5 * This file has definitions for some important file table
9 #include <linux/limits.h>
10 #include <linux/ioctl.h>
13 * It's silly to have NR_OPEN bigger than NR_FILE, but you can change
14 * the file limit at runtime and only root can increase the per-process
15 * nr_file rlimit, so it's safe to set up a ridiculously high absolute
16 * upper limit on files-per-process.
18 * Some programs (notably those using select()) may have to be
19 * recompiled to take full advantage of the new limits..
22 /* Fixed constants first: */
24 #define INR_OPEN 1024 /* Initial setting for nfile rlimits */
26 #define BLOCK_SIZE_BITS 10
27 #define BLOCK_SIZE (1<<BLOCK_SIZE_BITS)
29 #define SEEK_SET 0 /* seek relative to beginning of file */
30 #define SEEK_CUR 1 /* seek relative to current file position */
31 #define SEEK_END 2 /* seek relative to end of file */
32 #define SEEK_MAX SEEK_END
34 /* And dynamically-tunable limits and defaults: */
35 struct files_stat_struct
{
36 int nr_files
; /* read only */
37 int nr_free_files
; /* read only */
38 int max_files
; /* tunable */
41 struct inodes_stat_t
{
44 int dummy
[5]; /* padding for sysctl ABI compatibility */
48 #define NR_FILE 8192 /* this can well be larger on a larger system */
58 * flags in file.f_mode. Note that FMODE_READ and FMODE_WRITE must correspond
59 * to O_WRONLY and O_RDWR via the strange trick in __dentry_open()
62 /* file is open for reading */
63 #define FMODE_READ ((__force fmode_t)1)
64 /* file is open for writing */
65 #define FMODE_WRITE ((__force fmode_t)2)
66 /* file is seekable */
67 #define FMODE_LSEEK ((__force fmode_t)4)
68 /* file can be accessed using pread */
69 #define FMODE_PREAD ((__force fmode_t)8)
70 /* file can be accessed using pwrite */
71 #define FMODE_PWRITE ((__force fmode_t)16)
72 /* File is opened for execution with sys_execve / sys_uselib */
73 #define FMODE_EXEC ((__force fmode_t)32)
74 /* File is opened with O_NDELAY (only set for block devices) */
75 #define FMODE_NDELAY ((__force fmode_t)64)
76 /* File is opened with O_EXCL (only set for block devices) */
77 #define FMODE_EXCL ((__force fmode_t)128)
78 /* File is opened using open(.., 3, ..) and is writeable only for ioctls
79 (specialy hack for floppy.c) */
80 #define FMODE_WRITE_IOCTL ((__force fmode_t)256)
83 * Don't update ctime and mtime.
85 * Currently a special hack for the XFS open_by_handle ioctl, but we'll
86 * hopefully graduate it to a proper O_CMTIME flag supported by open(2) soon.
88 #define FMODE_NOCMTIME ((__force fmode_t)2048)
91 * The below are the various read and write types that we support. Some of
92 * them include behavioral modifiers that send information down to the
93 * block layer and IO scheduler. Terminology:
95 * The block layer uses device plugging to defer IO a little bit, in
96 * the hope that we will see more IO very shortly. This increases
97 * coalescing of adjacent IO and thus reduces the number of IOs we
98 * have to send to the device. It also allows for better queuing,
99 * if the IO isn't mergeable. If the caller is going to be waiting
100 * for the IO, then he must ensure that the device is unplugged so
101 * that the IO is dispatched to the driver.
103 * All IO is handled async in Linux. This is fine for background
104 * writes, but for reads or writes that someone waits for completion
105 * on, we want to notify the block layer and IO scheduler so that they
106 * know about it. That allows them to make better scheduling
107 * decisions. So when the below references 'sync' and 'async', it
108 * is referencing this priority hint.
110 * With that in mind, the available types are:
112 * READ A normal read operation. Device will be plugged.
113 * READ_SYNC A synchronous read. Device is not plugged, caller can
114 * immediately wait on this read without caring about
116 * READA Used for read-ahead operations. Lower priority, and the
117 * block layer could (in theory) choose to ignore this
118 * request if it runs into resource problems.
119 * WRITE A normal async write. Device will be plugged.
120 * SWRITE Like WRITE, but a special case for ll_rw_block() that
121 * tells it to lock the buffer first. Normally a buffer
122 * must be locked before doing IO.
123 * WRITE_SYNC_PLUG Synchronous write. Identical to WRITE, but passes down
124 * the hint that someone will be waiting on this IO
125 * shortly. The device must still be unplugged explicitly,
126 * WRITE_SYNC_PLUG does not do this as we could be
127 * submitting more writes before we actually wait on any
129 * WRITE_SYNC Like WRITE_SYNC_PLUG, but also unplugs the device
130 * immediately after submission. The write equivalent
132 * WRITE_ODIRECT_PLUG Special case write for O_DIRECT only.
134 * SWRITE_SYNC_PLUG Like WRITE_SYNC/WRITE_SYNC_PLUG, but locks the buffer.
136 * WRITE_BARRIER Like WRITE, but tells the block layer that all
137 * previously submitted writes must be safely on storage
138 * before this one is started. Also guarantees that when
139 * this write is complete, it itself is also safely on
140 * storage. Prevents reordering of writes on both sides
148 #define READA 2 /* read-ahead - don't block if no resources */
149 #define SWRITE 3 /* for ll_rw_block() - wait for buffer lock */
150 #define READ_SYNC (READ | (1 << BIO_RW_SYNCIO) | (1 << BIO_RW_UNPLUG))
151 #define READ_META (READ | (1 << BIO_RW_META))
152 #define WRITE_SYNC_PLUG (WRITE | (1 << BIO_RW_SYNCIO) | (1 << BIO_RW_NOIDLE))
153 #define WRITE_SYNC (WRITE_SYNC_PLUG | (1 << BIO_RW_UNPLUG))
154 #define WRITE_ODIRECT_PLUG (WRITE | (1 << BIO_RW_SYNCIO))
155 #define SWRITE_SYNC_PLUG \
156 (SWRITE | (1 << BIO_RW_SYNCIO) | (1 << BIO_RW_NOIDLE))
157 #define SWRITE_SYNC (SWRITE_SYNC_PLUG | (1 << BIO_RW_UNPLUG))
158 #define WRITE_BARRIER (WRITE | (1 << BIO_RW_BARRIER))
161 * These aren't really reads or writes, they pass down information about
162 * parts of device that are now unused by the file system.
164 #define DISCARD_NOBARRIER (WRITE | (1 << BIO_RW_DISCARD))
165 #define DISCARD_BARRIER (DISCARD_NOBARRIER | (1 << BIO_RW_BARRIER))
171 /* public flags for file_system_type */
172 #define FS_REQUIRES_DEV 1
173 #define FS_BINARY_MOUNTDATA 2
174 #define FS_HAS_SUBTYPE 4
175 #define FS_REVAL_DOT 16384 /* Check the paths ".", ".." for staleness */
176 #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move()
177 * during rename() internally.
181 * These are the fs-independent mount-flags: up to 32 flags are supported
183 #define MS_RDONLY 1 /* Mount read-only */
184 #define MS_NOSUID 2 /* Ignore suid and sgid bits */
185 #define MS_NODEV 4 /* Disallow access to device special files */
186 #define MS_NOEXEC 8 /* Disallow program execution */
187 #define MS_SYNCHRONOUS 16 /* Writes are synced at once */
188 #define MS_REMOUNT 32 /* Alter flags of a mounted FS */
189 #define MS_MANDLOCK 64 /* Allow mandatory locks on an FS */
190 #define MS_DIRSYNC 128 /* Directory modifications are synchronous */
191 #define MS_NOATIME 1024 /* Do not update access times. */
192 #define MS_NODIRATIME 2048 /* Do not update directory access times */
196 #define MS_VERBOSE 32768 /* War is peace. Verbosity is silence.
197 MS_VERBOSE is deprecated. */
198 #define MS_SILENT 32768
199 #define MS_POSIXACL (1<<16) /* VFS does not apply the umask */
200 #define MS_UNBINDABLE (1<<17) /* change to unbindable */
201 #define MS_PRIVATE (1<<18) /* change to private */
202 #define MS_SLAVE (1<<19) /* change to slave */
203 #define MS_SHARED (1<<20) /* change to shared */
204 #define MS_RELATIME (1<<21) /* Update atime relative to mtime/ctime. */
205 #define MS_KERNMOUNT (1<<22) /* this is a kern_mount call */
206 #define MS_I_VERSION (1<<23) /* Update inode I_version field */
207 #define MS_STRICTATIME (1<<24) /* Always perform atime updates */
208 #define MS_ACTIVE (1<<30)
209 #define MS_NOUSER (1<<31)
212 * Superblock flags that can be altered by MS_REMOUNT
214 #define MS_RMT_MASK (MS_RDONLY|MS_SYNCHRONOUS|MS_MANDLOCK|MS_I_VERSION)
217 * Old magic mount flag and mask
219 #define MS_MGC_VAL 0xC0ED0000
220 #define MS_MGC_MSK 0xffff0000
222 /* Inode flags - they have nothing to superblock flags now */
224 #define S_SYNC 1 /* Writes are synced at once */
225 #define S_NOATIME 2 /* Do not update access times */
226 #define S_APPEND 4 /* Append-only file */
227 #define S_IMMUTABLE 8 /* Immutable file */
228 #define S_DEAD 16 /* removed, but still open directory */
229 #define S_NOQUOTA 32 /* Inode is not counted to quota */
230 #define S_DIRSYNC 64 /* Directory modifications are synchronous */
231 #define S_NOCMTIME 128 /* Do not update file c/mtime */
232 #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
233 #define S_PRIVATE 512 /* Inode is fs-internal */
236 * Note that nosuid etc flags are inode-specific: setting some file-system
237 * flags just means all the inodes inherit those flags by default. It might be
238 * possible to override it selectively if you really wanted to with some
239 * ioctl() that is not currently implemented.
241 * Exception: MS_RDONLY is always applied to the entire file system.
243 * Unfortunately, it is possible to change a filesystems flags with it mounted
244 * with files in use. This means that all of the inodes will not have their
245 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
246 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
248 #define __IS_FLG(inode,flg) ((inode)->i_sb->s_flags & (flg))
250 #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
251 #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
252 ((inode)->i_flags & S_SYNC))
253 #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
254 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
255 #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
256 #define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
257 #define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
259 #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
260 #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
261 #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
262 #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
264 #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
265 #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
266 #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
267 #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
269 /* the read-only stuff doesn't really belong here, but any other place is
270 probably as bad and I don't want to create yet another include file. */
272 #define BLKROSET _IO(0x12,93) /* set device read-only (0 = read-write) */
273 #define BLKROGET _IO(0x12,94) /* get read-only status (0 = read_write) */
274 #define BLKRRPART _IO(0x12,95) /* re-read partition table */
275 #define BLKGETSIZE _IO(0x12,96) /* return device size /512 (long *arg) */
276 #define BLKFLSBUF _IO(0x12,97) /* flush buffer cache */
277 #define BLKRASET _IO(0x12,98) /* set read ahead for block device */
278 #define BLKRAGET _IO(0x12,99) /* get current read ahead setting */
279 #define BLKFRASET _IO(0x12,100)/* set filesystem (mm/filemap.c) read-ahead */
280 #define BLKFRAGET _IO(0x12,101)/* get filesystem (mm/filemap.c) read-ahead */
281 #define BLKSECTSET _IO(0x12,102)/* set max sectors per request (ll_rw_blk.c) */
282 #define BLKSECTGET _IO(0x12,103)/* get max sectors per request (ll_rw_blk.c) */
283 #define BLKSSZGET _IO(0x12,104)/* get block device sector size */
285 #define BLKPG _IO(0x12,105)/* See blkpg.h */
287 /* Some people are morons. Do not use sizeof! */
289 #define BLKELVGET _IOR(0x12,106,size_t)/* elevator get */
290 #define BLKELVSET _IOW(0x12,107,size_t)/* elevator set */
291 /* This was here just to show that the number is taken -
292 probably all these _IO(0x12,*) ioctls should be moved to blkpg.h. */
294 /* A jump here: 108-111 have been used for various private purposes. */
295 #define BLKBSZGET _IOR(0x12,112,size_t)
296 #define BLKBSZSET _IOW(0x12,113,size_t)
297 #define BLKGETSIZE64 _IOR(0x12,114,size_t) /* return device size in bytes (u64 *arg) */
298 #define BLKTRACESETUP _IOWR(0x12,115,struct blk_user_trace_setup)
299 #define BLKTRACESTART _IO(0x12,116)
300 #define BLKTRACESTOP _IO(0x12,117)
301 #define BLKTRACETEARDOWN _IO(0x12,118)
302 #define BLKDISCARD _IO(0x12,119)
303 #define BLKIOMIN _IO(0x12,120)
304 #define BLKIOOPT _IO(0x12,121)
305 #define BLKALIGNOFF _IO(0x12,122)
306 #define BLKPBSZGET _IO(0x12,123)
307 #define BLKDISCARDZEROES _IO(0x12,124)
309 #define BMAP_IOCTL 1 /* obsolete - kept for compatibility */
310 #define FIBMAP _IO(0x00,1) /* bmap access */
311 #define FIGETBSZ _IO(0x00,2) /* get the block size used for bmap */
312 #define FIFREEZE _IOWR('X', 119, int) /* Freeze */
313 #define FITHAW _IOWR('X', 120, int) /* Thaw */
315 #define FS_IOC_GETFLAGS _IOR('f', 1, long)
316 #define FS_IOC_SETFLAGS _IOW('f', 2, long)
317 #define FS_IOC_GETVERSION _IOR('v', 1, long)
318 #define FS_IOC_SETVERSION _IOW('v', 2, long)
319 #define FS_IOC_FIEMAP _IOWR('f', 11, struct fiemap)
320 #define FS_IOC32_GETFLAGS _IOR('f', 1, int)
321 #define FS_IOC32_SETFLAGS _IOW('f', 2, int)
322 #define FS_IOC32_GETVERSION _IOR('v', 1, int)
323 #define FS_IOC32_SETVERSION _IOW('v', 2, int)
326 * Inode flags (FS_IOC_GETFLAGS / FS_IOC_SETFLAGS)
328 #define FS_SECRM_FL 0x00000001 /* Secure deletion */
329 #define FS_UNRM_FL 0x00000002 /* Undelete */
330 #define FS_COMPR_FL 0x00000004 /* Compress file */
331 #define FS_SYNC_FL 0x00000008 /* Synchronous updates */
332 #define FS_IMMUTABLE_FL 0x00000010 /* Immutable file */
333 #define FS_APPEND_FL 0x00000020 /* writes to file may only append */
334 #define FS_NODUMP_FL 0x00000040 /* do not dump file */
335 #define FS_NOATIME_FL 0x00000080 /* do not update atime */
336 /* Reserved for compression usage... */
337 #define FS_DIRTY_FL 0x00000100
338 #define FS_COMPRBLK_FL 0x00000200 /* One or more compressed clusters */
339 #define FS_NOCOMP_FL 0x00000400 /* Don't compress */
340 #define FS_ECOMPR_FL 0x00000800 /* Compression error */
341 /* End compression flags --- maybe not all used */
342 #define FS_BTREE_FL 0x00001000 /* btree format dir */
343 #define FS_INDEX_FL 0x00001000 /* hash-indexed directory */
344 #define FS_IMAGIC_FL 0x00002000 /* AFS directory */
345 #define FS_JOURNAL_DATA_FL 0x00004000 /* Reserved for ext3 */
346 #define FS_NOTAIL_FL 0x00008000 /* file tail should not be merged */
347 #define FS_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */
348 #define FS_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/
349 #define FS_EXTENT_FL 0x00080000 /* Extents */
350 #define FS_DIRECTIO_FL 0x00100000 /* Use direct i/o */
351 #define FS_RESERVED_FL 0x80000000 /* reserved for ext2 lib */
353 #define FS_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */
354 #define FS_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */
357 #define SYNC_FILE_RANGE_WAIT_BEFORE 1
358 #define SYNC_FILE_RANGE_WRITE 2
359 #define SYNC_FILE_RANGE_WAIT_AFTER 4
363 #include <linux/linkage.h>
364 #include <linux/wait.h>
365 #include <linux/types.h>
366 #include <linux/kdev_t.h>
367 #include <linux/dcache.h>
368 #include <linux/path.h>
369 #include <linux/stat.h>
370 #include <linux/cache.h>
371 #include <linux/kobject.h>
372 #include <linux/list.h>
373 #include <linux/radix-tree.h>
374 #include <linux/prio_tree.h>
375 #include <linux/init.h>
376 #include <linux/pid.h>
377 #include <linux/mutex.h>
378 #include <linux/capability.h>
379 #include <linux/semaphore.h>
380 #include <linux/fiemap.h>
382 #include <asm/atomic.h>
383 #include <asm/byteorder.h>
385 struct export_operations
;
390 struct pipe_inode_info
;
391 struct poll_table_struct
;
393 struct vm_area_struct
;
397 extern void __init
inode_init(void);
398 extern void __init
inode_init_early(void);
399 extern void __init
files_init(unsigned long);
401 extern struct files_stat_struct files_stat
;
402 extern int get_max_files(void);
403 extern int sysctl_nr_open
;
404 extern struct inodes_stat_t inodes_stat
;
405 extern int leases_enable
, lease_break_time
;
406 #ifdef CONFIG_DNOTIFY
407 extern int dir_notify_enable
;
411 typedef int (get_block_t
)(struct inode
*inode
, sector_t iblock
,
412 struct buffer_head
*bh_result
, int create
);
413 typedef void (dio_iodone_t
)(struct kiocb
*iocb
, loff_t offset
,
414 ssize_t bytes
, void *private);
417 * Attribute flags. These should be or-ed together to figure out what
420 #define ATTR_MODE (1 << 0)
421 #define ATTR_UID (1 << 1)
422 #define ATTR_GID (1 << 2)
423 #define ATTR_SIZE (1 << 3)
424 #define ATTR_ATIME (1 << 4)
425 #define ATTR_MTIME (1 << 5)
426 #define ATTR_CTIME (1 << 6)
427 #define ATTR_ATIME_SET (1 << 7)
428 #define ATTR_MTIME_SET (1 << 8)
429 #define ATTR_FORCE (1 << 9) /* Not a change, but a change it */
430 #define ATTR_ATTR_FLAG (1 << 10)
431 #define ATTR_KILL_SUID (1 << 11)
432 #define ATTR_KILL_SGID (1 << 12)
433 #define ATTR_FILE (1 << 13)
434 #define ATTR_KILL_PRIV (1 << 14)
435 #define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */
436 #define ATTR_TIMES_SET (1 << 16)
439 * This is the Inode Attributes structure, used for notify_change(). It
440 * uses the above definitions as flags, to know which values have changed.
441 * Also, in this manner, a Filesystem can look at only the values it cares
442 * about. Basically, these are the attributes that the VFS layer can
443 * request to change from the FS layer.
445 * Derek Atkins <warlord@MIT.EDU> 94-10-20
448 unsigned int ia_valid
;
453 struct timespec ia_atime
;
454 struct timespec ia_mtime
;
455 struct timespec ia_ctime
;
458 * Not an attribute, but an auxilary info for filesystems wanting to
459 * implement an ftruncate() like method. NOTE: filesystem should
460 * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
462 struct file
*ia_file
;
466 * Includes for diskquotas.
468 #include <linux/quota.h>
471 * enum positive_aop_returns - aop return codes with specific semantics
473 * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
474 * completed, that the page is still locked, and
475 * should be considered active. The VM uses this hint
476 * to return the page to the active list -- it won't
477 * be a candidate for writeback again in the near
478 * future. Other callers must be careful to unlock
479 * the page if they get this return. Returned by
482 * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
483 * unlocked it and the page might have been truncated.
484 * The caller should back up to acquiring a new page and
485 * trying again. The aop will be taking reasonable
486 * precautions not to livelock. If the caller held a page
487 * reference, it should drop it before retrying. Returned
490 * address_space_operation functions return these large constants to indicate
491 * special semantics to the caller. These are much larger than the bytes in a
492 * page to allow for functions that return the number of bytes operated on in a
496 enum positive_aop_returns
{
497 AOP_WRITEPAGE_ACTIVATE
= 0x80000,
498 AOP_TRUNCATED_PAGE
= 0x80001,
501 #define AOP_FLAG_UNINTERRUPTIBLE 0x0001 /* will not do a short write */
502 #define AOP_FLAG_CONT_EXPAND 0x0002 /* called from cont_expand */
503 #define AOP_FLAG_NOFS 0x0004 /* used by filesystem to direct
504 * helper code (eg buffer layer)
505 * to clear GFP_FS from alloc */
508 * oh the beauties of C type declarations.
511 struct address_space
;
512 struct writeback_control
;
515 const struct iovec
*iov
;
516 unsigned long nr_segs
;
521 size_t iov_iter_copy_from_user_atomic(struct page
*page
,
522 struct iov_iter
*i
, unsigned long offset
, size_t bytes
);
523 size_t iov_iter_copy_from_user(struct page
*page
,
524 struct iov_iter
*i
, unsigned long offset
, size_t bytes
);
525 void iov_iter_advance(struct iov_iter
*i
, size_t bytes
);
526 int iov_iter_fault_in_readable(struct iov_iter
*i
, size_t bytes
);
527 size_t iov_iter_single_seg_count(struct iov_iter
*i
);
529 static inline void iov_iter_init(struct iov_iter
*i
,
530 const struct iovec
*iov
, unsigned long nr_segs
,
531 size_t count
, size_t written
)
534 i
->nr_segs
= nr_segs
;
536 i
->count
= count
+ written
;
538 iov_iter_advance(i
, written
);
541 static inline size_t iov_iter_count(struct iov_iter
*i
)
547 * "descriptor" for what we're up to with a read.
548 * This allows us to use the same read code yet
549 * have multiple different users of the data that
550 * we read from a file.
552 * The simplest case just copies the data to user
565 typedef int (*read_actor_t
)(read_descriptor_t
*, struct page
*,
566 unsigned long, unsigned long);
568 struct address_space_operations
{
569 int (*writepage
)(struct page
*page
, struct writeback_control
*wbc
);
570 int (*readpage
)(struct file
*, struct page
*);
571 void (*sync_page
)(struct page
*);
573 /* Write back some dirty pages from this mapping. */
574 int (*writepages
)(struct address_space
*, struct writeback_control
*);
576 /* Set a page dirty. Return true if this dirtied it */
577 int (*set_page_dirty
)(struct page
*page
);
579 int (*readpages
)(struct file
*filp
, struct address_space
*mapping
,
580 struct list_head
*pages
, unsigned nr_pages
);
582 int (*write_begin
)(struct file
*, struct address_space
*mapping
,
583 loff_t pos
, unsigned len
, unsigned flags
,
584 struct page
**pagep
, void **fsdata
);
585 int (*write_end
)(struct file
*, struct address_space
*mapping
,
586 loff_t pos
, unsigned len
, unsigned copied
,
587 struct page
*page
, void *fsdata
);
589 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
590 sector_t (*bmap
)(struct address_space
*, sector_t
);
591 void (*invalidatepage
) (struct page
*, unsigned long);
592 int (*releasepage
) (struct page
*, gfp_t
);
593 ssize_t (*direct_IO
)(int, struct kiocb
*, const struct iovec
*iov
,
594 loff_t offset
, unsigned long nr_segs
);
595 int (*get_xip_mem
)(struct address_space
*, pgoff_t
, int,
596 void **, unsigned long *);
597 /* migrate the contents of a page to the specified target */
598 int (*migratepage
) (struct address_space
*,
599 struct page
*, struct page
*);
600 int (*launder_page
) (struct page
*);
601 int (*is_partially_uptodate
) (struct page
*, read_descriptor_t
*,
603 int (*error_remove_page
)(struct address_space
*, struct page
*);
607 * pagecache_write_begin/pagecache_write_end must be used by general code
608 * to write into the pagecache.
610 int pagecache_write_begin(struct file
*, struct address_space
*mapping
,
611 loff_t pos
, unsigned len
, unsigned flags
,
612 struct page
**pagep
, void **fsdata
);
614 int pagecache_write_end(struct file
*, struct address_space
*mapping
,
615 loff_t pos
, unsigned len
, unsigned copied
,
616 struct page
*page
, void *fsdata
);
618 struct backing_dev_info
;
619 struct address_space
{
620 struct inode
*host
; /* owner: inode, block_device */
621 struct radix_tree_root page_tree
; /* radix tree of all pages */
622 spinlock_t tree_lock
; /* and lock protecting it */
623 unsigned int i_mmap_writable
;/* count VM_SHARED mappings */
624 struct prio_tree_root i_mmap
; /* tree of private and shared mappings */
625 struct list_head i_mmap_nonlinear
;/*list VM_NONLINEAR mappings */
626 spinlock_t i_mmap_lock
; /* protect tree, count, list */
627 unsigned int truncate_count
; /* Cover race condition with truncate */
628 unsigned long nrpages
; /* number of total pages */
629 pgoff_t writeback_index
;/* writeback starts here */
630 const struct address_space_operations
*a_ops
; /* methods */
631 unsigned long flags
; /* error bits/gfp mask */
632 struct backing_dev_info
*backing_dev_info
; /* device readahead, etc */
633 spinlock_t private_lock
; /* for use by the address_space */
634 struct list_head private_list
; /* ditto */
635 struct address_space
*assoc_mapping
; /* ditto */
636 } __attribute__((aligned(sizeof(long))));
638 * On most architectures that alignment is already the case; but
639 * must be enforced here for CRIS, to let the least signficant bit
640 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
643 struct block_device
{
644 dev_t bd_dev
; /* not a kdev_t - it's a search key */
645 struct inode
* bd_inode
; /* will die */
646 struct super_block
* bd_super
;
648 struct mutex bd_mutex
; /* open/close mutex */
649 struct list_head bd_inodes
;
653 struct list_head bd_holder_list
;
655 struct block_device
* bd_contains
;
656 unsigned bd_block_size
;
657 struct hd_struct
* bd_part
;
658 /* number of times partitions within this device have been opened. */
659 unsigned bd_part_count
;
661 struct gendisk
* bd_disk
;
662 struct list_head bd_list
;
664 * Private data. You must have bd_claim'ed the block_device
665 * to use this. NOTE: bd_claim allows an owner to claim
666 * the same device multiple times, the owner must take special
667 * care to not mess up bd_private for that case.
669 unsigned long bd_private
;
671 /* The counter of freeze processes */
672 int bd_fsfreeze_count
;
673 /* Mutex for freeze */
674 struct mutex bd_fsfreeze_mutex
;
678 * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
681 #define PAGECACHE_TAG_DIRTY 0
682 #define PAGECACHE_TAG_WRITEBACK 1
684 int mapping_tagged(struct address_space
*mapping
, int tag
);
687 * Might pages of this file be mapped into userspace?
689 static inline int mapping_mapped(struct address_space
*mapping
)
691 return !prio_tree_empty(&mapping
->i_mmap
) ||
692 !list_empty(&mapping
->i_mmap_nonlinear
);
696 * Might pages of this file have been modified in userspace?
697 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
698 * marks vma as VM_SHARED if it is shared, and the file was opened for
699 * writing i.e. vma may be mprotected writable even if now readonly.
701 static inline int mapping_writably_mapped(struct address_space
*mapping
)
703 return mapping
->i_mmap_writable
!= 0;
707 * Use sequence counter to get consistent i_size on 32-bit processors.
709 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
710 #include <linux/seqlock.h>
711 #define __NEED_I_SIZE_ORDERED
712 #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
714 #define i_size_ordered_init(inode) do { } while (0)
718 #define ACL_NOT_CACHED ((void *)(-1))
721 struct hlist_node i_hash
;
722 struct list_head i_list
; /* backing dev IO list */
723 struct list_head i_sb_list
;
724 struct list_head i_dentry
;
727 unsigned int i_nlink
;
733 #ifdef __NEED_I_SIZE_ORDERED
734 seqcount_t i_size_seqcount
;
736 struct timespec i_atime
;
737 struct timespec i_mtime
;
738 struct timespec i_ctime
;
740 unsigned int i_blkbits
;
741 unsigned short i_bytes
;
743 spinlock_t i_lock
; /* i_blocks, i_bytes, maybe i_size */
744 struct mutex i_mutex
;
745 struct rw_semaphore i_alloc_sem
;
746 const struct inode_operations
*i_op
;
747 const struct file_operations
*i_fop
; /* former ->i_op->default_file_ops */
748 struct super_block
*i_sb
;
749 struct file_lock
*i_flock
;
750 struct address_space
*i_mapping
;
751 struct address_space i_data
;
753 struct dquot
*i_dquot
[MAXQUOTAS
];
755 struct list_head i_devices
;
757 struct pipe_inode_info
*i_pipe
;
758 struct block_device
*i_bdev
;
764 #ifdef CONFIG_FSNOTIFY
765 __u32 i_fsnotify_mask
; /* all events this inode cares about */
766 struct hlist_head i_fsnotify_mark_entries
; /* fsnotify mark entries */
769 #ifdef CONFIG_INOTIFY
770 struct list_head inotify_watches
; /* watches on this inode */
771 struct mutex inotify_mutex
; /* protects the watches list */
774 unsigned long i_state
;
775 unsigned long dirtied_when
; /* jiffies of first dirtying */
777 unsigned int i_flags
;
779 atomic_t i_writecount
;
780 #ifdef CONFIG_SECURITY
783 #ifdef CONFIG_FS_POSIX_ACL
784 struct posix_acl
*i_acl
;
785 struct posix_acl
*i_default_acl
;
787 void *i_private
; /* fs or device private pointer */
791 * inode->i_mutex nesting subclasses for the lock validator:
793 * 0: the object of the current VFS operation
798 * The locking order between these classes is
799 * parent -> child -> normal -> xattr -> quota
801 enum inode_i_mutex_lock_class
811 * NOTE: in a 32bit arch with a preemptable kernel and
812 * an UP compile the i_size_read/write must be atomic
813 * with respect to the local cpu (unlike with preempt disabled),
814 * but they don't need to be atomic with respect to other cpus like in
815 * true SMP (so they need either to either locally disable irq around
816 * the read or for example on x86 they can be still implemented as a
817 * cmpxchg8b without the need of the lock prefix). For SMP compiles
818 * and 64bit archs it makes no difference if preempt is enabled or not.
820 static inline loff_t
i_size_read(const struct inode
*inode
)
822 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
827 seq
= read_seqcount_begin(&inode
->i_size_seqcount
);
828 i_size
= inode
->i_size
;
829 } while (read_seqcount_retry(&inode
->i_size_seqcount
, seq
));
831 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
835 i_size
= inode
->i_size
;
839 return inode
->i_size
;
844 * NOTE: unlike i_size_read(), i_size_write() does need locking around it
845 * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
846 * can be lost, resulting in subsequent i_size_read() calls spinning forever.
848 static inline void i_size_write(struct inode
*inode
, loff_t i_size
)
850 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
851 write_seqcount_begin(&inode
->i_size_seqcount
);
852 inode
->i_size
= i_size
;
853 write_seqcount_end(&inode
->i_size_seqcount
);
854 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
856 inode
->i_size
= i_size
;
859 inode
->i_size
= i_size
;
863 static inline unsigned iminor(const struct inode
*inode
)
865 return MINOR(inode
->i_rdev
);
868 static inline unsigned imajor(const struct inode
*inode
)
870 return MAJOR(inode
->i_rdev
);
873 extern struct block_device
*I_BDEV(struct inode
*inode
);
876 rwlock_t lock
; /* protects pid, uid, euid fields */
877 struct pid
*pid
; /* pid or -pgrp where SIGIO should be sent */
878 enum pid_type pid_type
; /* Kind of process group SIGIO should be sent to */
879 uid_t uid
, euid
; /* uid/euid of process setting the owner */
880 int signum
; /* posix.1b rt signal to be delivered on IO */
884 * Track a single file's readahead state
886 struct file_ra_state
{
887 pgoff_t start
; /* where readahead started */
888 unsigned int size
; /* # of readahead pages */
889 unsigned int async_size
; /* do asynchronous readahead when
890 there are only # of pages ahead */
892 unsigned int ra_pages
; /* Maximum readahead window */
893 unsigned int mmap_miss
; /* Cache miss stat for mmap accesses */
894 loff_t prev_pos
; /* Cache last read() position */
898 * Check if @index falls in the readahead windows.
900 static inline int ra_has_index(struct file_ra_state
*ra
, pgoff_t index
)
902 return (index
>= ra
->start
&&
903 index
< ra
->start
+ ra
->size
);
906 #define FILE_MNT_WRITE_TAKEN 1
907 #define FILE_MNT_WRITE_RELEASED 2
911 * fu_list becomes invalid after file_free is called and queued via
912 * fu_rcuhead for RCU freeing
915 struct list_head fu_list
;
916 struct rcu_head fu_rcuhead
;
919 #define f_dentry f_path.dentry
920 #define f_vfsmnt f_path.mnt
921 const struct file_operations
*f_op
;
922 spinlock_t f_lock
; /* f_ep_links, f_flags, no IRQ */
923 atomic_long_t f_count
;
924 unsigned int f_flags
;
927 struct fown_struct f_owner
;
928 const struct cred
*f_cred
;
929 struct file_ra_state f_ra
;
932 #ifdef CONFIG_SECURITY
935 /* needed for tty driver, and maybe others */
939 /* Used by fs/eventpoll.c to link all the hooks to this file */
940 struct list_head f_ep_links
;
941 #endif /* #ifdef CONFIG_EPOLL */
942 struct address_space
*f_mapping
;
943 #ifdef CONFIG_DEBUG_WRITECOUNT
944 unsigned long f_mnt_write_state
;
947 extern spinlock_t files_lock
;
948 #define file_list_lock() spin_lock(&files_lock);
949 #define file_list_unlock() spin_unlock(&files_lock);
951 #define get_file(x) atomic_long_inc(&(x)->f_count)
952 #define file_count(x) atomic_long_read(&(x)->f_count)
954 #ifdef CONFIG_DEBUG_WRITECOUNT
955 static inline void file_take_write(struct file
*f
)
957 WARN_ON(f
->f_mnt_write_state
!= 0);
958 f
->f_mnt_write_state
= FILE_MNT_WRITE_TAKEN
;
960 static inline void file_release_write(struct file
*f
)
962 f
->f_mnt_write_state
|= FILE_MNT_WRITE_RELEASED
;
964 static inline void file_reset_write(struct file
*f
)
966 f
->f_mnt_write_state
= 0;
968 static inline void file_check_state(struct file
*f
)
971 * At this point, either both or neither of these bits
974 WARN_ON(f
->f_mnt_write_state
== FILE_MNT_WRITE_TAKEN
);
975 WARN_ON(f
->f_mnt_write_state
== FILE_MNT_WRITE_RELEASED
);
977 static inline int file_check_writeable(struct file
*f
)
979 if (f
->f_mnt_write_state
== FILE_MNT_WRITE_TAKEN
)
981 printk(KERN_WARNING
"writeable file with no "
982 "mnt_want_write()\n");
986 #else /* !CONFIG_DEBUG_WRITECOUNT */
987 static inline void file_take_write(struct file
*filp
) {}
988 static inline void file_release_write(struct file
*filp
) {}
989 static inline void file_reset_write(struct file
*filp
) {}
990 static inline void file_check_state(struct file
*filp
) {}
991 static inline int file_check_writeable(struct file
*filp
)
995 #endif /* CONFIG_DEBUG_WRITECOUNT */
997 #define MAX_NON_LFS ((1UL<<31) - 1)
999 /* Page cache limit. The filesystems should put that into their s_maxbytes
1000 limits, otherwise bad things can happen in VM. */
1001 #if BITS_PER_LONG==32
1002 #define MAX_LFS_FILESIZE (((u64)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
1003 #elif BITS_PER_LONG==64
1004 #define MAX_LFS_FILESIZE 0x7fffffffffffffffUL
1009 #define FL_ACCESS 8 /* not trying to lock, just looking */
1010 #define FL_EXISTS 16 /* when unlocking, test for existence */
1011 #define FL_LEASE 32 /* lease held on this file */
1012 #define FL_CLOSE 64 /* unlock on close */
1013 #define FL_SLEEP 128 /* A blocking lock */
1016 * Special return value from posix_lock_file() and vfs_lock_file() for
1017 * asynchronous locking.
1019 #define FILE_LOCK_DEFERRED 1
1022 * The POSIX file lock owner is determined by
1023 * the "struct files_struct" in the thread group
1024 * (or NULL for no owner - BSD locks).
1026 * Lockd stuffs a "host" pointer into this.
1028 typedef struct files_struct
*fl_owner_t
;
1030 struct file_lock_operations
{
1031 void (*fl_copy_lock
)(struct file_lock
*, struct file_lock
*);
1032 void (*fl_release_private
)(struct file_lock
*);
1035 struct lock_manager_operations
{
1036 int (*fl_compare_owner
)(struct file_lock
*, struct file_lock
*);
1037 void (*fl_notify
)(struct file_lock
*); /* unblock callback */
1038 int (*fl_grant
)(struct file_lock
*, struct file_lock
*, int);
1039 void (*fl_copy_lock
)(struct file_lock
*, struct file_lock
*);
1040 void (*fl_release_private
)(struct file_lock
*);
1041 void (*fl_break
)(struct file_lock
*);
1042 int (*fl_mylease
)(struct file_lock
*, struct file_lock
*);
1043 int (*fl_change
)(struct file_lock
**, int);
1046 struct lock_manager
{
1047 struct list_head list
;
1050 void locks_start_grace(struct lock_manager
*);
1051 void locks_end_grace(struct lock_manager
*);
1052 int locks_in_grace(void);
1054 /* that will die - we need it for nfs_lock_info */
1055 #include <linux/nfs_fs_i.h>
1058 struct file_lock
*fl_next
; /* singly linked list for this inode */
1059 struct list_head fl_link
; /* doubly linked list of all locks */
1060 struct list_head fl_block
; /* circular list of blocked processes */
1061 fl_owner_t fl_owner
;
1062 unsigned char fl_flags
;
1063 unsigned char fl_type
;
1064 unsigned int fl_pid
;
1065 struct pid
*fl_nspid
;
1066 wait_queue_head_t fl_wait
;
1067 struct file
*fl_file
;
1071 struct fasync_struct
* fl_fasync
; /* for lease break notifications */
1072 unsigned long fl_break_time
; /* for nonblocking lease breaks */
1074 const struct file_lock_operations
*fl_ops
; /* Callbacks for filesystems */
1075 const struct lock_manager_operations
*fl_lmops
; /* Callbacks for lockmanagers */
1077 struct nfs_lock_info nfs_fl
;
1078 struct nfs4_lock_info nfs4_fl
;
1080 struct list_head link
; /* link in AFS vnode's pending_locks list */
1081 int state
; /* state of grant or error if -ve */
1086 /* The following constant reflects the upper bound of the file/locking space */
1088 #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
1089 #define OFFSET_MAX INT_LIMIT(loff_t)
1090 #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
1093 #include <linux/fcntl.h>
1095 extern void send_sigio(struct fown_struct
*fown
, int fd
, int band
);
1098 extern int do_sync_mapping_range(struct address_space
*mapping
, loff_t offset
,
1099 loff_t endbyte
, unsigned int flags
);
1101 #ifdef CONFIG_FILE_LOCKING
1102 extern int fcntl_getlk(struct file
*, struct flock __user
*);
1103 extern int fcntl_setlk(unsigned int, struct file
*, unsigned int,
1104 struct flock __user
*);
1106 #if BITS_PER_LONG == 32
1107 extern int fcntl_getlk64(struct file
*, struct flock64 __user
*);
1108 extern int fcntl_setlk64(unsigned int, struct file
*, unsigned int,
1109 struct flock64 __user
*);
1112 extern int fcntl_setlease(unsigned int fd
, struct file
*filp
, long arg
);
1113 extern int fcntl_getlease(struct file
*filp
);
1116 extern void locks_init_lock(struct file_lock
*);
1117 extern void locks_copy_lock(struct file_lock
*, struct file_lock
*);
1118 extern void __locks_copy_lock(struct file_lock
*, const struct file_lock
*);
1119 extern void locks_remove_posix(struct file
*, fl_owner_t
);
1120 extern void locks_remove_flock(struct file
*);
1121 extern void locks_release_private(struct file_lock
*);
1122 extern void posix_test_lock(struct file
*, struct file_lock
*);
1123 extern int posix_lock_file(struct file
*, struct file_lock
*, struct file_lock
*);
1124 extern int posix_lock_file_wait(struct file
*, struct file_lock
*);
1125 extern int posix_unblock_lock(struct file
*, struct file_lock
*);
1126 extern int vfs_test_lock(struct file
*, struct file_lock
*);
1127 extern int vfs_lock_file(struct file
*, unsigned int, struct file_lock
*, struct file_lock
*);
1128 extern int vfs_cancel_lock(struct file
*filp
, struct file_lock
*fl
);
1129 extern int flock_lock_file_wait(struct file
*filp
, struct file_lock
*fl
);
1130 extern int __break_lease(struct inode
*inode
, unsigned int flags
);
1131 extern void lease_get_mtime(struct inode
*, struct timespec
*time
);
1132 extern int generic_setlease(struct file
*, long, struct file_lock
**);
1133 extern int vfs_setlease(struct file
*, long, struct file_lock
**);
1134 extern int lease_modify(struct file_lock
**, int);
1135 extern int lock_may_read(struct inode
*, loff_t start
, unsigned long count
);
1136 extern int lock_may_write(struct inode
*, loff_t start
, unsigned long count
);
1137 #else /* !CONFIG_FILE_LOCKING */
1138 static inline int fcntl_getlk(struct file
*file
, struct flock __user
*user
)
1143 static inline int fcntl_setlk(unsigned int fd
, struct file
*file
,
1144 unsigned int cmd
, struct flock __user
*user
)
1149 #if BITS_PER_LONG == 32
1150 static inline int fcntl_getlk64(struct file
*file
, struct flock64 __user
*user
)
1155 static inline int fcntl_setlk64(unsigned int fd
, struct file
*file
,
1156 unsigned int cmd
, struct flock64 __user
*user
)
1161 static inline int fcntl_setlease(unsigned int fd
, struct file
*filp
, long arg
)
1166 static inline int fcntl_getlease(struct file
*filp
)
1171 static inline void locks_init_lock(struct file_lock
*fl
)
1176 static inline void __locks_copy_lock(struct file_lock
*new, struct file_lock
*fl
)
1181 static inline void locks_copy_lock(struct file_lock
*new, struct file_lock
*fl
)
1186 static inline void locks_remove_posix(struct file
*filp
, fl_owner_t owner
)
1191 static inline void locks_remove_flock(struct file
*filp
)
1196 static inline void posix_test_lock(struct file
*filp
, struct file_lock
*fl
)
1201 static inline int posix_lock_file(struct file
*filp
, struct file_lock
*fl
,
1202 struct file_lock
*conflock
)
1207 static inline int posix_lock_file_wait(struct file
*filp
, struct file_lock
*fl
)
1212 static inline int posix_unblock_lock(struct file
*filp
,
1213 struct file_lock
*waiter
)
1218 static inline int vfs_test_lock(struct file
*filp
, struct file_lock
*fl
)
1223 static inline int vfs_lock_file(struct file
*filp
, unsigned int cmd
,
1224 struct file_lock
*fl
, struct file_lock
*conf
)
1229 static inline int vfs_cancel_lock(struct file
*filp
, struct file_lock
*fl
)
1234 static inline int flock_lock_file_wait(struct file
*filp
,
1235 struct file_lock
*request
)
1240 static inline int __break_lease(struct inode
*inode
, unsigned int mode
)
1245 static inline void lease_get_mtime(struct inode
*inode
, struct timespec
*time
)
1250 static inline int generic_setlease(struct file
*filp
, long arg
,
1251 struct file_lock
**flp
)
1256 static inline int vfs_setlease(struct file
*filp
, long arg
,
1257 struct file_lock
**lease
)
1262 static inline int lease_modify(struct file_lock
**before
, int arg
)
1267 static inline int lock_may_read(struct inode
*inode
, loff_t start
,
1273 static inline int lock_may_write(struct inode
*inode
, loff_t start
,
1279 #endif /* !CONFIG_FILE_LOCKING */
1282 struct fasync_struct
{
1285 struct fasync_struct
*fa_next
; /* singly linked list */
1286 struct file
*fa_file
;
1289 #define FASYNC_MAGIC 0x4601
1291 /* SMP safe fasync helpers: */
1292 extern int fasync_helper(int, struct file
*, int, struct fasync_struct
**);
1293 /* can be called from interrupts */
1294 extern void kill_fasync(struct fasync_struct
**, int, int);
1295 /* only for net: no internal synchronization */
1296 extern void __kill_fasync(struct fasync_struct
*, int, int);
1298 extern int __f_setown(struct file
*filp
, struct pid
*, enum pid_type
, int force
);
1299 extern int f_setown(struct file
*filp
, unsigned long arg
, int force
);
1300 extern void f_delown(struct file
*filp
);
1301 extern pid_t
f_getown(struct file
*filp
);
1302 extern int send_sigurg(struct fown_struct
*fown
);
1308 #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
1309 #define MNT_DETACH 0x00000002 /* Just detach from the tree */
1310 #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
1312 extern struct list_head super_blocks
;
1313 extern spinlock_t sb_lock
;
1315 #define sb_entry(list) list_entry((list), struct super_block, s_list)
1316 #define S_BIAS (1<<30)
1317 struct super_block
{
1318 struct list_head s_list
; /* Keep this first */
1319 dev_t s_dev
; /* search index; _not_ kdev_t */
1320 unsigned long s_blocksize
;
1321 unsigned char s_blocksize_bits
;
1322 unsigned char s_dirt
;
1323 loff_t s_maxbytes
; /* Max file size */
1324 struct file_system_type
*s_type
;
1325 const struct super_operations
*s_op
;
1326 const struct dquot_operations
*dq_op
;
1327 const struct quotactl_ops
*s_qcop
;
1328 const struct export_operations
*s_export_op
;
1329 unsigned long s_flags
;
1330 unsigned long s_magic
;
1331 struct dentry
*s_root
;
1332 struct rw_semaphore s_umount
;
1333 struct mutex s_lock
;
1337 #ifdef CONFIG_SECURITY
1340 struct xattr_handler
**s_xattr
;
1342 struct list_head s_inodes
; /* all inodes */
1343 struct hlist_head s_anon
; /* anonymous dentries for (nfs) exporting */
1344 struct list_head s_files
;
1345 /* s_dentry_lru and s_nr_dentry_unused are protected by dcache_lock */
1346 struct list_head s_dentry_lru
; /* unused dentry lru */
1347 int s_nr_dentry_unused
; /* # of dentry on lru */
1349 struct block_device
*s_bdev
;
1350 struct backing_dev_info
*s_bdi
;
1351 struct mtd_info
*s_mtd
;
1352 struct list_head s_instances
;
1353 struct quota_info s_dquot
; /* Diskquota specific options */
1356 wait_queue_head_t s_wait_unfrozen
;
1358 char s_id
[32]; /* Informational name */
1360 void *s_fs_info
; /* Filesystem private info */
1364 * The next field is for VFS *only*. No filesystems have any business
1365 * even looking at it. You had been warned.
1367 struct mutex s_vfs_rename_mutex
; /* Kludge */
1369 /* Granularity of c/m/atime in ns.
1370 Cannot be worse than a second */
1374 * Filesystem subtype. If non-empty the filesystem type field
1375 * in /proc/mounts will be "type.subtype"
1380 * Saved mount options for lazy filesystems using
1381 * generic_show_options()
1386 extern struct timespec
current_fs_time(struct super_block
*sb
);
1389 * Snapshotting support.
1393 SB_FREEZE_WRITE
= 1,
1394 SB_FREEZE_TRANS
= 2,
1397 #define vfs_check_frozen(sb, level) \
1398 wait_event((sb)->s_wait_unfrozen, ((sb)->s_frozen < (level)))
1400 #define get_fs_excl() atomic_inc(¤t->fs_excl)
1401 #define put_fs_excl() atomic_dec(¤t->fs_excl)
1402 #define has_fs_excl() atomic_read(¤t->fs_excl)
1404 #define is_owner_or_cap(inode) \
1405 ((current_fsuid() == (inode)->i_uid) || capable(CAP_FOWNER))
1407 /* not quite ready to be deprecated, but... */
1408 extern void lock_super(struct super_block
*);
1409 extern void unlock_super(struct super_block
*);
1412 * VFS helper functions..
1414 extern int vfs_create(struct inode
*, struct dentry
*, int, struct nameidata
*);
1415 extern int vfs_mkdir(struct inode
*, struct dentry
*, int);
1416 extern int vfs_mknod(struct inode
*, struct dentry
*, int, dev_t
);
1417 extern int vfs_symlink(struct inode
*, struct dentry
*, const char *);
1418 extern int vfs_link(struct dentry
*, struct inode
*, struct dentry
*);
1419 extern int vfs_rmdir(struct inode
*, struct dentry
*);
1420 extern int vfs_unlink(struct inode
*, struct dentry
*);
1421 extern int vfs_rename(struct inode
*, struct dentry
*, struct inode
*, struct dentry
*);
1424 * VFS dentry helper functions.
1426 extern void dentry_unhash(struct dentry
*dentry
);
1429 * VFS file helper functions.
1431 extern int file_permission(struct file
*, int);
1434 * VFS FS_IOC_FIEMAP helper definitions.
1436 struct fiemap_extent_info
{
1437 unsigned int fi_flags
; /* Flags as passed from user */
1438 unsigned int fi_extents_mapped
; /* Number of mapped extents */
1439 unsigned int fi_extents_max
; /* Size of fiemap_extent array */
1440 struct fiemap_extent
*fi_extents_start
; /* Start of fiemap_extent
1443 int fiemap_fill_next_extent(struct fiemap_extent_info
*info
, u64 logical
,
1444 u64 phys
, u64 len
, u32 flags
);
1445 int fiemap_check_flags(struct fiemap_extent_info
*fieinfo
, u32 fs_flags
);
1450 * NOTE! These match bits 12..15 of stat.st_mode
1451 * (ie "(i_mode >> 12) & 15").
1453 #define DT_UNKNOWN 0
1464 * This is the "filldir" function type, used by readdir() to let
1465 * the kernel specify what kind of dirent layout it wants to have.
1466 * This allows the kernel to read directories into kernel space or
1467 * to have different dirent layouts depending on the binary type.
1469 typedef int (*filldir_t
)(void *, const char *, int, loff_t
, u64
, unsigned);
1470 struct block_device_operations
;
1472 /* These macros are for out of kernel modules to test that
1473 * the kernel supports the unlocked_ioctl and compat_ioctl
1474 * fields in struct file_operations. */
1475 #define HAVE_COMPAT_IOCTL 1
1476 #define HAVE_UNLOCKED_IOCTL 1
1480 * read, write, poll, fsync, readv, writev, unlocked_ioctl and compat_ioctl
1481 * can be called without the big kernel lock held in all filesystems.
1483 struct file_operations
{
1484 struct module
*owner
;
1485 loff_t (*llseek
) (struct file
*, loff_t
, int);
1486 ssize_t (*read
) (struct file
*, char __user
*, size_t, loff_t
*);
1487 ssize_t (*write
) (struct file
*, const char __user
*, size_t, loff_t
*);
1488 ssize_t (*aio_read
) (struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
1489 ssize_t (*aio_write
) (struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
1490 int (*readdir
) (struct file
*, void *, filldir_t
);
1491 unsigned int (*poll
) (struct file
*, struct poll_table_struct
*);
1492 int (*ioctl
) (struct inode
*, struct file
*, unsigned int, unsigned long);
1493 long (*unlocked_ioctl
) (struct file
*, unsigned int, unsigned long);
1494 long (*compat_ioctl
) (struct file
*, unsigned int, unsigned long);
1495 int (*mmap
) (struct file
*, struct vm_area_struct
*);
1496 int (*open
) (struct inode
*, struct file
*);
1497 int (*flush
) (struct file
*, fl_owner_t id
);
1498 int (*release
) (struct inode
*, struct file
*);
1499 int (*fsync
) (struct file
*, struct dentry
*, int datasync
);
1500 int (*aio_fsync
) (struct kiocb
*, int datasync
);
1501 int (*fasync
) (int, struct file
*, int);
1502 int (*lock
) (struct file
*, int, struct file_lock
*);
1503 ssize_t (*sendpage
) (struct file
*, struct page
*, int, size_t, loff_t
*, int);
1504 unsigned long (*get_unmapped_area
)(struct file
*, unsigned long, unsigned long, unsigned long, unsigned long);
1505 int (*check_flags
)(int);
1506 int (*flock
) (struct file
*, int, struct file_lock
*);
1507 ssize_t (*splice_write
)(struct pipe_inode_info
*, struct file
*, loff_t
*, size_t, unsigned int);
1508 ssize_t (*splice_read
)(struct file
*, loff_t
*, struct pipe_inode_info
*, size_t, unsigned int);
1509 int (*setlease
)(struct file
*, long, struct file_lock
**);
1512 struct inode_operations
{
1513 int (*create
) (struct inode
*,struct dentry
*,int, struct nameidata
*);
1514 struct dentry
* (*lookup
) (struct inode
*,struct dentry
*, struct nameidata
*);
1515 int (*link
) (struct dentry
*,struct inode
*,struct dentry
*);
1516 int (*unlink
) (struct inode
*,struct dentry
*);
1517 int (*symlink
) (struct inode
*,struct dentry
*,const char *);
1518 int (*mkdir
) (struct inode
*,struct dentry
*,int);
1519 int (*rmdir
) (struct inode
*,struct dentry
*);
1520 int (*mknod
) (struct inode
*,struct dentry
*,int,dev_t
);
1521 int (*rename
) (struct inode
*, struct dentry
*,
1522 struct inode
*, struct dentry
*);
1523 int (*readlink
) (struct dentry
*, char __user
*,int);
1524 void * (*follow_link
) (struct dentry
*, struct nameidata
*);
1525 void (*put_link
) (struct dentry
*, struct nameidata
*, void *);
1526 void (*truncate
) (struct inode
*);
1527 int (*permission
) (struct inode
*, int);
1528 int (*check_acl
)(struct inode
*, int);
1529 int (*setattr
) (struct dentry
*, struct iattr
*);
1530 int (*getattr
) (struct vfsmount
*mnt
, struct dentry
*, struct kstat
*);
1531 int (*setxattr
) (struct dentry
*, const char *,const void *,size_t,int);
1532 ssize_t (*getxattr
) (struct dentry
*, const char *, void *, size_t);
1533 ssize_t (*listxattr
) (struct dentry
*, char *, size_t);
1534 int (*removexattr
) (struct dentry
*, const char *);
1535 void (*truncate_range
)(struct inode
*, loff_t
, loff_t
);
1536 long (*fallocate
)(struct inode
*inode
, int mode
, loff_t offset
,
1538 int (*fiemap
)(struct inode
*, struct fiemap_extent_info
*, u64 start
,
1544 ssize_t
rw_copy_check_uvector(int type
, const struct iovec __user
* uvector
,
1545 unsigned long nr_segs
, unsigned long fast_segs
,
1546 struct iovec
*fast_pointer
,
1547 struct iovec
**ret_pointer
);
1549 extern ssize_t
vfs_read(struct file
*, char __user
*, size_t, loff_t
*);
1550 extern ssize_t
vfs_write(struct file
*, const char __user
*, size_t, loff_t
*);
1551 extern ssize_t
vfs_readv(struct file
*, const struct iovec __user
*,
1552 unsigned long, loff_t
*);
1553 extern ssize_t
vfs_writev(struct file
*, const struct iovec __user
*,
1554 unsigned long, loff_t
*);
1556 struct super_operations
{
1557 struct inode
*(*alloc_inode
)(struct super_block
*sb
);
1558 void (*destroy_inode
)(struct inode
*);
1560 void (*dirty_inode
) (struct inode
*);
1561 int (*write_inode
) (struct inode
*, int);
1562 void (*drop_inode
) (struct inode
*);
1563 void (*delete_inode
) (struct inode
*);
1564 void (*put_super
) (struct super_block
*);
1565 void (*write_super
) (struct super_block
*);
1566 int (*sync_fs
)(struct super_block
*sb
, int wait
);
1567 int (*freeze_fs
) (struct super_block
*);
1568 int (*unfreeze_fs
) (struct super_block
*);
1569 int (*statfs
) (struct dentry
*, struct kstatfs
*);
1570 int (*remount_fs
) (struct super_block
*, int *, char *);
1571 void (*clear_inode
) (struct inode
*);
1572 void (*umount_begin
) (struct super_block
*);
1574 int (*show_options
)(struct seq_file
*, struct vfsmount
*);
1575 int (*show_stats
)(struct seq_file
*, struct vfsmount
*);
1577 ssize_t (*quota_read
)(struct super_block
*, int, char *, size_t, loff_t
);
1578 ssize_t (*quota_write
)(struct super_block
*, int, const char *, size_t, loff_t
);
1580 int (*bdev_try_to_free_page
)(struct super_block
*, struct page
*, gfp_t
);
1584 * Inode state bits. Protected by inode_lock.
1586 * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
1587 * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
1589 * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
1590 * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
1591 * various stages of removing an inode.
1593 * Two bits are used for locking and completion notification, I_LOCK and I_SYNC.
1595 * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
1596 * fdatasync(). i_atime is the usual cause.
1597 * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
1598 * these changes separately from I_DIRTY_SYNC so that we
1599 * don't have to write inode on fdatasync() when only
1600 * mtime has changed in it.
1601 * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
1602 * I_NEW get_new_inode() sets i_state to I_LOCK|I_NEW. Both
1603 * are cleared by unlock_new_inode(), called from iget().
1604 * I_WILL_FREE Must be set when calling write_inode_now() if i_count
1605 * is zero. I_FREEING must be set when I_WILL_FREE is
1607 * I_FREEING Set when inode is about to be freed but still has dirty
1608 * pages or buffers attached or the inode itself is still
1610 * I_CLEAR Set by clear_inode(). In this state the inode is clean
1611 * and can be destroyed.
1613 * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
1614 * prohibited for many purposes. iget() must wait for
1615 * the inode to be completely released, then create it
1616 * anew. Other functions will just ignore such inodes,
1617 * if appropriate. I_LOCK is used for waiting.
1619 * I_LOCK Serves as both a mutex and completion notification.
1620 * New inodes set I_LOCK. If two processes both create
1621 * the same inode, one of them will release its inode and
1622 * wait for I_LOCK to be released before returning.
1623 * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
1624 * also cause waiting on I_LOCK, without I_LOCK actually
1625 * being set. find_inode() uses this to prevent returning
1626 * nearly-dead inodes.
1627 * I_SYNC Similar to I_LOCK, but limited in scope to writeback
1628 * of inode dirty data. Having a separate lock for this
1629 * purpose reduces latency and prevents some filesystem-
1630 * specific deadlocks.
1632 * Q: What is the difference between I_WILL_FREE and I_FREEING?
1633 * Q: igrab() only checks on (I_FREEING|I_WILL_FREE). Should it also check on
1634 * I_CLEAR? If not, why?
1636 #define I_DIRTY_SYNC 1
1637 #define I_DIRTY_DATASYNC 2
1638 #define I_DIRTY_PAGES 4
1640 #define I_WILL_FREE 16
1641 #define I_FREEING 32
1644 #define I_LOCK (1 << __I_LOCK)
1646 #define I_SYNC (1 << __I_SYNC)
1648 #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1650 extern void __mark_inode_dirty(struct inode
*, int);
1651 static inline void mark_inode_dirty(struct inode
*inode
)
1653 __mark_inode_dirty(inode
, I_DIRTY
);
1656 static inline void mark_inode_dirty_sync(struct inode
*inode
)
1658 __mark_inode_dirty(inode
, I_DIRTY_SYNC
);
1662 * inc_nlink - directly increment an inode's link count
1665 * This is a low-level filesystem helper to replace any
1666 * direct filesystem manipulation of i_nlink. Currently,
1667 * it is only here for parity with dec_nlink().
1669 static inline void inc_nlink(struct inode
*inode
)
1674 static inline void inode_inc_link_count(struct inode
*inode
)
1677 mark_inode_dirty(inode
);
1681 * drop_nlink - directly drop an inode's link count
1684 * This is a low-level filesystem helper to replace any
1685 * direct filesystem manipulation of i_nlink. In cases
1686 * where we are attempting to track writes to the
1687 * filesystem, a decrement to zero means an imminent
1688 * write when the file is truncated and actually unlinked
1689 * on the filesystem.
1691 static inline void drop_nlink(struct inode
*inode
)
1697 * clear_nlink - directly zero an inode's link count
1700 * This is a low-level filesystem helper to replace any
1701 * direct filesystem manipulation of i_nlink. See
1702 * drop_nlink() for why we care about i_nlink hitting zero.
1704 static inline void clear_nlink(struct inode
*inode
)
1709 static inline void inode_dec_link_count(struct inode
*inode
)
1712 mark_inode_dirty(inode
);
1716 * inode_inc_iversion - increments i_version
1717 * @inode: inode that need to be updated
1719 * Every time the inode is modified, the i_version field will be incremented.
1720 * The filesystem has to be mounted with i_version flag
1723 static inline void inode_inc_iversion(struct inode
*inode
)
1725 spin_lock(&inode
->i_lock
);
1727 spin_unlock(&inode
->i_lock
);
1730 extern void touch_atime(struct vfsmount
*mnt
, struct dentry
*dentry
);
1731 static inline void file_accessed(struct file
*file
)
1733 if (!(file
->f_flags
& O_NOATIME
))
1734 touch_atime(file
->f_path
.mnt
, file
->f_path
.dentry
);
1737 int sync_inode(struct inode
*inode
, struct writeback_control
*wbc
);
1739 struct file_system_type
{
1742 int (*get_sb
) (struct file_system_type
*, int,
1743 const char *, void *, struct vfsmount
*);
1744 void (*kill_sb
) (struct super_block
*);
1745 struct module
*owner
;
1746 struct file_system_type
* next
;
1747 struct list_head fs_supers
;
1749 struct lock_class_key s_lock_key
;
1750 struct lock_class_key s_umount_key
;
1752 struct lock_class_key i_lock_key
;
1753 struct lock_class_key i_mutex_key
;
1754 struct lock_class_key i_mutex_dir_key
;
1755 struct lock_class_key i_alloc_sem_key
;
1758 extern int get_sb_ns(struct file_system_type
*fs_type
, int flags
, void *data
,
1759 int (*fill_super
)(struct super_block
*, void *, int),
1760 struct vfsmount
*mnt
);
1761 extern int get_sb_bdev(struct file_system_type
*fs_type
,
1762 int flags
, const char *dev_name
, void *data
,
1763 int (*fill_super
)(struct super_block
*, void *, int),
1764 struct vfsmount
*mnt
);
1765 extern int get_sb_single(struct file_system_type
*fs_type
,
1766 int flags
, void *data
,
1767 int (*fill_super
)(struct super_block
*, void *, int),
1768 struct vfsmount
*mnt
);
1769 extern int get_sb_nodev(struct file_system_type
*fs_type
,
1770 int flags
, void *data
,
1771 int (*fill_super
)(struct super_block
*, void *, int),
1772 struct vfsmount
*mnt
);
1773 void generic_shutdown_super(struct super_block
*sb
);
1774 void kill_block_super(struct super_block
*sb
);
1775 void kill_anon_super(struct super_block
*sb
);
1776 void kill_litter_super(struct super_block
*sb
);
1777 void deactivate_super(struct super_block
*sb
);
1778 void deactivate_locked_super(struct super_block
*sb
);
1779 int set_anon_super(struct super_block
*s
, void *data
);
1780 struct super_block
*sget(struct file_system_type
*type
,
1781 int (*test
)(struct super_block
*,void *),
1782 int (*set
)(struct super_block
*,void *),
1784 extern int get_sb_pseudo(struct file_system_type
*, char *,
1785 const struct super_operations
*ops
, unsigned long,
1786 struct vfsmount
*mnt
);
1787 extern void simple_set_mnt(struct vfsmount
*mnt
, struct super_block
*sb
);
1788 int __put_super_and_need_restart(struct super_block
*sb
);
1789 void put_super(struct super_block
*sb
);
1791 /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
1792 #define fops_get(fops) \
1793 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
1794 #define fops_put(fops) \
1795 do { if (fops) module_put((fops)->owner); } while(0)
1797 extern int register_filesystem(struct file_system_type
*);
1798 extern int unregister_filesystem(struct file_system_type
*);
1799 extern struct vfsmount
*kern_mount_data(struct file_system_type
*, void *data
);
1800 #define kern_mount(type) kern_mount_data(type, NULL)
1801 extern int may_umount_tree(struct vfsmount
*);
1802 extern int may_umount(struct vfsmount
*);
1803 extern long do_mount(char *, char *, char *, unsigned long, void *);
1804 extern struct vfsmount
*collect_mounts(struct path
*);
1805 extern void drop_collected_mounts(struct vfsmount
*);
1807 extern int vfs_statfs(struct dentry
*, struct kstatfs
*);
1809 extern int current_umask(void);
1812 extern struct kobject
*fs_kobj
;
1814 extern int rw_verify_area(int, struct file
*, loff_t
*, size_t);
1816 #define FLOCK_VERIFY_READ 1
1817 #define FLOCK_VERIFY_WRITE 2
1819 #ifdef CONFIG_FILE_LOCKING
1820 extern int locks_mandatory_locked(struct inode
*);
1821 extern int locks_mandatory_area(int, struct inode
*, struct file
*, loff_t
, size_t);
1824 * Candidates for mandatory locking have the setgid bit set
1825 * but no group execute bit - an otherwise meaningless combination.
1828 static inline int __mandatory_lock(struct inode
*ino
)
1830 return (ino
->i_mode
& (S_ISGID
| S_IXGRP
)) == S_ISGID
;
1834 * ... and these candidates should be on MS_MANDLOCK mounted fs,
1835 * otherwise these will be advisory locks
1838 static inline int mandatory_lock(struct inode
*ino
)
1840 return IS_MANDLOCK(ino
) && __mandatory_lock(ino
);
1843 static inline int locks_verify_locked(struct inode
*inode
)
1845 if (mandatory_lock(inode
))
1846 return locks_mandatory_locked(inode
);
1850 static inline int locks_verify_truncate(struct inode
*inode
,
1854 if (inode
->i_flock
&& mandatory_lock(inode
))
1855 return locks_mandatory_area(
1856 FLOCK_VERIFY_WRITE
, inode
, filp
,
1857 size
< inode
->i_size
? size
: inode
->i_size
,
1858 (size
< inode
->i_size
? inode
->i_size
- size
1859 : size
- inode
->i_size
)
1864 static inline int break_lease(struct inode
*inode
, unsigned int mode
)
1867 return __break_lease(inode
, mode
);
1870 #else /* !CONFIG_FILE_LOCKING */
1871 static inline int locks_mandatory_locked(struct inode
*inode
)
1876 static inline int locks_mandatory_area(int rw
, struct inode
*inode
,
1877 struct file
*filp
, loff_t offset
,
1883 static inline int __mandatory_lock(struct inode
*inode
)
1888 static inline int mandatory_lock(struct inode
*inode
)
1893 static inline int locks_verify_locked(struct inode
*inode
)
1898 static inline int locks_verify_truncate(struct inode
*inode
, struct file
*filp
,
1904 static inline int break_lease(struct inode
*inode
, unsigned int mode
)
1909 #endif /* CONFIG_FILE_LOCKING */
1913 extern int do_truncate(struct dentry
*, loff_t start
, unsigned int time_attrs
,
1915 extern int do_fallocate(struct file
*file
, int mode
, loff_t offset
,
1917 extern long do_sys_open(int dfd
, const char __user
*filename
, int flags
,
1919 extern struct file
*filp_open(const char *, int, int);
1920 extern struct file
* dentry_open(struct dentry
*, struct vfsmount
*, int,
1921 const struct cred
*);
1922 extern int filp_close(struct file
*, fl_owner_t id
);
1923 extern char * getname(const char __user
*);
1927 extern int ioctl_preallocate(struct file
*filp
, void __user
*argp
);
1930 extern void __init
vfs_caches_init_early(void);
1931 extern void __init
vfs_caches_init(unsigned long);
1933 extern struct kmem_cache
*names_cachep
;
1935 #define __getname_gfp(gfp) kmem_cache_alloc(names_cachep, (gfp))
1936 #define __getname() __getname_gfp(GFP_KERNEL)
1937 #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
1938 #ifndef CONFIG_AUDITSYSCALL
1939 #define putname(name) __putname(name)
1941 extern void putname(const char *name
);
1945 extern int register_blkdev(unsigned int, const char *);
1946 extern void unregister_blkdev(unsigned int, const char *);
1947 extern struct block_device
*bdget(dev_t
);
1948 extern struct block_device
*bdgrab(struct block_device
*bdev
);
1949 extern void bd_set_size(struct block_device
*, loff_t size
);
1950 extern void bd_forget(struct inode
*inode
);
1951 extern void bdput(struct block_device
*);
1952 extern struct block_device
*open_by_devnum(dev_t
, fmode_t
);
1953 extern void invalidate_bdev(struct block_device
*);
1954 extern int sync_blockdev(struct block_device
*bdev
);
1955 extern struct super_block
*freeze_bdev(struct block_device
*);
1956 extern void emergency_thaw_all(void);
1957 extern int thaw_bdev(struct block_device
*bdev
, struct super_block
*sb
);
1958 extern int fsync_bdev(struct block_device
*);
1960 static inline void bd_forget(struct inode
*inode
) {}
1961 static inline int sync_blockdev(struct block_device
*bdev
) { return 0; }
1962 static inline void invalidate_bdev(struct block_device
*bdev
) {}
1964 static inline struct super_block
*freeze_bdev(struct block_device
*sb
)
1969 static inline int thaw_bdev(struct block_device
*bdev
, struct super_block
*sb
)
1974 extern int sync_filesystem(struct super_block
*);
1975 extern const struct file_operations def_blk_fops
;
1976 extern const struct file_operations def_chr_fops
;
1977 extern const struct file_operations bad_sock_fops
;
1978 extern const struct file_operations def_fifo_fops
;
1980 extern int ioctl_by_bdev(struct block_device
*, unsigned, unsigned long);
1981 extern int blkdev_ioctl(struct block_device
*, fmode_t
, unsigned, unsigned long);
1982 extern long compat_blkdev_ioctl(struct file
*, unsigned, unsigned long);
1983 extern int blkdev_get(struct block_device
*, fmode_t
);
1984 extern int blkdev_put(struct block_device
*, fmode_t
);
1985 extern int bd_claim(struct block_device
*, void *);
1986 extern void bd_release(struct block_device
*);
1988 extern int bd_claim_by_disk(struct block_device
*, void *, struct gendisk
*);
1989 extern void bd_release_from_disk(struct block_device
*, struct gendisk
*);
1991 #define bd_claim_by_disk(bdev, holder, disk) bd_claim(bdev, holder)
1992 #define bd_release_from_disk(bdev, disk) bd_release(bdev)
1997 #define CHRDEV_MAJOR_HASH_SIZE 255
1998 extern int alloc_chrdev_region(dev_t
*, unsigned, unsigned, const char *);
1999 extern int register_chrdev_region(dev_t
, unsigned, const char *);
2000 extern int __register_chrdev(unsigned int major
, unsigned int baseminor
,
2001 unsigned int count
, const char *name
,
2002 const struct file_operations
*fops
);
2003 extern void __unregister_chrdev(unsigned int major
, unsigned int baseminor
,
2004 unsigned int count
, const char *name
);
2005 extern void unregister_chrdev_region(dev_t
, unsigned);
2006 extern void chrdev_show(struct seq_file
*,off_t
);
2008 static inline int register_chrdev(unsigned int major
, const char *name
,
2009 const struct file_operations
*fops
)
2011 return __register_chrdev(major
, 0, 256, name
, fops
);
2014 static inline void unregister_chrdev(unsigned int major
, const char *name
)
2016 __unregister_chrdev(major
, 0, 256, name
);
2019 /* fs/block_dev.c */
2020 #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
2021 #define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
2024 #define BLKDEV_MAJOR_HASH_SIZE 255
2025 extern const char *__bdevname(dev_t
, char *buffer
);
2026 extern const char *bdevname(struct block_device
*bdev
, char *buffer
);
2027 extern struct block_device
*lookup_bdev(const char *);
2028 extern struct block_device
*open_bdev_exclusive(const char *, fmode_t
, void *);
2029 extern void close_bdev_exclusive(struct block_device
*, fmode_t
);
2030 extern void blkdev_show(struct seq_file
*,off_t
);
2033 #define BLKDEV_MAJOR_HASH_SIZE 0
2036 extern void init_special_inode(struct inode
*, umode_t
, dev_t
);
2038 /* Invalid inode operations -- fs/bad_inode.c */
2039 extern void make_bad_inode(struct inode
*);
2040 extern int is_bad_inode(struct inode
*);
2042 extern const struct file_operations read_pipefifo_fops
;
2043 extern const struct file_operations write_pipefifo_fops
;
2044 extern const struct file_operations rdwr_pipefifo_fops
;
2046 extern int fs_may_remount_ro(struct super_block
*);
2050 * return READ, READA, or WRITE
2052 #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
2055 * return data direction, READ or WRITE
2057 #define bio_data_dir(bio) ((bio)->bi_rw & 1)
2059 extern void check_disk_size_change(struct gendisk
*disk
,
2060 struct block_device
*bdev
);
2061 extern int revalidate_disk(struct gendisk
*);
2062 extern int check_disk_change(struct block_device
*);
2063 extern int __invalidate_device(struct block_device
*);
2064 extern int invalidate_partition(struct gendisk
*, int);
2066 extern int invalidate_inodes(struct super_block
*);
2067 unsigned long invalidate_mapping_pages(struct address_space
*mapping
,
2068 pgoff_t start
, pgoff_t end
);
2070 static inline unsigned long __deprecated
2071 invalidate_inode_pages(struct address_space
*mapping
)
2073 return invalidate_mapping_pages(mapping
, 0, ~0UL);
2076 static inline void invalidate_remote_inode(struct inode
*inode
)
2078 if (S_ISREG(inode
->i_mode
) || S_ISDIR(inode
->i_mode
) ||
2079 S_ISLNK(inode
->i_mode
))
2080 invalidate_mapping_pages(inode
->i_mapping
, 0, -1);
2082 extern int invalidate_inode_pages2(struct address_space
*mapping
);
2083 extern int invalidate_inode_pages2_range(struct address_space
*mapping
,
2084 pgoff_t start
, pgoff_t end
);
2085 extern int write_inode_now(struct inode
*, int);
2086 extern int filemap_fdatawrite(struct address_space
*);
2087 extern int filemap_flush(struct address_space
*);
2088 extern int filemap_fdatawait(struct address_space
*);
2089 extern int filemap_fdatawait_range(struct address_space
*, loff_t lstart
,
2091 extern int filemap_write_and_wait(struct address_space
*mapping
);
2092 extern int filemap_write_and_wait_range(struct address_space
*mapping
,
2093 loff_t lstart
, loff_t lend
);
2094 extern int __filemap_fdatawrite_range(struct address_space
*mapping
,
2095 loff_t start
, loff_t end
, int sync_mode
);
2096 extern int filemap_fdatawrite_range(struct address_space
*mapping
,
2097 loff_t start
, loff_t end
);
2099 extern int vfs_fsync_range(struct file
*file
, struct dentry
*dentry
,
2100 loff_t start
, loff_t end
, int datasync
);
2101 extern int vfs_fsync(struct file
*file
, struct dentry
*dentry
, int datasync
);
2102 extern int generic_write_sync(struct file
*file
, loff_t pos
, loff_t count
);
2103 extern void sync_supers(void);
2104 extern void emergency_sync(void);
2105 extern void emergency_remount(void);
2107 extern sector_t
bmap(struct inode
*, sector_t
);
2109 extern int notify_change(struct dentry
*, struct iattr
*);
2110 extern int inode_permission(struct inode
*, int);
2111 extern int generic_permission(struct inode
*, int,
2112 int (*check_acl
)(struct inode
*, int));
2114 static inline bool execute_ok(struct inode
*inode
)
2116 return (inode
->i_mode
& S_IXUGO
) || S_ISDIR(inode
->i_mode
);
2119 extern int get_write_access(struct inode
*);
2120 extern int deny_write_access(struct file
*);
2121 static inline void put_write_access(struct inode
* inode
)
2123 atomic_dec(&inode
->i_writecount
);
2125 static inline void allow_write_access(struct file
*file
)
2128 atomic_inc(&file
->f_path
.dentry
->d_inode
->i_writecount
);
2130 extern int do_pipe_flags(int *, int);
2131 extern struct file
*create_read_pipe(struct file
*f
, int flags
);
2132 extern struct file
*create_write_pipe(int flags
);
2133 extern void free_write_pipe(struct file
*);
2135 extern struct file
*do_filp_open(int dfd
, const char *pathname
,
2136 int open_flag
, int mode
, int acc_mode
);
2137 extern int may_open(struct path
*, int, int);
2139 extern int kernel_read(struct file
*, loff_t
, char *, unsigned long);
2140 extern struct file
* open_exec(const char *);
2142 /* fs/dcache.c -- generic fs support functions */
2143 extern int is_subdir(struct dentry
*, struct dentry
*);
2144 extern ino_t
find_inode_number(struct dentry
*, struct qstr
*);
2146 #include <linux/err.h>
2148 /* needed for stackable file system support */
2149 extern loff_t
default_llseek(struct file
*file
, loff_t offset
, int origin
);
2151 extern loff_t
vfs_llseek(struct file
*file
, loff_t offset
, int origin
);
2153 extern int inode_init_always(struct super_block
*, struct inode
*);
2154 extern void inode_init_once(struct inode
*);
2155 extern void inode_add_to_lists(struct super_block
*, struct inode
*);
2156 extern void iput(struct inode
*);
2157 extern struct inode
* igrab(struct inode
*);
2158 extern ino_t
iunique(struct super_block
*, ino_t
);
2159 extern int inode_needs_sync(struct inode
*inode
);
2160 extern void generic_delete_inode(struct inode
*inode
);
2161 extern void generic_drop_inode(struct inode
*inode
);
2162 extern int generic_detach_inode(struct inode
*inode
);
2164 extern struct inode
*ilookup5_nowait(struct super_block
*sb
,
2165 unsigned long hashval
, int (*test
)(struct inode
*, void *),
2167 extern struct inode
*ilookup5(struct super_block
*sb
, unsigned long hashval
,
2168 int (*test
)(struct inode
*, void *), void *data
);
2169 extern struct inode
*ilookup(struct super_block
*sb
, unsigned long ino
);
2171 extern struct inode
* iget5_locked(struct super_block
*, unsigned long, int (*test
)(struct inode
*, void *), int (*set
)(struct inode
*, void *), void *);
2172 extern struct inode
* iget_locked(struct super_block
*, unsigned long);
2173 extern int insert_inode_locked4(struct inode
*, unsigned long, int (*test
)(struct inode
*, void *), void *);
2174 extern int insert_inode_locked(struct inode
*);
2175 extern void unlock_new_inode(struct inode
*);
2177 extern void __iget(struct inode
* inode
);
2178 extern void iget_failed(struct inode
*);
2179 extern void clear_inode(struct inode
*);
2180 extern void destroy_inode(struct inode
*);
2181 extern void __destroy_inode(struct inode
*);
2182 extern struct inode
*new_inode(struct super_block
*);
2183 extern int should_remove_suid(struct dentry
*);
2184 extern int file_remove_suid(struct file
*);
2186 extern void __insert_inode_hash(struct inode
*, unsigned long hashval
);
2187 extern void remove_inode_hash(struct inode
*);
2188 static inline void insert_inode_hash(struct inode
*inode
) {
2189 __insert_inode_hash(inode
, inode
->i_ino
);
2192 extern struct file
* get_empty_filp(void);
2193 extern void file_move(struct file
*f
, struct list_head
*list
);
2194 extern void file_kill(struct file
*f
);
2197 extern void submit_bio(int, struct bio
*);
2198 extern int bdev_read_only(struct block_device
*);
2200 extern int set_blocksize(struct block_device
*, int);
2201 extern int sb_set_blocksize(struct super_block
*, int);
2202 extern int sb_min_blocksize(struct super_block
*, int);
2204 extern int generic_file_mmap(struct file
*, struct vm_area_struct
*);
2205 extern int generic_file_readonly_mmap(struct file
*, struct vm_area_struct
*);
2206 extern int file_read_actor(read_descriptor_t
* desc
, struct page
*page
, unsigned long offset
, unsigned long size
);
2207 int generic_write_checks(struct file
*file
, loff_t
*pos
, size_t *count
, int isblk
);
2208 extern ssize_t
generic_file_aio_read(struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
2209 extern ssize_t
__generic_file_aio_write(struct kiocb
*, const struct iovec
*, unsigned long,
2211 extern ssize_t
generic_file_aio_write(struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
2212 extern ssize_t
generic_file_direct_write(struct kiocb
*, const struct iovec
*,
2213 unsigned long *, loff_t
, loff_t
*, size_t, size_t);
2214 extern ssize_t
generic_file_buffered_write(struct kiocb
*, const struct iovec
*,
2215 unsigned long, loff_t
, loff_t
*, size_t, ssize_t
);
2216 extern ssize_t
do_sync_read(struct file
*filp
, char __user
*buf
, size_t len
, loff_t
*ppos
);
2217 extern ssize_t
do_sync_write(struct file
*filp
, const char __user
*buf
, size_t len
, loff_t
*ppos
);
2218 extern int generic_segment_checks(const struct iovec
*iov
,
2219 unsigned long *nr_segs
, size_t *count
, int access_flags
);
2221 /* fs/block_dev.c */
2222 extern ssize_t
blkdev_aio_write(struct kiocb
*iocb
, const struct iovec
*iov
,
2223 unsigned long nr_segs
, loff_t pos
);
2226 extern ssize_t
generic_file_splice_read(struct file
*, loff_t
*,
2227 struct pipe_inode_info
*, size_t, unsigned int);
2228 extern ssize_t
default_file_splice_read(struct file
*, loff_t
*,
2229 struct pipe_inode_info
*, size_t, unsigned int);
2230 extern ssize_t
generic_file_splice_write(struct pipe_inode_info
*,
2231 struct file
*, loff_t
*, size_t, unsigned int);
2232 extern ssize_t
generic_splice_sendpage(struct pipe_inode_info
*pipe
,
2233 struct file
*out
, loff_t
*, size_t len
, unsigned int flags
);
2234 extern long do_splice_direct(struct file
*in
, loff_t
*ppos
, struct file
*out
,
2235 size_t len
, unsigned int flags
);
2238 file_ra_state_init(struct file_ra_state
*ra
, struct address_space
*mapping
);
2239 extern loff_t
no_llseek(struct file
*file
, loff_t offset
, int origin
);
2240 extern loff_t
generic_file_llseek(struct file
*file
, loff_t offset
, int origin
);
2241 extern loff_t
generic_file_llseek_unlocked(struct file
*file
, loff_t offset
,
2243 extern int generic_file_open(struct inode
* inode
, struct file
* filp
);
2244 extern int nonseekable_open(struct inode
* inode
, struct file
* filp
);
2246 #ifdef CONFIG_FS_XIP
2247 extern ssize_t
xip_file_read(struct file
*filp
, char __user
*buf
, size_t len
,
2249 extern int xip_file_mmap(struct file
* file
, struct vm_area_struct
* vma
);
2250 extern ssize_t
xip_file_write(struct file
*filp
, const char __user
*buf
,
2251 size_t len
, loff_t
*ppos
);
2252 extern int xip_truncate_page(struct address_space
*mapping
, loff_t from
);
2254 static inline int xip_truncate_page(struct address_space
*mapping
, loff_t from
)
2261 ssize_t
__blockdev_direct_IO(int rw
, struct kiocb
*iocb
, struct inode
*inode
,
2262 struct block_device
*bdev
, const struct iovec
*iov
, loff_t offset
,
2263 unsigned long nr_segs
, get_block_t get_block
, dio_iodone_t end_io
,
2267 /* need locking between buffered and direct access */
2270 /* filesystem does not support filling holes */
2271 DIO_SKIP_HOLES
= 0x02,
2274 static inline ssize_t
blockdev_direct_IO(int rw
, struct kiocb
*iocb
,
2275 struct inode
*inode
, struct block_device
*bdev
, const struct iovec
*iov
,
2276 loff_t offset
, unsigned long nr_segs
, get_block_t get_block
,
2277 dio_iodone_t end_io
)
2279 return __blockdev_direct_IO(rw
, iocb
, inode
, bdev
, iov
, offset
,
2280 nr_segs
, get_block
, end_io
,
2281 DIO_LOCKING
| DIO_SKIP_HOLES
);
2284 static inline ssize_t
blockdev_direct_IO_no_locking(int rw
, struct kiocb
*iocb
,
2285 struct inode
*inode
, struct block_device
*bdev
, const struct iovec
*iov
,
2286 loff_t offset
, unsigned long nr_segs
, get_block_t get_block
,
2287 dio_iodone_t end_io
)
2289 return __blockdev_direct_IO(rw
, iocb
, inode
, bdev
, iov
, offset
,
2290 nr_segs
, get_block
, end_io
, 0);
2294 extern const struct file_operations generic_ro_fops
;
2296 #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
2298 extern int vfs_readlink(struct dentry
*, char __user
*, int, const char *);
2299 extern int vfs_follow_link(struct nameidata
*, const char *);
2300 extern int page_readlink(struct dentry
*, char __user
*, int);
2301 extern void *page_follow_link_light(struct dentry
*, struct nameidata
*);
2302 extern void page_put_link(struct dentry
*, struct nameidata
*, void *);
2303 extern int __page_symlink(struct inode
*inode
, const char *symname
, int len
,
2305 extern int page_symlink(struct inode
*inode
, const char *symname
, int len
);
2306 extern const struct inode_operations page_symlink_inode_operations
;
2307 extern int generic_readlink(struct dentry
*, char __user
*, int);
2308 extern void generic_fillattr(struct inode
*, struct kstat
*);
2309 extern int vfs_getattr(struct vfsmount
*, struct dentry
*, struct kstat
*);
2310 void inode_add_bytes(struct inode
*inode
, loff_t bytes
);
2311 void inode_sub_bytes(struct inode
*inode
, loff_t bytes
);
2312 loff_t
inode_get_bytes(struct inode
*inode
);
2313 void inode_set_bytes(struct inode
*inode
, loff_t bytes
);
2315 extern int vfs_readdir(struct file
*, filldir_t
, void *);
2317 extern int vfs_stat(char __user
*, struct kstat
*);
2318 extern int vfs_lstat(char __user
*, struct kstat
*);
2319 extern int vfs_fstat(unsigned int, struct kstat
*);
2320 extern int vfs_fstatat(int , char __user
*, struct kstat
*, int);
2322 extern int do_vfs_ioctl(struct file
*filp
, unsigned int fd
, unsigned int cmd
,
2324 extern int __generic_block_fiemap(struct inode
*inode
,
2325 struct fiemap_extent_info
*fieinfo
, u64 start
,
2326 u64 len
, get_block_t
*get_block
);
2327 extern int generic_block_fiemap(struct inode
*inode
,
2328 struct fiemap_extent_info
*fieinfo
, u64 start
,
2329 u64 len
, get_block_t
*get_block
);
2331 extern void get_filesystem(struct file_system_type
*fs
);
2332 extern void put_filesystem(struct file_system_type
*fs
);
2333 extern struct file_system_type
*get_fs_type(const char *name
);
2334 extern struct super_block
*get_super(struct block_device
*);
2335 extern struct super_block
*get_active_super(struct block_device
*bdev
);
2336 extern struct super_block
*user_get_super(dev_t
);
2337 extern void drop_super(struct super_block
*sb
);
2339 extern int dcache_dir_open(struct inode
*, struct file
*);
2340 extern int dcache_dir_close(struct inode
*, struct file
*);
2341 extern loff_t
dcache_dir_lseek(struct file
*, loff_t
, int);
2342 extern int dcache_readdir(struct file
*, void *, filldir_t
);
2343 extern int simple_getattr(struct vfsmount
*, struct dentry
*, struct kstat
*);
2344 extern int simple_statfs(struct dentry
*, struct kstatfs
*);
2345 extern int simple_link(struct dentry
*, struct inode
*, struct dentry
*);
2346 extern int simple_unlink(struct inode
*, struct dentry
*);
2347 extern int simple_rmdir(struct inode
*, struct dentry
*);
2348 extern int simple_rename(struct inode
*, struct dentry
*, struct inode
*, struct dentry
*);
2349 extern int simple_sync_file(struct file
*, struct dentry
*, int);
2350 extern int simple_empty(struct dentry
*);
2351 extern int simple_readpage(struct file
*file
, struct page
*page
);
2352 extern int simple_prepare_write(struct file
*file
, struct page
*page
,
2353 unsigned offset
, unsigned to
);
2354 extern int simple_write_begin(struct file
*file
, struct address_space
*mapping
,
2355 loff_t pos
, unsigned len
, unsigned flags
,
2356 struct page
**pagep
, void **fsdata
);
2357 extern int simple_write_end(struct file
*file
, struct address_space
*mapping
,
2358 loff_t pos
, unsigned len
, unsigned copied
,
2359 struct page
*page
, void *fsdata
);
2361 extern struct dentry
*simple_lookup(struct inode
*, struct dentry
*, struct nameidata
*);
2362 extern ssize_t
generic_read_dir(struct file
*, char __user
*, size_t, loff_t
*);
2363 extern const struct file_operations simple_dir_operations
;
2364 extern const struct inode_operations simple_dir_inode_operations
;
2365 struct tree_descr
{ char *name
; const struct file_operations
*ops
; int mode
; };
2366 struct dentry
*d_alloc_name(struct dentry
*, const char *);
2367 extern int simple_fill_super(struct super_block
*, int, struct tree_descr
*);
2368 extern int simple_pin_fs(struct file_system_type
*, struct vfsmount
**mount
, int *count
);
2369 extern void simple_release_fs(struct vfsmount
**mount
, int *count
);
2371 extern ssize_t
simple_read_from_buffer(void __user
*to
, size_t count
,
2372 loff_t
*ppos
, const void *from
, size_t available
);
2374 extern int simple_fsync(struct file
*, struct dentry
*, int);
2376 #ifdef CONFIG_MIGRATION
2377 extern int buffer_migrate_page(struct address_space
*,
2378 struct page
*, struct page
*);
2380 #define buffer_migrate_page NULL
2383 extern int inode_change_ok(const struct inode
*, struct iattr
*);
2384 extern int inode_newsize_ok(const struct inode
*, loff_t offset
);
2385 extern int __must_check
inode_setattr(struct inode
*, struct iattr
*);
2387 extern void file_update_time(struct file
*file
);
2389 extern int generic_show_options(struct seq_file
*m
, struct vfsmount
*mnt
);
2390 extern void save_mount_options(struct super_block
*sb
, char *options
);
2391 extern void replace_mount_options(struct super_block
*sb
, char *options
);
2393 static inline ino_t
parent_ino(struct dentry
*dentry
)
2397 spin_lock(&dentry
->d_lock
);
2398 res
= dentry
->d_parent
->d_inode
->i_ino
;
2399 spin_unlock(&dentry
->d_lock
);
2403 /* Transaction based IO helpers */
2406 * An argresp is stored in an allocated page and holds the
2407 * size of the argument or response, along with its content
2409 struct simple_transaction_argresp
{
2414 #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
2416 char *simple_transaction_get(struct file
*file
, const char __user
*buf
,
2418 ssize_t
simple_transaction_read(struct file
*file
, char __user
*buf
,
2419 size_t size
, loff_t
*pos
);
2420 int simple_transaction_release(struct inode
*inode
, struct file
*file
);
2422 void simple_transaction_set(struct file
*file
, size_t n
);
2425 * simple attribute files
2427 * These attributes behave similar to those in sysfs:
2429 * Writing to an attribute immediately sets a value, an open file can be
2430 * written to multiple times.
2432 * Reading from an attribute creates a buffer from the value that might get
2433 * read with multiple read calls. When the attribute has been read
2434 * completely, no further read calls are possible until the file is opened
2437 * All attributes contain a text representation of a numeric value
2438 * that are accessed with the get() and set() functions.
2440 #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
2441 static int __fops ## _open(struct inode *inode, struct file *file) \
2443 __simple_attr_check_format(__fmt, 0ull); \
2444 return simple_attr_open(inode, file, __get, __set, __fmt); \
2446 static const struct file_operations __fops = { \
2447 .owner = THIS_MODULE, \
2448 .open = __fops ## _open, \
2449 .release = simple_attr_release, \
2450 .read = simple_attr_read, \
2451 .write = simple_attr_write, \
2454 static inline void __attribute__((format(printf
, 1, 2)))
2455 __simple_attr_check_format(const char *fmt
, ...)
2457 /* don't do anything, just let the compiler check the arguments; */
2460 int simple_attr_open(struct inode
*inode
, struct file
*file
,
2461 int (*get
)(void *, u64
*), int (*set
)(void *, u64
),
2463 int simple_attr_release(struct inode
*inode
, struct file
*file
);
2464 ssize_t
simple_attr_read(struct file
*file
, char __user
*buf
,
2465 size_t len
, loff_t
*ppos
);
2466 ssize_t
simple_attr_write(struct file
*file
, const char __user
*buf
,
2467 size_t len
, loff_t
*ppos
);
2470 int proc_nr_files(struct ctl_table
*table
, int write
,
2471 void __user
*buffer
, size_t *lenp
, loff_t
*ppos
);
2473 int __init
get_filesystem_list(char *buf
);
2475 #endif /* __KERNEL__ */
2476 #endif /* _LINUX_FS_H */