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)
94 #define READA 2 /* read-ahead - don't block if no resources */
95 #define SWRITE 3 /* for ll_rw_block() - wait for buffer lock */
96 #define READ_SYNC (READ | (1 << BIO_RW_SYNCIO) | (1 << BIO_RW_UNPLUG))
97 #define READ_META (READ | (1 << BIO_RW_META))
98 #define WRITE_SYNC_PLUG (WRITE | (1 << BIO_RW_SYNCIO) | (1 << BIO_RW_NOIDLE))
99 #define WRITE_SYNC (WRITE_SYNC_PLUG | (1 << BIO_RW_UNPLUG))
100 #define WRITE_ODIRECT (WRITE | (1 << BIO_RW_SYNCIO) | (1 << BIO_RW_UNPLUG))
101 #define SWRITE_SYNC_PLUG \
102 (SWRITE | (1 << BIO_RW_SYNCIO) | (1 << BIO_RW_NOIDLE))
103 #define SWRITE_SYNC (SWRITE_SYNC_PLUG | (1 << BIO_RW_UNPLUG))
104 #define WRITE_BARRIER (WRITE | (1 << BIO_RW_BARRIER))
105 #define DISCARD_NOBARRIER (1 << BIO_RW_DISCARD)
106 #define DISCARD_BARRIER ((1 << BIO_RW_DISCARD) | (1 << BIO_RW_BARRIER))
112 /* public flags for file_system_type */
113 #define FS_REQUIRES_DEV 1
114 #define FS_BINARY_MOUNTDATA 2
115 #define FS_HAS_SUBTYPE 4
116 #define FS_REVAL_DOT 16384 /* Check the paths ".", ".." for staleness */
117 #define FS_RENAME_DOES_D_MOVE 32768 /* FS will handle d_move()
118 * during rename() internally.
122 * These are the fs-independent mount-flags: up to 32 flags are supported
124 #define MS_RDONLY 1 /* Mount read-only */
125 #define MS_NOSUID 2 /* Ignore suid and sgid bits */
126 #define MS_NODEV 4 /* Disallow access to device special files */
127 #define MS_NOEXEC 8 /* Disallow program execution */
128 #define MS_SYNCHRONOUS 16 /* Writes are synced at once */
129 #define MS_REMOUNT 32 /* Alter flags of a mounted FS */
130 #define MS_MANDLOCK 64 /* Allow mandatory locks on an FS */
131 #define MS_DIRSYNC 128 /* Directory modifications are synchronous */
132 #define MS_NOATIME 1024 /* Do not update access times. */
133 #define MS_NODIRATIME 2048 /* Do not update directory access times */
137 #define MS_VERBOSE 32768 /* War is peace. Verbosity is silence.
138 MS_VERBOSE is deprecated. */
139 #define MS_SILENT 32768
140 #define MS_POSIXACL (1<<16) /* VFS does not apply the umask */
141 #define MS_UNBINDABLE (1<<17) /* change to unbindable */
142 #define MS_PRIVATE (1<<18) /* change to private */
143 #define MS_SLAVE (1<<19) /* change to slave */
144 #define MS_SHARED (1<<20) /* change to shared */
145 #define MS_RELATIME (1<<21) /* Update atime relative to mtime/ctime. */
146 #define MS_KERNMOUNT (1<<22) /* this is a kern_mount call */
147 #define MS_I_VERSION (1<<23) /* Update inode I_version field */
148 #define MS_STRICTATIME (1<<24) /* Always perform atime updates */
149 #define MS_ACTIVE (1<<30)
150 #define MS_NOUSER (1<<31)
153 * Superblock flags that can be altered by MS_REMOUNT
155 #define MS_RMT_MASK (MS_RDONLY|MS_SYNCHRONOUS|MS_MANDLOCK|MS_I_VERSION)
158 * Old magic mount flag and mask
160 #define MS_MGC_VAL 0xC0ED0000
161 #define MS_MGC_MSK 0xffff0000
163 /* Inode flags - they have nothing to superblock flags now */
165 #define S_SYNC 1 /* Writes are synced at once */
166 #define S_NOATIME 2 /* Do not update access times */
167 #define S_APPEND 4 /* Append-only file */
168 #define S_IMMUTABLE 8 /* Immutable file */
169 #define S_DEAD 16 /* removed, but still open directory */
170 #define S_NOQUOTA 32 /* Inode is not counted to quota */
171 #define S_DIRSYNC 64 /* Directory modifications are synchronous */
172 #define S_NOCMTIME 128 /* Do not update file c/mtime */
173 #define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
174 #define S_PRIVATE 512 /* Inode is fs-internal */
177 * Note that nosuid etc flags are inode-specific: setting some file-system
178 * flags just means all the inodes inherit those flags by default. It might be
179 * possible to override it selectively if you really wanted to with some
180 * ioctl() that is not currently implemented.
182 * Exception: MS_RDONLY is always applied to the entire file system.
184 * Unfortunately, it is possible to change a filesystems flags with it mounted
185 * with files in use. This means that all of the inodes will not have their
186 * i_flags updated. Hence, i_flags no longer inherit the superblock mount
187 * flags, so these have to be checked separately. -- rmk@arm.uk.linux.org
189 #define __IS_FLG(inode,flg) ((inode)->i_sb->s_flags & (flg))
191 #define IS_RDONLY(inode) ((inode)->i_sb->s_flags & MS_RDONLY)
192 #define IS_SYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS) || \
193 ((inode)->i_flags & S_SYNC))
194 #define IS_DIRSYNC(inode) (__IS_FLG(inode, MS_SYNCHRONOUS|MS_DIRSYNC) || \
195 ((inode)->i_flags & (S_SYNC|S_DIRSYNC)))
196 #define IS_MANDLOCK(inode) __IS_FLG(inode, MS_MANDLOCK)
197 #define IS_NOATIME(inode) __IS_FLG(inode, MS_RDONLY|MS_NOATIME)
198 #define IS_I_VERSION(inode) __IS_FLG(inode, MS_I_VERSION)
200 #define IS_NOQUOTA(inode) ((inode)->i_flags & S_NOQUOTA)
201 #define IS_APPEND(inode) ((inode)->i_flags & S_APPEND)
202 #define IS_IMMUTABLE(inode) ((inode)->i_flags & S_IMMUTABLE)
203 #define IS_POSIXACL(inode) __IS_FLG(inode, MS_POSIXACL)
205 #define IS_DEADDIR(inode) ((inode)->i_flags & S_DEAD)
206 #define IS_NOCMTIME(inode) ((inode)->i_flags & S_NOCMTIME)
207 #define IS_SWAPFILE(inode) ((inode)->i_flags & S_SWAPFILE)
208 #define IS_PRIVATE(inode) ((inode)->i_flags & S_PRIVATE)
210 /* the read-only stuff doesn't really belong here, but any other place is
211 probably as bad and I don't want to create yet another include file. */
213 #define BLKROSET _IO(0x12,93) /* set device read-only (0 = read-write) */
214 #define BLKROGET _IO(0x12,94) /* get read-only status (0 = read_write) */
215 #define BLKRRPART _IO(0x12,95) /* re-read partition table */
216 #define BLKGETSIZE _IO(0x12,96) /* return device size /512 (long *arg) */
217 #define BLKFLSBUF _IO(0x12,97) /* flush buffer cache */
218 #define BLKRASET _IO(0x12,98) /* set read ahead for block device */
219 #define BLKRAGET _IO(0x12,99) /* get current read ahead setting */
220 #define BLKFRASET _IO(0x12,100)/* set filesystem (mm/filemap.c) read-ahead */
221 #define BLKFRAGET _IO(0x12,101)/* get filesystem (mm/filemap.c) read-ahead */
222 #define BLKSECTSET _IO(0x12,102)/* set max sectors per request (ll_rw_blk.c) */
223 #define BLKSECTGET _IO(0x12,103)/* get max sectors per request (ll_rw_blk.c) */
224 #define BLKSSZGET _IO(0x12,104)/* get block device sector size */
226 #define BLKPG _IO(0x12,105)/* See blkpg.h */
228 /* Some people are morons. Do not use sizeof! */
230 #define BLKELVGET _IOR(0x12,106,size_t)/* elevator get */
231 #define BLKELVSET _IOW(0x12,107,size_t)/* elevator set */
232 /* This was here just to show that the number is taken -
233 probably all these _IO(0x12,*) ioctls should be moved to blkpg.h. */
235 /* A jump here: 108-111 have been used for various private purposes. */
236 #define BLKBSZGET _IOR(0x12,112,size_t)
237 #define BLKBSZSET _IOW(0x12,113,size_t)
238 #define BLKGETSIZE64 _IOR(0x12,114,size_t) /* return device size in bytes (u64 *arg) */
239 #define BLKTRACESETUP _IOWR(0x12,115,struct blk_user_trace_setup)
240 #define BLKTRACESTART _IO(0x12,116)
241 #define BLKTRACESTOP _IO(0x12,117)
242 #define BLKTRACETEARDOWN _IO(0x12,118)
243 #define BLKDISCARD _IO(0x12,119)
245 #define BMAP_IOCTL 1 /* obsolete - kept for compatibility */
246 #define FIBMAP _IO(0x00,1) /* bmap access */
247 #define FIGETBSZ _IO(0x00,2) /* get the block size used for bmap */
248 #define FIFREEZE _IOWR('X', 119, int) /* Freeze */
249 #define FITHAW _IOWR('X', 120, int) /* Thaw */
251 #define FS_IOC_GETFLAGS _IOR('f', 1, long)
252 #define FS_IOC_SETFLAGS _IOW('f', 2, long)
253 #define FS_IOC_GETVERSION _IOR('v', 1, long)
254 #define FS_IOC_SETVERSION _IOW('v', 2, long)
255 #define FS_IOC_FIEMAP _IOWR('f', 11, struct fiemap)
256 #define FS_IOC32_GETFLAGS _IOR('f', 1, int)
257 #define FS_IOC32_SETFLAGS _IOW('f', 2, int)
258 #define FS_IOC32_GETVERSION _IOR('v', 1, int)
259 #define FS_IOC32_SETVERSION _IOW('v', 2, int)
262 * Inode flags (FS_IOC_GETFLAGS / FS_IOC_SETFLAGS)
264 #define FS_SECRM_FL 0x00000001 /* Secure deletion */
265 #define FS_UNRM_FL 0x00000002 /* Undelete */
266 #define FS_COMPR_FL 0x00000004 /* Compress file */
267 #define FS_SYNC_FL 0x00000008 /* Synchronous updates */
268 #define FS_IMMUTABLE_FL 0x00000010 /* Immutable file */
269 #define FS_APPEND_FL 0x00000020 /* writes to file may only append */
270 #define FS_NODUMP_FL 0x00000040 /* do not dump file */
271 #define FS_NOATIME_FL 0x00000080 /* do not update atime */
272 /* Reserved for compression usage... */
273 #define FS_DIRTY_FL 0x00000100
274 #define FS_COMPRBLK_FL 0x00000200 /* One or more compressed clusters */
275 #define FS_NOCOMP_FL 0x00000400 /* Don't compress */
276 #define FS_ECOMPR_FL 0x00000800 /* Compression error */
277 /* End compression flags --- maybe not all used */
278 #define FS_BTREE_FL 0x00001000 /* btree format dir */
279 #define FS_INDEX_FL 0x00001000 /* hash-indexed directory */
280 #define FS_IMAGIC_FL 0x00002000 /* AFS directory */
281 #define FS_JOURNAL_DATA_FL 0x00004000 /* Reserved for ext3 */
282 #define FS_NOTAIL_FL 0x00008000 /* file tail should not be merged */
283 #define FS_DIRSYNC_FL 0x00010000 /* dirsync behaviour (directories only) */
284 #define FS_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/
285 #define FS_EXTENT_FL 0x00080000 /* Extents */
286 #define FS_DIRECTIO_FL 0x00100000 /* Use direct i/o */
287 #define FS_RESERVED_FL 0x80000000 /* reserved for ext2 lib */
289 #define FS_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */
290 #define FS_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */
293 #define SYNC_FILE_RANGE_WAIT_BEFORE 1
294 #define SYNC_FILE_RANGE_WRITE 2
295 #define SYNC_FILE_RANGE_WAIT_AFTER 4
299 #include <linux/linkage.h>
300 #include <linux/wait.h>
301 #include <linux/types.h>
302 #include <linux/kdev_t.h>
303 #include <linux/dcache.h>
304 #include <linux/path.h>
305 #include <linux/stat.h>
306 #include <linux/cache.h>
307 #include <linux/kobject.h>
308 #include <linux/list.h>
309 #include <linux/radix-tree.h>
310 #include <linux/prio_tree.h>
311 #include <linux/init.h>
312 #include <linux/pid.h>
313 #include <linux/mutex.h>
314 #include <linux/capability.h>
315 #include <linux/semaphore.h>
316 #include <linux/fiemap.h>
318 #include <asm/atomic.h>
319 #include <asm/byteorder.h>
321 struct export_operations
;
326 struct pipe_inode_info
;
327 struct poll_table_struct
;
329 struct vm_area_struct
;
333 extern void __init
inode_init(void);
334 extern void __init
inode_init_early(void);
335 extern void __init
files_init(unsigned long);
337 extern struct files_stat_struct files_stat
;
338 extern int get_max_files(void);
339 extern int sysctl_nr_open
;
340 extern struct inodes_stat_t inodes_stat
;
341 extern int leases_enable
, lease_break_time
;
342 #ifdef CONFIG_DNOTIFY
343 extern int dir_notify_enable
;
347 typedef int (get_block_t
)(struct inode
*inode
, sector_t iblock
,
348 struct buffer_head
*bh_result
, int create
);
349 typedef void (dio_iodone_t
)(struct kiocb
*iocb
, loff_t offset
,
350 ssize_t bytes
, void *private);
353 * Attribute flags. These should be or-ed together to figure out what
356 #define ATTR_MODE (1 << 0)
357 #define ATTR_UID (1 << 1)
358 #define ATTR_GID (1 << 2)
359 #define ATTR_SIZE (1 << 3)
360 #define ATTR_ATIME (1 << 4)
361 #define ATTR_MTIME (1 << 5)
362 #define ATTR_CTIME (1 << 6)
363 #define ATTR_ATIME_SET (1 << 7)
364 #define ATTR_MTIME_SET (1 << 8)
365 #define ATTR_FORCE (1 << 9) /* Not a change, but a change it */
366 #define ATTR_ATTR_FLAG (1 << 10)
367 #define ATTR_KILL_SUID (1 << 11)
368 #define ATTR_KILL_SGID (1 << 12)
369 #define ATTR_FILE (1 << 13)
370 #define ATTR_KILL_PRIV (1 << 14)
371 #define ATTR_OPEN (1 << 15) /* Truncating from open(O_TRUNC) */
372 #define ATTR_TIMES_SET (1 << 16)
375 * This is the Inode Attributes structure, used for notify_change(). It
376 * uses the above definitions as flags, to know which values have changed.
377 * Also, in this manner, a Filesystem can look at only the values it cares
378 * about. Basically, these are the attributes that the VFS layer can
379 * request to change from the FS layer.
381 * Derek Atkins <warlord@MIT.EDU> 94-10-20
384 unsigned int ia_valid
;
389 struct timespec ia_atime
;
390 struct timespec ia_mtime
;
391 struct timespec ia_ctime
;
394 * Not an attribute, but an auxilary info for filesystems wanting to
395 * implement an ftruncate() like method. NOTE: filesystem should
396 * check for (ia_valid & ATTR_FILE), and not for (ia_file != NULL).
398 struct file
*ia_file
;
402 * Includes for diskquotas.
404 #include <linux/quota.h>
407 * enum positive_aop_returns - aop return codes with specific semantics
409 * @AOP_WRITEPAGE_ACTIVATE: Informs the caller that page writeback has
410 * completed, that the page is still locked, and
411 * should be considered active. The VM uses this hint
412 * to return the page to the active list -- it won't
413 * be a candidate for writeback again in the near
414 * future. Other callers must be careful to unlock
415 * the page if they get this return. Returned by
418 * @AOP_TRUNCATED_PAGE: The AOP method that was handed a locked page has
419 * unlocked it and the page might have been truncated.
420 * The caller should back up to acquiring a new page and
421 * trying again. The aop will be taking reasonable
422 * precautions not to livelock. If the caller held a page
423 * reference, it should drop it before retrying. Returned
426 * address_space_operation functions return these large constants to indicate
427 * special semantics to the caller. These are much larger than the bytes in a
428 * page to allow for functions that return the number of bytes operated on in a
432 enum positive_aop_returns
{
433 AOP_WRITEPAGE_ACTIVATE
= 0x80000,
434 AOP_TRUNCATED_PAGE
= 0x80001,
437 #define AOP_FLAG_UNINTERRUPTIBLE 0x0001 /* will not do a short write */
438 #define AOP_FLAG_CONT_EXPAND 0x0002 /* called from cont_expand */
439 #define AOP_FLAG_NOFS 0x0004 /* used by filesystem to direct
440 * helper code (eg buffer layer)
441 * to clear GFP_FS from alloc */
444 * oh the beauties of C type declarations.
447 struct address_space
;
448 struct writeback_control
;
451 const struct iovec
*iov
;
452 unsigned long nr_segs
;
457 size_t iov_iter_copy_from_user_atomic(struct page
*page
,
458 struct iov_iter
*i
, unsigned long offset
, size_t bytes
);
459 size_t iov_iter_copy_from_user(struct page
*page
,
460 struct iov_iter
*i
, unsigned long offset
, size_t bytes
);
461 void iov_iter_advance(struct iov_iter
*i
, size_t bytes
);
462 int iov_iter_fault_in_readable(struct iov_iter
*i
, size_t bytes
);
463 size_t iov_iter_single_seg_count(struct iov_iter
*i
);
465 static inline void iov_iter_init(struct iov_iter
*i
,
466 const struct iovec
*iov
, unsigned long nr_segs
,
467 size_t count
, size_t written
)
470 i
->nr_segs
= nr_segs
;
472 i
->count
= count
+ written
;
474 iov_iter_advance(i
, written
);
477 static inline size_t iov_iter_count(struct iov_iter
*i
)
483 * "descriptor" for what we're up to with a read.
484 * This allows us to use the same read code yet
485 * have multiple different users of the data that
486 * we read from a file.
488 * The simplest case just copies the data to user
501 typedef int (*read_actor_t
)(read_descriptor_t
*, struct page
*,
502 unsigned long, unsigned long);
504 struct address_space_operations
{
505 int (*writepage
)(struct page
*page
, struct writeback_control
*wbc
);
506 int (*readpage
)(struct file
*, struct page
*);
507 void (*sync_page
)(struct page
*);
509 /* Write back some dirty pages from this mapping. */
510 int (*writepages
)(struct address_space
*, struct writeback_control
*);
512 /* Set a page dirty. Return true if this dirtied it */
513 int (*set_page_dirty
)(struct page
*page
);
515 int (*readpages
)(struct file
*filp
, struct address_space
*mapping
,
516 struct list_head
*pages
, unsigned nr_pages
);
518 int (*write_begin
)(struct file
*, struct address_space
*mapping
,
519 loff_t pos
, unsigned len
, unsigned flags
,
520 struct page
**pagep
, void **fsdata
);
521 int (*write_end
)(struct file
*, struct address_space
*mapping
,
522 loff_t pos
, unsigned len
, unsigned copied
,
523 struct page
*page
, void *fsdata
);
525 /* Unfortunately this kludge is needed for FIBMAP. Don't use it */
526 sector_t (*bmap
)(struct address_space
*, sector_t
);
527 void (*invalidatepage
) (struct page
*, unsigned long);
528 int (*releasepage
) (struct page
*, gfp_t
);
529 ssize_t (*direct_IO
)(int, struct kiocb
*, const struct iovec
*iov
,
530 loff_t offset
, unsigned long nr_segs
);
531 int (*get_xip_mem
)(struct address_space
*, pgoff_t
, int,
532 void **, unsigned long *);
533 /* migrate the contents of a page to the specified target */
534 int (*migratepage
) (struct address_space
*,
535 struct page
*, struct page
*);
536 int (*launder_page
) (struct page
*);
537 int (*is_partially_uptodate
) (struct page
*, read_descriptor_t
*,
542 * pagecache_write_begin/pagecache_write_end must be used by general code
543 * to write into the pagecache.
545 int pagecache_write_begin(struct file
*, struct address_space
*mapping
,
546 loff_t pos
, unsigned len
, unsigned flags
,
547 struct page
**pagep
, void **fsdata
);
549 int pagecache_write_end(struct file
*, struct address_space
*mapping
,
550 loff_t pos
, unsigned len
, unsigned copied
,
551 struct page
*page
, void *fsdata
);
553 struct backing_dev_info
;
554 struct address_space
{
555 struct inode
*host
; /* owner: inode, block_device */
556 struct radix_tree_root page_tree
; /* radix tree of all pages */
557 spinlock_t tree_lock
; /* and lock protecting it */
558 unsigned int i_mmap_writable
;/* count VM_SHARED mappings */
559 struct prio_tree_root i_mmap
; /* tree of private and shared mappings */
560 struct list_head i_mmap_nonlinear
;/*list VM_NONLINEAR mappings */
561 spinlock_t i_mmap_lock
; /* protect tree, count, list */
562 unsigned int truncate_count
; /* Cover race condition with truncate */
563 unsigned long nrpages
; /* number of total pages */
564 pgoff_t writeback_index
;/* writeback starts here */
565 const struct address_space_operations
*a_ops
; /* methods */
566 unsigned long flags
; /* error bits/gfp mask */
567 struct backing_dev_info
*backing_dev_info
; /* device readahead, etc */
568 spinlock_t private_lock
; /* for use by the address_space */
569 struct list_head private_list
; /* ditto */
570 struct address_space
*assoc_mapping
; /* ditto */
571 } __attribute__((aligned(sizeof(long))));
573 * On most architectures that alignment is already the case; but
574 * must be enforced here for CRIS, to let the least signficant bit
575 * of struct page's "mapping" pointer be used for PAGE_MAPPING_ANON.
578 struct block_device
{
579 dev_t bd_dev
; /* not a kdev_t - it's a search key */
580 struct inode
* bd_inode
; /* will die */
581 struct super_block
* bd_super
;
583 struct mutex bd_mutex
; /* open/close mutex */
584 struct semaphore bd_mount_sem
;
585 struct list_head bd_inodes
;
589 struct list_head bd_holder_list
;
591 struct block_device
* bd_contains
;
592 unsigned bd_block_size
;
593 struct hd_struct
* bd_part
;
594 /* number of times partitions within this device have been opened. */
595 unsigned bd_part_count
;
597 struct gendisk
* bd_disk
;
598 struct list_head bd_list
;
599 struct backing_dev_info
*bd_inode_backing_dev_info
;
601 * Private data. You must have bd_claim'ed the block_device
602 * to use this. NOTE: bd_claim allows an owner to claim
603 * the same device multiple times, the owner must take special
604 * care to not mess up bd_private for that case.
606 unsigned long bd_private
;
608 /* The counter of freeze processes */
609 int bd_fsfreeze_count
;
610 /* Mutex for freeze */
611 struct mutex bd_fsfreeze_mutex
;
615 * Radix-tree tags, for tagging dirty and writeback pages within the pagecache
618 #define PAGECACHE_TAG_DIRTY 0
619 #define PAGECACHE_TAG_WRITEBACK 1
621 int mapping_tagged(struct address_space
*mapping
, int tag
);
624 * Might pages of this file be mapped into userspace?
626 static inline int mapping_mapped(struct address_space
*mapping
)
628 return !prio_tree_empty(&mapping
->i_mmap
) ||
629 !list_empty(&mapping
->i_mmap_nonlinear
);
633 * Might pages of this file have been modified in userspace?
634 * Note that i_mmap_writable counts all VM_SHARED vmas: do_mmap_pgoff
635 * marks vma as VM_SHARED if it is shared, and the file was opened for
636 * writing i.e. vma may be mprotected writable even if now readonly.
638 static inline int mapping_writably_mapped(struct address_space
*mapping
)
640 return mapping
->i_mmap_writable
!= 0;
644 * Use sequence counter to get consistent i_size on 32-bit processors.
646 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
647 #include <linux/seqlock.h>
648 #define __NEED_I_SIZE_ORDERED
649 #define i_size_ordered_init(inode) seqcount_init(&inode->i_size_seqcount)
651 #define i_size_ordered_init(inode) do { } while (0)
655 struct hlist_node i_hash
;
656 struct list_head i_list
;
657 struct list_head i_sb_list
;
658 struct list_head i_dentry
;
661 unsigned int i_nlink
;
667 #ifdef __NEED_I_SIZE_ORDERED
668 seqcount_t i_size_seqcount
;
670 struct timespec i_atime
;
671 struct timespec i_mtime
;
672 struct timespec i_ctime
;
673 unsigned int i_blkbits
;
675 unsigned short i_bytes
;
677 spinlock_t i_lock
; /* i_blocks, i_bytes, maybe i_size */
678 struct mutex i_mutex
;
679 struct rw_semaphore i_alloc_sem
;
680 const struct inode_operations
*i_op
;
681 const struct file_operations
*i_fop
; /* former ->i_op->default_file_ops */
682 struct super_block
*i_sb
;
683 struct file_lock
*i_flock
;
684 struct address_space
*i_mapping
;
685 struct address_space i_data
;
687 struct dquot
*i_dquot
[MAXQUOTAS
];
689 struct list_head i_devices
;
691 struct pipe_inode_info
*i_pipe
;
692 struct block_device
*i_bdev
;
699 #ifdef CONFIG_DNOTIFY
700 unsigned long i_dnotify_mask
; /* Directory notify events */
701 struct dnotify_struct
*i_dnotify
; /* for directory notifications */
704 #ifdef CONFIG_INOTIFY
705 struct list_head inotify_watches
; /* watches on this inode */
706 struct mutex inotify_mutex
; /* protects the watches list */
709 unsigned long i_state
;
710 unsigned long dirtied_when
; /* jiffies of first dirtying */
712 unsigned int i_flags
;
714 atomic_t i_writecount
;
715 #ifdef CONFIG_SECURITY
718 void *i_private
; /* fs or device private pointer */
722 * inode->i_mutex nesting subclasses for the lock validator:
724 * 0: the object of the current VFS operation
729 * The locking order between these classes is
730 * parent -> child -> normal -> xattr -> quota
732 enum inode_i_mutex_lock_class
741 extern void inode_double_lock(struct inode
*inode1
, struct inode
*inode2
);
742 extern void inode_double_unlock(struct inode
*inode1
, struct inode
*inode2
);
745 * NOTE: in a 32bit arch with a preemptable kernel and
746 * an UP compile the i_size_read/write must be atomic
747 * with respect to the local cpu (unlike with preempt disabled),
748 * but they don't need to be atomic with respect to other cpus like in
749 * true SMP (so they need either to either locally disable irq around
750 * the read or for example on x86 they can be still implemented as a
751 * cmpxchg8b without the need of the lock prefix). For SMP compiles
752 * and 64bit archs it makes no difference if preempt is enabled or not.
754 static inline loff_t
i_size_read(const struct inode
*inode
)
756 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
761 seq
= read_seqcount_begin(&inode
->i_size_seqcount
);
762 i_size
= inode
->i_size
;
763 } while (read_seqcount_retry(&inode
->i_size_seqcount
, seq
));
765 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
769 i_size
= inode
->i_size
;
773 return inode
->i_size
;
778 * NOTE: unlike i_size_read(), i_size_write() does need locking around it
779 * (normally i_mutex), otherwise on 32bit/SMP an update of i_size_seqcount
780 * can be lost, resulting in subsequent i_size_read() calls spinning forever.
782 static inline void i_size_write(struct inode
*inode
, loff_t i_size
)
784 #if BITS_PER_LONG==32 && defined(CONFIG_SMP)
785 write_seqcount_begin(&inode
->i_size_seqcount
);
786 inode
->i_size
= i_size
;
787 write_seqcount_end(&inode
->i_size_seqcount
);
788 #elif BITS_PER_LONG==32 && defined(CONFIG_PREEMPT)
790 inode
->i_size
= i_size
;
793 inode
->i_size
= i_size
;
797 static inline unsigned iminor(const struct inode
*inode
)
799 return MINOR(inode
->i_rdev
);
802 static inline unsigned imajor(const struct inode
*inode
)
804 return MAJOR(inode
->i_rdev
);
807 extern struct block_device
*I_BDEV(struct inode
*inode
);
810 rwlock_t lock
; /* protects pid, uid, euid fields */
811 struct pid
*pid
; /* pid or -pgrp where SIGIO should be sent */
812 enum pid_type pid_type
; /* Kind of process group SIGIO should be sent to */
813 uid_t uid
, euid
; /* uid/euid of process setting the owner */
814 int signum
; /* posix.1b rt signal to be delivered on IO */
818 * Track a single file's readahead state
820 struct file_ra_state
{
821 pgoff_t start
; /* where readahead started */
822 unsigned int size
; /* # of readahead pages */
823 unsigned int async_size
; /* do asynchronous readahead when
824 there are only # of pages ahead */
826 unsigned int ra_pages
; /* Maximum readahead window */
827 int mmap_miss
; /* Cache miss stat for mmap accesses */
828 loff_t prev_pos
; /* Cache last read() position */
832 * Check if @index falls in the readahead windows.
834 static inline int ra_has_index(struct file_ra_state
*ra
, pgoff_t index
)
836 return (index
>= ra
->start
&&
837 index
< ra
->start
+ ra
->size
);
840 #define FILE_MNT_WRITE_TAKEN 1
841 #define FILE_MNT_WRITE_RELEASED 2
845 * fu_list becomes invalid after file_free is called and queued via
846 * fu_rcuhead for RCU freeing
849 struct list_head fu_list
;
850 struct rcu_head fu_rcuhead
;
853 #define f_dentry f_path.dentry
854 #define f_vfsmnt f_path.mnt
855 const struct file_operations
*f_op
;
856 spinlock_t f_lock
; /* f_ep_links, f_flags, no IRQ */
857 atomic_long_t f_count
;
858 unsigned int f_flags
;
861 struct fown_struct f_owner
;
862 const struct cred
*f_cred
;
863 struct file_ra_state f_ra
;
866 #ifdef CONFIG_SECURITY
869 /* needed for tty driver, and maybe others */
873 /* Used by fs/eventpoll.c to link all the hooks to this file */
874 struct list_head f_ep_links
;
875 #endif /* #ifdef CONFIG_EPOLL */
876 struct address_space
*f_mapping
;
877 #ifdef CONFIG_DEBUG_WRITECOUNT
878 unsigned long f_mnt_write_state
;
881 extern spinlock_t files_lock
;
882 #define file_list_lock() spin_lock(&files_lock);
883 #define file_list_unlock() spin_unlock(&files_lock);
885 #define get_file(x) atomic_long_inc(&(x)->f_count)
886 #define file_count(x) atomic_long_read(&(x)->f_count)
888 #ifdef CONFIG_DEBUG_WRITECOUNT
889 static inline void file_take_write(struct file
*f
)
891 WARN_ON(f
->f_mnt_write_state
!= 0);
892 f
->f_mnt_write_state
= FILE_MNT_WRITE_TAKEN
;
894 static inline void file_release_write(struct file
*f
)
896 f
->f_mnt_write_state
|= FILE_MNT_WRITE_RELEASED
;
898 static inline void file_reset_write(struct file
*f
)
900 f
->f_mnt_write_state
= 0;
902 static inline void file_check_state(struct file
*f
)
905 * At this point, either both or neither of these bits
908 WARN_ON(f
->f_mnt_write_state
== FILE_MNT_WRITE_TAKEN
);
909 WARN_ON(f
->f_mnt_write_state
== FILE_MNT_WRITE_RELEASED
);
911 static inline int file_check_writeable(struct file
*f
)
913 if (f
->f_mnt_write_state
== FILE_MNT_WRITE_TAKEN
)
915 printk(KERN_WARNING
"writeable file with no "
916 "mnt_want_write()\n");
920 #else /* !CONFIG_DEBUG_WRITECOUNT */
921 static inline void file_take_write(struct file
*filp
) {}
922 static inline void file_release_write(struct file
*filp
) {}
923 static inline void file_reset_write(struct file
*filp
) {}
924 static inline void file_check_state(struct file
*filp
) {}
925 static inline int file_check_writeable(struct file
*filp
)
929 #endif /* CONFIG_DEBUG_WRITECOUNT */
931 #define MAX_NON_LFS ((1UL<<31) - 1)
933 /* Page cache limit. The filesystems should put that into their s_maxbytes
934 limits, otherwise bad things can happen in VM. */
935 #if BITS_PER_LONG==32
936 #define MAX_LFS_FILESIZE (((u64)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
937 #elif BITS_PER_LONG==64
938 #define MAX_LFS_FILESIZE 0x7fffffffffffffffUL
943 #define FL_ACCESS 8 /* not trying to lock, just looking */
944 #define FL_EXISTS 16 /* when unlocking, test for existence */
945 #define FL_LEASE 32 /* lease held on this file */
946 #define FL_CLOSE 64 /* unlock on close */
947 #define FL_SLEEP 128 /* A blocking lock */
950 * Special return value from posix_lock_file() and vfs_lock_file() for
951 * asynchronous locking.
953 #define FILE_LOCK_DEFERRED 1
956 * The POSIX file lock owner is determined by
957 * the "struct files_struct" in the thread group
958 * (or NULL for no owner - BSD locks).
960 * Lockd stuffs a "host" pointer into this.
962 typedef struct files_struct
*fl_owner_t
;
964 struct file_lock_operations
{
965 void (*fl_copy_lock
)(struct file_lock
*, struct file_lock
*);
966 void (*fl_release_private
)(struct file_lock
*);
969 struct lock_manager_operations
{
970 int (*fl_compare_owner
)(struct file_lock
*, struct file_lock
*);
971 void (*fl_notify
)(struct file_lock
*); /* unblock callback */
972 int (*fl_grant
)(struct file_lock
*, struct file_lock
*, int);
973 void (*fl_copy_lock
)(struct file_lock
*, struct file_lock
*);
974 void (*fl_release_private
)(struct file_lock
*);
975 void (*fl_break
)(struct file_lock
*);
976 int (*fl_mylease
)(struct file_lock
*, struct file_lock
*);
977 int (*fl_change
)(struct file_lock
**, int);
980 struct lock_manager
{
981 struct list_head list
;
984 void locks_start_grace(struct lock_manager
*);
985 void locks_end_grace(struct lock_manager
*);
986 int locks_in_grace(void);
988 /* that will die - we need it for nfs_lock_info */
989 #include <linux/nfs_fs_i.h>
992 struct file_lock
*fl_next
; /* singly linked list for this inode */
993 struct list_head fl_link
; /* doubly linked list of all locks */
994 struct list_head fl_block
; /* circular list of blocked processes */
996 unsigned char fl_flags
;
997 unsigned char fl_type
;
999 struct pid
*fl_nspid
;
1000 wait_queue_head_t fl_wait
;
1001 struct file
*fl_file
;
1005 struct fasync_struct
* fl_fasync
; /* for lease break notifications */
1006 unsigned long fl_break_time
; /* for nonblocking lease breaks */
1008 struct file_lock_operations
*fl_ops
; /* Callbacks for filesystems */
1009 struct lock_manager_operations
*fl_lmops
; /* Callbacks for lockmanagers */
1011 struct nfs_lock_info nfs_fl
;
1012 struct nfs4_lock_info nfs4_fl
;
1014 struct list_head link
; /* link in AFS vnode's pending_locks list */
1015 int state
; /* state of grant or error if -ve */
1020 /* The following constant reflects the upper bound of the file/locking space */
1022 #define INT_LIMIT(x) (~((x)1 << (sizeof(x)*8 - 1)))
1023 #define OFFSET_MAX INT_LIMIT(loff_t)
1024 #define OFFT_OFFSET_MAX INT_LIMIT(off_t)
1027 #include <linux/fcntl.h>
1029 extern void send_sigio(struct fown_struct
*fown
, int fd
, int band
);
1032 extern int do_sync_mapping_range(struct address_space
*mapping
, loff_t offset
,
1033 loff_t endbyte
, unsigned int flags
);
1035 #ifdef CONFIG_FILE_LOCKING
1036 extern int fcntl_getlk(struct file
*, struct flock __user
*);
1037 extern int fcntl_setlk(unsigned int, struct file
*, unsigned int,
1038 struct flock __user
*);
1040 #if BITS_PER_LONG == 32
1041 extern int fcntl_getlk64(struct file
*, struct flock64 __user
*);
1042 extern int fcntl_setlk64(unsigned int, struct file
*, unsigned int,
1043 struct flock64 __user
*);
1046 extern int fcntl_setlease(unsigned int fd
, struct file
*filp
, long arg
);
1047 extern int fcntl_getlease(struct file
*filp
);
1050 extern void locks_init_lock(struct file_lock
*);
1051 extern void locks_copy_lock(struct file_lock
*, struct file_lock
*);
1052 extern void __locks_copy_lock(struct file_lock
*, const struct file_lock
*);
1053 extern void locks_remove_posix(struct file
*, fl_owner_t
);
1054 extern void locks_remove_flock(struct file
*);
1055 extern void posix_test_lock(struct file
*, struct file_lock
*);
1056 extern int posix_lock_file(struct file
*, struct file_lock
*, struct file_lock
*);
1057 extern int posix_lock_file_wait(struct file
*, struct file_lock
*);
1058 extern int posix_unblock_lock(struct file
*, struct file_lock
*);
1059 extern int vfs_test_lock(struct file
*, struct file_lock
*);
1060 extern int vfs_lock_file(struct file
*, unsigned int, struct file_lock
*, struct file_lock
*);
1061 extern int vfs_cancel_lock(struct file
*filp
, struct file_lock
*fl
);
1062 extern int flock_lock_file_wait(struct file
*filp
, struct file_lock
*fl
);
1063 extern int __break_lease(struct inode
*inode
, unsigned int flags
);
1064 extern void lease_get_mtime(struct inode
*, struct timespec
*time
);
1065 extern int generic_setlease(struct file
*, long, struct file_lock
**);
1066 extern int vfs_setlease(struct file
*, long, struct file_lock
**);
1067 extern int lease_modify(struct file_lock
**, int);
1068 extern int lock_may_read(struct inode
*, loff_t start
, unsigned long count
);
1069 extern int lock_may_write(struct inode
*, loff_t start
, unsigned long count
);
1070 #else /* !CONFIG_FILE_LOCKING */
1071 static inline int fcntl_getlk(struct file
*file
, struct flock __user
*user
)
1076 static inline int fcntl_setlk(unsigned int fd
, struct file
*file
,
1077 unsigned int cmd
, struct flock __user
*user
)
1082 #if BITS_PER_LONG == 32
1083 static inline int fcntl_getlk64(struct file
*file
, struct flock64 __user
*user
)
1088 static inline int fcntl_setlk64(unsigned int fd
, struct file
*file
,
1089 unsigned int cmd
, struct flock64 __user
*user
)
1094 static inline int fcntl_setlease(unsigned int fd
, struct file
*filp
, long arg
)
1099 static inline int fcntl_getlease(struct file
*filp
)
1104 static inline void locks_init_lock(struct file_lock
*fl
)
1109 static inline void __locks_copy_lock(struct file_lock
*new, struct file_lock
*fl
)
1114 static inline void locks_copy_lock(struct file_lock
*new, struct file_lock
*fl
)
1119 static inline void locks_remove_posix(struct file
*filp
, fl_owner_t owner
)
1124 static inline void locks_remove_flock(struct file
*filp
)
1129 static inline void posix_test_lock(struct file
*filp
, struct file_lock
*fl
)
1134 static inline int posix_lock_file(struct file
*filp
, struct file_lock
*fl
,
1135 struct file_lock
*conflock
)
1140 static inline int posix_lock_file_wait(struct file
*filp
, struct file_lock
*fl
)
1145 static inline int posix_unblock_lock(struct file
*filp
,
1146 struct file_lock
*waiter
)
1151 static inline int vfs_test_lock(struct file
*filp
, struct file_lock
*fl
)
1156 static inline int vfs_lock_file(struct file
*filp
, unsigned int cmd
,
1157 struct file_lock
*fl
, struct file_lock
*conf
)
1162 static inline int vfs_cancel_lock(struct file
*filp
, struct file_lock
*fl
)
1167 static inline int flock_lock_file_wait(struct file
*filp
,
1168 struct file_lock
*request
)
1173 static inline int __break_lease(struct inode
*inode
, unsigned int mode
)
1178 static inline void lease_get_mtime(struct inode
*inode
, struct timespec
*time
)
1183 static inline int generic_setlease(struct file
*filp
, long arg
,
1184 struct file_lock
**flp
)
1189 static inline int vfs_setlease(struct file
*filp
, long arg
,
1190 struct file_lock
**lease
)
1195 static inline int lease_modify(struct file_lock
**before
, int arg
)
1200 static inline int lock_may_read(struct inode
*inode
, loff_t start
,
1206 static inline int lock_may_write(struct inode
*inode
, loff_t start
,
1212 #endif /* !CONFIG_FILE_LOCKING */
1215 struct fasync_struct
{
1218 struct fasync_struct
*fa_next
; /* singly linked list */
1219 struct file
*fa_file
;
1222 #define FASYNC_MAGIC 0x4601
1224 /* SMP safe fasync helpers: */
1225 extern int fasync_helper(int, struct file
*, int, struct fasync_struct
**);
1226 /* can be called from interrupts */
1227 extern void kill_fasync(struct fasync_struct
**, int, int);
1228 /* only for net: no internal synchronization */
1229 extern void __kill_fasync(struct fasync_struct
*, int, int);
1231 extern int __f_setown(struct file
*filp
, struct pid
*, enum pid_type
, int force
);
1232 extern int f_setown(struct file
*filp
, unsigned long arg
, int force
);
1233 extern void f_delown(struct file
*filp
);
1234 extern pid_t
f_getown(struct file
*filp
);
1235 extern int send_sigurg(struct fown_struct
*fown
);
1241 #define MNT_FORCE 0x00000001 /* Attempt to forcibily umount */
1242 #define MNT_DETACH 0x00000002 /* Just detach from the tree */
1243 #define MNT_EXPIRE 0x00000004 /* Mark for expiry */
1245 extern struct list_head super_blocks
;
1246 extern spinlock_t sb_lock
;
1248 #define sb_entry(list) list_entry((list), struct super_block, s_list)
1249 #define S_BIAS (1<<30)
1250 struct super_block
{
1251 struct list_head s_list
; /* Keep this first */
1252 dev_t s_dev
; /* search index; _not_ kdev_t */
1253 unsigned long s_blocksize
;
1254 unsigned char s_blocksize_bits
;
1255 unsigned char s_dirt
;
1256 unsigned long long s_maxbytes
; /* Max file size */
1257 struct file_system_type
*s_type
;
1258 const struct super_operations
*s_op
;
1259 struct dquot_operations
*dq_op
;
1260 struct quotactl_ops
*s_qcop
;
1261 const struct export_operations
*s_export_op
;
1262 unsigned long s_flags
;
1263 unsigned long s_magic
;
1264 struct dentry
*s_root
;
1265 struct rw_semaphore s_umount
;
1266 struct mutex s_lock
;
1270 #ifdef CONFIG_SECURITY
1273 struct xattr_handler
**s_xattr
;
1275 struct list_head s_inodes
; /* all inodes */
1276 struct list_head s_dirty
; /* dirty inodes */
1277 struct list_head s_io
; /* parked for writeback */
1278 struct list_head s_more_io
; /* parked for more writeback */
1279 struct hlist_head s_anon
; /* anonymous dentries for (nfs) exporting */
1280 struct list_head s_files
;
1281 /* s_dentry_lru and s_nr_dentry_unused are protected by dcache_lock */
1282 struct list_head s_dentry_lru
; /* unused dentry lru */
1283 int s_nr_dentry_unused
; /* # of dentry on lru */
1285 struct block_device
*s_bdev
;
1286 struct mtd_info
*s_mtd
;
1287 struct list_head s_instances
;
1288 struct quota_info s_dquot
; /* Diskquota specific options */
1291 wait_queue_head_t s_wait_unfrozen
;
1293 char s_id
[32]; /* Informational name */
1295 void *s_fs_info
; /* Filesystem private info */
1299 * The next field is for VFS *only*. No filesystems have any business
1300 * even looking at it. You had been warned.
1302 struct mutex s_vfs_rename_mutex
; /* Kludge */
1304 /* Granularity of c/m/atime in ns.
1305 Cannot be worse than a second */
1309 * Filesystem subtype. If non-empty the filesystem type field
1310 * in /proc/mounts will be "type.subtype"
1315 * Saved mount options for lazy filesystems using
1316 * generic_show_options()
1321 * storage for asynchronous operations
1323 struct list_head s_async_list
;
1326 extern struct timespec
current_fs_time(struct super_block
*sb
);
1329 * Snapshotting support.
1333 SB_FREEZE_WRITE
= 1,
1334 SB_FREEZE_TRANS
= 2,
1337 #define vfs_check_frozen(sb, level) \
1338 wait_event((sb)->s_wait_unfrozen, ((sb)->s_frozen < (level)))
1340 #define get_fs_excl() atomic_inc(¤t->fs_excl)
1341 #define put_fs_excl() atomic_dec(¤t->fs_excl)
1342 #define has_fs_excl() atomic_read(¤t->fs_excl)
1344 #define is_owner_or_cap(inode) \
1345 ((current_fsuid() == (inode)->i_uid) || capable(CAP_FOWNER))
1347 /* not quite ready to be deprecated, but... */
1348 extern void lock_super(struct super_block
*);
1349 extern void unlock_super(struct super_block
*);
1352 * VFS helper functions..
1354 extern int vfs_create(struct inode
*, struct dentry
*, int, struct nameidata
*);
1355 extern int vfs_mkdir(struct inode
*, struct dentry
*, int);
1356 extern int vfs_mknod(struct inode
*, struct dentry
*, int, dev_t
);
1357 extern int vfs_symlink(struct inode
*, struct dentry
*, const char *);
1358 extern int vfs_link(struct dentry
*, struct inode
*, struct dentry
*);
1359 extern int vfs_rmdir(struct inode
*, struct dentry
*);
1360 extern int vfs_unlink(struct inode
*, struct dentry
*);
1361 extern int vfs_rename(struct inode
*, struct dentry
*, struct inode
*, struct dentry
*);
1364 * VFS dentry helper functions.
1366 extern void dentry_unhash(struct dentry
*dentry
);
1369 * VFS file helper functions.
1371 extern int file_permission(struct file
*, int);
1374 * VFS FS_IOC_FIEMAP helper definitions.
1376 struct fiemap_extent_info
{
1377 unsigned int fi_flags
; /* Flags as passed from user */
1378 unsigned int fi_extents_mapped
; /* Number of mapped extents */
1379 unsigned int fi_extents_max
; /* Size of fiemap_extent array */
1380 struct fiemap_extent
*fi_extents_start
; /* Start of fiemap_extent
1383 int fiemap_fill_next_extent(struct fiemap_extent_info
*info
, u64 logical
,
1384 u64 phys
, u64 len
, u32 flags
);
1385 int fiemap_check_flags(struct fiemap_extent_info
*fieinfo
, u32 fs_flags
);
1390 * NOTE! These match bits 12..15 of stat.st_mode
1391 * (ie "(i_mode >> 12) & 15").
1393 #define DT_UNKNOWN 0
1403 #define OSYNC_METADATA (1<<0)
1404 #define OSYNC_DATA (1<<1)
1405 #define OSYNC_INODE (1<<2)
1406 int generic_osync_inode(struct inode
*, struct address_space
*, int);
1409 * This is the "filldir" function type, used by readdir() to let
1410 * the kernel specify what kind of dirent layout it wants to have.
1411 * This allows the kernel to read directories into kernel space or
1412 * to have different dirent layouts depending on the binary type.
1414 typedef int (*filldir_t
)(void *, const char *, int, loff_t
, u64
, unsigned);
1415 struct block_device_operations
;
1417 /* These macros are for out of kernel modules to test that
1418 * the kernel supports the unlocked_ioctl and compat_ioctl
1419 * fields in struct file_operations. */
1420 #define HAVE_COMPAT_IOCTL 1
1421 #define HAVE_UNLOCKED_IOCTL 1
1425 * read, write, poll, fsync, readv, writev, unlocked_ioctl and compat_ioctl
1426 * can be called without the big kernel lock held in all filesystems.
1428 struct file_operations
{
1429 struct module
*owner
;
1430 loff_t (*llseek
) (struct file
*, loff_t
, int);
1431 ssize_t (*read
) (struct file
*, char __user
*, size_t, loff_t
*);
1432 ssize_t (*write
) (struct file
*, const char __user
*, size_t, loff_t
*);
1433 ssize_t (*aio_read
) (struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
1434 ssize_t (*aio_write
) (struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
1435 int (*readdir
) (struct file
*, void *, filldir_t
);
1436 unsigned int (*poll
) (struct file
*, struct poll_table_struct
*);
1437 int (*ioctl
) (struct inode
*, struct file
*, unsigned int, unsigned long);
1438 long (*unlocked_ioctl
) (struct file
*, unsigned int, unsigned long);
1439 long (*compat_ioctl
) (struct file
*, unsigned int, unsigned long);
1440 int (*mmap
) (struct file
*, struct vm_area_struct
*);
1441 int (*open
) (struct inode
*, struct file
*);
1442 int (*flush
) (struct file
*, fl_owner_t id
);
1443 int (*release
) (struct inode
*, struct file
*);
1444 int (*fsync
) (struct file
*, struct dentry
*, int datasync
);
1445 int (*aio_fsync
) (struct kiocb
*, int datasync
);
1446 int (*fasync
) (int, struct file
*, int);
1447 int (*lock
) (struct file
*, int, struct file_lock
*);
1448 ssize_t (*sendpage
) (struct file
*, struct page
*, int, size_t, loff_t
*, int);
1449 unsigned long (*get_unmapped_area
)(struct file
*, unsigned long, unsigned long, unsigned long, unsigned long);
1450 int (*check_flags
)(int);
1451 int (*flock
) (struct file
*, int, struct file_lock
*);
1452 ssize_t (*splice_write
)(struct pipe_inode_info
*, struct file
*, loff_t
*, size_t, unsigned int);
1453 ssize_t (*splice_read
)(struct file
*, loff_t
*, struct pipe_inode_info
*, size_t, unsigned int);
1454 int (*setlease
)(struct file
*, long, struct file_lock
**);
1457 struct inode_operations
{
1458 int (*create
) (struct inode
*,struct dentry
*,int, struct nameidata
*);
1459 struct dentry
* (*lookup
) (struct inode
*,struct dentry
*, struct nameidata
*);
1460 int (*link
) (struct dentry
*,struct inode
*,struct dentry
*);
1461 int (*unlink
) (struct inode
*,struct dentry
*);
1462 int (*symlink
) (struct inode
*,struct dentry
*,const char *);
1463 int (*mkdir
) (struct inode
*,struct dentry
*,int);
1464 int (*rmdir
) (struct inode
*,struct dentry
*);
1465 int (*mknod
) (struct inode
*,struct dentry
*,int,dev_t
);
1466 int (*rename
) (struct inode
*, struct dentry
*,
1467 struct inode
*, struct dentry
*);
1468 int (*readlink
) (struct dentry
*, char __user
*,int);
1469 void * (*follow_link
) (struct dentry
*, struct nameidata
*);
1470 void (*put_link
) (struct dentry
*, struct nameidata
*, void *);
1471 void (*truncate
) (struct inode
*);
1472 int (*permission
) (struct inode
*, int);
1473 int (*setattr
) (struct dentry
*, struct iattr
*);
1474 int (*getattr
) (struct vfsmount
*mnt
, struct dentry
*, struct kstat
*);
1475 int (*setxattr
) (struct dentry
*, const char *,const void *,size_t,int);
1476 ssize_t (*getxattr
) (struct dentry
*, const char *, void *, size_t);
1477 ssize_t (*listxattr
) (struct dentry
*, char *, size_t);
1478 int (*removexattr
) (struct dentry
*, const char *);
1479 void (*truncate_range
)(struct inode
*, loff_t
, loff_t
);
1480 long (*fallocate
)(struct inode
*inode
, int mode
, loff_t offset
,
1482 int (*fiemap
)(struct inode
*, struct fiemap_extent_info
*, u64 start
,
1488 ssize_t
rw_copy_check_uvector(int type
, const struct iovec __user
* uvector
,
1489 unsigned long nr_segs
, unsigned long fast_segs
,
1490 struct iovec
*fast_pointer
,
1491 struct iovec
**ret_pointer
);
1493 extern ssize_t
vfs_read(struct file
*, char __user
*, size_t, loff_t
*);
1494 extern ssize_t
vfs_write(struct file
*, const char __user
*, size_t, loff_t
*);
1495 extern ssize_t
vfs_readv(struct file
*, const struct iovec __user
*,
1496 unsigned long, loff_t
*);
1497 extern ssize_t
vfs_writev(struct file
*, const struct iovec __user
*,
1498 unsigned long, loff_t
*);
1500 struct super_operations
{
1501 struct inode
*(*alloc_inode
)(struct super_block
*sb
);
1502 void (*destroy_inode
)(struct inode
*);
1504 void (*dirty_inode
) (struct inode
*);
1505 int (*write_inode
) (struct inode
*, int);
1506 void (*drop_inode
) (struct inode
*);
1507 void (*delete_inode
) (struct inode
*);
1508 void (*put_super
) (struct super_block
*);
1509 void (*write_super
) (struct super_block
*);
1510 int (*sync_fs
)(struct super_block
*sb
, int wait
);
1511 int (*freeze_fs
) (struct super_block
*);
1512 int (*unfreeze_fs
) (struct super_block
*);
1513 int (*statfs
) (struct dentry
*, struct kstatfs
*);
1514 int (*remount_fs
) (struct super_block
*, int *, char *);
1515 void (*clear_inode
) (struct inode
*);
1516 void (*umount_begin
) (struct super_block
*);
1518 int (*show_options
)(struct seq_file
*, struct vfsmount
*);
1519 int (*show_stats
)(struct seq_file
*, struct vfsmount
*);
1521 ssize_t (*quota_read
)(struct super_block
*, int, char *, size_t, loff_t
);
1522 ssize_t (*quota_write
)(struct super_block
*, int, const char *, size_t, loff_t
);
1524 int (*bdev_try_to_free_page
)(struct super_block
*, struct page
*, gfp_t
);
1528 * Inode state bits. Protected by inode_lock.
1530 * Three bits determine the dirty state of the inode, I_DIRTY_SYNC,
1531 * I_DIRTY_DATASYNC and I_DIRTY_PAGES.
1533 * Four bits define the lifetime of an inode. Initially, inodes are I_NEW,
1534 * until that flag is cleared. I_WILL_FREE, I_FREEING and I_CLEAR are set at
1535 * various stages of removing an inode.
1537 * Two bits are used for locking and completion notification, I_LOCK and I_SYNC.
1539 * I_DIRTY_SYNC Inode is dirty, but doesn't have to be written on
1540 * fdatasync(). i_atime is the usual cause.
1541 * I_DIRTY_DATASYNC Data-related inode changes pending. We keep track of
1542 * these changes separately from I_DIRTY_SYNC so that we
1543 * don't have to write inode on fdatasync() when only
1544 * mtime has changed in it.
1545 * I_DIRTY_PAGES Inode has dirty pages. Inode itself may be clean.
1546 * I_NEW get_new_inode() sets i_state to I_LOCK|I_NEW. Both
1547 * are cleared by unlock_new_inode(), called from iget().
1548 * I_WILL_FREE Must be set when calling write_inode_now() if i_count
1549 * is zero. I_FREEING must be set when I_WILL_FREE is
1551 * I_FREEING Set when inode is about to be freed but still has dirty
1552 * pages or buffers attached or the inode itself is still
1554 * I_CLEAR Set by clear_inode(). In this state the inode is clean
1555 * and can be destroyed.
1557 * Inodes that are I_WILL_FREE, I_FREEING or I_CLEAR are
1558 * prohibited for many purposes. iget() must wait for
1559 * the inode to be completely released, then create it
1560 * anew. Other functions will just ignore such inodes,
1561 * if appropriate. I_LOCK is used for waiting.
1563 * I_LOCK Serves as both a mutex and completion notification.
1564 * New inodes set I_LOCK. If two processes both create
1565 * the same inode, one of them will release its inode and
1566 * wait for I_LOCK to be released before returning.
1567 * Inodes in I_WILL_FREE, I_FREEING or I_CLEAR state can
1568 * also cause waiting on I_LOCK, without I_LOCK actually
1569 * being set. find_inode() uses this to prevent returning
1570 * nearly-dead inodes.
1571 * I_SYNC Similar to I_LOCK, but limited in scope to writeback
1572 * of inode dirty data. Having a separate lock for this
1573 * purpose reduces latency and prevents some filesystem-
1574 * specific deadlocks.
1576 * Q: What is the difference between I_WILL_FREE and I_FREEING?
1577 * Q: igrab() only checks on (I_FREEING|I_WILL_FREE). Should it also check on
1578 * I_CLEAR? If not, why?
1580 #define I_DIRTY_SYNC 1
1581 #define I_DIRTY_DATASYNC 2
1582 #define I_DIRTY_PAGES 4
1584 #define I_WILL_FREE 16
1585 #define I_FREEING 32
1588 #define I_LOCK (1 << __I_LOCK)
1590 #define I_SYNC (1 << __I_SYNC)
1592 #define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
1594 extern void __mark_inode_dirty(struct inode
*, int);
1595 static inline void mark_inode_dirty(struct inode
*inode
)
1597 __mark_inode_dirty(inode
, I_DIRTY
);
1600 static inline void mark_inode_dirty_sync(struct inode
*inode
)
1602 __mark_inode_dirty(inode
, I_DIRTY_SYNC
);
1606 * inc_nlink - directly increment an inode's link count
1609 * This is a low-level filesystem helper to replace any
1610 * direct filesystem manipulation of i_nlink. Currently,
1611 * it is only here for parity with dec_nlink().
1613 static inline void inc_nlink(struct inode
*inode
)
1618 static inline void inode_inc_link_count(struct inode
*inode
)
1621 mark_inode_dirty(inode
);
1625 * drop_nlink - directly drop an inode's link count
1628 * This is a low-level filesystem helper to replace any
1629 * direct filesystem manipulation of i_nlink. In cases
1630 * where we are attempting to track writes to the
1631 * filesystem, a decrement to zero means an imminent
1632 * write when the file is truncated and actually unlinked
1633 * on the filesystem.
1635 static inline void drop_nlink(struct inode
*inode
)
1641 * clear_nlink - directly zero an inode's link count
1644 * This is a low-level filesystem helper to replace any
1645 * direct filesystem manipulation of i_nlink. See
1646 * drop_nlink() for why we care about i_nlink hitting zero.
1648 static inline void clear_nlink(struct inode
*inode
)
1653 static inline void inode_dec_link_count(struct inode
*inode
)
1656 mark_inode_dirty(inode
);
1660 * inode_inc_iversion - increments i_version
1661 * @inode: inode that need to be updated
1663 * Every time the inode is modified, the i_version field will be incremented.
1664 * The filesystem has to be mounted with i_version flag
1667 static inline void inode_inc_iversion(struct inode
*inode
)
1669 spin_lock(&inode
->i_lock
);
1671 spin_unlock(&inode
->i_lock
);
1674 extern void touch_atime(struct vfsmount
*mnt
, struct dentry
*dentry
);
1675 static inline void file_accessed(struct file
*file
)
1677 if (!(file
->f_flags
& O_NOATIME
))
1678 touch_atime(file
->f_path
.mnt
, file
->f_path
.dentry
);
1681 int sync_inode(struct inode
*inode
, struct writeback_control
*wbc
);
1683 struct file_system_type
{
1686 int (*get_sb
) (struct file_system_type
*, int,
1687 const char *, void *, struct vfsmount
*);
1688 void (*kill_sb
) (struct super_block
*);
1689 struct module
*owner
;
1690 struct file_system_type
* next
;
1691 struct list_head fs_supers
;
1693 struct lock_class_key s_lock_key
;
1694 struct lock_class_key s_umount_key
;
1696 struct lock_class_key i_lock_key
;
1697 struct lock_class_key i_mutex_key
;
1698 struct lock_class_key i_mutex_dir_key
;
1699 struct lock_class_key i_alloc_sem_key
;
1702 extern int get_sb_ns(struct file_system_type
*fs_type
, int flags
, void *data
,
1703 int (*fill_super
)(struct super_block
*, void *, int),
1704 struct vfsmount
*mnt
);
1705 extern int get_sb_bdev(struct file_system_type
*fs_type
,
1706 int flags
, const char *dev_name
, void *data
,
1707 int (*fill_super
)(struct super_block
*, void *, int),
1708 struct vfsmount
*mnt
);
1709 extern int get_sb_single(struct file_system_type
*fs_type
,
1710 int flags
, void *data
,
1711 int (*fill_super
)(struct super_block
*, void *, int),
1712 struct vfsmount
*mnt
);
1713 extern int get_sb_nodev(struct file_system_type
*fs_type
,
1714 int flags
, void *data
,
1715 int (*fill_super
)(struct super_block
*, void *, int),
1716 struct vfsmount
*mnt
);
1717 void generic_shutdown_super(struct super_block
*sb
);
1718 void kill_block_super(struct super_block
*sb
);
1719 void kill_anon_super(struct super_block
*sb
);
1720 void kill_litter_super(struct super_block
*sb
);
1721 void deactivate_super(struct super_block
*sb
);
1722 int set_anon_super(struct super_block
*s
, void *data
);
1723 struct super_block
*sget(struct file_system_type
*type
,
1724 int (*test
)(struct super_block
*,void *),
1725 int (*set
)(struct super_block
*,void *),
1727 extern int get_sb_pseudo(struct file_system_type
*, char *,
1728 const struct super_operations
*ops
, unsigned long,
1729 struct vfsmount
*mnt
);
1730 extern void simple_set_mnt(struct vfsmount
*mnt
, struct super_block
*sb
);
1731 int __put_super_and_need_restart(struct super_block
*sb
);
1733 /* Alas, no aliases. Too much hassle with bringing module.h everywhere */
1734 #define fops_get(fops) \
1735 (((fops) && try_module_get((fops)->owner) ? (fops) : NULL))
1736 #define fops_put(fops) \
1737 do { if (fops) module_put((fops)->owner); } while(0)
1739 extern int register_filesystem(struct file_system_type
*);
1740 extern int unregister_filesystem(struct file_system_type
*);
1741 extern struct vfsmount
*kern_mount_data(struct file_system_type
*, void *data
);
1742 #define kern_mount(type) kern_mount_data(type, NULL)
1743 extern int may_umount_tree(struct vfsmount
*);
1744 extern int may_umount(struct vfsmount
*);
1745 extern long do_mount(char *, char *, char *, unsigned long, void *);
1746 extern struct vfsmount
*collect_mounts(struct vfsmount
*, struct dentry
*);
1747 extern void drop_collected_mounts(struct vfsmount
*);
1749 extern int vfs_statfs(struct dentry
*, struct kstatfs
*);
1751 extern int current_umask(void);
1754 extern struct kobject
*fs_kobj
;
1756 extern int rw_verify_area(int, struct file
*, loff_t
*, size_t);
1758 #define FLOCK_VERIFY_READ 1
1759 #define FLOCK_VERIFY_WRITE 2
1761 #ifdef CONFIG_FILE_LOCKING
1762 extern int locks_mandatory_locked(struct inode
*);
1763 extern int locks_mandatory_area(int, struct inode
*, struct file
*, loff_t
, size_t);
1766 * Candidates for mandatory locking have the setgid bit set
1767 * but no group execute bit - an otherwise meaningless combination.
1770 static inline int __mandatory_lock(struct inode
*ino
)
1772 return (ino
->i_mode
& (S_ISGID
| S_IXGRP
)) == S_ISGID
;
1776 * ... and these candidates should be on MS_MANDLOCK mounted fs,
1777 * otherwise these will be advisory locks
1780 static inline int mandatory_lock(struct inode
*ino
)
1782 return IS_MANDLOCK(ino
) && __mandatory_lock(ino
);
1785 static inline int locks_verify_locked(struct inode
*inode
)
1787 if (mandatory_lock(inode
))
1788 return locks_mandatory_locked(inode
);
1792 static inline int locks_verify_truncate(struct inode
*inode
,
1796 if (inode
->i_flock
&& mandatory_lock(inode
))
1797 return locks_mandatory_area(
1798 FLOCK_VERIFY_WRITE
, inode
, filp
,
1799 size
< inode
->i_size
? size
: inode
->i_size
,
1800 (size
< inode
->i_size
? inode
->i_size
- size
1801 : size
- inode
->i_size
)
1806 static inline int break_lease(struct inode
*inode
, unsigned int mode
)
1809 return __break_lease(inode
, mode
);
1812 #else /* !CONFIG_FILE_LOCKING */
1813 static inline int locks_mandatory_locked(struct inode
*inode
)
1818 static inline int locks_mandatory_area(int rw
, struct inode
*inode
,
1819 struct file
*filp
, loff_t offset
,
1825 static inline int __mandatory_lock(struct inode
*inode
)
1830 static inline int mandatory_lock(struct inode
*inode
)
1835 static inline int locks_verify_locked(struct inode
*inode
)
1840 static inline int locks_verify_truncate(struct inode
*inode
, struct file
*filp
,
1846 static inline int break_lease(struct inode
*inode
, unsigned int mode
)
1851 #endif /* CONFIG_FILE_LOCKING */
1855 extern int do_truncate(struct dentry
*, loff_t start
, unsigned int time_attrs
,
1857 extern long do_sys_open(int dfd
, const char __user
*filename
, int flags
,
1859 extern struct file
*filp_open(const char *, int, int);
1860 extern struct file
* dentry_open(struct dentry
*, struct vfsmount
*, int,
1861 const struct cred
*);
1862 extern int filp_close(struct file
*, fl_owner_t id
);
1863 extern char * getname(const char __user
*);
1866 extern void __init
vfs_caches_init_early(void);
1867 extern void __init
vfs_caches_init(unsigned long);
1869 extern struct kmem_cache
*names_cachep
;
1871 #define __getname() kmem_cache_alloc(names_cachep, GFP_KERNEL)
1872 #define __putname(name) kmem_cache_free(names_cachep, (void *)(name))
1873 #ifndef CONFIG_AUDITSYSCALL
1874 #define putname(name) __putname(name)
1876 extern void putname(const char *name
);
1880 extern int register_blkdev(unsigned int, const char *);
1881 extern void unregister_blkdev(unsigned int, const char *);
1882 extern struct block_device
*bdget(dev_t
);
1883 extern void bd_set_size(struct block_device
*, loff_t size
);
1884 extern void bd_forget(struct inode
*inode
);
1885 extern void bdput(struct block_device
*);
1886 extern struct block_device
*open_by_devnum(dev_t
, fmode_t
);
1887 extern void invalidate_bdev(struct block_device
*);
1888 extern int sync_blockdev(struct block_device
*bdev
);
1889 extern struct super_block
*freeze_bdev(struct block_device
*);
1890 extern void emergency_thaw_all(void);
1891 extern int thaw_bdev(struct block_device
*bdev
, struct super_block
*sb
);
1892 extern int fsync_bdev(struct block_device
*);
1893 extern int fsync_super(struct super_block
*);
1894 extern int fsync_no_super(struct block_device
*);
1896 static inline void bd_forget(struct inode
*inode
) {}
1897 static inline int sync_blockdev(struct block_device
*bdev
) { return 0; }
1898 static inline void invalidate_bdev(struct block_device
*bdev
) {}
1900 static inline struct super_block
*freeze_bdev(struct block_device
*sb
)
1905 static inline int thaw_bdev(struct block_device
*bdev
, struct super_block
*sb
)
1910 extern const struct file_operations def_blk_fops
;
1911 extern const struct file_operations def_chr_fops
;
1912 extern const struct file_operations bad_sock_fops
;
1913 extern const struct file_operations def_fifo_fops
;
1915 extern int ioctl_by_bdev(struct block_device
*, unsigned, unsigned long);
1916 extern int blkdev_ioctl(struct block_device
*, fmode_t
, unsigned, unsigned long);
1917 extern long compat_blkdev_ioctl(struct file
*, unsigned, unsigned long);
1918 extern int blkdev_get(struct block_device
*, fmode_t
);
1919 extern int blkdev_put(struct block_device
*, fmode_t
);
1920 extern int bd_claim(struct block_device
*, void *);
1921 extern void bd_release(struct block_device
*);
1923 extern int bd_claim_by_disk(struct block_device
*, void *, struct gendisk
*);
1924 extern void bd_release_from_disk(struct block_device
*, struct gendisk
*);
1926 #define bd_claim_by_disk(bdev, holder, disk) bd_claim(bdev, holder)
1927 #define bd_release_from_disk(bdev, disk) bd_release(bdev)
1932 #define CHRDEV_MAJOR_HASH_SIZE 255
1933 extern int alloc_chrdev_region(dev_t
*, unsigned, unsigned, const char *);
1934 extern int register_chrdev_region(dev_t
, unsigned, const char *);
1935 extern int register_chrdev(unsigned int, const char *,
1936 const struct file_operations
*);
1937 extern void unregister_chrdev(unsigned int, const char *);
1938 extern void unregister_chrdev_region(dev_t
, unsigned);
1939 extern void chrdev_show(struct seq_file
*,off_t
);
1941 /* fs/block_dev.c */
1942 #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
1943 #define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
1946 #define BLKDEV_MAJOR_HASH_SIZE 255
1947 extern const char *__bdevname(dev_t
, char *buffer
);
1948 extern const char *bdevname(struct block_device
*bdev
, char *buffer
);
1949 extern struct block_device
*lookup_bdev(const char *);
1950 extern struct block_device
*open_bdev_exclusive(const char *, fmode_t
, void *);
1951 extern void close_bdev_exclusive(struct block_device
*, fmode_t
);
1952 extern void blkdev_show(struct seq_file
*,off_t
);
1955 #define BLKDEV_MAJOR_HASH_SIZE 0
1958 extern void init_special_inode(struct inode
*, umode_t
, dev_t
);
1960 /* Invalid inode operations -- fs/bad_inode.c */
1961 extern void make_bad_inode(struct inode
*);
1962 extern int is_bad_inode(struct inode
*);
1964 extern const struct file_operations read_pipefifo_fops
;
1965 extern const struct file_operations write_pipefifo_fops
;
1966 extern const struct file_operations rdwr_pipefifo_fops
;
1968 extern int fs_may_remount_ro(struct super_block
*);
1972 * return READ, READA, or WRITE
1974 #define bio_rw(bio) ((bio)->bi_rw & (RW_MASK | RWA_MASK))
1977 * return data direction, READ or WRITE
1979 #define bio_data_dir(bio) ((bio)->bi_rw & 1)
1981 extern void check_disk_size_change(struct gendisk
*disk
,
1982 struct block_device
*bdev
);
1983 extern int revalidate_disk(struct gendisk
*);
1984 extern int check_disk_change(struct block_device
*);
1985 extern int __invalidate_device(struct block_device
*);
1986 extern int invalidate_partition(struct gendisk
*, int);
1988 extern int invalidate_inodes(struct super_block
*);
1989 unsigned long __invalidate_mapping_pages(struct address_space
*mapping
,
1990 pgoff_t start
, pgoff_t end
,
1992 unsigned long invalidate_mapping_pages(struct address_space
*mapping
,
1993 pgoff_t start
, pgoff_t end
);
1995 static inline unsigned long __deprecated
1996 invalidate_inode_pages(struct address_space
*mapping
)
1998 return invalidate_mapping_pages(mapping
, 0, ~0UL);
2001 static inline void invalidate_remote_inode(struct inode
*inode
)
2003 if (S_ISREG(inode
->i_mode
) || S_ISDIR(inode
->i_mode
) ||
2004 S_ISLNK(inode
->i_mode
))
2005 invalidate_mapping_pages(inode
->i_mapping
, 0, -1);
2007 extern int invalidate_inode_pages2(struct address_space
*mapping
);
2008 extern int invalidate_inode_pages2_range(struct address_space
*mapping
,
2009 pgoff_t start
, pgoff_t end
);
2010 extern void generic_sync_sb_inodes(struct super_block
*sb
,
2011 struct writeback_control
*wbc
);
2012 extern int write_inode_now(struct inode
*, int);
2013 extern int filemap_fdatawrite(struct address_space
*);
2014 extern int filemap_flush(struct address_space
*);
2015 extern int filemap_fdatawait(struct address_space
*);
2016 extern int filemap_write_and_wait(struct address_space
*mapping
);
2017 extern int filemap_write_and_wait_range(struct address_space
*mapping
,
2018 loff_t lstart
, loff_t lend
);
2019 extern int wait_on_page_writeback_range(struct address_space
*mapping
,
2020 pgoff_t start
, pgoff_t end
);
2021 extern int __filemap_fdatawrite_range(struct address_space
*mapping
,
2022 loff_t start
, loff_t end
, int sync_mode
);
2023 extern int filemap_fdatawrite_range(struct address_space
*mapping
,
2024 loff_t start
, loff_t end
);
2026 extern int vfs_fsync(struct file
*file
, struct dentry
*dentry
, int datasync
);
2027 extern void sync_supers(void);
2028 extern void sync_filesystems(int wait
);
2029 extern void __fsync_super(struct super_block
*sb
);
2030 extern void emergency_sync(void);
2031 extern void emergency_remount(void);
2032 extern int do_remount_sb(struct super_block
*sb
, int flags
,
2033 void *data
, int force
);
2035 extern sector_t
bmap(struct inode
*, sector_t
);
2037 extern int notify_change(struct dentry
*, struct iattr
*);
2038 extern int inode_permission(struct inode
*, int);
2039 extern int generic_permission(struct inode
*, int,
2040 int (*check_acl
)(struct inode
*, int));
2042 static inline bool execute_ok(struct inode
*inode
)
2044 return (inode
->i_mode
& S_IXUGO
) || S_ISDIR(inode
->i_mode
);
2047 extern int get_write_access(struct inode
*);
2048 extern int deny_write_access(struct file
*);
2049 static inline void put_write_access(struct inode
* inode
)
2051 atomic_dec(&inode
->i_writecount
);
2053 static inline void allow_write_access(struct file
*file
)
2056 atomic_inc(&file
->f_path
.dentry
->d_inode
->i_writecount
);
2058 extern int do_pipe_flags(int *, int);
2059 extern struct file
*create_read_pipe(struct file
*f
, int flags
);
2060 extern struct file
*create_write_pipe(int flags
);
2061 extern void free_write_pipe(struct file
*);
2063 extern struct file
*do_filp_open(int dfd
, const char *pathname
,
2064 int open_flag
, int mode
);
2065 extern int may_open(struct path
*, int, int);
2067 extern int kernel_read(struct file
*, unsigned long, char *, unsigned long);
2068 extern struct file
* open_exec(const char *);
2070 /* fs/dcache.c -- generic fs support functions */
2071 extern int is_subdir(struct dentry
*, struct dentry
*);
2072 extern ino_t
find_inode_number(struct dentry
*, struct qstr
*);
2074 #include <linux/err.h>
2076 /* needed for stackable file system support */
2077 extern loff_t
default_llseek(struct file
*file
, loff_t offset
, int origin
);
2079 extern loff_t
vfs_llseek(struct file
*file
, loff_t offset
, int origin
);
2081 extern struct inode
* inode_init_always(struct super_block
*, struct inode
*);
2082 extern void inode_init_once(struct inode
*);
2083 extern void inode_add_to_lists(struct super_block
*, struct inode
*);
2084 extern void iput(struct inode
*);
2085 extern struct inode
* igrab(struct inode
*);
2086 extern ino_t
iunique(struct super_block
*, ino_t
);
2087 extern int inode_needs_sync(struct inode
*inode
);
2088 extern void generic_delete_inode(struct inode
*inode
);
2089 extern void generic_drop_inode(struct inode
*inode
);
2091 extern struct inode
*ilookup5_nowait(struct super_block
*sb
,
2092 unsigned long hashval
, int (*test
)(struct inode
*, void *),
2094 extern struct inode
*ilookup5(struct super_block
*sb
, unsigned long hashval
,
2095 int (*test
)(struct inode
*, void *), void *data
);
2096 extern struct inode
*ilookup(struct super_block
*sb
, unsigned long ino
);
2098 extern struct inode
* iget5_locked(struct super_block
*, unsigned long, int (*test
)(struct inode
*, void *), int (*set
)(struct inode
*, void *), void *);
2099 extern struct inode
* iget_locked(struct super_block
*, unsigned long);
2100 extern int insert_inode_locked4(struct inode
*, unsigned long, int (*test
)(struct inode
*, void *), void *);
2101 extern int insert_inode_locked(struct inode
*);
2102 extern void unlock_new_inode(struct inode
*);
2104 extern void __iget(struct inode
* inode
);
2105 extern void iget_failed(struct inode
*);
2106 extern void clear_inode(struct inode
*);
2107 extern void destroy_inode(struct inode
*);
2108 extern struct inode
*new_inode(struct super_block
*);
2109 extern int should_remove_suid(struct dentry
*);
2110 extern int file_remove_suid(struct file
*);
2112 extern void __insert_inode_hash(struct inode
*, unsigned long hashval
);
2113 extern void remove_inode_hash(struct inode
*);
2114 static inline void insert_inode_hash(struct inode
*inode
) {
2115 __insert_inode_hash(inode
, inode
->i_ino
);
2118 extern struct file
* get_empty_filp(void);
2119 extern void file_move(struct file
*f
, struct list_head
*list
);
2120 extern void file_kill(struct file
*f
);
2123 extern void submit_bio(int, struct bio
*);
2124 extern int bdev_read_only(struct block_device
*);
2126 extern int set_blocksize(struct block_device
*, int);
2127 extern int sb_set_blocksize(struct super_block
*, int);
2128 extern int sb_min_blocksize(struct super_block
*, int);
2129 extern int sb_has_dirty_inodes(struct super_block
*);
2131 extern int generic_file_mmap(struct file
*, struct vm_area_struct
*);
2132 extern int generic_file_readonly_mmap(struct file
*, struct vm_area_struct
*);
2133 extern int file_read_actor(read_descriptor_t
* desc
, struct page
*page
, unsigned long offset
, unsigned long size
);
2134 int generic_write_checks(struct file
*file
, loff_t
*pos
, size_t *count
, int isblk
);
2135 extern ssize_t
generic_file_aio_read(struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
2136 extern ssize_t
generic_file_aio_write(struct kiocb
*, const struct iovec
*, unsigned long, loff_t
);
2137 extern ssize_t
generic_file_aio_write_nolock(struct kiocb
*, const struct iovec
*,
2138 unsigned long, loff_t
);
2139 extern ssize_t
generic_file_direct_write(struct kiocb
*, const struct iovec
*,
2140 unsigned long *, loff_t
, loff_t
*, size_t, size_t);
2141 extern ssize_t
generic_file_buffered_write(struct kiocb
*, const struct iovec
*,
2142 unsigned long, loff_t
, loff_t
*, size_t, ssize_t
);
2143 extern ssize_t
do_sync_read(struct file
*filp
, char __user
*buf
, size_t len
, loff_t
*ppos
);
2144 extern ssize_t
do_sync_write(struct file
*filp
, const char __user
*buf
, size_t len
, loff_t
*ppos
);
2145 extern int generic_segment_checks(const struct iovec
*iov
,
2146 unsigned long *nr_segs
, size_t *count
, int access_flags
);
2149 extern ssize_t
generic_file_splice_read(struct file
*, loff_t
*,
2150 struct pipe_inode_info
*, size_t, unsigned int);
2151 extern ssize_t
generic_file_splice_write(struct pipe_inode_info
*,
2152 struct file
*, loff_t
*, size_t, unsigned int);
2153 extern ssize_t
generic_file_splice_write_nolock(struct pipe_inode_info
*,
2154 struct file
*, loff_t
*, size_t, unsigned int);
2155 extern ssize_t
generic_splice_sendpage(struct pipe_inode_info
*pipe
,
2156 struct file
*out
, loff_t
*, size_t len
, unsigned int flags
);
2157 extern long do_splice_direct(struct file
*in
, loff_t
*ppos
, struct file
*out
,
2158 size_t len
, unsigned int flags
);
2161 file_ra_state_init(struct file_ra_state
*ra
, struct address_space
*mapping
);
2162 extern loff_t
no_llseek(struct file
*file
, loff_t offset
, int origin
);
2163 extern loff_t
generic_file_llseek(struct file
*file
, loff_t offset
, int origin
);
2164 extern loff_t
generic_file_llseek_unlocked(struct file
*file
, loff_t offset
,
2166 extern int generic_file_open(struct inode
* inode
, struct file
* filp
);
2167 extern int nonseekable_open(struct inode
* inode
, struct file
* filp
);
2169 #ifdef CONFIG_FS_XIP
2170 extern ssize_t
xip_file_read(struct file
*filp
, char __user
*buf
, size_t len
,
2172 extern int xip_file_mmap(struct file
* file
, struct vm_area_struct
* vma
);
2173 extern ssize_t
xip_file_write(struct file
*filp
, const char __user
*buf
,
2174 size_t len
, loff_t
*ppos
);
2175 extern int xip_truncate_page(struct address_space
*mapping
, loff_t from
);
2177 static inline int xip_truncate_page(struct address_space
*mapping
, loff_t from
)
2184 ssize_t
__blockdev_direct_IO(int rw
, struct kiocb
*iocb
, struct inode
*inode
,
2185 struct block_device
*bdev
, const struct iovec
*iov
, loff_t offset
,
2186 unsigned long nr_segs
, get_block_t get_block
, dio_iodone_t end_io
,
2190 DIO_LOCKING
= 1, /* need locking between buffered and direct access */
2191 DIO_NO_LOCKING
, /* bdev; no locking at all between buffered/direct */
2192 DIO_OWN_LOCKING
, /* filesystem locks buffered and direct internally */
2195 static inline ssize_t
blockdev_direct_IO(int rw
, struct kiocb
*iocb
,
2196 struct inode
*inode
, struct block_device
*bdev
, const struct iovec
*iov
,
2197 loff_t offset
, unsigned long nr_segs
, get_block_t get_block
,
2198 dio_iodone_t end_io
)
2200 return __blockdev_direct_IO(rw
, iocb
, inode
, bdev
, iov
, offset
,
2201 nr_segs
, get_block
, end_io
, DIO_LOCKING
);
2204 static inline ssize_t
blockdev_direct_IO_no_locking(int rw
, struct kiocb
*iocb
,
2205 struct inode
*inode
, struct block_device
*bdev
, const struct iovec
*iov
,
2206 loff_t offset
, unsigned long nr_segs
, get_block_t get_block
,
2207 dio_iodone_t end_io
)
2209 return __blockdev_direct_IO(rw
, iocb
, inode
, bdev
, iov
, offset
,
2210 nr_segs
, get_block
, end_io
, DIO_NO_LOCKING
);
2213 static inline ssize_t
blockdev_direct_IO_own_locking(int rw
, struct kiocb
*iocb
,
2214 struct inode
*inode
, struct block_device
*bdev
, const struct iovec
*iov
,
2215 loff_t offset
, unsigned long nr_segs
, get_block_t get_block
,
2216 dio_iodone_t end_io
)
2218 return __blockdev_direct_IO(rw
, iocb
, inode
, bdev
, iov
, offset
,
2219 nr_segs
, get_block
, end_io
, DIO_OWN_LOCKING
);
2223 extern const struct file_operations generic_ro_fops
;
2225 #define special_file(m) (S_ISCHR(m)||S_ISBLK(m)||S_ISFIFO(m)||S_ISSOCK(m))
2227 extern int vfs_readlink(struct dentry
*, char __user
*, int, const char *);
2228 extern int vfs_follow_link(struct nameidata
*, const char *);
2229 extern int page_readlink(struct dentry
*, char __user
*, int);
2230 extern void *page_follow_link_light(struct dentry
*, struct nameidata
*);
2231 extern void page_put_link(struct dentry
*, struct nameidata
*, void *);
2232 extern int __page_symlink(struct inode
*inode
, const char *symname
, int len
,
2234 extern int page_symlink(struct inode
*inode
, const char *symname
, int len
);
2235 extern const struct inode_operations page_symlink_inode_operations
;
2236 extern int generic_readlink(struct dentry
*, char __user
*, int);
2237 extern void generic_fillattr(struct inode
*, struct kstat
*);
2238 extern int vfs_getattr(struct vfsmount
*, struct dentry
*, struct kstat
*);
2239 void inode_add_bytes(struct inode
*inode
, loff_t bytes
);
2240 void inode_sub_bytes(struct inode
*inode
, loff_t bytes
);
2241 loff_t
inode_get_bytes(struct inode
*inode
);
2242 void inode_set_bytes(struct inode
*inode
, loff_t bytes
);
2244 extern int vfs_readdir(struct file
*, filldir_t
, void *);
2246 extern int vfs_stat(char __user
*, struct kstat
*);
2247 extern int vfs_lstat(char __user
*, struct kstat
*);
2248 extern int vfs_stat_fd(int dfd
, char __user
*, struct kstat
*);
2249 extern int vfs_lstat_fd(int dfd
, char __user
*, struct kstat
*);
2250 extern int vfs_fstat(unsigned int, struct kstat
*);
2252 extern int do_vfs_ioctl(struct file
*filp
, unsigned int fd
, unsigned int cmd
,
2254 extern int __generic_block_fiemap(struct inode
*inode
,
2255 struct fiemap_extent_info
*fieinfo
, u64 start
,
2256 u64 len
, get_block_t
*get_block
);
2257 extern int generic_block_fiemap(struct inode
*inode
,
2258 struct fiemap_extent_info
*fieinfo
, u64 start
,
2259 u64 len
, get_block_t
*get_block
);
2261 extern void get_filesystem(struct file_system_type
*fs
);
2262 extern void put_filesystem(struct file_system_type
*fs
);
2263 extern struct file_system_type
*get_fs_type(const char *name
);
2264 extern struct super_block
*get_super(struct block_device
*);
2265 extern struct super_block
*user_get_super(dev_t
);
2266 extern void drop_super(struct super_block
*sb
);
2268 extern int dcache_dir_open(struct inode
*, struct file
*);
2269 extern int dcache_dir_close(struct inode
*, struct file
*);
2270 extern loff_t
dcache_dir_lseek(struct file
*, loff_t
, int);
2271 extern int dcache_readdir(struct file
*, void *, filldir_t
);
2272 extern int simple_getattr(struct vfsmount
*, struct dentry
*, struct kstat
*);
2273 extern int simple_statfs(struct dentry
*, struct kstatfs
*);
2274 extern int simple_link(struct dentry
*, struct inode
*, struct dentry
*);
2275 extern int simple_unlink(struct inode
*, struct dentry
*);
2276 extern int simple_rmdir(struct inode
*, struct dentry
*);
2277 extern int simple_rename(struct inode
*, struct dentry
*, struct inode
*, struct dentry
*);
2278 extern int simple_sync_file(struct file
*, struct dentry
*, int);
2279 extern int simple_empty(struct dentry
*);
2280 extern int simple_readpage(struct file
*file
, struct page
*page
);
2281 extern int simple_prepare_write(struct file
*file
, struct page
*page
,
2282 unsigned offset
, unsigned to
);
2283 extern int simple_write_begin(struct file
*file
, struct address_space
*mapping
,
2284 loff_t pos
, unsigned len
, unsigned flags
,
2285 struct page
**pagep
, void **fsdata
);
2286 extern int simple_write_end(struct file
*file
, struct address_space
*mapping
,
2287 loff_t pos
, unsigned len
, unsigned copied
,
2288 struct page
*page
, void *fsdata
);
2290 extern struct dentry
*simple_lookup(struct inode
*, struct dentry
*, struct nameidata
*);
2291 extern ssize_t
generic_read_dir(struct file
*, char __user
*, size_t, loff_t
*);
2292 extern const struct file_operations simple_dir_operations
;
2293 extern const struct inode_operations simple_dir_inode_operations
;
2294 struct tree_descr
{ char *name
; const struct file_operations
*ops
; int mode
; };
2295 struct dentry
*d_alloc_name(struct dentry
*, const char *);
2296 extern int simple_fill_super(struct super_block
*, int, struct tree_descr
*);
2297 extern int simple_pin_fs(struct file_system_type
*, struct vfsmount
**mount
, int *count
);
2298 extern void simple_release_fs(struct vfsmount
**mount
, int *count
);
2300 extern ssize_t
simple_read_from_buffer(void __user
*to
, size_t count
,
2301 loff_t
*ppos
, const void *from
, size_t available
);
2303 #ifdef CONFIG_MIGRATION
2304 extern int buffer_migrate_page(struct address_space
*,
2305 struct page
*, struct page
*);
2307 #define buffer_migrate_page NULL
2310 extern int inode_change_ok(struct inode
*, struct iattr
*);
2311 extern int __must_check
inode_setattr(struct inode
*, struct iattr
*);
2313 extern void file_update_time(struct file
*file
);
2315 extern int generic_show_options(struct seq_file
*m
, struct vfsmount
*mnt
);
2316 extern void save_mount_options(struct super_block
*sb
, char *options
);
2318 static inline ino_t
parent_ino(struct dentry
*dentry
)
2322 spin_lock(&dentry
->d_lock
);
2323 res
= dentry
->d_parent
->d_inode
->i_ino
;
2324 spin_unlock(&dentry
->d_lock
);
2328 /* Transaction based IO helpers */
2331 * An argresp is stored in an allocated page and holds the
2332 * size of the argument or response, along with its content
2334 struct simple_transaction_argresp
{
2339 #define SIMPLE_TRANSACTION_LIMIT (PAGE_SIZE - sizeof(struct simple_transaction_argresp))
2341 char *simple_transaction_get(struct file
*file
, const char __user
*buf
,
2343 ssize_t
simple_transaction_read(struct file
*file
, char __user
*buf
,
2344 size_t size
, loff_t
*pos
);
2345 int simple_transaction_release(struct inode
*inode
, struct file
*file
);
2347 void simple_transaction_set(struct file
*file
, size_t n
);
2350 * simple attribute files
2352 * These attributes behave similar to those in sysfs:
2354 * Writing to an attribute immediately sets a value, an open file can be
2355 * written to multiple times.
2357 * Reading from an attribute creates a buffer from the value that might get
2358 * read with multiple read calls. When the attribute has been read
2359 * completely, no further read calls are possible until the file is opened
2362 * All attributes contain a text representation of a numeric value
2363 * that are accessed with the get() and set() functions.
2365 #define DEFINE_SIMPLE_ATTRIBUTE(__fops, __get, __set, __fmt) \
2366 static int __fops ## _open(struct inode *inode, struct file *file) \
2368 __simple_attr_check_format(__fmt, 0ull); \
2369 return simple_attr_open(inode, file, __get, __set, __fmt); \
2371 static struct file_operations __fops = { \
2372 .owner = THIS_MODULE, \
2373 .open = __fops ## _open, \
2374 .release = simple_attr_release, \
2375 .read = simple_attr_read, \
2376 .write = simple_attr_write, \
2379 static inline void __attribute__((format(printf
, 1, 2)))
2380 __simple_attr_check_format(const char *fmt
, ...)
2382 /* don't do anything, just let the compiler check the arguments; */
2385 int simple_attr_open(struct inode
*inode
, struct file
*file
,
2386 int (*get
)(void *, u64
*), int (*set
)(void *, u64
),
2388 int simple_attr_release(struct inode
*inode
, struct file
*file
);
2389 ssize_t
simple_attr_read(struct file
*file
, char __user
*buf
,
2390 size_t len
, loff_t
*ppos
);
2391 ssize_t
simple_attr_write(struct file
*file
, const char __user
*buf
,
2392 size_t len
, loff_t
*ppos
);
2395 int proc_nr_files(struct ctl_table
*table
, int write
, struct file
*filp
,
2396 void __user
*buffer
, size_t *lenp
, loff_t
*ppos
);
2398 int get_filesystem_list(char * buf
);
2400 #endif /* __KERNEL__ */
2401 #endif /* _LINUX_FS_H */