
1660 * set during data writeback, and cleared with a wakeup
1661 * on the bit address once it is done.
1663 * Q: What is the difference between I_WILL_FREE and I_FREEING?
1665 #define I_DIRTY_SYNC (1 << 0)
1666 #define I_DIRTY_DATASYNC (1 << 1)
1667 #define I_DIRTY_PAGES (1 << 2)
1669 #define I_NEW (1 << __I_NEW)
1670 #define I_WILL_FREE (1 << 4)
1671 #define I_FREEING (1 << 5)
1672 #define I_CLEAR (1 << 6)
1674 #define I_SYNC (1 << __I_SYNC)
1675 #define I_REFERENCED (1 << 8)
1677 #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1679 extern void __mark_inode_dirty(struct inode *, int);
1680 static inline void mark_inode_dirty(struct inode *inode)
1682 __mark_inode_dirty(inode, I_DIRTY);
1685 static inline void mark_inode_dirty_sync(struct inode *inode)
1687 __mark_inode_dirty(inode, I_DIRTY_SYNC);
1691 * inc_nlink - directly increment an inode's link count
1694 * This is a low-level filesystem helper to replace any
1695 * direct filesystem manipulation of i_nlink. Currently,
1696 * it is only here for parity with dec_nlink().
1698 static inline void inc_nlink(struct inode *inode)
1703 static inline void inode_inc_link_count(struct inode *inode)
1706 mark_inode_dirty(inode);
1710 * drop_nlink - directly drop an inode's link count
1713 * This is a low-level filesystem helper to replace any
1714 * direct filesystem manipulation of i_nlink. In cases
1715 * where we are attempting to track writes to the
1716 * filesystem, a decrement to zero means an imminent
1717 * write when the file is truncated and actually unlinked
1718 * on the filesystem.
1720 static inline void drop_nlink(struct inode *inode)
1726 * clear_nlink - directly zero an inode's link count
1729 * This is a low-level filesystem helper to replace any
1730 * direct filesystem manipulation of i_nlink. See
1731 * drop_nlink() for why we care about i_nlink hitting zero.
1733 static inline void clear_nlink(struct inode *inode)
1738 static inline void inode_dec_link_count(struct inode *inode)
1741 mark_inode_dirty(inode);
1745 * inode_inc_iversion - increments i_version
1746 * @inode: inode that need to be updated
1748 * Every time the inode is modified, the i_version field will be incremented.
1749 * The filesystem has to be mounted with i_version flag
1752 static inline void inode_inc_iversion(struct inode *inode)
1754 spin_lock(&inode->i_lock);
1756 spin_unlock(&inode->i_lock);
1759 extern void touch_atime(struct vfsmount *mnt, struct dentry *dentry);
1760 static inline void file_accessed(struct file *file)
1762 if (!(file->f_flags & O_NOATIME))
1763 touch_atime(file->f_path.mnt, file->f_path.dentry);
1766 int sync_inode(struct inode *inode, struct writeback_control *wbc);
1767 int sync_inode_metadata(struct inode *inode, int wait);
1769 struct file_system_type {
1772 int (*get_sb) (struct file_system_type *, int,
1773 const char *, void *, struct vfsmount *);
1774 void (*kill_sb) (struct super_block *);
1775 struct module *owner;
1776 struct file_system_type * next;
1777 struct list_head fs_supers;
1779 struct lock_class_key s_lock_key;
1780 struct lock_class_key s_umount_key;
1781 struct lock_class_key s_vfs_rename_key;
1783 struct lock_class_key i_lock_key;
1784 struct lock_class_key i_mutex_key;
1785 struct lock_class_key i_mutex_dir_key;
1786 struct lock_class_key i_alloc_sem_key;
1789 extern int get_sb_ns(struct file_system_type *fs_type, int flags, void *data,
1790 int (*fill_super)(struct super_block *, void *, int),
1791 struct vfsmount *mnt);
1792 extern int get_sb_bdev(struct file_system_type *fs_type,
1793 int flags, const char *dev_name, void *data,
1794 int (*fill_super)(struct super_block *, void *, int),
1795 struct vfsmount *mnt);
1796 extern int get_sb_single(struct file_system_type *fs_type,
1797 int flags, void *data,
1798 int (*fill_super)(struct super_block *, void *, int),
1799 struct vfsmount *mnt);
1800 extern int get_sb_nodev(struct file_system_type *fs_type,
1801 int flags, void *data,
1802 int (*fill_super)(struct super_block *, void *, int),
1803 struct vfsmount *mnt);
1804 void generic_shutdown_super(struct super_block *sb);
1805 void kill_block_super(struct super_block *sb);
1806 void kill_anon_super(struct super_block *sb);
1807 void kill_litter_super(struct super_block *sb);
1808 void deactivate_super(struct super_block *sb);
1809 void deactivate_locked_super(struct super_block *sb);
1810 int set_anon_super(struct super_block *s, void *data);
1811 struct super_block *sget(struct file_system_type *type,
1812 int (*test)(struct super_block *,void *),
1813 int (*set)(struct super_block *,void *),
1815 extern int get_sb_pseudo(struct file_system_type *, char *,
1816 const struct super_operations *ops, unsigned long,
1817 struct vfsmount *mnt);
1818 extern void simple_set_mnt(struct vfsmount *mnt, struct super_block *sb);
1820 static inline void sb_mark_dirty(struct super_block *sb)
1824 static inline void sb_mark_clean(struct super_block *sb)
1828 static inline int sb_is_dirty(struct super_block *sb)
1833 /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
1834 #define fops_get(fops) \
1835 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
1836 #define fops_put(fops) \
1837 do { if (fops) module_put((fops)->owner); } while(0)
1839 extern int register_filesystem(struct file_system_type *);
1840 extern int unregister_filesystem(struct file_system_type *);
1841 extern struct vfsmount *kern_mount_data(struct file_system_type *, void *data);
1842 #define kern_mount(type) kern_mount_data(type, NULL)
1843 extern int may_umount_tree(struct vfsmount *);
1844 extern int may_umount(struct vfsmount *);
1845 extern long do_mount(char *, char *, char *, unsigned long, void *);
1846 extern struct vfsmount *collect_mounts(struct path *);
1847 extern void drop_collected_mounts(struct vfsmount *);
1848 extern int iterate_mounts(int (*)(struct vfsmount *, void *), void *,
1850 extern int vfs_statfs(struct path *, struct kstatfs *);
1851 extern int statfs_by_dentry(struct dentry *, struct kstatfs *);
1852 extern int freeze_super(struct super_block *super);
1853 extern int thaw_super(struct super_block *super);
1855 extern int current_umask(void);
1858 extern struct kobject *fs_kobj;
1860 extern int rw_verify_area(int, struct file *, loff_t *, size_t);
1862 #define FLOCK_VERIFY_READ 1
1863 #define FLOCK_VERIFY_WRITE 2
1865 #ifdef CONFIG_FILE_LOCKING
1866 extern int locks_mandatory_locked(struct inode *);
1867 extern int locks_mandatory_area(int, struct inode *, struct file *, loff_t, size_t);
1870 * Candidates for mandatory locking have the setgid bit set
1871 * but no group execute bit - an otherwise meaningless combination.
1874 static inline int __mandatory_lock(struct inode *ino)
1876 return (ino->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID;
1880 * ... and these candidates should be on MS_MANDLOCK mounted fs,
1881 * otherwise these will be advisory locks
1884 static inline int mandatory_lock(struct inode *ino)
1886 return IS_MANDLOCK(ino) && __mandatory_lock(ino);
1889 static inline int locks_verify_locked(struct inode *inode)
1891 if (mandatory_lock(inode))
1892 return locks_mandatory_locked(inode);
1896 static inline int locks_verify_truncate(struct inode *inode,
1900 if (inode->i_flock && mandatory_lock(inode))
1901 return locks_mandatory_area(
1902 FLOCK_VERIFY_WRITE, inode, filp,
1903 size < inode->i_size ? size : inode->i_size,
1904 (size < inode->i_size ? inode->i_size - size
1905 : size - inode->i_size)
1910 static inline int break_lease(struct inode *inode, unsigned int mode)
1913 return __break_lease(inode, mode);
1916 #else /* !CONFIG_FILE_LOCKING */
1917 static inline int locks_mandatory_locked(struct inode *inode)
1922 static inline int locks_mandatory_area(int rw, struct inode *inode,
1923 struct file *filp, loff_t offset,
1929 static inline int __mandatory_lock(struct inode *inode)
1934 static inline int mandatory_lock(struct inode *inode)
1939 static inline int locks_verify_locked(struct inode *inode)
1944 static inline int locks_verify_truncate(struct inode *inode, struct file *filp,
1950 static inline int break_lease(struct inode *inode, unsigned int mode)
1955 #endif /* CONFIG_FILE_LOCKING */
1959 extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs,
1961 extern int do_fallocate(struct file *file, int mode, loff_t offset,
1963 extern long do_sys_open(int dfd, const char __user *filename, int flags,
1965 extern struct file *filp_open(const char *, int, int);
1966 extern struct file * dentry_open(struct dentry *, struct vfsmount *, int,
1967 const struct cred *);
1968 extern int filp_close(struct file *, fl_owner_t id);
1969 extern char * getname(const char __user *);
1973 extern int ioctl_preallocate(struct file *filp, void __user *argp);
1976 extern void __init vfs_caches_init_early(void);
1977 extern void __init vfs_caches_init(unsigned long);
1979 extern struct kmem_cache *names_cachep;
1981 #define __getname_gfp(gfp) kmem_cache_alloc(names_cachep, (gfp))
1982 #define __getname() __getname_gfp(GFP_KERNEL)
1983 #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
1984 #ifndef CONFIG_AUDITSYSCALL
1985 #define putname(name) __putname(name)
1987 extern void putname(const char *name);
1991 extern int register_blkdev(unsigned int, const char *);
1992 extern void unregister_blkdev(unsigned int, const char *);
1993 extern struct block_device *bdget(dev_t);
1994 extern struct block_device *bdgrab(struct block_device *bdev);
1995 extern void bd_set_size(struct block_device *, loff_t size);
1996 extern void bd_forget(struct inode *inode);
1997 extern void bdput(struct block_device *);
1998 extern struct block_device *open_by_devnum(dev_t, fmode_t);
1999 extern void invalidate_bdev(struct block_device *);
2000 extern int sync_blockdev(struct block_device *bdev);
2001 extern struct super_block *freeze_bdev(struct block_device *);
2002 extern void emergency_thaw_all(void);
2003 extern int thaw_bdev(struct block_device *bdev, struct super_block *sb);
2004 extern int fsync_bdev(struct block_device *);
2006 static inline void bd_forget(struct inode *inode) {}
2007 static inline int sync_blockdev(struct block_device *bdev) { return 0; }
2008 static inline void invalidate_bdev(struct block_device *bdev) {}
2010 static inline struct super_block *freeze_bdev(struct block_device *sb)
2015 static inline int thaw_bdev(struct block_device *bdev, struct super_block *sb)
2020 extern int sync_filesystem(struct super_block *);
2021 extern const struct file_operations def_blk_fops;
2022 extern const struct file_operations def_chr_fops;
2023 extern const struct file_operations bad_sock_fops;
2024 extern const struct file_operations def_fifo_fops;
2026 extern int ioctl_by_bdev(struct block_device *, unsigned, unsigned long);
2027 extern int blkdev_ioctl(struct block_device *, fmode_t, unsigned, unsigned long);
2028 extern long compat_blkdev_ioctl(struct file *, unsigned, unsigned long);
2029 extern int blkdev_get(struct block_device *, fmode_t);
2030 extern int blkdev_put(struct block_device *, fmode_t);
2031 extern int bd_claim(struct block_device *, void *);
2032 extern void bd_release(struct block_device *);
2034 extern int bd_claim_by_disk(struct block_device *, void *, struct gendisk *);
2035 extern void bd_release_from_disk(struct block_device *, struct gendisk *);
2037 #define bd_claim_by_disk(bdev, holder, disk) bd_claim(bdev, holder)
2038 #define bd_release_from_disk(bdev, disk) bd_release(bdev)
2043 #define CHRDEV_MAJOR_HASH_SIZE 255
2044 extern int alloc_chrdev_region(dev_t *, unsigned, unsigned, const char *);
2045 extern int register_chrdev_region(dev_t, unsigned, const char *);
2046 extern int __register_chrdev(unsigned int major, unsigned int baseminor,
2047 unsigned int count, const char *name,
2048 const struct file_operations *fops);
2049 extern void __unregister_chrdev(unsigned int major, unsigned int baseminor,
2050 unsigned int count, const char *name);
2051 extern void unregister_chrdev_region(dev_t, unsigned);
2052 extern void chrdev_show(struct seq_file *,off_t);
2054 static inline int register_chrdev(unsigned int major, const char *name,
2055 const struct file_operations *fops)
2057 return __register_chrdev(major, 0, 256, name, fops);
2060 static inline void unregister_chrdev(unsigned int major, const char *name)
2062 __unregister_chrdev(major, 0, 256, name);
2065 /* fs/block_dev.c */
2066 #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
2067 #define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
2070 #define BLKDEV_MAJOR_HASH_SIZE 255
2071 extern const char *__bdevname(dev_t, char *buffer);
2072 extern const char *bdevname(struct block_device *bdev, char *buffer);
2073 extern struct block_device *lookup_bdev(const char *);
2074 extern struct block_device *open_bdev_exclusive(const char *, fmode_t, void *);
2075 extern void close_bdev_exclusive(struct block_device *, fmode_t);
2076 extern void blkdev_show(struct seq_file *,off_t);
2079 #define BLKDEV_MAJOR_HASH_SIZE 0
2082 extern void init_special_inode(struct inode *, umode_t, dev_t);
2084 /* Invalid inode operations -- fs/bad_inode.c */
2085 extern void make_bad_inode(struct inode *);
2086 extern int is_bad_inode(struct inode *);
2088 extern const struct file_operations read_pipefifo_fops;
2089 extern const struct file_operations write_pipefifo_fops;
2090 extern const struct file_operations rdwr_pipefifo_fops;
2092 extern int fs_may_remount_ro(struct super_block *);
2096 * return READ, READA, or WRITE
2098 #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
2101 * return data direction, READ or WRITE
2103 #define bio_data_dir(bio) ((bio)->bi_rw & 1)
2105 extern void check_disk_size_change(struct gendisk *disk,
2106 struct block_device *bdev);
2107 extern int revalidate_disk(struct gendisk *);
2108 extern int check_disk_change(struct block_device *);
2109 extern int __invalidate_device(struct block_device *);
2110 extern int invalidate_partition(struct gendisk *, int);
2112 unsigned long invalidate_mapping_pages(struct address_space *mapping,
2113 pgoff_t start, pgoff_t end);
2115 static inline void invalidate_remote_inode(struct inode *inode)
2117 if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
2118 S_ISLNK(inode->i_mode))
2119 invalidate_mapping_pages(inode->i_mapping, 0, -1);
2121 extern int invalidate_inode_pages2(struct address_space *mapping);
2122 extern int invalidate_inode_pages2_range(struct address_space *mapping,
2123 pgoff_t start, pgoff_t end);
2124 extern int write_inode_now(struct inode *, int);
2125 extern int filemap_fdatawrite(struct address_space *);
2126 extern int filemap_flush(struct address_space *);
2127 extern int filemap_fdatawait(struct address_space *);
2128 extern int filemap_fdatawait_range(struct address_space *, loff_t lstart,
2130 extern int filemap_write_and_wait(struct address_space *mapping);
2131 extern int filemap_write_and_wait_range(struct address_space *mapping,
2132 loff_t lstart, loff_t lend);
2133 extern int __filemap_fdatawrite_range(struct address_space *mapping,
2134 loff_t start, loff_t end, int sync_mode);
2135 extern int filemap_fdatawrite_range(struct address_space *mapping,
2136 loff_t start, loff_t end);
2138 extern int vfs_fsync_range(struct file *file, loff_t start, loff_t end,
2140 extern int vfs_fsync(struct file *file, int datasync);
2141 extern int generic_write_sync(struct file *file, loff_t pos, loff_t count);
2142 extern void sync_supers(void);
2143 extern void emergency_sync(void);
2144 extern void emergency_remount(void);
2146 extern sector_t bmap(struct inode *, sector_t);
2148 extern int notify_change(struct dentry *, struct iattr *);
2149 extern int inode_permission(struct inode *, int);
2150 extern int generic_permission(struct inode *, int,
2151 int (*check_acl)(struct inode *, int));
2153 static inline bool execute_ok(struct inode *inode)
2155 return (inode->i_mode & S_IXUGO) || S_ISDIR(inode->i_mode);
2158 extern int get_write_access(struct inode *);
2159 extern int deny_write_access(struct file *);
2160 static inline void put_write_access(struct inode * inode)
2162 atomic_dec(&inode->i_writecount);
2164 static inline void allow_write_access(struct file *file)
2167 atomic_inc(&file->f_path.dentry->d_inode->i_writecount);
2169 extern int do_pipe_flags(int *, int);
2170 extern struct file *create_read_pipe(struct file *f, int flags);
2171 extern struct file *create_write_pipe(int flags);
2172 extern void free_write_pipe(struct file *);
2174 extern struct file *do_filp_open(int dfd, const char *pathname,
2175 int open_flag, int mode, int acc_mode);
2176 extern int may_open(struct path *, int, int);
2178 extern int kernel_read(struct file *, loff_t, char *, unsigned long);
2179 extern struct file * open_exec(const char *);
2181 /* fs/dcache.c -- generic fs support functions */
2182 extern int is_subdir(struct dentry *, struct dentry *);
2183 extern int path_is_under(struct path *, struct path *);
2184 extern ino_t find_inode_number(struct dentry *, struct qstr *);
2186 #include <linux/err.h>
2188 /* needed for stackable file system support */
2189 extern loff_t default_llseek(struct file *file, loff_t offset, int origin);
2191 extern loff_t vfs_llseek(struct file *file, loff_t offset, int origin);
2193 extern int inode_init_always(struct super_block *, struct inode *);
2194 extern void inode_init_once(struct inode *);
2195 extern void ihold(struct inode * inode);
2196 extern void iput(struct inode *);
2197 extern struct inode * igrab(struct inode *);
2198 extern ino_t iunique(struct super_block *, ino_t);
2199 extern int inode_needs_sync(struct inode *inode);
2200 extern int generic_delete_inode(struct inode *inode);
2201 extern int generic_drop_inode(struct inode *inode);
2203 extern struct inode *ilookup5_nowait(struct super_block *sb,
2204 unsigned long hashval, int (*test)(struct inode *, void *),
2206 extern struct inode *ilookup5(struct super_block *sb, unsigned long hashval,
2207 int (*test)(struct inode *, void *), void *data);
2208 extern struct inode *ilookup(struct super_block *sb, unsigned long ino);
2210 extern struct inode * iget5_locked(struct super_block *, unsigned long, int (*test)(struct inode *, void *), int (*set)(struct inode *, void *), void *);
2211 extern struct inode * iget_locked(struct super_block *, unsigned long);
2212 extern int insert_inode_locked4(struct inode *, unsigned long, int (*test)(struct inode *, void *), void *);
2213 extern int insert_inode_locked(struct inode *);
2214 extern void unlock_new_inode(struct inode *);
2215 extern unsigned int get_next_ino(void);
2217 extern void __iget(struct inode * inode);
2218 extern void iget_failed(struct inode *);
2219 extern void end_writeback(struct inode *);
2220 extern void __destroy_inode(struct inode *);
2221 extern struct inode *new_inode(struct super_block *);
2222 extern int should_remove_suid(struct dentry *);
2223 extern int file_remove_suid(struct file *);
2225 extern void __insert_inode_hash(struct inode *, unsigned long hashval);
2226 extern void remove_inode_hash(struct inode *);
2227 static inline void insert_inode_hash(struct inode *inode)
2229 __insert_inode_hash(inode, inode->i_ino);
2231 extern void inode_sb_list_add(struct inode *inode);
2234 extern void submit_bio(int, struct bio *);
2235 extern int bdev_read_only(struct block_device *);
2237 extern int set_blocksize(struct block_device *, int);
2238 extern int sb_set_blocksize(struct super_block *, int);
2239 extern int sb_min_blocksize(struct super_block *, int);
2241 extern int generic_file_mmap(struct file *, struct vm_area_struct *);
2242 extern int generic_file_readonly_mmap(struct file *, struct vm_area_struct *);
2243 extern int file_read_actor(read_descriptor_t * desc, struct page *page, unsigned long offset, unsigned long size);
2244 int generic_write_checks(struct file *file, loff_t *pos, size_t *count, int isblk);
2245 extern ssize_t generic_file_aio_read(struct kiocb *, const struct iovec *, unsigned long, loff_t);
2246 extern ssize_t __generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long,
2248 extern ssize_t generic_file_aio_write(struct kiocb *, const struct iovec *, unsigned long, loff_t);
2249 extern ssize_t generic_file_direct_write(struct kiocb *, const struct iovec *,
2250 unsigned long *, loff_t, loff_t *, size_t, size_t);
2251 extern ssize_t generic_file_buffered_write(struct kiocb *, const struct iovec *,
2252 unsigned long, loff_t, loff_t *, size_t, ssize_t);
2253 extern ssize_t do_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos);
2254 extern ssize_t do_sync_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos);
2255 extern int generic_segment_checks(const struct iovec *iov,
2256 unsigned long *nr_segs, size_t *count, int access_flags);
2258 /* fs/block_dev.c */
2259 extern ssize_t blkdev_aio_write(struct kiocb *iocb, const struct iovec *iov,
2260 unsigned long nr_segs, loff_t pos);
2261 extern int blkdev_fsync(struct file *filp, int datasync);
2264 extern ssize_t generic_file_splice_read(struct file *, loff_t *,
2265 struct pipe_inode_info *, size_t, unsigned int);
2266 extern ssize_t default_file_splice_read(struct file *, loff_t *,
2267 struct pipe_inode_info *, size_t, unsigned int);
2268 extern ssize_t generic_file_splice_write(struct pipe_inode_info *,
2269 struct file *, loff_t *, size_t, unsigned int);
2270 extern ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe,
2271 struct file *out, loff_t *, size_t len, unsigned int flags);
2272 extern long do_splice_direct(struct file *in, loff_t *ppos, struct file *out,
2273 size_t len, unsigned int flags);
2276 file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping);
2277 extern loff_t noop_llseek(struct file *file, loff_t offset, int origin);
2278 extern loff_t no_llseek(struct file *file, loff_t offset, int origin);
2279 extern loff_t generic_file_llseek(struct file *file, loff_t offset, int origin);
2280 extern loff_t generic_file_llseek_unlocked(struct file *file, loff_t offset,
2282 extern int generic_file_open(struct inode * inode, struct file * filp);
2283 extern int nonseekable_open(struct inode * inode, struct file * filp);
2285 #ifdef CONFIG_FS_XIP
2286 extern ssize_t xip_file_read(struct file *filp, char __user *buf, size_t len,
2288 extern int xip_file_mmap(struct file * file, struct vm_area_struct * vma);
2289 extern ssize_t xip_file_write(struct file *filp, const char __user *buf,
2290 size_t len, loff_t *ppos);
2291 extern int xip_truncate_page(struct address_space *mapping, loff_t from);
2293 static inline int xip_truncate_page(struct address_space *mapping, loff_t from)
2300 typedef void (dio_submit_t)(int rw, struct bio *bio, struct inode *inode,
2301 loff_t file_offset);
2304 /* need locking between buffered and direct access */
2307 /* filesystem does not support filling holes */
2308 DIO_SKIP_HOLES = 0x02,
2311 void dio_end_io(struct bio *bio, int error);
2313 ssize_t __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
2314 struct block_device *bdev, const struct iovec *iov, loff_t offset,
2315 unsigned long nr_segs, get_block_t get_block, dio_iodone_t end_io,
2316 dio_submit_t submit_io, int flags);
2318 static inline ssize_t blockdev_direct_IO(int rw, struct kiocb *iocb,
2319 struct inode *inode, struct block_device *bdev, const struct iovec *iov,
2320 loff_t offset, unsigned long nr_segs, get_block_t get_block,
2321 dio_iodone_t end_io)
2323 return __blockdev_direct_IO(rw, iocb, inode, bdev, iov, offset,
2324 nr_segs, get_block, end_io, NULL,
2325 DIO_LOCKING | DIO_SKIP_HOLES);
2329 extern const struct file_operations generic_ro_fops;
2331 #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
2333 extern int vfs_readlink(struct dentry *, char __user *, int, const char *);
2334 extern int vfs_follow_link(struct nameidata *, const char *);
2335 extern int page_readlink(struct dentry *, char __user *, int);
2336 extern void *page_follow_link_light(struct dentry *, struct nameidata *);
2337 extern void page_put_link(struct dentry *, struct nameidata *, void *);
2338 extern int __page_symlink(struct inode *inode, const char *symname, int len,
2340 extern int page_symlink(struct inode *inode, const char *symname, int len);
2341 extern const struct inode_operations page_symlink_inode_operations;
2342 extern int generic_readlink(struct dentry *, char __user *, int);
2343 extern void generic_fillattr(struct inode *, struct kstat *);
2344 extern int vfs_getattr(struct vfsmount *, struct dentry *, struct kstat *);
2345 void __inode_add_bytes(struct inode *inode, loff_t bytes);
2346 void inode_add_bytes(struct inode *inode, loff_t bytes);
2347 void inode_sub_bytes(struct inode *inode, loff_t bytes);
2348 loff_t inode_get_bytes(struct inode *inode);
2349 void inode_set_bytes(struct inode *inode, loff_t bytes);
2351 extern int vfs_readdir(struct file *, filldir_t, void *);
2353 extern int vfs_stat(const char __user *, struct kstat *);
2354 extern int vfs_lstat(const char __user *, struct kstat *);
2355 extern int vfs_fstat(unsigned int, struct kstat *);
2356 extern int vfs_fstatat(int , const char __user *, struct kstat *, int);
2358 extern int do_vfs_ioctl(struct file *filp, unsigned int fd, unsigned int cmd,
2360 extern int __generic_block_fiemap(struct inode *inode,
2361 struct fiemap_extent_info *fieinfo,
2362 loff_t start, loff_t len,
2363 get_block_t *get_block);
2364 extern int generic_block_fiemap(struct inode *inode,
2365 struct fiemap_extent_info *fieinfo, u64 start,
2366 u64 len, get_block_t *get_block);
2368 extern void get_filesystem(struct file_system_type *fs);
2369 extern void put_filesystem(struct file_system_type *fs);
2370 extern struct file_system_type *get_fs_type(const char *name);
2371 extern struct super_block *get_super(struct block_device *);
2372 extern struct super_block *get_active_super(struct block_device *bdev);
2373 extern struct super_block *user_get_super(dev_t);
2374 extern void drop_super(struct super_block *sb);
2375 extern void iterate_supers(void (*)(struct super_block *, void *), void *);
2377 extern int dcache_dir_open(struct inode *, struct file *);
2378 extern int dcache_dir_close(struct inode *, struct file *);
2379 extern loff_t dcache_dir_lseek(struct file *, loff_t, int);
2380 extern int dcache_readdir(struct file *, void *, filldir_t);
2381 extern int simple_setattr(struct dentry *, struct iattr *);
2382 extern int simple_getattr(struct vfsmount *, struct dentry *, struct kstat *);
2383 extern int simple_statfs(struct dentry *, struct kstatfs *);
2384 extern int simple_link(struct dentry *, struct inode *, struct dentry *);
2385 extern int simple_unlink(struct inode *, struct dentry *);
2386 extern int simple_rmdir(struct inode *, struct dentry *);
2387 extern int simple_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
2388 extern int noop_fsync(struct file *, int);
2389 extern int simple_empty(struct dentry *);
2390 extern int simple_readpage(struct file *file, struct page *page);
2391 extern int simple_write_begin(struct file *file, struct address_space *mapping,
2392 loff_t pos, unsigned len, unsigned flags,
2393 struct page **pagep, void **fsdata);
2394 extern int simple_write_end(struct file *file, struct address_space *mapping,
2395 loff_t pos, unsigned len, unsigned copied,
2396 struct page *page, void *fsdata);
2398 extern struct dentry *simple_lookup(struct inode *, struct dentry *, struct nameidata *);
2399 extern ssize_t generic_read_dir(struct file *, char __user *, size_t, loff_t *);
2400 extern const struct file_operations simple_dir_operations;
2401 extern const struct inode_operations simple_dir_inode_operations;
2402 struct tree_descr { char *name; const struct file_operations *ops; int mode; };
2403 struct dentry *d_alloc_name(struct dentry *, const char *);
2404 extern int simple_fill_super(struct super_block *, unsigned long, struct tree_descr *);
2405 extern int simple_pin_fs(struct file_system_type *, struct vfsmount **mount, int *count);
2406 extern void simple_release_fs(struct vfsmount **mount, int *count);
2408 extern ssize_t simple_read_from_buffer(void __user *to, size_t count,
2409 loff_t *ppos, const void *from, size_t available);
2410 extern ssize_t simple_write_to_buffer(void *to, size_t available, loff_t *ppos,
2411 const void __user *from, size_t count);
2413 extern int generic_file_fsync(struct file *, int);
2415 extern int generic_check_addressable(unsigned, u64);
2417 #ifdef CONFIG_MIGRATION
2418 extern int buffer_migrate_page(struct address_space *,
2419 struct page *, struct page *);
2421 #define buffer_migrate_page NULL
2424 extern int inode_change_ok(const struct inode *, struct iattr *);
2425 extern int inode_newsize_ok(const struct inode *, loff_t offset);
2426 extern void setattr_copy(struct inode *inode, const struct iattr *attr);
2428 extern void file_update_time(struct file *file);
2430 extern int generic_show_options(struct seq_file *m, struct vfsmount *mnt);
2431 extern void save_mount_options(struct super_block *sb, char *options);
2432 extern void replace_mount_options(struct super_block *sb, char *options);
2434 static inline ino_t parent_ino(struct dentry *dentry)
2438 spin_lock(&dentry->d_lock);
2439 res = dentry->d_parent->d_inode->i_ino;
2440 spin_unlock(&dentry->d_lock);
2444 /* Transaction based IO helpers */
2447 * An argresp is stored in an allocated page and holds the
2448 * size of the argument or response, along with its content
2450 struct simple_transaction_argresp {
2455 #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
2457 char *simple_transaction_get(struct file *file, const char __user *buf,
2459 ssize_t simple_transaction_read(struct file *file, char __user *buf,
2460 size_t size, loff_t *pos);
2461 int simple_transaction_release(struct inode *inode, struct file *file);
2463 void simple_transaction_set(struct file *file, size_t n);
2466 * simple attribute files
2468 * These attributes behave similar to those in sysfs:
2470 * Writing to an attribute immediately sets a value, an open file can be
2471 * written to multiple times.
2473 * Reading from an attribute creates a buffer from the value that might get
2474 * read with multiple read calls. When the attribute has been read
2475 * completely, no further read calls are possible until the file is opened
2478 * All attributes contain a text representation of a numeric value
2479 * that are accessed with the get() and set() functions.
2481 #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
2482 static int __fops ## _open(struct inode *inode, struct file *file) \
2484 __simple_attr_check_format(__fmt, 0ull); \
2485 return simple_attr_open(inode, file, __get, __set, __fmt); \
2487 static const struct file_operations __fops = { \
2488 .owner = THIS_MODULE, \
2489 .open = __fops ## _open, \
2490 .release = simple_attr_release, \
2491 .read = simple_attr_read, \
2492 .write = simple_attr_write, \
2493 .llseek = generic_file_llseek, \
2496 static inline void __attribute__((format(printf, 1, 2)))
2497 __simple_attr_check_format(const char *fmt, ...)
2499 /* don't do anything, just let the compiler check the arguments; */
2502 int simple_attr_open(struct inode *inode, struct file *file,
2503 int (*get)(void *, u64 *), int (*set)(void *, u64),
2505 int simple_attr_release(struct inode *inode, struct file *file);
2506 ssize_t simple_attr_read(struct file *file, char __user *buf,
2507 size_t len, loff_t *ppos);
2508 ssize_t simple_attr_write(struct file *file, const char __user *buf,
2509 size_t len, loff_t *ppos);
2512 int proc_nr_files(struct ctl_table *table, int write,
2513 void __user *buffer, size_t *lenp, loff_t *ppos);
2514 int proc_nr_dentry(struct ctl_table *table, int write,
2515 void __user *buffer, size_t *lenp, loff_t *ppos);
2516 int proc_nr_inodes(struct ctl_table *table, int write,
2517 void __user *buffer, size_t *lenp, loff_t *ppos);
2518 int __init get_filesystem_list(char *buf);
2520 #define ACC_MODE(x) ("\004\002\006\006"[(x)&O_ACCMODE])
2521 #define OPEN_FMODE(flag) ((__force fmode_t)(((flag + 1) & O_ACCMODE) | \
2522 (flag & FMODE_NONOTIFY)))
2524 #endif /* __KERNEL__ */
2525 #endif /* _LINUX_FS_H */